Re: BBB 주요 세포 pericyte 병리 2024 Cell

[문형철]

ReviewVolume 34, Issue 1p58-71January 2024Open access

Download Full Issue

Pericytes in the disease spotlight

Hielke van Splunder1,5 ∙ Pilar Villacampa2,5 ∙ Anabel Martínez-Romero1 ∙ Mariona Graupera1,3,4 mgraupera@carrerasresearch.org

Affiliations & NotesArticle Info

• Download PDF
• Cite

• Share

• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Highlights

Molecular and functional pericyte studies at single-cell resolution are providing new insights into long-standing questions about pericyte heterogeneity.

Pericytes are not identified by a single marker but instead by gene expression‎ signatures that show substantial inter-organ differences.

Pericytes orchestrate and precede endothelial cell responses during angiogenesis.

Pericyte degeneration and dysfunction, that are triggered by the onset of some diseases, contribute to the progression of those diseases in both vascular and non-vascular contexts.

The number of diseases with pericyte dysfunction continues to expand, thereby anticipating a promising future for pericyte-focused therapy.

하이라이트

단일 세포 해상도에서의 분자적 및 기능적 페리사이트 연구는

페리사이트 이질성에 관한 오랜 의문에 대한 새로운 통찰력을 제공하고 있습니다.

페리사이트는 단일 마커로 식별되는 것이 아니라,

장기 간 상당한 차이를 보이는 유전자 발현 시그니처로 식별됩니다.

페리사이트는

혈관 신생 과정에서 내피 세포 반응을 조율하고 선행한다.

일부 질환의 발병으로 유발되는

페리사이트 퇴행 및 기능 장애는

혈관 및 비혈관 환경 모두에서 해당 질환의 진행에 기여한다.

페리사이트 기능 장애가 관여하는 질환의 범위는

지속적으로 확대되고 있어,

페리사이트 중심 치료법의 유망한 미래를 예고하고 있다.

Abstract

Pericytes are known as the mural cells in small-caliber vessels that interact closely with the endothelium. Pericytes play a key role in vasculature formation and homeostasis, and when dysfunctional contribute to vasculature-related diseases such as diabetic retinopathy and neurodegenerative conditions. In addition, significant extravascular roles of pathological pericytes are being discovered with relevant implications for cancer and fibrosis. Pericyte research is challenged by the lack of consistent molecular markers and clear discrimination criteria versus other (mural) cells. However, advances in single-cell approaches are uncovering and clarifying mural cell identities, biological functions, and ontogeny across organs. We discuss the latest developments in pericyte pathobiology to inform future research directions and potential outcomes.

초록

페리사이트는

작은 구경 혈관 내벽 세포로 알려져 있으며

내피세포와 밀접하게 상호작용한다.

페리사이트는

혈관 형성 및 항상성 유지에 핵심적인 역할을 수행하며,

기능 장애 시 당뇨병성 망막병증 및 신경퇴행성 질환과 같은

혈관 관련 질환의 발병에 기여한다.

또한

병리학적 페리사이트의 중요한 혈관 외 역할이 발견되며,

이는 암 및 섬유화와 관련이 있습니다.

페리사이트 연구는

일관된 분자 표지자의 부재와 다른 (벽세포)와의 명확한 구분 기준 부족으로 어려움을 겪고 있습니다.

그러나

단일 세포 접근법의 발전으로 장기 전반에 걸친 벽세포의 정체성,

생물학적 기능 및 발생학적 특성이 밝혀지고 있습니다.

향후 연구 방향과 잠재적 결과를 제시하기 위해 페리사이트 병리생물학의 최신 발전 사항을 논의합니다.

키워드

Keywords

1. pericyte
2. vascular disease
3. pericrine signaling
4. fibrosis

Multifaceted roles of mural cells in health and disease

Pericytes are classically defined as mural cells (see Glossary) that envelop the endothelium of small caliber blood vessels, the so-called capillaries. Pericytes are embedded within the same basement membrane as endothelial cells (ECs) and interact closely with them [1,2]. By contrast, vascular smooth muscle cells (vSMCs), the other mural cell type, cover large arteries and veins, and are physically separated from the endothelium by an intimal layer of extracellular matrix (ECM). Of note, lymphatic capillaries lack pericytes under physiological conditions, although collecting lymphatic vessels contain vSMCs [3].

A fundamental function of mural cells is to regulate the stabilization and function of blood vessels. It is therefore not surprising that pericyte loss and dysfunction were linked to several diseases including cancer and cerebrovascular diseases more than a decade ago [4,5]. However, pericyte-focused therapies have been poorly explored. Instead, most studies on vascular-directed therapeutic strategies have been on ECs – the central components that build blood vessels. Emerging data are, however, changing the perception of pericytes from mere supporting vascular cells that are recruited at the final stage of vessel formation to essential elements in the early phases of angiogenesis that anticipate and orchestrate EC behavior. In addition, recent research is revealing novel pathological roles for pericytes beyond their implications in the vasculature. Collectively, we believe that these data open exciting avenues for pericyte-focused therapeutic approaches and call for a broader understanding of these cells in disease progression.

We provide here a global overview of recent significant advances regarding our understanding of the role of pericytes in different pathobiological scenarios and discuss the field's current paradigms and controversies. First, we address new insights into the functions associated with pericytes during physiological vascular responses. Second, we discuss evidence supporting a role of pericytes in disease, including pericyte cell-autonomous implications beyond the vasculature. For comprehensive details on pericyte biology, function ontology, and specific signaling pathways, we refer the reader to [1,2,5]. Of importance, some of the emerging concepts in pericyte biology described in the following sections have only been studied in one specific tissue. To avoid confusion about the generalizability of pericyte properties, we frame each function by considering the relevant organ of study.

건강과 질병에서 벽세포의 다면적 역할

페리사이트는 전통적으로 소형 혈관(모세혈관)의 내피를 둘러싸는

벽세포(용어집 참조)로 정의됩니다.

페리사이트는

내피세포(ECs)와 동일한 기저막 내에 위치하며

밀접하게 상호작용합니다[1,2].

https://www.mdpi.com/1422-0067/25/12/6592

Signaling Role of Pericytes in Vascular Health and Tissue Homeostasis

반면, 다른 벽세포 유형인 혈관 평활근 세포(vSMCs)는

대동맥과 대정맥을 덮으며,

내피세포로부터 세포외기질(ECM)로 이루어진 내막층에 의해 물리적으로 분리되어 있습니다.

주목할 점은,

림프관 모세혈관은 생리적 조건에서 페리사이트가 없지만,

집합 림프관은 vSMCs를 포함한다는 것입니다[3].

혈관벽 세포의 근본적 기능은

혈관의 안정화와 기능을 조절하는 것이다.

따라서 10여 년 전부터

암 및 뇌혈관 질환을 포함한 여러 질환과

페리사이트 손실 및 기능 장애가 연관되었다는 사실은 놀랍지 않다[4,5].

그러나

페리사이트 중심 치료법은 거의 연구되지 않았다.

대신 혈관 형성 치료 전략에 관한 대부분의 연구는

혈관을 구성하는 핵심 요소인 내피세포(ECs)에 집중되었다.

그러나 새롭게 등장하는 데이터는

혈관형성의 최종 단계에서 동원되는 단순한 보조 혈관 세포라는 기존 인식에서 벗어나,

혈관신생 초기 단계에서 EC의 행동을 예측하고 조율하는 필수 요소로서의

주변세포의 역할을 제시하고 있다.

또한 최근 연구는

혈관계에서의 역할 외에도 주변세포의 새로운 병리학적 기능을 밝혀내고 있다.

종합적으로,

이러한 데이터는 주변세포 중심 치료 접근법의 흥미로운 가능성을 열어주며,

질병 진행 과정에서 이 세포들에 대한 보다 포괄적인 이해를 요구한다고 믿는다.

본고에서는

다양한 병리생물학적 시나리오에서 페리사이트의 역할에 대한 최근 주요 진전을 종합적으로 개관하고,

해당 분야의 현재 패러다임과 논쟁점을 논의한다.

첫째,

생리적 혈관 반응 과정에서 페리사이트와 연관된 기능에 대한 새로운 통찰을 다룬다.

둘째,

혈관 외 영역에서의 페리사이트 세포 자율적 함의를 포함하여

질병에서 페리사이트의 역할을 뒷받침하는 증거를 논의한다.

페리사이트 생물학, 기능 온톨로지 및 특정 신호 전달 경로에 대한 포괄적인 세부 사항은 [1,2,5]를 참조하십시오. 중요한 점은, 다음 섹션에서 설명하는 페리사이트 생물학의 일부 새로운 개념들은 특정 조직에서만 연구되었다는 것입니다. 페리사이트 특성의 일반화에 대한 혼란을 피하기 위해, 우리는 각 기능을 연구 대상 관련 장기를 고려하여 설명합니다.

Figure 1 Schematic representation of mural cell zonation in the adult mouse brain.

Show full captionFigure viewer

Key concepts about pericytes in physiologyPericytes: a particular subtype of mural cells

Pericytes exhibit significant inter- and intra-tissue molecular differences and exert tissue-specific functions [2]. Their molecular, morphological, and functional heterogeneity is inextricably linked to their diverse developmental origins, modes of vessel recruitment, and specific anatomical localization. For example, pericytes of the central nervous system (CNS) microvasculature are firmly and continuously invested around the endothelium to support vascular barrier properties, whereas liver pericytes, commonly referred to as hepatic stellate cells (HSCs), reside in the perisinusoidal space, are loosely and discontinuously associated to ECs, and hold a unique vitamin A storage capacity [2]. To meet tissue-specific demands, pericyte distribution and density are variable among organs and vascular beds, with the CNS microvasculature showing the greatest pericyte-to-EC abundance. From a molecular standpoint there is no single molecular marker that can exclusively identify pericytes (Box 1), albeit the emergence of single-cell techniques is shedding light on tissue-specific pericyte molecular markers and functions. For example, the first molecular atlas of vascular cell types in the brain of adult mice by single-cell RNA sequencing (scRNA-seq) revealed that mural cells follow a gradient of transitional phenotypes. This gradient occurs at the interface of precapillary arterioles, capillaries, and postcapillary venules, and does not follow a single continuum along the arteriovenous axis (Figure 1 and Box 1) [6]. Whether this gradient of transitional phenotypes is specifically restricted to the brain vasculature or is also present in other vascular beds remains to be determined. Indeed, pericytes exhibit many organotypic differences in the expression‎ of molecular markers (Figure 2 illustrates three top-ranked pericyte markers with enriched expression‎ per organ), of which the expression‎ of transporters and components of the contractile machinery exhibit the greatest differences between organs [7]. Another intriguing observation is that pericytes exhibit more cross-organ heterogeneity than vSMCs [7,8]. Currently, the inter-tissue differences in the behavior of the two main mural cell types are not completely understood. However, this may be because pericytes exhibit a greater cell-intrinsic plasticity to adapt their molecular portfolio and function to tissue-specific demands, whereas vSMCs fulfill a more universal function across tissues. In contrast to the tissue-specific transcriptomic differences, the expression‎ of transcription factors appears to be relatively conserved in mural cells across organs, thereby suggesting that mural cell subtypes are defined by epigenetic mechanisms [7]. Accordingly, DNA hypermethylation was recently found to control alpha smooth muscle actin (αSMA) expression‎ in renal mural cells after ischemia [9]. This indicates that methods such as assay for transposase-accessible chromatin sequencing (ATAC-seq) will be instrumental to further understand mural cell phenotypes.

생리학에서 페리사이트에 관한 핵심 개념페리사이트: 벽세포의 특정 하위 유형

페리사이트는

조직 간 및 조직 내 분자적 차이를 현저히 보이며

조직 특이적 기능을 발휘한다 [2].

그들의 분자적, 형태학적, 기능적 이질성은

다양한 발생 기원, 혈관 유입 방식, 특정 해부학적 위치와 불가분의 관계를 맺고 있다.

예를 들어,

중추신경계(CNS) 미세혈관의 페리사이트는

혈관 장벽 특성을 지원하기 위해 내피 세포 주위를 단단하고 지속적으로 둘러싸고 있는 반면,

간 페리사이트(일반적으로 간성상세포(HSCs)로 불림)는 간소혈관 주위 공간에 위치하며,

내피 세포(ECs)와 느슨하고 불연속적으로 연관되어 있으며,

독특한 비타민 A 저장 능력을 보유합니다[2].

조직별 요구를 충족시키기 위해,

주변세포의 분포와 밀도는 장기 및 혈관층마다 다양하며,

중추신경계 미세혈관계에서 주변세포 대 내피세포 비율이 가장 높습니다.

분자적 관점에서 볼 때,

단일 분자 마커로만 주변세포를 식별할 수는 없지만(박스 1),

단일 세포 기술의 출현으로 조직 특이적 주변세포 분자 마커와 기능에 대한 이해가 깊어지고 있습니다.

예를 들어, 단일 세포 RNA 시퀀싱(scRNA-seq)을 통한 성인 마우스 뇌의 혈관 세포 유형에 대한 최초의 분자 아틀라스는 벽세포가 과도기적 표현형의 경사를 따르는 것을 밝혀냈습니다. 이 경사는 전모세관 동맥, 모세관, 후모세관 정맥의 경계면에서 발생하며, 동정맥 축을 따라 단일 연속체를 따르지 않는다(그림 1 및 박스 1) [6]. 이러한 과도형 표현형의 경사가 뇌 혈관에 특이적으로 제한되는지, 아니면 다른 혈관층에서도 존재하는지는 아직 밝혀지지 않았다. 실제로, 페리사이트는 분자 마커 발현에 있어 많은 장기 특이적 차이를 보이며(그림 2는 장기별로 풍부하게 발현되는 상위 3개 페리사이트 마커를 보여줌), 그중 수송체와 수축 기구의 구성 요소 발현이 장기 간 가장 큰 차이를 보인다[7].

또 다른 흥미로운 관찰 결과는

페리사이트가 혈관 평활근 세포(vSMCs)보다 장기 간 이질성이 더 크다는 점이다[7,8].

현재 두 주요 벽세포 유형의 행동에 대한 조직 간 차이는 완전히 이해되지 않았다.

그러나

이는 페리사이트가 분자 구성과 기능을 조직 특이적 요구에 적응시키기 위해

더 큰 세포 내재적 가소성을 보이는 반면,

vSMCs는 조직 전반에 걸쳐 보다 보편적인 기능을 수행하기 때문일 수 있다.

조직 특이적 전사체 차이와 대조적으로, 전사 인자의 발현은

장기 전반에 걸쳐 벽세포에서 상대적으로 보존되어 있는 것으로 보이며,

이는 벽세포 하위 유형이 후성유전학적 메커니즘에 의해 정의된다는 것을 시사한다 [7].

이에 따라, 최근 허혈 후 신장 벽세포에서 알파 평활근 액틴(αSMA) 발현을 제어하는 데 DNA 과메틸화가 관여한다는 사실이 밝혀졌다 [9]. 이는 전좌효소 접근성 염색질 시퀀싱(ATAC-seq)과 같은 방법이 벽세포 표현형을 더 깊이 이해하는 데 유용할 것임을 시사한다.

Figure 2 Organotypic heterogeneity of pericyte markers.

Show full captionFigure viewer

Box 1

Unraveling the identity of pericytes

The identification of pericytes remains a challenging task. Despite ongoing efforts, there is no consensus regarding unambiguous criteria for pericyte identification. To date no single molecular marker can exclusively identify all pericytes or distinguish pericytes from other cell types, although scRNA-seq is now providing new opportunities to discern pericyte marker heterogeneity and tissue specificity [6,8,71,93]. The use of transgenic reporter mouse models has been instrumental to label, trace, and locate different mural cell populations in vivo. A combination of multiple reporter lines is often necessary to properly identify and discriminate pericytes from endothelial cells (ECs) and other perivascular cells [6–8]. Mural cells are highly plastic cells; phenotypic zonation of mouse brain mural cells has revealed that these cells do not follow a single continuum along the arteriovenous axis (see Figure 1A,B in main text) [6]. From a transcriptional point of view, there are two distinct continuums of mural cells: (i) capillary pericytes and venous smooth muscle cells (SMCs), where pericytes gradually transition to a venous SMC phenotype, and (ii) arterial SMCs which transition in an distinct pattern towards arteriole SMCs. The transcriptional resemblance between mouse brain pericytes and venular mural cells [6], as well as the lack of classic pericytes in several organs [7,8], have led to the hypothesis that capillary pericytes are transcriptionally and morphologically similar to venous SMCs in some tissues. Human brain mural cells recapitulate the mouse zonation pattern, although human pericytes are evenly distributed over capillaries and veins [50,94]. Unlike the anatomical separation of pericytes and venous SMCs in the mouse brain, subtypes of human pericytes are discerned by functionality marked by solute transport and extracellular matrix (ECM) organization [50]. Unfortunately, the ability of mouse markers to predict the presence of human pericytes remains limited, and only a select few retain adequate specificity. The use of zebrafish models may provide a better alternative to study conserved pericyte genes [95]. We believe that RGS5, NDUFA4L2, KNCJ8, HIGD1B, ABCC9, NOTCH3, and PDGFRB are currently the most organ and species conserved pericyte markers, although detailed intra-tissue characterization remains necessary when studying pericytes (see Figure 2 in main text).

페리사이트 정체성 규명

페리사이트의 식별은

여전히 어려운 과제이다.

지속적인 노력에도 불구하고, 페리사이트 식별을 위한 명확한 기준에 대한 합의는 없다.

현재까지 단일 분자 표지자가

모든 페리사이트를 독점적으로 식별하거나 페리사이트를 다른 세포 유형과 구별할 수 없었으나,

scRNA-seq은 이제 페리사이트 표지자의 이질성과 조직 특이성을 식별하는 새로운 기회를 제공하고 있다 [6,8,71,93].

트랜스제닉 리포터 마우스 모델의 사용은 생체 내에서 다양한 벽세포 집단을 표지하고 추적하며 위치를 파악하는 데 중요한 역할을 해왔다. 내피세포(ECs) 및 기타 혈관 주위 세포와 페리사이트를 적절히 식별하고 구분하기 위해서는 종종 여러 리포터 계통의 조합이 필요하다[6–8]. 벽세포는 높은 가소성을 지닌 세포이다; 마우스 뇌 벽세포의 표현형 구획화는 이 세포들이 동정맥 축을 따라 단일 연속체를 따르지 않음을 밝혀냈다(본문 그림 1A,B 참조)[6].

전사적 관점에서 벽세포는 두 가지 뚜렷한 연속체를 이룬다:

(i) 모세혈관 페리사이트와 정맥 평활근세포(SMCs)로, 페리사이트가 점진적으로 정맥 SMC 표현형으로 전환되는 경우, 그리고

(ii) 동맥 SMCs로, 이들은 사구체 SMCs로 향하는 독특한 패턴으로 전환된다.

마우스 뇌 페리사이트와 소정맥 벽 세포 간의 전사적 유사성[6], 그리고 여러 장기에서 전형적인 페리사이트의 부재[7,8]는 일부 조직에서 모세혈관 페리사이트가 전사적·형태학적으로 정맥 SMC와 유사하다는 가설을 낳았다. 인간 뇌 벽 세포는 마우스의 구획화 패턴을 재현하지만, 인간 페리사이트는 모세혈관과 정맥에 고르게 분포한다[50,94] . 생쥐 뇌에서 관세포와 정맥 평활근 세포의 해부학적 분리와는 달리, 인간 관세포의 하위 유형은 용질 수송 및 세포외 기질(ECM) 조직화를 특징으로 하는 기능성에 의해 구분된다[50]. 안타깝게도 생쥐 표지자가 인간 관세포의 존재를 예측하는 능력은 여전히 제한적이며, 극소수만이 충분한 특이성을 유지한다. 제브라피시 모델을 활용하면 보존된 페리사이트 유전자 연구에 더 나은 대안이 될 수 있다[95]. 우리는 RGS5, NDUFA4L2, KNCJ8, HIGD1B, ABCC9, NOTCH3, PDGFRB가 현재 가장 기관 및 종 보존성이 높은 페리사이트 마커라고 믿지만, 페리사이트 연구 시 조직 내 상세한 특성화는 여전히 필요하다(본문 그림 2 참조).

Pericytes at play during vascular growth

Many studies have documented that pericytes contribute to angiogenesis [10]. The historical view proposes that pericytes mainly contribute to the late stages of vessel formation [2,10]. By taking advantage of the mouse retina as a paradigmatic experimental model of developmental angiogenesis, this concept has been challenged [11–16]. Indeed, these studies showed that, during the early phases of developmental angiogenesis, pericytes, which have not yet achieved the maturity seen in formed vessels, are permissive to cell-cycle progression, morphological adaptation, and migration [12,13]. In this setting, pericyte growth precedes the expansion of ECs, although it is still unclear why. One possibility is that, by expanding rapidly, pericytes ensure the production of sufficient EC growth signals, a hypothesis which is coherent with the observation that inhibition of pericyte activation blocks EC proliferation [12] and induces nuclear translocation of FOXO1 [11], the master regulator of EC quiescence [17]. Another study that examined the brain vasculature showed that, when pericytes are absent, ECs become angiogenic but are not able to proliferate [18], thereby supporting a model in which ECs require the presence of pericytes to expand. Nonetheless, it is fair to acknowledge that other studies have shown that reduced pericyte coverage leads to increased EC proliferation [19]. Although these discrepancies highlight that pericyte–EC interactions are complex, they may be explained by the differences between the animal models and genetic strategies used to interfere with pericytes. Importantly, pericyte behaviors during angiogenesis have been mostly described in tissues belonging to the CNS. Hence, given the high abundance of pericytes in the CNS, it is possible that angiogenic pericytes fulfill different roles in tissues where ECs substantially outnumber them. Another interesting observation is that, during angiogenesis, immature pericytes remain in close contact with ECs, although they do not cover them in their entirety [12,20]. This suggests that pericyte–EC communication during angiogenesis relies on both paracrine and juxtracrine signaling, and may explain why pericyte loss [11,16,21,22] and impaired transition to a fully maturate state [12] lead to distinct endothelial phenotypes during angiogenesis. scRNA-seq analysis of prenatal developing human brains confirmed that angiogenesis is supported by immature mural cells [20]. Consistent with mouse data [12], the state of mature human pericytes correlates with the progression of angiogenesis. Furthermore, the gene expression‎ profiles of these cells show involvement in processes related to the transport across the blood–brain barrier (BBB) and the synthesis of ECM components [20]. Together, these data support a model in which pericytes modulate the early phases of angiogenesis by directly regulating EC behavior. Intriguingly, however, detailed ultrastructural analysis of angiogenic vessels in human brain distinguishes only a single mural cell population, compared to three distinct EC populations [20]. This suggests that ultrastructural features do not define subtype specification in the mural cell compartment, and that molecular and structural features are not necessarily associated with each other.

혈관 성장 과정에서 작용하는 페리사이트

많은 연구에서 페리사이트가 혈관신생에 기여한다는 사실이 입증되었습니다[10]. 기존 관점은 페리사이트가 주로 혈관 형성의 후기 단계에 기여한다고 제안합니다[2,10]. 발달기 혈관신생의 대표적인 실험 모델인 마우스 망막을 활용함으로써, 이 개념은 도전받았습니다[11–16]. 실제로 이 연구들은 발달 혈관신생 초기 단계에서, 아직 형성된 혈관에서 관찰되는 성숙도에 이르지 않은 페리사이트가 세포주기 진행, 형태학적 적응 및 이동에 관여할 수 있음을 보여주었다[12,13]. 이러한 환경에서, 그 이유는 아직 명확하지 않지만, 페리사이트의 성장은 혈관 내피세포(ECs)의 확장보다 선행한다. 한 가지 가능성은 페리사이트가 빠르게 확장함으로써 충분한 EC 성장 신호의 생산을 보장한다는 가설이다. 이는 페리사이트 활성화 억제가 EC 증식을 차단한다는 관찰[12]과 EC 휴지 상태의 주요 조절인자인 FOXO1의 핵 이동을 유도한다는 관찰[11]과 일관성을 보인다[17]. 뇌 혈관을 조사한 또 다른 연구에서는 페리사이트가 결핍될 경우 내피세포가 혈관신생 능력을 획득하지만 증식할 수 없음을 보여주었다[18]. 이는 내피세포가 확장하기 위해 페리사이트의 존재가 필요하다는 모델을 지지한다. 그럼에도 불구하고, 페리사이트의 덮임이 감소하면 내피세포 증식이 증가한다는 다른 연구 결과도 존재함을 인정하는 것이 타당하다[19]. 이러한 상이점은 페리사이트-내피세포 상호작용이 복잡함을 부각시키지만, 페리사이트를 방해하기 위해 사용된 동물 모델과 유전적 전략의 차이로 설명될 수 있다. 중요한 점은 혈관신생 과정 중 페리사이트의 행동이 주로 중추신경계(CNS) 조직에서 기술되었다는 것이다. 따라서 CNS 내 페리사이트의 높은 풍부도를 고려할 때, 혈관신생성 페리사이트는 내피세포가 그 수를 현저히 초과하는 조직에서 다른 역할을 수행할 가능성이 있다. 또 다른 흥미로운 관찰 결과는 혈관신생 과정에서 미성숙 페리사이트가 EC를 완전히 덮지는 않더라도 EC와 밀접하게 접촉을 유지한다는 점이다 [12,20]. 이는 혈관신생 과정에서의 페리사이트-내피세포 간 소통이 파라크라인 및 쥬스트라크라인 신호전달에 의존함을 시사하며, 페리사이트 손실[11,16,21,22] 및 완전 성숙 상태로의 전환 장애[12]가 혈관신생 중 특이적인 내피세포 표현형을 유발하는 이유를 설명할 수 있다. 태아기 발달 중인 인간 뇌에 대한 scRNA-seq 분석은 혈관신생이 미성숙 벽세포에 의해 지원됨을 확인하였다[20].

마우스 데이터[12]와 일관되게,

성숙한 인간 페리사이트의 상태는

혈관신생 진행과 상관관계를 보인다.

또한 이 세포들의 유전자 발현 프로파일은

혈뇌장벽(BBB)을 통한 수송 및 ECM 구성 요소 합성과 관련된 과정에

관여함을 보여준다[20].

종합하면, 이러한 데이터는

페리사이트가 EC 행동을 직접 조절함으로써

혈관신생 초기 단계를 조절한다는 모델을 지지한다.

그러나 흥미롭게도, 인간 뇌의 혈관신생 혈관에 대한 상세한 초미세구조 분석은 세 가지 별개의 EC 집단[20]과 달리 단일 벽세포 집단만을 구분한다. 이는 초미세구조적 특징이 벽세포 구획 내 하위 유형 특이화를 정의하지 않으며, 분자적 및 구조적 특징이 반드시 서로 연관되지 않음을 시사한다.

Brain pericytes and vessel contraction: a matter of transitional phenotypes

Although the regulation of vascular tone through pericyte contractility is considered to be an important function of cardiac, renal, and pulmonary pericytes, as well as of HSCs [2], there has been a long-standing debate in the field as to whether pericytes actively modulate cerebral blood flow [23–25]. For instance, by using optical imaging, Hill et al. suggested that neural/glia antigen 2-positive (NG2+) αSMA− pericytes are not contractile and do not actively modulate the capillary diameter [26]. Instead, by similar optogenetic approaches Hartmann et al. proposed that pericytes do constrict, although they require prolonged and more intense stimulation than αSMA+ mural cells located at larger vessels [27]. Although no consensus has been established, the opposing results between studies may simply reflect heterogeneities in the type of blood vessels and mural cells analyzed. A recent report has shown that NG2+αSMA+ mural cells, located at the transitional segment between arteries and capillaries, regulate the vascular tone and contractility [28]. This suggests that the transition of functional phenotypes between mural cells covering distinct types of blood vessels is tightly regulated.

scRNA-seq analysis of brain mural cells has revealed an abrupt change in the molecular signatures of pericytes and mural cells located in arteries, even from cells residing in proximity on the vasculature, thereby supporting the existence of a blunt transition [6]. Taken together, one can speculate that, in addition to defined vSMC types, there is a subtype of mural cells that exhibit some traits, but not all, of classic pericytes, and are located at transitional vessels and can modulate the vascular tone. Given the ability of pericytes to adapt their phenotype to various microenvironmental conditions [1,2], it is also possible that regulation of blood flow may only occur under specific circumstances. However, one should consider that some of the data disputing pericyte contractility may relate to experimental artefacts, and it should be stressed that most analyses were conducted in the cerebral vasculature as a prototypical example of a vascular bed that is highly sensitive to contraction [25]. An important observation is that pericytes exhibit significant organotypic differences in the basal expression‎ of contractility genes, and pericytes in the bladder and colon express considerable levels of Myh11, Tagln, and Acta2 (αSMA), whereas pericytes in the brain, lung, and heart express negligible amounts of these contractile genes [6,7]. This highlights a conundrum regarding how brain pericytes regulate vessel contractility when typical contractility genes are not expressed.

뇌 페리사이트와 혈관 수축: 과도기적 표현형의 문제

페리사이트 수축성을 통한 혈관 긴장도 조절은 심장, 신장, 폐 페리사이트 및 HSC의 중요한 기능으로 간주되지만[2], 페리사이트가 뇌혈류를 능동적으로 조절하는지에 대해서는 해당 분야에서 오랜 논쟁이 지속되어 왔다[23–25]. 예를 들어, Hill 등은 광학 영상법을 활용하여 신경/교세포 항원 2 양성(NG2+) αSMA− 페리사이트는 수축 능력이 없으며 모세혈관 직경을 능동적으로 조절하지 않는다고 제안했다[26]. 반면 Hartmann 등은 유사한 광유전학적 접근법을 통해 페리사이트가 수축한다는 가설을 제시했으나, 이는 대형 혈관에 위치한 αSMA+ 벽세포보다 더 길고 강도 높은 자극이 필요하다고 밝혔다[27]. 아직 합의가 이루어지진 않았지만, 연구 간 상반된 결과는 단순히 분석된 혈관 유형과 벽세포의 이질성을 반영할 수 있다. 최근 보고에 따르면 동맥과 모세혈관 사이의 과도 구간에 위치한 NG2+αSMA+ 벽세포가 혈관 긴장도와 수축성을 조절한다[28]. 이는 서로 다른 유형의 혈관을 덮는 벽세포 간 기능적 표현형의 전환이 엄격히 조절됨을 시사한다.

뇌 벽세포에 대한 scRNA-seq 분석은 동맥에 위치한 페리사이트와 벽세포의 분자적 특징이 급격히 변화함을 밝혀냈으며, 이는 혈관 구조상 근접한 위치에 있는 세포들 사이에서도 관찰되어 급격한 전환의 존재를 뒷받침한다 [6]의 존재를 뒷받침한다.

종합하면,

정의된 혈관 평활근 세포 유형 외에도,

고전적 페리사이트의 일부 특성은 나타내지만 전부는 아닌 하위 유형의 벽세포가 존재하며,

이들은 과도기 혈관에 위치하여 혈관 긴장도를 조절할 수 있다고 추측할 수 있다.

페리사이트가 다양한 미세환경 조건에 따라 표현형을 적응시킬 수 있는 능력[1,2]을 고려할 때,

혈류 조절은 특정 상황에서만 발생할 가능성도 있다.

그러나

주변세포 수축성을 반박하는 일부 데이터는

실험적 오류와 관련될 수 있음을 고려해야 하며,

대부분의 분석이 수축에 매우 민감한 혈관층의 전형적 사례인 뇌혈관계에서 수행되었다는 점을

강조해야 한다[25].

중요한 관찰 결과는 페리사이트가 수축성 유전자의 기초 발현에서 상당한 장기 특이적 차이를 보이며, 방광과 결장의 페리사이트는 상당한 수준의 Myh11, Tagln 및 Acta2 (αSMA)를 발현하는 반면, 뇌, 폐 및 심장의 페리사이트는 이러한 수축성 유전자를 미미한 수준만 발현한다는 점이다 [6,7]. 이는 전형적인 수축성 유전자가 발현되지 않을 때 뇌 페리사이트가 어떻게 혈관 수축성을 조절하는지에 대한 난제를 부각시킨다.

Pericyte safeguarding the capillary brain bed by a special touch

An essential function of pericytes is to regulate the BBB by controlling the passage of fluid and substances into the parenchymal space [22,29]. Hence, defective pericyte coverage caused by pericyte dysfunction, impaired pericyte recruitment, and pericyte loss all lead to increased EC transcytosis and permeability [22,29]. Aberrant platelet-derived growth factor B (PDGF-B)/platelet-derived growth factor receptor beta (PDGFRβ) signaling is sufficient to experimentally reduce pericyte abundance and the subsequent loss of BBB properties [22,29]. In addition, proper ECM deposition by pericytes (among other cell types composing the neurovascular unit) plays an essential role in maintaining the integrity of the vascular barrier. Indeed, pericyte-derived vitronectin prevents endothelial transcytosis by binding to integrin α5 subunit on ECs [30], and pericyte-secreted laminin interacts with the dystrophin–glycoprotein complex in astrocytes and regulates their endfeet polarization [5,31].

To serve as guardians of the capillary bed, pericytes also establish physical interactions with ECs and form a continuous chain-like network along the capillaries of the cerebral vasculature. Adequate coverage of the endothelium is sustained by active remodeling of distal pericyte processes through cytoskeletal rearrangements [32]. Of relevance, pericyte remodeling capabilities become exhausted with age [33], and this may explain why pericyte coverage is diminished in the vasculature of old mice [33,34]. An interesting observation is that pericyte depletion in adult mice leads to relatively mild BBB defects in different experimental models [35,36]. This includes adult induced Pdgfb ablation [36] and diphtheria toxin A (DTA) expression‎ in PDGFRβ+ cells [35]. Currently it is not clear why loss of pericytes leads to different vascular barrier phenotypes in development and adulthood. Given that the BBB becomes functional during late embryonic development, one can speculate that defects in pericyte coverage are only significant before the onset of BBB formation. Another possibility is that pericyte coverage determines the threshold for BBB defects, and Vazquez-Liebanas et al. showed that only <50% longitudinal pericyte coverage in adult brains leads to significant leakage defects [36]. This is coherent with previous observations of brain vessel phenotypes during development which demonstrated that pericyte coverage is positively correlated with BBB integrity [22]. Choe et al. also reported that DTA-induced loss of pericytes leads to capillary stalling due to increased interactions between ECs and leukocytes. However, because this effect was not observed in other adult pericyte depletion models [35], one should acknowledge that it is possible that the expression‎ of DTA generated unintended toxic effects beyond pericytes.

특별한 접촉으로 뇌 모세혈관층을 보호하는 페리사이트

페리사이트의 핵심 기능은

실질 공간으로의 체액 및 물질 이동을 제어하여

혈뇌 장벽(BBB)을 조절하는 것이다 [22,29].

따라서, 주변세포 기능 장애,

주변세포 모집 장애 및 주변세포 손실로 인한

주변세포 피복 결함은

모두 EC 세포간 이송 및 투과성 증가로 이어집니다 [22,29].

비정상적인 혈소판 유래 성장 인자 B (PDGF-B)/혈소판 유래 성장 인자 수용체 베타 (PDGFRβ) 신호 전달은

실험적으로 주변세포의 풍부함을 감소시키고

그에 따른 BBB 특성의 상실을 감소시키기에 충분합니다 [22,29].

또한, 신경혈관 단위를 구성하는 다른 세포 유형들 중에서도

특히 페리사이트에 의한 적절한 ECM(세포외 기질) 침착은

혈관 장벽의 무결성을 유지하는 데 필수적인 역할을 합니다.

실제로, 주변 세포에서 유래된 비트로넥틴은 내피 세포의 인테그린 α5 서브유닛에 결합하여 내피 세포의 트랜스사이토시스를 방지합니다 [30]. 또한 주변 세포가 분비하는 라미닌은 성상 세포의 디스트로핀-당단백질 복합체와 상호작용하여 성상 세포의 말단 발(endfeet) 분극화를 조절합니다 [5,31].

모세혈관층의 수호자 역할을 수행하기 위해, 페리사이트는 또한 EC와 물리적 상호작용을 확립하고 뇌혈관의 모세혈관을 따라 연속적인 사슬형 네트워크를 형성한다. 적절한 내피 세포 커버리지는 세포골격 재구성을 통한 원위부 페리사이트 돌기의 능동적 재형성으로 유지된다 [32]. 관련하여, 페리사이트의 재구성 능력은 노화와 함께 소진된다[33], 이는 노화 마우스 혈관계에서 페리사이트 커버리지가 감소하는 이유를 설명할 수 있다[33,34]. 흥미로운 점은 성체 생쥐에서 페리사이트가 고갈되면 다양한 실험 모델에서 상대적으로 경미한 혈뇌장벽 결함이 발생한다는 것이다[35,36]. 여기에는 성체 유도 Pdgfb 결손[36] 및 PDGFRβ+ 세포에서의 디프테리아 독소 A(DTA) 발현[35]이 포함된다. 현재 왜 주변세포 손실이 발달기 및 성인기에서 서로 다른 혈관 장벽 표현형을 유발하는지는 명확하지 않다. 혈뇌장벽이 후기 배아 발달 중에 기능화됨을 고려할 때, 주변세포 덮개 결함은 혈뇌장벽 형성 시작 전에만 유의미하다고 추측할 수 있다. 또 다른 가능성은 주변세포 덮개가 혈뇌장벽 결함의 임계값을 결정한다는 점이며, 바스케스-리바나스 등은 성인 뇌에서 종방향 주변세포 덮개가 50% 미만일 때만 유의미한 누출 결함이 발생함을 보여주었다 [36]. 이는 발달기 뇌 혈관 표현형에 대한 기존 관찰 결과와 일치하는데, 당시 연구에서는 페리사이트 덮개가 BBB 무결성과 양의 상관관계를 보인다는 점이 입증되었다[22]. Choe 등은 또한 DTA로 유발된 페리사이트 손실이 혈관내피세포(ECs)와 백혈구 간의 상호작용 증가로 인해 모세혈관 정지를 초래한다고 보고했다. 그러나 이 효과는 다른 성인 페리사이트 고갈 모델에서는 관찰되지 않았기 때문에[35], DTA 발현이 페리사이트 외에 의도하지 않은 독성 효과를 유발했을 가능성을 인정해야 한다.

Pericytes in disease

Pericyte dysfunction is a hallmark of various diseases (Figure 3). For a long time it was believed that maladaptive pericytes mainly affect vascular homeostasis because pericyte and EC functions are interdependent and require bidirectional communication (Box 2). However, there is growing evidence that pericytes have roles in processes beyond the vasculature. As such, pericyte-derived signals (hereafter referred to as pericrine signaling) also modulate tissue function in both physiology and disease. In the following section, we capture recent data showing new observations that link pericyte dysfunction and loss in vascular and non-vascular-related diseases.

질환에서의 페리사이트

페리사이트 기능 장애는 다양한 질환의 특징이다(그림 3). 오랫동안 페리사이트와 내피세포 기능은 상호 의존적이며 양방향 소통을 필요로 하기 때문에(박스 2), 부적응적 페리사이트가 주로 혈관 항상성에 영향을 미친다고 여겨져 왔다. 그러나 페리사이트가 혈관 외 과정에서도 역할을 한다는 증거가 점차 증가하고 있다. 따라서 페리사이트 유래 신호(이하 페리크린 신호전달이라 함)는 생리학적 및 병리학적 상황에서 조직 기능을 조절한다. 다음 섹션에서는 혈관 관련 및 비혈관 관련 질환에서 페리사이트 기능 장애와 소실을 연결하는 새로운 관찰 결과를 보여주는 최근 데이터를 정리한다.

Figure 3 Dysfunctional pericytes in disease.

Show full captionFigure viewer

Box 2

Key signaling pathways that orchestrate pericyte–EC crosstalk

Given the close relationship between pericytes and ECs, it is not surprising that bidirectional communication and regulation between them are crucial during vessel formation and maintenance. During angiogenesis, established examples of pericyte–EC communication include the PDGFRβ, transforming growth factor β1 (TGF-β1), ANG1, and NOTCH3 pathways [1]. PDGF-B production from tip ECs is the master signal that recruits PDGFRβ-expressing pericytes to newly formed vessels [1], together with CD146 (MCAM), which acts as a coreceptor for PDGFRβ [96]. Recent advances have shown that NCK1 and NCK2 promote phosphorylation of PDGFRβ in response to PDGF-BB and stimulate pericyte migration by inducing MRTF translocation to the nucleus where they interact with the serum response transcription factor (SRF) [13,21]. Similarly, jagged 1 (JAG1) expressed by ECs activates NOTCH3 in pericytes and promotes pericyte maturation [14,97] and the expression‎ of PDGFRβ [98]. Conversely, ANG1 is secreted by pericytes, activates the tyrosine receptor TIE2 in ECs, and promotes EC maturation and vascular integrity [2,15]. TGF-β exerts complex effects on ECs and pericytes, and TGF-B receptor 1 (also known as ALK5) plays a dominant role in these interactions. Indeed, deletion of ALK5 in ECs leads to pericyte dysfunction and hemorrhagic vascular malformations [99]. Instead, deletion of ALK5 in pericytes results in increased EC proliferation, reduced collagen deposition, and enhanced matrix metalloproteinase activity [19]. Of note, pericytes also express canonical EC receptors such as VEGF-R1 [15,16] and TIE2 which allow pericytes to modulate intrinsic EC signaling.

페리사이트-내피세포 간 교신을 조율하는 주요 신호 전달 경로

페리사이트와 내피세포(ECs)의 긴밀한 관계를 고려할 때,

혈관 형성 및 유지 과정에서 양방향적 소통과 조절이 중요하다는 것은 당연하다.

혈관신생 과정에서 확립된 페리사이트-내피세포 간 소통의 예로는

PDGFRβ, 변형성장인자 β1(TGF-β1), ANG1, NOTCH3 경로가 있다[1].

끝 EC에서 생성된 PDGF-B는 PDGFRβ 발현 페리사이트를 새로 형성된 혈관으로 모집하는 주요 신호이며[1], PDGFRβ의 공동 수용체 역할을 하는 CD146(MCAM)과 함께 작용합니다[96]. 최근 연구에 따르면 NCK1 및 NCK2는 PDGF-BB에 반응하여 PDGFRβ의 인산화를 촉진하고, MRTF를 핵으로 이동시켜 혈관주위세포의 이동을 자극하며, 핵 내에서 혈청 반응 전사 인자(SRF)와 상호작용합니다 [13,21]. 마찬가지로, 내피세포(ECs)가 발현하는 자그드 1(JAG1)은 페리사이트에서 NOTCH3를 활성화하여 페리사이트 성숙[14,97]과 PDGFRβ 발현[98]을 촉진한다. 반대로, 페리사이트가 분비하는 ANG1은 내피세포에서 티로신 수용체 TIE2를 활성화하여 내피세포 성숙과 혈관 무결성을 촉진한다[2,15]. TGF-β는 내피세포와 주변세포에 복합적인 영향을 미치며, TGF-β 수용체 1(ALK5)이 이러한 상호작용에서 주도적인 역할을 한다. 실제로 내피세포에서 ALK5를 제거하면 주변세포 기능 장애와 출혈성 혈관 기형이 발생한다[99]. 반면, 페리사이트에서 ALK5를 삭제하면 EC 증식이 증가하고, 콜라겐 침착이 감소하며, 매트릭스 메탈로프로테이나제 활성이 증가한다[19]. 주목할 점은 페리사이트가 VEGF-R1[15,16] 및 TIE2와 같은 정형적인 EC 수용체도 발현하여 페리사이트가 내재적 EC 신호전달을 조절할 수 있다는 것이다.

The CNS: a hotspot of pericyte-related vascular diseases

Pericyte-related vascular defects have been reported in various CNS diseases, including Alzheimer's disease (AD), Parkinson's disease, dementia, stroke, diabetic retinopathy, glaucoma, and intracranial vascular malformations [5,37–39]. The involvement of pericytes in several CNS-related diseases is partially explained by their abundance within the brain vasculature and their key role in maintaining the BBB, where barrier breakdown precedes neurodegeneration. Other phenotypes linking pericytes dysfunction and CNS disease include neuron death [40] and impaired neurovascular coupling [41,42]. Intriguingly, NG2+ retinal pericytes orchestrate neurovascular coupling through closed-ended nanotubes between pericytes on adjacent capillaries, even when they are positioned far apart. These nanotubes terminate in a gap junction at the recipient pericyte, which permits rapid fluxes of small molecules and calcium ions, thereby allowing pericytes to coordinate neuronal activity [41]. Indeed, maintaining adequate calcium levels is essential to sustain pericyte function in the CNS, and aberrant levels of calcium in NG2+ pericytes lead to poor recovery after ischemic stroke [23,43] or neovascular dysfunction and neuronal death in glaucoma [42].

AD is the prototypical example of a CNS disease associated with aberrant vascular function and BBB breakdown linked to pericyte dysfunction and loss [38,44]. Although the involvement of pericytes in AD has been recognized for several years [5,44], new insights have challenged the timeframe in which patients suffering from AD develop pericyte dysfunction and BBB impairment. Indeed, it is now understood that BBB breakdown is an early event in AD, and these defects are used as an early biomarker of cognitive decline [45]. We highlight recent observations which support the involvement of pericytes in the onset of AD. For instance, Nortley et al. showed that the reduction in cerebral blood flow, that is considered to be the first clinical manifestation of AD, is caused by amyloid-β-induced pericyte contraction in brain capillaries [46]. Another study indicated that cognitive decline and BBB disruption in AD are linked to accelerated pericyte degeneration in carriers of AD susceptibility allele apolipoprotein E4 (APOE4) [47], a process which occurs independently of amyloid-β pathology. In this context, APOE4 carriers show high baseline cerebrospinal fluid levels of soluble (s)PDGFRβ which can be used as a BBB pericyte injury biomarker [47]. Intriguingly, analysis of the cortex of APOE4 transgenic mice using single-nucleus (sn)RNA-seq and phosphoproteomics revealed profound molecular changes related to progressive BBB failure in both ECs and pericytes [48]. Nonetheless, because only a constitutive APOE4-expressing transgenic line was included in the study, it remains unclear whether the molecular alterations of ECs and pericytes solely comprise cell-autonomous effects. In addition, one should not forget that mice do not fully recapitulate all traits of AD. It has been recently noted that pericytes and microglia associations (described in both physiological mouse and human brains) are diminished in the brain capillaries of individuals with AD, and this may also have implications for BBB breakdown [49]. In human brain, two types of pericytes have been identified that are distinguished by solute transport and ECM organization (Box 1). Intriguingly, the second type seems to be selectively affected in AD [50]. Thus, identifying methods to specifically target this cluster of pericytes may provide new ways to maintain vascular fitness in AD.

Pericyte degeneration and death also encompasses early phases of diabetic retinopathy, in which pericytes are primary targets of hyperglycemic damage. Recent findings suggest that, upon initiation of hyperglycemia, pericytes shift towards cell-bridging positions, resulting in physical detachment from ECs [51]. Whether this remodeling is independent of pericyte death or is related to the initiation of that process needs further investigation. Mechanistically, pericyte detachment and shifting are induced by exogenous factors such as angiopoietin 2 (ANG2) and PDGF-B, and are reversed by insulin treatment, illustrating the dynamic behavior of pericytes in the microvasculature [51,52]. In line with this, PDGFB signaling through PDGFRβ and NCKs in pericytes that cover sprouting vessels during experimental proliferative retinopathy [oxygen-induced retinopathy (OIR) model] activates ectopic αSMA expression‎ and promotes pathological neovascularization [21]. Interestingly, depletion of retinal pericytes in adulthood does not phenocopy retinopathy unless another stimulus is present (e.g., vascular endothelial growth factor A, VEGF-A). Upon depletion of pericytes, either during vessel development or in adulthood followed by VEGF addition, inhibition of ANG2 action restrains the severity of the diabetic retinopathy-like phenotypes [11]. Molecular effectors governing the early phases of diabetic retinopathy have remained elusive, precluding the development of drugs aiming to halt disease onset. These data suggest that targeting pericyte adhesion and migration capacities may be of therapeutic interest. Furthermore, pericyte loss in diabetic retinopathy was recently associated with aberrant levels of circular RNAs [53], thereby suggesting the use of circular RNAs as a diagnostic biomarker for early pericyte dysfunction in disease.

Finally, we would like to stress that familial mutations in essential pericyte genes have also been linked to CNS abnormalities. Well-known examples include loss-of-function mutations in NOTCH3 as a cause of CADASIL [54], and mutations in PDGFRB as a cause of brain calcifications [55], neurological deterioration, and white matter lesions [56]. Of note, these genes are equally relevant for pericyte and vSMC biology, and it is unclear whether these mutations lead to distinct phenotypes in mural cells. Current next-generation sequencing approaches allow the discovery of somatic mutations present in pericytes at low allelic frequency. In line with this, it has been proposed that PIK3CA- and AKT-related somatic cerebral cavernous malformations in mice emerge from mutant pericytes [39,57]. However, these data have some caveats because the lineage-tracing experiments used to support these findings were performed with a CRE-recombinase mouse line that is neither pericyte-specific nor inducible.

중추신경계(CNS): 페리사이트 관련 혈관 질환의 주요 발생지

페리사이트 관련 혈관 결함은

알츠하이머병(AD), 파킨슨병, 치매, 뇌졸중, 당뇨병성 망막병증,

녹내장, 두개내 혈관 기형 등

다양한 중추신경계 질환에서 보고되었다 [5,37–39].

뇌혈관 내 풍부한 분포와 신경퇴행 전 단계에서

장벽 파괴가 발생하는 혈뇌장벽(BBB) 유지의 핵심 역할로 인해,

여러 중추신경계 관련 질환에서의 페리사이트 관여가 부분적으로 설명된다.

페리사이트 기능 장애와 중추신경계 질환을 연결하는 다른 표현형으로는

신경세포 사멸[40] 및 손상된 신경혈관 결합[41,42]이 포함된다.

흥미롭게도,

NG2+ 망막 페리사이트는 인접한 모세혈관의 페리사이트 간에 닫힌 끝의 나노튜브를 통해

신경혈관 결합을 조율하는데,

이는 페리사이트들이 멀리 떨어져 있을 때도 마찬가지이다.

이러한 나노튜브는 수신 페리사이트의 갭 접합부에서 종결되며, 이는 소분자 및 칼슘 이온의 신속한 유동을 허용하여 페리사이트가 신경 세포 활동을 조정할 수 있게 한다 [41]. 실제로, 적절한 칼슘 수치를 유지하는 것은 중추신경계에서 페리사이트 기능을 유지하는 데 필수적이며, NG2+ 페리사이트의 비정상적인 칼슘 수치는 허혈성 뇌졸중 후 회복 불량[23,43] 또는 녹내장에서 신생혈관 기능 장애 및 신경세포 사멸[42]을 초래한다.

알츠하이머병(AD)은

페리사이트 기능 장애 및 손실과 관련된 비정상적인 혈관 기능 및 혈뇌장벽(BBB) 파괴와 연관된

중추신경계 질환의 전형적인 사례이다[38,44].

비세포의 AD 관여는 수년간 인식되어 왔음에도[5,44], 새로운 연구 결과는 AD 환자가 비세포 기능 장애 및 BBB 손상을 발현하는 시간대에 대한 기존 관점을 재고하게 했다. 실제로 BBB 파괴는 AD의 초기 사건으로 이해되며, 이러한 결함은 인지 기능 저하의 조기 생체표지자로 활용된다[45]. 우리는 AD 발병에 페리사이트가 관여한다는 점을 뒷받침하는 최근 관찰 결과를 강조한다. 예를 들어, Nortley 등은 AD의 첫 번째 임상 증상으로 간주되는 뇌혈류 감소가 뇌 모세혈관에서 아밀로이드-β에 의해 유발된 페리사이트 수축에 기인함을 보여주었다[46]. 또 다른 연구에서는 AD의 인지 기능 저하와 혈뇌장벽 손상이 아밀로이드-β 병리와 무관하게 발생하는 과정인, AD 감수성 대립유전자 아포지단백질 E4(APOE4) 보유자에서 가속화된 페리사이트 퇴화와 연관되어 있음을 시사하였다[47]. 이러한 맥락에서, APOE4 보유자는 혈뇌장벽 페리사이트 손상 바이오마커로 활용될 수 있는 용해성(s)PDGFRβ의 높은 기준선 뇌척수액 수치를 보인다[47]. 흥미롭게도, 단일핵(sn)RNA-seq 및 인산화 단백질체학 분석을 통해 APOE4 트랜스제닉 마우스의 피질에서 내피세포(ECs)와 페리사이트 모두에서 진행성 BBB 기능 부전과 관련된 심대한 분자적 변화가 발견되었다[48]. 그러나 연구에 포함된 트랜스제닉 계통이 APOE4를 지속적으로 발현하는 계통뿐이었기 때문에, 내피세포와 페리사이트의 분자적 변화가 순수하게 세포 자율적 효과로만 구성되는지는 여전히 불분명하다. 또한, 생쥐가 AD의 모든 특성을 완전히 재현하지는 않는다는 점을 간과해서는 안 된다. 최근 AD 환자의 뇌 모세혈관에서는 생리적 생쥐 및 인간 뇌에서 모두 관찰되는 페리사이트와 미세아교세포의 연관성이 감소한다는 점이 지적되었으며, 이는 혈뇌장벽 붕괴에도 영향을 미칠 수 있다[49]. 인간 뇌에서는 용질 수송 및 ECM 조직화에 따라 두 가지 유형의 페리사이트가 확인되었다(Box 1). 흥미롭게도 두 번째 유형은 AD에서 선택적으로 영향을 받는 것으로 보인다[50]. 따라서 이 페리사이트 군집을 특이적으로 표적화하는 방법을 규명하는 것이 AD에서 혈관 건강을 유지하는 새로운 방안을 제시할 수 있을 것이다.

페리사이트 퇴행 및 사멸은 당뇨병성 망막병증의 초기 단계에서도 관찰되며, 이 경우 페리사이트는 고혈당 손상의 주요 표적이 된다. 최근 연구에 따르면 고혈당이 시작되면 페리사이트가 세포 연결 위치로 이동하여 혈관내피세포(ECs)로부터 물리적으로 분리된다[51]. 이러한 재구성이 페리사이트 사멸과 무관한지, 아니면 그 과정의 시작과 관련이 있는지는 추가 연구가 필요하다. 기전적으로, 페리사이트 이탈 및 이동은 안지오포이에틴 2(ANG2) 및 PDGF-B와 같은 외인성 인자에 의해 유도되며, 인슐린 치료로 역전되어 미세혈관 내 페리사이트의 동적 행동을 보여준다[51,52]. 이와 일치하게, 실험적 증식성 망막병증(산소유발성 망막병증(OIR) 모델) 동안 발아 혈관을 덮고 있는 페리사이트에서 PDGFRβ 및 NCKs를 통한 PDGFB 신호전달은 이소성 αSMA 발현을 활성화하고 병리학적 신생혈관 형성을 촉진한다[21]. 흥미롭게도, 성체에서 망막 주위세포를 제거해도 다른 자극(예: 혈관내피성장인자 A, VEGF-A)이 없는 한 망막병증을 재현하지 못한다. 혈관 발달 중이거나 성체에서 주위세포를 제거한 후 VEGF를 추가할 경우, ANG2 작용을 억제하면 당뇨병성 망막병증 유사 표현형의 심각성을 제한한다[11]. 당뇨병성 망막병증 초기 단계를 지배하는 분자적 효과기는 여전히 불분명하여, 질병 발병을 차단하는 약물 개발을 가로막고 있다. 이러한 데이터는 페리사이트의 접착 및 이동 능력을 표적으로 삼는 것이 치료적 관심을 가질 수 있음을 시사한다. 또한, 당뇨병성 망막병증에서의 페리사이트 손실은 최근 비정상적인 수준의 원형 RNA [53]와 연관되어, 질병 초기 페리사이트 기능 장애의 진단적 바이오마커로 원형 RNA를 활용할 수 있음을 시사한다.

마지막으로, 필수적인 페리사이트 유전자의 가족성 돌연변이가 중추신경계(CNS) 이상과도 연관되어 있음을 강조하고자 한다. 잘 알려진 예로는 NOTCH3의 기능 상실 돌연변이가 CADASIL의 원인[54]이며, PDGFRB 돌연변이가 뇌 석회화[55], 신경학적 악화 및 백질 병변[56]의 원인으로 알려져 있다. 주목할 점은, 이러한 유전자들이 페리사이트와 혈관 평활근 세포(vSMC) 생물학에 동등하게 관련되어 있으며, 이러한 돌연변이가 벽세포에서 서로 다른 표현형을 유발하는지 여부는 불분명하다는 것이다. 현재의 차세대 시퀀싱 접근법은 낮은 대립유전자 빈도로 페리사이트에 존재하는 체세포 돌연변이를 발견할 수 있게 한다. 이와 일치하게, 쥐에서 PIK3CA 및 AKT 관련 체세포 뇌동정맥기형이 돌연변이 페리사이트에서 발생한다는 제안이 있다[39,57]. 그러나 이러한 결과를 뒷받침하기 위해 사용된 계통 추적 실험은 페리사이트 특이적이지도 유도 가능하지도 않은 CRE 재조합효소 마우스 계통으로 수행되었기 때문에 이 데이터에는 몇 가지 주의사항이 있다.

Pathobiological pericytes beyond the CNS

Although the implications of pericytes in diseases beyond the CNS are less well studied, the number of diseases demonstrating the involvement of pericytes continues to expand. We discuss here emerging evidence supporting a relevant role of pericytes in myocardial infarction [58], acute lung injury [59], and diabetes [60] as prototypical examples. For instance, after myocardial infarction, pericytes regulate inflammation and immune cell trafficking, and modulate ECM remodeling and revascularization [61]. In line with this, molecular reprogramming of PDGFRβ+NG2+ cardiac pericytes into vSMCs through inhibition of MEK1/2 improved the functional cardiac response by promoting revascularization [58]. In acute lung inflammation, the crosstalk between endothelium-derived nitric oxide (NO) and pericyte soluble guanylate cyclase (sGC) is impaired, leading to elevated vascular permeability [59]. Pharmacological activation of the NO–sGC axis led to an improved pericyte-driven inflammatory response. Moreover, pericytes in pancreatic islets exert vascular control of hormone secretion and glucose homeostasis, and pericyte alteration has been linked to diabetic islet dysfunction [60]. Interestingly, pericyte effects on islet functionality are not limited to vascular support for insulin secretion because pancreatic β cell maturation and functionality rely on pericyte-derived bone morphogenic protein 4 (BMP4). Recently, other pericrine signaling molecules have been identified as key players in the functional regulation of tissue parenchyma encompassing a range of organ-specific functions in both vascular and non-vascular interfaces. Two interesting examples are that leptin receptor-expressing pericytes in the mediobasal hypothalamus mediate energy balance via neuronal leptin signaling [62], and that the Hippo–YAP/TAZ pathway in pericytes generates essential pericrine signals to epithelial and ECs during lung morphogenesis [63]. All things considered, these studies suggest that restoring the physiological functions of pericytes improves blood vessel performance and disease outcomes in various contexts, which may encourage the development of novel pericyte-focused therapies.

중추신경계를 넘어선 병리생물학적 페리사이트

중추신경계 외 질환에서의 페리사이트 역할은 덜 연구되었지만,

페리사이트 관여가 입증된 질환의 수는 계속 증가하고 있다.

본고에서는

심근경색[58], 급성 폐손상[59], 당뇨병[60]을 대표적 사례로

페리사이트의 관련성을 뒷받침하는 새로운 증거를 논의한다.

예를 들어,

심근경색 후 페리사이트는 염증과 면역 세포 이동을 조절하며,

ECM 재형성과 재혈관화를 조절한다[61].

이와 일치하게,

MEK1/2 억제를 통한 PDGFRβ+NG2+ 심장 페리사이트의 vSMC로의 분자적 재프로그래밍은

재혈관화를 촉진하여 기능적 심장 반응을 개선했다[58].

급성 폐 염증에서는

내피 유래 산화질소(NO)와 페리사이트 용해성 구아닐산 사이클라제(sGC) 간의 교신이 손상되어

혈관 투과성이 증가한다[59]. N

O-sGC 축의 약리학적 활성화는

페리사이트에 의한 염증 반응을 개선시켰다.

또한 췌도 내 페리사이트는

호르몬 분비와 포도당 항상성에 대한 혈관 조절 기능을 수행하며,

페리사이트 변화는 당뇨병성 췌도 기능 장애와 연관되어 있다[60].

흥미롭게도 페리사이트가

췌도 기능에 미치는 영향은 인슐린 분비를 위한 혈관 지원에 국한되지 않는다.

췌장 β 세포의 성숙과 기능성은

페리사이트 유래 골형성단백질 4(BMP4)에 의존하기 때문이다.

최근 다른 페리크린 신호 분자들이 혈관 및 비혈관 인터페이스에서 다양한 장기 특이적 기능을 포괄하는 조직 실질의 기능적 조절에 핵심 역할을 한다는 것이 밝혀졌다. 두 가지 흥미로운 사례는 중기저 시상하부의 렙틴 수용체 발현 페리사이트가 신경 세포 렙틴 신호 전달을 통해 에너지 균형을 매개한다는 점[62], 폐 형태 발생 과정에서 페리사이트 내 Hippo–YAP/TAZ 경로가 상피세포 및 내피세포에 필수적인 페리크린 신호를 생성한다는 점이다[63]. 종합해 볼 때, 이러한 연구들은 페리사이트의 생리적 기능 회복이 다양한 맥락에서 혈관 성능과 질병 예후를 개선함을 시사하며, 이는 새로운 페리사이트 중심 치료법 개발을 촉진할 수 있다.

Tumor pericytes: loss or change of identity?

Many preclinical studies over the past decade showed that pericyte dysfunction is involved in cancer progression [64]. In this context, it is now well established that tumoral vessels are poorly covered by pericytes [65,66], that diminished pericyte recruitment and maturation lead to enhanced vascular and tumor growth [15,66], and that pericyte depletion favors metastasis. However, there is no consensus about whether defective pericyte–EC interaction in tumors relates to loss of pericyte molecular identity [67], pericyte transdifferentiation into fibroblasts [68], poor pericyte recruitment [66], or a combination of these phenotypes. Functionally, metabolic reprogramming of tumor pericytes has also been implicated in abnormal blood vessel contraction and unfavorable patient outcomes [69].

scRNA-seq analyses have provided a new layer of information about tumor pericytes by showing that tumor pericytes are, in fact, relatively homogeneous [70–72]. However, one should acknowledge that most tumor scRNA-seq datasets include low numbers of pericytes, and the presence of tumor pericyte subclusters may have been obscured by the granularity of the data. One way to overcome this limitation would be to establish spatially resolved pericyte-focused atlases by using multiomic approaches. An interesting observation is that pericyte phenotypes are distinct depending on the mechanisms by which tumor vessels form. While angiogenic pericytes persist in an active immature state characterized by a signature of motility and ECM organization, pericytes covering co-opted vessels remain largely quiescent [71,73]. Together, these studies support a model in which pericytes undergo genetic and molecular reprogramming in cancer, which in turn has negative consequences for disease progression. Nonetheless, one should not forget that most of these data refer to mouse preclinical studies, and adequate longitudinal studies in humans are missing. Of note, the so-called pericrine signaling response is also at play in cancer. Indeed, in hepatocellular carcinoma it has been shown that metabolic reprogramming in tumor cells activates HSCs which in turn promote tumorigenesis through the secretion of senescence-associated factors [74]. Another study has described that loss of integrin β3 in tumor pericytes leads to enhanced focal adhesion kinase (FAK)-mediated cytokine release by pericytes, which subsequently stimulates tumor survival and growth [75].

종양 페리사이트: 정체성 상실인가 변화인가?

지난 10년간의 다수 전임상 연구에서 혈관주위세포 기능 장애가 암 진행과 연관됨이 밝혀졌다[64]. 이러한 맥락에서, 종양 혈관은 혈관주위세포로 제대로 덮이지 않으며[65,66], 혈관주위세포의 모집 및 성숙 감소가 혈관 및 종양 성장을 촉진하며[15,66], 혈관주위세포 고갈이 전이를 촉진한다는 점이 현재 확립된 사실이다. 그러나 종양 내 페리사이트-내피세포(EC) 상호작용 결함이 페리사이트 분자 정체성 상실[67], 섬유아세포로의 페리사이트 전분화[68], 페리사이트 모집 불량[66], 또는 이러한 표현형의 복합적 결과인지에 대해서는 합의가 이루어지지 않았다. 기능적으로, 종양 페리사이트의 대사적 재프로그래밍은 비정상적인 혈관 수축 및 불리한 환자 예후와도 연관되어 있다[69].

scRNA-seq 분석은 종양 페리사이트가 사실 상대적으로 동질적임을 보여줌으로써 종양 페리사이트에 대한 새로운 정보 계층을 제공했다[70–72]. 그러나 대부분의 종양 scRNA-seq 데이터셋에는 페리사이트 수가 적게 포함되어 있으며, 데이터의 세분화 수준으로 인해 종양 페리사이트 하위 클러스터의 존재가 가려졌을 수 있음을 인정해야 한다. 이러한 한계를 극복하는 한 가지 방법은 다중체 접근법을 활용하여 공간적으로 분해능이 높은 페리사이트 중심 아틀라스를 구축하는 것이다. 흥미로운 관찰 결과는 종양 혈관이 형성되는 메커니즘에 따라 페리사이트 표현형이 다르다는 점이다. 혈관신생성 페리사이트는 이동성과 ECM 조직화의 특징을 보이는 활성 미성숙 상태를 유지하는 반면, 편입된 혈관을 덮고 있는 페리사이트는 대체로 휴면 상태를 유지한다[71,73].

이러한 연구들은 총체적으로, 페리사이트가 암에서 유전적·분자적 재프로그래밍을 겪으며 이는 질병 진행에 부정적 영향을 미친다는 모델을 지지한다. 그럼에도 불구하고, 대부분의 데이터가 생쥐 전임상 연구에 기반하며 인간 대상의 적절한 종단 연구가 부족하다는 점을 간과해서는 안 된다. 주목할 점은 소위 페리크린 신호 전달 반응 역시 암에서 작용한다는 것이다. 실제로 간세포암종에서 종양 세포의 대사적 재프로그래밍이 HSC를 활성화시키며, 이는 노화 관련 인자의 분비를 통해 종양 형성을 촉진하는 것으로 밝혀졌다[74]. 또 다른 연구에서는 종양 페리사이트에서 인테그린 β3의 상실이 페리사이트에 의한 초점 접착 키나아제(FAK) 매개 사이토카인 방출을 증가시켜 종양 생존 및 성장을 촉진한다고 기술하였다[75].

Pericyte immunomodulatory properties

Emerging evidence suggests that pericytes form an integral part of the immune surveillance unit rather than solely performing complementary functions. Upon proinflammatory stimuli, pericytes promote endothelial expression‎ of the leukocyte adhesion molecules vascular cell adhesion protein 1 (VCAM-1) and/or intracellular adhesion molecule 1 (ICAM-1) in the CNS, lung, skin, or muscle that subsequently promote T cell or macrophage infiltration into the affected tissue [59,76,77]. The Rgs5+ and Col1a1+ subgroups of PDGFRβ+ perivascular cells seem to be early responders to neuroinflammation [78]. Of note, inflammation per se induces pericyte detachment from the endothelium and impairs barrier properties [59]. In addition to the physical interaction with leukocytes, pericytes also secrete and respond to cytokines that further regulate immune cell functions, including both innate and adaptive responses. The chemotactic migration and effector functions of neutrophils, T cells, and macrophages are dependent on these early pericrine signals, including MIF, CXCL1, and CCL2 [77–80].

It is now believed that modulation of the immune-related functions of pericytes could affect the outcome of disease progression. For example, pericyte-deficient Pdgfbret/ret mice exhibit increased leukocyte infiltration and activation, leading to aberrant inflammation in a model of experimental autoimmune encephalomyelitis [76]. Treatment with antagonistic VCAM-1 and ICAM-1 antibodies partially rescued the excessive inflammatory phenotype in Pdgfbret/ret mice. Similarly, activating sGC in acute lung injury improved disease outcomes by increasing pericyte interaction with ECs [59]. Conversely, in the tumor microenvironment, tumor cells induce autophagy of NG2+ pericytes which equips them with immunosuppressive properties that favor tumor cell survival and prevent antitumor T cell responses [81]. Pericytes have also been implicated in the underlying pathophysiology of emergent infectious diseases such as coronavirus disease 2019 (COVID-19) [82], thus presenting an additional niche of investigation for the field in the coming years. In conclusion, the importance of pericytes during various inflammatory processes is growing in recognition, although the modulation of immunity by pericytes can have double-edged outcomes.

페리사이트의 면역 조절 특성

새로운 증거에 따르면,

페리사이트는 단순히 보조 기능을 수행하는 것이 아니라

면역 감시 단위의 필수적인 부분을 형성하는 것으로 보입니다.

전염증성 자극에 반응하여,

페리사이트는 중추신경계(CNS),

폐, 피부 또는 근육에서 백혈구 접착 분자인 혈관세포접착단백질 1(VCAM-1) 및/또는

세포내접착분자 1(ICAM-1)의 내피세포 발현을 촉진하며,

이는 이후 T 세포 또는 대식세포의 손상 조직으로의 침투를 촉진한다[59,76,77].

PDGFRβ+ 혈관주위 세포의 Rgs5+ 및 Col1a1+ 하위군은 신경염증에 대한 초기 반응자로 보인다[78]. 주목할 점은 염증 자체가 내피로부터 페리사이트의 이탈을 유발하고 장벽 기능을 손상시킨다는 것이다[59]. 백혈구와의 물리적 상호작용 외에도, 페리사이트는 선천적 및 적응적 반응을 포함한 면역 세포 기능을 추가로 조절하는 사이토카인을 분비하고 이에 반응한다. 호중구, T 세포 및 대식세포의 화학유인성 이동 및 효과기 기능은 MIF, CXCL1 및 CCL2를 포함한 이러한 초기 페리크린 신호에 의존한다 [77–80].

현재는 페리사이트의 면역 관련 기능 조절이 질병 진행 결과에 영향을 미칠 수 있다고 여겨진다. 예를 들어, 페리사이트 결핍 Pdgfbret/ret 마우스는 실험적 자가면역 뇌척수염 모델에서 백혈구 침윤 및 활성화가 증가하여 비정상적인 염증을 보인다[76]. VCAM-1 및 ICAM-1 항체 치료는 Pdgfbret/ret 마우스의 과도한 염증 표현형을 부분적으로 회복시켰다. 마찬가지로, 급성 폐 손상에서 sGC를 활성화하면 내피세포와의 상호작용이 증가하여 질병 예후가 개선되었다[59]. 반대로 종양 미세환경에서는 종양 세포가 NG2+ 페리사이트의 자가포식을 유도하여, 이들이 종양 세포 생존에 유리하고 항종양 T 세포 반응을 억제하는 면역억제 특성을 갖도록 한다[81]. 페리사이트는 또한 코로나바이러스 감염증-2019(COVID-19) [82]와 같은 신종 감염병의 근본적 병리생리학에도 관여하는 것으로 밝혀져 향후 연구 분야의 추가적 탐구 영역을 제시한다. 결론적으로, 페리사이트에 의한 면역 조절이 양날의 검과 같은 결과를 초래할 수 있음에도 불구하고, 다양한 염증 과정에서의 페리사이트의 중요성은 점차 인식되고 있다.

Do pericytes contribute to fibrosis?

A dysregulated tissue repair response after acute or chronic injury can lead to the onset of fibrosis associated with the abnormal accumulation of activated and contractile αSMA+ myofibroblasts [83]. Myofibroblasts secrete high amounts of inflammatory mediators, growth factors, and ECM components, and promote aberrant ECM remodeling. The current consensus places fibroblasts as the predominant origin of myofibroblasts, although various studies have found alternative cellular origins [83]. Indeed, the existence of pericyte-to-myofibroblast transition has been proposed as a contributing factor in several fibrotic contexts [2], but the promiscuity of cell markers within the mesenchymal compartments across tissues has led to ambiguous and contradictory observations. In the latest developments, scRNA-seq, ATAC-seq, and spatial transcriptomics provide new insights into this conundrum which support a pericyte origin of myofibroblasts in the fibrotic liver, colon, and kidney. For instance, central vein-associated Rgs5+ HSCs are thought to be the dominant origin of ECM-producing myofibroblasts in fibrotic mouse liver [84]. In human colorectal cancer, a subset of periostin (POSTN)+ myofibroblasts seem to originate from RGS5+ pericytes [70]; similarly, NOTCH3+RGS5+PDGFRβ+ human pericytes contribute to the generation of POSTN+PDGFRα+NKD2+ myofibroblasts during kidney fibrosis [85]. Surprisingly, however, the same authors did not capture the existence of profibrotic pericytes during myocardial infarction when sequencing the entire heart [86]. Although transdifferentiation from pericytes to myofibroblasts may be tissue-specific, it is fair to acknowledge that the latter study did not include pericytes in the trajectory analysis that predicted the origins of myofibroblasts.

Lineage-tracing experiments in mice have also cast some light onto the role of pericytes in fibrosis. For instance, Pham et al. showed that myofibroblast genes are enriched in Tbx18+ pericytes from injured mouse hearts and brains [87], and Dias et al. reported that GLAST+PDGFRβ+ perivascular cells also contribute to fibrosis in the post-stroke brain [88]. In line with this, depletion of GLAST+PDGFRβ+ perivascular cells in the spinal cord leads to reduced fibrotic scar after injury and improves neuronal function [89]. Of note, GLAST+PDGFRβ+ perivascular cells were characterized as spinal cord pericytes by the researchers, but there is insufficient evidence to rule out that these cells might be fibroblasts or astrocytes. This further exemplifies that the shared marker expression‎ profiles of pericytes and other perivascular residing cells still hamper the design of robust pericyte reporter models. Moreover, multiple studies have found no significant pericyte origin for myofibroblasts in distinct fibrosis models of the CNS [90,91] and heart [92]. Although the discrepancies between studies may reflect the lack of robust pericyte identification strategies, it appears that the occurrence of pericyte-to-myofibroblast transition is tissue-specific, and is contingent both on the local microenvironment and on the extent of the injurious stimuli [83]. This emphasizes the need to further unravel the fibrotic pericyte responses in different organs and prompts the question of how myofibroblast origins relate to different pathophysiological phenotypes. All things considered, the origin of myofibroblasts may involve distinct precursor cells depending on the circumstances, although a role for pericytes seems to be indisputable (Figure 4).

페리사이트는 섬유화에 기여하는가?

급성 또는 만성 손상 후 조절되지 않은 조직 수리 반응은 활성화되고

수축성 αSMA+ 근섬유아세포의 비정상적 축적과 관련된 섬유화의 발병으로 이어질 수 있다[83].

근섬유아세포는

다량의 염증 매개체, 성장 인자 및 ECM 성분을 분비하며

비정상적인 ECM 재구성을 촉진한다.

현재의 합의는 섬유모세포가 근섬유모세포의 주요 기원 세포로 간주되지만, 다양한 연구에서 대체 세포 기원을 제시한 바 있다[83]. 실제로, 여러 섬유화 맥락에서 주변세포-근섬유모세포 전환의 존재가 기여 요인으로 제안되었으나[2], 조직 간 중간엽 구획 내에서 세포 표지자의 다중성으로 인해 모호하고 모순된 관찰 결과가 발생했다. 최신 연구에서 단일세포 RNA-seq, ATAC-seq 및 공간 전사체학은 이 난제에 대한 새로운 통찰력을 제공하며, 섬유화 간, 대장 및 신장에서의 근섬유아세포의 페리사이트 기원을 지지한다. 예를 들어, 중앙정맥 관련 Rgs5+ 간성상세포(HSCs)는 섬유화 마우스 간에서 ECM 생성 근섬유아세포의 주요 기원인 것으로 여겨진다[84]. 인간 대장암에서는 페리오스틴(POSTN)+ 근섬유아세포의 일부가 RGS5+ 페리사이트에서 유래하는 것으로 보인다[70]; 마찬가지로, NOTCH3+RGS5+PDGFRβ+ 인간 페리사이트는 신장 섬유화 과정에서 POSTN+PDGFRα+NKD2+ 근섬유아세포 생성에 기여한다[85]. 그러나 놀랍게도 동일한 저자들은 심장 전체를 시퀀싱할 때 심근경색 중 섬유화 촉진성 페리사이트의 존재를 포착하지 못했습니다[86]. 페리사이트에서 근섬유아세포로의 전분화는 조직 특이적일 수 있지만, 후자의 연구가 근섬유아세포의 기원을 예측하는 궤적 분석에 페리사이트를 포함하지 않았다는 점을 인정하는 것이 타당합니다.

마우스에서의 계통 추적 실험 또한 섬유화에서 페리사이트의 역할에 대한 일부 통찰을 제공했다. 예를 들어, Pham 등은 손상된 마우스 심장과 뇌에서 유래한 Tbx18+ 페리사이트에서 근섬유모세포 유전자들이 풍부하게 존재함을 보여주었다[87], 그리고 Dias 등은 GLAST+PDGFRβ+ 혈관 주위 세포들도 뇌졸중 후 뇌의 섬유화에 기여한다고 보고했다[88]. 이와 일치하게, 척수 내 GLAST+PDGFRβ+ 혈관주위세포를 제거하면 손상 후 섬유성 흉터가 감소하고 신경 기능이 개선된다[89]. 주목할 점은 연구진이 GLAST+PDGFRβ+ 혈관주위세포를 척수 페리사이트로 특성화했으나, 이 세포들이 섬유아세포나 성상세포일 가능성을 배제할 만한 증거는 불충분하다는 것이다. 이는 혈관주위세포와 다른 혈관주위 거주 세포의 공유된 표지자 발현 프로파일이 여전히 강력한 혈관주위세포 리포터 모델 설계를 방해함을 다시 한번 보여준다. 더욱이 여러 연구에서 중추신경계[90,91] 및 심장[92]의 다양한 섬유화 모델에서 근섬유아세포의 유의미한 혈관주위세포 기원을 발견하지 못했다. 연구 간 불일치는 강력한 페리사이트 식별 전략의 부재를 반영할 수 있으나, 페리사이트에서 근섬유아세포로의 전환 발생은 조직 특이적이며 국소 미세환경과 손상 자극의 정도에 모두 좌우되는 것으로 보인다[83]. 이는 다양한 장기에서 섬유화성 페리사이트 반응을 더 깊이 규명할 필요성을 강조하며, 근섬유아세포의 기원이 다양한 병리생리학적 표현형과 어떻게 연관되는지에 대한 질문을 제기한다. 종합해 볼 때, 페리사이트의 역할은 분명해 보이지만(그림 4), 근섬유아세포의 기원은 상황에 따라 서로 다른 전구 세포를 포함할 수 있다.

그림 4 섬유화에서 근섬유모세포의 기원으로서의 페리사이트.

결론

Figure 4 Pericytes as a source of myofibroblasts in fibrosis.

Show full captionFigure viewer

Concluding remarks

It is widely recognized that pericytes play an important role in blood vessel formation, stabilization, and function, and that degeneration or loss of brain pericytes impairs their protective barrier properties. Recent advances have revealed novel and crucial roles for pericytes across tissues in a variety of vascular and non-vascular processes. Single-cell technology is becoming more commonly used to better understand the molecular processes that define pericytes in health and disease. Although the molecular and functional attributes of pericytes are not fully elucidated, deep sequencing has revealed organotypic pericyte heterogeneities and new criteria to distinguish pericytes from other cell types. Despite the numerous suggested roles of pericytes in various diseases and physiological processes, including neurodegeneration, cancer, fibrosis, blood flow regulation, and inflammation, the underlying organotypic mechanisms of these contributions are not yet fully understood. The discrepancies between some studies highlight the importance of designing suitable mouse models for eval‎uating the specific mechanisms by which pericytes impact on these processes, and future cross-validations with human data are warranted to ascertain the clinical relevance of pathological pericytes (see Outstanding questions). Overall, based on the emerging evidence on the contribution of pericytes to several diseases, we anticipate an increasing emphasis on pericyte-oriented research in vascular (and non-vascular) studies in the coming years.

결론

페리사이트가 혈관 형성, 안정화 및 기능에 중요한 역할을 하며, 뇌 페리사이트의 퇴화 또는 손실이 보호 장벽 특성을 손상시킨다는 점은 널리 인정받고 있다. 최근 연구를 통해 다양한 조직에서 혈관 및 비혈관 과정 전반에 걸쳐 페리사이트의 새롭고 중요한 역할이 밝혀졌다. 건강 및 질병 상태에서 페리사이트를 정의하는 분자적 과정을 더 잘 이해하기 위해 단일 세포 기술이 점점 더 널리 사용되고 있다. 비록 페리사이트의 분자적·기능적 특성이 완전히 규명되지는 않았으나, 심층 시퀀싱을 통해 장기유사 페리사이트의 이질성과 다른 세포 유형과 페리사이트를 구분하는 새로운 기준이 밝혀졌다. 신경퇴행, 암, 섬유화, 혈류 조절, 염증 등 다양한 질환 및 생리적 과정에서의 페리사이트 역할이 다수 제안되었음에도, 이러한 기여의 근본적인 장기유사 기전은 아직 완전히 이해되지 않았다. 일부 연구 간의 불일치는 페리사이트가 이러한 과정에 영향을 미치는 특정 메커니즘을 평가하기 위한 적절한 마우스 모델 설계의 중요성을 강조하며, 병리학적 페리사이트의 임상적 관련성을 확인하기 위해서는 향후 인간 데이터와의 교차 검증이 필요합니다(미해결 질문 참조). 전반적으로, 여러 질환에 대한 페리사이트의 기여에 관한 새로운 증거를 바탕으로, 향후 몇 년간 혈관(및 비혈관) 연구에서 페리사이트 중심 연구에 대한 강조가 증가할 것으로 예상됩니다.

Outstanding questions

Why are pericytes molecularly and functionally promiscuous and heterogeneous between distinct tissues? If the transcription factors that regulate pericyte differentiation and function are similar across tissues, will epigenetic mechanisms provide cues into the mechanisms of pericyte identity?

Will spatial molecular atlases focused on pericytes address the key functional and identity conundrums posed by transitioning phenotypes? Will this suffice, or are complementary morphological and functional studies also necessary?

Will pericyte-focused therapy provide new means to stimulate functional angiogenesis in pathology? Given that pericytes are associated with many diseases, will (and how broadly can) pericyte-focused therapies improve patient outcomes?

Neurodegenerative diseases are age-related diseases, and pericyte degeneration is an aging process. Hence, is pericyte degeneration a confounding factor in the development of age-related neurodegeneration? Will maintaining healthy pericyte function promote healthy aging?

미해결 과제

왜 페리사이트는 분자적·기능적으로 다중적이며 조직 간 이질성을 보일까? 페리사이트 분화 및 기능을 조절하는 전사 인자가 조직 전반에 유사하다면, 후성유전학적 기전이 페리사이트 정체성 메커니즘에 대한 단서를 제공할 수 있을까?

페리사이트에 초점을 맞춘 공간적 분자 지도(molecular atlas)가 전환형 표현형이 제기하는 핵심 기능적·정체성 난제를 해결할 수 있을까? 이것만으로도 충분한가, 아니면 보완적인 형태학적·기능적 연구도 필요한가?

페리사이트 중심 치료법이 병리학적 상황에서 기능적 혈관신생을 촉진하는 새로운 수단이 될 수 있을까? 페리사이트가 다양한 질환과 연관된다는 점을 고려할 때, 페리사이트 중심 치료법이 환자 예후를 개선할 수 있을까(그리고 어느 정도까지 가능할까)?

신경퇴행성 질환은 노화 관련 질환이며, 페리사이트 퇴화는 노화 과정이다. 따라서 페리사이트 퇴화는 노화 관련 신경퇴행성 질환 발병의 혼란 요인일까? 건강한 페리사이트 기능 유지가 건강한 노화를 촉진할 수 있을까?

AcknowledgmentsAcknowledgments

We would like to thank Sandra D. Castillo, Ana Angulo-Urarte, and Leonor Gouveia for their valuable feedback. Figures were created with BioRender.com. We thank the Centres de Recerca de Catalunya (CERCA) Program/Generalitat de Catalunya and the Josep Carreras Foundation for institutional support. Work-related to this publication in the laboratory of M.G. is supported by research grants from la Asociación Española Contra el Cancer (AECC)-Grupos Traslacionales (GCTRA18006CARR); and by Worldwide Cancer Research (WCR 21-0159). H.v.S. received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 955951; P.V. recieved funding from AECC (AECC-INVES211084VILL) and from the Spanish Ministry of Science and Innovation (RYC2020-029929-I). We apologize to the many authors whose primary papers could not be cited owing to space constraints.

Declaration of interests

The authors declare no conflicts of interest.

References

1.

Armulik, A. ...

Pericytes: developmental, physiological, and pathological perspectives, problems, and promises

Dev. Cell. 2011; 21:193-215

Full Text

Full Text (PDF)

Scopus (1996)

PubMed

Google Scholar

2.

Holm, A. ...

Microvascular mural cell organotypic heterogeneity and functional plasticity

Trends Cell Biol. 2018; 28:302-316

Full Text

Full Text (PDF)

Scopus (85)

PubMed

Google Scholar

3.

Petrova, T.V. ∙ Koh, G.Y.

Biological functions of lymphatic vessels

Science. 2020; 369, eaax4063

Crossref

Scopus (234)

PubMed

Google Scholar

ReviewVolume 34, Issue 1p58-71January 2024Open access

Download Full Issue

Pericytes in the disease spotlight

Hielke van Splunder1,5 ∙ Pilar Villacampa2,5 ∙ Anabel Martínez-Romero1 ∙ Mariona Graupera1,3,4 mgraupera@carrerasresearch.org

Affiliations & NotesArticle Info

• Download PDF
• Cite

• Share

• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Highlights

Molecular and functional pericyte studies at single-cell resolution are providing new insights into long-standing questions about pericyte heterogeneity.

Pericytes are not identified by a single marker but instead by gene expression‎ signatures that show substantial inter-organ differences.

Pericytes orchestrate and precede endothelial cell responses during angiogenesis.

Pericyte degeneration and dysfunction, that are triggered by the onset of some diseases, contribute to the progression of those diseases in both vascular and non-vascular contexts.

The number of diseases with pericyte dysfunction continues to expand, thereby anticipating a promising future for pericyte-focused therapy.

Abstract

Pericytes are known as the mural cells in small-caliber vessels that interact closely with the endothelium. Pericytes play a key role in vasculature formation and homeostasis, and when dysfunctional contribute to vasculature-related diseases such as diabetic retinopathy and neurodegenerative conditions. In addition, significant extravascular roles of pathological pericytes are being discovered with relevant implications for cancer and fibrosis. Pericyte research is challenged by the lack of consistent molecular markers and clear discrimination criteria versus other (mural) cells. However, advances in single-cell approaches are uncovering and clarifying mural cell identities, biological functions, and ontogeny across organs. We discuss the latest developments in pericyte pathobiology to inform future research directions and potential outcomes.

Keywords

1. pericyte
2. vascular disease
3. pericrine signaling
4. fibrosis

Multifaceted roles of mural cells in health and disease

Pericytes are classically defined as mural cells (see Glossary) that envelop the endothelium of small caliber blood vessels, the so-called capillaries. Pericytes are embedded within the same basement membrane as endothelial cells (ECs) and interact closely with them [1,2]. By contrast, vascular smooth muscle cells (vSMCs), the other mural cell type, cover large arteries and veins, and are physically separated from the endothelium by an intimal layer of extracellular matrix (ECM). Of note, lymphatic capillaries lack pericytes under physiological conditions, although collecting lymphatic vessels contain vSMCs [3].

A fundamental function of mural cells is to regulate the stabilization and function of blood vessels. It is therefore not surprising that pericyte loss and dysfunction were linked to several diseases including cancer and cerebrovascular diseases more than a decade ago [4,5]. However, pericyte-focused therapies have been poorly explored. Instead, most studies on vascular-directed therapeutic strategies have been on ECs – the central components that build blood vessels. Emerging data are, however, changing the perception of pericytes from mere supporting vascular cells that are recruited at the final stage of vessel formation to essential elements in the early phases of angiogenesis that anticipate and orchestrate EC behavior. In addition, recent research is revealing novel pathological roles for pericytes beyond their implications in the vasculature. Collectively, we believe that these data open exciting avenues for pericyte-focused therapeutic approaches and call for a broader understanding of these cells in disease progression.

We provide here a global overview of recent significant advances regarding our understanding of the role of pericytes in different pathobiological scenarios and discuss the field's current paradigms and controversies. First, we address new insights into the functions associated with pericytes during physiological vascular responses. Second, we discuss evidence supporting a role of pericytes in disease, including pericyte cell-autonomous implications beyond the vasculature. For comprehensive details on pericyte biology, function ontology, and specific signaling pathways, we refer the reader to [1,2,5]. Of importance, some of the emerging concepts in pericyte biology described in the following sections have only been studied in one specific tissue. To avoid confusion about the generalizability of pericyte properties, we frame each function by considering the relevant organ of study.

Key concepts about pericytes in physiologyPericytes: a particular subtype of mural cells

Pericytes exhibit significant inter- and intra-tissue molecular differences and exert tissue-specific functions [2]. Their molecular, morphological, and functional heterogeneity is inextricably linked to their diverse developmental origins, modes of vessel recruitment, and specific anatomical localization. For example, pericytes of the central nervous system (CNS) microvasculature are firmly and continuously invested around the endothelium to support vascular barrier properties, whereas liver pericytes, commonly referred to as hepatic stellate cells (HSCs), reside in the perisinusoidal space, are loosely and discontinuously associated to ECs, and hold a unique vitamin A storage capacity [2]. To meet tissue-specific demands, pericyte distribution and density are variable among organs and vascular beds, with the CNS microvasculature showing the greatest pericyte-to-EC abundance. From a molecular standpoint there is no single molecular marker that can exclusively identify pericytes (Box 1), albeit the emergence of single-cell techniques is shedding light on tissue-specific pericyte molecular markers and functions. For example, the first molecular atlas of vascular cell types in the brain of adult mice by single-cell RNA sequencing (scRNA-seq) revealed that mural cells follow a gradient of transitional phenotypes. This gradient occurs at the interface of precapillary arterioles, capillaries, and postcapillary venules, and does not follow a single continuum along the arteriovenous axis (Figure 1 and Box 1) [6]. Whether this gradient of transitional phenotypes is specifically restricted to the brain vasculature or is also present in other vascular beds remains to be determined. Indeed, pericytes exhibit many organotypic differences in the expression‎ of molecular markers (Figure 2 illustrates three top-ranked pericyte markers with enriched expression‎ per organ), of which the expression‎ of transporters and components of the contractile machinery exhibit the greatest differences between organs [7]. Another intriguing observation is that pericytes exhibit more cross-organ heterogeneity than vSMCs [7,8]. Currently, the inter-tissue differences in the behavior of the two main mural cell types are not completely understood. However, this may be because pericytes exhibit a greater cell-intrinsic plasticity to adapt their molecular portfolio and function to tissue-specific demands, whereas vSMCs fulfill a more universal function across tissues. In contrast to the tissue-specific transcriptomic differences, the expression‎ of transcription factors appears to be relatively conserved in mural cells across organs, thereby suggesting that mural cell subtypes are defined by epigenetic mechanisms [7]. Accordingly, DNA hypermethylation was recently found to control alpha smooth muscle actin (αSMA) expression‎ in renal mural cells after ischemia [9]. This indicates that methods such as assay for transposase-accessible chromatin sequencing (ATAC-seq) will be instrumental to further understand mural cell phenotypes.

Figure 1 Schematic representation of mural cell zonation in the adult mouse brain.

Show full captionFigure viewer

Figure 2 Organotypic heterogeneity of pericyte markers.

Show full captionFigure viewer

Box 1

Unraveling the identity of pericytes

The identification of pericytes remains a challenging task. Despite ongoing efforts, there is no consensus regarding unambiguous criteria for pericyte identification. To date no single molecular marker can exclusively identify all pericytes or distinguish pericytes from other cell types, although scRNA-seq is now providing new opportunities to discern pericyte marker heterogeneity and tissue specificity [6,8,71,93]. The use of transgenic reporter mouse models has been instrumental to label, trace, and locate different mural cell populations in vivo. A combination of multiple reporter lines is often necessary to properly identify and discriminate pericytes from endothelial cells (ECs) and other perivascular cells [6–8]. Mural cells are highly plastic cells; phenotypic zonation of mouse brain mural cells has revealed that these cells do not follow a single continuum along the arteriovenous axis (see Figure 1A,B in main text) [6]. From a transcriptional point of view, there are two distinct continuums of mural cells: (i) capillary pericytes and venous smooth muscle cells (SMCs), where pericytes gradually transition to a venous SMC phenotype, and (ii) arterial SMCs which transition in an distinct pattern towards arteriole SMCs. The transcriptional resemblance between mouse brain pericytes and venular mural cells [6], as well as the lack of classic pericytes in several organs [7,8], have led to the hypothesis that capillary pericytes are transcriptionally and morphologically similar to venous SMCs in some tissues. Human brain mural cells recapitulate the mouse zonation pattern, although human pericytes are evenly distributed over capillaries and veins [50,94]. Unlike the anatomical separation of pericytes and venous SMCs in the mouse brain, subtypes of human pericytes are discerned by functionality marked by solute transport and extracellular matrix (ECM) organization [50]. Unfortunately, the ability of mouse markers to predict the presence of human pericytes remains limited, and only a select few retain adequate specificity. The use of zebrafish models may provide a better alternative to study conserved pericyte genes [95]. We believe that RGS5, NDUFA4L2, KNCJ8, HIGD1B, ABCC9, NOTCH3, and PDGFRB are currently the most organ and species conserved pericyte markers, although detailed intra-tissue characterization remains necessary when studying pericytes (see Figure 2 in main text).

Pericytes at play during vascular growth

Many studies have documented that pericytes contribute to angiogenesis [10]. The historical view proposes that pericytes mainly contribute to the late stages of vessel formation [2,10]. By taking advantage of the mouse retina as a paradigmatic experimental model of developmental angiogenesis, this concept has been challenged [11–16]. Indeed, these studies showed that, during the early phases of developmental angiogenesis, pericytes, which have not yet achieved the maturity seen in formed vessels, are permissive to cell-cycle progression, morphological adaptation, and migration [12,13]. In this setting, pericyte growth precedes the expansion of ECs, although it is still unclear why. One possibility is that, by expanding rapidly, pericytes ensure the production of sufficient EC growth signals, a hypothesis which is coherent with the observation that inhibition of pericyte activation blocks EC proliferation [12] and induces nuclear translocation of FOXO1 [11], the master regulator of EC quiescence [17]. Another study that examined the brain vasculature showed that, when pericytes are absent, ECs become angiogenic but are not able to proliferate [18], thereby supporting a model in which ECs require the presence of pericytes to expand. Nonetheless, it is fair to acknowledge that other studies have shown that reduced pericyte coverage leads to increased EC proliferation [19]. Although these discrepancies highlight that pericyte–EC interactions are complex, they may be explained by the differences between the animal models and genetic strategies used to interfere with pericytes. Importantly, pericyte behaviors during angiogenesis have been mostly described in tissues belonging to the CNS. Hence, given the high abundance of pericytes in the CNS, it is possible that angiogenic pericytes fulfill different roles in tissues where ECs substantially outnumber them. Another interesting observation is that, during angiogenesis, immature pericytes remain in close contact with ECs, although they do not cover them in their entirety [12,20]. This suggests that pericyte–EC communication during angiogenesis relies on both paracrine and juxtracrine signaling, and may explain why pericyte loss [11,16,21,22] and impaired transition to a fully maturate state [12] lead to distinct endothelial phenotypes during angiogenesis. scRNA-seq analysis of prenatal developing human brains confirmed that angiogenesis is supported by immature mural cells [20]. Consistent with mouse data [12], the state of mature human pericytes correlates with the progression of angiogenesis. Furthermore, the gene expression‎ profiles of these cells show involvement in processes related to the transport across the blood–brain barrier (BBB) and the synthesis of ECM components [20]. Together, these data support a model in which pericytes modulate the early phases of angiogenesis by directly regulating EC behavior. Intriguingly, however, detailed ultrastructural analysis of angiogenic vessels in human brain distinguishes only a single mural cell population, compared to three distinct EC populations [20]. This suggests that ultrastructural features do not define subtype specification in the mural cell compartment, and that molecular and structural features are not necessarily associated with each other.

Brain pericytes and vessel contraction: a matter of transitional phenotypes

Although the regulation of vascular tone through pericyte contractility is considered to be an important function of cardiac, renal, and pulmonary pericytes, as well as of HSCs [2], there has been a long-standing debate in the field as to whether pericytes actively modulate cerebral blood flow [23–25]. For instance, by using optical imaging, Hill et al. suggested that neural/glia antigen 2-positive (NG2+) αSMA− pericytes are not contractile and do not actively modulate the capillary diameter [26]. Instead, by similar optogenetic approaches Hartmann et al. proposed that pericytes do constrict, although they require prolonged and more intense stimulation than αSMA+ mural cells located at larger vessels [27]. Although no consensus has been established, the opposing results between studies may simply reflect heterogeneities in the type of blood vessels and mural cells analyzed. A recent report has shown that NG2+αSMA+ mural cells, located at the transitional segment between arteries and capillaries, regulate the vascular tone and contractility [28]. This suggests that the transition of functional phenotypes between mural cells covering distinct types of blood vessels is tightly regulated.

scRNA-seq analysis of brain mural cells has revealed an abrupt change in the molecular signatures of pericytes and mural cells located in arteries, even from cells residing in proximity on the vasculature, thereby supporting the existence of a blunt transition [6]. Taken together, one can speculate that, in addition to defined vSMC types, there is a subtype of mural cells that exhibit some traits, but not all, of classic pericytes, and are located at transitional vessels and can modulate the vascular tone. Given the ability of pericytes to adapt their phenotype to various microenvironmental conditions [1,2], it is also possible that regulation of blood flow may only occur under specific circumstances. However, one should consider that some of the data disputing pericyte contractility may relate to experimental artefacts, and it should be stressed that most analyses were conducted in the cerebral vasculature as a prototypical example of a vascular bed that is highly sensitive to contraction [25]. An important observation is that pericytes exhibit significant organotypic differences in the basal expression‎ of contractility genes, and pericytes in the bladder and colon express considerable levels of Myh11, Tagln, and Acta2 (αSMA), whereas pericytes in the brain, lung, and heart express negligible amounts of these contractile genes [6,7]. This highlights a conundrum regarding how brain pericytes regulate vessel contractility when typical contractility genes are not expressed.

Pericyte safeguarding the capillary brain bed by a special touch

An essential function of pericytes is to regulate the BBB by controlling the passage of fluid and substances into the parenchymal space [22,29]. Hence, defective pericyte coverage caused by pericyte dysfunction, impaired pericyte recruitment, and pericyte loss all lead to increased EC transcytosis and permeability [22,29]. Aberrant platelet-derived growth factor B (PDGF-B)/platelet-derived growth factor receptor beta (PDGFRβ) signaling is sufficient to experimentally reduce pericyte abundance and the subsequent loss of BBB properties [22,29]. In addition, proper ECM deposition by pericytes (among other cell types composing the neurovascular unit) plays an essential role in maintaining the integrity of the vascular barrier. Indeed, pericyte-derived vitronectin prevents endothelial transcytosis by binding to integrin α5 subunit on ECs [30], and pericyte-secreted laminin interacts with the dystrophin–glycoprotein complex in astrocytes and regulates their endfeet polarization [5,31].

To serve as guardians of the capillary bed, pericytes also establish physical interactions with ECs and form a continuous chain-like network along the capillaries of the cerebral vasculature. Adequate coverage of the endothelium is sustained by active remodeling of distal pericyte processes through cytoskeletal rearrangements [32]. Of relevance, pericyte remodeling capabilities become exhausted with age [33], and this may explain why pericyte coverage is diminished in the vasculature of old mice [33,34]. An interesting observation is that pericyte depletion in adult mice leads to relatively mild BBB defects in different experimental models [35,36]. This includes adult induced Pdgfb ablation [36] and diphtheria toxin A (DTA) expression‎ in PDGFRβ+ cells [35]. Currently it is not clear why loss of pericytes leads to different vascular barrier phenotypes in development and adulthood. Given that the BBB becomes functional during late embryonic development, one can speculate that defects in pericyte coverage are only significant before the onset of BBB formation. Another possibility is that pericyte coverage determines the threshold for BBB defects, and Vazquez-Liebanas et al. showed that only <50% longitudinal pericyte coverage in adult brains leads to significant leakage defects [36]. This is coherent with previous observations of brain vessel phenotypes during development which demonstrated that pericyte coverage is positively correlated with BBB integrity [22]. Choe et al. also reported that DTA-induced loss of pericytes leads to capillary stalling due to increased interactions between ECs and leukocytes. However, because this effect was not observed in other adult pericyte depletion models [35], one should acknowledge that it is possible that the expression‎ of DTA generated unintended toxic effects beyond pericytes.

Pericytes in disease

Pericyte dysfunction is a hallmark of various diseases (Figure 3). For a long time it was believed that maladaptive pericytes mainly affect vascular homeostasis because pericyte and EC functions are interdependent and require bidirectional communication (Box 2). However, there is growing evidence that pericytes have roles in processes beyond the vasculature. As such, pericyte-derived signals (hereafter referred to as pericrine signaling) also modulate tissue function in both physiology and disease. In the following section, we capture recent data showing new observations that link pericyte dysfunction and loss in vascular and non-vascular-related diseases.

Figure 3 Dysfunctional pericytes in disease.

Show full captionFigure viewer

Box 2

Key signaling pathways that orchestrate pericyte–EC crosstalk

Given the close relationship between pericytes and ECs, it is not surprising that bidirectional communication and regulation between them are crucial during vessel formation and maintenance. During angiogenesis, established examples of pericyte–EC communication include the PDGFRβ, transforming growth factor β1 (TGF-β1), ANG1, and NOTCH3 pathways [1]. PDGF-B production from tip ECs is the master signal that recruits PDGFRβ-expressing pericytes to newly formed vessels [1], together with CD146 (MCAM), which acts as a coreceptor for PDGFRβ [96]. Recent advances have shown that NCK1 and NCK2 promote phosphorylation of PDGFRβ in response to PDGF-BB and stimulate pericyte migration by inducing MRTF translocation to the nucleus where they interact with the serum response transcription factor (SRF) [13,21]. Similarly, jagged 1 (JAG1) expressed by ECs activates NOTCH3 in pericytes and promotes pericyte maturation [14,97] and the expression‎ of PDGFRβ [98]. Conversely, ANG1 is secreted by pericytes, activates the tyrosine receptor TIE2 in ECs, and promotes EC maturation and vascular integrity [2,15]. TGF-β exerts complex effects on ECs and pericytes, and TGF-B receptor 1 (also known as ALK5) plays a dominant role in these interactions. Indeed, deletion of ALK5 in ECs leads to pericyte dysfunction and hemorrhagic vascular malformations [99]. Instead, deletion of ALK5 in pericytes results in increased EC proliferation, reduced collagen deposition, and enhanced matrix metalloproteinase activity [19]. Of note, pericytes also express canonical EC receptors such as VEGF-R1 [15,16] and TIE2 which allow pericytes to modulate intrinsic EC signaling.

The CNS: a hotspot of pericyte-related vascular diseases

Pericyte-related vascular defects have been reported in various CNS diseases, including Alzheimer's disease (AD), Parkinson's disease, dementia, stroke, diabetic retinopathy, glaucoma, and intracranial vascular malformations [5,37–39]. The involvement of pericytes in several CNS-related diseases is partially explained by their abundance within the brain vasculature and their key role in maintaining the BBB, where barrier breakdown precedes neurodegeneration. Other phenotypes linking pericytes dysfunction and CNS disease include neuron death [40] and impaired neurovascular coupling [41,42]. Intriguingly, NG2+ retinal pericytes orchestrate neurovascular coupling through closed-ended nanotubes between pericytes on adjacent capillaries, even when they are positioned far apart. These nanotubes terminate in a gap junction at the recipient pericyte, which permits rapid fluxes of small molecules and calcium ions, thereby allowing pericytes to coordinate neuronal activity [41]. Indeed, maintaining adequate calcium levels is essential to sustain pericyte function in the CNS, and aberrant levels of calcium in NG2+ pericytes lead to poor recovery after ischemic stroke [23,43] or neovascular dysfunction and neuronal death in glaucoma [42].

AD is the prototypical example of a CNS disease associated with aberrant vascular function and BBB breakdown linked to pericyte dysfunction and loss [38,44]. Although the involvement of pericytes in AD has been recognized for several years [5,44], new insights have challenged the timeframe in which patients suffering from AD develop pericyte dysfunction and BBB impairment. Indeed, it is now understood that BBB breakdown is an early event in AD, and these defects are used as an early biomarker of cognitive decline [45]. We highlight recent observations which support the involvement of pericytes in the onset of AD. For instance, Nortley et al. showed that the reduction in cerebral blood flow, that is considered to be the first clinical manifestation of AD, is caused by amyloid-β-induced pericyte contraction in brain capillaries [46]. Another study indicated that cognitive decline and BBB disruption in AD are linked to accelerated pericyte degeneration in carriers of AD susceptibility allele apolipoprotein E4 (APOE4) [47], a process which occurs independently of amyloid-β pathology. In this context, APOE4 carriers show high baseline cerebrospinal fluid levels of soluble (s)PDGFRβ which can be used as a BBB pericyte injury biomarker [47]. Intriguingly, analysis of the cortex of APOE4 transgenic mice using single-nucleus (sn)RNA-seq and phosphoproteomics revealed profound molecular changes related to progressive BBB failure in both ECs and pericytes [48]. Nonetheless, because only a constitutive APOE4-expressing transgenic line was included in the study, it remains unclear whether the molecular alterations of ECs and pericytes solely comprise cell-autonomous effects. In addition, one should not forget that mice do not fully recapitulate all traits of AD. It has been recently noted that pericytes and microglia associations (described in both physiological mouse and human brains) are diminished in the brain capillaries of individuals with AD, and this may also have implications for BBB breakdown [49]. In human brain, two types of pericytes have been identified that are distinguished by solute transport and ECM organization (Box 1). Intriguingly, the second type seems to be selectively affected in AD [50]. Thus, identifying methods to specifically target this cluster of pericytes may provide new ways to maintain vascular fitness in AD.

Pericyte degeneration and death also encompasses early phases of diabetic retinopathy, in which pericytes are primary targets of hyperglycemic damage. Recent findings suggest that, upon initiation of hyperglycemia, pericytes shift towards cell-bridging positions, resulting in physical detachment from ECs [51]. Whether this remodeling is independent of pericyte death or is related to the initiation of that process needs further investigation. Mechanistically, pericyte detachment and shifting are induced by exogenous factors such as angiopoietin 2 (ANG2) and PDGF-B, and are reversed by insulin treatment, illustrating the dynamic behavior of pericytes in the microvasculature [51,52]. In line with this, PDGFB signaling through PDGFRβ and NCKs in pericytes that cover sprouting vessels during experimental proliferative retinopathy [oxygen-induced retinopathy (OIR) model] activates ectopic αSMA expression‎ and promotes pathological neovascularization [21]. Interestingly, depletion of retinal pericytes in adulthood does not phenocopy retinopathy unless another stimulus is present (e.g., vascular endothelial growth factor A, VEGF-A). Upon depletion of pericytes, either during vessel development or in adulthood followed by VEGF addition, inhibition of ANG2 action restrains the severity of the diabetic retinopathy-like phenotypes [11]. Molecular effectors governing the early phases of diabetic retinopathy have remained elusive, precluding the development of drugs aiming to halt disease onset. These data suggest that targeting pericyte adhesion and migration capacities may be of therapeutic interest. Furthermore, pericyte loss in diabetic retinopathy was recently associated with aberrant levels of circular RNAs [53], thereby suggesting the use of circular RNAs as a diagnostic biomarker for early pericyte dysfunction in disease.

Finally, we would like to stress that familial mutations in essential pericyte genes have also been linked to CNS abnormalities. Well-known examples include loss-of-function mutations in NOTCH3 as a cause of CADASIL [54], and mutations in PDGFRB as a cause of brain calcifications [55], neurological deterioration, and white matter lesions [56]. Of note, these genes are equally relevant for pericyte and vSMC biology, and it is unclear whether these mutations lead to distinct phenotypes in mural cells. Current next-generation sequencing approaches allow the discovery of somatic mutations present in pericytes at low allelic frequency. In line with this, it has been proposed that PIK3CA- and AKT-related somatic cerebral cavernous malformations in mice emerge from mutant pericytes [39,57]. However, these data have some caveats because the lineage-tracing experiments used to support these findings were performed with a CRE-recombinase mouse line that is neither pericyte-specific nor inducible.

Pathobiological pericytes beyond the CNS

Although the implications of pericytes in diseases beyond the CNS are less well studied, the number of diseases demonstrating the involvement of pericytes continues to expand. We discuss here emerging evidence supporting a relevant role of pericytes in myocardial infarction [58], acute lung injury [59], and diabetes [60] as prototypical examples. For instance, after myocardial infarction, pericytes regulate inflammation and immune cell trafficking, and modulate ECM remodeling and revascularization [61]. In line with this, molecular reprogramming of PDGFRβ+NG2+ cardiac pericytes into vSMCs through inhibition of MEK1/2 improved the functional cardiac response by promoting revascularization [58]. In acute lung inflammation, the crosstalk between endothelium-derived nitric oxide (NO) and pericyte soluble guanylate cyclase (sGC) is impaired, leading to elevated vascular permeability [59]. Pharmacological activation of the NO–sGC axis led to an improved pericyte-driven inflammatory response. Moreover, pericytes in pancreatic islets exert vascular control of hormone secretion and glucose homeostasis, and pericyte alteration has been linked to diabetic islet dysfunction [60]. Interestingly, pericyte effects on islet functionality are not limited to vascular support for insulin secretion because pancreatic β cell maturation and functionality rely on pericyte-derived bone morphogenic protein 4 (BMP4). Recently, other pericrine signaling molecules have been identified as key players in the functional regulation of tissue parenchyma encompassing a range of organ-specific functions in both vascular and non-vascular interfaces. Two interesting examples are that leptin receptor-expressing pericytes in the mediobasal hypothalamus mediate energy balance via neuronal leptin signaling [62], and that the Hippo–YAP/TAZ pathway in pericytes generates essential pericrine signals to epithelial and ECs during lung morphogenesis [63]. All things considered, these studies suggest that restoring the physiological functions of pericytes improves blood vessel performance and disease outcomes in various contexts, which may encourage the development of novel pericyte-focused therapies.

Tumor pericytes: loss or change of identity?

Many preclinical studies over the past decade showed that pericyte dysfunction is involved in cancer progression [64]. In this context, it is now well established that tumoral vessels are poorly covered by pericytes [65,66], that diminished pericyte recruitment and maturation lead to enhanced vascular and tumor growth [15,66], and that pericyte depletion favors metastasis. However, there is no consensus about whether defective pericyte–EC interaction in tumors relates to loss of pericyte molecular identity [67], pericyte transdifferentiation into fibroblasts [68], poor pericyte recruitment [66], or a combination of these phenotypes. Functionally, metabolic reprogramming of tumor pericytes has also been implicated in abnormal blood vessel contraction and unfavorable patient outcomes [69].

scRNA-seq analyses have provided a new layer of information about tumor pericytes by showing that tumor pericytes are, in fact, relatively homogeneous [70–72]. However, one should acknowledge that most tumor scRNA-seq datasets include low numbers of pericytes, and the presence of tumor pericyte subclusters may have been obscured by the granularity of the data. One way to overcome this limitation would be to establish spatially resolved pericyte-focused atlases by using multiomic approaches. An interesting observation is that pericyte phenotypes are distinct depending on the mechanisms by which tumor vessels form. While angiogenic pericytes persist in an active immature state characterized by a signature of motility and ECM organization, pericytes covering co-opted vessels remain largely quiescent [71,73]. Together, these studies support a model in which pericytes undergo genetic and molecular reprogramming in cancer, which in turn has negative consequences for disease progression. Nonetheless, one should not forget that most of these data refer to mouse preclinical studies, and adequate longitudinal studies in humans are missing. Of note, the so-called pericrine signaling response is also at play in cancer. Indeed, in hepatocellular carcinoma it has been shown that metabolic reprogramming in tumor cells activates HSCs which in turn promote tumorigenesis through the secretion of senescence-associated factors [74]. Another study has described that loss of integrin β3 in tumor pericytes leads to enhanced focal adhesion kinase (FAK)-mediated cytokine release by pericytes, which subsequently stimulates tumor survival and growth [75].

Pericyte immunomodulatory properties

Emerging evidence suggests that pericytes form an integral part of the immune surveillance unit rather than solely performing complementary functions. Upon proinflammatory stimuli, pericytes promote endothelial expression‎ of the leukocyte adhesion molecules vascular cell adhesion protein 1 (VCAM-1) and/or intracellular adhesion molecule 1 (ICAM-1) in the CNS, lung, skin, or muscle that subsequently promote T cell or macrophage infiltration into the affected tissue [59,76,77]. The Rgs5+ and Col1a1+ subgroups of PDGFRβ+ perivascular cells seem to be early responders to neuroinflammation [78]. Of note, inflammation per se induces pericyte detachment from the endothelium and impairs barrier properties [59]. In addition to the physical interaction with leukocytes, pericytes also secrete and respond to cytokines that further regulate immune cell functions, including both innate and adaptive responses. The chemotactic migration and effector functions of neutrophils, T cells, and macrophages are dependent on these early pericrine signals, including MIF, CXCL1, and CCL2 [77–80].

It is now believed that modulation of the immune-related functions of pericytes could affect the outcome of disease progression. For example, pericyte-deficient Pdgfbret/ret mice exhibit increased leukocyte infiltration and activation, leading to aberrant inflammation in a model of experimental autoimmune encephalomyelitis [76]. Treatment with antagonistic VCAM-1 and ICAM-1 antibodies partially rescued the excessive inflammatory phenotype in Pdgfbret/ret mice. Similarly, activating sGC in acute lung injury improved disease outcomes by increasing pericyte interaction with ECs [59]. Conversely, in the tumor microenvironment, tumor cells induce autophagy of NG2+ pericytes which equips them with immunosuppressive properties that favor tumor cell survival and prevent antitumor T cell responses [81]. Pericytes have also been implicated in the underlying pathophysiology of emergent infectious diseases such as coronavirus disease 2019 (COVID-19) [82], thus presenting an additional niche of investigation for the field in the coming years. In conclusion, the importance of pericytes during various inflammatory processes is growing in recognition, although the modulation of immunity by pericytes can have double-edged outcomes.

Do pericytes contribute to fibrosis?

A dysregulated tissue repair response after acute or chronic injury can lead to the onset of fibrosis associated with the abnormal accumulation of activated and contractile αSMA+ myofibroblasts [83]. Myofibroblasts secrete high amounts of inflammatory mediators, growth factors, and ECM components, and promote aberrant ECM remodeling. The current consensus places fibroblasts as the predominant origin of myofibroblasts, although various studies have found alternative cellular origins [83]. Indeed, the existence of pericyte-to-myofibroblast transition has been proposed as a contributing factor in several fibrotic contexts [2], but the promiscuity of cell markers within the mesenchymal compartments across tissues has led to ambiguous and contradictory observations. In the latest developments, scRNA-seq, ATAC-seq, and spatial transcriptomics provide new insights into this conundrum which support a pericyte origin of myofibroblasts in the fibrotic liver, colon, and kidney. For instance, central vein-associated Rgs5+ HSCs are thought to be the dominant origin of ECM-producing myofibroblasts in fibrotic mouse liver [84]. In human colorectal cancer, a subset of periostin (POSTN)+ myofibroblasts seem to originate from RGS5+ pericytes [70]; similarly, NOTCH3+RGS5+PDGFRβ+ human pericytes contribute to the generation of POSTN+PDGFRα+NKD2+ myofibroblasts during kidney fibrosis [85]. Surprisingly, however, the same authors did not capture the existence of profibrotic pericytes during myocardial infarction when sequencing the entire heart [86]. Although transdifferentiation from pericytes to myofibroblasts may be tissue-specific, it is fair to acknowledge that the latter study did not include pericytes in the trajectory analysis that predicted the origins of myofibroblasts.

Lineage-tracing experiments in mice have also cast some light onto the role of pericytes in fibrosis. For instance, Pham et al. showed that myofibroblast genes are enriched in Tbx18+ pericytes from injured mouse hearts and brains [87], and Dias et al. reported that GLAST+PDGFRβ+ perivascular cells also contribute to fibrosis in the post-stroke brain [88]. In line with this, depletion of GLAST+PDGFRβ+ perivascular cells in the spinal cord leads to reduced fibrotic scar after injury and improves neuronal function [89]. Of note, GLAST+PDGFRβ+ perivascular cells were characterized as spinal cord pericytes by the researchers, but there is insufficient evidence to rule out that these cells might be fibroblasts or astrocytes. This further exemplifies that the shared marker expression‎ profiles of pericytes and other perivascular residing cells still hamper the design of robust pericyte reporter models. Moreover, multiple studies have found no significant pericyte origin for myofibroblasts in distinct fibrosis models of the CNS [90,91] and heart [92]. Although the discrepancies between studies may reflect the lack of robust pericyte identification strategies, it appears that the occurrence of pericyte-to-myofibroblast transition is tissue-specific, and is contingent both on the local microenvironment and on the extent of the injurious stimuli [83]. This emphasizes the need to further unravel the fibrotic pericyte responses in different organs and prompts the question of how myofibroblast origins relate to different pathophysiological phenotypes. All things considered, the origin of myofibroblasts may involve distinct precursor cells depending on the circumstances, although a role for pericytes seems to be indisputable (Figure 4).

Figure 4 Pericytes as a source of myofibroblasts in fibrosis.

Show full captionFigure viewer

Concluding remarks

It is widely recognized that pericytes play an important role in blood vessel formation, stabilization, and function, and that degeneration or loss of brain pericytes impairs their protective barrier properties. Recent advances have revealed novel and crucial roles for pericytes across tissues in a variety of vascular and non-vascular processes. Single-cell technology is becoming more commonly used to better understand the molecular processes that define pericytes in health and disease. Although the molecular and functional attributes of pericytes are not fully elucidated, deep sequencing has revealed organotypic pericyte heterogeneities and new criteria to distinguish pericytes from other cell types. Despite the numerous suggested roles of pericytes in various diseases and physiological processes, including neurodegeneration, cancer, fibrosis, blood flow regulation, and inflammation, the underlying organotypic mechanisms of these contributions are not yet fully understood. The discrepancies between some studies highlight the importance of designing suitable mouse models for eval‎uating the specific mechanisms by which pericytes impact on these processes, and future cross-validations with human data are warranted to ascertain the clinical relevance of pathological pericytes (see Outstanding questions). Overall, based on the emerging evidence on the contribution of pericytes to several diseases, we anticipate an increasing emphasis on pericyte-oriented research in vascular (and non-vascular) studies in the coming years.

Outstanding questions

Why are pericytes molecularly and functionally promiscuous and heterogeneous between distinct tissues? If the transcription factors that regulate pericyte differentiation and function are similar across tissues, will epigenetic mechanisms provide cues into the mechanisms of pericyte identity?

Will spatial molecular atlases focused on pericytes address the key functional and identity conundrums posed by transitioning phenotypes? Will this suffice, or are complementary morphological and functional studies also necessary?

Will pericyte-focused therapy provide new means to stimulate functional angiogenesis in pathology? Given that pericytes are associated with many diseases, will (and how broadly can) pericyte-focused therapies improve patient outcomes?

Neurodegenerative diseases are age-related diseases, and pericyte degeneration is an aging process. Hence, is pericyte degeneration a confounding factor in the development of age-related neurodegeneration? Will maintaining healthy pericyte function promote healthy aging?

AcknowledgmentsAcknowledgments

We would like to thank Sandra D. Castillo, Ana Angulo-Urarte, and Leonor Gouveia for their valuable feedback. Figures were created with BioRender.com. We thank the Centres de Recerca de Catalunya (CERCA) Program/Generalitat de Catalunya and the Josep Carreras Foundation for institutional support. Work-related to this publication in the laboratory of M.G. is supported by research grants from la Asociación Española Contra el Cancer (AECC)-Grupos Traslacionales (GCTRA18006CARR); and by Worldwide Cancer Research (WCR 21-0159). H.v.S. received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 955951; P.V. recieved funding from AECC (AECC-INVES211084VILL) and from the Spanish Ministry of Science and Innovation (RYC2020-029929-I). We apologize to the many authors whose primary papers could not be cited owing to space constraints.

Declaration of interests

The authors declare no conflicts of interest.

References

1.

Armulik, A. ...

Pericytes: developmental, physiological, and pathological perspectives, problems, and promises

Dev. Cell. 2011; 21:193-215

Full Text

Full Text (PDF)

Scopus (1996)

PubMed

Google Scholar

2.

Holm, A. ...

Microvascular mural cell organotypic heterogeneity and functional plasticity

Trends Cell Biol. 2018; 28:302-316

Full Text

Full Text (PDF)

Scopus (85)

PubMed

Google Scholar

3.

Petrova, T.V. ∙ Koh, G.Y.

Biological functions of lymphatic vessels

Science. 2020; 369, eaax4063

Crossref

Scopus (234)

PubMed

Google Scholar

ReviewVolume 34, Issue 1p58-71January 2024Open access

Download Full Issue

Pericytes in the disease spotlight

Hielke van Splunder1,5 ∙ Pilar Villacampa2,5 ∙ Anabel Martínez-Romero1 ∙ Mariona Graupera1,3,4 mgraupera@carrerasresearch.org

Affiliations & NotesArticle Info

• Download PDF
• Cite

• Share

• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Highlights

Molecular and functional pericyte studies at single-cell resolution are providing new insights into long-standing questions about pericyte heterogeneity.

Pericytes are not identified by a single marker but instead by gene expression‎ signatures that show substantial inter-organ differences.

Pericytes orchestrate and precede endothelial cell responses during angiogenesis.

Pericyte degeneration and dysfunction, that are triggered by the onset of some diseases, contribute to the progression of those diseases in both vascular and non-vascular contexts.

The number of diseases with pericyte dysfunction continues to expand, thereby anticipating a promising future for pericyte-focused therapy.

Abstract

Pericytes are known as the mural cells in small-caliber vessels that interact closely with the endothelium. Pericytes play a key role in vasculature formation and homeostasis, and when dysfunctional contribute to vasculature-related diseases such as diabetic retinopathy and neurodegenerative conditions. In addition, significant extravascular roles of pathological pericytes are being discovered with relevant implications for cancer and fibrosis. Pericyte research is challenged by the lack of consistent molecular markers and clear discrimination criteria versus other (mural) cells. However, advances in single-cell approaches are uncovering and clarifying mural cell identities, biological functions, and ontogeny across organs. We discuss the latest developments in pericyte pathobiology to inform future research directions and potential outcomes.

Keywords

1. pericyte
2. vascular disease
3. pericrine signaling
4. fibrosis

Multifaceted roles of mural cells in health and disease

Pericytes are classically defined as mural cells (see Glossary) that envelop the endothelium of small caliber blood vessels, the so-called capillaries. Pericytes are embedded within the same basement membrane as endothelial cells (ECs) and interact closely with them [1,2]. By contrast, vascular smooth muscle cells (vSMCs), the other mural cell type, cover large arteries and veins, and are physically separated from the endothelium by an intimal layer of extracellular matrix (ECM). Of note, lymphatic capillaries lack pericytes under physiological conditions, although collecting lymphatic vessels contain vSMCs [3].

A fundamental function of mural cells is to regulate the stabilization and function of blood vessels. It is therefore not surprising that pericyte loss and dysfunction were linked to several diseases including cancer and cerebrovascular diseases more than a decade ago [4,5]. However, pericyte-focused therapies have been poorly explored. Instead, most studies on vascular-directed therapeutic strategies have been on ECs – the central components that build blood vessels. Emerging data are, however, changing the perception of pericytes from mere supporting vascular cells that are recruited at the final stage of vessel formation to essential elements in the early phases of angiogenesis that anticipate and orchestrate EC behavior. In addition, recent research is revealing novel pathological roles for pericytes beyond their implications in the vasculature. Collectively, we believe that these data open exciting avenues for pericyte-focused therapeutic approaches and call for a broader understanding of these cells in disease progression.

We provide here a global overview of recent significant advances regarding our understanding of the role of pericytes in different pathobiological scenarios and discuss the field's current paradigms and controversies. First, we address new insights into the functions associated with pericytes during physiological vascular responses. Second, we discuss evidence supporting a role of pericytes in disease, including pericyte cell-autonomous implications beyond the vasculature. For comprehensive details on pericyte biology, function ontology, and specific signaling pathways, we refer the reader to [1,2,5]. Of importance, some of the emerging concepts in pericyte biology described in the following sections have only been studied in one specific tissue. To avoid confusion about the generalizability of pericyte properties, we frame each function by considering the relevant organ of study.

Key concepts about pericytes in physiologyPericytes: a particular subtype of mural cells

Pericytes exhibit significant inter- and intra-tissue molecular differences and exert tissue-specific functions [2]. Their molecular, morphological, and functional heterogeneity is inextricably linked to their diverse developmental origins, modes of vessel recruitment, and specific anatomical localization. For example, pericytes of the central nervous system (CNS) microvasculature are firmly and continuously invested around the endothelium to support vascular barrier properties, whereas liver pericytes, commonly referred to as hepatic stellate cells (HSCs), reside in the perisinusoidal space, are loosely and discontinuously associated to ECs, and hold a unique vitamin A storage capacity [2]. To meet tissue-specific demands, pericyte distribution and density are variable among organs and vascular beds, with the CNS microvasculature showing the greatest pericyte-to-EC abundance. From a molecular standpoint there is no single molecular marker that can exclusively identify pericytes (Box 1), albeit the emergence of single-cell techniques is shedding light on tissue-specific pericyte molecular markers and functions. For example, the first molecular atlas of vascular cell types in the brain of adult mice by single-cell RNA sequencing (scRNA-seq) revealed that mural cells follow a gradient of transitional phenotypes. This gradient occurs at the interface of precapillary arterioles, capillaries, and postcapillary venules, and does not follow a single continuum along the arteriovenous axis (Figure 1 and Box 1) [6]. Whether this gradient of transitional phenotypes is specifically restricted to the brain vasculature or is also present in other vascular beds remains to be determined. Indeed, pericytes exhibit many organotypic differences in the expression‎ of molecular markers (Figure 2 illustrates three top-ranked pericyte markers with enriched expression‎ per organ), of which the expression‎ of transporters and components of the contractile machinery exhibit the greatest differences between organs [7]. Another intriguing observation is that pericytes exhibit more cross-organ heterogeneity than vSMCs [7,8]. Currently, the inter-tissue differences in the behavior of the two main mural cell types are not completely understood. However, this may be because pericytes exhibit a greater cell-intrinsic plasticity to adapt their molecular portfolio and function to tissue-specific demands, whereas vSMCs fulfill a more universal function across tissues. In contrast to the tissue-specific transcriptomic differences, the expression‎ of transcription factors appears to be relatively conserved in mural cells across organs, thereby suggesting that mural cell subtypes are defined by epigenetic mechanisms [7]. Accordingly, DNA hypermethylation was recently found to control alpha smooth muscle actin (αSMA) expression‎ in renal mural cells after ischemia [9]. This indicates that methods such as assay for transposase-accessible chromatin sequencing (ATAC-seq) will be instrumental to further understand mural cell phenotypes.

Figure 1 Schematic representation of mural cell zonation in the adult mouse brain.

Show full captionFigure viewer

Figure 2 Organotypic heterogeneity of pericyte markers.

Show full captionFigure viewer

Box 1

Unraveling the identity of pericytes

The identification of pericytes remains a challenging task. Despite ongoing efforts, there is no consensus regarding unambiguous criteria for pericyte identification. To date no single molecular marker can exclusively identify all pericytes or distinguish pericytes from other cell types, although scRNA-seq is now providing new opportunities to discern pericyte marker heterogeneity and tissue specificity [6,8,71,93]. The use of transgenic reporter mouse models has been instrumental to label, trace, and locate different mural cell populations in vivo. A combination of multiple reporter lines is often necessary to properly identify and discriminate pericytes from endothelial cells (ECs) and other perivascular cells [6–8]. Mural cells are highly plastic cells; phenotypic zonation of mouse brain mural cells has revealed that these cells do not follow a single continuum along the arteriovenous axis (see Figure 1A,B in main text) [6]. From a transcriptional point of view, there are two distinct continuums of mural cells: (i) capillary pericytes and venous smooth muscle cells (SMCs), where pericytes gradually transition to a venous SMC phenotype, and (ii) arterial SMCs which transition in an distinct pattern towards arteriole SMCs. The transcriptional resemblance between mouse brain pericytes and venular mural cells [6], as well as the lack of classic pericytes in several organs [7,8], have led to the hypothesis that capillary pericytes are transcriptionally and morphologically similar to venous SMCs in some tissues. Human brain mural cells recapitulate the mouse zonation pattern, although human pericytes are evenly distributed over capillaries and veins [50,94]. Unlike the anatomical separation of pericytes and venous SMCs in the mouse brain, subtypes of human pericytes are discerned by functionality marked by solute transport and extracellular matrix (ECM) organization [50]. Unfortunately, the ability of mouse markers to predict the presence of human pericytes remains limited, and only a select few retain adequate specificity. The use of zebrafish models may provide a better alternative to study conserved pericyte genes [95]. We believe that RGS5, NDUFA4L2, KNCJ8, HIGD1B, ABCC9, NOTCH3, and PDGFRB are currently the most organ and species conserved pericyte markers, although detailed intra-tissue characterization remains necessary when studying pericytes (see Figure 2 in main text).

Pericytes at play during vascular growth

Many studies have documented that pericytes contribute to angiogenesis [10]. The historical view proposes that pericytes mainly contribute to the late stages of vessel formation [2,10]. By taking advantage of the mouse retina as a paradigmatic experimental model of developmental angiogenesis, this concept has been challenged [11–16]. Indeed, these studies showed that, during the early phases of developmental angiogenesis, pericytes, which have not yet achieved the maturity seen in formed vessels, are permissive to cell-cycle progression, morphological adaptation, and migration [12,13]. In this setting, pericyte growth precedes the expansion of ECs, although it is still unclear why. One possibility is that, by expanding rapidly, pericytes ensure the production of sufficient EC growth signals, a hypothesis which is coherent with the observation that inhibition of pericyte activation blocks EC proliferation [12] and induces nuclear translocation of FOXO1 [11], the master regulator of EC quiescence [17]. Another study that examined the brain vasculature showed that, when pericytes are absent, ECs become angiogenic but are not able to proliferate [18], thereby supporting a model in which ECs require the presence of pericytes to expand. Nonetheless, it is fair to acknowledge that other studies have shown that reduced pericyte coverage leads to increased EC proliferation [19]. Although these discrepancies highlight that pericyte–EC interactions are complex, they may be explained by the differences between the animal models and genetic strategies used to interfere with pericytes. Importantly, pericyte behaviors during angiogenesis have been mostly described in tissues belonging to the CNS. Hence, given the high abundance of pericytes in the CNS, it is possible that angiogenic pericytes fulfill different roles in tissues where ECs substantially outnumber them. Another interesting observation is that, during angiogenesis, immature pericytes remain in close contact with ECs, although they do not cover them in their entirety [12,20]. This suggests that pericyte–EC communication during angiogenesis relies on both paracrine and juxtracrine signaling, and may explain why pericyte loss [11,16,21,22] and impaired transition to a fully maturate state [12] lead to distinct endothelial phenotypes during angiogenesis. scRNA-seq analysis of prenatal developing human brains confirmed that angiogenesis is supported by immature mural cells [20]. Consistent with mouse data [12], the state of mature human pericytes correlates with the progression of angiogenesis. Furthermore, the gene expression‎ profiles of these cells show involvement in processes related to the transport across the blood–brain barrier (BBB) and the synthesis of ECM components [20]. Together, these data support a model in which pericytes modulate the early phases of angiogenesis by directly regulating EC behavior. Intriguingly, however, detailed ultrastructural analysis of angiogenic vessels in human brain distinguishes only a single mural cell population, compared to three distinct EC populations [20]. This suggests that ultrastructural features do not define subtype specification in the mural cell compartment, and that molecular and structural features are not necessarily associated with each other.

Brain pericytes and vessel contraction: a matter of transitional phenotypes

Although the regulation of vascular tone through pericyte contractility is considered to be an important function of cardiac, renal, and pulmonary pericytes, as well as of HSCs [2], there has been a long-standing debate in the field as to whether pericytes actively modulate cerebral blood flow [23–25]. For instance, by using optical imaging, Hill et al. suggested that neural/glia antigen 2-positive (NG2+) αSMA− pericytes are not contractile and do not actively modulate the capillary diameter [26]. Instead, by similar optogenetic approaches Hartmann et al. proposed that pericytes do constrict, although they require prolonged and more intense stimulation than αSMA+ mural cells located at larger vessels [27]. Although no consensus has been established, the opposing results between studies may simply reflect heterogeneities in the type of blood vessels and mural cells analyzed. A recent report has shown that NG2+αSMA+ mural cells, located at the transitional segment between arteries and capillaries, regulate the vascular tone and contractility [28]. This suggests that the transition of functional phenotypes between mural cells covering distinct types of blood vessels is tightly regulated.

scRNA-seq analysis of brain mural cells has revealed an abrupt change in the molecular signatures of pericytes and mural cells located in arteries, even from cells residing in proximity on the vasculature, thereby supporting the existence of a blunt transition [6]. Taken together, one can speculate that, in addition to defined vSMC types, there is a subtype of mural cells that exhibit some traits, but not all, of classic pericytes, and are located at transitional vessels and can modulate the vascular tone. Given the ability of pericytes to adapt their phenotype to various microenvironmental conditions [1,2], it is also possible that regulation of blood flow may only occur under specific circumstances. However, one should consider that some of the data disputing pericyte contractility may relate to experimental artefacts, and it should be stressed that most analyses were conducted in the cerebral vasculature as a prototypical example of a vascular bed that is highly sensitive to contraction [25]. An important observation is that pericytes exhibit significant organotypic differences in the basal expression‎ of contractility genes, and pericytes in the bladder and colon express considerable levels of Myh11, Tagln, and Acta2 (αSMA), whereas pericytes in the brain, lung, and heart express negligible amounts of these contractile genes [6,7]. This highlights a conundrum regarding how brain pericytes regulate vessel contractility when typical contractility genes are not expressed.

Pericyte safeguarding the capillary brain bed by a special touch

An essential function of pericytes is to regulate the BBB by controlling the passage of fluid and substances into the parenchymal space [22,29]. Hence, defective pericyte coverage caused by pericyte dysfunction, impaired pericyte recruitment, and pericyte loss all lead to increased EC transcytosis and permeability [22,29]. Aberrant platelet-derived growth factor B (PDGF-B)/platelet-derived growth factor receptor beta (PDGFRβ) signaling is sufficient to experimentally reduce pericyte abundance and the subsequent loss of BBB properties [22,29]. In addition, proper ECM deposition by pericytes (among other cell types composing the neurovascular unit) plays an essential role in maintaining the integrity of the vascular barrier. Indeed, pericyte-derived vitronectin prevents endothelial transcytosis by binding to integrin α5 subunit on ECs [30], and pericyte-secreted laminin interacts with the dystrophin–glycoprotein complex in astrocytes and regulates their endfeet polarization [5,31].

To serve as guardians of the capillary bed, pericytes also establish physical interactions with ECs and form a continuous chain-like network along the capillaries of the cerebral vasculature. Adequate coverage of the endothelium is sustained by active remodeling of distal pericyte processes through cytoskeletal rearrangements [32]. Of relevance, pericyte remodeling capabilities become exhausted with age [33], and this may explain why pericyte coverage is diminished in the vasculature of old mice [33,34]. An interesting observation is that pericyte depletion in adult mice leads to relatively mild BBB defects in different experimental models [35,36]. This includes adult induced Pdgfb ablation [36] and diphtheria toxin A (DTA) expression‎ in PDGFRβ+ cells [35]. Currently it is not clear why loss of pericytes leads to different vascular barrier phenotypes in development and adulthood. Given that the BBB becomes functional during late embryonic development, one can speculate that defects in pericyte coverage are only significant before the onset of BBB formation. Another possibility is that pericyte coverage determines the threshold for BBB defects, and Vazquez-Liebanas et al. showed that only <50% longitudinal pericyte coverage in adult brains leads to significant leakage defects [36]. This is coherent with previous observations of brain vessel phenotypes during development which demonstrated that pericyte coverage is positively correlated with BBB integrity [22]. Choe et al. also reported that DTA-induced loss of pericytes leads to capillary stalling due to increased interactions between ECs and leukocytes. However, because this effect was not observed in other adult pericyte depletion models [35], one should acknowledge that it is possible that the expression‎ of DTA generated unintended toxic effects beyond pericytes.

Pericytes in disease

Pericyte dysfunction is a hallmark of various diseases (Figure 3). For a long time it was believed that maladaptive pericytes mainly affect vascular homeostasis because pericyte and EC functions are interdependent and require bidirectional communication (Box 2). However, there is growing evidence that pericytes have roles in processes beyond the vasculature. As such, pericyte-derived signals (hereafter referred to as pericrine signaling) also modulate tissue function in both physiology and disease. In the following section, we capture recent data showing new observations that link pericyte dysfunction and loss in vascular and non-vascular-related diseases.

Figure 3 Dysfunctional pericytes in disease.

Show full captionFigure viewer

Box 2

Key signaling pathways that orchestrate pericyte–EC crosstalk

Given the close relationship between pericytes and ECs, it is not surprising that bidirectional communication and regulation between them are crucial during vessel formation and maintenance. During angiogenesis, established examples of pericyte–EC communication include the PDGFRβ, transforming growth factor β1 (TGF-β1), ANG1, and NOTCH3 pathways [1]. PDGF-B production from tip ECs is the master signal that recruits PDGFRβ-expressing pericytes to newly formed vessels [1], together with CD146 (MCAM), which acts as a coreceptor for PDGFRβ [96]. Recent advances have shown that NCK1 and NCK2 promote phosphorylation of PDGFRβ in response to PDGF-BB and stimulate pericyte migration by inducing MRTF translocation to the nucleus where they interact with the serum response transcription factor (SRF) [13,21]. Similarly, jagged 1 (JAG1) expressed by ECs activates NOTCH3 in pericytes and promotes pericyte maturation [14,97] and the expression‎ of PDGFRβ [98]. Conversely, ANG1 is secreted by pericytes, activates the tyrosine receptor TIE2 in ECs, and promotes EC maturation and vascular integrity [2,15]. TGF-β exerts complex effects on ECs and pericytes, and TGF-B receptor 1 (also known as ALK5) plays a dominant role in these interactions. Indeed, deletion of ALK5 in ECs leads to pericyte dysfunction and hemorrhagic vascular malformations [99]. Instead, deletion of ALK5 in pericytes results in increased EC proliferation, reduced collagen deposition, and enhanced matrix metalloproteinase activity [19]. Of note, pericytes also express canonical EC receptors such as VEGF-R1 [15,16] and TIE2 which allow pericytes to modulate intrinsic EC signaling.

The CNS: a hotspot of pericyte-related vascular diseases

Pericyte-related vascular defects have been reported in various CNS diseases, including Alzheimer's disease (AD), Parkinson's disease, dementia, stroke, diabetic retinopathy, glaucoma, and intracranial vascular malformations [5,37–39]. The involvement of pericytes in several CNS-related diseases is partially explained by their abundance within the brain vasculature and their key role in maintaining the BBB, where barrier breakdown precedes neurodegeneration. Other phenotypes linking pericytes dysfunction and CNS disease include neuron death [40] and impaired neurovascular coupling [41,42]. Intriguingly, NG2+ retinal pericytes orchestrate neurovascular coupling through closed-ended nanotubes between pericytes on adjacent capillaries, even when they are positioned far apart. These nanotubes terminate in a gap junction at the recipient pericyte, which permits rapid fluxes of small molecules and calcium ions, thereby allowing pericytes to coordinate neuronal activity [41]. Indeed, maintaining adequate calcium levels is essential to sustain pericyte function in the CNS, and aberrant levels of calcium in NG2+ pericytes lead to poor recovery after ischemic stroke [23,43] or neovascular dysfunction and neuronal death in glaucoma [42].

AD is the prototypical example of a CNS disease associated with aberrant vascular function and BBB breakdown linked to pericyte dysfunction and loss [38,44]. Although the involvement of pericytes in AD has been recognized for several years [5,44], new insights have challenged the timeframe in which patients suffering from AD develop pericyte dysfunction and BBB impairment. Indeed, it is now understood that BBB breakdown is an early event in AD, and these defects are used as an early biomarker of cognitive decline [45]. We highlight recent observations which support the involvement of pericytes in the onset of AD. For instance, Nortley et al. showed that the reduction in cerebral blood flow, that is considered to be the first clinical manifestation of AD, is caused by amyloid-β-induced pericyte contraction in brain capillaries [46]. Another study indicated that cognitive decline and BBB disruption in AD are linked to accelerated pericyte degeneration in carriers of AD susceptibility allele apolipoprotein E4 (APOE4) [47], a process which occurs independently of amyloid-β pathology. In this context, APOE4 carriers show high baseline cerebrospinal fluid levels of soluble (s)PDGFRβ which can be used as a BBB pericyte injury biomarker [47]. Intriguingly, analysis of the cortex of APOE4 transgenic mice using single-nucleus (sn)RNA-seq and phosphoproteomics revealed profound molecular changes related to progressive BBB failure in both ECs and pericytes [48]. Nonetheless, because only a constitutive APOE4-expressing transgenic line was included in the study, it remains unclear whether the molecular alterations of ECs and pericytes solely comprise cell-autonomous effects. In addition, one should not forget that mice do not fully recapitulate all traits of AD. It has been recently noted that pericytes and microglia associations (described in both physiological mouse and human brains) are diminished in the brain capillaries of individuals with AD, and this may also have implications for BBB breakdown [49]. In human brain, two types of pericytes have been identified that are distinguished by solute transport and ECM organization (Box 1). Intriguingly, the second type seems to be selectively affected in AD [50]. Thus, identifying methods to specifically target this cluster of pericytes may provide new ways to maintain vascular fitness in AD.

Pericyte degeneration and death also encompasses early phases of diabetic retinopathy, in which pericytes are primary targets of hyperglycemic damage. Recent findings suggest that, upon initiation of hyperglycemia, pericytes shift towards cell-bridging positions, resulting in physical detachment from ECs [51]. Whether this remodeling is independent of pericyte death or is related to the initiation of that process needs further investigation. Mechanistically, pericyte detachment and shifting are induced by exogenous factors such as angiopoietin 2 (ANG2) and PDGF-B, and are reversed by insulin treatment, illustrating the dynamic behavior of pericytes in the microvasculature [51,52]. In line with this, PDGFB signaling through PDGFRβ and NCKs in pericytes that cover sprouting vessels during experimental proliferative retinopathy [oxygen-induced retinopathy (OIR) model] activates ectopic αSMA expression‎ and promotes pathological neovascularization [21]. Interestingly, depletion of retinal pericytes in adulthood does not phenocopy retinopathy unless another stimulus is present (e.g., vascular endothelial growth factor A, VEGF-A). Upon depletion of pericytes, either during vessel development or in adulthood followed by VEGF addition, inhibition of ANG2 action restrains the severity of the diabetic retinopathy-like phenotypes [11]. Molecular effectors governing the early phases of diabetic retinopathy have remained elusive, precluding the development of drugs aiming to halt disease onset. These data suggest that targeting pericyte adhesion and migration capacities may be of therapeutic interest. Furthermore, pericyte loss in diabetic retinopathy was recently associated with aberrant levels of circular RNAs [53], thereby suggesting the use of circular RNAs as a diagnostic biomarker for early pericyte dysfunction in disease.

Finally, we would like to stress that familial mutations in essential pericyte genes have also been linked to CNS abnormalities. Well-known examples include loss-of-function mutations in NOTCH3 as a cause of CADASIL [54], and mutations in PDGFRB as a cause of brain calcifications [55], neurological deterioration, and white matter lesions [56]. Of note, these genes are equally relevant for pericyte and vSMC biology, and it is unclear whether these mutations lead to distinct phenotypes in mural cells. Current next-generation sequencing approaches allow the discovery of somatic mutations present in pericytes at low allelic frequency. In line with this, it has been proposed that PIK3CA- and AKT-related somatic cerebral cavernous malformations in mice emerge from mutant pericytes [39,57]. However, these data have some caveats because the lineage-tracing experiments used to support these findings were performed with a CRE-recombinase mouse line that is neither pericyte-specific nor inducible.

Pathobiological pericytes beyond the CNS

Although the implications of pericytes in diseases beyond the CNS are less well studied, the number of diseases demonstrating the involvement of pericytes continues to expand. We discuss here emerging evidence supporting a relevant role of pericytes in myocardial infarction [58], acute lung injury [59], and diabetes [60] as prototypical examples. For instance, after myocardial infarction, pericytes regulate inflammation and immune cell trafficking, and modulate ECM remodeling and revascularization [61]. In line with this, molecular reprogramming of PDGFRβ+NG2+ cardiac pericytes into vSMCs through inhibition of MEK1/2 improved the functional cardiac response by promoting revascularization [58]. In acute lung inflammation, the crosstalk between endothelium-derived nitric oxide (NO) and pericyte soluble guanylate cyclase (sGC) is impaired, leading to elevated vascular permeability [59]. Pharmacological activation of the NO–sGC axis led to an improved pericyte-driven inflammatory response. Moreover, pericytes in pancreatic islets exert vascular control of hormone secretion and glucose homeostasis, and pericyte alteration has been linked to diabetic islet dysfunction [60]. Interestingly, pericyte effects on islet functionality are not limited to vascular support for insulin secretion because pancreatic β cell maturation and functionality rely on pericyte-derived bone morphogenic protein 4 (BMP4). Recently, other pericrine signaling molecules have been identified as key players in the functional regulation of tissue parenchyma encompassing a range of organ-specific functions in both vascular and non-vascular interfaces. Two interesting examples are that leptin receptor-expressing pericytes in the mediobasal hypothalamus mediate energy balance via neuronal leptin signaling [62], and that the Hippo–YAP/TAZ pathway in pericytes generates essential pericrine signals to epithelial and ECs during lung morphogenesis [63]. All things considered, these studies suggest that restoring the physiological functions of pericytes improves blood vessel performance and disease outcomes in various contexts, which may encourage the development of novel pericyte-focused therapies.

Tumor pericytes: loss or change of identity?

Many preclinical studies over the past decade showed that pericyte dysfunction is involved in cancer progression [64]. In this context, it is now well established that tumoral vessels are poorly covered by pericytes [65,66], that diminished pericyte recruitment and maturation lead to enhanced vascular and tumor growth [15,66], and that pericyte depletion favors metastasis. However, there is no consensus about whether defective pericyte–EC interaction in tumors relates to loss of pericyte molecular identity [67], pericyte transdifferentiation into fibroblasts [68], poor pericyte recruitment [66], or a combination of these phenotypes. Functionally, metabolic reprogramming of tumor pericytes has also been implicated in abnormal blood vessel contraction and unfavorable patient outcomes [69].

scRNA-seq analyses have provided a new layer of information about tumor pericytes by showing that tumor pericytes are, in fact, relatively homogeneous [70–72]. However, one should acknowledge that most tumor scRNA-seq datasets include low numbers of pericytes, and the presence of tumor pericyte subclusters may have been obscured by the granularity of the data. One way to overcome this limitation would be to establish spatially resolved pericyte-focused atlases by using multiomic approaches. An interesting observation is that pericyte phenotypes are distinct depending on the mechanisms by which tumor vessels form. While angiogenic pericytes persist in an active immature state characterized by a signature of motility and ECM organization, pericytes covering co-opted vessels remain largely quiescent [71,73]. Together, these studies support a model in which pericytes undergo genetic and molecular reprogramming in cancer, which in turn has negative consequences for disease progression. Nonetheless, one should not forget that most of these data refer to mouse preclinical studies, and adequate longitudinal studies in humans are missing. Of note, the so-called pericrine signaling response is also at play in cancer. Indeed, in hepatocellular carcinoma it has been shown that metabolic reprogramming in tumor cells activates HSCs which in turn promote tumorigenesis through the secretion of senescence-associated factors [74]. Another study has described that loss of integrin β3 in tumor pericytes leads to enhanced focal adhesion kinase (FAK)-mediated cytokine release by pericytes, which subsequently stimulates tumor survival and growth [75].

Pericyte immunomodulatory properties

Emerging evidence suggests that pericytes form an integral part of the immune surveillance unit rather than solely performing complementary functions. Upon proinflammatory stimuli, pericytes promote endothelial expression‎ of the leukocyte adhesion molecules vascular cell adhesion protein 1 (VCAM-1) and/or intracellular adhesion molecule 1 (ICAM-1) in the CNS, lung, skin, or muscle that subsequently promote T cell or macrophage infiltration into the affected tissue [59,76,77]. The Rgs5+ and Col1a1+ subgroups of PDGFRβ+ perivascular cells seem to be early responders to neuroinflammation [78]. Of note, inflammation per se induces pericyte detachment from the endothelium and impairs barrier properties [59]. In addition to the physical interaction with leukocytes, pericytes also secrete and respond to cytokines that further regulate immune cell functions, including both innate and adaptive responses. The chemotactic migration and effector functions of neutrophils, T cells, and macrophages are dependent on these early pericrine signals, including MIF, CXCL1, and CCL2 [77–80].

It is now believed that modulation of the immune-related functions of pericytes could affect the outcome of disease progression. For example, pericyte-deficient Pdgfbret/ret mice exhibit increased leukocyte infiltration and activation, leading to aberrant inflammation in a model of experimental autoimmune encephalomyelitis [76]. Treatment with antagonistic VCAM-1 and ICAM-1 antibodies partially rescued the excessive inflammatory phenotype in Pdgfbret/ret mice. Similarly, activating sGC in acute lung injury improved disease outcomes by increasing pericyte interaction with ECs [59]. Conversely, in the tumor microenvironment, tumor cells induce autophagy of NG2+ pericytes which equips them with immunosuppressive properties that favor tumor cell survival and prevent antitumor T cell responses [81]. Pericytes have also been implicated in the underlying pathophysiology of emergent infectious diseases such as coronavirus disease 2019 (COVID-19) [82], thus presenting an additional niche of investigation for the field in the coming years. In conclusion, the importance of pericytes during various inflammatory processes is growing in recognition, although the modulation of immunity by pericytes can have double-edged outcomes.

Do pericytes contribute to fibrosis?

A dysregulated tissue repair response after acute or chronic injury can lead to the onset of fibrosis associated with the abnormal accumulation of activated and contractile αSMA+ myofibroblasts [83]. Myofibroblasts secrete high amounts of inflammatory mediators, growth factors, and ECM components, and promote aberrant ECM remodeling. The current consensus places fibroblasts as the predominant origin of myofibroblasts, although various studies have found alternative cellular origins [83]. Indeed, the existence of pericyte-to-myofibroblast transition has been proposed as a contributing factor in several fibrotic contexts [2], but the promiscuity of cell markers within the mesenchymal compartments across tissues has led to ambiguous and contradictory observations. In the latest developments, scRNA-seq, ATAC-seq, and spatial transcriptomics provide new insights into this conundrum which support a pericyte origin of myofibroblasts in the fibrotic liver, colon, and kidney. For instance, central vein-associated Rgs5+ HSCs are thought to be the dominant origin of ECM-producing myofibroblasts in fibrotic mouse liver [84]. In human colorectal cancer, a subset of periostin (POSTN)+ myofibroblasts seem to originate from RGS5+ pericytes [70]; similarly, NOTCH3+RGS5+PDGFRβ+ human pericytes contribute to the generation of POSTN+PDGFRα+NKD2+ myofibroblasts during kidney fibrosis [85]. Surprisingly, however, the same authors did not capture the existence of profibrotic pericytes during myocardial infarction when sequencing the entire heart [86]. Although transdifferentiation from pericytes to myofibroblasts may be tissue-specific, it is fair to acknowledge that the latter study did not include pericytes in the trajectory analysis that predicted the origins of myofibroblasts.

Lineage-tracing experiments in mice have also cast some light onto the role of pericytes in fibrosis. For instance, Pham et al. showed that myofibroblast genes are enriched in Tbx18+ pericytes from injured mouse hearts and brains [87], and Dias et al. reported that GLAST+PDGFRβ+ perivascular cells also contribute to fibrosis in the post-stroke brain [88]. In line with this, depletion of GLAST+PDGFRβ+ perivascular cells in the spinal cord leads to reduced fibrotic scar after injury and improves neuronal function [89]. Of note, GLAST+PDGFRβ+ perivascular cells were characterized as spinal cord pericytes by the researchers, but there is insufficient evidence to rule out that these cells might be fibroblasts or astrocytes. This further exemplifies that the shared marker expression‎ profiles of pericytes and other perivascular residing cells still hamper the design of robust pericyte reporter models. Moreover, multiple studies have found no significant pericyte origin for myofibroblasts in distinct fibrosis models of the CNS [90,91] and heart [92]. Although the discrepancies between studies may reflect the lack of robust pericyte identification strategies, it appears that the occurrence of pericyte-to-myofibroblast transition is tissue-specific, and is contingent both on the local microenvironment and on the extent of the injurious stimuli [83]. This emphasizes the need to further unravel the fibrotic pericyte responses in different organs and prompts the question of how myofibroblast origins relate to different pathophysiological phenotypes. All things considered, the origin of myofibroblasts may involve distinct precursor cells depending on the circumstances, although a role for pericytes seems to be indisputable (Figure 4).

Figure 4 Pericytes as a source of myofibroblasts in fibrosis.

Show full captionFigure viewer

Concluding remarks

It is widely recognized that pericytes play an important role in blood vessel formation, stabilization, and function, and that degeneration or loss of brain pericytes impairs their protective barrier properties. Recent advances have revealed novel and crucial roles for pericytes across tissues in a variety of vascular and non-vascular processes. Single-cell technology is becoming more commonly used to better understand the molecular processes that define pericytes in health and disease. Although the molecular and functional attributes of pericytes are not fully elucidated, deep sequencing has revealed organotypic pericyte heterogeneities and new criteria to distinguish pericytes from other cell types. Despite the numerous suggested roles of pericytes in various diseases and physiological processes, including neurodegeneration, cancer, fibrosis, blood flow regulation, and inflammation, the underlying organotypic mechanisms of these contributions are not yet fully understood. The discrepancies between some studies highlight the importance of designing suitable mouse models for eval‎uating the specific mechanisms by which pericytes impact on these processes, and future cross-validations with human data are warranted to ascertain the clinical relevance of pathological pericytes (see Outstanding questions). Overall, based on the emerging evidence on the contribution of pericytes to several diseases, we anticipate an increasing emphasis on pericyte-oriented research in vascular (and non-vascular) studies in the coming years.

Outstanding questions

Why are pericytes molecularly and functionally promiscuous and heterogeneous between distinct tissues? If the transcription factors that regulate pericyte differentiation and function are similar across tissues, will epigenetic mechanisms provide cues into the mechanisms of pericyte identity?

Will spatial molecular atlases focused on pericytes address the key functional and identity conundrums posed by transitioning phenotypes? Will this suffice, or are complementary morphological and functional studies also necessary?

Will pericyte-focused therapy provide new means to stimulate functional angiogenesis in pathology? Given that pericytes are associated with many diseases, will (and how broadly can) pericyte-focused therapies improve patient outcomes?

Neurodegenerative diseases are age-related diseases, and pericyte degeneration is an aging process. Hence, is pericyte degeneration a confounding factor in the development of age-related neurodegeneration? Will maintaining healthy pericyte function promote healthy aging?

AcknowledgmentsAcknowledgments

We would like to thank Sandra D. Castillo, Ana Angulo-Urarte, and Leonor Gouveia for their valuable feedback. Figures were created with BioRender.com. We thank the Centres de Recerca de Catalunya (CERCA) Program/Generalitat de Catalunya and the Josep Carreras Foundation for institutional support. Work-related to this publication in the laboratory of M.G. is supported by research grants from la Asociación Española Contra el Cancer (AECC)-Grupos Traslacionales (GCTRA18006CARR); and by Worldwide Cancer Research (WCR 21-0159). H.v.S. received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 955951; P.V. recieved funding from AECC (AECC-INVES211084VILL) and from the Spanish Ministry of Science and Innovation (RYC2020-029929-I). We apologize to the many authors whose primary papers could not be cited owing to space constraints.

Declaration of interests

The authors declare no conflicts of interest.

References

1.

Armulik, A. ...

Pericytes: developmental, physiological, and pathological perspectives, problems, and promises

Dev. Cell. 2011; 21:193-215

Full Text

Full Text (PDF)

Scopus (1996)

PubMed

Google Scholar

2.

Holm, A. ...

Microvascular mural cell organotypic heterogeneity and functional plasticity

Trends Cell Biol. 2018; 28:302-316

Full Text

Full Text (PDF)

Scopus (85)

PubMed

Google Scholar

3.

Petrova, T.V. ∙ Koh, G.Y.

Biological functions of lymphatic vessels

Science. 2020; 369, eaax4063

Crossref

Scopus (234)

PubMed

Google Scholar

ReviewVolume 34, Issue 1p58-71January 2024Open access

Download Full Issue

Pericytes in the disease spotlight

Hielke van Splunder1,5 ∙ Pilar Villacampa2,5 ∙ Anabel Martínez-Romero1 ∙ Mariona Graupera1,3,4 mgraupera@carrerasresearch.org

Affiliations & NotesArticle Info

• Download PDF
• Cite

• Share

• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Highlights

Molecular and functional pericyte studies at single-cell resolution are providing new insights into long-standing questions about pericyte heterogeneity.

Pericytes are not identified by a single marker but instead by gene expression‎ signatures that show substantial inter-organ differences.

Pericytes orchestrate and precede endothelial cell responses during angiogenesis.

Pericyte degeneration and dysfunction, that are triggered by the onset of some diseases, contribute to the progression of those diseases in both vascular and non-vascular contexts.

The number of diseases with pericyte dysfunction continues to expand, thereby anticipating a promising future for pericyte-focused therapy.

Abstract

Pericytes are known as the mural cells in small-caliber vessels that interact closely with the endothelium. Pericytes play a key role in vasculature formation and homeostasis, and when dysfunctional contribute to vasculature-related diseases such as diabetic retinopathy and neurodegenerative conditions. In addition, significant extravascular roles of pathological pericytes are being discovered with relevant implications for cancer and fibrosis. Pericyte research is challenged by the lack of consistent molecular markers and clear discrimination criteria versus other (mural) cells. However, advances in single-cell approaches are uncovering and clarifying mural cell identities, biological functions, and ontogeny across organs. We discuss the latest developments in pericyte pathobiology to inform future research directions and potential outcomes.

Keywords

1. pericyte
2. vascular disease
3. pericrine signaling
4. fibrosis

Multifaceted roles of mural cells in health and disease

Pericytes are classically defined as mural cells (see Glossary) that envelop the endothelium of small caliber blood vessels, the so-called capillaries. Pericytes are embedded within the same basement membrane as endothelial cells (ECs) and interact closely with them [1,2]. By contrast, vascular smooth muscle cells (vSMCs), the other mural cell type, cover large arteries and veins, and are physically separated from the endothelium by an intimal layer of extracellular matrix (ECM). Of note, lymphatic capillaries lack pericytes under physiological conditions, although collecting lymphatic vessels contain vSMCs [3].

A fundamental function of mural cells is to regulate the stabilization and function of blood vessels. It is therefore not surprising that pericyte loss and dysfunction were linked to several diseases including cancer and cerebrovascular diseases more than a decade ago [4,5]. However, pericyte-focused therapies have been poorly explored. Instead, most studies on vascular-directed therapeutic strategies have been on ECs – the central components that build blood vessels. Emerging data are, however, changing the perception of pericytes from mere supporting vascular cells that are recruited at the final stage of vessel formation to essential elements in the early phases of angiogenesis that anticipate and orchestrate EC behavior. In addition, recent research is revealing novel pathological roles for pericytes beyond their implications in the vasculature. Collectively, we believe that these data open exciting avenues for pericyte-focused therapeutic approaches and call for a broader understanding of these cells in disease progression.

We provide here a global overview of recent significant advances regarding our understanding of the role of pericytes in different pathobiological scenarios and discuss the field's current paradigms and controversies. First, we address new insights into the functions associated with pericytes during physiological vascular responses. Second, we discuss evidence supporting a role of pericytes in disease, including pericyte cell-autonomous implications beyond the vasculature. For comprehensive details on pericyte biology, function ontology, and specific signaling pathways, we refer the reader to [1,2,5]. Of importance, some of the emerging concepts in pericyte biology described in the following sections have only been studied in one specific tissue. To avoid confusion about the generalizability of pericyte properties, we frame each function by considering the relevant organ of study.

Key concepts about pericytes in physiologyPericytes: a particular subtype of mural cells

Pericytes exhibit significant inter- and intra-tissue molecular differences and exert tissue-specific functions [2]. Their molecular, morphological, and functional heterogeneity is inextricably linked to their diverse developmental origins, modes of vessel recruitment, and specific anatomical localization. For example, pericytes of the central nervous system (CNS) microvasculature are firmly and continuously invested around the endothelium to support vascular barrier properties, whereas liver pericytes, commonly referred to as hepatic stellate cells (HSCs), reside in the perisinusoidal space, are loosely and discontinuously associated to ECs, and hold a unique vitamin A storage capacity [2]. To meet tissue-specific demands, pericyte distribution and density are variable among organs and vascular beds, with the CNS microvasculature showing the greatest pericyte-to-EC abundance. From a molecular standpoint there is no single molecular marker that can exclusively identify pericytes (Box 1), albeit the emergence of single-cell techniques is shedding light on tissue-specific pericyte molecular markers and functions. For example, the first molecular atlas of vascular cell types in the brain of adult mice by single-cell RNA sequencing (scRNA-seq) revealed that mural cells follow a gradient of transitional phenotypes. This gradient occurs at the interface of precapillary arterioles, capillaries, and postcapillary venules, and does not follow a single continuum along the arteriovenous axis (Figure 1 and Box 1) [6]. Whether this gradient of transitional phenotypes is specifically restricted to the brain vasculature or is also present in other vascular beds remains to be determined. Indeed, pericytes exhibit many organotypic differences in the expression‎ of molecular markers (Figure 2 illustrates three top-ranked pericyte markers with enriched expression‎ per organ), of which the expression‎ of transporters and components of the contractile machinery exhibit the greatest differences between organs [7]. Another intriguing observation is that pericytes exhibit more cross-organ heterogeneity than vSMCs [7,8]. Currently, the inter-tissue differences in the behavior of the two main mural cell types are not completely understood. However, this may be because pericytes exhibit a greater cell-intrinsic plasticity to adapt their molecular portfolio and function to tissue-specific demands, whereas vSMCs fulfill a more universal function across tissues. In contrast to the tissue-specific transcriptomic differences, the expression‎ of transcription factors appears to be relatively conserved in mural cells across organs, thereby suggesting that mural cell subtypes are defined by epigenetic mechanisms [7]. Accordingly, DNA hypermethylation was recently found to control alpha smooth muscle actin (αSMA) expression‎ in renal mural cells after ischemia [9]. This indicates that methods such as assay for transposase-accessible chromatin sequencing (ATAC-seq) will be instrumental to further understand mural cell phenotypes.

Figure 1 Schematic representation of mural cell zonation in the adult mouse brain.

Show full captionFigure viewer

Figure 2 Organotypic heterogeneity of pericyte markers.

Show full captionFigure viewer

Box 1

Unraveling the identity of pericytes

The identification of pericytes remains a challenging task. Despite ongoing efforts, there is no consensus regarding unambiguous criteria for pericyte identification. To date no single molecular marker can exclusively identify all pericytes or distinguish pericytes from other cell types, although scRNA-seq is now providing new opportunities to discern pericyte marker heterogeneity and tissue specificity [6,8,71,93]. The use of transgenic reporter mouse models has been instrumental to label, trace, and locate different mural cell populations in vivo. A combination of multiple reporter lines is often necessary to properly identify and discriminate pericytes from endothelial cells (ECs) and other perivascular cells [6–8]. Mural cells are highly plastic cells; phenotypic zonation of mouse brain mural cells has revealed that these cells do not follow a single continuum along the arteriovenous axis (see Figure 1A,B in main text) [6]. From a transcriptional point of view, there are two distinct continuums of mural cells: (i) capillary pericytes and venous smooth muscle cells (SMCs), where pericytes gradually transition to a venous SMC phenotype, and (ii) arterial SMCs which transition in an distinct pattern towards arteriole SMCs. The transcriptional resemblance between mouse brain pericytes and venular mural cells [6], as well as the lack of classic pericytes in several organs [7,8], have led to the hypothesis that capillary pericytes are transcriptionally and morphologically similar to venous SMCs in some tissues. Human brain mural cells recapitulate the mouse zonation pattern, although human pericytes are evenly distributed over capillaries and veins [50,94]. Unlike the anatomical separation of pericytes and venous SMCs in the mouse brain, subtypes of human pericytes are discerned by functionality marked by solute transport and extracellular matrix (ECM) organization [50]. Unfortunately, the ability of mouse markers to predict the presence of human pericytes remains limited, and only a select few retain adequate specificity. The use of zebrafish models may provide a better alternative to study conserved pericyte genes [95]. We believe that RGS5, NDUFA4L2, KNCJ8, HIGD1B, ABCC9, NOTCH3, and PDGFRB are currently the most organ and species conserved pericyte markers, although detailed intra-tissue characterization remains necessary when studying pericytes (see Figure 2 in main text).

Pericytes at play during vascular growth

Many studies have documented that pericytes contribute to angiogenesis [10]. The historical view proposes that pericytes mainly contribute to the late stages of vessel formation [2,10]. By taking advantage of the mouse retina as a paradigmatic experimental model of developmental angiogenesis, this concept has been challenged [11–16]. Indeed, these studies showed that, during the early phases of developmental angiogenesis, pericytes, which have not yet achieved the maturity seen in formed vessels, are permissive to cell-cycle progression, morphological adaptation, and migration [12,13]. In this setting, pericyte growth precedes the expansion of ECs, although it is still unclear why. One possibility is that, by expanding rapidly, pericytes ensure the production of sufficient EC growth signals, a hypothesis which is coherent with the observation that inhibition of pericyte activation blocks EC proliferation [12] and induces nuclear translocation of FOXO1 [11], the master regulator of EC quiescence [17]. Another study that examined the brain vasculature showed that, when pericytes are absent, ECs become angiogenic but are not able to proliferate [18], thereby supporting a model in which ECs require the presence of pericytes to expand. Nonetheless, it is fair to acknowledge that other studies have shown that reduced pericyte coverage leads to increased EC proliferation [19]. Although these discrepancies highlight that pericyte–EC interactions are complex, they may be explained by the differences between the animal models and genetic strategies used to interfere with pericytes. Importantly, pericyte behaviors during angiogenesis have been mostly described in tissues belonging to the CNS. Hence, given the high abundance of pericytes in the CNS, it is possible that angiogenic pericytes fulfill different roles in tissues where ECs substantially outnumber them. Another interesting observation is that, during angiogenesis, immature pericytes remain in close contact with ECs, although they do not cover them in their entirety [12,20]. This suggests that pericyte–EC communication during angiogenesis relies on both paracrine and juxtracrine signaling, and may explain why pericyte loss [11,16,21,22] and impaired transition to a fully maturate state [12] lead to distinct endothelial phenotypes during angiogenesis. scRNA-seq analysis of prenatal developing human brains confirmed that angiogenesis is supported by immature mural cells [20]. Consistent with mouse data [12], the state of mature human pericytes correlates with the progression of angiogenesis. Furthermore, the gene expression‎ profiles of these cells show involvement in processes related to the transport across the blood–brain barrier (BBB) and the synthesis of ECM components [20]. Together, these data support a model in which pericytes modulate the early phases of angiogenesis by directly regulating EC behavior. Intriguingly, however, detailed ultrastructural analysis of angiogenic vessels in human brain distinguishes only a single mural cell population, compared to three distinct EC populations [20]. This suggests that ultrastructural features do not define subtype specification in the mural cell compartment, and that molecular and structural features are not necessarily associated with each other.

Brain pericytes and vessel contraction: a matter of transitional phenotypes

Although the regulation of vascular tone through pericyte contractility is considered to be an important function of cardiac, renal, and pulmonary pericytes, as well as of HSCs [2], there has been a long-standing debate in the field as to whether pericytes actively modulate cerebral blood flow [23–25]. For instance, by using optical imaging, Hill et al. suggested that neural/glia antigen 2-positive (NG2+) αSMA− pericytes are not contractile and do not actively modulate the capillary diameter [26]. Instead, by similar optogenetic approaches Hartmann et al. proposed that pericytes do constrict, although they require prolonged and more intense stimulation than αSMA+ mural cells located at larger vessels [27]. Although no consensus has been established, the opposing results between studies may simply reflect heterogeneities in the type of blood vessels and mural cells analyzed. A recent report has shown that NG2+αSMA+ mural cells, located at the transitional segment between arteries and capillaries, regulate the vascular tone and contractility [28]. This suggests that the transition of functional phenotypes between mural cells covering distinct types of blood vessels is tightly regulated.

scRNA-seq analysis of brain mural cells has revealed an abrupt change in the molecular signatures of pericytes and mural cells located in arteries, even from cells residing in proximity on the vasculature, thereby supporting the existence of a blunt transition [6]. Taken together, one can speculate that, in addition to defined vSMC types, there is a subtype of mural cells that exhibit some traits, but not all, of classic pericytes, and are located at transitional vessels and can modulate the vascular tone. Given the ability of pericytes to adapt their phenotype to various microenvironmental conditions [1,2], it is also possible that regulation of blood flow may only occur under specific circumstances. However, one should consider that some of the data disputing pericyte contractility may relate to experimental artefacts, and it should be stressed that most analyses were conducted in the cerebral vasculature as a prototypical example of a vascular bed that is highly sensitive to contraction [25]. An important observation is that pericytes exhibit significant organotypic differences in the basal expression‎ of contractility genes, and pericytes in the bladder and colon express considerable levels of Myh11, Tagln, and Acta2 (αSMA), whereas pericytes in the brain, lung, and heart express negligible amounts of these contractile genes [6,7]. This highlights a conundrum regarding how brain pericytes regulate vessel contractility when typical contractility genes are not expressed.

Pericyte safeguarding the capillary brain bed by a special touch

An essential function of pericytes is to regulate the BBB by controlling the passage of fluid and substances into the parenchymal space [22,29]. Hence, defective pericyte coverage caused by pericyte dysfunction, impaired pericyte recruitment, and pericyte loss all lead to increased EC transcytosis and permeability [22,29]. Aberrant platelet-derived growth factor B (PDGF-B)/platelet-derived growth factor receptor beta (PDGFRβ) signaling is sufficient to experimentally reduce pericyte abundance and the subsequent loss of BBB properties [22,29]. In addition, proper ECM deposition by pericytes (among other cell types composing the neurovascular unit) plays an essential role in maintaining the integrity of the vascular barrier. Indeed, pericyte-derived vitronectin prevents endothelial transcytosis by binding to integrin α5 subunit on ECs [30], and pericyte-secreted laminin interacts with the dystrophin–glycoprotein complex in astrocytes and regulates their endfeet polarization [5,31].

To serve as guardians of the capillary bed, pericytes also establish physical interactions with ECs and form a continuous chain-like network along the capillaries of the cerebral vasculature. Adequate coverage of the endothelium is sustained by active remodeling of distal pericyte processes through cytoskeletal rearrangements [32]. Of relevance, pericyte remodeling capabilities become exhausted with age [33], and this may explain why pericyte coverage is diminished in the vasculature of old mice [33,34]. An interesting observation is that pericyte depletion in adult mice leads to relatively mild BBB defects in different experimental models [35,36]. This includes adult induced Pdgfb ablation [36] and diphtheria toxin A (DTA) expression‎ in PDGFRβ+ cells [35]. Currently it is not clear why loss of pericytes leads to different vascular barrier phenotypes in development and adulthood. Given that the BBB becomes functional during late embryonic development, one can speculate that defects in pericyte coverage are only significant before the onset of BBB formation. Another possibility is that pericyte coverage determines the threshold for BBB defects, and Vazquez-Liebanas et al. showed that only <50% longitudinal pericyte coverage in adult brains leads to significant leakage defects [36]. This is coherent with previous observations of brain vessel phenotypes during development which demonstrated that pericyte coverage is positively correlated with BBB integrity [22]. Choe et al. also reported that DTA-induced loss of pericytes leads to capillary stalling due to increased interactions between ECs and leukocytes. However, because this effect was not observed in other adult pericyte depletion models [35], one should acknowledge that it is possible that the expression‎ of DTA generated unintended toxic effects beyond pericytes.

Pericytes in disease

Pericyte dysfunction is a hallmark of various diseases (Figure 3). For a long time it was believed that maladaptive pericytes mainly affect vascular homeostasis because pericyte and EC functions are interdependent and require bidirectional communication (Box 2). However, there is growing evidence that pericytes have roles in processes beyond the vasculature. As such, pericyte-derived signals (hereafter referred to as pericrine signaling) also modulate tissue function in both physiology and disease. In the following section, we capture recent data showing new observations that link pericyte dysfunction and loss in vascular and non-vascular-related diseases.

Figure 3 Dysfunctional pericytes in disease.

Show full captionFigure viewer

Box 2

Key signaling pathways that orchestrate pericyte–EC crosstalk

Given the close relationship between pericytes and ECs, it is not surprising that bidirectional communication and regulation between them are crucial during vessel formation and maintenance. During angiogenesis, established examples of pericyte–EC communication include the PDGFRβ, transforming growth factor β1 (TGF-β1), ANG1, and NOTCH3 pathways [1]. PDGF-B production from tip ECs is the master signal that recruits PDGFRβ-expressing pericytes to newly formed vessels [1], together with CD146 (MCAM), which acts as a coreceptor for PDGFRβ [96]. Recent advances have shown that NCK1 and NCK2 promote phosphorylation of PDGFRβ in response to PDGF-BB and stimulate pericyte migration by inducing MRTF translocation to the nucleus where they interact with the serum response transcription factor (SRF) [13,21]. Similarly, jagged 1 (JAG1) expressed by ECs activates NOTCH3 in pericytes and promotes pericyte maturation [14,97] and the expression‎ of PDGFRβ [98]. Conversely, ANG1 is secreted by pericytes, activates the tyrosine receptor TIE2 in ECs, and promotes EC maturation and vascular integrity [2,15]. TGF-β exerts complex effects on ECs and pericytes, and TGF-B receptor 1 (also known as ALK5) plays a dominant role in these interactions. Indeed, deletion of ALK5 in ECs leads to pericyte dysfunction and hemorrhagic vascular malformations [99]. Instead, deletion of ALK5 in pericytes results in increased EC proliferation, reduced collagen deposition, and enhanced matrix metalloproteinase activity [19]. Of note, pericytes also express canonical EC receptors such as VEGF-R1 [15,16] and TIE2 which allow pericytes to modulate intrinsic EC signaling.

The CNS: a hotspot of pericyte-related vascular diseases

Pericyte-related vascular defects have been reported in various CNS diseases, including Alzheimer's disease (AD), Parkinson's disease, dementia, stroke, diabetic retinopathy, glaucoma, and intracranial vascular malformations [5,37–39]. The involvement of pericytes in several CNS-related diseases is partially explained by their abundance within the brain vasculature and their key role in maintaining the BBB, where barrier breakdown precedes neurodegeneration. Other phenotypes linking pericytes dysfunction and CNS disease include neuron death [40] and impaired neurovascular coupling [41,42]. Intriguingly, NG2+ retinal pericytes orchestrate neurovascular coupling through closed-ended nanotubes between pericytes on adjacent capillaries, even when they are positioned far apart. These nanotubes terminate in a gap junction at the recipient pericyte, which permits rapid fluxes of small molecules and calcium ions, thereby allowing pericytes to coordinate neuronal activity [41]. Indeed, maintaining adequate calcium levels is essential to sustain pericyte function in the CNS, and aberrant levels of calcium in NG2+ pericytes lead to poor recovery after ischemic stroke [23,43] or neovascular dysfunction and neuronal death in glaucoma [42].

AD is the prototypical example of a CNS disease associated with aberrant vascular function and BBB breakdown linked to pericyte dysfunction and loss [38,44]. Although the involvement of pericytes in AD has been recognized for several years [5,44], new insights have challenged the timeframe in which patients suffering from AD develop pericyte dysfunction and BBB impairment. Indeed, it is now understood that BBB breakdown is an early event in AD, and these defects are used as an early biomarker of cognitive decline [45]. We highlight recent observations which support the involvement of pericytes in the onset of AD. For instance, Nortley et al. showed that the reduction in cerebral blood flow, that is considered to be the first clinical manifestation of AD, is caused by amyloid-β-induced pericyte contraction in brain capillaries [46]. Another study indicated that cognitive decline and BBB disruption in AD are linked to accelerated pericyte degeneration in carriers of AD susceptibility allele apolipoprotein E4 (APOE4) [47], a process which occurs independently of amyloid-β pathology. In this context, APOE4 carriers show high baseline cerebrospinal fluid levels of soluble (s)PDGFRβ which can be used as a BBB pericyte injury biomarker [47]. Intriguingly, analysis of the cortex of APOE4 transgenic mice using single-nucleus (sn)RNA-seq and phosphoproteomics revealed profound molecular changes related to progressive BBB failure in both ECs and pericytes [48]. Nonetheless, because only a constitutive APOE4-expressing transgenic line was included in the study, it remains unclear whether the molecular alterations of ECs and pericytes solely comprise cell-autonomous effects. In addition, one should not forget that mice do not fully recapitulate all traits of AD. It has been recently noted that pericytes and microglia associations (described in both physiological mouse and human brains) are diminished in the brain capillaries of individuals with AD, and this may also have implications for BBB breakdown [49]. In human brain, two types of pericytes have been identified that are distinguished by solute transport and ECM organization (Box 1). Intriguingly, the second type seems to be selectively affected in AD [50]. Thus, identifying methods to specifically target this cluster of pericytes may provide new ways to maintain vascular fitness in AD.

Pericyte degeneration and death also encompasses early phases of diabetic retinopathy, in which pericytes are primary targets of hyperglycemic damage. Recent findings suggest that, upon initiation of hyperglycemia, pericytes shift towards cell-bridging positions, resulting in physical detachment from ECs [51]. Whether this remodeling is independent of pericyte death or is related to the initiation of that process needs further investigation. Mechanistically, pericyte detachment and shifting are induced by exogenous factors such as angiopoietin 2 (ANG2) and PDGF-B, and are reversed by insulin treatment, illustrating the dynamic behavior of pericytes in the microvasculature [51,52]. In line with this, PDGFB signaling through PDGFRβ and NCKs in pericytes that cover sprouting vessels during experimental proliferative retinopathy [oxygen-induced retinopathy (OIR) model] activates ectopic αSMA expression‎ and promotes pathological neovascularization [21]. Interestingly, depletion of retinal pericytes in adulthood does not phenocopy retinopathy unless another stimulus is present (e.g., vascular endothelial growth factor A, VEGF-A). Upon depletion of pericytes, either during vessel development or in adulthood followed by VEGF addition, inhibition of ANG2 action restrains the severity of the diabetic retinopathy-like phenotypes [11]. Molecular effectors governing the early phases of diabetic retinopathy have remained elusive, precluding the development of drugs aiming to halt disease onset. These data suggest that targeting pericyte adhesion and migration capacities may be of therapeutic interest. Furthermore, pericyte loss in diabetic retinopathy was recently associated with aberrant levels of circular RNAs [53], thereby suggesting the use of circular RNAs as a diagnostic biomarker for early pericyte dysfunction in disease.

Finally, we would like to stress that familial mutations in essential pericyte genes have also been linked to CNS abnormalities. Well-known examples include loss-of-function mutations in NOTCH3 as a cause of CADASIL [54], and mutations in PDGFRB as a cause of brain calcifications [55], neurological deterioration, and white matter lesions [56]. Of note, these genes are equally relevant for pericyte and vSMC biology, and it is unclear whether these mutations lead to distinct phenotypes in mural cells. Current next-generation sequencing approaches allow the discovery of somatic mutations present in pericytes at low allelic frequency. In line with this, it has been proposed that PIK3CA- and AKT-related somatic cerebral cavernous malformations in mice emerge from mutant pericytes [39,57]. However, these data have some caveats because the lineage-tracing experiments used to support these findings were performed with a CRE-recombinase mouse line that is neither pericyte-specific nor inducible.

Pathobiological pericytes beyond the CNS

Although the implications of pericytes in diseases beyond the CNS are less well studied, the number of diseases demonstrating the involvement of pericytes continues to expand. We discuss here emerging evidence supporting a relevant role of pericytes in myocardial infarction [58], acute lung injury [59], and diabetes [60] as prototypical examples. For instance, after myocardial infarction, pericytes regulate inflammation and immune cell trafficking, and modulate ECM remodeling and revascularization [61]. In line with this, molecular reprogramming of PDGFRβ+NG2+ cardiac pericytes into vSMCs through inhibition of MEK1/2 improved the functional cardiac response by promoting revascularization [58]. In acute lung inflammation, the crosstalk between endothelium-derived nitric oxide (NO) and pericyte soluble guanylate cyclase (sGC) is impaired, leading to elevated vascular permeability [59]. Pharmacological activation of the NO–sGC axis led to an improved pericyte-driven inflammatory response. Moreover, pericytes in pancreatic islets exert vascular control of hormone secretion and glucose homeostasis, and pericyte alteration has been linked to diabetic islet dysfunction [60]. Interestingly, pericyte effects on islet functionality are not limited to vascular support for insulin secretion because pancreatic β cell maturation and functionality rely on pericyte-derived bone morphogenic protein 4 (BMP4). Recently, other pericrine signaling molecules have been identified as key players in the functional regulation of tissue parenchyma encompassing a range of organ-specific functions in both vascular and non-vascular interfaces. Two interesting examples are that leptin receptor-expressing pericytes in the mediobasal hypothalamus mediate energy balance via neuronal leptin signaling [62], and that the Hippo–YAP/TAZ pathway in pericytes generates essential pericrine signals to epithelial and ECs during lung morphogenesis [63]. All things considered, these studies suggest that restoring the physiological functions of pericytes improves blood vessel performance and disease outcomes in various contexts, which may encourage the development of novel pericyte-focused therapies.

Tumor pericytes: loss or change of identity?

Many preclinical studies over the past decade showed that pericyte dysfunction is involved in cancer progression [64]. In this context, it is now well established that tumoral vessels are poorly covered by pericytes [65,66], that diminished pericyte recruitment and maturation lead to enhanced vascular and tumor growth [15,66], and that pericyte depletion favors metastasis. However, there is no consensus about whether defective pericyte–EC interaction in tumors relates to loss of pericyte molecular identity [67], pericyte transdifferentiation into fibroblasts [68], poor pericyte recruitment [66], or a combination of these phenotypes. Functionally, metabolic reprogramming of tumor pericytes has also been implicated in abnormal blood vessel contraction and unfavorable patient outcomes [69].

scRNA-seq analyses have provided a new layer of information about tumor pericytes by showing that tumor pericytes are, in fact, relatively homogeneous [70–72]. However, one should acknowledge that most tumor scRNA-seq datasets include low numbers of pericytes, and the presence of tumor pericyte subclusters may have been obscured by the granularity of the data. One way to overcome this limitation would be to establish spatially resolved pericyte-focused atlases by using multiomic approaches. An interesting observation is that pericyte phenotypes are distinct depending on the mechanisms by which tumor vessels form. While angiogenic pericytes persist in an active immature state characterized by a signature of motility and ECM organization, pericytes covering co-opted vessels remain largely quiescent [71,73]. Together, these studies support a model in which pericytes undergo genetic and molecular reprogramming in cancer, which in turn has negative consequences for disease progression. Nonetheless, one should not forget that most of these data refer to mouse preclinical studies, and adequate longitudinal studies in humans are missing. Of note, the so-called pericrine signaling response is also at play in cancer. Indeed, in hepatocellular carcinoma it has been shown that metabolic reprogramming in tumor cells activates HSCs which in turn promote tumorigenesis through the secretion of senescence-associated factors [74]. Another study has described that loss of integrin β3 in tumor pericytes leads to enhanced focal adhesion kinase (FAK)-mediated cytokine release by pericytes, which subsequently stimulates tumor survival and growth [75].

Pericyte immunomodulatory properties

Emerging evidence suggests that pericytes form an integral part of the immune surveillance unit rather than solely performing complementary functions. Upon proinflammatory stimuli, pericytes promote endothelial expression‎ of the leukocyte adhesion molecules vascular cell adhesion protein 1 (VCAM-1) and/or intracellular adhesion molecule 1 (ICAM-1) in the CNS, lung, skin, or muscle that subsequently promote T cell or macrophage infiltration into the affected tissue [59,76,77]. The Rgs5+ and Col1a1+ subgroups of PDGFRβ+ perivascular cells seem to be early responders to neuroinflammation [78]. Of note, inflammation per se induces pericyte detachment from the endothelium and impairs barrier properties [59]. In addition to the physical interaction with leukocytes, pericytes also secrete and respond to cytokines that further regulate immune cell functions, including both innate and adaptive responses. The chemotactic migration and effector functions of neutrophils, T cells, and macrophages are dependent on these early pericrine signals, including MIF, CXCL1, and CCL2 [77–80].

It is now believed that modulation of the immune-related functions of pericytes could affect the outcome of disease progression. For example, pericyte-deficient Pdgfbret/ret mice exhibit increased leukocyte infiltration and activation, leading to aberrant inflammation in a model of experimental autoimmune encephalomyelitis [76]. Treatment with antagonistic VCAM-1 and ICAM-1 antibodies partially rescued the excessive inflammatory phenotype in Pdgfbret/ret mice. Similarly, activating sGC in acute lung injury improved disease outcomes by increasing pericyte interaction with ECs [59]. Conversely, in the tumor microenvironment, tumor cells induce autophagy of NG2+ pericytes which equips them with immunosuppressive properties that favor tumor cell survival and prevent antitumor T cell responses [81]. Pericytes have also been implicated in the underlying pathophysiology of emergent infectious diseases such as coronavirus disease 2019 (COVID-19) [82], thus presenting an additional niche of investigation for the field in the coming years. In conclusion, the importance of pericytes during various inflammatory processes is growing in recognition, although the modulation of immunity by pericytes can have double-edged outcomes.

Do pericytes contribute to fibrosis?

A dysregulated tissue repair response after acute or chronic injury can lead to the onset of fibrosis associated with the abnormal accumulation of activated and contractile αSMA+ myofibroblasts [83]. Myofibroblasts secrete high amounts of inflammatory mediators, growth factors, and ECM components, and promote aberrant ECM remodeling. The current consensus places fibroblasts as the predominant origin of myofibroblasts, although various studies have found alternative cellular origins [83]. Indeed, the existence of pericyte-to-myofibroblast transition has been proposed as a contributing factor in several fibrotic contexts [2], but the promiscuity of cell markers within the mesenchymal compartments across tissues has led to ambiguous and contradictory observations. In the latest developments, scRNA-seq, ATAC-seq, and spatial transcriptomics provide new insights into this conundrum which support a pericyte origin of myofibroblasts in the fibrotic liver, colon, and kidney. For instance, central vein-associated Rgs5+ HSCs are thought to be the dominant origin of ECM-producing myofibroblasts in fibrotic mouse liver [84]. In human colorectal cancer, a subset of periostin (POSTN)+ myofibroblasts seem to originate from RGS5+ pericytes [70]; similarly, NOTCH3+RGS5+PDGFRβ+ human pericytes contribute to the generation of POSTN+PDGFRα+NKD2+ myofibroblasts during kidney fibrosis [85]. Surprisingly, however, the same authors did not capture the existence of profibrotic pericytes during myocardial infarction when sequencing the entire heart [86]. Although transdifferentiation from pericytes to myofibroblasts may be tissue-specific, it is fair to acknowledge that the latter study did not include pericytes in the trajectory analysis that predicted the origins of myofibroblasts.

Lineage-tracing experiments in mice have also cast some light onto the role of pericytes in fibrosis. For instance, Pham et al. showed that myofibroblast genes are enriched in Tbx18+ pericytes from injured mouse hearts and brains [87], and Dias et al. reported that GLAST+PDGFRβ+ perivascular cells also contribute to fibrosis in the post-stroke brain [88]. In line with this, depletion of GLAST+PDGFRβ+ perivascular cells in the spinal cord leads to reduced fibrotic scar after injury and improves neuronal function [89]. Of note, GLAST+PDGFRβ+ perivascular cells were characterized as spinal cord pericytes by the researchers, but there is insufficient evidence to rule out that these cells might be fibroblasts or astrocytes. This further exemplifies that the shared marker expression‎ profiles of pericytes and other perivascular residing cells still hamper the design of robust pericyte reporter models. Moreover, multiple studies have found no significant pericyte origin for myofibroblasts in distinct fibrosis models of the CNS [90,91] and heart [92]. Although the discrepancies between studies may reflect the lack of robust pericyte identification strategies, it appears that the occurrence of pericyte-to-myofibroblast transition is tissue-specific, and is contingent both on the local microenvironment and on the extent of the injurious stimuli [83]. This emphasizes the need to further unravel the fibrotic pericyte responses in different organs and prompts the question of how myofibroblast origins relate to different pathophysiological phenotypes. All things considered, the origin of myofibroblasts may involve distinct precursor cells depending on the circumstances, although a role for pericytes seems to be indisputable (Figure 4).

Figure 4 Pericytes as a source of myofibroblasts in fibrosis.

Show full captionFigure viewer

Concluding remarks

It is widely recognized that pericytes play an important role in blood vessel formation, stabilization, and function, and that degeneration or loss of brain pericytes impairs their protective barrier properties. Recent advances have revealed novel and crucial roles for pericytes across tissues in a variety of vascular and non-vascular processes. Single-cell technology is becoming more commonly used to better understand the molecular processes that define pericytes in health and disease. Although the molecular and functional attributes of pericytes are not fully elucidated, deep sequencing has revealed organotypic pericyte heterogeneities and new criteria to distinguish pericytes from other cell types. Despite the numerous suggested roles of pericytes in various diseases and physiological processes, including neurodegeneration, cancer, fibrosis, blood flow regulation, and inflammation, the underlying organotypic mechanisms of these contributions are not yet fully understood. The discrepancies between some studies highlight the importance of designing suitable mouse models for eval‎uating the specific mechanisms by which pericytes impact on these processes, and future cross-validations with human data are warranted to ascertain the clinical relevance of pathological pericytes (see Outstanding questions). Overall, based on the emerging evidence on the contribution of pericytes to several diseases, we anticipate an increasing emphasis on pericyte-oriented research in vascular (and non-vascular) studies in the coming years.

Outstanding questions

Why are pericytes molecularly and functionally promiscuous and heterogeneous between distinct tissues? If the transcription factors that regulate pericyte differentiation and function are similar across tissues, will epigenetic mechanisms provide cues into the mechanisms of pericyte identity?

Will spatial molecular atlases focused on pericytes address the key functional and identity conundrums posed by transitioning phenotypes? Will this suffice, or are complementary morphological and functional studies also necessary?

Will pericyte-focused therapy provide new means to stimulate functional angiogenesis in pathology? Given that pericytes are associated with many diseases, will (and how broadly can) pericyte-focused therapies improve patient outcomes?

Neurodegenerative diseases are age-related diseases, and pericyte degeneration is an aging process. Hence, is pericyte degeneration a confounding factor in the development of age-related neurodegeneration? Will maintaining healthy pericyte function promote healthy aging?

AcknowledgmentsAcknowledgments

We would like to thank Sandra D. Castillo, Ana Angulo-Urarte, and Leonor Gouveia for their valuable feedback. Figures were created with BioRender.com. We thank the Centres de Recerca de Catalunya (CERCA) Program/Generalitat de Catalunya and the Josep Carreras Foundation for institutional support. Work-related to this publication in the laboratory of M.G. is supported by research grants from la Asociación Española Contra el Cancer (AECC)-Grupos Traslacionales (GCTRA18006CARR); and by Worldwide Cancer Research (WCR 21-0159). H.v.S. received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 955951; P.V. recieved funding from AECC (AECC-INVES211084VILL) and from the Spanish Ministry of Science and Innovation (RYC2020-029929-I). We apologize to the many authors whose primary papers could not be cited owing to space constraints.

Declaration of interests

The authors declare no conflicts of interest.

References

1.

Armulik, A. ...

Pericytes: developmental, physiological, and pathological perspectives, problems, and promises

Dev. Cell. 2011; 21:193-215

Full Text

Full Text (PDF)

Scopus (1996)

PubMed

Google Scholar

2.

Holm, A. ...

Microvascular mural cell organotypic heterogeneity and functional plasticity

Trends Cell Biol. 2018; 28:302-316

Full Text

Full Text (PDF)

Scopus (85)

PubMed

Google Scholar

3.

Petrova, T.V. ∙ Koh, G.Y.

Biological functions of lymphatic vessels

Science. 2020; 369, eaax4063

Crossref

Scopus (234)

PubMed

Google Scholar

ReviewVolume 34, Issue 1p58-71January 2024Open access

Download Full Issue

Pericytes in the disease spotlight

Hielke van Splunder1,5 ∙ Pilar Villacampa2,5 ∙ Anabel Martínez-Romero1 ∙ Mariona Graupera1,3,4 mgraupera@carrerasresearch.org

Affiliations & NotesArticle Info

• Download PDF
• Cite

• Share

• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Highlights

Molecular and functional pericyte studies at single-cell resolution are providing new insights into long-standing questions about pericyte heterogeneity.

Pericytes are not identified by a single marker but instead by gene expression‎ signatures that show substantial inter-organ differences.

Pericytes orchestrate and precede endothelial cell responses during angiogenesis.

Pericyte degeneration and dysfunction, that are triggered by the onset of some diseases, contribute to the progression of those diseases in both vascular and non-vascular contexts.

The number of diseases with pericyte dysfunction continues to expand, thereby anticipating a promising future for pericyte-focused therapy.

Abstract

Pericytes are known as the mural cells in small-caliber vessels that interact closely with the endothelium. Pericytes play a key role in vasculature formation and homeostasis, and when dysfunctional contribute to vasculature-related diseases such as diabetic retinopathy and neurodegenerative conditions. In addition, significant extravascular roles of pathological pericytes are being discovered with relevant implications for cancer and fibrosis. Pericyte research is challenged by the lack of consistent molecular markers and clear discrimination criteria versus other (mural) cells. However, advances in single-cell approaches are uncovering and clarifying mural cell identities, biological functions, and ontogeny across organs. We discuss the latest developments in pericyte pathobiology to inform future research directions and potential outcomes.

Keywords

1. pericyte
2. vascular disease
3. pericrine signaling
4. fibrosis

Multifaceted roles of mural cells in health and disease

Pericytes are classically defined as mural cells (see Glossary) that envelop the endothelium of small caliber blood vessels, the so-called capillaries. Pericytes are embedded within the same basement membrane as endothelial cells (ECs) and interact closely with them [1,2]. By contrast, vascular smooth muscle cells (vSMCs), the other mural cell type, cover large arteries and veins, and are physically separated from the endothelium by an intimal layer of extracellular matrix (ECM). Of note, lymphatic capillaries lack pericytes under physiological conditions, although collecting lymphatic vessels contain vSMCs [3].

A fundamental function of mural cells is to regulate the stabilization and function of blood vessels. It is therefore not surprising that pericyte loss and dysfunction were linked to several diseases including cancer and cerebrovascular diseases more than a decade ago [4,5]. However, pericyte-focused therapies have been poorly explored. Instead, most studies on vascular-directed therapeutic strategies have been on ECs – the central components that build blood vessels. Emerging data are, however, changing the perception of pericytes from mere supporting vascular cells that are recruited at the final stage of vessel formation to essential elements in the early phases of angiogenesis that anticipate and orchestrate EC behavior. In addition, recent research is revealing novel pathological roles for pericytes beyond their implications in the vasculature. Collectively, we believe that these data open exciting avenues for pericyte-focused therapeutic approaches and call for a broader understanding of these cells in disease progression.

We provide here a global overview of recent significant advances regarding our understanding of the role of pericytes in different pathobiological scenarios and discuss the field's current paradigms and controversies. First, we address new insights into the functions associated with pericytes during physiological vascular responses. Second, we discuss evidence supporting a role of pericytes in disease, including pericyte cell-autonomous implications beyond the vasculature. For comprehensive details on pericyte biology, function ontology, and specific signaling pathways, we refer the reader to [1,2,5]. Of importance, some of the emerging concepts in pericyte biology described in the following sections have only been studied in one specific tissue. To avoid confusion about the generalizability of pericyte properties, we frame each function by considering the relevant organ of study.

Key concepts about pericytes in physiologyPericytes: a particular subtype of mural cells

Pericytes exhibit significant inter- and intra-tissue molecular differences and exert tissue-specific functions [2]. Their molecular, morphological, and functional heterogeneity is inextricably linked to their diverse developmental origins, modes of vessel recruitment, and specific anatomical localization. For example, pericytes of the central nervous system (CNS) microvasculature are firmly and continuously invested around the endothelium to support vascular barrier properties, whereas liver pericytes, commonly referred to as hepatic stellate cells (HSCs), reside in the perisinusoidal space, are loosely and discontinuously associated to ECs, and hold a unique vitamin A storage capacity [2]. To meet tissue-specific demands, pericyte distribution and density are variable among organs and vascular beds, with the CNS microvasculature showing the greatest pericyte-to-EC abundance. From a molecular standpoint there is no single molecular marker that can exclusively identify pericytes (Box 1), albeit the emergence of single-cell techniques is shedding light on tissue-specific pericyte molecular markers and functions. For example, the first molecular atlas of vascular cell types in the brain of adult mice by single-cell RNA sequencing (scRNA-seq) revealed that mural cells follow a gradient of transitional phenotypes. This gradient occurs at the interface of precapillary arterioles, capillaries, and postcapillary venules, and does not follow a single continuum along the arteriovenous axis (Figure 1 and Box 1) [6]. Whether this gradient of transitional phenotypes is specifically restricted to the brain vasculature or is also present in other vascular beds remains to be determined. Indeed, pericytes exhibit many organotypic differences in the expression‎ of molecular markers (Figure 2 illustrates three top-ranked pericyte markers with enriched expression‎ per organ), of which the expression‎ of transporters and components of the contractile machinery exhibit the greatest differences between organs [7]. Another intriguing observation is that pericytes exhibit more cross-organ heterogeneity than vSMCs [7,8]. Currently, the inter-tissue differences in the behavior of the two main mural cell types are not completely understood. However, this may be because pericytes exhibit a greater cell-intrinsic plasticity to adapt their molecular portfolio and function to tissue-specific demands, whereas vSMCs fulfill a more universal function across tissues. In contrast to the tissue-specific transcriptomic differences, the expression‎ of transcription factors appears to be relatively conserved in mural cells across organs, thereby suggesting that mural cell subtypes are defined by epigenetic mechanisms [7]. Accordingly, DNA hypermethylation was recently found to control alpha smooth muscle actin (αSMA) expression‎ in renal mural cells after ischemia [9]. This indicates that methods such as assay for transposase-accessible chromatin sequencing (ATAC-seq) will be instrumental to further understand mural cell phenotypes.

Figure 1 Schematic representation of mural cell zonation in the adult mouse brain.

Show full captionFigure viewer

Figure 2 Organotypic heterogeneity of pericyte markers.

Show full captionFigure viewer

Box 1

Unraveling the identity of pericytes

The identification of pericytes remains a challenging task. Despite ongoing efforts, there is no consensus regarding unambiguous criteria for pericyte identification. To date no single molecular marker can exclusively identify all pericytes or distinguish pericytes from other cell types, although scRNA-seq is now providing new opportunities to discern pericyte marker heterogeneity and tissue specificity [6,8,71,93]. The use of transgenic reporter mouse models has been instrumental to label, trace, and locate different mural cell populations in vivo. A combination of multiple reporter lines is often necessary to properly identify and discriminate pericytes from endothelial cells (ECs) and other perivascular cells [6–8]. Mural cells are highly plastic cells; phenotypic zonation of mouse brain mural cells has revealed that these cells do not follow a single continuum along the arteriovenous axis (see Figure 1A,B in main text) [6]. From a transcriptional point of view, there are two distinct continuums of mural cells: (i) capillary pericytes and venous smooth muscle cells (SMCs), where pericytes gradually transition to a venous SMC phenotype, and (ii) arterial SMCs which transition in an distinct pattern towards arteriole SMCs. The transcriptional resemblance between mouse brain pericytes and venular mural cells [6], as well as the lack of classic pericytes in several organs [7,8], have led to the hypothesis that capillary pericytes are transcriptionally and morphologically similar to venous SMCs in some tissues. Human brain mural cells recapitulate the mouse zonation pattern, although human pericytes are evenly distributed over capillaries and veins [50,94]. Unlike the anatomical separation of pericytes and venous SMCs in the mouse brain, subtypes of human pericytes are discerned by functionality marked by solute transport and extracellular matrix (ECM) organization [50]. Unfortunately, the ability of mouse markers to predict the presence of human pericytes remains limited, and only a select few retain adequate specificity. The use of zebrafish models may provide a better alternative to study conserved pericyte genes [95]. We believe that RGS5, NDUFA4L2, KNCJ8, HIGD1B, ABCC9, NOTCH3, and PDGFRB are currently the most organ and species conserved pericyte markers, although detailed intra-tissue characterization remains necessary when studying pericytes (see Figure 2 in main text).

Pericytes at play during vascular growth

Many studies have documented that pericytes contribute to angiogenesis [10]. The historical view proposes that pericytes mainly contribute to the late stages of vessel formation [2,10]. By taking advantage of the mouse retina as a paradigmatic experimental model of developmental angiogenesis, this concept has been challenged [11–16]. Indeed, these studies showed that, during the early phases of developmental angiogenesis, pericytes, which have not yet achieved the maturity seen in formed vessels, are permissive to cell-cycle progression, morphological adaptation, and migration [12,13]. In this setting, pericyte growth precedes the expansion of ECs, although it is still unclear why. One possibility is that, by expanding rapidly, pericytes ensure the production of sufficient EC growth signals, a hypothesis which is coherent with the observation that inhibition of pericyte activation blocks EC proliferation [12] and induces nuclear translocation of FOXO1 [11], the master regulator of EC quiescence [17]. Another study that examined the brain vasculature showed that, when pericytes are absent, ECs become angiogenic but are not able to proliferate [18], thereby supporting a model in which ECs require the presence of pericytes to expand. Nonetheless, it is fair to acknowledge that other studies have shown that reduced pericyte coverage leads to increased EC proliferation [19]. Although these discrepancies highlight that pericyte–EC interactions are complex, they may be explained by the differences between the animal models and genetic strategies used to interfere with pericytes. Importantly, pericyte behaviors during angiogenesis have been mostly described in tissues belonging to the CNS. Hence, given the high abundance of pericytes in the CNS, it is possible that angiogenic pericytes fulfill different roles in tissues where ECs substantially outnumber them. Another interesting observation is that, during angiogenesis, immature pericytes remain in close contact with ECs, although they do not cover them in their entirety [12,20]. This suggests that pericyte–EC communication during angiogenesis relies on both paracrine and juxtracrine signaling, and may explain why pericyte loss [11,16,21,22] and impaired transition to a fully maturate state [12] lead to distinct endothelial phenotypes during angiogenesis. scRNA-seq analysis of prenatal developing human brains confirmed that angiogenesis is supported by immature mural cells [20]. Consistent with mouse data [12], the state of mature human pericytes correlates with the progression of angiogenesis. Furthermore, the gene expression‎ profiles of these cells show involvement in processes related to the transport across the blood–brain barrier (BBB) and the synthesis of ECM components [20]. Together, these data support a model in which pericytes modulate the early phases of angiogenesis by directly regulating EC behavior. Intriguingly, however, detailed ultrastructural analysis of angiogenic vessels in human brain distinguishes only a single mural cell population, compared to three distinct EC populations [20]. This suggests that ultrastructural features do not define subtype specification in the mural cell compartment, and that molecular and structural features are not necessarily associated with each other.

Brain pericytes and vessel contraction: a matter of transitional phenotypes

Although the regulation of vascular tone through pericyte contractility is considered to be an important function of cardiac, renal, and pulmonary pericytes, as well as of HSCs [2], there has been a long-standing debate in the field as to whether pericytes actively modulate cerebral blood flow [23–25]. For instance, by using optical imaging, Hill et al. suggested that neural/glia antigen 2-positive (NG2+) αSMA− pericytes are not contractile and do not actively modulate the capillary diameter [26]. Instead, by similar optogenetic approaches Hartmann et al. proposed that pericytes do constrict, although they require prolonged and more intense stimulation than αSMA+ mural cells located at larger vessels [27]. Although no consensus has been established, the opposing results between studies may simply reflect heterogeneities in the type of blood vessels and mural cells analyzed. A recent report has shown that NG2+αSMA+ mural cells, located at the transitional segment between arteries and capillaries, regulate the vascular tone and contractility [28]. This suggests that the transition of functional phenotypes between mural cells covering distinct types of blood vessels is tightly regulated.

scRNA-seq analysis of brain mural cells has revealed an abrupt change in the molecular signatures of pericytes and mural cells located in arteries, even from cells residing in proximity on the vasculature, thereby supporting the existence of a blunt transition [6]. Taken together, one can speculate that, in addition to defined vSMC types, there is a subtype of mural cells that exhibit some traits, but not all, of classic pericytes, and are located at transitional vessels and can modulate the vascular tone. Given the ability of pericytes to adapt their phenotype to various microenvironmental conditions [1,2], it is also possible that regulation of blood flow may only occur under specific circumstances. However, one should consider that some of the data disputing pericyte contractility may relate to experimental artefacts, and it should be stressed that most analyses were conducted in the cerebral vasculature as a prototypical example of a vascular bed that is highly sensitive to contraction [25]. An important observation is that pericytes exhibit significant organotypic differences in the basal expression‎ of contractility genes, and pericytes in the bladder and colon express considerable levels of Myh11, Tagln, and Acta2 (αSMA), whereas pericytes in the brain, lung, and heart express negligible amounts of these contractile genes [6,7]. This highlights a conundrum regarding how brain pericytes regulate vessel contractility when typical contractility genes are not expressed.

Pericyte safeguarding the capillary brain bed by a special touch

An essential function of pericytes is to regulate the BBB by controlling the passage of fluid and substances into the parenchymal space [22,29]. Hence, defective pericyte coverage caused by pericyte dysfunction, impaired pericyte recruitment, and pericyte loss all lead to increased EC transcytosis and permeability [22,29]. Aberrant platelet-derived growth factor B (PDGF-B)/platelet-derived growth factor receptor beta (PDGFRβ) signaling is sufficient to experimentally reduce pericyte abundance and the subsequent loss of BBB properties [22,29]. In addition, proper ECM deposition by pericytes (among other cell types composing the neurovascular unit) plays an essential role in maintaining the integrity of the vascular barrier. Indeed, pericyte-derived vitronectin prevents endothelial transcytosis by binding to integrin α5 subunit on ECs [30], and pericyte-secreted laminin interacts with the dystrophin–glycoprotein complex in astrocytes and regulates their endfeet polarization [5,31].

To serve as guardians of the capillary bed, pericytes also establish physical interactions with ECs and form a continuous chain-like network along the capillaries of the cerebral vasculature. Adequate coverage of the endothelium is sustained by active remodeling of distal pericyte processes through cytoskeletal rearrangements [32]. Of relevance, pericyte remodeling capabilities become exhausted with age [33], and this may explain why pericyte coverage is diminished in the vasculature of old mice [33,34]. An interesting observation is that pericyte depletion in adult mice leads to relatively mild BBB defects in different experimental models [35,36]. This includes adult induced Pdgfb ablation [36] and diphtheria toxin A (DTA) expression‎ in PDGFRβ+ cells [35]. Currently it is not clear why loss of pericytes leads to different vascular barrier phenotypes in development and adulthood. Given that the BBB becomes functional during late embryonic development, one can speculate that defects in pericyte coverage are only significant before the onset of BBB formation. Another possibility is that pericyte coverage determines the threshold for BBB defects, and Vazquez-Liebanas et al. showed that only <50% longitudinal pericyte coverage in adult brains leads to significant leakage defects [36]. This is coherent with previous observations of brain vessel phenotypes during development which demonstrated that pericyte coverage is positively correlated with BBB integrity [22]. Choe et al. also reported that DTA-induced loss of pericytes leads to capillary stalling due to increased interactions between ECs and leukocytes. However, because this effect was not observed in other adult pericyte depletion models [35], one should acknowledge that it is possible that the expression‎ of DTA generated unintended toxic effects beyond pericytes.

Pericytes in disease

Pericyte dysfunction is a hallmark of various diseases (Figure 3). For a long time it was believed that maladaptive pericytes mainly affect vascular homeostasis because pericyte and EC functions are interdependent and require bidirectional communication (Box 2). However, there is growing evidence that pericytes have roles in processes beyond the vasculature. As such, pericyte-derived signals (hereafter referred to as pericrine signaling) also modulate tissue function in both physiology and disease. In the following section, we capture recent data showing new observations that link pericyte dysfunction and loss in vascular and non-vascular-related diseases.

Figure 3 Dysfunctional pericytes in disease.

Show full captionFigure viewer

Box 2

Key signaling pathways that orchestrate pericyte–EC crosstalk

Given the close relationship between pericytes and ECs, it is not surprising that bidirectional communication and regulation between them are crucial during vessel formation and maintenance. During angiogenesis, established examples of pericyte–EC communication include the PDGFRβ, transforming growth factor β1 (TGF-β1), ANG1, and NOTCH3 pathways [1]. PDGF-B production from tip ECs is the master signal that recruits PDGFRβ-expressing pericytes to newly formed vessels [1], together with CD146 (MCAM), which acts as a coreceptor for PDGFRβ [96]. Recent advances have shown that NCK1 and NCK2 promote phosphorylation of PDGFRβ in response to PDGF-BB and stimulate pericyte migration by inducing MRTF translocation to the nucleus where they interact with the serum response transcription factor (SRF) [13,21]. Similarly, jagged 1 (JAG1) expressed by ECs activates NOTCH3 in pericytes and promotes pericyte maturation [14,97] and the expression‎ of PDGFRβ [98]. Conversely, ANG1 is secreted by pericytes, activates the tyrosine receptor TIE2 in ECs, and promotes EC maturation and vascular integrity [2,15]. TGF-β exerts complex effects on ECs and pericytes, and TGF-B receptor 1 (also known as ALK5) plays a dominant role in these interactions. Indeed, deletion of ALK5 in ECs leads to pericyte dysfunction and hemorrhagic vascular malformations [99]. Instead, deletion of ALK5 in pericytes results in increased EC proliferation, reduced collagen deposition, and enhanced matrix metalloproteinase activity [19]. Of note, pericytes also express canonical EC receptors such as VEGF-R1 [15,16] and TIE2 which allow pericytes to modulate intrinsic EC signaling.

The CNS: a hotspot of pericyte-related vascular diseases

Pericyte-related vascular defects have been reported in various CNS diseases, including Alzheimer's disease (AD), Parkinson's disease, dementia, stroke, diabetic retinopathy, glaucoma, and intracranial vascular malformations [5,37–39]. The involvement of pericytes in several CNS-related diseases is partially explained by their abundance within the brain vasculature and their key role in maintaining the BBB, where barrier breakdown precedes neurodegeneration. Other phenotypes linking pericytes dysfunction and CNS disease include neuron death [40] and impaired neurovascular coupling [41,42]. Intriguingly, NG2+ retinal pericytes orchestrate neurovascular coupling through closed-ended nanotubes between pericytes on adjacent capillaries, even when they are positioned far apart. These nanotubes terminate in a gap junction at the recipient pericyte, which permits rapid fluxes of small molecules and calcium ions, thereby allowing pericytes to coordinate neuronal activity [41]. Indeed, maintaining adequate calcium levels is essential to sustain pericyte function in the CNS, and aberrant levels of calcium in NG2+ pericytes lead to poor recovery after ischemic stroke [23,43] or neovascular dysfunction and neuronal death in glaucoma [42].

AD is the prototypical example of a CNS disease associated with aberrant vascular function and BBB breakdown linked to pericyte dysfunction and loss [38,44]. Although the involvement of pericytes in AD has been recognized for several years [5,44], new insights have challenged the timeframe in which patients suffering from AD develop pericyte dysfunction and BBB impairment. Indeed, it is now understood that BBB breakdown is an early event in AD, and these defects are used as an early biomarker of cognitive decline [45]. We highlight recent observations which support the involvement of pericytes in the onset of AD. For instance, Nortley et al. showed that the reduction in cerebral blood flow, that is considered to be the first clinical manifestation of AD, is caused by amyloid-β-induced pericyte contraction in brain capillaries [46]. Another study indicated that cognitive decline and BBB disruption in AD are linked to accelerated pericyte degeneration in carriers of AD susceptibility allele apolipoprotein E4 (APOE4) [47], a process which occurs independently of amyloid-β pathology. In this context, APOE4 carriers show high baseline cerebrospinal fluid levels of soluble (s)PDGFRβ which can be used as a BBB pericyte injury biomarker [47]. Intriguingly, analysis of the cortex of APOE4 transgenic mice using single-nucleus (sn)RNA-seq and phosphoproteomics revealed profound molecular changes related to progressive BBB failure in both ECs and pericytes [48]. Nonetheless, because only a constitutive APOE4-expressing transgenic line was included in the study, it remains unclear whether the molecular alterations of ECs and pericytes solely comprise cell-autonomous effects. In addition, one should not forget that mice do not fully recapitulate all traits of AD. It has been recently noted that pericytes and microglia associations (described in both physiological mouse and human brains) are diminished in the brain capillaries of individuals with AD, and this may also have implications for BBB breakdown [49]. In human brain, two types of pericytes have been identified that are distinguished by solute transport and ECM organization (Box 1). Intriguingly, the second type seems to be selectively affected in AD [50]. Thus, identifying methods to specifically target this cluster of pericytes may provide new ways to maintain vascular fitness in AD.

Pericyte degeneration and death also encompasses early phases of diabetic retinopathy, in which pericytes are primary targets of hyperglycemic damage. Recent findings suggest that, upon initiation of hyperglycemia, pericytes shift towards cell-bridging positions, resulting in physical detachment from ECs [51]. Whether this remodeling is independent of pericyte death or is related to the initiation of that process needs further investigation. Mechanistically, pericyte detachment and shifting are induced by exogenous factors such as angiopoietin 2 (ANG2) and PDGF-B, and are reversed by insulin treatment, illustrating the dynamic behavior of pericytes in the microvasculature [51,52]. In line with this, PDGFB signaling through PDGFRβ and NCKs in pericytes that cover sprouting vessels during experimental proliferative retinopathy [oxygen-induced retinopathy (OIR) model] activates ectopic αSMA expression‎ and promotes pathological neovascularization [21]. Interestingly, depletion of retinal pericytes in adulthood does not phenocopy retinopathy unless another stimulus is present (e.g., vascular endothelial growth factor A, VEGF-A). Upon depletion of pericytes, either during vessel development or in adulthood followed by VEGF addition, inhibition of ANG2 action restrains the severity of the diabetic retinopathy-like phenotypes [11]. Molecular effectors governing the early phases of diabetic retinopathy have remained elusive, precluding the development of drugs aiming to halt disease onset. These data suggest that targeting pericyte adhesion and migration capacities may be of therapeutic interest. Furthermore, pericyte loss in diabetic retinopathy was recently associated with aberrant levels of circular RNAs [53], thereby suggesting the use of circular RNAs as a diagnostic biomarker for early pericyte dysfunction in disease.

Finally, we would like to stress that familial mutations in essential pericyte genes have also been linked to CNS abnormalities. Well-known examples include loss-of-function mutations in NOTCH3 as a cause of CADASIL [54], and mutations in PDGFRB as a cause of brain calcifications [55], neurological deterioration, and white matter lesions [56]. Of note, these genes are equally relevant for pericyte and vSMC biology, and it is unclear whether these mutations lead to distinct phenotypes in mural cells. Current next-generation sequencing approaches allow the discovery of somatic mutations present in pericytes at low allelic frequency. In line with this, it has been proposed that PIK3CA- and AKT-related somatic cerebral cavernous malformations in mice emerge from mutant pericytes [39,57]. However, these data have some caveats because the lineage-tracing experiments used to support these findings were performed with a CRE-recombinase mouse line that is neither pericyte-specific nor inducible.

Pathobiological pericytes beyond the CNS

Although the implications of pericytes in diseases beyond the CNS are less well studied, the number of diseases demonstrating the involvement of pericytes continues to expand. We discuss here emerging evidence supporting a relevant role of pericytes in myocardial infarction [58], acute lung injury [59], and diabetes [60] as prototypical examples. For instance, after myocardial infarction, pericytes regulate inflammation and immune cell trafficking, and modulate ECM remodeling and revascularization [61]. In line with this, molecular reprogramming of PDGFRβ+NG2+ cardiac pericytes into vSMCs through inhibition of MEK1/2 improved the functional cardiac response by promoting revascularization [58]. In acute lung inflammation, the crosstalk between endothelium-derived nitric oxide (NO) and pericyte soluble guanylate cyclase (sGC) is impaired, leading to elevated vascular permeability [59]. Pharmacological activation of the NO–sGC axis led to an improved pericyte-driven inflammatory response. Moreover, pericytes in pancreatic islets exert vascular control of hormone secretion and glucose homeostasis, and pericyte alteration has been linked to diabetic islet dysfunction [60]. Interestingly, pericyte effects on islet functionality are not limited to vascular support for insulin secretion because pancreatic β cell maturation and functionality rely on pericyte-derived bone morphogenic protein 4 (BMP4). Recently, other pericrine signaling molecules have been identified as key players in the functional regulation of tissue parenchyma encompassing a range of organ-specific functions in both vascular and non-vascular interfaces. Two interesting examples are that leptin receptor-expressing pericytes in the mediobasal hypothalamus mediate energy balance via neuronal leptin signaling [62], and that the Hippo–YAP/TAZ pathway in pericytes generates essential pericrine signals to epithelial and ECs during lung morphogenesis [63]. All things considered, these studies suggest that restoring the physiological functions of pericytes improves blood vessel performance and disease outcomes in various contexts, which may encourage the development of novel pericyte-focused therapies.

Tumor pericytes: loss or change of identity?

Many preclinical studies over the past decade showed that pericyte dysfunction is involved in cancer progression [64]. In this context, it is now well established that tumoral vessels are poorly covered by pericytes [65,66], that diminished pericyte recruitment and maturation lead to enhanced vascular and tumor growth [15,66], and that pericyte depletion favors metastasis. However, there is no consensus about whether defective pericyte–EC interaction in tumors relates to loss of pericyte molecular identity [67], pericyte transdifferentiation into fibroblasts [68], poor pericyte recruitment [66], or a combination of these phenotypes. Functionally, metabolic reprogramming of tumor pericytes has also been implicated in abnormal blood vessel contraction and unfavorable patient outcomes [69].

scRNA-seq analyses have provided a new layer of information about tumor pericytes by showing that tumor pericytes are, in fact, relatively homogeneous [70–72]. However, one should acknowledge that most tumor scRNA-seq datasets include low numbers of pericytes, and the presence of tumor pericyte subclusters may have been obscured by the granularity of the data. One way to overcome this limitation would be to establish spatially resolved pericyte-focused atlases by using multiomic approaches. An interesting observation is that pericyte phenotypes are distinct depending on the mechanisms by which tumor vessels form. While angiogenic pericytes persist in an active immature state characterized by a signature of motility and ECM organization, pericytes covering co-opted vessels remain largely quiescent [71,73]. Together, these studies support a model in which pericytes undergo genetic and molecular reprogramming in cancer, which in turn has negative consequences for disease progression. Nonetheless, one should not forget that most of these data refer to mouse preclinical studies, and adequate longitudinal studies in humans are missing. Of note, the so-called pericrine signaling response is also at play in cancer. Indeed, in hepatocellular carcinoma it has been shown that metabolic reprogramming in tumor cells activates HSCs which in turn promote tumorigenesis through the secretion of senescence-associated factors [74]. Another study has described that loss of integrin β3 in tumor pericytes leads to enhanced focal adhesion kinase (FAK)-mediated cytokine release by pericytes, which subsequently stimulates tumor survival and growth [75].

Pericyte immunomodulatory properties

Emerging evidence suggests that pericytes form an integral part of the immune surveillance unit rather than solely performing complementary functions. Upon proinflammatory stimuli, pericytes promote endothelial expression‎ of the leukocyte adhesion molecules vascular cell adhesion protein 1 (VCAM-1) and/or intracellular adhesion molecule 1 (ICAM-1) in the CNS, lung, skin, or muscle that subsequently promote T cell or macrophage infiltration into the affected tissue [59,76,77]. The Rgs5+ and Col1a1+ subgroups of PDGFRβ+ perivascular cells seem to be early responders to neuroinflammation [78]. Of note, inflammation per se induces pericyte detachment from the endothelium and impairs barrier properties [59]. In addition to the physical interaction with leukocytes, pericytes also secrete and respond to cytokines that further regulate immune cell functions, including both innate and adaptive responses. The chemotactic migration and effector functions of neutrophils, T cells, and macrophages are dependent on these early pericrine signals, including MIF, CXCL1, and CCL2 [77–80].

It is now believed that modulation of the immune-related functions of pericytes could affect the outcome of disease progression. For example, pericyte-deficient Pdgfbret/ret mice exhibit increased leukocyte infiltration and activation, leading to aberrant inflammation in a model of experimental autoimmune encephalomyelitis [76]. Treatment with antagonistic VCAM-1 and ICAM-1 antibodies partially rescued the excessive inflammatory phenotype in Pdgfbret/ret mice. Similarly, activating sGC in acute lung injury improved disease outcomes by increasing pericyte interaction with ECs [59]. Conversely, in the tumor microenvironment, tumor cells induce autophagy of NG2+ pericytes which equips them with immunosuppressive properties that favor tumor cell survival and prevent antitumor T cell responses [81]. Pericytes have also been implicated in the underlying pathophysiology of emergent infectious diseases such as coronavirus disease 2019 (COVID-19) [82], thus presenting an additional niche of investigation for the field in the coming years. In conclusion, the importance of pericytes during various inflammatory processes is growing in recognition, although the modulation of immunity by pericytes can have double-edged outcomes.

Do pericytes contribute to fibrosis?

A dysregulated tissue repair response after acute or chronic injury can lead to the onset of fibrosis associated with the abnormal accumulation of activated and contractile αSMA+ myofibroblasts [83]. Myofibroblasts secrete high amounts of inflammatory mediators, growth factors, and ECM components, and promote aberrant ECM remodeling. The current consensus places fibroblasts as the predominant origin of myofibroblasts, although various studies have found alternative cellular origins [83]. Indeed, the existence of pericyte-to-myofibroblast transition has been proposed as a contributing factor in several fibrotic contexts [2], but the promiscuity of cell markers within the mesenchymal compartments across tissues has led to ambiguous and contradictory observations. In the latest developments, scRNA-seq, ATAC-seq, and spatial transcriptomics provide new insights into this conundrum which support a pericyte origin of myofibroblasts in the fibrotic liver, colon, and kidney. For instance, central vein-associated Rgs5+ HSCs are thought to be the dominant origin of ECM-producing myofibroblasts in fibrotic mouse liver [84]. In human colorectal cancer, a subset of periostin (POSTN)+ myofibroblasts seem to originate from RGS5+ pericytes [70]; similarly, NOTCH3+RGS5+PDGFRβ+ human pericytes contribute to the generation of POSTN+PDGFRα+NKD2+ myofibroblasts during kidney fibrosis [85]. Surprisingly, however, the same authors did not capture the existence of profibrotic pericytes during myocardial infarction when sequencing the entire heart [86]. Although transdifferentiation from pericytes to myofibroblasts may be tissue-specific, it is fair to acknowledge that the latter study did not include pericytes in the trajectory analysis that predicted the origins of myofibroblasts.

Lineage-tracing experiments in mice have also cast some light onto the role of pericytes in fibrosis. For instance, Pham et al. showed that myofibroblast genes are enriched in Tbx18+ pericytes from injured mouse hearts and brains [87], and Dias et al. reported that GLAST+PDGFRβ+ perivascular cells also contribute to fibrosis in the post-stroke brain [88]. In line with this, depletion of GLAST+PDGFRβ+ perivascular cells in the spinal cord leads to reduced fibrotic scar after injury and improves neuronal function [89]. Of note, GLAST+PDGFRβ+ perivascular cells were characterized as spinal cord pericytes by the researchers, but there is insufficient evidence to rule out that these cells might be fibroblasts or astrocytes. This further exemplifies that the shared marker expression‎ profiles of pericytes and other perivascular residing cells still hamper the design of robust pericyte reporter models. Moreover, multiple studies have found no significant pericyte origin for myofibroblasts in distinct fibrosis models of the CNS [90,91] and heart [92]. Although the discrepancies between studies may reflect the lack of robust pericyte identification strategies, it appears that the occurrence of pericyte-to-myofibroblast transition is tissue-specific, and is contingent both on the local microenvironment and on the extent of the injurious stimuli [83]. This emphasizes the need to further unravel the fibrotic pericyte responses in different organs and prompts the question of how myofibroblast origins relate to different pathophysiological phenotypes. All things considered, the origin of myofibroblasts may involve distinct precursor cells depending on the circumstances, although a role for pericytes seems to be indisputable (Figure 4).

Figure 4 Pericytes as a source of myofibroblasts in fibrosis.

Show full captionFigure viewer

Concluding remarks

It is widely recognized that pericytes play an important role in blood vessel formation, stabilization, and function, and that degeneration or loss of brain pericytes impairs their protective barrier properties. Recent advances have revealed novel and crucial roles for pericytes across tissues in a variety of vascular and non-vascular processes. Single-cell technology is becoming more commonly used to better understand the molecular processes that define pericytes in health and disease. Although the molecular and functional attributes of pericytes are not fully elucidated, deep sequencing has revealed organotypic pericyte heterogeneities and new criteria to distinguish pericytes from other cell types. Despite the numerous suggested roles of pericytes in various diseases and physiological processes, including neurodegeneration, cancer, fibrosis, blood flow regulation, and inflammation, the underlying organotypic mechanisms of these contributions are not yet fully understood. The discrepancies between some studies highlight the importance of designing suitable mouse models for eval‎uating the specific mechanisms by which pericytes impact on these processes, and future cross-validations with human data are warranted to ascertain the clinical relevance of pathological pericytes (see Outstanding questions). Overall, based on the emerging evidence on the contribution of pericytes to several diseases, we anticipate an increasing emphasis on pericyte-oriented research in vascular (and non-vascular) studies in the coming years.

Outstanding questions

Why are pericytes molecularly and functionally promiscuous and heterogeneous between distinct tissues? If the transcription factors that regulate pericyte differentiation and function are similar across tissues, will epigenetic mechanisms provide cues into the mechanisms of pericyte identity?

Will spatial molecular atlases focused on pericytes address the key functional and identity conundrums posed by transitioning phenotypes? Will this suffice, or are complementary morphological and functional studies also necessary?

Will pericyte-focused therapy provide new means to stimulate functional angiogenesis in pathology? Given that pericytes are associated with many diseases, will (and how broadly can) pericyte-focused therapies improve patient outcomes?

Neurodegenerative diseases are age-related diseases, and pericyte degeneration is an aging process. Hence, is pericyte degeneration a confounding factor in the development of age-related neurodegeneration? Will maintaining healthy pericyte function promote healthy aging?

AcknowledgmentsAcknowledgments

We would like to thank Sandra D. Castillo, Ana Angulo-Urarte, and Leonor Gouveia for their valuable feedback. Figures were created with BioRender.com. We thank the Centres de Recerca de Catalunya (CERCA) Program/Generalitat de Catalunya and the Josep Carreras Foundation for institutional support. Work-related to this publication in the laboratory of M.G. is supported by research grants from la Asociación Española Contra el Cancer (AECC)-Grupos Traslacionales (GCTRA18006CARR); and by Worldwide Cancer Research (WCR 21-0159). H.v.S. received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 955951; P.V. recieved funding from AECC (AECC-INVES211084VILL) and from the Spanish Ministry of Science and Innovation (RYC2020-029929-I). We apologize to the many authors whose primary papers could not be cited owing to space constraints.

Declaration of interests

The authors declare no conflicts of interest.

References

1.

Armulik, A. ...

Pericytes: developmental, physiological, and pathological perspectives, problems, and promises

Dev. Cell. 2011; 21:193-215

Full Text

Full Text (PDF)

Scopus (1996)

PubMed

Google Scholar

2.

Holm, A. ...

Microvascular mural cell organotypic heterogeneity and functional plasticity

Trends Cell Biol. 2018; 28:302-316

Full Text

Full Text (PDF)

Scopus (85)

PubMed

Google Scholar

3.

Petrova, T.V. ∙ Koh, G.Y.

Biological functions of lymphatic vessels

Science. 2020; 369, eaax4063

Crossref

Scopus (234)

PubMed

Google Scholar