Re: 인슐린 100주년 기념 2022 nature, 만성염증 관련 질환 2022년 Cell press 리뷰 논문..

[문형철]

이 논문은

비만(obesity), 제2형 당뇨병(T2D),

그리고 관련 합병증에서 나타나는 만성 저등급 염증(chronic low-grade inflammation)의 메커니즘을

체계적으로 검토한 리뷰 논문입니다.

주요 내용은

비만으로 인한 조직 특이적 염증이

어떻게 인슐린 저항성(insulin resistance, IR),

베타세포 기능 장애,

그리고 전신 합병증으로 이어지는지를 분자·세포·조직 수준에서 다룹니다.

1. 핵심 요약 (Summary)

비만은

지방조직(adipose tissue)의 과도한 에너지 축적으로 인해

만성적인 면역 활성화를 유발합니다.

이는 초기에는 적응적(energy storage 보호)이지만,

지속되면 병리적 염증으로 전환되어:

• 인슐린 저항성(IR) → glucose intolerance
• 췌장 베타세포(β-cell) 기능 저하 → T2D 발병
• NAFLD(비알코올성 지방간), 심혈관질환, 신장병, 망막병증 등 합병증
• 알츠하이머, PCOS, 통풍, 류마티스 관절염 등과의 연관성

을 초래합니다.

논문은

이 염증의 조직 특이적 메커니즘과 치료적 타겟을 중점적으로 논의합니다.

1. 간(Liver) 부위

• Lean liver: Kupffer cells (KCs, 간 고유 대식세포)이 주로 존재. 이들은 주로 M2-like (항염증성) phenotype을 유지하며 조직 항상성을 유지.
• NAFLD (비알코올성 지방간):
  • Monocyte-derived recruited hepatic macrophages (RHMs)가 대폭 증가.
  • M1-like RHMs↑ (pro-inflammatory phenotype)
  • CD9⁺ Trem2⁺ macrophage subpopulation이 특징적으로 증가 (이들은 lipid-laden, pro-fibrotic 특성을 가짐).
• 결과: 간 내 염증 증폭 → 지방축적, 섬유화, gluconeogenesis 증가.

2. 지방조직 (Adipose Tissue, AT) 부위

• Lean AT:
  • Adipocyte 주위에 M2-like ATMs (adipose tissue macrophages)가 존재 (CD206⁺ marker).
  • 조직 항상성과 lipid buffering 역할을 함.
• Obese AT:
  • Adipocyte hypertrophy → dead/dying adipocytes 증가.
  • Crown-like structures (CLS): 죽은 지방세포 주위에 대식세포가 둘러싸는 전형적인 병리학적 구조.
  • M1-like ATMs↑ (CD11c⁺ 또는 CD9⁺)
  • LAMs (Lipid-associated macrophages)↑ (CD9⁺ Trem2⁺) — 최근 주목받는 subpopulation으로, lipid uptake와 lysosomal remodeling에 관여하지만 만성적으로는 염증 증폭.
• 결과: 지방조직 내 강한 pro-inflammatory milieu 형성 (TNF-α, IL-1β, IL-6 등 사이토카인 폭발).

3. 공통 병태생리 과정

• Monocytes (혈액에서 유래)가 조직으로 recruitment되어 지방조직과 간의 염증성 macrophage pool을 확대.
• 하단에 표시된 pancreas, gut, spleen 등 다른 장기도 조직 염증에 기여.
• 최종 결과: Tissue inflammation↑ + Macrophage activation↑ → Insulin resistance & Glucose intolerance

2. 비만에서 염증의 발생 기전 (Mechanisms of Inflammation in Obesity)

비만 시 지방조직에서 일어나는 주요 변화:

• 지방세포(adipocyte) hypertrophy: 세포가 커지면서 hypoxia(저산소증), ER stress(소포체 스트레스), lipotoxicity(지방 독성)가 발생. 이는 DAMPs(damage-associated molecular patterns)를 방출하여 innate immune activation을 유발.
• 면역세포 침윤:
  • Macrophage: M1-like pro-inflammatory phenotype으로 전환 (crown-like structures 형성). TNF-α, IL-6, IL-1β 등 사이토카인 과다 분비.
  • T cells: Th1, Th17 증가; Treg 감소 → imbalance.
  • B cells, ILCs(inmate lymphoid cells) 등도 관여.
• 신호 경로: JNK, NF-κB, NLRP3 inflammasome 활성화. 특히 IL-1β가 핵심 mediator로, IRS-1(insulin receptor substrate-1) serine phosphorylation을 통해 insulin signaling을 억제.
• 기타 조직: 간(hepatocyte), 골격근, 췌장, 시상하부(hypothalamus)에서도 유사한 low-grade inflammation 발생. 예: hypothalamic inflammation → leptin resistance 및 energy imbalance 악화.

이 과정은 meta-inflammation(metabolic inflammation)으로 불리며, 전통적 급성 염증과 달리 지속적·저강도입니다.

이 그림은엑소좀(exosomes)이
비만·당뇨병에서 어떻게 세포 간(intercellular)
그리고 장기 간(interorgan) crosstalk을 매개하는지를 보여주는 핵심 모식도.

특히 adipose tissue, liver, pancreatic islet에서 유래한 exosomes가
특정 microRNA(miRNA)를 통해 insulin sensitivity, NASH, β-cell apoptosis 등을 조절하는 메커니즘을 강조.

1. Adipose Tissue (지방조직)에서의 Exosome 역할

• M2-like ATMs (항염증성 대식세포):
  • miR-690를 포함한 exosomes 분비
  • 효과: Insulin sensitivity ↑ (전신 인슐린 감수성 향상)
• M1-like ATMs (염증성 대식세포):
  • miR-155를 포함한 exosomes 분비
  • 효과: Insulin sensitivity ↓ (인슐린 저항성 악화)

→ 지방조직 macrophage phenotype에 따라 exosomes의 내용물이 달라지며,
    이는 protective vs. pathogenic 신호를 다른 장기로 전달합니다.

2. Liver (간)에서의 Exosome 역할

• 정상 Hepatocytes:
  • miR-3075 포함 exosomes
  • 효과: Insulin sensitivity ↑
• Steatotic hepatocytes (지방축적 간세포, NAFLD 상태):
  • miR-434-5p 포함 exosomes → Insulin sensitivity ↓
  • miR-128-3p, miR-1 포함 exosomes → NASH progression ↑ (염증·섬유화 촉진)

→ 지방간이 진행되면서
   hepatocyte-derived exosomes의 miRNA profile이 바뀌어,
   인슐린 저항성을 악화시키고 비알코올성 지방간염(NASH)으로의 진행을 가속화.

3. Pancreatic Islet (췌도)에서의 Exosome 역할

• Stressed β-cells (스트레스 받은 베타세포):
  • miR-21-5p 포함 exosomes 분비
  • 효과: β-cell apoptosis ↑ (베타세포 사멸 증가)

→ 고혈당·지방독성 등 스트레스 하에서 베타세포가
    miR-21-5p를 통해 자가 사멸을 증폭시키는
    autocrine/paracrine 신호를 보냅니다.

3. 염증이 인슐린 저항성과 T2D로 이어지는 과정

• Peripheral IR: Adipose tissue, liver, muscle에서 insulin signaling (PI3K-Akt pathway) 억제 → glucose uptake 감소, gluconeogenesis 증가.
• β-cell dysfunction: 만성 IL-1β 노출 → β-cell apoptosis, dedifferentiation. 초기에는 compensatory hyperinsulinemia지만, 결국 exhaustion.
• Glucose intolerance → overt T2D: 염증이 vicious cycle을 형성 (hyperglycemia 자체가 NLRP3 활성화 → 더 많은 IL-1β).

논문은 tissue-specific 접근을 강조:

예를 들어, visceral fat vs. subcutaneous fat의 염증 정도 차이, 또는 liver-specific Kupffer cell 역할 등.

Figure 3 요약: Islet Inflammation (췌도 염증)

이 그림은
비만 → 전당뇨 → 제2형 당뇨병으로 진행되면서
췌장 췌도(pancreatic islet)에서 발생하는 염증의 시간적·단계적 변화를 보여주는 모식도.

핵심은
resident macrophage → recruited inflammatory macrophage의 전환과,
IL-1β / IL-1Ra 불균형 및 amyloid deposition이
베타세포 기능 저하와 insulin secretion failure로 이어지는 과정.

1. Normal (정상) 상태

• 췌도 내 resident macrophages (조직 상주 대식세포)가 소수 존재.
• 이들은 IL-1Ra (IL-1 receptor antagonist)를 주로 분비 → IL-1β의 과도한 신호를 억제하여 β-cell protection 역할.
• Insulin 분비: 정상 수준 (→).
• 전체적으로 anti-inflammatory, homeostatic microenvironment 유지.

2. Pre-diabetes (전당뇨) 단계

• Macrophages ↑ (혈액 유래 monocyte recruitment 증가).
• miR-21-5p ↑ (macrophage-derived exosome를 통해 β-cell에 전달, 이전 Figure 2와 연계).
• Amyloid deposition ↑ (IAPP/islet amyloid polypeptide 축적).
• IL-1β ↑, IL-1Ra는 아직 유지되거나 약간 증가.
• Insulin: 보상성 과분비 (↑) — β-cell이 초기 스트레스에 대해 hyperfunction.
• 특징: 아직 overt hyperglycemia는 없으나, low-grade islet inflammation이 시작되는 tipping point.

3. Diabetes (당뇨병) 단계

• Macrophages 대폭 증가 (강한 recruitment).
• miR-21-5p ↑↑, Amyloid ↑↑.
• IL-1β ↑↑ (강력한 pro-inflammatory cytokine).
• IL-1Ra ↓ (보호 인자의 감소) → IL-1 signaling이 unchecked.
• Insulin: 분비 저하 (↓) — β-cell exhaustion 및 apoptosis.
• 결과: β-cell mass 감소, dedifferentiation, 기능 부전.

4. 당뇨 합병증과 동반 질환에서의 염증 역할

• Macrovascular/CV disease: Endothelial dysfunction, atherosclerosis plaque formation (macrophage foam cells).
• Microvascular: Retinopathy, nephropathy (glomerular inflammation).
• NAFLD/NASH: Kupffer cell activation, stellate cell fibrosis.
• Neuroinflammation: Alzheimer’s disease와의 link (brain insulin resistance, microglial activation).
• 기타: Gout (IL-1β-driven crystal inflammation), RA overlap 등.

염증은 T2D의 공통 병태생리로, 합병증 간 연관성을 설명합니다.

이 그림은
비만·제2형 당뇨병의 만성 염증(meta-inflammation)이
어떻게 다양한 합병증으로 이어지는지를 중심으로,
면역조절(immunomodulatory) 치료 타겟을 체계적으로 정리한 모식도.

중앙의 대사 스트레스 요인들이
여러 사이토카인·염증 경로를 활성화시켜 장기별 합병증을 유발하고,
이를 약물적으로 타겟팅할 수 있는 가능성을 보여줍니다.

1. 중앙 대사 스트레스 요인 (Central Drivers)

• Glucose (고혈당)
• sFA (saturated fatty acids, 포화지방산)
• Hypoxia (저산소증)
• Cholesterol (콜레스테롤)

이 요인들은 NLRP3 inflammasome, NF-κB 등을 활성화시켜 IL-1β, IL-6, TNF-α 등의 pro-inflammatory mediator를 대량 생산합니다.

2. 주요 합병증과 핵심 면역 타겟

• Heart failure (심부전) → IL-1β, NLRP3
• Retinopathy (망막증) → IL-1β, NLRP3, IL-6
• NASH (비알코올성 지방간염) → TNF-α, NLRP3, IL-1β, IL-6
• Nephropathy (신장병) → IL-1β
• Neuropathy (신경병증) → IL-6
• Atherosclerosis (동맥경화) → IL-1β, NLRP3, IL-6 (Possible, Validated by clinical trials)
• Defective insulin secretion (인슐린 분비 장애) → IL-1β, NLRP3, IL-22, IL-33
• Insulin resistance (인슐린 저항성) → IL-1β, NLRP3, IL-22, IL-33

3. Graduate-level 핵심 Insight

1. 공통 병태생리 축: 거의 모든 합병증에서 IL-1β / NLRP3 inflammasome이 핵심 공통 mediator로 등장합니다. 이는 고혈당·지방독성·저산소가 NLRP3를 직접 활성화시켜 IL-1β maturation을 유발하기 때문입니다.
2. IL-1 signaling의 중심성:
   • 췌도 β-cell 사멸 (Figure 3)
   • 혈관 내피 기능 장애
   • 대식세포 foam cell 형성
   • 신장 mesangial cell activation 등 에서 광범위하게 작용합니다.
3. Clinical Evidence 수준 표시:
   • Atherosclerosis 부위에 “Possible Validated by clinical trials” 표시 → CANTOS trial (canakinumab, IL-1β monoclonal antibody)에서 심혈관 사건 감소가 입증된 대표 사례.
   • 다른 합병증들은 아직 preclinical~early clinical 단계이지만, 기전적으로 매우 유망함.
4. 치료 전략 방향:
   • Specific targeting: IL-1β blocker (canakinumab, anakinra), NLRP3 inhibitor, IL-6 inhibitor (tocilizumab), TNF-α inhibitor 등.
   • Multi-target: GLP-1 receptor agonist (semaglutide 등)는 체중 감소 + 직접적 anti-inflammatory 효과를 동시에 가짐.
   • Combination therapy: 기존 대사약물 + immunomodulator.

5. 치료적 함의와 Clinical Trials

논문은 anti-inflammatory therapy의 가능성을 강조:

• IL-1β antagonism (canakinumab 등): β-cell 보호, CV event 감소 (CANTOS trial 등 참조).
• 기타 타겟: TNF-α, IL-6, JAK inhibitors, NLRP3 inhibitors.
• 기존 약물의 anti-inflammatory 효과: GLP-1RA (semaglutide 등)는 체중 감소 외에 직접적 면역 조절 효과. SGLT2i도 일부 염증 감소.
• 한계: Non-specific immunosuppression (e.g., methotrexate)는 실패. Mechanism-based targeting이 중요.

Table 1에 다양한 clinical trials 요약이 있으며, IL-1 pathway targeting이 가장 유망합니다.

6. 결론 및 Future Directions (Concluding Remarks)

비만과 T2D는 이제 chronic inflammatory disease로 재정의됩니다.

초기 adaptive inflammation이 chronic pathogenic state로 전환되는 tipping point를 이해하는 것이 핵심.

미래 연구 방향:

• Single-cell RNA-seq 등으로 tissue-specific immune landscape mapping.
• Personalized anti-inflammatory therapy (biomarker-driven).
• Combination therapy (GLP-1 + IL-1 blocker 등).

이 논문은 면역-대사 crosstalk의 최신 understanding을 종합하며, translational research의 중요한 bridge 역할을 합니다. 특히 Immunity 저널 특성상 molecular immunology 관점에서 깊이 있게 다룹니다.

이 논문은

인슐린 저항성(Insulin Resistance, IR)의 분자 메커니즘을 체계적으로 정리하고,

조직 간 crosstalk, 비유전적 요인, 치료 전략까지 포괄적으로 다룬 리뷰 논문.

Insulin 발견 100주년을 맞아

IR을 단순한 ‘대사 장애’가 아닌 다중 경로(multifactorial) 질환으로 재정의하며,

translational 연구 방향을 제시.

1. 핵심 개념 및 Insulin Signaling Pathway

• IR 정의: Insulin 자극에 대한 생물학적 반응이 손상된 상태.
• 주요 표적 조직(liver, skeletal muscle, adipose tissue)에서 insulin signaling pathway의 다양한 단계가 붕괴됨.
• 주요 Signaling:
  • Insulin → INSR (Insulin Receptor) → IRS (IRS1/2) → PI3K → AKT
  • Downstream 효과:
    • Glucose uptake: GLUT4 translocation (muscle, adipose)
    • Hepatic gluconeogenesis 억제: FOXO1 phosphorylation
    • Glycogen synthesis: GSK3 inhibition
    • Lipogenesis: SREBP-1c, ChREBP 활성화
    • Lipolysis 억제: PDE3B, ABHD15

IR 발생 시:

Proximal (INSR, IRS, PI3K, AKT) 및 Distal (GLUT4 vesicle trafficking, FOXO1 nuclear retention 등) 단계 모두에서 defect 발생.

이 그림은
인슐린 저항성(Insulin Resistance, IR)이
간(Liver), 골격근(Skeletal muscle), 지방조직(Adipose tissue)에서
어떻게 동시에 발생하고 상호 연결되는지를 한 장에 통합적으로 보여주는 매우 핵심적인 모식도.

중심에 Overnutrition (과영양) → Glucose ↑가 자리 잡고,
대식세포(Macrophage)에서 유래한 염증 신호(TNF-α, IL-1β 등)가
모든 조직을 연결하는 meta-inflammation의 네트워크를 명확하게 드러냅니다.

1. 전체 구조와 공통 메커니즘

• 중심 Trigger: Overnutrition → 고혈당 + 지방산 과다
• 공통 Inhibitory Pathway:
  • JNK activation (c-Jun N-terminal kinase)
  • IRS1/2 serine phosphorylation → insulin signaling 차단
  • DAG (diacylglycerol) / PKC (Protein kinase C) activation
  • Ceramide 축적
  • TNF-α / TNFR signaling → NF-κB, JNK 루프

이 신호들은 insulin → INSR → IRS1/2 → PI3K → pAKT 경로를 광범위하게 억제합니다.

2. 조직별 Insulin Resistance 메커니즘 ① Liver (간) – Hepatic Insulin Resistance

• Insulin이 IRS1/2 → pAKT2를 통해 FOXO1을 인산화하지 못함 → FOXO1 nuclear retention → Gluconeogenesis ↑ (PC, PEPCK 등 효소 발현 증가)
• Glycogen synthesis ↓
• 지방산 → β-oxidation 증가하지만, 과잉 시 DAG/PKCε 활성화 → IR 악화
• 결과: 고혈당 유지 + VLDL 분비 증가

② Skeletal Muscle – Muscle Insulin Resistance

• Insulin → IRS1/2 → pAKT2 경로 억제
• GLUT4 vesicle (GSV) translocation 실패 → glucose uptake ↓
• Glycogen synthesis ↓
• 지방산 축적 → ceramide, DAG/PKCδ → JNK 활성화
• PP2A activation → pAKT dephosphorylation

③ Adipose Tissue – Adipose Insulin Resistance

• pAKT ↓ → PDE3B ↓ → cAMP/PKA ↑ → lipolysis ↑ (HSL activation)
• Fatty acids + Glycerol 방출 → 다른 조직으로 lipotoxicity 전달
• SREBP1c ↓ → De novo lipogenesis ↓ (역설적으로 초기에는 증가하나 만성 IR에서는 억제)
• TAG (triacylglycerol) 축적

3. Macrophage와 Inter-tissue Crosstalk

• Macrophage (지방조직 내 crown-like structure 등에서 활성화)
  • TNF-α, IL-1β, 기타 cytokines 분비
  • 모든 조직(간, 근육, 지방)에 paracrine/endocrine 방식으로 작용
  • JNK, TNFR signaling을 통해 IRS inhibitory phosphorylation 증폭
• Bidirectional vicious cycle: Adipose lipolysis → ectopic lipid (ceramide, DAG) → 다른 조직 IR → 더 많은 hyperglycemia → macrophage activation

2. IR의 주요 메커니즘 (Mechanisms)

논문은 IR을 직접적 signaling defect와 간접적 요인으로 분류:

• 직접적 Insulin Signaling Defect:
  • INSR downregulation / kinase activity 저하
  • IRS serine phosphorylation (JNK, IKK 등에 의한 inhibitory phosphorylation)
  • PI3K/AKT pathway impairment
  • GLUT4 translocation failure
• 역학적·환경적 요인:
  • Sex difference (premenopausal female protective effect via estrogen/ERα)
  • Ethnicity difference (South Asian, Aboriginal populations에서 높은 IR preval‎ence)
  • Lifestyle: high-fat diet, sedentary life, sleep disruption, stress, smoking
• Interorgan Metabolic Crosstalk:
  • Adipose tissue → liver/muscle로의 adipokine (adiponectin↓, leptin resistance), lipotoxic lipids (ceramide, diacylglycerol), exosome-mediated 신호
  • Liver → peripheral tissues (hepatic steatosis → systemic IR)
  • Muscle → adipose/hepatic feedback
• Metabolic Mediators:
  • Inflammation (TNF-α, IL-6, IL-1β → NF-κB, JNK activation)
  • Oxidative stress, ER stress, mitochondrial dysfunction
  • Ectopic lipid accumulation (lipotoxicity)
• 유전·후성유전적 요인:
  • Genetic mutation (INSR, IRS, AKT 등)
  • Epigenetic dysregulation (DNA methylation, histone modification)
  • Non-coding RNAs: miRNA (miR-155, miR-33 등), lncRNA (H19, MALAT1), circRNA (circHIPK3 등) — IR과 β-cell dysfunction 조절
  • Gut Microbiota Dysbiosis: Firmicutes/Bacteroidetes ratio 변화, SCFA 생산 감소, LPS (endotoxemia) → systemic inflammation

이 그림은
비만(obesity)을 중심으로,
지방독성(lipotoxicity),
만성 저등급 염증(meta-inflammation),
인슐린 저항성(insulin resistance), proinflammatory adipocytokines가
어떻게 전신 여러 장기에 병리적 영향을 미쳐 다양한 합병증을 유발하는지를
한눈에 보여주는 통합적 개요도.

비만을 단순한 ‘체중 증가’가 아닌
다장기(multisystem) 대사-염증 질환으로 재정의하는 데
핵심적인 시각 자료입니다.

1. 중심 병태생리 (Core Pathophysiology)

• Obesity → Proinflammatory Adipocytokines (TNF-α, IL-6, leptin resistance, adiponectin ↓ 등) 분비 증가
• Lipotoxicity + Glucose intolerance → Type 2 Diabetes 악화
• 이 두 축이 insulin resistance를 전신적으로 증폭시키며, vicious cycle을 형성

2. 장기별 주요 합병증 심혈관계 (Cardiovascular)

• Artery: Atherosclerosis (동맥경화), Hypertension (고혈압)
• Heart: Cardiomyopathy (심근병증) — diabetic cardiomyopathy 포함
• 기전: Endothelial dysfunction, foam cell formation, cardiac lipotoxicity, fibrosis

간 (Liver)

• MAFLD (Metabolic dysfunction-associated fatty liver disease, 과거 NAFLD)
  • Steatosis → Steatohepatitis (NASH) → Cirrhosis (간경변)
• ↑ Insulin resistance + ↓ Lipid metabolism
• Lipotoxicity와 inflammation이 핵심 driver

신장 (Kidney)

• Hyperfiltration (과여과) → Chronic Kidney Disease (CKD)
• 기전: Glomerular hypertension, podocyte injury, tubulointerstitial inflammation

중추신경계 (Central Nervous System)

• Impaired energy homeostasis (시상하부 leptin resistance)
• Disturbed brain development (특히 소아비만)
• Alzheimer’s disease (brain insulin resistance = “Type 3 diabetes”)

암 (Cancer)

• breast, uterus, cervix, colon, esophagus, pancreas, kidney, prostate 등 다수 solid tumor 위험 증가
• 기전: Hyperinsulinemia (mitogenic effect), chronic inflammation, adipokine imbalance, sex hormone dysregulation

부인과/생식계 (Gynecologic abnormalities)

• Polycystic Ovary Syndrome (PCOS)
  • Androgen production 증가
  • Abnormal menses, infertility, polycystic ovarian morphology

기타

• Skeletal muscle: Lipotoxicity → glucose intolerance
• 전반적 metabolic dysfunction-associated fatty liver disease와의 연계 강조

3. IR 관련 질환

• Obesity, T2DM, MAFLD/NASH, Cardiovascular disease, PCOS, Cancer 등
• IR이 공통 병태생리로 작용하며, vicious cycle 형성

4. IR 측정 및 Biomarker

• Gold standard: Hyperinsulinemic-euglycemic clamp
• Surrogate: HOMA-IR, QUICKI, OGTT-derived indices, TyG index 등
• Emerging biomarkers: Adipokines, inflammatory cytokines, exosomal miRNAs, gut microbiota signatures

5. 치료 전략 (Therapeutic Approaches)

• Lifestyle Intervention: Calorie restriction, Mediterranean diet, aerobic + resistance exercise (가장 기본적이고 효과적)
• Pharmacological Interventions:
  • Metformin (AMPK activation)
  • GLP-1RA, SGLT2i (체중 감소 + 직접적 IR 개선)
  • Thiazolidinediones (PPARγ agonist, adipocyte remodeling)
  • Emerging: PTP1B inhibitor, IKK/JNK inhibitor, anti-inflammatory agents, miRNA-based therapy, gut microbiota modulation (probiotics, FMT)
• 미래 방향: Personalized medicine (genetic/epigenetic profiling 기반), multi-organ targeting combination therapy

ReviewVolume 55, Issue 1p31-55January 11, 2022Open Archive

Download Full Issue

Inflammation in obesity, diabetes, and related disorders

Theresa V. Rohm1,4 ∙ Daniel T. Meier2,3,4 ∙ Jerrold M. Olefsky1,4 ∙ Marc Y. Donath2,3,4 Send email to marc.donath@usb.ch

Affiliations & Notes

1Division of Endocrinology and Metabolism, Department of Medicine, University of California, San Diego, La Jolla, CA 92093, USA

2Clinic of Endocrinology, Diabetes and Metabolism, University Hospital Basel, Basel 4031, Switzerland

3Department of Biomedicine (DBM), University of Basel, University Hospital Basel, Basel 4031, Switzerland

4

These authors contributed equally

Article Info

DOI: 10.1016/j.immuni.2021.12.013 External LinkAlso available on ScienceDirect External Link

Copyright: © 2021 Elsevier Inc.

User License: Elsevier user license | Elsevier's open access license policy

• Download PDF
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Summary

Obesity leads to chronic, systemic inflammation and can lead to insulin resistance (IR), β-cell dysfunction, and ultimately type 2 diabetes (T2D). This chronic inflammatory state contributes to long-term complications of diabetes, including non-alcoholic fatty liver disease (NAFLD), retinopathy, cardiovascular disease, and nephropathy, and may underlie the association of type 2 diabetes with other conditions such as Alzheimer’s disease, polycystic ovarian syndrome, gout, and rheumatoid arthritis. Here, we review the current understanding of the mechanisms underlying inflammation in obesity, T2D, and related disorders. We discuss how chronic tissue inflammation results in IR, impaired insulin secretion, glucose intolerance, and T2D and review the effect of inflammation on diabetic complications and on the relationship between T2D and other pathologies. In this context, we discuss current therapeutic options for the treatment of metabolic disease, advances in the clinic and the potential of immune-modulatory approaches.

Introduction

Obesity along with genetic predisposition, as observed in some populations such as Pima Indians, Latinos, North Africans, and Indians, is the most common cause of insulin resistance (IR). Additionally, IR is a key underlying etiology of type 2 diabetes (T2D). Before the onset of this disease, compensatory hyperinsulinemia is observed, which declines as a result of progressive β-cell dysfunction leading to deterioration of glucose homeostasis and eventually diabetes. In this way, the development of T2D is usually because of a combination of both IR and β-cell dysfunction. Over time, T2D leads to a number of long-term complications, including microvascular disease (retinopathy, nephropathy, and neuropathy), macrovascular disease (stroke, myocardial infarction, and peripheral arterial disease), heart failure, and nonalcoholic fatty liver disease (NAFLD). Furthermore, it may contribute to the progression of Alzheimer`s disease, polycystic ovarian syndrome, gout, and rheumatoid arthritis among other diseases promoted by inflammation.

Although there is a significant palette of anti-diabetic drugs available, most subjects do not achieve optimal levels of glucose control. Additionally, glucose control per se is not always sufficient to prevent the long-term complications of diabetes. Indeed, only three classes of glucose-lowering drugs have shown prevention of complications; metformin, glucagon-like 1 (GLP-1) analogs, and sodium-dependent glucose transporter 2 (SGLT-2) inhibitors. While GLP-1 analogs mitigate cardiovascular diseases and SGLT-2 inhibitors prevent heart failure and nephropathy, none of these drugs has demonstrated efficiency against other complications such as retinopathy, NAFLD, and diabetes-associated inflammatory diseases. Most importantly, except for metformin, progression of diabetes is not delayed by the available drugs. This highlights the importance of improved understanding of the underlying mechanisms of IR and β-cell dysfunction.

In this regard, chronic tissue inflammation has emerged as a key feature of obesity and T2D and is observed in insulin target tissues, such as adipose tissue, liver, muscle, and pancreatic islets. These observations have led to the term “immunometabolism,” which incorporates the underlying interplay between immunologic processes and metabolic defects. The recruitment, accumulation, and activation of pro-inflammatory macrophages in metabolic tissues is the ultimate driver of this chronic low-grade inflammation. While the macrophage is the major effector cell type, other types of immune cells participate in these inflammatory processes.

This review examines recent advances and underlying mechanisms of inflammation in obesity, T2D, and related disorders. We describe how this chronic tissue inflammation includes multi-organ cross-talk causing IR, impaired insulin secretion, glucose intolerance, and T2D. We also review the effect of inflammation on diabetic complications and how obesity and diabetes are connected to other inflammatory diseases. In particular, we span the spectrum from basic research to translational findings, including a review of current and potential immune-modulatory therapeutic options for metabolic disease aimed to treat disease progression and complications.

Inflammation in obesity and the pathogenesis of T2D

The potential contribution of chronic tissue inflammation to metabolic disease has been suggested for many years (Cook et al., 2000; Dandona et al., 2004; Donath and Shoelson, 2011). There are a number of key studies published over the past two decades that have provided foundational insights to the immunometabolism field. For example, Feingold et al. discovered that tumor necrosis factor (TNF-α) leads to glucose intolerance in rodents (Feingold et al., 1989). The observation that adipose tissue in obesity expresses high levels of TNF-α and that its neutralization improves insulin sensitivity and glucose intolerance were key findings in establishing the association between immune cells and metabolic dysfunction (Hotamisligil et al., 1993; Yuan et al., 2001). Pro-inflammatory cytokines, like TNF-α, can activate various intracellular signaling molecules, such as JNK and IKKβ, that are critical components of the inflammatory signaling system, leading to impaired insulin action (Arkan et al., 2005; Hirosumi et al., 2002). Activation of IKKβ leads to nuclear translocation of nuclear factor κB (NF-κB), which drives increased expression‎ of inflammatory mediators, including chemokines and cytokines (Chen et al., 2012b; Shoelson et al., 2006; Shoelson et al., 2003).

An additional key component of inflammatory activation is a multimeric protein complex termed “inflammasome,” which is activated by cell nutrients, such as glucose and free fatty acids, inducing interleukin-1 beta (IL-1β) production (Böni-Schnetzler et al., 2009; Maedler et al., 2002). Intracellularly, the inflammasome regulates the activation of caspase-1. Activated caspase-1 cleaves precursor cytokines, such as pro-IL-1β, leading to increased tissue levels of active IL-1β (Thornberry et al., 1992). Accordingly, binding of lipopolysaccharide (LPS) or saturated fatty acids (SFAs) to Toll-Like receptor-4 (TLR-4) and glucose via oxidative stress stimulate NLRP3, mediating activation of caspase-1 (Okla et al., 2018; Paik et al., 2021; Zhou et al., 2010). Although some studies have indicated that IL-1β can blunt insulin action, the most prominent effects of IL-1β are in the islets, where it modulates β-cell mass and function.

Another key finding was the simultaneous discovery by Weisberg et al. and Xu et al. that macrophages accumulate in both mouse and human adipose tissue during the course of obesity (Weisberg et al., 2003; Xu et al., 2003). The pro-inflammatory polarization state of these cells leads to the release of a number of inflammatory cytokines and other factors that contribute to decreased insulin signaling. Unlike the classification of in vitro polarized pro-inflammatory (M1) and anti-inflammatory, alternatively activated (M2) macrophages (Martinez and Gordon, 2014; Orecchioni et al., 2019), these terms do not strictly apply to the in vivo situation. In recognition of phenotypic, transcriptomic, and polarization differences between in vitro and in vivo macrophages, the terms “M1-like” and “M2-like” will be used as general descriptors throughout the rest of this review, when appropriate.

Adipose tissue inflammation

Following the initial reports of increased numbers of adipose tissue macrophages (ATMs) in both mouse and human obesity (Weisberg et al., 2003; Xu et al., 2003), additional studies showed that these cells contribute to the insulin-resistant state (Bigornia et al., 2012; Lackey and Olefsky, 2016; Xu et al., 2003). Indeed, prevention of ATM accumulation or pro-inflammatory macrophage signaling protects obese mice from glucose intolerance and IR (Desai et al., 2017; Dror et al., 2017; Lee and Olefsky, 2021; Takikawa et al., 2016). In addition to macrophages, there are other immune cell types that participate in adipose tissue inflammation during obesity. However, while certain T and B cell subsets play important regulatory roles, it is generally thought that macrophages are the major effector cells leading to decreased insulin signaling (Chawla et al., 2011; McLaughlin et al., 2017). In lean mice and humans, ATMs make up 10% to 15% of the stromovascular cells and largely display an M2-like polarization state (Weisberg et al., 2003; Xu et al., 2003) (Figure 1, lean AT). In obesity, the number of ATMs increases and can comprise up to 40% of all adipose tissue cells (Lumeng et al., 2007). Most of these obesity-induced ATMs are pro-inflammatory (M1-like) (Lumeng et al., 2007; Nguyen et al., 2007) (Figure 1, obese AT) and similar findings have been shown in human ATMs (Fuchs et al., 2021; Wentworth et al., 2010). The phenotype of these macrophages is not fixed and can be modified by SFAs acting through a TLR-4 mechanism that stimulates the M1-like polarization state (Shi et al., 2006). Omega-3 fatty acids (fish oils) inhibit this inflammatory state by binding to the cell surface receptor GPR120 (Oh et al., 2010).

Figure 1 Chronic tissue inflammation in obesity

Show full captionFigure viewer

An interesting study by Wernstedt Asterholm et al. (2014) points out the marked functional differences between acute and chronic inflammation specifically in adipocytes. This paper used genetic manipulations to express specific anti-inflammatory proteins within adipocytes and demonstrated that this affected only adipocytes and not ATMs. They went on to show that inhibiting acute inflammation within adipocytes during high-fat diet (HFD) led to a marked decrease in adipogenesis and angiogenesis. This principle finding supports their main conclusion, which is that “acute” inflammation promotes adipogenesis under the challenge of HFD. Interestingly, this also led to an increase in ATMs and crown-like structures with a deterioration of glucose tolerance and insulin sensitivity. The changes in glucose homeostasis might be explained by the fact that these mice developed a “leaky” gut. The authors found that the anti-inflammatory effects in adipocytes were most marked in mesenteric WAT, which, in turn, could affect the gastro-intestinal epithelium leading to increased permeability. They suggest that LPS and perhaps other factors that leak into the circulation could be responsible for worsening glucose tolerance (Wernstedt Asterholm et al., 2014).

ATMs are the most abundant immune cells in adipose tissue and tend to accumulate in crown-like structures, which are observed surrounding enlarged and dying adipocytes (Murano et al., 2008; Strissel et al., 2007). In obesity, ATMs span a wide polarization spectrum (Hill et al., 2018; Jaitin et al., 2019). Single-cell RNA sequencing studies have shown that ATMs classified as “M1” or “M2” by using specific surface markers are actually quite heterogeneous and comprise several different discrete clusters of cells within each category in mouse and human adipose tissue (Cox et al., 2021; Hildreth et al., 2021; Hill et al., 2018; Jaitin et al., 2019; Kratz et al., 2014; Shaul et al., 2010; Xu et al., 2013). ATMs within obese adipose tissue are generally pro-inflammatory and CD11c is a typical marker used to identify these M1-like ATMs (Wentworth et al., 2010). However, Hill et al. showed that CD9 expression‎ can be used to further define M1-like ATMs (Hill et al., 2018). These CD9+ ATMs exhibit unique locational and morphologic features, expressing high levels of pro-inflammatory cytokines, and contain prominent intracellular lipid droplets. CD9+ ATMs are found within crown-like structures and are CD11c positive or negative. There is also a cluster of ATMs in obesity that display the monocyte marker Ly6C (Hill et al., 2018). These cells are derived from recently recruited monocytes and are undergoing macrophage differentiation (Yang et al., 2014; Yona et al., 2013). Over a period of several days, these cells lose Ly6C expression‎ and become mature macrophages found within and outside of the crown-like structures (Hill et al., 2018; Oh et al., 2012). The great majority of these CD9+ or Ly6C+ cells are bone marrow derived and form distinct clusters depending on whether they express the lipid receptor TREM2. Double-positive TREM2+CD9+ ATMs actively participate in lipid metabolism and have been termed lipid-associated macrophages (LAMs) (Jaitin et al., 2019). Interestingly, accumulation of LAMs is functionally dependent on TREM2 because genetic deletion of TREM2 in obese mice promotes obesity, dyslipidemia, and glucose intolerance (Jaitin et al., 2019). These latter findings suggest that TREM2+ LAMs can actually mitigate obesity and the associated metabolic abnormalities. Thus, depending on the relative expression‎ of CD11c, CD9, and TREM2, ATMs show distinct phenotypes and further studies of the functional properties of these subsets will be of great interest.

There are several other immune cell types that participate in the overall state of adipose tissue inflammation in obesity. For example, CD8+ T cells are more abundant and can promote monocyte chemotaxis followed by differentiation into ATMs (Nishimura et al., 2009). Pro-inflammatory CD3+CD4+ T helper (Th1) cells are increased in obese adipose tissue and also contribute to the pro-inflammatory state via interferon (IFN)-γ production (Winer et al., 2009). Of interest, several studies have examined the content and function of Tregs. These CD3+CD4+FOXP3+ Tregs are the most abundant (40%–80%) adipose CD4+ T cell subset in normal adult mice (Bapat et al., 2015; Feuerer et al., 2009; Li et al., 2020). They can modulate the activity of other T cells and, importantly, inhibit monocyte immigration and drive them toward an anti-inflammatory polarization state (Romano et al., 2018). In obese fat, the number of Tregs decreases, and this promotes increased ATM-mediated chronic inflammation (Feuerer et al., 2009). Interestingly, the transcriptomic profile of adipose Tregs is quite different from Tregs in other tissues. An important difference is the expression‎ of peroxisome proliferator activated receptor γ (PPAR-γ), by adipose tissue Tregs (Cipolletta et al., 2012). Thiazolidinediones (TZDs) are insulin sensitizing PPAR-γ agonists and exert substantial anti-inflammatory effects via direct action on adipose Tregs. Interestingly, age appears to have independent effects on AT Tregs. Bapat et al. have found that Tregs in mouse visceral fat increase as a function of age, independent of obesity, and that Treg depletion in aged mice promotes insulin sensitivity (Bapat et al., 2015). Other immune cell types, such as mast cells, innate lymphoid cell type 1 (ILC1) cells, neutrophils, eosinophils, natural killer (NK) cells, invariant natural killer T (iNKT) cells, and B lymphocytes also participate in the overall inflammatory tone of adipose tissue largely by modifying ATM recruitment or phenotypes. Their role in the inflammatory process has been recently published and reviewed in detail elsewhere (Boulenouar et al., 2017; Kane and Lynch, 2019; Lee and Dixit, 2020; Lee et al., 2016; Lee and Olefsky, 2021; O’Sullivan et al., 2016; Wang et al., 2019; Wensveen et al., 2015; Zatterale et al., 2020)

Mechanism of adipose tissue macrophage accumulation

Increased monocyte chemotaxis into adipose tissue is the major mechanism leading to ATM accumulation (Kamei et al., 2006; Kanda et al., 2006; Nagareddy et al., 2014; Weisberg et al., 2006). As one example, injection of fluorescently labeled monocytes into obese mice showed that enhanced monocyte uptake in obesity is induced via chemokines such as CCL2/MCP-1 and leukotriene B4 and galectin-3 (Amano et al., 2014; Li et al., 2015, 2016). Another study shows that IL-6-mediated trans-signaling through the soluble form of the IL-6 receptor (sIL-6R) interacting with glycoprotein 130 can also produce chemotactic effects for ATM accumulation (Kraakman et al., 2015). There are additional chemokines beyond these that can regulate macrophage recruitment and insulin resistance (Han and Levings, 2013; Xu et al., 2015). The majority of the literature show that genetic deletion or pharmacological inhibition of chemokine receptors reduces monocyte influx into obese adipose tissue (Kanda et al., 2006; Li et al., 2015; Lumeng et al., 2007; Spite et al., 2011; Weisberg et al., 2006). After arriving in the adipose tissue, monocytes can efflux into surrounding lymph nodes or back into the circulation (Auffray et al., 2007; Palframan et al., 2001). In obesity, there is increased adipose tissue expression‎ of Semaphorin 3E and Netrin-1, which inhibit this emigration process, causing macrophage retention and further increasing ATM content (Nakano et al., 2013; Ramkhelawon et al., 2014; Shimizu et al., 2013; Wanschel et al., 2013). Resident ATMs, as well as blood monocyte-derived ATMs, also undergo increased proliferation in the context of obesity, further contributing to the increase in ATMs (Amano et al., 2014; Haase et al., 2014; Zamarron et al., 2017; Zheng et al., 2016). It will be of interest to assess the functional properties and transcriptomic profiles of proliferated versus non-proliferated macrophages in comparison to ATMs derived from chemotaxis.

Initiating mechanisms for chronic tissue inflammation

Several initiating mechanisms have been proposed to generate chronic tissue inflammation, and these mechanisms are not mutually exclusive and could work in a complementary manner. Regardless of the specific mechanism, the ultimate inflammation triggers would have to initiate either proliferation of ATMs or monocyte chemotaxis into the tissue. Increased influx of nutrients is an obvious feature of obesity and can participate in the production and release of reactive oxygen species (ROS), resulting in oxidative stress (Houstis et al., 2006). Nutrient overload can also promote endoplasmatic reticulum (ER) stress and an unfolded protein response, activating intracellular inflammatory pathways (Hotamisligil and Erbay, 2008). Glucose and free fatty acids activate inflammasomes and Toll-like receptors (TLRs), specifically TLR-4, with stimulation of the downstream inflammatory pathways (Böni-Schnetzler et al., 2009; Maedler et al., 2002; Shi et al., 2006; Zhou et al., 2010).

Hypoxia is another activator of inflammation, and particularly in obesity, it is a very early event in the initiation of adipose tissue inflammation (Halberg et al., 2009; Hosogai et al., 2007; Lee et al., 2014). It has been well described in both mouse and human adipose tissue that oxygen (O2) tension decreases in obesity (Halberg et al., 2009; Hosogai et al., 2007; Lawler et al., 2016; Pasarica et al., 2009, 2010; Seo et al., 2019; Smith et al., 2019). As adipocytes and adipose tissue expand, angiogenesis can lag behind with impaired capillary density leading to less interstitial perfusion of O2. Importantly, interstitial and intracellular O2 tension may not always go hand in hand, as the intracellular O2 levels represent the balance between O2 supply and demand. This is important because recent studies indicate that obesity leads to an increase in intracellular adipocyte O2 consumption, accounting for about 40% of the decrease in the interstitial O2 tension (Lee et al., 2014). The enhanced adipocyte O2 consumption is secondary to activated uncoupled oxidative phosphorylation, which, in turn, is due to increased free fatty acid stimulation of a mitochondrial protein termed Adenine nucleotide translocase (ANT2) (Lee et al., 2014; Shabalina et al., 2006). ANT2 is highly activated by the abundant SFAs in obese adipocytes, leading to proton leakage from the intermembrane space and to uncoupled mitochondrial respiration and greater O2 consumption. Therefore, stimulation of ANT2 serves to uncouple O2 utilization from ATP production in the mitochondria, lowering intracellular O2 levels. The intra-adipocyte hypoxic state triggers induction of hypoxia inducible factor 1 alpha (HIF-1α), primarily by inhibiting degradation of HIF-1α (Schödel and Ratcliffe, 2019; Semenza, 2019; Seo et al., 2019). HIF-1α within adipocytes can then initiate an inflammatory response by inducing transcription of chemokines (Imtiyaz and Simon, 2010). These chemokines initiate monocyte recruitment and differentiation into pro-inflammatory M1-like ATMs. Supporting this mechanism, deletion of adipocyte ANT2 prevents increased oxygen consumption, blocks HIF-1α induction and protects from adipose tissue inflammation, glucose intolerance, and IR (Seo et al., 2019). Likewise, adipocyte-specific deletion of HIF-1α prevents the development of adipose tissue inflammation in obese mice mitigating IR and glucose intolerance (Jiang et al., 2011; Sun et al., 2013).

How does inflammation cause IR?

The underlying mechanisms by which ATMs cause IR have been widely studied. The general idea is that ATMs produce factors that work in a paracrine or systemic manner and interrupt insulin signaling in target cells. Because these M1-like macrophages produce a variety of chemokines and cytokines, significant attention has been focused on these molecules as causes of decreased insulin sensitivity. Chemokines stimulate chemotaxis of circulating monocytes and other immune cells by providing a concentration gradient between the blood and the interstitial space. While some of these chemokines can leak into the circulation in obesity, this would not generate a necessary concentration gradient to stimulate monocyte chemotaxis. Furthermore, the circulating concentrations of chemokines are much lower than the biologically relevant levels in the interstitial space of obese adipose tissue, so circulating chemokines are a marker of tissue inflammation but are quite unlikely to cause IR.

Therefore, the effects of cytokines to mediate tissue inflammation have been intensively studied. TNF-α, the most well-examined of these cytokines, reduces insulin sensitivity (Hotamisligil et al., 1994). TNF-α promotes inhibitory phosphorylation of insulin receptor substrate (IRS) proteins, enhances ceramide synthesis and adipocyte lipolysis, and inhibits PPARγ expression‎ (Guilherme et al., 2008; Stephens et al., 1997). Normal levels of PPARγ are necessary to maintain insulin sensitivity and the effects of ceramides to inhibit AKT phosphorylation and insulin action are well-known. However, the circulating levels of TNF-α in obese states are below the concentrations that impair insulin signaling (Amar et al., 2007; McGillicuddy et al., 2011; Stephens et al., 1997). Because of this, it is unlikely that TNF-α “leakage” out of obese adipose tissue into the circulation is responsible for the decreased muscle and liver insulin sensitivity. However, locally produced high levels of TNF-α in the liver and muscle tissues could cause IR (Lang et al., 1992). Another cytokine released by M1-like ATMs is IL-6. Unlike other cytokines, blood IL-6 circulates at biologically active levels. However, the effects of IL-6 on insulin signaling are unclear because some studies have reported IL-6-mediated IR while others suggest that IL-6 has insulin-like actions with improved insulin sensitivity (Carey et al., 2006; Franckhauser et al., 2008). Another prominently released cytokine is IL-1β. It is well known that islet macrophage-derived IL-1β impairs insulin secretion contributing to the development of glucose intolerance and T2D (Eguchi et al., 2012; Maedler et al., 2002). Further, IL-1β was also shown to promote IR (Stienstra et al., 2010; Vandanmagsar et al., 2011; Wen et al., 2011; Zhou et al., 2010). In adipose tissue, IL-1β impairs adipocyte insulin signaling (Jager et al., 2007; Lagathu et al., 2006; Stienstra et al., 2010). In humans, IL-1β release from visceral ATMs is enhanced with glycemic deterioration and decreases after gastric bypass surgery (Dalmas et al., 2014). However, clinical evidence for a role of IL-1β in IR remains to be clarified. In a clinical study of IL-1Ra in patients with T2D, no changes in insulin sensitivity could be found, although HbA1c decreased (Larsen et al., 2007). However, insulin sensitivity was only assessed in a small patient subgroup in this study. In contrast, in an elegant study in patients with T1D who also have IR caused by obesity, treatment with IL-1Ra improved insulin sensitivity and glycemic control (van Asseldonk et al., 2015). Finally, in a recent study of patients with rheumatoid arthritis and diabetes, IL-1Ra improved insulin sensitivity (Ruscitti et al., 2019). Therefore, it is likely that IL-1β plays also a role in IR.

Galectin-3, a member of the lectin family, is a factor which is exclusively macrophage-derived and achieves biologically effective concentrations in the circulation of obese mice and humans (Li et al., 2016). Studies have shown that galectin-3 can directly cause macrophage chemotaxis and impaired insulin signaling, while myeloid-specific deletion of galectin-3 inhibits HFD-induced glucose intolerance and IR (Li et al., 2016).

ATMs release exosomes, which work locally in a paracrine fashion and can also enter the circulation in sufficient concentrations to have distal effects on insulin sensitive tissues. Exosomes are a component of secreted extracellular vesicles (EVs) with a diameter of ∼150 nm and can be produced by almost all cell types (Pegtel and Gould, 2019). Exosomal cargo consists of a large number of proteins, lipids, micro RNAs (miRNAs), mRNAs, and noncoding RNA species. EV/exosome preparations obtained from obese human adipose tissue express increased levels of several miRNAs that negatively interact with the insulin signaling pathway (Kita et al., 2019). Consistent with this, exosomes prepared from obese adipose tissue cause IR, and miR-141-3p has been implicated in this process (Dang et al., 2019; Kranendonk et al., 2014). However, because all of the cell types within adipose tissue release exosomes, studies using exosome preparations derived from whole adipose tissue do not reveal which cell type produces the biologically relevant exosomal particles. Recent studies, specifically focused on macrophages, have shown that ATM-derived exosomes isolated from obese mice directly cause decreased insulin signaling in adipocytes, myocytes, or primary hepatocytes in vitro (Ying et al., 2017) (Figure 2, adipose tissue). These “obese” exosomes contain high levels of miR-155, which represses PPAR-γ expression‎, thereby contributing to IR. Lean mice treated intravenously with these “obese” ATM exosomes develop glucose intolerance, hyperinsulinemia, and IR comparable with obese mice, without changes in body weight. Interestingly, lean chow-fed mice produce ATM exosomes that cause the opposite effects. In other words, “lean” ATM exosomes directly enhance insulin signaling in vitro and, when given to obese mice, markedly improve insulin sensitivity and glucose tolerance. These effects were entirely attributed to the miRNA cargo of the exosomes, and miRNA-690 was identified as the predominant miRNA within M2-like macrophage exosomes causing these beneficial effects (Ying et al., 2021). Indeed, in vivo treatment with a miR-690 mimic recapitulates the insulin sensitizing effects of “lean” ATM exosomes, whereas treatment with an antagomir blocks miR-690-mediated effects. In this way, M2-like macrophage exosomes and miR-690 mimics have emerged as promising insulin sensitizers.

Figure 2 Exosomes as regulators of intercellular and interorgan crosstalk in metabolism

Show full captionFigure viewer

Hepatic inflammation

Obesity characteristically leads to hepatic steatosis and liver inflammation, commonly termed NAFLD (Polyzos et al., 2019). This is a very common condition, and some estimates suggest there are one billion NAFLD patients worldwide (Younossi et al., 2019). Hepatocyte lipid accumulation is multi-factorial. In obesity, adipocytes become insulin resistant, leading to impaired suppression of lipolysis with increased free fatty acid release. This enhanced influx of adipocyte-derived free fatty acids, as well as dietary fat entering the liver via gastrointestinal-derived chylomicrons, are important pro-steatotic mechanisms (Donnelly et al., 2005; Meex and Watt, 2017). In addition, within the hepatocyte, de novo lipogenesis pathways are increased in obesity and this is coupled with decreased fatty acid oxidation (Song et al., 2018). All of these factors drive steatosis and NAFLD.

There are two major categories of liver macrophages: Kupfer cells (KCs) and recruited hepatic macrophages (RHMs) (Morinaga et al., 2015; Seidman et al., 2020). KCs are the resident liver macrophages and are derived from the embryonic yolk sac (Naito et al., 1989; Perdiguero et al., 2015). A second category of hepatic macrophages are bone marrow-derived and arise from circulating monocytes that enter the liver in response to chemokines, primarily CCL2/MCP-1, which can be secreted from steatotic hepatocytes and KCs (Morinaga et al., 2015; Obstfeld et al., 2010). These monocytes differentiate into M1-like pro-inflammatory RHMs (Morinaga et al., 2015; Seidman et al., 2020). In obesity, KC numbers are relatively unchanged, but there is a large increase in RHMs (Morinaga et al., 2015; Obstfeld et al., 2010) (Figure 1, NAFLD). RHMs also produce chemokines, setting up a feed-forward loop, in which newly arrived RHMs produce chemokines, which attract additional monocytes into the liver (Morinaga et al., 2015). KCs perform scavenger and phagocytic functions, such as removing senescent red blood cells to ensure iron conservation, filtering gut-derived factors like LPS, and participating in tissue homeostasis and repair (Nguyen-Lefebvre and Horuzsko, 2015). These cells comprise up to 10% of all liver cells and their transcriptomic profile does not fall into typical M1 or M2 categories (Bouwens et al., 1986; Morinaga et al., 2015). In obesity, RHMs are predominantly pro-inflammatory and are responsible for the major component of obesity-mediated hepatic inflammation (Morinaga et al., 2015). KCs and RHMs participate in intrahepatic signaling because cytokines derived from hepatic macrophages can stimulate hepatocyte de novo lipogenesis and ceramide biosynthesis (Chavez and Summers, 2012; Holland et al., 2011; Sanders and Griffin, 2016). Macrophages are not the only immune cells participating in hepatic inflammation, as obesity also leads to an increase in liver neutrophils and CD8+ T cells (Ghazarian et al., 2017; Talukdar et al., 2012). Neutrophil elastase can be released and taken up into hepatocytes (Talukdar et al., 2012), promoting intracellular degradation of insulin receptor substrate 2 (IRS-2) and thus potentiating hepatocyte IR.

Some estimates indicate that up to 37% of obese NAFLD subjects progress to nonalcoholic steatohepatitis (NASH) (Abrams et al., 2004; Gholam et al., 2007; Machado et al., 2006; Ong et al., 2005), which potentially leads to cirrhosis with severely impaired liver function (Loomba et al., 2021). Further, NASH is increasingly recognized as a common complication of obesity (Akshintala et al., 2000). Given the commonality of obesity to NASH and T2D, NASH can also be thought of as a complication of T2D and studies have indicated that 20%–40% of subjects with diabetes also have NASH (Younossi et al., 2019). Inflammation and ER stress are important drivers of NAFLD and the progression to NASH (Puri et al., 2008; Rutkowski et al., 2008). Fu et al. reported that hepatic ER stress is caused by significant changes in ER fatty acid and lipid composition, resulting in inhibition of calcium ATPase (SERCA) activity (Fu et al., 2011). In addition, the progression of NASH is partially dependent on pro-inflammatory hepatic macrophages producing TNF (Nakagawa et al., 2014; Todoric et al., 2020).

Recent studies using a variety of monocyte tracking techniques coupled with RNA sequencing have revealed new insights into the phenotypes of KCs and RHMs in NASH (Seidman et al., 2020). As NASH develops, the original resident yolk sac-derived KCs undergo gradual apoptosis and are replaced by monocyte-derived RHMs, which accumulate in NASH livers. These RHMs can enter the sinusoidal niche where they respond to local environmental clues and develop a KC-like transcriptomic profile. These “KC-like” cells in NASH livers are similar, but not identical, to original healthy KCs. Thus, apoptosis and defects in normal KC function represent an important feature of NASH. The exact pathophysiologic consequences of this form of KC dysfunction are currently unknown.

A key element of the NASH liver phenotype is collagen accumulation (Loomba et al., 2018, 2021). Thus, it is important to understand the various determinants which convert normal stellate cells from the quiescent or inactive state into activated cells that proliferate and elaborate excess collagen that deposits in the extracellular matrix. Transforming growth factor beta (TGF-β) and connective tissue growth factor (CTGF) can be released by hepatocytes or hepatic macrophages and stimulate the stellate cell fibrogenic program (Gressner and Gressner, 2008; Hellerbrand et al., 1999). Indeed, a subpopulation of CD9+Trem2+ bone marrow-derived hepatic macrophages are found in close proximity to collagen fibrils and have been termed scar-associated macrophages (SAMs) (Seidman et al., 2020). These cells may play a role in promoting collagen production by stellate cells (Fallowfield et al., 2007). Hepatocytes under early HFD stress (4 weeks HFD) release exosomes that promotes insulin sensitizing effects via miR-3075 (Ji et al., 2021) (Figure 2, liver). In addition, several reports have shown that exosomes released by steatotic hepatocytes stimulate stellate cell fibrogenesis through their exosomal cargo, including miR-128-3p, which suppress PPAR-γ signaling (Povero et al., 2015) and miR-1 via activation of the NF-κB pathway (Jiang et al., 2020). Apart from these intrahepatic signals, it is also likely that extrahepatic signals impinge on the liver to promote the NASH phenotype. For example, obesity is a frequent concomitant in subjects with NASH, and it is well known that obesity leads to a “leaky” gut (see intestinal inflammation), allowing LPS and other bacterial byproducts, or perhaps bacteria themselves to enter the bloodstream and target the liver. Normally, KCs filter out gut-derived products, but it is possible that the dysfunctional KCs observed in NASH livers are deficient in this function (Bennett et al., 2021). In this case, these gut-derived products may provide signals to other cells in the liver to promote NASH. Furthermore, adipose tissue is well-known to secrete adipocytokines such as leptin, adiponectin, and galectin-3, all of which influence steatosis and hepatic inflammation.

Muscle inflammation

The skeletal muscle (SM) is a key organ in the development of metabolic disease because it is responsible for the majority (70%–80%) of insulin-stimulated glucose disposal (Ferrannini et al., 1988). Therefore, IR in SM is an important determinant of metabolic dysregulation in T2D (DeFronzo and Tripathy, 2009). Proposed mechanisms for SM IR include mitochondrial defects, ectopic fat accumulation, and intrinsic dysregulation of insulin action, and these have been reviewed elsewhere (Di Meo et al., 2017; Samuel and Shulman, 2016). In this section, we will focus on the role of muscle inflammation as a possible mediator of IR. This subject is less well-studied compared with fat tissue and liver, but several reports suggest a role of chronic SM inflammation in IR (Boon et al., 2015; Varma et al., 2009). These reports demonstrated increased inflammatory markers, activated pro-inflammatory pathways, and immune cell accumulation in SM in obesity and diabetes (Fink et al., 2014; Hong et al., 2009; Patsouris et al., 2014). Similar to visceral fat, muscle macrophages are increased in obesity in the intermyocellular/intermuscular adipose tissue (IMAT) between the muscle fibers as well as in perimuscular adipose tissue (PMAT) (Khan et al., 2015). These IMAT and PMAT macrophages exhibit a CD11c+ pro-inflammatory, M1-like phenotype (Fink et al., 2014; Khan et al., 2015; Patsouris et al., 2014). SM Th1, CD4+, and CD8+ T cells are also more abundant while Tregs are decreased, all of which characterize chronic SM inflammation and point to similar mechanisms as observed for visceral fat (Khan et al., 2015). These immune cells contribute to higher levels of pro-inflammatory cytokines, such as TNF-α, IL-6, IL-1β, and CCL2/MCP-1 in SM. SM can also be a secretory organ. Exercise and muscle contraction leads to increased secretion of “myokines” such as IL-6 with potential beneficial effects on lipid metabolism and inflammation (Eckardt et al., 2014; Pedersen and Febbraio, 2012). In the presence of the high SFA levels in obesity, myocyte TLR-2 and TLR-4 are stimulated (Wei et al., 2008). This causes them to secrete increased levels of chemokines, promoting monocyte migration and SM inflammation. In support of this, deletion of TLR-4 from cultured myocytes protects them from lipid-induced IR both in vitro and in vivo (Radin et al., 2008). IL-6 is probably the most well-studied myocyte cytokine and is a mediator of macrophage infiltration and polarization, as well as muscle repair (Muñoz-Cánoves et al., 2013). Several studies show that IL-6 has pro-inflammatory effects and can induce IR in obesity (Rehman et al., 2017; Rotter et al., 2003). However, acute IL-6 treatment increases both basal and insulin-stimulated SM glucose uptake and improves systemic glucose homeostasis (Carey et al., 2006; Pedersen and Febbraio, 2012). In vitro, TNF-α, or conditioned media from Th1 cells or pro-inflammatory macrophages induce IR in myocyte cultures (de Alvaro et al., 2004). TNF-α infusion in healthy humans also results in SM IR (Plomgaard et al., 2005). Interestingly, the anti-inflammatory cytokine IL-10 protects myocytes from the development of IR. Indeed, muscle-specific IL-10 expression‎ attenuates muscle inflammation and IR in vivo, while genetic deletion of IL-10 promotes these metabolic abnormalities (Dagdeviren et al., 2016; Hong et al., 2009).

Intestinal inflammation

Obesity leads to several pro-inflammatory alterations in the gastrointestinal tract. The gut contains an extensive microbiome repertoire, which is altered in human and mouse obesity, a finding termed dysbiosis. Different analyses of microbiota composition have been extensively reported, and the reader is referred to several excellent reviews on this subject (Choi et al., 2020; Fan and Pedersen, 2021; Khan et al., 2021).

The first mechanical barrier, preventing entry to the systemic circulation, is the intestinal epithelia, a single layer of cells, joined by tight junctions, as well as a mucosal layer, produced by goblet cells. In obesity, leakage of microbial factors or bacterial products, such as LPS occurs due to higher intestinal permeability (Amar et al., 2011; Cani et al., 2007, 2008; de La Serre et al., 2010). These microbial products and metabolites, such as short-chain FAs, can activate G-protein-coupled receptors (GPCRs), or inflammasomes, triggering inflammation and impaired glucose metabolism (Kim et al., 2012; Kimura et al., 2013; Tolhurst et al., 2012). For instance, bile acids improve glucose and insulin sensitivity in obese and diabetic mice through activation of the nuclear farnesoid X receptor (FXR) and Takeda G-protein receptor (TGR5) (Swann et al., 2011; Thomas et al., 2009).

As the gut is the first organ exposed to food and microbial antigens, it contains a robust and complex immune system, located in the subepithelial lamina propria. Several changes in adaptive and intestinal innate immunity have been reported in obesity. IFN-γ and IL-1β-producing immune cells increase barrier permeability in obesity, while IL-22-secreting innate lymphoid cells (ILCs) and Tregs promote production of mucin and anti-microbial immunoglobulin A (IgA) (Al-Sadi and Ma, 2007; Cong et al., 2009; Luck et al., 2015; Wang et al., 2014). Furthermore, depletion of intestinal eosinophils by HFD points toward a protective function of these cells on the gut barrier (Johnson et al., 2015). In addition, HFD induces a pro-inflammatory shift in T cells, characterized by higher INF-γ+ Th1 and CD8+ T cells, along with reduced IL-10+ Tregs and IL17+ Th17 cells (Luck et al., 2015). Obese subjects also display an accumulation of intraepithelial CD8αβ+ T cells (Monteiro-Sepulveda et al., 2015). A lower level of IgA+ antibody-secreting cells and B cells in HFD-fed mice causes a reduction in colonic secretory IgA+ and might promote IR (Luck et al., 2019). Obesity is also associated with reduced IL-22- and IL-17-secreting intestinal group 3 ILCs (ILC3s), which are important for barrier integrity (Luck et al., 2015). In contrast, ILC2s seem to have a detrimental effect, because mice deficient in ILC2 are resistant to diet-induced obesity (Sasaki et al., 2019). In humans, intestinal ILC1 and ILC2 numbers negatively correlate with body-mass index (Yudanin et al., 2019). Recent reviews (Febbraio and Karin, 2021; Herman and Birnbaum, 2021) suggest that fructose is a metabolic toxin causing impaired intestinal barrier integrity and endotoxemia. Fructose was thought to be exclusively metabolized in the liver, but it is now known that it can also be metabolized at its site of absorption in the small intestine (Goncalves et al., 2019; Jang et al., 2020; Taylor et al., 2021). Fructose enhances nutrient absorption (Taylor et al., 2021) and induces dyslipidemia and hepatic ER stress, contributing to steatosis and NASH via the gut-liver axis (see section on hepatic inflammation) (Jang et al., 2020; Todoric et al., 2020; Zhao et al., 2020).

HFD leads to a lack of CD103+CD11b+conventional dendritic cells (DCs) (Stagg, 2018) and a reduction in intestinal innate-like T cells, called mucosal-associated invariant T (MAIT) cells (Toubal et al., 2020). These MAIT cells regulate intestinal barrier function, and the activated CD44+ phenotype of these cells in obesity promotes both intestinal and adipose tissue inflammation (Toubal et al., 2020; Varelias et al., 2018).

As in other tissues, intestinal macrophages are heterogeneous and are sub-grouped based on phenotypic markers, and functional differences. After circulating blood monocytes enter the gut lumen, they initially display high expression‎ levels of Ly6C and CCR2, but gradually differentiate toward resident mature macrophages, losing Ly6C/CCR2 and gaining CD64/MHII expression‎ (Bain et al., 2013, 2014; Bain and Mowat, 2014; Tamoutounour et al., 2013; Tamoutounour et al., 2012). These “monocyte-like” transition cells are characterized by upregulation of TNF-α, IL-1β, or IL-6 and hyper-responsiveness to inflammatory stimuli (Bain et al., 2013, 2014). In obese mice, these gut CCR2+ pro-inflammatory monocytes/macrophages increase (Kawano et al., 2016; Rohm et al., 2017). In contrast, resident intestinal macrophages display anti-inflammatory features, such as IL-10 secretion, phagocytic activity, tissue repair, and anergy toward TLR stimulation (Bain et al., 2013). Mature CX3CR1+ macrophages are reduced in obesity (Hong et al., 2017; Luck et al., 2019). One of the initiating signals priming obesity-induced monocyte recruitment into the gut is mediated by the chemokine gradient generated by epithelial cells, which produce higher levels of CCL2/MCP-1. This was demonstrated in mouse models where macrophages lacked CCR2 and intestinal epithelial cells lacked CCL2/MCP-1, resulting in reduced colonic macrophage accumulation, intestinal permeability, inflammasome activation, and improved glucose metabolism (Kawano et al., 2016). In human obese subjects, CD14hi pro-inflammatory macrophages accumulate along the gastro-intestinal tract because of higher blood monocyte recruitment (Rohm et al., 2021). Moreover, greater body-mass index or waist-to-height ratio, as well as sedentary lifestyle, positively correlate with intestinal inflammation (Rohm et al., 2021). This correlation was confirmed by colon-specific macrophage depletion improving insulin sensitivity and glucose intolerance in obese mice (Rohm et al., 2018). Bacteria might be a prerequisite for intestinal inflammation and glucose dysregulation since germ-free mice display lower colonic macrophage numbers (Bain et al., 2014) and are partially protected from adiposity and glucose intolerance (Bäckhed et al., 2004, 2007).

Islet inflammation and impaired insulin production

Chronic low-grade inflammation in pancreatic islets is a hallmark of T2D, characterized by an increased number of innate immune cells, cytokines, and chemokines (Ehses et al., 2007; Maedler et al., 2002) (Figure 3). Following these initial reports, immunohistological and flow cytometric analysis confirmed and further characterized islet immune subsets in T2D (Butcher et al., 2014; Kamata et al., 2014; Martino et al., 2015; Nordmann et al., 2017; Richardson et al., 2009; Rodriguez-Calvo et al., 2014). Most cells express markers of M1-like macrophages but low numbers of other immune cells such as T and B cells may also be increased in islets in T2D (Böni-Schnetzler and Meier, 2019; Butcher et al., 2014; Ehses et al., 2007; Kamata et al., 2014; Martino et al., 2015; Nordmann et al., 2017; Richardson et al., 2009; Rodriguez-Calvo et al., 2014). Interestingly, increased frequency of immune cells (mostly CD8+ T cells) was also reported in the exocrine pancreas (Rodriguez-Calvo et al., 2014). Rodent models of obesity and T2D confirm‎ these observations. GK rats, leptin receptor-deficient db/db mice, and HFD-fed wild-type or KKAy mice show increased insulitis (Eguchi et al., 2012; Homo-Delarche et al., 2006). In the setting of HFD, the major mechanism of increased intra-islet macrophage numbers seems to be proliferation of resident macrophages (Ying et al., 2019).

Figure 3 Islet inflammation

Show full captionFigure viewer

These immune cells secrete cytokines and the first one described in islets from patients with T2D was IL-1β (Maedler et al., 2002). Islet IL-1β is derived from NLRP3 inflammasomes in intra-islet macrophages, and in the setting of T2D, mostly glucose, free fatty acids, and islet amyloid polypeptide activate this sensor of metabolic stress (Böni-Schnetzler et al., 2009; Maedler et al., 2002; Masters et al., 2010; Wen et al., 2011). Further, insulin was shown to promote IL-1β maturation in a glucose-dependent manner (Dror et al., 2017). Blocking IL-1 signaling in human islet cultures using IL-1Ra prevents the induction of IL-6, IL-8, CXCL1, and TNF-α and the chemokine CCL2/MCP-1, suggesting that these pro-inflammatory mediators are downstream of IL-1 receptor activation (Böni-Schnetzler et al., 2009; Igoillo-Esteve et al., 2010). Thus, it makes sense that inhibition of IL-1 signaling also prevents immune cell infiltration into the islet (Ehses et al., 2009; Sauter et al., 2015). Immune cell-derived cytokines, such as IL-1β, activate the highly expressed IL-1R1 on neighboring insulin-producing β-cells. Although, IL-1β acutely potentiates insulin secretion (Dror et al., 2017), prolonged activation of the IL-1 system in β-cells impairs insulin secretion and has a negative impact on β-cell mass, leading to β-cell failure and eventually hyperglycemia. Important for this concept, macrophage depletion, IL-1, and, to a certain extent, TNF-α blockade improve β-cell function and reduce glycemia in rodent models (Eguchi et al., 2012; Ehses et al., 2009; Nordmann et al., 2017; Sauter et al., 2015). This effect was also observed in a human study, in which the IL-1 pathway was blocked with IL-Ra (Larsen et al., 2007, 2009). This led to an increase in the ratio of circulating insulin/proinsulin in patients with T2D, indicative for improved β-cell function.

Islet resident macrophages are in a preactivated state even under normal conditions and IL-1β induces its own expression‎. Further, β-cell expression‎ of the counterregulatory IL-1R antagonist (IL-1Ra) is reduced in T2D (Maedler et al., 2004), tilting the IL-1/IL-1Ra balance toward a pro-inflammatory state. Supporting this concept, β-cell-specific IL-1Ra ablation reduces β-cell proliferation and mass and impairs insulin secretion (Böni-Schnetzler et al., 2018). Further, in response to inflammatory stimuli, β-cells secrete exosomes and miR-21-5p was identified to induce apoptosis in surrounding cells (Guay et al., 2015; Javeed et al., 2021; Lakhter et al., 2018) (Figure 2, pancreatic islet).

An important aspect of islet inflammation in the context of T2D is islet amyloidosis. The β-cell product islet amyloid polypeptide is a strong inducer of IL-1β in islets. It triggers the NLRP3 inflammasome to release mature IL-1β (Masters et al., 2010) and has the propensity to aggregate and form amyloid deposits. These are found in most patients with T2D (Clark et al., 1988) and are associated with β-cell apoptosis and a reduction in β-cell mass (Jurgens et al., 2011). Islet amyloid polypeptide secretion is increased in prediabetes because it is functionally coupled to the increased insulin demand. It not only activates immune cells but also recruits additional immune cells to the islet by triggering chemokine release from β-cells. Interestingly, in a human T2D study, only amyloid-positive islets showed increased immune cell infiltration (Kamata et al., 2014). The contribution of islet amyloidosis is generally underappreciated, as rodents do not show amyloid deposits due to differences in the polypeptide sequence. In support of this, transgenic expression‎ of human islet amyloid polypeptide leads to increased chemokine and cytokine expression‎ in a rodent model of long-term HFD feeding (Meier et al., 2014). Importantly, IL-1 blockade is able to mitigate islet amyloid pathogenesis (Westwell-Roper et al., 2015).

Taken together, in T2D, islets show increased immune cell infiltration and cytokine release, which in turn directly impairs β-cell mass and function. So far, IL-1β is identified as the major cytokine to mediate these pathologies because blocking IL-1 activation prevents deterioration of β-cell mass and function and improves glycemia. A detailed review of islet inflammation comparing human and rodent data was recently published (Böni-Schnetzler and Meier, 2019).

Beneficial physiological effects of inflammation on metabolism

Metabolism and the immune system are usually considered as two distinct entities. However, their functions are physiologically linked, supporting the concept of immunometabolism. Indeed, a key aspect of the immune system is to secure tissue integrity and repair, preserving homeostasis. In this function, immune cells consume a substantial amount of energy that needs to be mobilized and transported to these cells. Beyond the supply of energy, glucose also has a signaling function, activating a response to pathogens (Bantug et al., 2018; Murphy and O’Neill, 2018). If food is not available in sufficient quantities, as in starvation, glucose concentrations may be too low to support an immune response, thus saving energy at the risk of uncontrolled infection. Conversely, the immune system needs to control metabolism to secure its energy supply. Using infection as an example, an important component of the innate immune response to pathogens, is mediated through IL-1β, which is a strong stimulator of ACTH and cortisol release (Bataillard et al., 1992). Cortisol will induce IR with subsequent increased glucose concentrations that can fuel immune cells.

This competition or optimization of the distribution of cell nutrients regulated by the immune and endocrine-metabolic systems is apparent under several physiological situations such as during a meal or exercise. Indeed, following food ingestion, the gut flora and glucose stimulate the release of IL-1β by peritoneal macrophages (Dror et al., 2017; Traba et al., 2015). IL-1β then potentiates glucose-induced insulin secretion (Dror et al., 2017; Hajmrle et al., 2016; Maedler et al., 2002). Next, both insulin and IL-1β can regulate glucose disposal because IL-1β facilitates glucose uptake into immune cells. Thereby, IL-1β alerts and fuels the immune system, possibly to prevent the dissemination of microorganisms contained in the food. Concomitantly, insulin signals the availability of glucose and regulates its distribution to the SM, while surplus calories are stored as fat (Dror et al., 2017).

Another example of adaptive cross-talk between metabolism and immunity is the role of IL-6 during exercise. Beyond its function as a cytokine, IL-6 can be viewed as a myokine because it is produced and released by contracting SM during exercise (Pedersen and Febbraio, 2008). Interestingly, IL-6 released during physical activity also increases circulating GLP-1 levels (Adam and Westerterp-Plantenga, 2004; Ellingsgaard et al., 2011; Traub et al., 2017). Indeed, IL-6 stimulates GLP-1 secretion from intestinal L-cells and re-programs pancreatic α-cells to process proglucagon to GLP-1 via upregulation of the prohormone convertase 1/3 (Ellingsgaard et al., 2011; Traub et al., 2017). Thus, during exercise, IL-6-induced GLP-1 leads to metabolic adaptation by reducing gut motility and appetite. This also provides a trophic effect on the islets to prepare them for the insulin secretion needed for the expected following meal. Additionally, exercise-mediated mobilization of visceral fat requires IL-6 signaling (Wedell-Neergaard et al., 2018).

Several other cytokines contribute to the physiological regulation of metabolism. For example, IL-33 promotes insulin secretion via islet-resident group 2 ILCs (Dalmas et al., 2017). Another example is IL-22, which contributes to the preservation of the gut mucosal barrier and, thus, mitigates endotoxemia-induced IR (Wang et al., 2014). Furthermore, IL-22 protects β-cells from oxidative and endoplasmic reticulum stress (Hasnain et al., 2014).

Finally, intra- and peri-islet macrophages participate in the adaptive increase of β-cell mass in response to high calorie intake by promoting β-cell proliferation (Ying et al., 2019). M2-like macrophages may promote β-cell replication by upregulation of SMAD7 (Cao et al., 2014; Xiao et al., 2014).

Chronic effects of inflammation on the complications of diabetes

In contrast to the early beneficial regulatory effects of the immune system on metabolism described above, prolonged and exaggerated metabolic stress can lead to deleterious inflammatory reactions and precipitate autoinflammatory diseases. Although specific to each organ, a similar pattern is seen in many tissues affected by metabolic stress. Thus, increased nutrient exposure leads to cytokine and chemokine release, followed by immune cell recruitment and/or proliferation. In some cases, disproportional expansion of the tissue (e.g., adipocytes) or microvascular occlusions (retina, peripheral neurons) will lead to hypoxia and trigger an immune response. The inflammasome is a critical sensor of all these changes. Indeed, the NLRP3 inflammasome is activated by increased levels of glucose, fatty acids, cholesterol, and uric acid as well as by hypoxia (Duewell et al., 2010; Guo et al., 2015; Martinon et al., 2002, 2006; Stienstra et al., 2010; Tschöp and Thomas, 2006; Zhou et al., 2010). Depending on the specific tissue involved, this will lead to distinct pathologies, but the pattern of metabolic stress-induced inflammation remains similar. As detailed below, this can be observed in cardiovascular disease (CVD), NAFLD, retinopathy, nephropathy, neuropathy, and Alzheimer`s disease, among others.

Inflammation and cardiovascular disease

Almost 200 years ago, Carl von Rokitansky and Rudolf Virchow described atherosclerosis as an inflammatory process (Mayerl et al., 2006). Over the years, this view has been confirmed. Indeed, recent studies show that atherosclerosis is not only a complication of T2D as part of the metabolic syndrome, but also shares a similar pattern of metabolic stress-induced inflammation. Myeloid cells play a critical role in CVD inflammation, and an initiating trigger can be the IL-1 system in sensing metabolic changes.

Early in the development of atherosclerosis, monocytes adhere to the endothelial surface and then migrate into the artery wall, where they can become foam cells that accumulate around endothelial lesions (Gerrity, 1981). This is initiated by the NLRP3 inflammasome, which is activated by crystalline cholesterol and SFAs, whereas, replacement of saturated by monounsaturated FAs reduces IL-1β priming and secretion (Bevilacqua et al., 1985; Duewell et al., 2010; Finucane et al., 2015; Libby et al., 1985; Wen et al., 2011). IL-β can then modulate chemotaxis and adhesion of monocytes (Dinarello, 2009). Further proof for the role of IL-1β in CVD has been generated by the CANTOS study (see below).

The role of IL-1α is less clear, although it has been implicated in the development of vascular smooth muscle cell senescence as a contributor to atherosclerosis (Gardner et al., 2015). IL-1α may also promote later stage development of ischemic events (Cohen et al., 2010). Therefore, the beneficial effects of IL-1 receptor blockade after acute myocardial infarction may be due to antagonism of both IL-1β and IL-1α (Abbate et al., 2008).

The role of IL-6 in metabolism and in the context of CVD is complex (Febbraio, 2014). Because circulating IL-6 is elevated during obesity and correlates with the onset of diabetes (Spranger et al., 2003), it is often thought to have a negative effect on metabolism. This is supported by studies showing that IL-6 induces hepatic IR (Klover et al., 2003a, 2003b; Lagathu et al., 2003). This is in apparent contradiction to the physiological release of IL-6 by SM cells in response to muscle contraction (Febbraio and Pedersen, 2002) because IR decreases with exercise (Wojtaszewski et al., 2000). Furthermore, IL-6 is associated with reduced visceral adipose tissue mass, and blocking IL-6 leads to dyslipidemia (Wedell-Neergaard et al., 2018). A large-scale CV outcomes trial with an anti-IL-6 antibody is ongoing and should provide further answers on this subject (Ridker et al., 2021).

In summary, the role of inflammation in CVD with a particular role for macrophages and IL-1β has been demonstrated. It is likely that other cytokines also contribute and ongoing clinical trials will provide further insights.

Inflammation and retinopathy, nephropathy, neuropathy, and wound healing

The role of inflammation in the pathogenesis of diabetic retinopathy and macular edema is well established (Joussen et al., 2004; Mesquida et al., 2019). Chronic retinal inflammation is detectable from the early phases to the sight-threatening advanced stages of diabetic retinopathy (Rübsam et al., 2018). Key cytokines important to this process include IL-1β, IL-1-dependent IL-6, IL-8, and TNF-α, as well as monocyte adhesion to the endothelial wall followed by chemotaxis into the subendothelial space (Dong et al., 2013; Joussen et al., 2004; Simó-Servat et al., 2012). Important factors promoting these processes appear to be hypoxia due to microvasculature occlusions and hyperglycemia. The induced cytokines can damage the retinal vasculature, disrupting the blood-retinal barrier with subsequent macular edema and retinal neovascularization. A key mediator of these events is vascular endothelial growth factor (VEGF), which may be directly induced via hypoxia, IL-1β, IL-6, insulin, insulin-like growth factor-1 (IGF-1), and fibroblast growth factor (FGF) (Nagineni et al., 2012; Simó et al., 2014). The importance of cytokines in this process may explain why VEGF is overexpressed in diabetic macular edema despite the absence of overt hypoxia (Simó et al., 2014).

Innate immunity is also involved in the development and progression of diabetic nephropathy (Tang and Yiu, 2020). Mechanistically, TLR4 is overexpressed in kidneys of patients with T2D and correlates positively with HbA1c levels and negatively with renal function (Lin et al., 2012). Furthermore, animal studies support the participation of TLR2 and TLR4 in diabetic kidney disease (Jialal et al., 2014; Lin et al., 2012; Ma et al., 2014a, 2014b). In addition, NLRP3 inflammasome activation of IL-1β plays an important role by sensing metabolic stress in the diabetic kidney (Fang et al., 2013; Shahzad et al., 2016; Tang and Yiu, 2020; Vilaysane et al., 2010).

Although diabetic polyneuropathy has major clinical consequences such as diabetic foot ulcers and amputation, research on its pathogenesis is limited. Traditionally, diabetic neuropathy is classified as non-inflammatory in contrast to neuropathies such as Guillain-Barré syndrome or demyelinating neuropathy (Herder et al., 2019). However, several recent studies have uncovered strong correlations between inflammation and diabetic neuropathy (Bönhof et al., 2019; Herder et al., 2017, 2018, 2019). Nevertheless, mechanistic data are limited.

Wound healing is delayed in T2D and worldwide approximately 20% of patients with T2D develop diabetic wounds. Overactive pro-inflammatory cytokine pathways are a major contributor to this T2D co-morbidity (Geng et al., 2021).

Inflammation and Alzheimer’s disease

The correlation between T2D and the risk of Alzheimer’s disease (AD) is well-known, and some have suggested that AD could be considered type 3 diabetes (Nguyen et al., 2020). Two major mechanisms have been widely reported in regard to the etiology of AD. One is the aggregation of Aβ monomers into amyloid fibrils or plaques following extracellular proteolytic cleavage of β-amyloid precursor protein (APP) (Braak and Braak, 1991; Kocahan and Doğan, 2017; Murphy and LeVine, 2010; Selkoe, 1994). Second, hyperphosphorylation of TAU leads to the formation of extracellular tangles as well as neuronal microtubular dysfunction (Alonso et al., 1994, 1996, 1997, 2018). Chronic neuroinflammation is now recognized as an additional feature of Alzheimer’s, and this is due to activation of brain resident macrophages, termed microglia (Akiyama et al., 2000; Kinney et al., 2018; Leng and Edison, 2021). These microglia release a variety of cytokines and other factors that might play a functional role in the pathogenesis of AD and have been referred to as disease-associated microglia (Keren-Shaul et al., 2017). Although neuroinflammation with microglia activation may not be the initiating mechanism for AD, they appear to play a role in augmenting the Aβ and TAU etiologies (Dani et al., 2018; Hanisch, 2002; Hayes et al., 2002).

In the normal state, microglia are inactive and function to assess the local microenvironment for harmful molecules and communicate to neuronal cells (Arcuri et al., 2017). The increased presence of Aβ may serve as an activating signal for microglia, and these cells then help clear excessive Aβ from the brain. During the chronically inflamed and activated state, the ability of the microglia to clear Aβ declines, which would promote the development of fibrils and plaques (Bolmont et al., 2008; Heneka et al., 2015a; Hickman et al., 2008). Several genetic studies show that a large proportion of the AD risk alleles are associated with genes that are primarily or exclusively expressed in microglia (Hansen et al., 2018). Most of these variants are located in noncoding areas and might function to promote the AD phenotype due to effects on enhancer and or promoter activity (Nott et al., 2019). Many of these genes reside in microglial pathways that involve phagocytosis, autophagy, or lysosomal function (Nott et al., 2019). Reduced activity of these pathways can be a cause of increased cytokine production by microglia. There is also evidence that neuroinflammation can cause bone marrow-derived macrophage recruitment into the brain and participate in AD pathology (Thériault et al., 2015). These concepts are supported by the identification of genetic variants in the macrophage receptor TREM2, which cause enhanced risk for AD development (Guerreiro et al., 2013; Jonsson et al., 2013; Jonsson and Stefansson, 2013; Sims et al., 2017). TREM2 is highly expressed in microglia, but the exact mechanisms underlying the enhanced risk for AD remains unclear (Keren-Shaul et al., 2017). Because TREM2 is a receptor for a yet-to-be-identified lipid species, it is possible that activation of TREM2 with a ligand or an activating antibody could represent a potential therapeutic approach.

Similar to macrophages throughout the body, activated microglia also release a number of cytokines that may participate in AD pathogenesis (Del Bo et al., 1995; Hanisch, 2002; Heneka et al., 2013, 2015b; Smith et al., 2012). IL-1β might be the most important cytokine released in this context and is elevated in the brains of AD patients (Liu and Quan, 2018). IL-1β can cause increased secretion of Aβ precursor APP and promote enhanced processing to Aβ, further increasing fibril and plaque formation (Dash and Moore, 1995). Because Aβ also activates the NLRP3 inflammasome in microglia promoting the release of IL-1β, this could provide a feed-forward mechanism sustaining or increasing AD (Goldgaber et al., 1989; Yates et al., 2000). TNF-α and IL-6 are also released by activated microglia and lead to increased APP production, hyperphosphorylation of TAU, and conversion of APP to Aβ by promoting β-secretase production (Chen et al., 2012a; Quintanilla et al., 2004). In addition to microglia, astrocytes can play a role in neuroinflammation, as activated microglial cells release cytokines that promote the development of reactive astrocytes that actively participate in neuronal cell death and damage (Liddelow et al., 2017).

The prevailing view is that Aβ monomers derived from APP proteolysis and hyperphosphorylated TAU protein are the prevailing mechanisms underlying the tangles and amyloid plaques, which define AD brain pathology. However, the resident microglia, as well as brain astrocytes, appear to be important in promoting the processes that ultimately lead to AD and all of its clinical sequalae.

Additional obesity and diabetes-related inflammatory disorders

As discussed above, obesity and T2D have strong inflammatory components and chronic activation of the innate immune system may induce or precipitate other conditions driven by inflammation. Some examples are given below.

Polycystic ovary syndrome (PCOS) is a common disease defined by a combination of androgen excess and ovarian dysfunction, leading to hirsutism, anovulation, and infertility (Escobar-Morreale, 2018). Interestingly, PCOS is strongly associated with IR and diabetes. A possible explanation for this is that inflammation might participate in PCOS. Indeed, women with PCOS show increased circulating markers of inflammation (Kelly et al., 2001). Furthermore, the ovary expresses IL-1β and IL-1 receptor type 1 with peri-ovulatory changes compatible with the notion that it has effects on ovulation (Kol et al., 1999). This is supported by several studies showing the impact of IL-1β on steroidogenesis (Popovic et al., 2019). Clinical intervention studies to determine whether IL-1β is causative in PCOS are ongoing (https://clinicaltrials.gov/ct2/show/NCT03578497).

Similar to the association of PCOS with T2D in women, obesity and T2D in men are frequently associated with hypogonadism (Rao et al., 2013). The association between low testosterone and T2D might also be partly explained by inflammation. Thus, TNF-α, IL-6, and IL-1β can indirectly inhibit the secretion of testosterone by reducing hypothalamic-pituitary gonadotropin production and can also directly inhibit testicular and androgen production (Jones and Kennedy, 1993). This idea has recently been supported by a clinical study showing that treatment with an IL-1 antagonist increased testosterone levels in obese men with testosterone deficiency (Ebrahimi et al., 2018).

Some inflammatory diseases with an etiology unrelated to T2D may be exacerbated or precipitated by the concomitant occurrence of both conditions. A striking example is rheumatoid arthritis. Its primary pathology is autoimmune joint inflammation (Klareskog et al., 2009). Although the etiology of rheumatoid arthritis is unrelated to T2D, the preval‎ence of both conditions together is increased (Ruscitti et al., 2017a; Ruscitti et al., 2017b). The molecular mechanism could involve cytokines, such as TNF-α and IL-1β, promoting both rheumatoid arthritis and T2D (Dinarello, 2011). Interestingly, monocytes from patients with rheumatoid arthritis and T2D display increased production of IL-1β via NLRP3-inflammasome activation compared with monocytes from patients with only one of the two conditions (Ruscitti et al., 2015). Accordingly, a recent clinical study of IL-1 antagonism in patients with rheumatoid arthritis and T2D showed a remarkable improvement in both glycemia and joint disease activity ((Ruscitti et al., 2019); for details, see below). Autoimmune skin diseases such as psoriasis are also associated with a profound increase in preval‎ence of T2D. A large longitudinal study found a preval‎ence of 11.4% of T2D in patients with psoriasis (Holm and Thomsen, 2019). Other examples of diseases that may be exacerbated by T2D are gout and Crohn disease (Donath, 2014).

An unexpected role of inflammation has been reported in postprandial hypoglycemia after bariatric surgery. This condition presents acutely after meal ingestion with disabling neuroglycopenic symptoms. This may result in a long-term increase in food intake and subsequent weight regain. The underlying mechanism is related to the rapid transit of food without gastric delay because of the anatomical changes, leading to an early peak in glycemia. This, in turn, causes an exaggerated induction of the postprandial rise in IL-1β-induced insulin secretion (see Dror et al., 2017; Maedler et al., 2002; and above), leading to an overshoot of the insulin response, followed by hypoglycemia. This is supported by a proof-of-concept clinical study, which tackled this process at two steps (Hepprich et al., 2020). The initial hyperglycemic peak was reduced via SGLT-2-inhibition while the effect of IL-1β was blocked with an IL-1 receptor antagonist. Both drugs reduced postprandial insulin release and prevented episodes of hypoglycemia.

Therapeutic approaches targeting inflammation

The first clinical evidence for the beneficial effect of anti-inflammatory treatment in diabetes was published in 1876 (Ebstein, 1876). This empirical study showed that sodium salicylate improved glycemia in patients with diabetes. More than a hundred years later, Shoelson and colleagues showed that this anti-diabetic effect is mediated via inhibition of the NF-κB pathway (Yuan et al., 2001). They went on to validate the effectiveness of salsalate to improve glycemia in proof-of-concept studies (Faghihimani et al., 2013; Fleischman et al., 2008; Goldfine et al., 2008, 2013a; Koska et al., 2009) followed by larger placebo-controlled studies (Goldfine et al., 2010, 2013b). The only safety signal was a small increase in low-density lipoprotein cholesterol levels and a reversible rise in urinary albumin. These clinical studies add to the evidence for a role of chronic inflammation in diabetes.

The initial discovery by Hotamisligil and Spiegelman on the effect of TNF-α in IR (Hotamisligil et al., 1993) was followed by a rapid attempt to translate the findings into the clinic. Unfortunately, these studies were not adequately designed, with low sample size and duration of treatment e.g., a single dose in 7–10 patients for a few days (Bernstein et al., 2006; Dominguez et al., 2005; Ofei et al., 1996; Paquot et al., 2000). Taking into account the number of variable factors that can influence a metabolic response, including genetic background, body weight, food intake, and exercise, it is likely that the design of these studies was insufficient to detect changes, explaining the apparent lack of metabolic effects of TNF-α antagonism. For details on the limitations of these studies, see Donath (2014). However, indirect evidence supports the clinical potential of TNF-α antagonists in T2D. Thus, these drugs improve glycemia in obese subjects without diabetes as well as in patients with psoriasis, rheumatoid arthritis, and Crohn's disease (Gonzalez-Gay et al., 2006; Huvers et al., 2007; Kiortsis et al., 2005; Marra et al., 2007; Stanley et al., 2011; Timper et al., 2013; Yazdani-Biuki et al., 2004, 2006). Furthermore, TNF-α antagonism reduces the incidence of T2D in patients with rheumatoid arthritis or psoriasis (Antohe et al., 2012; Solomon et al., 2011). However, it is unclear whether the beneficial effect on glycemia was direct or due to an improvement in the underlying disease with subsequent enhanced exercise capacity. On the other hand, a recent study directly compared TNF-α versus IL-1 blockade in patients with rheumatoid arthritis and T2D. TNF-α antagonism failed to decrease glycemia (Ruscitti et al., 2019). Therefore, a well-designed diabetes-focused study of TNF antagonism, ideally with clinically relevant outcomes, such as CVD, is warranted.

The most advanced clinical translation of inflammation and T2D has been achieved with IL-1 antagonism. An initial trial in patients with T2D showed that anakinra (a recombinant human IL-1Ra) improved β-cell secretory function and reduced glycemia (Larsen et al., 2007). The ability of IL-1 inhibitors to improve insulin secretion was confirmed in several follow-up studies using anakinra (van Asseldonk et al., 2011; van Poppel et al., 2014). Similarly, treatment of T2D subjects with neutralizing anti-IL-1β antibodies demonstrated beneficial effects, although the magnitude varied depending on baseline HbA1c levels and sample size (Cavelti-Weder et al., 2012; Rissanen et al., 2012; Ruscitti et al., 2019; Sloan-Lancaster et al., 2013).

In a large CV outcome study (CANTOS), treatment with an anti-IL-1β antibody prevented CV events (Ridker et al., 2017). A sub-analysis showed that blocking IL-1β also significantly decreased HbA1c during the first 6–9 months of treatment, with waning effect toward the end of the study (Everett et al., 2018). This attenuation of the glucose lowering effect may be due to the design of the study, which allowed for lifestyle interventions and adaptations of standard anti-diabetic drugs. Supporting this assumption, in the non-diabetic patients, anti-IL-1β treatment decreased HbA1c for the duration of the study. Furthermore, IL-1 antagonism prevented new onset of diabetes for about 4 years. After 4 years, the number of patients followed decreased by 90% and the effect of anti-IL-1β could no longer be detected. The observed prevention of diabetes suggests an ongoing islet inflammatory process in patients with prediabetes. Most importantly, this outcome study showed that in a patient population with T2D, CV complications can be prevented (Everett et al., 2018). An additional beneficial effect of IL-1β inhibition was prevention of heart failure, particularly in patients with higher body-mass index and diabetes (Everett et al., 2019), confirm‎ing previous clinical proof-of-concept studies (Abbate et al., 2010, 2015; Van Tassell et al., 2017, 2018). Although the CANTOS study showed the overall safety of prolonged IL-1β inhibition, in patients with severe infections necessitating inpatient treatment, IL-1β antagonism was associated with a higher incidence of fatal infections, which warrants caution in patients at risk.

Following these studies, a meta-analysis involving 2,921 cases with T2D treated with IL-1-blocking therapy demonstrated a highly significant reduction in HbA1c (p < 0.00001) (Kataria et al., 2019).

Recently, a new designer ligand for the gp130 receptor with CNTF-like properties that signals in an IL-6-receptor-dependent manner was shown to improve glycemia and prevent body weight gain in HFD-fed mice (Findeisen et al., 2019), showing that designer approaches targeting specific pathways but avoiding activation of others have great potential for future human studies.

A list of anti-inflammatory clinical trials targeting T2D and its associated pathologies is given in Table 1.

TargetsDrugsReferences

Table 1

Clinical studies of anti-inflammatory treatments in patients with type 2 diabetes and associated complications

• Open table in a new tab

Concluding remarks

At this point, the only anti-diabetic drugs demonstrating improvement in macrovascular disease are GLP-1 analogs (Marso et al., 2016), while SGLT-2 inhibitors are the only drug class demonstrating protective effects against heart failure and nephropathy (Zinman et al., 2015). This explains why these two classes of drugs are currently often used as first line anti-diabetic treatment as opposed to insulin, sulfonylureas, and DPP-4 inhibitors, which only lower glucose. An anti-inflammatory drug may provide additional benefits beyond improving glycemia. Indeed, as detailed above, IL-1β antagonism mitigated β-cell dysfunction and improved glycemia, CV complications, and heart failure (Abbate et al., 2010, 2015; Cavelti-Weder et al., 2012; Donath, 2014; Everett et al., 2019; Larsen et al., 2007; Ridker et al., 2017, 2018; Rissanen et al., 2012; Ruscitti et al., 2019; Sloan-Lancaster et al., 2013; Van Tassell et al., 2017, 2018). Future studies are needed to determine whether anti-inflammatory treatment may counteract other complications of diabetes, including NASH, nephropathy, fatigue, polyneuropathy, retinopathy, and macular edema (Cavelti-Weder et al., 2011; Febbraio et al., 2019; Herder et al., 2009, 2013, 2017, 2018; Lehrskov et al., 2018; Mesquida et al., 2014; Schlesinger et al., 2019; Stahel et al., 2016; Sun and Karin, 2012; Tesch, 2017). Beyond IL-1β antagonism, other immunomodulatory drugs, either alone or in combination, may be more effective. However, the choice of an anti-inflammatory drug should be based on a pathophysiologic mechanistic understanding and not on a non-specific effect, as with methotrexate, which did not decrease IL-1β and failed to prevent CV complications in patients with metabolic syndrome (Ridker et al., 2019). An overview of validated and potential future pro-inflammatory targets in human T2DM and associated diseases is given in Figure 4.

Figure 4 Immunomodulatory targets for the treatment of diabetes and its complications

Show full captionFigure viewer

The association of diabetes with other inflammatory diseases such as gout or rheumatoid arthritis allows treatment of two conditions with one therapeutic agent. This has been demonstrated by Ruscitti and Giacomelli in patients with rheumatoid arthritis and T2D (Ruscitti et al., 2019). Treatment of these patients with anakinra was highly effective, achieving a decrease in HbA1c by >1% after 6 months, while simultaneously reducing rheumatoid disease activity. Therefore, anti-inflammatory treatment can improve both conditions and is already doable in countries where IL-1 antagonism is approved for rheumatoid arthritis or gout.

Several companies are developing inhibitors of the NLRP3-IL-1β pathway, including small molecules that can be taken orally, avoiding injection site reactions of IL-1Ra. The short half-life of these oral agents would also help in the management of severe infections. Hopefully, in the near future, these inhibitors will be tested in patients with T2D and associated complications. An encouraging phase 1b study using an oral NLRP3 inhibitor, dapansutrile, showed signs of improved heart failure and significantly decreased blood glucose levels (Wohlford et al., 2020).

Acknowlegments

Funding supporting the work in this review was from the Swiss National Foundation (SNF) Early Postdoc Mobility grant (P2BSP3_200177 to T.R.) and by the U.S. National Institute of Diabetes and Digestive and Kidney Diseases (P30 DK063491 and DK101395 to J.M.O.).

Declaration of interests

M.Y.D. is listed as the inventor on a patent filed in 2003 for the use of an IL-1 receptor antagonist for the treatment of or prophylaxis for type 2 diabetes.

References

1.

Abbate, A. ∙ Salloum, F.N. ∙ Vecile, E. ...

Anakinra, a recombinant human interleukin-1 receptor antagonist, inhibits apoptosis in experimental acute myocardial infarction

Circulation. 2008; 117:2670-2683

Google Scholar

2.

Abbate, A. ∙ Kontos, M.C. ∙ Grizzard, J.D. ..., VCU-ART Investigators

Interleukin-1 blockade with anakinra to prevent adverse cardiac remodeling after acute myocardial infarction (Virginia Commonwealth University Anakinra Remodeling Trial [VCU-ART] Pilot study)

Am J Cardiol. 2010; 105:1371-1377

Google Scholar

3.

Abbate, A. ∙ Kontos, M.C. ∙ Abouzaki, N.A. ...

Comparative safety of interleukin-1 blockade with anakinra in patients with ST-segment elevation acute myocardial infarction (from the VCU-ART and VCU-ART2 pilot studies)

Am. J. Cardiol. 2015; 115:288-292

Google Scholar

4.

Abrams, G.A. ∙ Kunde, S.S. ∙ Lazenby, A.J. ...

Portal fibrosis and hepatic steatosis in morbidly obese subjects: A spectrum of nonalcoholic fatty liver disease

Hepatology. 2004; 40:475-483

Google Scholar

5.

Adam, T.C. ∙ Westerterp-Plantenga, M.S.

Activity-induced GLP-1 release in lean and obese subjects

Physiol. Behav. 2004; 83:459-466

Google Scholar

6.

Akiyama, H. ∙ Barger, S. ∙ Barnum, S. ...

Inflammation and Alzheimer’s disease

Neurobiol. Aging. 2000; 21:383-421

Google Scholar

7.

Al-Sadi, R.M. ∙ Ma, T.Y.

IL-1beta causes an increase in intestinal epithelial tight junction permeability

J. Immunol. 2007; 178:4641-4649

Google Scholar

8.

Alonso, A.C. ∙ Zaidi, T. ∙ Grundke-Iqbal, I. ...

Role of abnormally phosphorylated tau in the breakdown of microtubules in Alzheimer disease

Proc. Natl. Acad. Sci. USA. 1994; 91:5562-5566

Google Scholar

9.

Alonso, A.C. ∙ Grundke-Iqbal, I. ∙ Iqbal, K.

Alzheimer’s disease hyperphosphorylated tau sequesters normal tau into tangles of filaments and disassembles microtubules

Nat. Med. 1996; 2:783-787

Google Scholar

10.

Alonso, A.D. ∙ Grundke-Iqbal, I. ∙ Barra, H.S. ...

Abnormal phosphorylation of tau and the mechanism of Alzheimer neurofibrillary degeneration: sequestration of microtubule-associated proteins 1 and 2 and the disassembly of microtubules by the abnormal tau

Proc. Natl. Acad. Sci. USA. 1997; 94:298-303

Google Scholar