Re: 패혈증 PICS 중환자 단백질 분해 기전 2025 cell

[문형철]

거의 대부분 난치병 환자는 체중이 40~50kg 정도다

대개 60~80kg 체중에서 20kg이상 체중이 빠진다.

그 기전을 탐구중이다.

https://cafe.daum.net/panicbird/OXYY/84

Re: 패혈증의 면역반응패턴(면역과잉--＞ 면역마비) PICS 중환자 단백질 분해 이유!!

이 논문은

Cell Press의 Med 저널 (Impact Factor 높은 고영향력 저널)에

2025년 1월 16일 온라인 공개 (정식 발행 2025년 5월 9일, Volume 6, Issue 5, Article 100569)된 최신 연구입니다.

제목: Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis (패혈증 후 지속 염증·면역억제·분해 증후군(PICS)의 면역 세포 서명).

저자: Xing-Feng Sun, Wen-Chen Luo 등 (주요 교신저자: Zhi-Xin Qiu, Jing Zhong, Chang-Hong Miao; Fudan University Zhongshan Hospital 중심 중국 연구팀).

연구 배경 및 목적

패혈증 후 PICS는 ICU 환자에서

지속 염증 + 면역 억제 + catabolism(소모/체중 감소 등)으로

삶의 질 저하와 사망을 초래하지만,

단일 세포 수준 면역 landscape가 불분명했습니다.

이 연구는 단일 세포 RNA 시퀀싱(scRNA-seq)으로

PICS 환자의 말초 혈액 면역 세포 변화를 분석해 면역 이질성(heterogeneity)을 밝히고,

예후 예측 및 치료 타깃을 탐색했습니다.

방법

• 환자군: 65세 이상 노인 대상
  • PICS 환자 6명 (PICS-death 2명, PICS-alive 4명)
  • 급성 패혈증 환자 5명
  • 건강 대조군(HC) 5명
• PBMC(말초 혈액 단핵구 세포) scRNA-seq (10x Genomics Chromium + Illumina 시퀀싱)
• 분석: Seurat 클러스터링, 차등 발현, GO/GSEA, 세포 간 상호작용(CellphoneDB/CellChat)
• 검증: 쥐 CLP(cecal ligation and puncture) 모델로 PICS 유발 후 megakaryocyte(MK) 기능 확인

주요 결과 (면역 세포 서명 핵심)

1. 단핵구(Monocytes) – MDSC-like 특징
   • Mono1 & Mono4 서브셋: sepsis에서 면역 기능 크게 억제 (HLA-DR ↓, S100A8/S100A9 ↑) → PICS에서 부분 회복되지만 HC보다 열등
   • MDSC-like (myeloid-derived suppressor cell) 특성: B 세포 & CD8 T 세포에 면역 억제 + pro-apoptotic(세포 사멸 유도) 효과
   • PICS-death 그룹: Mono1-6 비율 ↓ → 예후 불량 예측 바이오마커 가능
2. B 세포 & Plasma cells
   • sepsis/PICS: naive B & memory B ↓, plasma cells 증식
   • PICS: naive/memory B의 antigen processing/presentation 유전자 활성화 (sepsis보다 ↑)
   • 예후 좋은 PICS-alive: active memory B & IGHA1-plasma cells ↑
   • PICS-death: IGHG1-plasma cells ↑ (항체 생산 과다 → 예후 불량?)
3. T 세포
   • CD8 TEMRA: PICS-death에서 proliferation & immune dysfunction 징후
   • Treg: PICS-death에서 ↓, DUSP1/FOS ↑ → 증식 억제 & apoptosis 촉진
   • CD8 T: LAG3/TIGIT 등 exhaustion marker ↑
4. Megakaryocytes (MK, 거대핵세포)
   • PICS & sepsis: MK proliferation ↑ (HC 대비)
   • MK1 서브셋: IFITM2/3, S100A9 등 면역 조절 유전자 ↑ → anti-inflammatory & immunomodulatory 효과
   • 쥐 모델 검증: PICS 유래 MK → TNF-α/IL-6 ↓ (항염증 확인)
5. 세포 간 상호작용
   • Mono1/4 → TNFSF13-TNFRSF14 등으로 B/T 세포에 pro-apoptotic/억제 신호
   • PICS-death: 상호작용 강도 ↓ → 면역 기능 더 억제

논의 및 결론

• PICS는 면역 세포 재프로그래밍(재배치)이 핵심: MDSC-like monocyte 억제, B 세포 dysfunction, T 세포 exhaustion, MK 항염증 역할
• IFN-γ 신호 장애 → IFN-γ 치료 잠재적
• 특정 프로파일 (e.g., IGHA1-plasma ↑ in alive, Mono 비율 ↓ in death)이 예후 바이오마커
• PICS 면역 이질성 이해 → personalized intervention (면역 강화/조절) 기반 마련

PICS(Persistent Inflammation, Immunosuppression, and Catabolism Syndrome)에서
면역 세포의 재프로그래밍(reprogramming)은
세포의 기능적·전사적 변화(예: 유전자 발현 패턴 재배치)를 의미하며,

이는
패혈증 후 지속적인 염증, 면역 억제,
그리고 catabolism(체중/근육 소실 등 소모 증상)을 유지하는 핵심 기전입니다.

이는 최근 2025년 Cell Med 논문(Sun et al.)에서
단일 세포 RNA 시퀀싱(scRNA-seq)을 통해 밝혀진 내용으로,
PICS 환자의 말초 혈액 면역 세포가
건강인이나 급성 패혈증 환자와 다르게 재구성되어 악순환을 형성합니다.

아래에서 요청하신
포인트(MDSC-like monocyte 억제, B 세포 dysfunction, T 세포 exhaustion, MK 항염증 역할)를
하나씩 자세히 설명하겠습니다.

간단히 요약하면
PICS는 초기 과염증 후 면역 세포가 "재설정"되어 회복되지 않고
만성 상태로 고착된다는 것입니다.

1. MDSC-like Monocyte 억제 (MDSC-like 단핵구의 억제 역할)

• 기본 개념: MDSC(myeloid-derived suppressor cells)는 골수 유래 억제 세포로, 정상적으로 면역 과반응을 막지만 PICS처럼 과도하게 확장되면 T/B 세포를 억제해 면역 마비(immunoparalysis)를 유발합니다. 여기서 "MDSC-like monocyte"는 단핵구(monocyte)가 MDSC와 유사한 특성을 띠는 것을 의미합니다.
• PICS에서의 재프로그래밍: 논문에서 Mono1과 Mono4 서브셋(단핵구 하위군)이 주요 타깃으로, 패혈증 후 PICS로 진행 시 HLA-DR(항원 제시 관련 분자) 발현이 감소하고 S100A8/S100A9(염증 관련 단백질) 발현이 증가합니다. 이는 세포가 "억제 모드"로 재프로그래밍되어 B 세포와 CD8 T 세포에 pro-apoptotic(세포 사멸 유도) 신호를 보냅니다.
• 상세 기전:
  • 세포 간 상호작용: Mono1/4가 TNFSF13-TNFRSF14 등의 리간드-수용체 쌍을 통해 다른 면역 세포를 억제. 이는 CellphoneDB 분석으로 확인됨.
  • 예후 연관: PICS 사망군(death group)에서 Mono1-6 비율이 현저히 감소 → 이는 면역 기능이 더 약해진 상태를 반영. 건강인 대비 PICS 환자에서 부분 회복되지만 여전히 열등(inferior)합니다.
• 임상 의미: 이 억제가 지속되면 재발 감염(opportunistic infection)이 증가하고 catabolism(체중 감소)이 악화됩니다. 치료 타깃으로 MDSC 억제제(예: all-trans retinoic acid)가 유망합니다.

2. B 세포 Dysfunction (B 세포 기능 장애)

• 기본 개념: B 세포는 항체 생산과 면역 기억을 담당하지만, PICS에서 기능이 재프로그래밍되어 항원 처리/제시 능력이 왜곡됩니다.
• PICS에서의 재프로그래밍: 논문에서 naive B 세포와 memory B 세포 수가 감소하고, plasma cells(플라즈마 세포, 항체 생산 세포)가 증식합니다. 특히 PICS 환자에서 naive/memory B의 항원 처리 관련 유전자(antigen processing/presentation genes)가 과활성화되지만, 이는 역설적으로 면역 불균형을 초래합니다.
• 상세 기전:
  • 서브셋 변화: PICS 생존군(alive)에서 active memory B와 IGHA1-plasma cells(주로 IgA 생산)이 증가 → 보호적 역할(항체 생산 강화).
  • PICS 사망군: IGHG1-plasma cells(IgG 생산)이 과다 → 과도한 항체 생산이 염증을 유지하거나 면역 피로를 유발.
  • GO/GSEA 분석: B 세포 경로에서 면역 dysfunction 관련 유전자가 업레귤레이션(upregulation)됨.
• 임상 의미: B 세포 dysfunction은 항체 의존적 감염 방어가 약해져 2차 감염을 증가시킵니다. 이는 PICS의 면역 억제 측면을 강조하며, IVIG(면역글로불린) 치료나 B 세포 자극제가 잠재적 옵션입니다.

3. T 세포 Exhaustion (T 세포 고갈/피로)

• 기본 개념: T 세포 exhaustion은 지속 자극으로 T 세포가 기능적으로 소진되는 상태로, 억제 마커(PD-1, LAG3 등) 발현 증가와 cytokine 생산 감소를 특징합니다.
• PICS에서의 재프로그래밍: 논문에서 CD8 TEMRA(효과기 기억 T 세포)와 Treg(조절 T 세포)가 주요 변화. PICS 사망군에서 CD8 TEMRA의 proliferation(증식) 관련 유전자가 억제되고, exhaustion marker(LAG3, TIGIT)가 증가합니다. Treg은 DUSP1/FOS 유전자 발현 ↑로 apoptosis가 촉진됩니다.
• 상세 기전:
  • CD8 T 세포: naive/memory CD8 손실 → lymphocytopenia(림프구 감소증) 유발. PICS에서 TIGIT 등 checkpoint molecule ↑ → 세포가 "피로" 상태로 재프로그래밍.
  • Treg 변화: PICS 사망군에서 Treg 수가 감소하지만, 남은 Treg이 더 강한 억제 기능을 발휘 → 전체 면역 균형 붕괴.
  • 상호작용: MDSC-like monocyte로부터 억제 신호를 받아 T 세포 exhaustion 강화.
• 임상 의미: T 세포 exhaustion은 병원체 제거 능력을 떨어뜨려 장기 사망률을 높입니다. 치료로 immune checkpoint inhibitors(anti-PD-1 등)나 IL-7( T 세포 활성화)이 연구 중입니다.

4. MK 항염증 역할 (Megakaryocytes의 항염증 역할)

• 기본 개념: MK(거대핵세포)는 혈소판 생산 세포지만, 최근 면역 조절 역할(항염증)이 밝혀지고 있습니다.
• PICS에서의 재프로그래밍: 논문에서 PICS와 패혈증 환자에서 MK 수가 증가(proliferation ↑)하며, MK1 서브셋이 IFITM2/3, S100A9 등 유전자를 과발현합니다. 이는 anti-inflammatory(항염증) 및 immunomodulatory(면역 조절) 효과를 발휘합니다.
• 상세 기전:
  • MK1 특징: interferon-related genes(IFITM) ↑ → 바이러스/염증 방어.
  • 쥐 모델 검증: PICS 유래 MK를 이식하면 TNF-α/IL-6 같은 프로염증 cytokine가 감소 → 항염증 확인.
  • 전체 역할: MK가 면역 세포 재프로그래밍의 "브레이크" 역할을 하지만, PICS에서 과도한 MK는 혈전(thrombosis) 위험을 높일 수 있음.
• 임상 의미: MK의 항염증 역할은 PICS의 지속 염증을 완화할 수 있지만, 균형이 깨지면 해로울 수 있습니다. 치료 타깃으로 MK 조절 요법(예: thrombopoietin agonist)이 탐색 중입니다.

전체 요약 및 임상 함의

이 재프로그래밍은
PICS의 악순환을 유지:
MDSC-like monocyte와 T/B 세포 dysfunction이 면역 억제를 주도하고,
MK가 이를 부분적으로 완화하지만 불완전합니다.

결과적으로
IFN-γ 신호 장애가 발생해 면역 재건이 어렵습니다.

논문은
이를 바탕으로 개인화 치료(예: MDSC 타깃 약물, IFN-γ 보충)를 제안합니다.

이는 PICS 환자(특히 노인)의 예후 예측(예: Mono 비율, IGHA1-plasma)에 유용하며,
미래 연구로 다인종/다양 연령 코호트 확장이 필요합니다.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

패혈증 후 지속성 염증·면역억제·이화증후군(PICS)은

중환자실 환자들에게 여전히 큰 도전 과제입니다.

그러나

PICS의 면역 landscape는 아직 잘 알려져 있지 않습니다.

본 연구에서는

단일세포 해상도(single-cell resolution) 프로파일링을 통해

패혈증 및 PICS 환자의 말초혈액 내 면역세포 변화를 상세히 분석하였습니다.

연구팀은

PICS에서 면역 기능 장애를 일으키는 주요 면역세포 클러스터를 규명하였으며,

특히 골수유래 억제세포(myeloid-derived suppressor cell)와 유사한 단핵구(monocyte) 아집단,

B세포,

조절 T세포(regulatory T cell),

거핵세포(megakaryocyte) 등의 특징을 밝혔습니다.

또한

PICS 환자의 예후와 관련된

특정 면역 프로파일을 확인하였습니다.

이러한 결과는

패혈증 후 PICS의 면역 이질성(heterogeneity)에 대한 새로운 통찰을 제공하며,

향후 이들 환자를 위한 치료제 개발에 중요한 지침이 될 수 있습니다.

주요 내용(Highlights)

• 단일세포 프로파일링을 통해 PICS 환자 말초혈액에서 면역세포 재프로그래밍(immune-cell reprogramming)이 드러남
• 패혈증 환자에서 단핵구 아집단이 크게 억제되었으며, PICS 단계에서 부분적으로 회복됨
• 예후가 좋은 PICS 환자에서는 활성 기억 B세포(memory B cell)와 IGHA1-플라스마세포가 더 많음
• 거핵세포(megakaryocytes)는 PICS에서 항염증 및 면역조절 효과를 나타냄

Summary

Background

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

요약(Summary)

배경 패혈증 후 지속성 염증·면역억제·이화증후군(PICS)은 중환자실 환자들에게 삶의 질 저하와 사망 위험을 초래하는 어려운 상태입니다. 그러나 패혈증 후 PICS 환자의 면역세포 특징(immune-cell signatures)은 아직 명확히 밝혀지지 않았습니다.

방법 패혈증 후 PICS의 면역세포 특징을 단일세포 수준에서 규명하였습니다. 검증을 위해 마우스 cecal ligation and puncture (CLP) 모델을 PICS 모델로 적용하였습니다.

결과 패혈증 환자에서 두 개의 풍부한 단핵구 아집단(Mono1 및 Mono4)의 면역 기능이 현저히 억제되었으며, 패혈증 후 PICS 단계에서는 부분적으로 회복되었습니다. 이들 단핵구는 B세포와 CD8 T세포에 대해 면역억제 및 세포사멸 촉진 효과를 나타냈습니다. 패혈증 및 PICS 환자 모두에서 초기(naive) 및 기억 B세포가 감소하고, 플라스마세포 증식이 줄었습니다. 그러나 PICS 환자의 초기·기억 B세포는 패혈증 환자에 비해 항원 처리·제시(antigen processing and presentation) 관련 유전자 발현이 활성화된 특징을 보였습니다. 예후가 좋은 PICS 환자군에서는 활성 기억 B세포와 IGHA1-플라스마세포가 더 많았습니다. PICS 사망군에서는 CD8 TEMRA 세포가 증식과 동시에 면역 기능 장애 징후를 보인 반면, 생존군에서는 그렇지 않았습니다. 거핵세포의 증식은 건강 대조군에 비해 패혈증 및 PICS 환자에서 더 두드러졌으며, 특히 PICS에서 항염증·면역조절 효과의 뚜렷한 변화가 관찰되었고 이는 마우스 모델에서도 확인되었습니다.

결론 본 연구는 패혈증 후 PICS를 단일세포 수준에서 평가하여 면역세포 아집단 내 이질성을 규명하였으며, 이는 질병 진행 예측과 효과적인 중재 개발에 기여할 수 있습니다.

Graphical abstract

• 패혈증 급성기 → 단핵구(Mono1/4) 강한 억제, B세포 감소, 플라스마세포 증가
• PICS로 진행 → 단핵구 부분 회복, B세포 기능 일부 활성화(특히 antigen presentation), 하지만 전체 면역 억제 지속
• 예후 좋은 PICS-alive → 기억 B세포 + IGHA1-plasma cell 활성 ↑, CD8TEMRA 기능 유지
• 예후 나쁜 PICS-death → CD8TEMRA 기능 장애 + MK 기능 저하
• 거핵세포(MK)의 새로운 역할 강조: 단순 혈소판 만드는 세포가 아니라, PICS에서 항염증·면역조절 기능으로 보호 역할 가능 (마우스에서도 확인)

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

서론 (Introduction)

패혈증(sepsis)은

전 세계적으로 수백만 명의 환자에게 영향을 미치는 공중보건 비상사태이며,

주요 사망 원인 중 하나입니다.

일부 중증 환자들은 급성 위기를 극복하고 생존하지만,

약 40%는 완전히 회복되지 못하고 장기적인 생명 유지 치료가 필요하며,

이는 만성 중증 질환(chronic critical illness, CCI)으로 이어집니다.1,2,3

CCI 환자 중 30%–50%4는

장기 입원, 지속적인 염증, 관리 가능한 장기 부전, 단백질 이화(protein catabolism)를 경험하게 되며,

이는 상처 치유 지연, 체중 감소, 면역억제를 초래합니다.

이러한 상태를

지속성 염증·면역억제·이화증후군

(persistent inflammation, immunosuppression, and catabolism syndrome, PICS)이라고 합니다.1,4,5,6

PICS 가설은

2012년 이후 성인에서 검증되었습니다.7,8

PICS는 사망의 독립적 위험 인자이며,9

특히 고령 환자에서 더욱 두드러집니다.10

PICS 환자의 중환자실 사망률은

PICS가 없는 환자에 비해 높습니다(24.2% vs. 12.3%).9

PICS의 불량 예후는

숙주 면역의 조절 이상(dysregulated host immunity)과 관련이 있습니다.1

그러나

PICS와 관련된 전신 면역 장애의 유전적·분자적 변화는

아직 잘 이해되지 않고 있습니다.11,12

면역 장애의 기전 이해가 제한적이기 때문에

PICS에 대한 특이적 치료법이 부족하며,13

높은 민감도와 특이도를 가진 바이오마커를 식별하기 위한

검증 연구가 여전히 필요합니다.6

최근 단일세포 RNA 시퀀싱(single-cell RNA sequencing, scRNA-seq)과 같은

첨단 기술의 발전으로 패혈증에서의 면역 반응이 점차 규명되고 있습니다.14,15,16

패혈증 환자의

말초혈액 단핵구 세포(peripheral blood mononuclear cells, PBMCs)에 대한

세포 유형과 분자 프로파일의 전경(panoramic picture)이 밝혀지면서,14

단일세포 유전체학을 이용해

세균성 감염 관련 패혈증에서 질병 관련 세포학적 특징을 발견하고

면역 조절 이상의 세포적 기초를 이해하는 데 통찰을 제공하였습니다.15

이러한 연구들은

주로 패혈증 환자의 단핵구(monocyte) 특성 변화에 초점을 맞추었습니다.

그러나

PICS와 패혈증에서 단핵구, B세포, T세포, 거핵세포(megakaryocytes, MKs) 등의

면역세포 구성과 아집단 간 상호작용을 단일세포 전사체 수준에서

포괄적으로 이해하는 것은 여전히 제한적입니다.

따라서

본 연구는

패혈증 후 PICS 발병에 관여하는 면역세포 특징(immune-cell signatures)을 조사하는 것을 목적으로 하였습니다.

이를 위해 고령

성인 참가자(PICS 환자, 급성 패혈증 환자, 건강 대조군)를 대상으로

말초혈액 단핵구 세포(PBMCs)를 단일세포 전사체 분석(single-cell transcriptomic analysis)으로

체계적으로 프로파일링하였습니다.

본 연구 결과는

패혈증 후 PICS 환자에서 면역세포 재프로그래밍(immune-cell reprogramming)에 대한

새로운 통찰을 제공합니다.

Results

scRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

결과 (Results)

PICS와 패혈증에서의 면역 landscape에 대한 scRNA-seq 프로파일링

(scRNA-seq profiling of the immune landscape in PICS and sepsis)

급성 패혈증 및 PICS 환자의 면역 세포 다양성과 분자적 특징을 탐색하기 위해,

패혈증에 이차적인 PICS 환자 6명,

급성 패혈증 환자 5명,

건강 대조군(healthy controls, HCs) 5명의 PBMCs를 분리하여

scRNA-seq으로 프로파일링하였습니다(Figure 1A).

임상 결과와 관련된 하위군 분석을 용이하게 하기 위해

PICS 환자 6명을 두 개의 하위군으로 분류하였습니다:

PICS 사망군(PICS-death group, n=2; PD-01 및 PD-02)과

PICS 생존군(PICS-alive group, n=4; PA-01-04).

총 16명의 참가자로부터

91,180개의 고품질 세포(high-quality cells)가 프로파일링되었습니다.

모든 등록 참가자의 임상 특성과 scRNA-seq 수량은

Table S1에 요약되어 있습니다.

총 23개의 세포 클러스터가 식별되었으며(Figures 1B 및 S1),

주요 배치 효과(major batch effects)는 관찰되지 않았습니다(Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

6Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai 200032, China

7Department of Anesthesiology, Zhongshan Hospital, Institute for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, MOE Innovative Center for New Drug Development of Immune Inflammatory Diseases, Fudan University, Shanghai 200032, China

8Laboratory of Perioperative Stress and Protection, Shanghai 200032, China

9

These authors contributed equally

10

Lead contact

Article Info

Publication History:

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

•  
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

SummaryBackground

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

Graphical abstract

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

ResultsscRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

6Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai 200032, China

7Department of Anesthesiology, Zhongshan Hospital, Institute for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, MOE Innovative Center for New Drug Development of Immune Inflammatory Diseases, Fudan University, Shanghai 200032, China

8Laboratory of Perioperative Stress and Protection, Shanghai 200032, China

9

These authors contributed equally

10

Lead contact

Article Info

Publication History:

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

•  
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

SummaryBackground

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

Graphical abstract

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

ResultsscRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

6Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai 200032, China

7Department of Anesthesiology, Zhongshan Hospital, Institute for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, MOE Innovative Center for New Drug Development of Immune Inflammatory Diseases, Fudan University, Shanghai 200032, China

8Laboratory of Perioperative Stress and Protection, Shanghai 200032, China

9

These authors contributed equally

10

Lead contact

Article Info

Publication History:

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

•  
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

SummaryBackground

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

Graphical abstract

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

ResultsscRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

6Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai 200032, China

7Department of Anesthesiology, Zhongshan Hospital, Institute for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, MOE Innovative Center for New Drug Development of Immune Inflammatory Diseases, Fudan University, Shanghai 200032, China

8Laboratory of Perioperative Stress and Protection, Shanghai 200032, China

9

These authors contributed equally

10

Lead contact

Article Info

Publication History:

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

•  
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

SummaryBackground

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

Graphical abstract

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

ResultsscRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

6Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai 200032, China

7Department of Anesthesiology, Zhongshan Hospital, Institute for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, MOE Innovative Center for New Drug Development of Immune Inflammatory Diseases, Fudan University, Shanghai 200032, China

8Laboratory of Perioperative Stress and Protection, Shanghai 200032, China

9

These authors contributed equally

10

Lead contact

Article Info

Publication History:

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

•  
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

SummaryBackground

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

Graphical abstract

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

ResultsscRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

ArticleVolume 6, Issue 5100569May 09, 2025Open access

Download Full Issue

Immune-cell signatures of persistent inflammation, immunosuppression, and catabolism syndrome after sepsis

Xing-Feng Sun1,2,9 ∙ Wen-Chen Luo1,9 ∙ Shao-Qiang Huang2,9 ∙ … ∙ Zhi-Xin Qiu1,7 qiuzhixin15@163.com ∙ Jing Zhong1,10 jzhong12@fudan.edu.cn ∙ Chang-Hong Miao1,8 ziteng1934@aliyun.com … Show more

Affiliations & Notes

1Department of Anesthesiology, Zhongshan Hospital Fudan University, Shanghai 200032, China

2Department of Anesthesiology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai 200438, China

3Department of Critical Care and Pain Medicine, Fudan University Shanghai Cancer Center, Shanghai 200032, China

4The State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, and the Institutes of Brain Science, Fudan University, Shanghai 200032, China

5Shanghai Key Laboratory of Medical Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China

6Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai 200032, China

7Department of Anesthesiology, Zhongshan Hospital, Institute for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, MOE Innovative Center for New Drug Development of Immune Inflammatory Diseases, Fudan University, Shanghai 200032, China

8Laboratory of Perioperative Stress and Protection, Shanghai 200032, China

9

These authors contributed equally

10

Lead contact

Article Info

Publication History:

Received May 7, 2024; Revised October 13, 2024; Accepted December 12, 2024; Published online January 16, 2025

DOI: 10.1016/j.medj.2024.12.003 External LinkAlso available on ScienceDirect External Link

Copyright: © 2024 The Author(s). Published by Elsevier Inc.

User License: Creative Commons Attribution (CC BY 4.0) | Elsevier's open access license policy

Published: January 16, 2025

•  
• Cite
•  
• Set Alert
• Get Rights
• Reprints

Download Full Issue

Previous articleNext article

Show Outline

Context and significance

Persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in intensive care units. However, the immune landscape of PICS is poorly understood. Here, the authors performed single-cell resolution profiling to dissect immune-cell alterations in the peripheral blood of patients with sepsis and PICS. They described the specific immune-cell clusters contributing to the immune dysfunction in PICS, highlighting the characteristics of different immune-cell subsets like monocyte reminiscent of myeloid-derived suppressor cell, B cell, regulatory T cell, and megakaryocyte, and identified a specific immune profile associated with prognosis of patients with PICS. These findings provide new insights into the immune heterogeneity in PICS after sepsis and may guide the development of future therapies for these patients.

Highlights

•

Single-cell profiling reveals immune-cell reprogramming in peripheral blood of PICS

•

Monocyte subsets are suppressed in patients with sepsis and partially restored in PICS

•

PICS patients with better prognoses have more active memory B and IGHA1-plasma cells

•

Megakaryocytes exhibit anti-inflammatory and immunomodulatory effects in PICS

SummaryBackground

Management of persistent inflammation, immunosuppression, and catabolism syndrome (PICS) after sepsis remains challenging for patients in the intensive care unit, experiencing poor quality of life and death. However, immune-cell signatures in patients with PICS after sepsis remain unclear.

Methods

We determined immune-cell signatures of PICS after sepsis at single-cell resolution. Murine cecal ligation and puncture models of PICS were applied for validation.

Findings

Immune functions of two enriched monocyte subpopulations, Mono1 and Mono4, were suppressed substantially in patients with sepsis and were partially restored in patients with PICS after sepsis and exhibited immunosuppressive and pro-apoptotic effects on B and CD8T cells. Patients with PICS and sepsis had reduced naive and memory B cells and proliferated plasma cells. Besides, naive and memory B cells in patients with PICS showed an active antigen processing and presentation gene signature compared to those with sepsis. PICS patients with better prognoses exhibited more active memory B cells and IGHA1-plasma cells. CD8TEMRA displayed signs of proliferation and immune dysfunction in the PICS-death group in contrast with the PICS-alive group. Megakaryocytes proliferation was more pronounced in patients with PICS and sepsis than in healthy controls, with notable changes in the anti-inflammatory and immunomodulatory effects observed in patients with PICS and verified in mice models.

Conclusions

Our study eval‎uated PICS after sepsis at the single-cell level, identifying the heterogeneity present within immune-cell subsets, facilitating the prediction of disease progression and the development of effective intervention.

Funding

This work was supported by the National Natural Science Foundation of China, Shanghai Municipal Health Commission “Yiyuan New Star” Youth Medical Talent Cultivating Program, and Shanghai Clinical Research Center for Anesthesiology.

Graphical abstract

Keywords

1. immune function
2. sepsis
3. single-cell RNA sequencing
4. persistent inflammation, immunosuppression, and catabolism syndrome

CAT Scale

1. Translation to patients

Introduction

Sepsis is a global public health emergency that affects millions of patients and is a leading cause of death. Although certain critically ill patients survive the acute crisis, approximately 40% do not fully recover and require long-term life support, leading to chronic critical illness (CCI).1,2,3 In patients with CCI, 30%–50%4 may experience prolonged hospitalization, ongoing inflammation, manageable organ failure, and protein catabolism, leading to poor wound healing, weight loss, and immunosuppression, described as persistent inflammation, immunosuppression, and catabolism syndrome (PICS).1,4,5,6 The PICS hypothesis has been validated in adults since 2012.7,8 PICS is an independent risk factor for death,9 particularly in elderly patients.10 The intensive care unit (ICU) mortality rate of patients with PICS is higher than that of patients without PICS (24.2% vs. 12.3%).9

Poor prognosis of PICS is associated with dysregulated host immunity.1 However, the genetic and molecular changes of systemic immune disorders associated with PICS remain poorly understood.11,12 Specific treatments for PICS are lacking based on the limited understanding underlying immune disorders,13 and validation studies remain warranted to identify biomarkers with high sensitivity and specificity.6

Immune responses in sepsis have been characterized gradually following recent development of cutting-edge techniques, such as single-cell RNA sequencing (scRNA-seq).14,15,16 A panoramic picture of cell types and molecular profiles of peripheral blood mononuclear cells (PBMCs) from patients with sepsis14 allowed the discovery of disease-associated cytological features using single-cell genomics that provided insights into the cellular basis of immune dysregulation in bacterial tract infection-associated sepsis.15 Such studies have focused primarily on the changes in monocyte characteristics in patients with sepsis. However, a comprehensive understanding of immune-cell composition, such as monocytes, B cells, T cells, and megakaryocytes (MKs), and interactions among immune-cell subpopulations in PICS and sepsis at the single-cell transcriptome level, remains limited.

Therefore, this study aimed to investigate the immune-cell signatures involved in PICS pathogenesis after sepsis. To this end, we systematically profiled PBMCs of elderly adult participants, including patients with PICS, patients with acute sepsis, and healthy individuals, through single-cell transcriptomic analysis. Our results provide novel insights into immune-cell reprogramming in patients with PICS following sepsis.

ResultsscRNA-seq profiling of the immune landscape in PICS and sepsis

To explore immune cellular diversity and molecular signatures in patients with acute sepsis and PICS, PBMCs from six patients with PICS secondary to sepsis, five with acute sepsis, and five healthy controls (HCs) were isolated and profiled by scRNA-seq (Figure 1A). To facilitate subgroup analysis associated with clinical outcome, the six patients with PICS were further classified into two subgroups: the PICS-death group (n = 2; PD-01 and PD-02) and the PICS-alive group (n = 4; PA-01-04). A total of 91,180 high-quality cells were profiled from 16 participants. The clinical characteristics and scRNA-seq quantities of all enrolled participants are summarized in Table S1. A total of 23 cell clusters were identified (Figures 1B and S1), and no major batch effects were observed (Figure 1C).

Figure 1 Study design and the single-cell transcriptomic atlas of PBMCs obtained from study subjects

Show full captionFigure viewer

Based on the expression‎ of canonical lineage markers, cells in different clusters were assigned to eight distinct cell types (Figures 1B and 1D): B cells (CD79A and CD19), dendritic cells (DCs; CLEC10A and CD1C), MKs (PF4 and PPBP), monocytes (VCAN, FCN1, CD14 and LYZ), natural killer cells (NK cells; GNLY and KLRF1), plasmacytoid dendritic cells (pDCs; SOX4 and CLEC4C), plasma cells (CD79A, JCHAIN, and IGHA1), and T cells (CD3D and CD3E). The distribution of cell types exhibited notable variation (Figure 1E and Table S2), indicating substantial cell heterogeneity among patients with PICS, sepsis, and HCs. Additionally, the proportions of monocytes, T cells, B cells, plasma cells, and MKs differed between patients with PICS and sepsis, suggesting a marked shift in peripheral immune response following septic insults and the potential for major immune-cell types to distinguish endotypes among patients with PICS.

Classifications and annotations of monocyte clusters in patients with PICS and sepsis

We classified monocytes according to the marker genes, VCAN, FCN1, CD14, and LYZ,17,18 as shown in Figure 1D. The percentages of monocytes in the PICS, sepsis, and HC groups were 22.5%, 28.3%, and 10.3%, respectively (Figure 1E). We re-clustered monocytes and identified six distinct clusters, Mono1 to 6 (Figure 2A). Based on the expression‎ of marker genes, we elucidated the possible contribution of the six monocyte subtypes to disease progression (Figure 2B). Substantial enrichment of Mono1 and Mono4 was observed in patients with PICS and sepsis (Figure 2C, left and upper right). Both subsets exhibited high expression‎ of S100A8 and S100A9, along with low expression‎ of HLA-DR, and showed obvious activation of CLU, which was related to worse clinical outcomes in patients with sepsis.19 Mono2-expressing CD14 with low or no expression‎ of FCGR3A did not proliferate notably in patients with PICS and sepsis, corresponding to classical monocytes. Mono6 and Mono3 expressed high levels of HLA-DR, which closely resembled CD14+HLA-DRhigh inflammatory monocytes. Mono1 and Mono5 showed marked reductions in patients with PICS compared with those with sepsis, manifesting divergent tendencies compared with other subsets. Mono5 was enriched with immune-related genes, as evidenced by high levels of FCGR3A and MS4A7 expression‎ associated with the macrophage phenotype. Subgroup analysis showed substantial decreases in the proportions of Mono1 to 6 in the PICS-death group than in the PICS-alive group (Figure 2C lower right; Table S3), which may aid in predicting sepsis progression and clinical prognosis.

Figure 2 Characterization of the monocyte subsets

Show full captionFigure viewer

The expansion of myeloid-derived suppressor cells (MDSCs) has been hypothesized as a central mechanism in CCI and PICS.6 In patients with sepsis, an expanded CD14+ monocyte state MS1, which is reminiscent of MDSCs,15,20 was identified as conserved HLA-DRlowS100Ahigh monocyte subset with immunosuppressive function.21 These monocytes exhibited lower expression‎ of HLA-DR,2,22 typical of MDSCs. From our findings, Mono1 and Mono4 expressed a CD14+, S100A8/9+, HLA-DRB1−, HLA-DMA−, and HLA-DQB1− signature. Subsequently, Gene Ontology (GO) enrichment and gene set enrichment analyses (GSEA) of the cells were performed. As shown in Figure 2D, the GO terms including inflammatory response, defense response to bacterium, positive regulation of reactive oxygen species metabolic process, toll-like receptor signaling pathway, and items written in green, which are related to cell growth and development, were predominantly enriched in Mono1 and Mono4. GSEA showed that the biological functions enriched in Mono1 and Mono4 were diverse. Cytokine-cytokine receptor interaction activation was decreased significantly in Mono1 and Mono4 subsets in patients with sepsis compared with in patients with PICS or in HCs (Figure S2A). Similarly, single sample GSEA (ssGSEA) showed that the activation of this pathway in Mono1 and Mono4 in PICS was higher than that occurring in sepsis (Figure 2E). The T cell receptor signaling pathway in Mono1, B cell receptor signaling pathway, and NK cell-mediated cytotoxicity in Mono4 were downregulated in the PICS-death group, indicating impaired activation of immune pathways in patients with PICS with poor outcome (Figures S2B and S2C). A comparison of all the monocyte sub-clusters showed that Mono1 and Mono4 exhibited the dysfunctional phenotype of monocytes, supported by the reduced enrichment of items written in blue related to immune function, suggesting immunosuppressive activity (Figure 2D). Collectively, based on gene expression‎, GO, GSEA, and ssGSEA, Mono1, and Mono4 are considered to be reminiscent of MDSCs,20,21,23 and a monocytic (M-MDSC) subtype.24 Furthermore, we analyzed the reported expression‎ of MDSC functional genes25 in monocytes and observed that the expression‎ of RNASE2 (antivirus and immunomodulation),26 S100A10, and VCAN (cell adhesion and proliferation)27 increased in Mono1 and Mono4 in patients with PICS and sepsis, whereas HLA-DRA and HLA-DPA1 expression‎ decreased (Figure S2D). We then scored the related gene expressions of MS1 (an expanded CD14+ monocyte state, MDSC reported by Reyes et al.),15,20 and observed that the scores of Mono1 and Mono4 in patients with PICS and sepsis in our study were higher than that observed in HCs (Figure S2E). The above results further identified the MDSC characteristics of Mono 1 and Mono 4 in PICS and sepsis in our study.

By comparing up- and downregulated differentially expressed genes (DEGs) in patients with PICS and sepsis vs. HCs, Mono1 and Mono4 subsets in patients with sepsis exhibited elevated levels of genes associated with worse clinical outcomes, including PLAC8, RETN, CLU,19 and CD16328 compared with HCs, whereas decreased levels of CD163 were observed in patients with PICS than those with sepsis (Figure 2F). The expression‎ of PLAC8 in Mono4 exhibited a similar trend (Figure 2F). Our results suggested partial recovery of the immune function of Mono1 and Mono4 in patients PICS compared with patients with sepsis. Significantly, the expression‎ levels of inflammatory regulation-related genes (such as S100A8, S100A9, and S100A12) was upregulated in Mono1 and Mono4 subsets of patients with PICS, whereas the HLA family was downregulated compared with HCs. These changes aligned with earlier studies15,29 and are corroborated by research indicating that diminished HLA-DR expression‎ correlates with monocyte immaturity, leading to reduced responsiveness of monocytes to stimuli30,31 and elevated susceptibility to nosocomial infection and mortality in patients with sepsis.32,33 Moreover, compared with those in patients in the PICS-death group, monocytes from patients in the PICS-alive group exhibited elevated levels of TMEM176 A/B (Figure S2F), which is believed to play a role in antigen presentation34 and regulation of the inflammasome.35

Furthermore, trajectory analysis was conducted on monocyte scRNA-seq data using Monocle 236 to elucidate potential relationships among the six monocyte sub-clusters. Pseudotime ordering of all cells revealed three distinct states arranged along a single main branch. Notably, Mono1 and Mono4 were predominant in the early and medium phases, consistent with M-MDSC states that dominated in the early and medium disease stage in a previous study,37 whereas Mono 3 and Mono5 showed significant enrichment in later stages (Figures 2G and 2H).

In summary, the immune functions of Mono1 and Mono4 (M-MDSC) were notably suppressed in patients with sepsis. Upon transitioning to the PICS phase, these cell subsets exhibited partial restoration of immune functions, while remaining inferior to those of HCs. The reduction in monocyte proportions, coupled with pronounced immunoparalysis, is a potential indicator of unfavorable clinical prognosis in patients with PICS.

B cells and plasma cells were dysregulated in patients with PICS and sepsis

We performed re-clustering and divided B cells into the following five sub-clusters: naive B, memory B, IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells (Figure 3A).38,39 Naive B cells, expressing high levels of CD79A, TCL1A, and CD184, were involved in regulating the immune effector process, B cell proliferation, and lymphocytes. Memory B cells associated with B cell differentiation were identified based on the high expression‎ of CD79A, CD27, and AIM2. CD38, CD27, and JCHAIN were the marker genes used to identify plasma cells, whereas IGHA1-plasma cells, IGHA2-plasma cells, and IGHG1-plasma cells highly expressed IGHA1, IGHA2, and IGHG1, respectively (Figure 3B). A wide range of molecular functions was observed commonly in B cells, including positive immune regulation (CD82 and CD79b), negative immune regulation (CD23, CD32, and CD305), and antigen presentation (CD1c, CD40, and HLA-DRA). We plotted molecules and organized them by function (labeled by colors) (Figure S3A). Furthermore, the upregulated genes in naive B and memory B cells, including several pro-inflammatory genes (IFITM1, IFITM2, IFITM3, and AIM2) (Figure S3B), are critical for inflammation and activation of the innate immune response,40,41 indicating that naive B and memory B cells can upregulate pro-inflammatory genes and participate in the innate immune response.

Figure 3 Functional changes in B cells in different disease states

Show full captionFigure viewer

During sepsis, naive B and memory B cells were reported to exhibit increased apoptosis and depletion, leading to a decrease in their proportion, which is associated with poor prognosis.16,42 In the present study, patients with PICS and sepsis exhibited decreased counts and proportions of naive B and memory B cells, alongside notable proliferation of plasma cells compared with HCs (Figure 3C left; Table S3). The increased proportion and activation of plasma cells were observed in the acute and recovery stages in patients with sepsis,16 reinforcing the essential roles of plasma cell proliferation in sepsis. Additionally, compared with sepsis, a decline in the counts and proportions of plasma cells (Figure 3C left) and lower numbers of the three plasma cell subsets based on the total observed/expected (O/E) ratio (Figure S3C and Table S4) were observed in patients with PICS. The above phenomenon may be attributed to the enhanced differentiation of B cells into plasma cells in patients with PICS and sepsis compared with HCs and less humoral immune activation was observed in patients with PICS compared with sepsis.

GO analysis revealed that naive B and three plasma subsets (IGHA1-plasma, IGHA2-plasma, and IGHG1-plasma cells) were involved in more immune processes than memory B cells (Figure 3D). Naive B cells responded to relevant antigens during the immune challenge processes like PICS and sepsis and were selected to proliferate and differentiate into effector and memory classes, as revealed by trajectory analysis (Figure S3D). In naive B cells, GSEA showed that the T and B cell receptor signaling, chemokine signaling, and toll-like receptor signaling pathway-related genes were downregulated in patients with PICS and sepsis relative to HCs (Figure S3E). Additionally, antigen processing and presentation-related genes were downregulated in naive B and memory B cells in patients with PICS and sepsis compared with HCs. In comparison, the naive B and memory B cells exhibited an enriched gene signature related to antigen processing and presentation in patients with PICS than in patients with sepsis (Figure S3F). Consequently, the immune function of naive B and memory B cells was suppressed in patients with PICS and sepsis but was enhanced notably in the antigen processing and presentation pathways in patients with PICS compared with those with sepsis.

Compared with the PICS-death group, the PICS-alive group showed upregulated inflammatory mediators S100A8/S100A9 in all B cell subsets (Figure 3E) and genes related to immune function IGKV2-24, IGKV7-46, and IGKV2-8 upregulated significantly in memory B cells, IGHV3-33, IGKV4-1, and IGLVI-70 increased in IGHA1-plasma cells (Figure 3E). Altogether, the results indicated that patients with PICS with favorable prognosis exhibited more activated memory B and IGHA1-plasma subsets, and the ability of the IGHA1-plasma subset to secrete antibodies was enhanced in patients in the PICS-alive group. A recent study has shown that immunoglobulin production is positively correlated with the proportion of plasma cells in patients with sepsis during recovery.16 Such differences in plasma cells between the PICS-alive and PICS-death groups in this study linked the activation of B cells and the function of plasma cells with the PICS prognosis. However, the IGHG1-plasma subset exhibited a greater capacity for antibody production (upregulation of IGKV1-8, IGHG3, IGHV4-31, IGKV1D-13, and IGHG1) (Figure 3E), indicating the increased ability of IGHG1 subset to secrete immunoglobulin (Ig)G antibodies in the PICS-death group.

Collectively, the proportions of naive B and memory B cells decreased in patients with PICS and sepsis compared with HCs, and the proportion of plasma cells was the highest in patients with sepsis and the lowest in HCs. The naive B and memory B subsets had suppressed immune function in patients with PICS and sepsis compared with HCs and exhibited an active gene signature for antigen processing and presentation in patients with PICS compared with sepsis. Moreover, PICS patients with better prognosis showed more active memory B and IGHA1-plasma cells.

Dysfunctions of CD8TEMRA and regulatory T cells in the PICS-death group

T cells were the major cell type present in PBMCs in all participants and tended to decrease in patients with sepsis and PICS compared with HCs (46.25%, 52.83%, and 59.76%, respectively, Figure 1E). When we divided the T cells in PBMC into 10 sub-clusters, including four distinct CD4T cell and six CD8T cell sub-clusters annotated using specific gene markers (Figure 4A and Table S5), the O/E ratio showed that regulatory T cells (Tregs) did not demonstrate more depletion in patients with PICS than in patients with sepsis, and CD8TEMRA was more enriched in patients with sepsis than in patients with PICS (Figure S4A and Table S4). Subgroup analysis showed that the PICS-death group exhibited notable increases in the proportions of CD8TEMRA cells and a decrease in Tregs compared with the PICS-alive group, suggesting a marked heterogeneity in cellular immune infiltration in patients who died from PICS (Figure 4B).

Figure 4 Dysfunction of CD8TEMRA and Tregs in PICS-death patients

Show full captionFigure viewer

GO analysis revealed that genes related to NK and T cell activation, differentiation, and T cell secretory granule organization were upregulated in CD8TEMRA cells in patients in the PICS-death group (Figure 4C). Pathways associated with apoptosis and granzyme-mediated apoptosis were markedly activated. CD8TEMRA cells from patients in the PICS-death group had upregulation of genes associated with antigen processing and presentation, but downregulation of genes associated with cytokine-cytokine receptor interaction, according to GSEA (Figure S4B). The findings implied a dual role of CD8TEMRA cells in immune activation and immunosuppression in the PICS-death group.

Treg levels can be used to assess the course of sepsis.43 Expression‎ of CD69 (a marker of T cell activation)44,45 was elevated in the PICS-death group compared with the PICS-alive group and HC, but the level was similar to that in patients with sepsis, suggesting that the Treg immune response was activated in patients of the PICS-death group and patients with sepsis. The expression‎ levels of DUSP1 and FOS were significantly higher in patients of the PICS-death group. DUSP1 and FOS are related to the inhibition of cell proliferation and promoting apoptosis, respectively.46,47 These data indicate that Tregs in patients of the PICS-death group are activated in a sense but exhibit suppressed proliferation and increased apoptosis, resulting in an unfavorable prognosis. Meanwhile, genes associated with immune regulation (including HLA-DRB5, S100A8, IGKV3, and JCHAIN) were downregulated in patients in the PICS-death group, indicating a potential suppression of Treg function (Figure 4D).

In sepsis, immune suppression is characterized mainly by exhaustion and apoptosis of lymphocytes, particularly T cells.16,48,49 To provide additional insight into the subpopulations of depleted T cells, we analyzed the expression‎ of T cell exhaustion-associated markers, such as HAVCR2, LAG3, PDCD1, and TIGIT,50 and showed a significant upregulation of LAG3 expression‎ in CD8 T cell subtypes of patients with PICS, particularly in the PICS-death group (Figure 4E). TIGIT expression‎ was significantly elevated in Tregs across all participants. Moreover, pro-apoptotic genes (including TNFRSF14 and CD27) were upregulated in all T cell subtypes, particularly in the PICS-death group (Figure 4E).

Participation of MKs in inflammatory and immunomodulatory responses

Emerging evidence shows the participation of MKs in a range of biological processes, such as coagulation, hemostasis, inflammation, angiogenesis, and innate immunity.51,52,53 Transcriptional studies on human blood MKs from patients with PICS and sepsis are limited. In this study, 1379 MKs were collected and stratified into five clusters, termed MK1 to MK5, with different gene expression‎ patterns (Figures 5A and 5B). MK1 and MK2 accounted for most MKs (77.5%). In comparison to HCs, MKs were enriched in all diseased patients, particularly in the PICS-death subgroup (Figure 5C and Table S3).

Figure 5 Identification and characterization of MK subpopulations and their functions

Show full captionFigure viewer

GO analysis showed that MKs had potential functional heterogeneity (Figure 5D). MK1 showed strong enrichment of immune-associated gene sets, including neutrophil degranulation and antigen processing and presentation, suggesting that the MK1 subpopulation is associated with immune regulation. Notably, MK1 constituted approximately 46.5% of total MKs from patients, which was greater than the 7.1% reported in human bone marrow in a non-infectious state,52 suggesting that MK progenitor cells are likely to differentiate into immune subset MK1 in patients with PICS and sepsis. The MK2 subpopulation showed several enriched gene sets, such as platelet activation, blood coagulation, and protein transport, and was enriched by certain genes related to platelet function, including GP6, and genes important for thrombopoiesis, such as TUBB1 (Figure 5B), GATA1, MYLK, NFE2, and MEIS1 (Figure S5A), suggesting that MK2 is mainly composed of cells with strong platelet-producing ability and thrombopoiesis. MK3 and MK4 were enriched in translation, SRP-dependent cotranslational protein targeting to membrane, cytoplasmic translation, and nuclear-transcribed mRNA catabolic processes, indicating that they may represent a less-mature MK subpopulation. Pseudotime analysis of MKs indicated the presence of a single branch and two primary directions at the end of the developmental trajectory. Specifically, MK2 was mainly present along the “direction 1” branch, whereas MK3 and MK4 were concentrated mainly along the “direction 2” branch (Figure 5E). The results imply that the various subpopulations of MKs might be generated along distinct developmental routes. MK5 exhibited an enrichment of RNA splicing, mRNA processing, regulation of RNA splicing, and apoptotic process (Figure 5D); meanwhile, MK5 was mainly present along the “direction 1” and “direction 2” branches, indicating that these cells may exert the function of niche support in the blood.

We further compared MK expression‎ patterns from different disease states to explore the alterations in MK functions within distinct pathological conditions. The GO term and pathway analysis revealed that in the PICS-alive group, innate immune pathways like neutrophil aggregation and chemokine production were notably activated. In contrast, the PICS-death group showed significant activation of adaptive immune pathways, including CD8T cell activation, NK cell-mediated immunity, antibody-dependent cellular cytotoxicity, and antigen processing and presentation (Figures 5F and S5B). MKs overexpressing IFITM2 and IFITM3 can resist viral infections.54,55 S100A9 is a critical modulator of inflammation.56 In our study, IFITM2, IFITM3, and S100A9 were upregulated in patients with PICS and patients with sepsis (Figure 5G). The results indicated that the anti-infective function and inflammatory response regulation function of MKs were activated in PICS and sepsis.

Dynamics of communication patterns in different cell types

We explored intercellular communication between different cell populations to gain insight into the mutual regulation of human PBMCs. We characterized the intercellular ligand-receptor pairs of all cell types using cell-cell communication (including CellphoneDB and CellChat) to analyze intercellular communications from scRNA-seq data. MDSCs are the most discussed biological entities in immunology. Evidence supports a key role for MDSCs in suppressing adaptive immune cells, including B and T cells.25,57 Our CellphoneDB analysis showed that in PICS, sepsis, and HC, Mono1 and Mono4 (M-MDSC) exerted an inhibitory effect against naive B and memory B cells via TNFSF13 ≫ TNFRSF14,58 and that TNFSF10 ≫ RIPK1,59 TNFSF13 ≫ FAS,60 and TNFSF10 ≫ TNFRSF10A61 exerted a pro-apoptotic effect on the three plasma cell subpopulations. The observed pro-apoptotic effects were stronger in sepsis and weaker in PICS than in HC (Figure 6A upper). In addition, Mono1 and Mono4 significantly inhibited CD8 T cells more than CD4 T cells (Figure 6A middle and lower), exerted pro-apoptotic effects on CD8 T cell subsets mainly through TNFSF13 ≫ TNFRSF14,58 and inhibited CD8TEMRA via TGFB1 ≫ TGFBR362,63 in all three groups, most notably for PICS, and CD8TEMRA via TGFB1 ≫ TGFBR164 only in the PICS group. Furthermore, in the PICS group, Mono1 and Mono4 exerted inhibitory effects on three subpopulations of CD4T cells (CD4Tem, Treg, and CD4Tmix) via TGFB1 ≫ TGFBR1, TGFB1 ≫ TGFBR2,64 and TGFB1 ≫ TGFBR3 (Figure 6A middle and lower panels). The data suggested that there were active pro-apoptotic and inhibitory signals from Mono1 and Mono4 to B and T cell subsets in patients with PICS and sepsis. Together, potential cross-talks with the focus on the immunosuppressive effects of MDSCs on adaptive immunity can be abstracted from our data.

[홍민]

중요한 내용인거 같습니다