Gut-derived macrophages link intestinal damage to brain injury after cardiac arrest through TREM1 signaling

Abstract

Brain injury is the leading cause of death and disability in survivors of cardiac arrest, where neuroinflammation triggered by infiltrating macrophages plays a pivotal role. Here, we seek to elucidate the origin of macrophages infiltrating the brain and their mechanism of action after cardiac arrest/cardiopulmonary resuscitation (CA/CPR). Wild-type or photoconvertible Cd68-Cre:R26-LSL-KikGR mice were subjected to 10-min CA/CPR, and the migration of gut-derived macrophages into brain was assessed. Transcriptome sequencing was performed to identify the key proinflammatory signal of macrophages infiltrating the brain, triggering receptor expressed on myeloid cells 1 (TREM1). Upon drug intervention, the effects of TREM1 on post-CA/CPR brain injury were further evaluated. 16S rRNA sequencing was used to detect gut dysbiosis after CA/CPR. Through photoconversion experiments, we found that small intestine-derived macrophages infiltrated the brain and played a crucial role in triggering secondary brain injury after CA/CPR. The infiltrating peripheral macrophages showed upregulated TREM1 levels, and we further revealed the crucial role of gut-derived TREM1⁺ macrophages in post-CA/CPR brain injury through a drug intervention targeting TREM1. Moreover, a close correlation between upregulated TREM1 expression and poor neurological outcomes was observed in CA survivors. Mechanistically, CA/CPR caused a substantial expansion of Enterobacter at the early stage, which ignited intestinal TREM1 signaling via the activation of Toll-like receptor 4 on macrophages through the release of lipopolysaccharide. Our findings reveal essential crosstalk between the gut and brain after CA/CPR and underscore the potential of targeting TREM1⁺ small intestine-derived macrophages as a novel therapeutic strategy for mitigating post-CA/CPR brain injury.

Introduction

Brain injury after cardiac arrest (CA) remains a substantial cause of morbidity and mortality globally [1, 2]. The pathophysiology of post-CA brain injury includes the primary injury resulting from the immediate cessation of cerebral blood flow during CA and secondary injury that occurs hours to days after resuscitation [3]. Excess neuroinflammation plays an essential role in secondary brain injury after CA, causing cumulative neuronal death and poor neurological outcomes [4]. Although brain-resident microglia are early generators of neuroinflammation and potent effectors of tissue damage, the continuous inflammatory state, which is consistent with cumulative brain injury, suggests the participation of peripheral immune cells [5]. We and other researchers have reported that a large number of peripheral macrophages infiltrates the brain after cardiac arrest/cardiopulmonary resuscitation (CA/CPR), amplifying neuroinflammation and aggravating brain injury [4, 6, 7]. However, the origin of these macrophages and their mechanism of action remain poorly understood.

The small intestine stores many macrophages and is highly vulnerable to ischemia/reperfusion injury [8, 9]. In addition, the intestinal macrophage pool is constantly replenished by circulating monocytes, especially under inflammatory or other challenging conditions [10, 11]. After ischemic stroke, intestinal macrophages migrate from the small intestine to the brain, thereby worsening cerebral inflammatory injury [12, 13]. Unlike intestinal injury after ischemic stroke, which is caused mainly by sympathetic vasoconstriction, CA/CPR can directly lead to intestinal ischemia/reperfusion (II/R) injury and even prolonged intestinal hypoperfusion in addition to the activation of sympathetic inputs [14]. Indeed, substantial small intestinal injury occurs in both CA/CPR model rats and clinical CA survivors and is associated with multiple organ dysfunction and mortality [15–17]. However, whether small intestine-derived macrophages infiltrate the brain and how they contribute to post-CA/CPR brain injury remain elusive.

In this study, we aimed to investigate the role of gut-derived macrophages in post-CA/CPR brain injury. Using photoconvertible mice, we observed that gut-derived macrophages migrated to the brain after CA/CPR or II/R alone. We excluded the interference of ischemia/reperfusion in other peripheral organs on brain injury using fluorescence-activated cell sorting (FACS) to identify macrophages infiltrating the brains of II/R model mice. Using transcriptome sequencing, we further identified triggering receptor expressed on myeloid cells 1 (TREM1), a robust proinflammatory molecule that is upregulated in infiltrating macrophages, as a potential executor linking intestinal and brain injury. Blocking TREM1 limited the trafficking of gut-derived macrophages to the brain and alleviated intestinal and brain injury in CA/CPR mice, whereas increasing TREM1 expression had the opposite effect. Mechanistically, we elucidated that CA/CPR caused the rapid expansion of Enterobacter, leading to the significant release of lipopolysaccharide (LPS) and unchecked activation of Toll-like receptor 4 (TLR4), which activated TREM1 signaling. Our findings have valuable clinical implications and suggest that TREM1 is a potential target for treating survivors of CA.

Methods

For a detailed description of the methods, please see the Supplemental Materials. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Participants and blood samples

The study protocol was approved by the Ethics Committee of Nanfang Hospital, Southern Medical University (approval number NFEC-202311-K34). Written informed consent was obtained from all participants for the anonymous use of their clinical information, whole blood samples, and follow-up results. From March 2022 to August 2022, we continuously recruited adult out-of-hospital CA patients who had been hospitalized at Nanfang Hospital (Southern Medical University, Guangzhou, China) if they presented with hypoxic-ischemic encephalopathy after successful CPR. Participants were excluded if they (1) were < 18 years old; (2) were transferred from other hospitals; (3) were assumed to have a septic or anaphylactic etiology; (4) were pregnant; (5) decided to limit life-sustaining therapy upon arrival; (6) used antidiarrheals, laxatives or prebiotics within one week, or antibiotics within one month; (7) had a medical history of chronic intestinal disease or gastrointestinal surgery; or (8) underwent cardiothoracic surgery, extracorporeal membranous oxygenation or used a ventricular assist device. Finally, 10 eligible patients were enrolled in the CA/CPR group. In addition, we recruited 8 healthy volunteers in the control group.

Blood samples were collected in EDTA-containing plasma tubes for subsequent peripheral blood mononuclear cell (PBMC) isolation. The TREM1 expression levels in monocytes/macrophages in the blood of the two groups were detected by flow cytometry, followed by a correlation analysis with multiple clinical data to evaluate the role of peripheral TREM1 in the neurological outcomes and systemic inflammation, including acute physiology and chronic health evaluation II (APACHE II) scores, Glasgow Coma Scale (GCS) scores, and the serum levels of C-reactive protein (CRP) and interleukin-6 (IL-6), in resuscitated patients.

Animals

All the animal experiments were approved by the Institutional Animal Care and Use Committee (approval number L2018080) and were compliant with the Animal Research: Reporting of In Vivo Experiments guidelines. Only males (8–10 weeks old) were used for the experiments. Wild-type C57BL/6 mice were purchased from the Experimental Animal Center of Southern Medical University. Cd68-2A-Cre mice were purchased from Shanghai Model Organisms Center, Inc. C57BL/6J-Gt(ROSA)26Sor^{em1 (CAG-LSL-KikGR-WPRE-pA) 1Smoc} (referred to as R26-LSL-KikGR) mice were also constructed by Shanghai Model Organisms Center, Inc., in which the CAG-LSL-KikGR-WPRE-pA cassette was knocked into the Gt(ROSA)26Sor (ENSMUSG00000086429, referred to as Rosa26) locus via CRISPR/Cas9. Cd68-Cre:R26-LSL-KikGR transgenic mice were constructed by crossing R26-LSL-KikGR mice with Cd68-2A-Cre mice, promoting the deletion of the stop cassette and consequent expression of the Kikume green–red photoconvertible fluorescent protein in CD68⁺ cells. Upon exposure to violet light (405 nm), the Kikume-green (KikG) fluorescent protein switches to Kikume-red (KikR) fluorescence, thereby allowing the visualization of the migration process of macrophages in vivo. All the mice were housed in a specific pathogen-free facility on a strict 12-hour light/dark cycle with free access to food and water and were acclimated to our facility for 1 week prior to the experiments. All efforts were made to minimize the number of animals used and their suffering in this study.

Mouse model of CA/CPR

The mice were anesthetized, followed by orotracheal intubation with a 22 G cannula (BD, Suzhou, China) and connection to a ventilator (RWD, Shenzhen, China). Normothermia was maintained with a heating pad during the surgery. For drug delivery, an intravascular catheter (Smiths Medical, Ashford, UK) was inserted into the internal jugular vein. Electrocardiograms (ECGs) were monitored with subcutaneous needle electrodes throughout the surgical procedure. After 5 min of stabilization, CA was induced by the intravenous administration of potassium chloride (50 µL, 0.5 M), which was defined as the appearance of isoelectric tracing on the ECG monitor. During CA, the ventilator was disconnected. At the end of the 10-min CA, CPR was initiated by adequate ventilation with 100% oxygen, intravenous administration of epinephrine (0.5 mL, 16 µg/mL), and chest compressions (300 compressions/min). The return of spontaneous circulation (ROSC) was defined as stable spontaneous electrical activity on an ECG. After spontaneous respiration recovered, the mice were weaned from the ventilator and extubated. Mice that failed to undergo ROSC within 3.5 min were excluded from the subsequent experiments. At the end of the experimental period, the mice were placed in cages with easy access to food and water. Mice that underwent all procedures except CA/CPR were used as a sham control.

Statistical analysis

All the data are presented as the means ± SDs or medians and 25th to 75th percentiles. SPSS 20.0 and GraphPad Prism 8.0 were used for the statistical analyses. P < 0.05 was considered statistically significant. Detailed descriptions of the statistical tests are available in the Supplemental Materials.

Results

Brain injury after CA/CPR is characterized by delayed exacerbation

Given the increasing clinical evidence that the peak of post-CA/CPR brain injury appears to be delayed in most cases [18, 19], we first explored whether post-CA/CPR brain injury in mice exhibited a similar dynamic pattern. The mortality of mice increased sharply at 72–96 h after CA/CPR compared with that at 24–48 h (Fig. 1A). Consistently, the neurological deficits were also markedly aggravated at 72 h after CA/CPR (Fig. 1B). Coronal T2-weighted (T2WI) images revealed high-signal lesions in the cortex and hippocampus at 24–48 h after CA/CPR, with further enlargement at 72 h (Fig. 1C, D). Immunostaining revealed that CA/CPR elicited neuronal and dendritic loss, neuronal degeneration, cellular apoptosis, and gliosis in the hippocampus and cortex, which occurred after 24 h and worsened at 72 h (Fig. 1E–L). Overall, these results imply that CA/CPR causes sustained brain damage with delayed exacerbation at approximately 72 h.

Fig. 1.

Brain injury after CA/CPR is characterized by delayed exacerbation. A Kaplan‒Meier analysis of cumulative survival during the 7-day follow-up period after CA/CPR. n = 23. B Neurological function scores of the surviving mice at 24 h, 48 h, and 72 h after CA/CPR. The scale of the neurological function score consists of 5 components, namely, consciousness, corneal reflex, respiration, coordination, and movement/activity, of which a total score of 10 is considered normal (details in the Supplemental Materials). C, D Coronal T2WI images showing lesions reflected by high signals at 24 h, 48 h, and 72 h after CA/CPR. E‒G Representative photomicrographs of neuropathological damage characterized by staining for NeuN, MAP-2, Iba-1, and GFAP in the hippocampal CA1 region of the sham and experimental groups at 24 h, 48 h, and 72 h after CA/CPR. H‒L Representative photomicrographs of FJC and TUNEL staining in the cerebral cortex and hippocampal CA1 region of the sham group and the experimental groups at 24 h, 48 h, and 72 h after CA/CPR. Scale bars, 100 μm or 200 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the sham group; ##P < 0.01 and ###P < 0.001 compared with 48 h. n = 5 or 11 per group

CA/CPR induces rapid and severe small intestinal injury

Considering that small intestinal injury might serve as a key trigger of local and even systemic inflammation, we further investigated damage to the small intestine after CA/CPR. We first calculated the intestinal wet-to-dry weight ratio to determine intestinal edema and found that the ratio at 48–72 h after CA/CPR was significantly higher than that in the sham group (Fig. 2A). Lactic acid is a product of anaerobic glucose metabolism, and an increased lactic acid level reflects reduced tissue perfusion [20]. Here, we verified that the level of lactic acid in the small intestinal mucosa was elevated at 6 h but then decreased gradually, suggesting that the small intestine underwent ischemia and reperfusion after CA/CPR (Fig. 2B). Moreover, CA/CPR induced histological injury of the small intestine, manifesting mainly as severe mucosal villus edema at 6–12 h (Fig. 2C). The damage to the small intestinal mucosa was gradually aggravated, resulting in many denuded and severed villi at 48–72 h after modeling. The Chiu’s scores paralleled the histological changes in the small intestine (Fig. 2D). Similarly, mice presented fewer cells expressing Ki67 and goblet cells secreting acidic mucins at 48 h after CA/CPR, indicating impaired epithelial cell proliferation and increased glandular damage (Fig. 2E, F).

Fig. 2.

CA/CPR induces rapid and severe small intestinal injury. A Graph reflecting the dynamic changes in the small intestinal water content after CA/CPR. B Quantification of lactic acid levels in intestinal mucosal tissues after CA/CPR. C, D H&E staining and corresponding Chiu’s scores indicating intestinal mucosal injury at different time points after CA/CPR. E, F Representative images of Alcian blue and Ki67 staining and quantification of the numbers of goblet cells secreting acid mucins and Ki67⁺ cells after CA/CPR. G, H Representative photomicrographs of intestinal barrier damage characterized by staining for Occludin, Claudin1, and ZO-1 at 12 h, 24 h, 48 h, and 72 h after CA/CPR. I The dynamic trends in the serum levels of the intestinal mechanical barrier damage indicators DAO and I-FABP after CA/CPR. J‒L Quantification of the GSH/GSSG ratio, SOD activity, and MDA content in intestinal mucosal tissues was performed to evaluate oxidative stress-induced damage after ROSC. Scale bar, 100 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the sham group. n = 5 per group

We then detected the intestinal barrier integrity and observed a gradual decrease in the expression of tight junction proteins in the intestinal epithelial layer and increased levels of diamine oxidase (DAO) and intestinal fatty acid binding protein (I-FABP) in the serum after CA/CPR, indicating that CA/CPR led to damage to the gut barrier, which peaked at 24–48 h (Fig. 2G–I). CA/CPR also significantly reduced the ratio of glutathione/oxidized glutathione (GSH/GSSG) and the activity of superoxide dismutase (SOD) and increased the level of malondialdehyde (MDA), a lipid peroxidation product, indicating that drastic oxidative stress-induced damage occurred in the small intestinal mucosa after CA/CPR and peaked at approximately 12 h (Fig. 2J–L). Considering the absence of notable pathological changes in other peripheral organs and the colon in the early stage after CA/CPR, as well as the lack of significant alterations in the colonic glandular structures and the expression of tight junction proteins in the colonic epithelia during this early post-CA/CPR period (Fig. S1A, B), these results indicate that the small intestine may be the most prematurely damaged tissue among the peripheral organs after CA/CPR, of which the peak of injury occurs earlier than that of brain injury.

Gut-derived macrophages infiltrate the brain and exacerbate brain injury after CA/CPR

Since the small intestine serves as a reservoir for macrophages [8, 9], we wondered whether post-CA/CPR intestinal injury activated macrophages and whether the latter linked intestinal damage to brain injury through their migration into the brain. We first examined the dynamic changes in the percentages of cerebral and small intestinal macrophages. The number of macrophages infiltrating the brain gradually increased, with a sharp increase at 72 h after CA/CPR, the same time point as the peak of brain injury (Fig. 3A–C, Fig. S1C, D). Moreover, we observed that the number of microglia also increased at 72 h, suggesting that the rapid increase in the number of infiltrating macrophages at 72 h facilitated the formation of a proinflammatory microenvironment and thus evoked the expansion of microglia. Notably, the macrophages in the small intestine also continuously accumulated after CA/CPR but their numbers were decreased dramatically before the peak of post-CA/CPR brain injury, suggesting the possible migration of many intestinal macrophages into the brain (Fig. 3D–F, Fig. S1E, F). We constructed Cd68-Cre:R26-LSL-KikGR transgenic mice and conducted in vivo photoconversion to label gut-derived macrophages as KikR⁺ cells and visually monitor whether small intestine-derived macrophages migrated to the brain (Fig. 3G–J, Fig. S1G). After testing the efficiency of photoconversion (Fig. 3G–H), we revealed that the number of KikR⁺CX3CR1⁺ macrophages was markedly increased in the brain but not other vital peripheral organs at 72 h after CA/CPR (Fig. 3I–J, Fig. S1H), indicating that many gut-derived macrophages tended to infiltrate the brain before the delayed exacerbation of brain injury. Since only approximately 50% of intestinal macrophages were labeled KikR⁺ cells after in vivo photoconversion (Fig. 3G), the observation of 30.8% of KikR⁺ infiltrating macrophages in the brain (Fig. 3I) implied that gut-derived macrophages might account for more than 60% of cerebral infiltrating macrophages.

Fig. 3.

Gut-derived macrophages infiltrate the brain and amplify neuroinflammation after CA/CPR. A, B Changes in the proportions of cerebral macrophages (CD11b⁺CD45^high) and microglia (CD11b⁺CD45^int) at different time points after CA/CPR. C Changes in the proportions of cerebral macrophages after excluding monocytes (Ly6C⁺), NK cells (NK1.1⁺), neutrophils (Ly6G⁺), dendritic cells (CD11c⁺), and eosinophils (SiglecF⁺) in the brain after CA/CPR. D‒F Changes in the proportions of intestinal macrophages at different time points after CA/CPR, especially those of macrophages identified with a more precise intestinal surface marker panel (Ly6C^-F4/80⁺CX3CR1⁺MHC-II⁺). G Representative flow cytometry plots showing gut-derived CX3CR1⁺KikR⁺ macrophages in the small intestine after in vivo photoconversion. H Immunostaining showing the conversion of KikG⁺CX3CR1⁺ small intestinal macrophages to KikR⁺CX3CR1⁺ macrophages after in vivo photoconversion. I Representative flow cytometry plots showing gut-derived CX3CR1⁺KikR⁺ macrophages in the brain after in vivo photoconversion. J Immunostaining showing the infiltration of gut-derived KikR⁺CX3CR1⁺ macrophages in the brain after CA/CPR. K, L Heatmaps reflecting the changes in the expression of the indicated genes in FACS-sorted macrophages infiltrating the brain after CA/CPR. Scale bars, 50 or 200 μm. *P < 0.05 and ***P < 0.001 compared with the sham group; ### P < 0.001 compared with 48 h. n = 5 per group

We analyzed the dynamic changes in peripheral blood macrophages to elucidate the migratory path of gut-derived macrophages into the brain or other organs and discovered that the population also constantly increased until 48 h after modeling (Fig. 4A, B). The significant increase in the number of KikR⁺ macrophages in the blood of post-CA/CPR mice at 24 h after photoconversion (24 h group) further suggested that the gut-derived macrophages might be transported into the brain through the blood circulation after CA/CPR (Fig. 4C). In addition, the proportion of KikR⁺ macrophages was comparable between the sham and 0 h (data from post-CA/CPR mice obtained immediately after photoconversion) groups, indicating that no obvious direct labeling of circulating immune cells occurred during photoconversion. As an essential reserve for immune cells [21], splenic macrophages also persistently accumulated but were interrupted earlier (Fig. 4D, E). Combined with the increase in the number of KikR⁺ macrophages in the spleen in the early phase after CA/CPR (Fig. 4F), we speculated that the migratory gut-derived macrophages activated the splenic macrophages and further amplified systemic inflammation with the help of their outward migration.

Fig. 4.

Dynamic changes in gut-derived macrophages in the spleen and blood after CA/CPR. A, B Changes in the proportion of macrophages in the blood at different time points after CA/CPR, especially macrophages expressing surface markers similar to those of gut macrophages (CX3CR1⁺MHC-II⁺) (B). C Representative flow cytometry plots showing gut-derived CX3CR1⁺KikR⁺ macrophages in the blood of the sham, 0 h (data from post-CA/CPR mice obtained immediately after photoconversion), and 24 h (data from post-CA/CPR mice obtained 24 h after photoconversion) groups. D Changes in the proportion of macrophages in the spleen at different time points after CA/CPR. E Changes in the proportion of splenic macrophages after excluding monocytes, NK cells, neutrophils, and eosinophils. (F) Representative flow cytometry plots showing gut-derived CX3CR1⁺KikR⁺ macrophages in the spleen at 24 h after CA/CPR modeling. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the sham group; ### P < 0.001 compared with 0 h or 48 h. n = 5 per group

Afterward, we sought to determine the role of gut-derived macrophages in post-CA/CPR brain injury. First, the mRNA levels of multiple proinflammatory cytokines and chemokines in cerebral infiltrating macrophages obtained via FACS were increased and peaked at 72 h after CA/CPR (Fig. 3K–L, Fig. S2A). We subsequently depleted peripheral macrophages by injecting chlodronate liposomes to explore the effects of peripheral macrophages, including gut-derived macrophages, on post-CA/CPR brain injury. After verifying the depletion efficiency (Fig. S2B), immunostaining revealed that the liposome treatment significantly reversed neuropathological damage, including excessive glial activation, neuronal loss, dendritic injury, and cellular apoptosis, at 72 h after ROSC (Fig. S2C–H). In summary, these findings imply that gut-derived macrophages infiltrate the brain after CA/CPR and may participate in exacerbating secondary brain injury.

II/R results in brain injury via the migration of gut-derived macrophages to the brain

Since the CA/CPR model causes ischemia/reperfusion in multiple organs, we constructed an II/R model to further verify the causal relationship between small intestinal ischemia/reperfusion and brain injury. The histological examination revealed that small intestinal injury was gradually aggravated and peaked at 24 h after modeling, as evidenced by the denuded mucosal villi and the reduction in the number of acid mucin-producing goblet cells (Fig. S3A–D). We subsequently analyzed whether small intestinal ischemia/reperfusion injury alone was sufficient to cause brain injury. The immunofluorescence results revealed that microglia and astrocytes were overactivated at 24 h after II/R (Fig. S3E–H). In addition, II/R led to substantial neuronal loss, dendritic injury, and neuronal degeneration, accompanied by many apoptotic cells in the cortex (Fig. S3I–N). These results indicate that small intestine ischemia/reperfusion injury alone can initiate brain damage.

We elucidated whether gut-derived macrophages likewise mediate brain damage after II/R by detecting dynamic changes in macrophages in the brain, small intestine, and peripheral blood. The number of infiltrating macrophages in the brain increased and peaked at 24 h after II/R, which was the same time as the peak of brain injury (Fig. S4A–B). In contrast, macrophages continuously accumulated in the small intestine or blood after II/R but their numbers began to decrease before the peak of brain injury, supporting the hypothesis that gut-derived macrophages migrate to the brain through the blood circulation (Fig. S4C–E). Using Cd68-Cre:R26-LSL-KikGR mice and in vivo photoconversion, we obtained more potent evidence that after II/R, many gut-derived macrophages (KikR⁺CX3CR1⁺) were transferred into the brain at the peak of brain injury, and the number of KikR⁺CX3CR1⁺ macrophages was also increased in the blood before the peak of brain damage (Fig. S4F–J). Collectively, these data indicate that II/R triggers the trafficking of gut-derived macrophages into the brain.

II/R induces TREM1 expression in gut-derived macrophages that infiltrate the brain

We used FACS to sort macrophages infiltrating the brain after II/R for a high-throughput RNA-sequencing analysis to screen potential proinflammatory molecules expressed in gut-derived macrophages after intestinal ischemia/reperfusion injury. The samples were divided into the sham, 12 h, 24 h, and 72 h groups, and each sample was composed of a sufficient number of cells sorted from approximately 7–8 mice in the same group to ensure good representativeness for the subsequent analysis. We first evaluated the overall distribution of genes in each sample and confirmed that the samples were comparable (Fig. S5A). The clustering analysis verified the reliability of the groups, as indicated by the minimal variation within each group (Fig. S5B). Principal component analysis (PCA) revealed that the 24 h group was clearly separated from the sham group, suggesting the most significant differences in overall gene expression between these groups (Fig. S5C). Similarly, the weak correlation between the 24 h and sham groups was further illustrated by the Pearson correlation coefficient (Fig. S5D).

According to the established thresholds, we identified 1959 differentially expressed genes (DEGs) between the 24 h and sham groups, 410 DEGs between the 12 h and sham groups, and 564 DEGs between the 72 h and sham groups (Fig. S5E, F). Among the DEGs, the expression level of Trem1 was upregulated at 12 h, 24 h, and 72 h after modeling and peaked at 24 h, which was also the peak of brain injury after II/R (Fig. S5G, H). In addition, the expression of TREM1-related genes reported previously [22], such as Tyrobp (encoding DAP12) and Syk (encoding spleen tyrosine kinase), was significantly upregulated at 24 h after II/R, accompanied by an increase in the expression of a variety of proinflammatory cytokines, chemokines, and relevant receptors (Fig. S5I). The changes in gene expression were further confirmed by quantitative real-time polymerase chain reaction (qRT–PCR) (Fig. S5J). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed on the DEGs between the 24 h and sham groups to determine the specific biological processes and signaling pathways involved in brain damage after II/R (Fig. S5K, L). The GO analysis revealed that the DEGs were significantly enriched in the biological processes of the inflammatory response, myeloid leukocyte migration, and chemotaxis. The KEGG analysis revealed that the DEGs were enriched in several pathways closely related to TREM1, such as the TLR signaling and nuclear factor kappa-B (NF-κB) signaling pathways. These results suggest that TREM1 signaling might be involved in the brain injury caused by gut-derived macrophages after II/R.

We then examined the protein levels of TREM1 expressed in macrophages after II/R. Immunofluorescence staining revealed the increased accumulation of TREM1⁺CX3CR1⁺ cells in the brain (Fig. S6A). Notably, the expression of TREM1 on macrophages infiltrating the brain was increased at 24 h, which was consistent with the peak of brain damage after II/R (Fig. S6B, C). Western blotting of sorted macrophages from the brain further confirmed the increased expression of TREM1 (Fig. S6D). Surprisingly, the number of TREM1⁺ microglia was also increased at 24 h after model establishment (Fig. S6E, F), possibly because infiltrating macrophages produced inflammatory molecules that acted as ligands to further activate TREM1 on microglia. We also observed that TREM1⁺ macrophages accumulated in the lamina propria of the small intestine at the early stage after II/R and peaked before the peak of brain injury (Fig. S6G–K). Given that the changes in TREM1⁺ macrophages in the blood after II/R paralleled those in the small intestine (Fig. S6L, M), we speculated that II/R might prime macrophages with the proinflammatory TREM1 signal before they migrate to the blood and the brain. We validated this speculation by repeating the in vivo photoconversion experiments and discovered a significantly increased number of KikR⁺TREM1⁺ macrophages in the small intestine after II/R, accompanied by increased expression of TREM1 on KikR⁺ macrophages in the blood and brain but not on nongut-derived (KikR^-) macrophages in the brain (Fig. S6N–P). Together, these data imply that II/R induces TREM1 expression in gut-derived macrophages and promotes their migration to the brain via the circulation.

CA/CPR evokes the trafficking of TREM1⁺ gut-derived macrophages to the brain

We then sought to determine whether the expression of TREM1 on macrophages was also upregulated after CA/CPR in mice. We observed that an increasing number of TREM1⁺CX3CR1⁺ cells accumulated in the brain after CA/CPR, with a sharp increase observed at 72 h (Fig. 5A). The flow cytometry results revealed that TREM1 expression was significantly increased in macrophages infiltrating the brain at 12 h after CA/CPR and further increased at 72 h, which was consistent with the peak time of brain injury (Fig. 5B, C). Western blotting of sorted macrophages from the brain further confirmed the increased expression of TREM1 (Fig. 5D). Consistent with the findings observed after II/R, the dynamic changes in TREM1⁺ microglia paralleled those in TREM1⁺ macrophages infiltrating the brain (Fig. 5E–G). This finding implies that after CA/CPR, the infiltrating TREM1⁺ macrophages and microglia might create a vicious cycle in which microglia recruit peripheral macrophages to the brain [4, 23], while the macrophages, in turn, build a proinflammatory microenvironment that activates TREM1 on microglia.

Fig. 5.

Post-CA/CPR intestinal injury evokes the accumulation of TREM1⁺ gut-derived macrophages in the brain. A Immunofluorescence staining showing TREM1⁺CX3CR1⁺ cells in the brain after CA/CPR. B, C Flow cytometry plots indicating the dynamic changes in TREM1⁺ macrophages in the brain after CA/CPR. D Western blotting was used to detect TREM1 expression in cerebral macrophages sorted by FACS. E–G Gating strategy (E) and flow cytometry plots indicating the dynamic changes in TREM1⁺ microglia in the brain after CA/CPR. H Immunofluorescence staining showing TREM1⁺CX3CR1⁺ cells in the lamina propria of the small intestine after CA/CPR. I‒J Flow cytometry plots indicating the dynamic changes in TREM1⁺ macrophages in the small intestine after CA/CPR. K Western blotting was used to detect TREM1 expression in intestinal macrophages sorted by FACS. L Flow cytometry plots indicating the dynamic changes in TREM1 expression in intestinal macrophages identified with a more precise surface marker panel (Ly6C^-F4/80⁺CX3CR1⁺MHC-II⁺) after CA/CPR. Flow cytometry plots showing increased TREM1 expression in gut-derived KikR⁺ macrophages in the brain (M) and small intestine (N) after CA/CPR. Scale bars, 50 or 100 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the sham group; ### P < 0.001 compared with 48 h. n = 5 per group

We further evaluated TREM1 expression in macrophages in peripheral organs after CA/CPR. Immunostaining revealed an increase in the number of TREM1⁺CX3CR1⁺ cells in the lamina propria of the small intestine at the early stage (Fig. 5H). The number of TREM1⁺ macrophages in the small intestine increased very early at 3 h and peaked before the peak of brain injury (Fig. 5I–L). Similarly, the dynamic changes in TREM1⁺ macrophages in the blood and spleen aligned with those in the small intestine, but the increase started much later (Fig. S7A–F). We thus assumed that the proinflammatory TREM1 signal was first activated in the lamina propria of the small intestine, magnifying inflammation in the brain or other organs through the blood, whereas the splenic macrophages acted as secondary systemic inflammatory amplifiers fueled by the gut-derived TREM1 signal. The flow cytometry results partly support this assumption. We further conducted in vivo photoconversion experiments and found that TREM1 expression was upregulated in KikR⁺ macrophages in the small intestine after CA/CPR, as well as in KikR⁺ macrophages in the brain, spleen, and blood, but not in nongut-derived (KikR^-) macrophages in the brain (Fig. 5M, N, Fig. S7G–H). These findings indicate that CA/CPR evokes the priming of TREM1 expression in gut-derived macrophages and their trafficking to the brain.

As the transcriptome analysis revealed that the DEGs identified in cerebral macrophages after II/R were also enriched in neutrophil migration, we explored the potential role of neutrophils in post-CA/CPR brain injury. We revealed that the number of neutrophils infiltrating the brain was markedly increased after CA/CPR, with a rapid increase at the peak of brain injury (Fig. S8A–B). However, no difference in TREM1 expression in neutrophils was observed between CA/CPR and sham mice, which consistently remained at a high level but did not parallel the changing pattern of brain injury (Fig. S8C–D). Similar findings were observed in the II/R model (Fig. S8E–H). Collectively, these data suggest that infiltrating neutrophils may adversely affect brain injury after CA/CPR through their accumulation in the brain but may not depend on the activation of TREM1.

LP17 efficiently inhibits the accumulation of TREM1 in gut-derived macrophages in the brain after CA/CPR

We used the TREM1 inhibitory peptide LP17 and the specific TREM1 agonist rat TREM1 monoclonal antibody (rTREM1 mAb) to block or reinforce peripheral TREM1 signals, respectively, and further explore the role of TREM1⁺ infiltrating macrophages in post-CA/CPR brain injury [24, 25]. First, we sought to evaluate the intervention effects of these drugs. As shown in Fig. 6A–B, LP17 markedly inhibited the infiltration of macrophages into the brain after CA/CPR, whereas the rTREM1 mAb increased their migration. After in vivo photoconversion, we further found that LP17 efficiently reduced the number of KikR⁺ gut-derived macrophages in the brain after CA/CPR (Fig. 6C). In addition, LP17 suppressed the accumulation of TREM1⁺ macrophages in the brain after CA/CPR (Fig. 6D–F), as further shown by the reduced number of KikR⁺TREM1⁺ gut-derived macrophages in the brain after LP17 treatment (Fig. 6G–I). In contrast, the rTREM1 mAb facilitated the migration of gut-derived TREM1⁺ macrophages to the brain after CA/CPR. Interestingly, LP17 also efficiently suppressed the expression of TREM1 on microglia (Fig. 6J, K), which might be attributed to the reduction in inflammatory stimuli in the brain after blocking peripheral TREM1 signaling. Similarly, LP17 and the rTREM1 mAb also exerted opposite effects on the migration of TREM1⁺ macrophages to the brain after II/R (Fig. S9), where the infiltration of gut-derived macrophages into the brain and their TREM1 expression were suppressed by LP17.

Fig. 6.

LP17 efficiently inhibits the accumulation of TREM1 signals carried by gut-derived macrophages in the brain after CA/CPR. A Changes in the proportions of cerebral macrophages (CD11b⁺CD45^high) and microglia (CD11b⁺CD45^int) in mice treated with LP17 or the rTREM1 mAb after CA/CPR. B Changes in the proportions of cerebral macrophages after excluding monocytes (Ly6C⁺), NK cells (NK1.1⁺), neutrophils (Ly6G⁺), dendritic cells (CD11c⁺), and eosinophils (SiglecF⁺) in the brain after CA/CPR. C Changes in the proportions of gut-derived CX3CR1⁺KikR⁺ macrophages in the brain after LP17 treatment. D, E Flow cytometry plots showing the changes in the proportions of cerebral TREM1⁺ macrophages in mice treated with LP17 or the rTREM1 mAb after CA/CPR. F Western blotting was used to detect TREM1 expression in cerebral macrophages sorted by FACS after treatment with LP17 or the rTREM1 mAb. G, H Immunofluorescence staining showing gut-derived KikR⁺TREM1⁺ macrophages in the brain of mice treated with LP17 or the rTREM1 mAb after CA/CPR. I Flow cytometry plots showing the changes in TREM1 expression in gut-derived KikR⁺ cerebral macrophages from mice treated with LP17 after CA/CPR. J, K Flow cytometry plots showing the changes in the proportions of cerebral TREM1⁺ microglia in mice treated with LP17 or the rTREM1 mAb after CA/CPR. Scale bar, 50 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the sham group; #P < 0.05, ##P < 0.01, and ###P < 0.001 compared with the vehicle group or the indicated group; and δP < 0.05 and δδP < 0.01 compared with the rIgG2A group. n = 5 per group

We subsequently explored the direct effect of LP17 on peripheral TREM1 signaling. First, LP17 decreased the accumulation of macrophages in the small intestine after CA/CPR (Fig. 7A, B). Moreover, LP17 potently inhibited the upregulation of TREM1 in the small intestine after CA/CPR, as indicated by the decreased number of TREM1⁺CX3CR1⁺ cells in the lamina propria, reduced number of TREM1⁺ intestinal macrophages, and downregulated TREM1 expression in sorted intestinal macrophages (Fig. 7C–F). Reduced numbers of macrophages and TREM1 expression were also observed when macrophages in the blood and spleen were assessed after modeling (Fig. 7G–N). In contrast, the rTREM1 mAb significantly increased the expression of TREM1 in macrophages in the small intestine, spleen, and blood after CA/CPR. Concerning the II/R model, the efficient inhibitory effects of LP17 on peripheral TREM1 signaling were confirmed (Fig. S10). The above data indicate that LP17 can forcefully block the accumulation of TREM1 in the brain by inhibiting the expression of TREM1 in gut-derived macrophages and the trafficking of these cells, whereas the rTREM1 mAb has the opposite effect.

Fig. 7.

LP17 substantially inhibits the reinforcement and dissemination of gut-derived TREM1 signaling after CA/CPR. A, B Changes in the proportions of intestinal macrophages in mice treated with LP17 or the rTREM1 mAb after CA/CPR, especially in macrophages identified with a more precise intestinal surface marker panel (Ly6C^-F4/80⁺CX3CR1⁺MHC-II⁺). C Immunofluorescence staining showing TREM1⁺CX3CR1⁺ cells in the lamina propria of small intestine from mice treated with LP17 or the rTREM1 mAb after ROSC. D Flow cytometry plots indicating the changes in the proportions of intestinal TREM1⁺ macrophages in mice treated with LP17 or the rTREM1 mAb after ROSC. E Western blotting was used to detect TREM1 expression in intestinal macrophages sorted by FACS after treatment with LP17 or the rTREM1 mAb. F Flow cytometry plots indicating the dynamic changes in TREM1 expression in intestinal macrophages identified with a more precise surface marker panel (Ly6C^-F4/80⁺CX3CR1⁺MHC-II⁺) after treatment with LP17 or the rTREM1 mAb. G, H Changes in the proportion of macrophages in the blood of mice treated with LP17 or the rTREM1 mAb, especially macrophages expressing surface markers similar to those of gut macrophages (CX3CR1⁺MHC-II⁺). I, J Flow cytometry plots indicating the dynamic trend of the number of TREM1⁺ macrophages in the blood of mice treated with LP17 or the rTREM1 mAb after ROSC, especially macrophages expressing surface markers similar to those of gut macrophages (CX3CR1⁺MHC-II⁺). K Changes in the proportions of splenic macrophages in mice treated with LP17 or the rTREM1 mAb after CA/CPR. L Changes in the proportions of splenic macrophages after excluding monocytes, NK cells, neutrophils, and eosinophils. M Flow cytometry plots indicating the dynamic trend of changes in TREM1⁺ macrophages in the spleen after treatment with LP17 or the rTREM1 mAb after ROSC. N Western blotting was used to detect TREM1 expression in splenic macrophages sorted by FACS after treatment with LP17 or the rTREM1 mAb. Scale bar, 100 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the sham group; ##P < 0.01 and ###P < 0.001 compared with the indicated group; and δP < 0.05, δδP < 0.01, and δδδP < 0.001 compared with the rIgG2A group. n = 5 per group

LP17 exerts dual protective effects on the small intestine and brain after CA/CPR

We further evaluated the impacts of LP17 and the rTREM1 mAb on brain injury and intestinal injury after CA/CPR. The histological examination revealed that LP17 mitigated mucosal damage in the small intestine at 48 h after CA/CPR, which was exacerbated in the presence of the rTREM1 mAb (Fig. 8A). Similarly, the impaired intestinal barrier after CA/CPR was also repaired by LP17 but aggravated by the rTREM1 mAb (Fig. 8B, C). In addition, LP17 reversed the reduction in the number of Ki67⁺ cells in the small intestine, indicating that LP17 could improve the repair of the small intestine (Fig. 8B, C).

Fig. 8.

LP17 exerts dual protective effects on the small intestine and brain after CA/CPR. A H&E staining and corresponding Chiu’s scores indicating intestinal mucosal injury in mice treated with LP17 or the rTREM1 mAb after CA/CPR. B Representative images of immunohistochemical staining for tight junction proteins and Ki67 in mice treated with LP17 or the rTREM1 mAb after CA/CPR. C Changes in the serum levels of intestinal mechanical barrier damage indicators, including DAO and I-FABP, in mice treated with LP17 or the rTREM1 mAb after CA/CPR. D Heatmap reflecting the expression of proinflammatory genes in sorted cerebral macrophages from mice treated with LP17 or the rTREM1 mAb after CA/CPR. E, F Confocal microscopy was used to detect the colocalization of proinflammatory or anti-inflammatory markers with Iba-1 in mice treated with LP17 or the rTREM1 mAb after CA/CPR. G Kaplan‒Meier analysis of cumulative survival during the 7-day follow-up after CA/CPR. n = 23. H Neurological function scores of surviving mice in the indicated groups after CA/CPR. I Latency to drop from the accelerating rotarod. Latency to find the platform (J) and swimming distance (K) during the spatial exploration experiment. L Representative swimming paths of the mice during the probe trial. M Percentage of time spent in the target quadrant and platform crossings during the probe trial. N Swimming speed of the mice in the Morris water maze test. O Representative photomicrographs of post-CA/CPR neuropathological damage characterized by TUNEL and immunohistochemical staining for NeuN, MAP-2, Iba-1, and GFAP. Scale bars, 50 μm, 100 μm, or 200 μm. *P < 0.05 and ***P < 0.001 compared with the sham group; #P < 0.05, ##P < 0.01, and ### P < 0.001 compared with the CA/CPR group; and δP < 0.05 and δδP < 0.01 for the comparison of the rTREM1 mAb and sham groups. n = 5 per group

We then detected whether LP17 alleviated brain injury after CA/CPR. As shown in the heatmap, CA/CPR facilitated the transition of cerebral infiltrating macrophages to a proinflammatory state, which was reversed by LP17 treatment (Fig. 8D). Consistently, the regulation of post-CA/CPR cerebral macrophage functional status with LP17 was further suggested by the findings that LP17 substantially increased the number of macrophages positive for both the anti-inflammatory marker Arg1 and Iba-1 (ionized calcium-binding adapter molecule-1) but dramatically decreased the number of macrophages positive for both the proinflammatory marker CD16/32 and Iba-1 (Fig. 8E, F). In contrast, the rTREM1 mAb further strengthened the proinflammatory status of cerebral infiltrating macrophages.

Regarding the 7-day survival after CA/CPR, 39% (9 of 23) of the mice in the CA/CPR group survived, whereas 69% (16 of 23) of the mice in the LP17 group and 34% (8 of 23) of the mice in the rTREM1 mAb group survived (Fig. 8G). Furthermore, LP17 improved neurological outcomes after CA/CPR, as shown by higher neurological function scores than those of the CA/CPR and rTREM1 mAb groups (Fig. 8H). We subsequently performed several behavioral tests. LP17-treated mice displayed better performance on the rotarod test at 72 h after modeling (Fig. 8I). During the hidden-platform trial of the Morris water maze, LP17 markedly reversed the increases in both the time and swimming path length required to find the hidden platform in the CA/CPR group, whereas the rTREM1 mAb maintained the increases in these parameters (Fig. 8J, K). In the probe trial, LP17, but not the rTREM1 mAb, significantly reversed the reductions in both the target quadrant travel time and the number of platform crossings after CA/CPR (Fig. 8L, M). Notably, the swimming speed was comparable among the groups, indicating that the differences in performance on the Morris water maze test were not attributed to a motor impairment (Fig. 8N). These results confirm that LP17 treatment improves cognitive dysfunction after CA/CPR. In addition, LP17 markedly attenuated the neuronal loss, dendritic injury, cellular apoptosis, and glial activation caused by CA/CPR. In contrast, these injuries were even somewhat aggravated in the presence of the rTREM1 mAb (Fig. 8O). In summary, the above results indicate that LP17 exerts dual protective effects on post-CA/CPR intestinal and brain damage, further suggesting the critical role of gut-derived macrophages in linking intestinal and brain injury after CA/CPR through TREM1 signaling.

TREM1 expression in peripheral monocytes/macrophages in CA survivors is upregulated and closely related to the disease severity and systemic inflammation

We collected blood samples from 10 patients who were resuscitated from CA and 8 healthy controls without evidence of brain or intestinal injury (Table S1). After the PBMCs were isolated, we utilized flow cytometry to measure the levels of TREM1 expressed on monocytes/macrophages in the blood. As shown in Fig. 9A, B, compared with those in healthy controls, the number of TREM1⁺ monocytes/macrophages in the blood of CA patients was dramatically increased, along with upregulated TREM1 expression. Moreover, the TREM1 expression levels in monocytes/macrophages were positively correlated with the APACHE II score, reflecting the severity of the patient’s condition, and negatively correlated with the GCS score, with higher scores indicating a better state of consciousness (Fig. 9C). Consistent with the results of the mouse experiments, the upregulation of TREM1 on monocytes/macrophages was closely related to the levels of biomarkers of the systemic inflammatory response, CRP and IL-6 (Fig. 9D). The above results suggest that CA/CPR induces increased TREM1 expression in human peripheral monocytes/macrophages, which closely correlates with the disease severity and systemic inflammation.

Fig. 9.

TREM1 expression in the peripheral monocytes/macrophages in CA survivors is upregulated and closely related to the disease severity and systemic inflammation. A, B Flow cytometry plots showing the levels of TREM1 expressed on monocytes/macrophages in the blood of healthy controls and CA survivors. C The correlations between patients’ TREM1 levels in the blood and patients’ APACHE II or GCS scores. D The correlations between patients’ TREM1 levels in the blood and patients’ serum levels of CRP or IL-6. *P < 0.05 and ***P < 0.001 compared with the control group. n = 8 or 10 per group

CA/CPR induces rapid gut dysbiosis with Enterobacter blooming

We then sought to determine the mechanisms responsible for the upregulated expression of TREM1 in the small intestine after CA/CPR. Severe destruction of the small intestinal barrier often leads to intestinal flora disorders [26], which in turn activates the innate immune response through the release of pattern recognition molecules such as LPS, a potential bacteria-derived ligand of TREM1 [22, 27, 28]. Therefore, PCR and 16S rRNA gene sequencing were performed to explore the changes in the gut microbiota at the early stage (3 h) after CA/CPR. We first confirmed the rationality of the sequencing data (Fig. S11A, B). The differences in the rank–abundance curves and alpha diversity indices suggested altered community richness and diversity of the gut microbiota after CA/CPR (Fig. S11C, Fig. S12A).

As shown in Fig. S12B, analysis of similarities (ANOSIM) revealed significant differences between the sham and CA/CPR groups, as indicated by the hierarchical clustering tree and sample distance heatmap (Fig. S11D–F). Moreover, the results from PCA, principal coordinate analysis (PCoA), and nonmetric multidimensional scaling (NMDS) revealed that the gut microbiota of the sham and CA/CPR groups clustered in completely separate groups (Fig. S12C). We further explored the trends for the changes in the abundance of each species in the different groups and observed a significant increase in the relative abundance of Proteobacteria at the phylum level and Enterobacter at the genus level in the CA/CPR group compared with the sham group (Fig. S11G–I, Fig. S12D–E). Linear discriminant analysis effect size (LEfSe) analysis was employed to identify the key bacterial taxa responsible for discriminating between groups and confirmed that Enterobacter was the most significantly changed taxon after CA/CPR and likely played a key role in post-CA/CPR pathophysiological processes (Fig. S12F). In particular, PICRUSt2 analysis was performed to predict potential functional information about the altered microbial communities, indicating that the genomic abundance of some pathways, such as those related to energy substance metabolism, energy production, and defense mechanisms, was significantly impaired in the CA/CPR group (Fig. S12G). However, the pathways for LPS biosynthesis, the bacterial invasion process, and cationic antimicrobial peptide resistance were dramatically strengthened after CA/CPR, suggesting the possible involvement of LPS released by the disturbed intestinal flora in post-CA/CPR pathophysiological processes. Collectively, these data illustrate that CA/CPR rapidly causes gut microbiota disorders at the early stage, with a sharp expansion of Enterobacter.

Gut dysbiosis exacerbates brain injury after CA/CPR by priming gut-derived macrophages with TREM1 expression

ABX treatment was used to deplete the gut microbiota, with or without the recolonization of Enterobacter cloacae (E. cloacae), the type strain of Enterobacter, to clarify whether the blooming of Enterobacter after CA/CPR was associated with increased TREM1 expression in the small intestine. First, ABX-treated mice presented a decrease in the number of TREM1⁺CX3CR1⁺ cells in the lamina propria and reduced TREM1 expression in intestinal macrophages, as determined by FACS; these effects were almost completely reversed after E. cloacae recolonization (Fig. 10A, B, Fig. S13A). After ABX treatment, the severe mucosal injury, destruction of the intestinal barrier, and impaired repair of the small intestine at 48 h after CA/CPR were attenuated but recovered in the presence of E. cloacae (Fig. 10C, D, Fig. S13B–D), indicating the unwanted effects of Enterobacter expansion on post-CA/CPR small intestinal injury through an increase in the TREM1 levels carried by gut-derived macrophages.

Fig. 10.

Gut dysbiosis aggravates secondary brain injury after CA/CPR by intensifying TREM1 signaling in intestinal macrophages. A Immunofluorescence staining showing intestinal TREM1⁺CX3CR1⁺ cells from mice treated with ABX or E. cloacae after ROSC. B Western blotting was used to detect TREM1 expression in intestinal macrophages sorted by FACS. C H&E staining showing intestinal mucosal injury. D Representative images of immunohistochemical staining for the tight junction proteins and Ki67. E Immunofluorescence staining showing cerebral TREM1⁺CX3CR1⁺ cells from mice treated with ABX or E. cloacae after ROSC. F Western blotting was used to detect TREM1 expression in cerebral macrophages sorted by FACS. G Confocal microscopy was used to detect the colocalization of proinflammatory or anti-inflammatory markers with Iba-1. H Kaplan‒Meier analysis of cumulative survival during the 7-day follow-up after ROSC. n = 23. I Neurological function scores of the surviving mice in the indicated groups after ROSC. J Latency to drop from the accelerating rotarod. K Representative swimming paths of the mice during the probe trial. L Bar graphs of post-CA/CPR neuropathological damage in mice treated with ABX or E. cloacae. Scale bars, 50 μm, 100 μm, or 200 μm. ***P < 0.001 compared with the sham group; #P < 0.05 and ##P < 0.01 compared with the CA/CPR group. n = 5 per group

We further explored the role of sharply increasing Enterobacter abundance in post-CA/CPR brain injury. First, the accumulation of cerebral TREM1⁺CX3CR1⁺ cells and the increase in TREM1 expression in infiltrating macrophages in the brain were significantly inhibited by ABX treatment but dramatically restored by the recolonization of E. cloacae (Fig. 10E, F, Fig. S13E, F). In addition, the macrophages infiltrating the brains of ABX-treated mice presented an obvious anti-inflammatory phenotype after CA/CPR but switched to the proinflammatory state again in the presence of E. cloacae, as further indicated by the results of immunofluorescence colocalization (Fig. 10G, Fig. S13G, H).

Notably, ABX treatment (16 of 23) improved the 7-day survival after CA/CPR (7 of 23), but the survival rate decreased to 7/23 when E. cloacae were recolonized (Fig. 10H). Moreover, the ABX-treated mice also presented increased neurological function scores at 72 h after CA/CPR (Fig. 10I). However, the favorable impact of ABX treatment disappeared after the recolonization of E. cloacae. In the behavioral tests, ABX-treated mice performed better on the rotarod test at 72 h after CA/CPR, but the improved motor coordination was reversed again in the E. cloacae group (Fig. 10J). During the hidden platform trial, both the time and swimming path length when searching for the hidden platform were markedly shorter in the ABX group but increased again after E. cloacae recolonization (Fig. S13I). Similarly, ABX increased both the target quadrant travel time and the frequency of crossing the platform area in the probe trial, whereas these changes were largely reversed by E. cloacae (Fig. 10K, Fig. S13J). After the potential influence of a motor impairment was excluded (Fig. S13K), our results demonstrated that ABX treatment improved short-term spatial learning and memory abilities after CA/CPR. Finally, we confirmed that ABX treatment markedly attenuated post-CA/CPR neuropathological injury, which was aggravated with the recolonization of E. cloacae (Fig. 10L, Fig. S13L). These results suggest that gut microbiota disorders characterized by a sharp increase in Enterobacter abundance aggravate brain injury after CA/CPR by priming gut-derived macrophages with TREM1 expression.

LPS released by overgrown Enterobacter promotes TREM1 expression in intestinal macrophages via the activation of TLR4

The overgrowth of Enterobacter, a typical LPS-producing bacterial genus, may lead to a sharp increase in the levels of LPS, which has been reported to upregulate TREM1 expression in macrophages [22, 28]. Considering the previous finding that the activation of TLRs induced the expression of TREM1 by facilitating the nuclear translocation of NF-κB and activator protein-1 (AP-1) [22], we speculated that the release of LPS by blooming Enterobacter might upregulate TREM1 expression in intestinal macrophages via the nuclear translocation of NF-κB and AP-1 after the activation of TLR4, a crucial receptor that recognizes LPS [29]. We first confirmed the increased levels of LPS in the serum and small intestine after CA/CPR, accompanied by increased serum levels of LBP (lipopolysaccharide binding protein), whose activation assisted in stimulating the LPS-related TLR4 pathway [30]. The above changes were substantially reversed by ABX treatment but recovered after the recolonization of E. cloacae (Fig. 11A), suggesting that the blooming of Enterobacter after CA/CPR might be the primary source of LPS. After the intestinal macrophages were sorted, we discovered the increased nuclear translocation of the NF-κB subunit p65 and AP-1 in the intestinal macrophages after CA/CPR (Fig. 11B, C). Subsequently, we found that the increased expression levels of multiple proinflammatory genes, including Trem1, in the intestinal macrophages after CA/CPR were dramatically reversed by blocking TLR4-related signal transduction or weakening the DNA-binding ability of AP-1 and NF-κB subunit p65 (Fig. 11D). Similarly, the increased protein level of TREM1 in intestinal macrophages after CA/CPR was also decreased by inhibiting AP-1 and the NF-κB subunit p65 (Fig. 11E, F). In summary, these findings illustrate that the sharp expansion of Enterobacter after CA/CPR results in an increase in intestinal LPS levels, and the latter upregulates TREM1 expression in intestinal macrophages by promoting the nuclear translocation of NF-κB and AP-1 after the activation of TLR4.

Fig. 11.

LPS released by overgrown Enterobacter promotes TREM1 expression in intestinal macrophages via the activation of TLR4. A Bar graphs reflecting the levels of LPS and LBP in the serum and small intestine after CA/CPR. B, C Western blotting showing the nuclear translocation of NF-κB p65 and AP-1 in intestinal macrophages sorted via FACS. D Heatmap showing proinflammatory gene expression in sorted intestinal macrophages. E, F Western blotting showing the expression of TREM1 in intestinal macrophages sorted via FACS. ***P < 0.001 compared with the sham group; ##P < 0.01 and ### P < 0.001 compared with the CA/CPR or vehicle group. n = 5 per group

Discussion

After CA/CPR, peripheral macrophages infiltrate the brain, amplify neuroinflammation, and exacerbate secondary brain injury. However, the origin of these macrophages and their mechanism of action remain to be elucidated. In this study, using Cd68-Cre:R26-LSL-KikGR transgenic mice, a novel tool for tracking cell migration, we confirmed that many gut-derived macrophages migrated to the brain (possibly accounting for more than 60% of cerebral infiltrating macrophages) and mediated secondary brain injury after CA/CPR. Moreover, we identified TREM1 as a critical molecule expressed in gut-derived macrophages that links intestinal injury and brain damage. Furthermore, we elucidated that CA/CPR caused a sharp expansion of Enterobacter, which activated TLR4 and downstream signaling through the release of LPS, thereby triggering TREM1 signaling. To our knowledge, this study is the first to clarify the primary origin of macrophages trafficking into the brain after CA/CPR and their critical role in secondary brain injury.

The lamina propria of the small intestine serves as a large reservoir of macrophages that maintain an anti-inflammatory homeostatic microenvironment [31]. Therefore, once intestinal barrier dysfunction occurs, intestinal macrophages are likely to participate in initiating or reinforcing the systemic immune response. In the present study, CA/CPR led to severe mucosal injury and damage to the mechanical barrier of the small intestine. These pathological changes are responsible for the accumulation and outward migration of intestinal macrophages, which may represent a host defense against a weakened gut barrier but generate an unwelcome effect of worsening local and distant organ damage after CA/CPR. Consistent with our results, Schroeder et al. reported small intestinal tissue damage in a rat model exposed to CA/CPR for 6 min [15]. However, no apparent systemic inflammation was detected in their model. The possible reasons for these differences between the two CA/CPR models are listed below. First, our model causes a 10-min CA, with a longer duration of complete blood cessation; second, the diverse animal species used for modeling may have led to large differences, as evidenced by the lack of obvious delayed brain injury in our previous study on CA/CPR rats, suggesting possibly better tolerance to the ischemia/reperfusion of the gut and other peripheral organs in these rats. Previous studies have revealed that only sympathetic inputs after ischemic stroke are sufficient to evoke marked gut injury, which in turn exacerbates brain damage [13, 32]. In contrast, CA/CPR causes not only sympathetic vasoconstriction but also direct ischemia/reperfusion injury to the gut. Indeed, 4 min of CA results in hypoperfusion of the small intestine for up to 60 min and may generate much more severe intestinal damage than that caused by merely sympathetic activation [14]. We thus consider that small intestinal injury after CA/CPR plays a greater governing role in aggravating brain injury.

We then screened potential therapeutic targets for alleviating post-CA/CPR brain injury and identified the substantial upregulation of TREM1 expression in the lamina propria of the small intestine at a very early stage after CA/CPR. TREM1, a transmembrane immune receptor, is a potent amplifier of inflammation in multiple infectious or noninfectious diseases, such as sepsis [33], atherosclerosis [34], pancreatitis [35], and inflammatory bowel disease [36]. As shown in previous studies, TREM1 often enhances the innate immune response to eliminate pathogens and promote the resolution of inflammation in infectious diseases [37, 38]. In contrast, TREM1-mediated excessive neuroinflammation worsens neurological outcomes in mice with noninfectious central nervous system diseases such as ischemic stroke and intracerebral hemorrhage, which are markedly reversed by LP17, a short, highly conserved peptide that acts as a decoy receptor for natural TREM1 ligands [13, 39]. Liu et al. reported that increased numbers of TREM1⁺ macrophages migrate into the brain after ischemic stroke [13]. Similarly, for the first time, our study highlights that the proinflammatory TREM1 signals carried by gut-derived macrophages are transferred to the brain through the circulation after CA/CPR and are crucially involved in aggravating secondary brain injury, as shown by the solid neuroprotective effects of LP17 via the restriction of the activation and migration of gut-derived TREM1 signaling. Schenk et al. reported that the absence of TREM1 expression in lamina propria macrophages in the normal intestine was due to the presence of IL-10 and TGF-β, which prevented excessive inflammatory reactions [40]. Similarly, we found that the increase in TREM1 expression in infiltrating macrophages in the brain was accompanied by an increase in the expression of multiple proinflammatory markers but a decrease in the expression of IL-10 and TGF-β, suggesting a transition of the functional status of cerebral infiltrating macrophages after CA/CPR. However, LP17 accelerates the transformation of gut-derived macrophages into an anti-inflammatory state after CA/CPR to gradually restore the homeostatic microenvironment, which may be attributed to the inhibition of the synergistic effects of TLRs and TREM1 on facilitating the transcription of downstream proinflammatory genes after blocking TREM1-related signal transduction [13, 22].

Xu et al. reported that microglial TREM1 activates the nucleotide-binding oligomerization domain-like receptor containing pyrin domain 3 (NLRP3) inflammasome, leading to pyroptosis via an interaction with SYK [41]. Our recent studies further revealed that microglia play a cardinal role in initiating inflammatory cascades early after CA/CPR via the establishment of a scaffold for NLRP3 inflammasome activation and consequent pyroptosis [6, 42]. Intriguingly, our results revealed that the simultaneous increases in the numbers of TREM1⁺ microglia and cerebral TREM1⁺ macrophages after CA/CPR were accompanied by the dramatic upregulation of TREM1 expression in gut-derived macrophages after they entered the brain. Therefore, we propose that a vicious cycle between the migration of TREM1⁺ macrophages into the brain and the modulation of TREM1 expression in microglia occurs after CA/CPR (Fig. 12). In this detrimental cycle, microglia recruit peripheral macrophages into the brain through the release of chemokines [4, 23], whereas TREM1⁺ macrophages infiltrating the brain, in turn, exacerbate the imbalance of microenvironment homeostasis and promote the TLR-mediated expansion of microglial TREM1 [22]. Microglia then amplify neuroinflammation through TREM1-mediated NLRP3 activation and attract more macrophages. In addition, impaired neurons positively modulate TREM1 signaling in the brain by releasing many damage-associated molecular patterns (DAMPs) [4, 23]. LP17 is clearly capable of breaking this vicious cycle, and we further provided encouraging evidence that LP17 efficiently ameliorates the cerebral inflammatory microenvironment by modulating TREM1-related pathogenic processes, such as by inhibiting initial TREM1 activation in the small intestine, restricting the migration of peripheral macrophages to the brain, and downregulating TREM1 expression in both microglia and infiltrating macrophages.

Fig. 12.

Schematic representation of the detrimental role of gut-derived TREM1 signaling in gut‒brain crosstalk after CA/CPR. The gut dysbiosis characterized by the overgrowth of LPS-producing Enterobacter occurs in the early stage after CA/CPR, which is attributed to severe destruction of the small intestinal barrier caused by both direct ischemia/reperfusion and theoretical sympathetic vasoconstriction. In the context of gut flora disorders, TREM1 expression in small intestinal macrophages is upregulated following the increased release of LPS, which triggers the TLR4-mediated nuclear translocation of downstream transcription factors such as NF-κB and AP-1. In addition to amplifying local inflammation in the small intestine by synergizing with TLR4, gut-derived TREM1⁺ macrophages infiltrate the brain to create a vicious cycle, thus continuously worsening secondary brain injury after CA/CPR. During this detrimental cycle, TREM1⁺ macrophages exacerbate the imbalance of microenvironmental homeostasis and thus promote the TLR-mediated expansion of microglial TREM1 expression, whereas microglia, in turn, recruit more peripheral macrophages into the brain through the release of many chemokines mediated by TREM1. Moreover, impaired neurons positively modulate TREM1 signaling in gut-derived macrophages by releasing damage-associated molecular patterns (DAMPs). LP17 is a potent treatment to break this cycle

A disrupted blood–brain barrier (BBB) is an essential condition for the undue entry of peripheral immune cells into the brain after CA/CPR, as reported in previous studies [43, 44]. Given that excess neuroinflammation in turn drives BBB breakdown [45], targeting TREM1 with LP17 may also serve as a promising strategy to overcome this vicious cycle. First, LP17 can prevent BBB damage by reducing neuroinflammation. Second, LP17 promotes the transformation of cerebral macrophages into an anti-inflammatory state, which hastens BBB remodeling and rescues vascular leakage [46]. In addition, the functional transition of cerebral infiltrating macrophages after LP17 treatment likely potentiates their ability to clear cellular debris after CA/CPR, thereby allowing for neuroinflammation resolution and improved neurological recovery. Because of the dual effects of LP17 on strengthening brain repair and relieving inflammatory brain damage simultaneously, we are convinced that LP17 is a promising candidate for ameliorating secondary brain injury after CA/CPR. From a translational perspective, strategies targeting TREM1 have important clinical implications, given the extended window for intervening in post-CA/CPR neuroinflammation. Furthermore, compared with blocking pattern recognition receptors, targeting TREM1 does not entirely abolish the proper inflammatory response required for normal antimicrobial function. Notably, however, despite LP17 possessing the aforementioned advantages for clinical translation and having shown neuroprotective effects in this study and in previous studies involving small animals, much more work is needed before its clinical application can be realized. Therefore, large animal experiments and clinical trials must be conducted in sequence to test the efficacy and safety of this drug thoroughly.

As elucidated in previous studies, breach of the intestinal barrier results in gut dysbiosis and then triggers the activation of innate immune cells, including monocytes/macrophages [47]. Here, for the first time, we elucidated the pathogenic role of gut microbiota disorders in secondary brain injury after CA/CPR and clarified how gut dysbiosis triggers the activation of gut-derived TREM1 signaling. Compared with previous studies [32], we characterized CA/CPR-related dysbiosis of pathogenic bacteria at the genus level, revealing the substantial expansion of LPS-producing Enterobacter at the early stage. As expected, the serum level of LPS is increased in the early phase after CA/CPR, which is consistent with the findings of a previous study [48]. Xu et al. proposed that a large amount of LPS links the impaired gut barrier to anabatic brain infarction through systemic inflammation [32]. Here, the significance of gut-derived TREM1 signaling in the gut dysbiosis-related pathogenesis of secondary brain injury after CA/CPR was firmly proven by the comparable therapeutic effects between LP17 administration and the modulation of gut dysbiosis and the decreased TREM1 expression in both the brain and small intestine when intestinal flora disorders are disrupted. However, whether the systemic inflammation caused by bacterial product translocation independently functions as a harmful initial factor in post-CA/CPR brain injury remains to be further explored.

This study has several limitations. First, we discovered that TREM1 signaling accumulates in the spleen after CA/CPR, following the initial induction of intestinal TREM1 signaling. Considering the entry of TREM1⁺ gut-derived macrophages into the spleen and the dynamic trend that the number of splenic macrophages first increases and then decreases after CA/CPR, we propose that splenic macrophages may serve as a secondary amplifier of systemic inflammation through outward migration after being fueled by gut-derived TREM1 signaling. However, the specific role of the spleen in post-CA/CPR brain injury remains unknown, and whether the increase in TREM1 signaling in the spleen aggravates intestinal injury in turn and creates a vicious cycle remains to be explored. Second, microglia recruit peripheral CD8⁺ T cells to worsen brain injury after cerebral ischemia [49], but evidence for the role of peripheral T cells in secondary brain injury after CA/CPR is lacking in this study.

In conclusion, the proinflammatory TREM1 signals carried by gut-derived macrophages are transferred into the brain and are critically involved in the pathogenesis of secondary brain injury after CA/CPR. Mechanistically, the induction of intestinal TREM1 signaling closely correlates with gut microbiota disorders characterized by a substantial expansion of Enterobacter in the early stage after CA/CPR. This study thoroughly describes the therapeutic potential of LP17 in secondary brain injury after CA/CPR and extends its application to combat neurological diseases involving gut dysbiosis-related peripheral TREM1 activation. Nevertheless, the transition of LP17 from laboratory to clinical practice is not yet fully bridged. Moving forward, it is essential to systematically undertake large-scale animal studies followed by clinical trials to comprehensively evaluate the therapeutic efficacy and safety profile of LP17.

Significance of this study

What is already known about this subject?

• Brain injury is the leading cause of death and disability in survivors of cardiac arrest.

• Persistent neuroinflammation plays a pivotal role in the pathogenesis of brain injury after cardiac arrest, in which macrophages infiltrating the brain play a pivotal role.

• The origin of macrophages infiltrating the brain after cardiac arrest has not been clarified.

• •The small intestine stores many macrophages and is highly vulnerable to ischemia/reperfusion injury.

What are the new findings?

• Small intestine-derived macrophages infiltrate the brain and play a crucial role in triggering secondary brain injury after CA/CPR.

• The proinflammatory TREM1 signal carried by gut-derived macrophages acts as a critical executor of post-CA/CPR brain injury.

• Gut dysbiosis characterized by a sharp expansion of Enterobacter after CA/CPR triggers intestinal TREM1 signaling via the activation of Toll-like receptor 4 on macrophages through the release of lipopolysaccharide.

How might it impact clinical practice in the foreseeable future?

• This study links small intestinal injury to secondary brain injury after CA/CPR and may alert clinicians to the concern and management of gut injury after CA/CPR.

• This study may open the door to explore the role of peripheral TREM1 signaling as a target to disrupt the vicious cycle in the pathogenesis of post-CA/CPR brain injury and extend the clinical applications of LP17 to combat neurological diseases involving gut dysbiosis-related peripheral TREM1 activation.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41423-025-01263-0.