Inflammatory Monocytes and Subsets of Macrophages with Distinct Surface Phenotype Correlate with Specific Integrin Expression Profile during Murine Sepsis

Sepsis, a multifarious clinical syndrome, is a life-threatening organ dysfunction caused by the dysregulated host response to uncontrolled infection (1). It is a leading cause of death, affecting more than 26 million people worldwide and claims more than 210,000 lives every year in the United States, with similar rates in other countries (2, 3). A 2017 World Health Organization resolution recognized sepsis as a global health priority (4). Sepsis could result from bacterial, fungal, or viral infections and have devastating consequences, as seen with COVID-19 (5). The ambiguous pathobiology of the septic syndrome goes through two stages, namely an initial hyperinflammatory phase and a compensatory anti-inflammatory phase (6). Numerous reports have demonstrated an essential role of innate immune cells, such as neutrophils and macrophages, in driving sepsis-associated systemic inflammation (7). These hyperactive events culminate in a dysfunctional host response, leading to multiorgan failure and death. Despite numerous efforts in improving diagnosis and therapies, there are currently no specific drugs available to cure or slow down the progression of sepsis.

Monocytes and macrophages are a key subset of leukocytes that serve as sentinel and patrolling cell types and play an essential role in the induction and resolution of inflammatory responses. During sepsis, along with neutrophils, infiltrating monocytes are believed to contribute toward the initial cytokine storm that may lead to multiorgan dysfunction (8–10). Similarly, because of their plasticity property and ability to switch from a classical inflammatory (M1) to an alternative phenotype (M2), macrophages are also shown to drive the immunosuppressive responses during sepsis (11, 12). Integrins are the transmembrane heterodimeric glycoproteins that mediate extravasation and interstitial migration of immune cells during inflammation. In addition, binding of surface integrins to extracellular matrices also provides essential signals for the survival, differentiation, proliferation, and immunological functions of the infiltrating cells (13). Integrin receptors are divided into 24 subtypes based on different combinations of α and β subunits and are shown to be associated with distinct tissue trafficking and cellular functions of macrophages. For example, integrin CD11d (α_D), α₅β₁, α₄β₁, and α_vβ₃ are shown to mediate monocyte trafficking and extravasation in specific infections and tumor conditions (14–18). However, a lot more remains to be understood about the possible correlation between specific β₁ and β₃ integrin receptors expression on monocytes and their mobilization to different tissues in response to various stimuli. Second, it is also important to understand whether macrophages expressing particular β₁ or β₃ integrin receptors polarize into a regulatory phenotype during systemic inflammation.

Using two widely used murine sepsis models, namely cecal ligation and puncture (CLP) and LPS-induced endotoxemia, we have performed a detailed and systematic investigation of the monocyte migration and differentiation postinduction (p.i.) of sepsis along with its correlation with the expression dynamics of important β₁ and β₃ group of integrins. We have also evaluated the correlation of specific surface integrin receptors with the functional phenotype of macrophages in the inflamed peritoneum and lungs. Our results show prominent differences in the monocyte mobilization and recruitment pattern following induction of sepsis in both models. In this study, the inflammatory monocytes showed upregulation in the expression of integrin α₄β₁, and the macrophages in the peritoneum and lungs showed prominent integrin α₅β₁ and α_vβ₃ expression in both endotoxemia and CLP animals. The macrophages demonstrate an immunosuppressive phenotype in the lungs of the septic animals, as indicated by higher CD206 expression, which correlated with higher IL-10 levels and lower proinflammatory cytokine transcript levels in the lungs compared with mesentery. The F480⁺CD206^hi expressing macrophages in both peritoneum and lungs showed a higher level of α_vβ₃ expression. Furthermore, using bioinformatics tools, we have analyzed and validated the expression of integrin α_4, α₅, and, α_v with their possible association with inflammatory/anti-inflammatory functions of macrophages. The presented data will help improve the understanding of the monocyte and macrophage responses in two widely used murine models of sepsis and other infectious and inflammatory diseases.

CLP and LPS-induced endotoxemia was performed in 8- to 12-wk-old C57BL/6 male mice procured from Council of Scientific and Industrial Research–Institute of Microbial Technology (Chandigarh, India) and Indian Institute of Science Education and Research (Mohali, India), according to the protocol approved by the Institute Animal Ethics Committee, Indian Institute of Technology Roorkee. For the endotoxemia assay, LPS from Escherichia coli O55:B (Sigma-Aldrich, St. Louis, MO) was administered to achieve 70–80% mortality by i.p. injection (36.7 mg/kg body weight) as per the titration of the received lot of LPS. For CLP surgery, mice were anesthetized by i.p. administration of a mixture of ketamine and xylazine. After midline incision, the cecum was taken out and ligated. Using a 21-gauge needle, the cecum was punctured through and through following the ligation procedure. After wound closure using silk suture, lignocaine was applied to the incision site for local anesthesia. Dextrose normal saline was injected s.c. to resuscitate the animals following the surgery. During the entire experimentation period, access to soft food pellets and water was made available to the animals. Bone marrow (femur), lungs, blood, and peritoneal lavage were isolated at 6, 12, and 24 h post sepsis induction. Lungs were digested using collagenase IV (Sigma-Aldrich, St. Louis, MO). Isolated cells were processed for flow cytometric analysis. BD Horizon Fixable Viability Stain 450 was used to gate live cells (FVS450; BD Biosciences, San Jose, CA). For surface marker staining, purified anti-mouse FcR (CD16/CD32), Ly-6C–allophycocyanin/Cy7, F4/80–Alexa Fluor 488/PerCP, MHCII–PerCP, CD206(MMR)–PE/Cy7, CD11b– allophycocyanin, CD11b–FITC, CCR2–FITC, CD49b–PE, CD49c– allophycocyanin, CD49d–PE, CD49e–PE, CD49f– allophycocyanin, CD51–PE, CD29–PE (BioLegend, San Diego, CA) and TGFBR1 (Sigma-Aldrich, St. Louis, MO) Abs were used. For detection of TGFBR1, goat anti-rabbit cross-adsorbed secondary Ab-FITC (Thermo Fisher Scientific, Waltham, MA) was used. All samples were fixed with 1% formaldehyde and collected on a FACSVerse flow cytometer (BD Biosciences, San Diego, CA). The data were analyzed using FlowJo software.

For intracellular cytokine staining, isolated cells were incubated in brefeldin A (BD Biosciences, San Diego, CA) and stained further with anti-mouse FcR (CD16/CD32) (BioLegend, San Diego, CA), FVS450 (BD Biosciences, San Diego, CA), Ly-6C–allophycocyanin/Cy7, F4/80–Alexa Fluor 488, CD206 (MMR)–PE/Cy7, CD11b–allophycocyanin and CCR2–FITC (BioLegend, San Diego, CA), followed by treatment with fixation and permeabilization buffers (Thermo Fisher Scientific, Waltham, MA). The cells were further stained with IL-6–PE, IL-10–Alexa Fluor 488, IL-12–PE, TNF-α–Alexa Fluor 488 (BioLegend, San Diego, CA). For phospho-SRC (p-SRC) staining, the isolated cells were stained with surface markers and then stimulated with LPS (Sigma-Aldrich, St. Louis, MO) (50 ng/ml) for 15 min and immediately fixed and permeabilized using Intracellular Fixation & Permeabilization Buffer Set (Thermo Fisher Scientific, Waltham, MA). The cells were further stained with p-SRC–Alexa Fluor 488 (Thermo Fisher Scientific, Waltham, MA). The data were collected on a FACSVerse flow cytometer (BD Biosciences, San Diego, CA) and analyzed using FlowJo software.

Blood was collected from healthy donors according to the guidelines of the Institute Human Ethics Committee, Indian Institute of Technology Roorkee approved protocol, and PBMCs were isolated using Histopaque-1077 (Sigma-Aldrich, St. Louis, MO). The PBMC layer was gently removed, and monocytes were carefully separated using the plastic adherence method, as described previously, with modifications (19). Briefly, 5 × 10⁵ PBMCs were seeded in a 48-well flat-bottom plate in 400 µl of RPMI 1640 without FBS (Life Technologies, Carlsbad, CA) for 1 h. Tightly adhered monocytes were further cultured in RPMI 1640 (10% FBS) in the presence of MCSF (25 ng/ml) for 6 d to differentiate them into macrophages. For flow cytometry staining of integrins, macrophages were detached using cold PBS (purity ∼55%) and stimulated with LPS (100 ng/ml) and IL-1β (10 ng/ml) for 3 h in RPMI 1640 supplemented with 10% FBS at 37°C, and surface staining was performed using purified mouse anti-human integrins β₁ and α_v (BioLegend, San Diego, CA) primary, and PE/Cy7–labeled goat anti-mouse (BioLegend, San Diego, CA) secondary Abs.

For mRNA expression studies, on day 6 of culture, the adhered macrophages in the culture plate were washed multiple times with PBS to remove any that were loosely adhered and increase purity of macrophages (CD14⁺ cells) to over 90% before proceeding for stimulation. Subsequently, the total RNA was isolated from the monocyte-derived macrophages for human data using TRIzol reagent (Life Technologies, Carlsbad, CA). Similarly, total RNA was also isolated from the mesentery (the thin membrane that holds the intestine together) and lungs of naive and septic animals at different time points following sepsis. The RNA was reverse transcribed using iScript Reverse Transcription Supermix for RT-qPCR (Bio-Rad Laboratories, Hercules, CA), and the real-time PCR was performed using a StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA). The real-time PCR was performed using the SYBR Green PCR Master Mix (Bio-Rad Laboratories, Hercules, CA), and PCR amplification of a housekeeping gene (human GAPDH and mouse HPRT) was performed for each sample as a control for sample loading and to allow normalization between samples. The primers used were as follows: human GAPDH forward, 5′-TCCTCTGACTTCAACAGCGACAC-3′ and reverse, 5′-TCTCTCTTCCTCTTGTGCTCTTGC-3′; ITGB1 forward, 5′-GCACCAGCCCATTTAGCTACAA-3′ and reverse, 5′-TCGAAACCACCTTCTGGAGAATC; ITGAV forward, 5′-GCTGTCGGAGATTTCAATGGT-3′ and reverse, 5′-TCTGCTCGCCAGTAAAATTGT-3′; mouse HPRT forward, 5′-GTTAAGCAGTACAGCCCCAAA-3′and reverse, 5′-AGGGCATATCCAACAACAAACTT-3′; mouse CCL2 forward, 5′-AGCAGGTGTCCCAAAGAAGCT-3′ and reverse, 5′-GATCTCATTTGGTTCCGATCCA-3′; mouse IL-6 forward, 5′-GACAAAGCCAGAGTCCTTCAG-3′ and reverse 5′-GCATTGGAAATTGGGGTAGGA-3′; mouse IL-1β forward, 5′-GAAGTTGACGGACCCCAAAA-3′ and reverse, 5′-GTTGATGTGCTGCTGTGAGA-3′; mouse TNF-α forward, 5′-TATGGCTCAGGGTCCAACTC-3′ and reverse, 5′-CCTTGATGGTGGTGCATGAG-3′; and mouse IL-10 forward, 5′-TGCAGGACTTTAAGGGTTACTTGG-3′ and reverse, 5′-GGCCTTGTAGACACCTTGGTC-3′.

Analytical determination of the investigated proteins (integrin α₄, α₅, and α_v) for both physical (direct) and functional (indirect) aspects were analyzed through Search Tool for the Retrieval of Interacting Genes/Proteins (STRING; version 11.0) database (https://string-db.org) (20). Medium confidence (0.400) parameter was chosen for the minimum required interaction score. Also, the structure preview inside the network bubbles was disabled to simplify the display output. All the proteins were individually searched in the STRING database for Homo sapiens and Mus musculus separately to analyze their interacting partners and their respective functional relevance. Interacting proteins with functional relevance with respect to immune cells were shortlisted, and an interaction network was formed for each integrin for both H. sapiens and M. musculus.

The normality of the in vitro and in vivo data were tested using a Shapiro–Wilk test. For the normally distributed data, the statistical differences between two groups were analyzed by performing an unpaired t test with Welch correction, and the data with more than two groups were analyzed using one-way ANOVA with Bonferroni correction. For the data that did not pass the normality test, the statistical differences with more than two groups were analyzed using a Kruskal–Wallis with Dunn multiple comparisons test. All values presented are expressed as the mean ± SEM. A p value of < 0.05 was considered significant. All statistics were performed using GraphPad Prism (GraphPad software, La Jolla, CA).

To investigate the migration of inflammatory monocytes, sepsis was induced in mice using endotoxemia and CLP survival surgery, as described in the Materials and Methods. Flow cytometry was used to explore the changes in the distribution of monocytes and neutrophils in the bone marrow, blood, peritoneum, and lungs following induction of sepsis in the animals. At 6-, 12-, and 24-h time points, cells were isolated from the tissues and stained with CD11b and Ly-6C surface markers. FVS450 staining was used to gate on the live cells. As shown in (Fig. 1A, Ly-6C^hi CD11b^hi, and Ly-6C^int CD11b^hi cells were gated on live cells to differentiate between monocytes and neutrophils, as described earlier (21, 22). Both in CLP and endotoxemia animals, the frequencies (Fig. 1A, upper panel) and the total numbers (Fig. 1B–E) of these cells were reduced in the bone marrow at 6 h compared with naive mice. A moderate change was observed in the blood, suggesting that the neutrophils and monocytes have migrated to other organs within 6 h of p.i. (Fig. 1A, second panel). In support of this, an increase in the numbers of neutrophils and monocytes was observed in the lungs and inflamed peritoneum of septic animals at 6 h p.i. (Fig. 1A). Second, the number and frequencies of the neutrophils and the monocytes were higher in the peritoneum of CLP animals compared with the endotoxemic group of animals (Fig. 1A). The frequency and number of neutrophils increased at 6 h, followed by a mass accumulation of monocytes at the 12-h time point in the peritoneum of CLP animals, which decreased by the 24-h time point. Unlike the peritoneum, there was an initial surge in the frequencies and the total numbers of monocytes by 6 h p.i. in the lungs, followed by a reduction (Fig. 1A–E). This is suggestive of differentiation of infiltrated Ly-6C^hi monocytes into F4/80^hi macrophages.

The peritoneum contains two subsets of macrophages (i.e., large peritoneal macrophages [LPMs] and small peritoneal macrophages [SPMs]) (23). Therefore, to study these subsets of macrophages, Ly-6C^loF4/80^hiCD11b^hi and Ly-6C^loF4/80^loCD11b^lo cells were analyzed as LPMs and SPMs, respectively (Fig. 1F). The frequency and number of LPMs decreased substantially by 6 h in both CLP and endotoxemic animals compared with naive animals (0 h), possibly due to mobilization of existing cells and the concurrent influx of inflammatory neutrophils and monocytes (Fig 1G). In contrast, an increase in the number of SPMs was observed at 6 h, followed by reduction at later time points (Fig. 1H). However, a significant increase in the numbers of Ly-6C⁺ cells showing higher F4/80 and lower Ly-6C expression than inflammatory monocytes at 24 h p.i. appeared only in the CLP group (data not shown). Upon the analysis of cell viability in peritoneum, a gradual decrease in the frequency of live cells in the peritoneum was observed, which might be attributed to the inflammatory microenvironment-induced apoptosis (Fig. 1I). In contrast to the peritoneum, a gradual increase in the numbers of F4/80⁺ cells in the lungs was observed in CLP and endotoxemia groups of animals, suggesting a tissue-specific microenvironment in the lungs to support macrophage differentiation compared with the site of inflammation (Fig. 1J). Because in the endotoxemia group there is a faster onset of inflammatory responses, we checked whether the inflammatory cells could start infiltrating into a distant organ, such as lungs, as early as 2 h p.i, and as shown in (Fig. 1K, at this time point, a significantly higher number of CD11b⁺ cells could be seen in the lungs of endotoxemic animals than naive animals.

Integrin receptors are essential for both intravascular and extravascular migration of leukocytes under both physiological and inflammatory conditions. Previous reports have suggested an altered migration and localization of inflammatory cells in the visceral tissues. Whereas recent reports have emphasized the trafficking of neutrophils, less is known about the monocytes and macrophages. Therefore, using flow cytometry, we measured the expression of important β₁ and β₃ groups of integrins critical for extravasation and interstitial migration of leukocytes. For cell surface staining, cells were isolated from the bone marrow, blood, peritoneum, and lungs at 6, 12 and 24 h p.i. of sepsis and were stained with Ly-6C, CD11b, MHC class II (MHCII) and integrins α₂, α₃, α₄, α₅, α₆, α_v, and β₁. For analysis, cells were gated on Ly-6C^hiCD11b^hi live monocytes, and an increase in the mean fluorescence intensity (MFI) of various integrins was measured and presented in comparison with naive Ly-6C^hi cells. As shown in (Fig. 2A–D upper (CLP) and lower (endotoxemia) panels, among the β₁ group, integrin α₄β₁ was dramatically upregulated on inflammatory monocytes compared with other integrins. No change was observed in the expression of the integrin α₂, α₃, and α₆ (not shown). Additionally, MFI of CD11b/α_Mβ₂, a vital adhesion molecule and a marker of inflammatory phenotype, was also upregulated in bone marrow and blood (Fig. 2A, 2B). In contrast, the expression of CD11b was not much changed in the peritoneum and lungs of both CLP and endotoxemic animals (Fig. 2C, 2D). However, an upregulation in the expression of CD11b was observed at 24 h in the lungs of CLP animals. As seen in (Fig. 2E–H, the shift of the histograms and increase in the MFI values suggest that the entire inflammatory monocyte population exiting from bone marrow and infiltrating into peritoneum and lungs showed an increase in integrin α₄β₁ expression.

CCR2 is the receptor for MCP-1/CCL2 and is responsible for monocyte chemotaxis during inflammation (24). Therefore, we further investigated the expression of CCR2 on the α₄β₁-expressing inflammatory monocytes mobilized from the bone marrow into the circulation in response to inflammatory stimulation. To this end, cells were isolated from the blood and bone marrow of endotoxemic animals at 12 h post-LPS administration. Upon analysis of CD11b^hiLy-6C^hi inflammatory monocytes, two distinct populations based on CCR2 expression (i.e., Ly-6C^hiCCR2^hi and Ly-6C^hiCCR2^lo) were observed both in the bone marrow and blood (Fig. 3A). Between both populations, the Ly-6C^hiCCR2^hi monocytes showed significantly higher integrin α₄ expression compared with the Ly-6C^hiCCR2^lo cells (Fig. 3B, 3C) both in bone marrow and blood. Next, we compared the ability of both population of monocyte to produce inflammatory cytokines, and intracellular IL-6 staining was performed for flow cytometry analysis. As is shown in (Fig. 3D and 3E, whereas the Ly-6C^hiCCR2^hi monocytes showed higher intracellular IL-6 expression compared with the naive animals, no difference was observed in the IL-6 production by Ly-6C^hiCCR2^hi and Ly-6C^hiCCR2^lo monocytes. As shown in (Fig. 3F, both mesentery and lung tissues expressed higher levels of CCL2 (ligand for CCR2) transcript compared with respective naive tissues, although the mesentery had higher transcript expression compared with lungs. These findings suggest that integrin α₄ may assist in the CCL2–CCR2–mediated chemotaxis of the Ly-6C^hiCCR2^hi monocytes into the tissues during systemic inflammation.

After extravasation, monocytes migrate through the interstitial matrix to the site of inflammation. During this process, monocytes undergo differentiation into macrophages by receiving signals from various molecules such as cytokines (25, 26). Similarly, signals received from the binding of surface integrins and extracellular matrix (ECM) protein also have significant roles in the differentiation and cellular functions (13). Thus, it is possible that macrophages may modulate the expression of surface integrin according to the inflammatory milieu and may have unique profiles in different tissues. Therefore, integrin expression on F4/80⁺ cells in the peritoneum and lungs was further investigated at 6-, 12-, and 24-h time points using flow cytometry. As mentioned earlier, F4/80⁺ cells were fewer in both lungs and peritoneum compared with monocytes, possibly due to the shorter duration of the study period. A significant decrease in the LPMs with a moderate increase in the SPMs was observed upon induction of inflammation. Upon integrin analysis on SPMs (F4/80^loCD11b^lo), increased expression of α₅β₁ and α_vβ₃ were observed in both CLP and endotoxemic animals (Fig. 4A, 4B). Interestingly, a similarity was observed between the SPMs in the peritoneum and F4/80⁺ cells in the lungs of both the groups. As shown in (Fig. 4C and 4D, macrophages in the lungs show upregulated expression of α₅β₁ and α_vβ₃ integrins. Similarly, integrin expression analysis was performed in human blood monocyte-derived macrophages stimulated with LPS and IL-1β by qRT-PCR and flow cytometry. As shown in the (Fig. 4E and 4F, there was an increase in the transcripts of integrin β₁ upon stimulation with LPS and IL-1β compared with integrin α_v. Moreover, as seen in (Fig. 4G and 4H, a mild change was observed in the protein expression of integrin β₁ within 3 h of stimulation. These findings suggest a possible function of α₅β₁ and α_vβ₃ integrins in differentiation and other inflammatory functions of macrophages.

In any systemic inflammation, the lung is one of the primary visceral organs to be negatively affected by the inflammatory mediators released by innate immune cells such as neutrophils and macrophages (27, 28). However, after differentiation in the tissue, macrophages may undergo a switch from inflammatory (MHCII^hi) to an anti-inflammatory (CD206^hiMHCII^lo) phenotype by sensing the tissue microenvironment (29). Thus, we analyzed the characteristics of the F4/80⁺ population in the lungs at different time points p.i. of sepsis. In the previous section, the data showed a gradual increase in the number of F4/80⁺ cells in the lungs of both groups of animals. Various surface markers, such as CD11b, Ly-6C, CD206 (macrophage mannose receptor), and MHCII were used to further understand the functional characteristics of F4/80 expressing macrophages.

When gated on the F4/80⁺ cells in the lungs, a significant (more than 2-fold) increase in both frequencies (Fig. 5A) and the total number (Fig. 5B) of Ly-6C^loCD206^hi cells was observed in both models of sepsis compared with the naive animals. This population also expressed significantly higher levels of CD206 compared with naive animals (Fig. 5C). To further confirm whether such cells may have a regulatory or inflammatory phenotype, MHCII and CD11b expression was measured on the F480⁺Ly-6C^loCD206^hi cell population, and, as shown in the figure, such cells had significantly lower levels of MHCII in both models compared with naive animals (Fig. 5D). The CD11b expression was also reduced at 6 h in the endotoxemia model compared with naive animals. However, the change in CD11b expression was not significant in CLP animals compared with naive cells (Fig. 5E). Of note, there was also a moderate increase in the Ly-6C⁺CD206⁺ macrophages in the lungs, suggesting a possible conversion of inflammatory macrophages into CD206-expressing phenotype. Similar to the lungs, the F4/80⁺ cells in the peritoneum were also analyzed to check the presence of such phenotype of macrophages. In contrast to lungs, both frequencies and total number of Ly-6C^loCD206^hi cells were very small and showed a decreasing trend compared with such cells in naive animals (Fig. 5F, 5G). Such cells showed reduced CD206 expression only in the CLP group at 12 h compared with naive animals (Fig. 5H). In both groups, these cells showed relatively higher MHCII expression when compared with naive animals (Fig. 5I). However, CD11b expression was reduced in the septic animals compared with the naive group (Fig. 5J). These findings suggest that macrophages show a distinct difference in their CD206 expression and could exhibit regulatory functions in lungs. Second, this data may also suggest that macrophages may be initiating certain anti-inflammatory signals to protect the lungs from the detrimental effects of inflammation.

As discussed previously, integrin is essential for several functional attributes of immune cells (13, 30). Macrophages are highly plastic cells that play an essential role in the inflammatory and resolution events during an immune response. Our data showed a differential expression pattern of integrin expression in the monocytes and macrophages. Therefore, we analyzed the cells isolated from the peritoneum for their surface expression of F4/80, CD206, and specific integrins that were upregulated on F4/80⁺ cells at the 24-h time point because of their higher numbers in the tissues at this time point. The elevation in the expression of CD206 was limited to very few macrophages in the peritoneum (Fig. 6A, upper panel). As shown in (Fig. 6B, in both the murine models, the macrophages with higher expression of CD206 showed an elevated level of integrin α_v, without any significant changes in the expression of integrin β₁ compared with the macrophages with low CD206 expression (Fig. 6B, 6C; upper panels).

Similarly, F4/80⁺ cells in the lungs were analyzed at 12 h, as the frequencies of cells with Ly-6C^loF4/80⁺CD206^hi phenotype were significantly higher in the lungs of both the models. Approximately 70% of F4/80⁺ cells in the lungs showed higher expression of CD206 in both CLP and endotoxemic animals (Fig. 6A, lower panel). CD206^hi macrophages in the lungs showed higher α_v and β₁ expression compared with CD206^lo macrophages (Fig. 6B, 6C; lower panels). To understand the functional status of these macrophage subsets, intracellular expression of pro- (IL-6, IL-12, TNF-α) and anti-inflammatory (IL-10) cytokines was measured using flow cytometry in the CD206^hi and CD206^lo macrophage population in the peritoneum and lungs of endotoxemic animals at 20 h p.i. As shown in the (Fig. 6D, 6F, and 6G, no significant differences were observed in the expression of cytokines IL-6, IL-12, and TNF-α between both subsets of macrophages in the peritoneum. However, in the lungs, there was a significant increase in the proinflammatory cytokine IL-6 expression in the CD206^hi subset compared with CD206^lo population, with no remarkable differences in the expression of IL-12 and TNF-α. Importantly, the expression of anti-inflammatory cytokine IL-10 was dramatically higher in the CD206^hi population compared with the CD206^lo population in both peritoneum and lungs of endotoxemic animals (Fig. 6E). These data suggest that the F4/80⁺CD206^hi subsets of macrophages may have an inclination to form anti-inflammatory macrophages.

As mentioned earlier, macrophages are highly plastic cell types and are known to switch their functional phenotypes by sensing the cues from the tissue microenvironment. Because the flow cytometry data indicated that the macrophages in the lungs showed higher expression of CD206 than the peritoneum, we further investigated whether the inflammatory mediators in the microenvironment of lungs and peritoneum (mesentery was analyzed) are contributing to the phenotype of macrophages. To this end, we measured the kinetics of different cytokines, such as IL-6, IL-1β, TNF-α, and IL-10 at the mRNA level by qRT-PCR in these tissues. In naive animals, no significant differences were observed in the levels of the transcripts of the above-mentioned cytokines between lungs and mesentery (data not shown). As shown in (Fig. 6H and 6I, induction of sepsis led to the upregulation of the il-6 and il-1β mRNAs in both lungs and mesentery compared with the respective naive tissues, which gradually decreased in the lungs with the progression of sepsis. Although il-1β mRNA levels were higher in the lungs than in the mesentery, the il-6 mRNA was up to 2-fold higher in mesentery than that of lungs. The levels of tnf-α mRNA also increased in both lungs and mesentery, but the relative expression of this proinflammatory cytokine was higher in the mesentery (up to a 50- to 60-fold increase) compared with the lungs (up to a 5-fold increase) when compared with respective naive tissues (Fig. 6H, 6I).

Along with proinflammatory cytokines, a significant upregulation in the expression of mRNA of anti-inflammatory cytokine il-10 was observed in the lungs and mesentery of both CLP and endotoxemic animals with a more prominent increase in the lung tissues (Fig. 6I).

The il-10 level increased as early as 6 h in CLP animals and by 12 h in the mesentery of endotoxemic animals in comparison with naive mesentery tissues (Fig. 6H). Notably, the relative expression of il-10 was dramatically higher in the lungs (up to a 40- to 200-fold increase) compared with mesentery (up to a 2- to 6-fold increase) (Fig. 6H, 6I). As per the data, in the mesentery, the inflammatory cytokines, such as il-6 and tnf-α, levels were higher, and in the lungs, the level of anti-inflammatory cytokine il-10 was higher, which suggests that the tissue microenvironment is relatively inflammatory in the peritoneum compared with the lungs. This may relate to higher CD206-expressing macrophages in the lungs compared with the peritoneum.

Our data showed differential integrin expression among inflammatory monocytes and different phenotypes of F4/80⁺ macrophages. Whereas the Ly-6C^hi inflammatory monocytes showed higher α₄ integrin expression, the macrophages showed higher α_5, and α_v, and the F4/80⁺CD206^hi macrophages showed higher integrin α_v expression. To further analyze the correlation between the integrin expression and functional phenotype of monocytes and macrophages, bioinformatics analysis of mouse and human α₄, α₅, and α_v integrins was performed to explore their interacting partner molecules using STRING v11.0. All the integrin subunits showed to be associated with several other integrin subunits, cell adhesion molecules, ECM protein, and molecules involved in the different signaling cascades. Both in humans and mice, integrin α₄ showed association with fibronectin (FN1) as well as VCAM-1, which may be involved in cell adhesion and migration (Fig. 7A). Integrin α₅ showed interaction with FN1, osteopontin (SPP1), integrin-linked protein kinase (ILK), and Talin-1, etc., which are mainly associated with cell adhesion and integrin signaling. In addition, α₄ and α₅ showed interaction with tyrosine kinase Src, which in addition to cell adhesion and migration has been shown to be associated with inflammatory cytokine production (Fig. 7A, 7B) (31). Integrin α₅ showed interaction with TGFB1 as well, but the interaction was with low confidence, showing weak strength-of-data support. Similarly, α_v integrin showed interaction with fibronectin, vitronectin, and different cell adhesion molecules both in humans and mice. However, in contrast to integrin α₄ and α₅, integrin α_v showed a strong association with TGFB1, TGFBR1, and TGFBR2, which indicates that integrin α_v could be involved in the anti-inflammatory and resolution processes in the macrophages (Fig. 7C).

Next, to validate the findings obtained from the bioinformatics analysis, we performed flow cytometry analysis to evaluate the level of p-SRC in the α₄^hi and α₅^hi subsets of macrophages in naive and endotoxemic animals at 12 h following LPS administration. As shown in (Fig. 8A and 8B, the MFI of p-SRC in F4/80⁺α₄^hi cells as well as the frequency of the F4/80⁺α₄^hi p-SRC^hi cells among F4/80⁺α₄^hi cells in both peritoneum and lungs of endotoxemic animals were higher compared with the naive tissues. As shown in (Fig. 8C, whereas no differences in the absolute numbers of such cells were noticed in the peritoneum, the numbers of F4/80⁺α₄^hip-SRC^hi cells were significantly higher in the lungs of endotoxemic animals compared with naive animals. Similar to the F4/80⁺α₄^hi macrophages, F4/80⁺α₅^hi macrophages also showed elevated expression levels (MFI) of p-SRC. As shown in (Fig. 8D, whereas a similar pattern for the expression levels of F4/80⁺α₅^hip-SRC^hi cells was observed in the peritoneum of endotoxemic animals, a moderate upregulation (p = 0.07) was observed in the lungs. Following LPS administration, the frequencies of the F4/80⁺α₅^hip-SRC^hi cells increased among F4/80⁺α₅^hi cells in the peritoneum of endotoxemic animals, whereas no difference was observed in the case of lungs (Fig. 8E). However, the absolute numbers of the F4/80⁺α₅^hip-SRC^hi cells were again higher in the lungs compared with naive animals (p = 0.06), with no difference in the peritoneum (Fig. 8F).

Because the bioinformatics data showed interactions of TGFBR1 with integrin α_v, indicating a possible role in immunoregulation or resolution, we further analyzed the expression of the TGFBR1 receptor on the F4/80⁺α_v^hi cells in the peritoneum and lungs of endotoxemic animals at both 12 h and 20 h p.i. As shown in (Fig. 9A, the TGFBR1 was highly upregulated in the F4/80⁺α_v^hi macrophages, both in the lungs (12 h and 20 h) and peritoneum (20 h), with a strong expression in the lungs compared with that of the peritoneum. Second, higher frequencies of F4/80⁺α_v^hi TGFBR1^hi macrophages among the F4/80⁺α_v^hi subset was observed in the lungs and peritoneum of endotoxemic animals, although the numbers of such cells were relatively lower in the peritoneum of endotoxemic animals compared with naive peritoneum and inflamed lungs (Fig. 9B, 9C). Because CD206⁺ macrophages were associated with a regulatory phenotype and such cells expressed higher levels of integrin α_v, we next compared the expression of TGFBR1 on such cells. As shown in (Fig. 9D and 9E, F4/80⁺CD206^hi cells expressed relatively higher levels of TGFBR1 compared with F4/80⁺CD206^lo cells in both peritoneum and lungs. These findings are consistent with our hypothesis made from the bioinformatics study, which showed an association of p-SRC with the α₄ and α₅ integrins and TGFBR1 with α_v integrins.

Myeloid cells play a crucial role in driving severity, including multiple organ failure during sepsis. CLP and LPS-induced endotoxemia are two well-established and frequently used murine models for sepsis research. Previous studies have reported striking differences between these two models of sepsis. For example, in LPS-induced endotoxemia, the proinflammatory cytokine level in the plasma peaks earlier and at a very high level compared with the human sepsis (32, 33). However, CLP is often termed as the gold standard of polymicrobial sepsis, as it bears a close resemblance to the pathophysiology of human sepsis (34). In the current study, using these two models of sepsis (i.e., CLP and LPS-induced endotoxemia models), migration, distribution, and differentiation of monocyte and macrophage cell populations were evaluated. Our study also evaluated the gradual differentiation and functional switching of monocytes into macrophages of different phenotype following migration into different tissue compartments. In addition to showing multiple differences in the distribution and association of specific cell surface integrin on specific cellular and functional subsets of monocytes/macrophages, our results indicate an inflammatory phenotype of such cells in the peritoneum and an inclination to take up an immunosuppressive phenotype in the lungs. We have also validated such phenotypic and functional inclinations with bioinformatics and experimental studies.

Following induction of sepsis, there is mobilization of monocytes from the bone marrow and circulation into the tissues. Several signals could alter the migration behaviors of monocytes during a systemic inflammation. Most prominently, chemokines, such as CCL2 and heparin-binding protein/CAP37, are shown to be potent chemoattractants for monocytes during sepsis (35). Inflammation-induced remodeling of ECM proteins may also influence the migration behaviors of innate immune cells (36). We have recently shown major differences in the expression of ECM protein in LPS and CLP models, with an earlier upregulation of various ECM proteins in different tissues in endotoxemic animals compared with CLP (37). In our study, numbers of neutrophils and monocytes were reduced in the bone marrow at six hours compared with naive mice, which correlated with an increase in their number and frequencies in the peritoneum. Although, it is proposed that the rapid dissemination of LPS and faster progression of inflammatory events in endotoxemia may result in brisk recruitment of inflammatory cells in endotoxemia, our results show higher numbers of monocytes and neutrophils in the CLP animals compared with endotoxemic animals, which may be due to stronger stimulation in CLP animals. Peritoneal macrophages could be grouped into LPM and SPM (23). Following infection or inflammation, the LPMs disappear with an increase in the population of SPM, with massive influx of inflammatory cells (38). And the reason behind this disappearance of LPMs is shown to be migration of such cells into the omentum (23). In line with previous reports, our data showed rapid decrease in the LPM number and frequency with moderate increase in the SPMs upon induction of CLP and endotoxemia. Under inflammatory conditions, SPMs could be generated by the differentiation of inflammatory monocytes (23). In our study, at 24 hours post-CLP surgery, a unique population with higher F4/80 and lower Ly-6C was observed in the peritoneum (data not shown), suggesting the differentiation of inflammatory monocytes to SPMs by 24 hours post-CLP surgery that was not observed in the endotoxemia animals.

As discussed earlier, the initial hyperinflammatory phase of the sepsis is followed by an immunosuppressive phase, and at later stages, both pro- and anti-inflammatory events could occur concomitantly (39). During later stages of sepsis, the apoptosis of the activated M1 macrophages and the polarization of macrophages to the M2 phenotype are shown to contribute to the immunosuppression (40). The results of our study show an upregulation of CD206 expression by F4/80⁺ cells in the lungs by 6 hours in the endotoxemia model and by 12 hours in the CLP animals. The frequencies and the total number of F4/80⁺Ly-6C^loCD206^hi cells also increased at 12 hours in both the models. Moreover, these cells also reduced the expression of MHCII on their surface, showing a possibility of regulatory phenotype in the lungs in both the models. The number of F4/80⁺Ly-6C^loCD206^hi macrophages was significantly lower in the inflamed peritoneum of CLP and endotoxemic animals compared with naive peritoneum and inflamed lungs. These data on a differential pattern of the presence of CD206-expressing macrophages in the lungs and peritoneum correlated with the inflammatory environment in the respective tissues. For example, the data from the cytokine mRNA expression study indicated an inflammatory environment in the peritoneum as evident from the presence of higher IL-6 and TNF-α and lower IL-10 transcripts in the peritoneum compared with inflamed lungs. These data along with the abundance of the regulatory phenotype of macrophages in the lungs suggest an immunosuppressive microenvironment in the lungs. Interestingly, CD206^hi macrophages showed higher intracellular IL-6 compared with CD206^lo cells in the lungs at 20 hours p.i. of sepsis. However, the CD206^hi macrophages showed an increased level of anti-inflammatory cytokine IL-10 compared with CD206^lo macrophages. This may mean that the macrophages may polarize into an anti-inflammatory type after performing their inflammatory functions, as we observed higher IL-6 expression in the CD206^hi macrophages. Moreover, previous studies have shown IL-6 to promote anti-inflammatory responses in local or systemic acute inflammatory conditions (41, 42). In contrast to the lungs, the frequency and number of F4/80⁺Ly-6C^loCD206^hi macrophages decreased in the peritoneum. Our intracellular cytokine staining did not show any difference in IL-6, IL-12, and TNF-α level in the peritoneum, whereas the level of IL-10 was increased in the CD206^hi cells in both peritoneum and lungs.

Integrins are the important cell adhesion molecules that mediate the migration and functions of immune cells. Previous studies have advocated for altered migration of innate immune cells, such as neutrophils, and specific upregulation of integrin α₃ on hyperactive neutrophils during sepsis (43). Results from this study shows that, whereas the Ly-6C^hi monocytes predominantly expressed integrin α₄β₁ in both peritoneum and lungs, the SPMs in peritoneum and F4/80⁺ macrophages in the lungs showed higher expression of α₅β₁ and α_vβ₃ integrins in both CLP and endotoxemia. Whereas macrophages in the CLP animals showed gradual increase in the expression of integrins over time, macrophages in the endotoxemic animals showed upregulation in integrin expression as early as six hours. Such differences may be attributed to the differences in the nature of inflammatory stimulation in both models. CLP induces a polymicrobial septic response, whereas LPS triggers the TLR-4 receptors on the innate immune cells. Similarly, in the lungs, the F4/80⁺CD206^hi cells showed higher integrin α_v expression compared with the F4/80⁺CD206^lo cells in the lungs and peritoneum of both models. Thus, these findings are suggestive of association of integrin α₄ and α₅ with initial infiltration and inflammatory process, whereas the integrin α_v may be associated with further polarization of macrophages.

In support of the hypothesis, the bioinformatics analysis using STRING showed the interaction of integrin α₄, and α₅ with tyrosine kinase Src with a high confidence level. Consistent with our in silico findings, the level of p-SRC was higher in the F4/80⁺α₄^hi and F4/80⁺ α₅^hi macrophages in the septic animals compared with the naive animals. FAK and Src activation are important for the integrin mediated functional response and are also involved in the recruitment and activation of monocytes and macrophages, and their inhibition is shown to diminish tissue injury in several inflammatory diseases (44, 45). Downstream of integrins and TLRs, Src is involved in the signaling cascade responsible for cell adhesion, migration, and secretion of inflammatory proteins (31). Thus, our data suggest that the α₄β₁ and α₅β₁ integrins may be involved in the initial inflammatory functions of macrophages.

Our study showed higher integrin α_v expression by F4/80⁺CD206^hi macrophages, and bioinformatics analysis of both human and mouse showed interaction of integrin α_v with TGF-β and its receptors TGFBR1 and TGFBR2, which are involved in the resolution processes (46). TGF-β is also shown to suppress the monocyte and macrophage activity in a way similar to IL-10 (47). TGF-β is also known to inhibit the proinflammatory response of macrophages (48, 49). Previous reports showed involvement of several integrins in controlling TGF-β both in the Smad-dependent and Smad-independent manner (50). Consistent with the bioinformatics findings, our TGFBR1 expression data showed higher expression of TGFBR1 in F4/80⁺α_v^hi and F4/80⁺CD206^hi macrophages in both peritoneum and lungs. The number of F4/80⁺α_v^hiTGFBR1^hi macrophages was also relatively higher in the lungs of endotoxemic animals than naive animals, whereas in the peritoneum of endotoxemic animals, the number of such cells decreased significantly. These findings imply the dominance of immunosuppressive phenotype of macrophages in the lungs of septic animals and a correlation of integrin α_v and TGFBR1 signaling in the anti-inflammatory functions of macrophages.

In conclusion, our study shows striking differences in the migration and distribution pattern of inflammatory immune cells in two widely used murine models of sepsis. To the best of our knowledge, this is the first study to show innate immune cell mobilization and distribution along with expression kinetics of important β₁ and β₃ integrins on monocytes and macrophages during different stages of murine sepsis. Our findings provide a significant contribution to the understanding of the complex pathobiology of sepsis.

The authors have no financial conflicts of interest.