Inflammatory monocytes promote granuloma control of Yersinia infection

Abstract

Granulomas are organized immune cell aggregates formed in response to chronic infection or antigen persistence. The bacterial pathogen Yersinia pseudotuberculosis (Yp) blocks innate inflammatory signaling and immune defense, inducing neutrophil-rich pyogranulomas within lymphoid tissues. Here, we uncover that Yp also triggers pyogranuloma formation within the murine intestinal mucosa. Mice lacking circulating monocytes fail to form defined pyogranulomas, have defects in neutrophil activation, and succumb to Yp infection. Yersinia lacking virulence factors that target actin polymerization to block phagocytosis and reactive oxygen burst do not induce pyogranulomas, indicating that intestinal pyogranulomas form in response to Yp disruption of cytoskeletal dynamics. Notably, mutation of the virulence factor YopH restores pyogranuloma formation and control of Yp in mice lacking circulating monocytes, demonstrating that monocytes override YopH-dependent blockade of innate immune defense. This work reveals an unappreciated site of Yersinia intestinal invasion and defines host and pathogen drivers of intestinal granuloma formation.

Microbial pathogens utilize diverse mechanisms to subvert host immunity in order to replicate and spread to new hosts. While acute infections can be cleared rapidly by the immune system, some pathogens evade immune defenses to cause chronic disease. Chronic infections often result in formation of structures termed granulomas that limit pathogen dissemination and tissue damage¹. Granulomas are characterized by the presence of activated phagocytes, notably monocytes and macrophages, and form in response to a wide variety of infections². Monocytes are rapidly recruited to infected tissues where they produce inflammatory cytokines and anti-microbial effector molecules, contributing to defense against multiple pathogens^3–6. Some pathogens, however, exploit monocytes as a means of dissemination, including Salmonella enterica, Yersinia pestis, and Mycobacterium species^7–9. The pathogen-specific signals that induce granuloma formation remain poorly defined.

Enteropathogenic Yersinia, including Y. pseudotuberculosis (Yp) and Y. enterocolitica (Ye), cause self-limiting gastroenteritis and mesenteric lymphadenopathy following enteric infection^10,11. In immune-compromised patients, however, bacteria disseminate and cause a systemic plague-like disease, indicating that the intestinal immune system is critical for control of acute infection. Indeed, the intestine constitutes a bottleneck against Yersinia dissemination, as bacteria in systemic organs are thought to originate predominantly from the intestinal lumen rather than gut-associated lymphoid tissues¹². A hallmark of Yersinia infections is the presence of chronic pyogranulomas (PG) in lymphoid tissue, characterized by nodular infiltrates of activated monocytes and macrophages surrounding a core of activated neutrophils¹³. Notably, the contribution of monocytes to pyogranuloma formation and their role in Yersinia restriction are unclear^14,15.

Here we report that pyogranulomas form acutely in the murine intestinal mucosa during enteric Yersinia infection. PG are enriched in neutrophils and inflammatory monocytes, and contain live bacteria at levels comparable to Peyer’s patches (PP). Notably, CCR2-deficient mice, which lack circulating inflammatory monocytes^16,17, form disorganized necrosuppurative lesions rather than defined PG, are unable to contain bacteria within the lesions, and succumb rapidly to infection. Moreover, mice lacking circulating monocytes exhibit reduced levels of IL-1 cytokines and surface expression of the neutrophil activation marker CD11b within intestinal pyogranulomas. Yp lacking either the virulence plasmid encoding the type III secreted Yersinia Outer Proteins (Yops), or lacking Yops that block phagocytosis and the reactive oxygen burst, do not induce detectable pyogranulomas, indicating that pyogranulomas are induced in response to Yersinia blockade of innate immune defense. Notably, CCR2-deficient mice infected with bacteria lacking the virulence factor YopH, which blocks actin cytoskeleton dynamics, were able to restrict bacterial burdens and form defined granulomatous lesions, accompanied by restored neutrophil CD11b surface expression. Neutrophil depletion in CCR2-deficient animals abrogates control of YopH-mutant Yersinia, demonstrating that inflammatory monocytes overcome YopH-mediated disruption of neutrophil function. Altogether, our study identifies an unappreciated site of Yersinia colonization within the murine intestinal mucosa, and reveals an essential function for inflammatory monocytes in maintainance of PG architecture during Yp infection.

Results

Intestinal pyogranulomas form upon oral Yersinia infection

Yersinia pseudotuberculosis colonizes gut-associated lymphoid tissues, resulting in acute pyogranuloma (PG) formation following oral infection¹⁸. Interactions between Yersinia and immune cells within systemic tissues have been extensively documented^19–21. In our efforts to dissect intestinal immune responses to Yp, we observed numerous macroscopically visible nodular lesions in the gastrointestinal tract five days post-infection, which appeared as punctate areas of increased opacity (Fig. 1a). Lesions were most prevalent in the jejunum and ileum, ranged in number from two to over forty in individual mice, and also formed in response to Yersinia enterocolitica infection (Fig. 1b, c). Histology of Yp-infected intestines revealed focal inflammation characterized by crypt hyperplasia, edema, and submucosal to transmural cellular infiltration, whereas non-lesional areas of infected intestines appeared largely unaffected (Fig. 1d). The lesions contained infiltrates of macrophages and neutrophils surrounding colonies of coccobacilli, similar to structures that we and others observed in lymphoid tissues and have termed pyogranulomas (Fig. 1e)^15,18,22. Consistently, flow-cytometric analysis of intestinal punch biopsies containing pyogranulomas (PG+), adjacent non-granulomatous tissue (PG−), and uninfected control tissue (uninf) revealed that neutrophils were the most enriched cell type within PG+ tissue, followed by macrophages and inflammatory monocytes (Fig. 1f, g). We also observed increases in eosinophils, dendritic cells, and CD4⁺ T cells in PG+ tissue, although the relative frequencies of these populations were decreased, due to even larger increases in neutrophils, monocytes, and macrophages (Extended Data Fig. 1a, b).

Figure 1. Intestinal pyogranulomas form upon oral Yersinia infection.

(a) Small-intestinal segments from uninfected and Yp-infected mice with arrows depicting lesions, and a magnified lesion with dotted circle depicting size of punch biopsies. Scale bars = 3 mm (left) and 0.5 mm (right). Representative of >3 independent experiments.

(b) Frequency distribution of lesions along the intestine, with graphical key of anatomical segments. Each colored bar represents the mean frequency of lesions in a given segment (n = 20 mice). Only mice with >9 total lesions (>80% of mice) were included. Pooled from three independent experiments.

(c) Quantification of total intestinal lesions at day 5 post-infection with Yp or Ye. Each circle represents one mouse (n = 16-17). Lines represent median. Pooled from three independent experiments.

(d) H&E-stained paraffin-embedded longitudinal small-intestinal sections from uninf and Yp-infected mice. Scale bars = 200 μm. a = lesion, b = crypt hyperplasia, c = submucosal inflammation, d = edema. Representative of three independent experiments.

(e) H&E-stained paraffin-embedded small-intestinal pyogranuloma depicting an encircled bacterial colony. Scale bar = 10 μm. Representative of three independent experiments.

(f) Flow-cytometry plots identifying CD11b⁺Ly-6G⁺ neutrophils, CD11b⁺CD64⁺Ly-6C⁺ monocytes, and CD11b⁺CD64⁺Ly-6C⁻MHC-II⁺ macrophages in small-intestinal tissue. Representative of four independent experiments.

(g) Frequency and total number of neutrophils, monocytes and macrophages in small-intestinal tissue. Each circle represents one mouse (n = 15-19). Lines represent median. Pooled from four independent experiments.

(h) Fluorescently labeled small-intestinal pyogranuloma from a Ccr2^gfp/+ mouse. White (Yp-mCherry), magenta (Ly-6G-AF647), green (CCR2-GFP). Scale bars = 100 μm. Representative of two independent experiments.

(i) Bacterial burdens in small-intestinal tissue. Each circle represents one mouse (n = 24-25). Lines represent geometric mean. Dotted line represents detection limit. Pooled from four independent experiments.

(j) Cumulative bacterial burdens in PG+ tissue and Peyer’s patches. Each circle represents one mouse (n = 29-30). Lines represent geometric mean. Pooled from five independent experiments.

Wilcoxon test (two-tailed) was performed for paired analyses (PG− vs PG+). Mann-Whitney U test (two-tailed) was performed for remaining statistical analyses. * (p < 0.05), ** (p < 0.01), **** (p < 0.0001), NS (not significant, p > 0.05), ND (not detected).

In lymphoid tissues Yp pyogranulomas consist of a central bacterial colony, surrounded by neutrophils, which are bordered in turn by monocytes and macrophages^15,22. Confocal microscopy demonstrated that intestinal PG also contained a central Yp colony surrounded by a dense population of Ly-6G⁺ neutrophils (Fig. 1h). Interestingly, in contrast to lymphoid tissues, CCR2⁺ monocytes and macrophages formed a mesh-like network of cells that both overlapped with and bordered the neutrophils (Fig. 1h). Consistent with the presence of bacterial colonies detected by histology and fluorescence microscopy, PG+ tissue harbored high numbers of viable bacteria comparable to that found in the Peyer’s patches (Fig. 1i, j). Since PP are a major entry point and replicative niche for enteropathogenic Yersinia following oral innoculation^23,24, altogether, these findings reveal intestinal pyogranulomas as a previously unappreciated location of Yersinia invasion within the intestinal mucosa and a potential site of bacterial restriction or dissemination.

Intestinal inflammation is spatially restricted to PGs

To test whether PGs exhibit location-specific inflammatory responses, we next performed RNA sequencing of pyogranulomas, adjacent non-pyogranuloma tissue, and uninfected tissue. Principal component analysis showed distinct clustering by sample type (Fig. 2a), and comparison of PG+ and PG− samples revealed 355 upregulated and 363 downregulated genes (Fig. 2b). Top upregulated genes included granulocyte- and monocyte-recruiting chemokines (Cxcl1, Cxcl2, Cxcl3, Cxcl5), pro-inflammatory cytokines (Il1b, Il22), metal-sequestration proteins (S100a8, S100a9), and matrix metalloproteases (Mmp3) (Fig. 2c). Gene-ontology analysis indicated that chemotaxis of myeloid cells and defense against bacterial pathogens predominated the top upregulated responses within PG+ biopsies (Fig. 2d, Extended Data Table 1). These responses were strikingly similar to previously reported Yp-infected PP²⁵. Consistently, gene set-enrichment analysis indicated that a set of 50 upregulated genes previously reported in Yp-infected PP were also enriched in PG+ samples (Fig. 2e). Furthermore, protein levels of the pro-inflammatory cytokines IL-1α, IL-1β, IL-6, TNF, and CCL2 were significantly elevated within PG+ biopsies (Fig. 2f). Consistent with histology and microscopy, the inflammatory transcriptional response was localized to PG+ tissue, as PG− samples did not exhibit enrichment of genes or ontology terms related to myeloid cell migration or innate immune-cell activation (Extended Data Fig. 2, Extended Data Table 2). Likewise, production of pro-inflammatory cytokines was not detected in PG- tissue (Fig. 2e). Altogether, these data indicate that the pro-inflammatory response to Yp infection in the gut mucosa is spatially restricted to PG.

Figure 2. Intestinal inflammation is spatially restricted to PGs.

(a) Principal component analysis of PG+ (pink), PG− (green), and uninf (gray) samples at day 5 post-infection. Five biopsies were pooled per mouse.

(b) Heatmap of all differentially expressed genes in PG+ compared to PG− samples. False discovery rate < 0.05 using Benjamini-Hochberg procedure. Pink and orange bars denote two clusters grouped based on Pearson correlation.

(c) Heatmap of top 30 significantly upregulated genes in PG+ compared to PG− samples in descending order by fold change. False discovery rate < 0.05 using Benjamini-Hochberg procedure.

(d) Gene ontology analysis of top 30 upregulated genes by fold change only in PG+ compared to PG− samples.

(e) Gene set enrichment analysis (GSEA) of top 50 upregulated genes in Yp-infected Peyer’s patches. NES = normalized enrichment score; FDR = false discovery rate.

(f) Cytokine levels in homogenates of tissue punch biopsies at day 5 post-infection. Lines represent group mean. Each circle represents one mouse (n = 6-9). Statistical analysis by Mann-Whitney U test (two-tailed). * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05). Data from one or two pooled independent experiments.

Inflammatory monocytes maintain PGs to restrict infection

Inflammatory monocytes promote host defense by differentiating into phagocytes and antigen presenting cells^4,6,26,27, producing pro-inflammatory mediators³, and modulating other immune cell functions^5,28. Monocyte-derived cells can also promote pathogen replication or dissemination to new sites^7,29. Mice lacking the chemokine receptor CCR2 have a ten-fold reduction in circulating monocytes due to defective bone marrow-egress^16,17. Monocytes are rapidly recruited to intestinal PGs (Fig. 1) and sites of systemic Yp infection^15,18,22, raising the question of their role in Yp infection. CCR2-deficiency has been associated with more rapid bacterial clearance from the MLN following enteric infection¹⁵, but increased susceptibility to intravenous Yp infection¹⁴. Interestingly, we found that while Ccr2^gfp/gfp mice, which lack circulating mononcytes due to lack of CCR2 expression^30,31, had similar overall numbers of macroscopic intestinal lesions as WT mice (Fig. 3a, Extended Data Fig. 3a), their lesions had a disorganized appearance, and exhibited central caseation with tissue necrosis (Fig. 3b). In contrast to WT PG which exhibited robust inflammatory infiltrates and a defined cellular organization encapsulating central bacterial colonies, Ccr2^gfp/gfp intestinal lesions contained expanded coalescing bacterial colonies with limited immune cell recruitment (Fig. 3b, c). PP of Ccr2^gfp/gfp mice had similarly disorganized lesions with central tissue necrosis, suggesting that monocytes are required to establish or maintain organized PGs during Yp infection (Extended Data Fig. 3b). Two independent CCR2-deficient mouse lines showed significantly higher bacterial burdens in PG+ and PG− tissues compared to WT mice (Fig. 3d, Extended Data Fig. 3c). Moreover, acute depletion of monocytes in WT mice with anti-CCR2 specific antibodies resulted in increased PG− and PG+ bacterial burden (Extended Data Fig. 3d, e), demonstrating that the requirement for monocytes in control of Yp is not due to developmental defects in CCR2-deficient mice. Interestingly, at day 3 post-infection, Ccr2^gfp/gfp mice had elevated bacterial burdens in PG− tissue but not PG+ tissue and did not exhibit overt signs of tissue necrosis in PGs (Extended Data Fig. 3f, g), indicating that the defect in bacterial control and PG architecture develops between day 3 and 5. Together, these data suggest that monocytes enable maintenance of PG architecture and restrict Yp within intestinal PG.

Figure 3. Inflammatory monocytes maintain PGs to restrict infection.

(a) Quantification of total intestinal lesions at day 5 post-infection. Each circle represents one mouse (n = 20). Lines represent median. Pooled data from three independent experiments.

(b) H&E-stained small intestinal sections from Yp-infected mice at day 5 post-infection. 1 denotes the Yersinia microcolony, 2 denotes necrotic tissue. Scale bars = 250 μm (top) and 50 μm (bottom). Representative images of two independent experiments.

(c) Histopathological scores of small intestinal tissue at day 5 post-infection. Each mouse was given a score between 0-4 (healthy-severe) for presence of bacterial colonies free from immune cell infiltrate. Each circle represents one mouse (n = 7-9 for uninfected, 9-16 for Yp). Lines represent median. Pooled data from two independent experiments.

(d) Bacterial burdens in PG− and PG+ tissue at day 5 post-infection. Each circle represents the mean Yp-CFU of 3-5 pooled punch biopsies from one mouse (n = 31-37). Lines represent geometric mean. Pooled data from six independent experiments.

(e) Total numbers and frequencies of indicated cells in small intestinal uninf, PG−, and PG+ tissue at day 5 post-infection. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 24). Lines represent median. Pooled data from five independent experiments.

(f) Fluorescently labeled PG+ tissue from Yp-infected Ccr2^gfp/+ (top) and Ccr2^gfp/gfp (bottom) mice at day 5 post-infection. Scale bars = 100 μm. Representative images of two independent experiments.

(g) PG+ neutrophil surface CD11b expression at day 5 post-infection. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 6). Lines represent median. Representative of four independent experiments.

(h) Cytokine levels in homogenates of tissue punch biopsies were measured by cytometric bead array at day 5 post-infection. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 18-21). Lines represent median. Pooled data from three independent experiments.

Mann-Whitney U test (two-tailed) was performed for all statistical analyses. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05)

Consistent with their reduced overall cellularity, Ccr2^gfp/gfp lesions exhibited decreased frequency and numbers of viable CD45⁺ hematopoietic cells compared to WT PG (Fig. 3e). Consistent with the important role of CCR2 in promoting monocyte egress from bone marrow^16,17, Ccr2^gfp/gfp intestinal lesions contained significantly lower numbers of monocytes, macrophages, and dendritic cells compared with PG from WT mice (Fig. 3e, Extended Data Fig. 4a). Notably, CCR2 deficiency did not impact T or B cell numbers in intestinal PGs, indicating that the defect in enteric control of Yp in CCR2-deficient mice was independent of adaptive immune cells (Extended Data Fig. 4a). Ccr2^gfp/gfp intestinal lesions also showed a trend toward reduced neutrophil numbers (Fig. 3e), suggesting that monocytes promote recruitment, retention, or survival of neutrophils within intestinal PG during Yp infection. This defect was specific to the intestinal lesions, as we observed similar frequencies of neutrophils in the MLN and spleen (Extended Data Fig. 4b). Immunofluorescence microscopy of the lesions indicated that neutrophils were unable to effectively contain Yp in the absence of monocytes, as the bacterial colony expanded outside the range of the neutrophil marker Ly-6G. In contrast, Yp was fully encapsulated by neutrophils and CCR2⁺ cells within WT PGs (Fig. 3f).

Intriguingly, surface expression of the integrin CD11b, a well-established marker of neutrophil activation^32–34, was significantly reduced in both PGs and MLN of Ccr2^gfp/gfp mice compared to WT counterparts (Fig. 3g, Extended Data Fig. 4c), suggesting a defect in neutrophil activation in the absence of monocytes. CD11b is present on the neutrophil cell surface and membranes of intracellular granules, and increased CD11b surface expression occurs in numerous inflammatory settings^35–37. Notably, PG neutrophils in CCR2-deficient mice exhibited increased intracellular CD11b levels, suggesting that translocation of CD11b from intracellular granules to the cell surface is defective in the absence of monocytes (Extended Data Fig. 4d). Interestingly, total IL-1α and IL-1β levels were significantly reduced in Ccr2^gfp/gfp PGs, whereas other pro-inflammatory cytokines were unaffected (Fig. 3h, Extended Data Fig. 4e), indicating that monocytes or monocyte-derived cells specifically produce IL-1, or specifically promote IL-1 production by other cells within intestinal PGs. Notably, intracellular levels of IL-1 cytokines, TNF, and lipocalin were unaffected in PG neutrophils in Ccr2^gfp/gfp mice (Extended Data Fig. 4f), suggesting that inflammatory monocytes do not regulate neutrophil-intrinsic inflammatory cytokine and antimicrobial protein production. Altogether, these results demonstrate that inflammatory monocytes or monocyte-derived cells promote maintenance of functional granulomas that limit intestinal bacterial replication and dissemination.

Inflammatory monocytes control systemic Yersinia

Following systemic dissemination, Yp colonizes and induces PGs in lymphoid tissues^15,18,22. Critically, both CCR2-deficient and anti-CCR2 depleted mice had significantly higher Yp burdens in the MLN and systemic organs (Fig. 4a, Extended Data Fig. 5a, 5b). Similar to intestinal tissue, systemic bacterial burdens in Ccr2^gfp/gfp mice were unaffected at day 3 post-infection, indicating that this defect in control develops between day 3 and 5 (Extended Data Fig. 5c). Consistent with our findings that monocytes were required for maintenance of intestinal pyogranuloma architecture, infected Ccr2^gfp/gfp spleens exhibited widespread tissue necrosis, free bacterial colonies, and sparse immune cell recruitment, in contrast to WT spleens where neutrophils and monocytes effectively encapsulated Yp microcolonies within organized PG (Fig. 4b, c). Notably, mice lacking CCR2 succumbed rapidly to acute infection (Fig. 4d, Extended Data Fig. 5d). Importantly, co-housed littermate Ccr2^+/+ and Ccr2^gfp/+ mice were equally resistant to Yp infection while Ccr2^gfp/gfp littermates succumbed (Fig. 4e), indicating that increased susceptibility of Ccr2^gfp/gfp to Yp infection is not due to differences in composition of a vertically-transmitted intestinal microbiota. Collectively, our findings demonstrate that inflammatory monocytes maintain organized pyogranulomas in host tissues, thereby limiting tissue necrosis and systemic bacterial dissemination, ultimately enabling bacterial control and host survival following oral Yp infection.

Figure 4. Inflammatory monocytes control systemic Yersinia.

(a) Bacterial burdens in indicated organs at day 5 post-infection. Each circle represents one mouse (n = 20-25). Lines represent geometric mean. Pooled data from four independent experiments.

(b) H&E-stained paraffin-embedded longitudinal spleen sections from WT and Ccr2^gfp/gfp mice at day 5 post-infection. Dashed circle denotes area with bacterial microcolonies and neutrophils. Scale bars = 500 μm (top) and 50 μm (top). Representative images of two independent experiments.

(c) Histopathological scores of spleens from uninfected and Yp-infected mice at day 5 post-infection. Each mouse was given a score between 0-4 (healthy-severe) for presence of bacterial colonies free from immune cell infiltrate. Each circle represents one mouse (n = 9 for uninfected, 16 for Yp). Lines represent median. Pooled data from two independent experiments.

(d) Survival of infected WT (n=16) and Ccr2^gfp/gfp (n=13) mice. Pooled data from two independent experiments.

(e) Survival of infected littermate wild-type Ccr2^+/+ (n=20), heterozygous Ccr2^gfp/+ (n=19), and homozygous Ccr2^gfp/gfp (n=15) mice. Pooled data from three independent experiments.

Statistical analyses by (a, c) Mann-Whitney U test (two-tailed) and (d, e) Mantel-Cox test. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Yersinia virulence factors induce intestinal pyogranulomas

Yersinia inject Yersinia Outer Proteins (Yops), which are encoded on a virulence plasmid (pYV), into target cells³⁸ to block phagocytosis, production of reactive oxygen species (ROS), and degranulation²⁰. Interestingly, Yp lacking the virulence plasmid (pYV-) still colonize the intestine during the acute phase of infection without causing disease³⁹. Since granulomas form in response to pathogens that thwart immune defenses, we hypothesized that intestinal pyogranulomas may be triggered by the activity of Yp effector proteins. Indeed, even at a ten-fold higher infectious dose, we did not observe intestinal lesions at day 5 post-infection with pYV- bacteria (Fig. 5a), despite detectable (although reduced) intestinal colonization (Extended Data Fig. 6a).

Figure 5. Yersinia virulence factors induce intestinal pyogranulomas.

(a) Quantification of total intestinal lesions at day 5 post-Yp infection. Each symbol represents one mouse (n = 10-43). Lines represent median. Pooled from 2-6 independent experiments.

(b) H&E-stained paraffin-embedded transverse small intestinal sections from WT and Ccr2^gfp/gfp mice infected with WT or YopH^R409A Yp at day 5 post infection depicting encircled bacterial colonies. Scale bars = 100 μm. Representative of two independent experiments.

(c) Frequency and total number of neutrophils in small-intestinal PG+ tissue at day 5 post WT or YopH^R409A Yp infection. Each circle represents one mouse (n = 12-14). Lines represent median. Pooled from three independent experiments.

(d) Mean fluorescent intensity (MFI) of surface CD11b expression on neutrophils in PG+ tissue at day 5 post WT or YopH^R409A Yp infection. Each symbol represents one mouse (n = 7-10). Lines represent median. Pooled from two independent experiments.

(e) Bacterial burdens in small-intestinal PG− and PG+ tissue at day 5 post WT or YopH^R409A Yp infection. Each symbol represents one mouse (n = 25-26). Lines represent geometric mean. Pooled from four independent experiments.

(f) Bacterial burdens in indicated organs at day 5 post WT or YopH^R409A Yp infection. Each symbol represents one mouse (n = 15-22). Lines represent geometric mean. Pooled from four independent experiments.

(g) Survival of WT mice infected with WT (n = 20) or YopH^R409A (n = 20) Yp and Ccr2^−/− mice infected with WT (n = 10) or YopH^R409A (n = 17) Yp Pooled from two independent experiments.

Statistical analyses by (a, c, d, e, f) Kruskal-Wallis test with Dunn’s post-test, and (g) Mantel Cox test. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

pYV- bacteria still induced monocyte recruitment to both the intestinal mucosa and MLN, indicating that lack of detectable PG was not due to an absence of immune infiltration (Extended Data Fig. 6b). However, neutrophil accumulation was absent upon pYV- infection, demonstrating that neutrophil recruitment occurs in response to Yp virulence (Extended Data Fig. 6c). Although comparable in the intestinal mucosa, pYV- bacterial burdens were reduced in other gut-associated and systemic lymphoid tissues (Extended Data Fig. 6d), illustrating that intestinal PG formation is dispensable for control of pYV- Yersinia.

Intestinal lesions formed at WT levels in mice infected with Yp individually deficient in either YopM or YopJ enzymatic activity, which block pyrin inflammasome assembly or NF-κB and MAPK signaling^40–42, respectively, suggesting that neither YopM nor YopJ are singly required for pyogranuloma formation (Extended Data Fig. 6e). Several Yops function together to disrupt the actin cytoskeleton, thereby blocking phagocytosis and the reactive oxygen burst²⁰. These Yops (E, H, and T), have partially overlapping functions and can compensate for one another in certain settings²⁰. Notably, Yp with combined point mutations in each of the catalytic residues of these Yops (YopE^R144A, YopT^C139A, and YopH^R409A) did not induce intestinal lesions and had burdens similar to pYV- infection (Fig. 5a and Extended Data Fig. 6a, d). In contrast, bacteria lacking YopE and YopT, or YopE alone, induced WT numbers of intestinal lesions (Fig. 5a) and were only attenuated in systemic organs (Extended Data Fig. 6d), indicating that YopH is sufficient, in the absence of YopE and YopT, to induce pyogranuloma formation. Interestingly, YopH^R409A mutants that lack YopH tyrosine phosphatase activity induced fewer pyogranulomas than wild-type bacteria (Fig. 5a), suggesting that YopH is sufficient and partially responsible for pyogranuloma formation. In addition, MLN colonization was abrogated in the absence of YopH activity, whereas YopE was dispensable for MLN colonization (Extended Data Fig. 6d). Intriguingly, bacteria lacking both YopE and YopH (yopEH) induced no detectable lesions (Fig. 5a) and showed similar levels of colonization to yopETH bacteria (Extended Data Fig. 6a, d), suggesting that the host response to YopE and YopH drives pyogranuloma formation.

YopH blocks innate cell phagocytosis and ROS production^20,43–46. YopH-deficient Yp are therefore more susceptible to neutrophil killing in vitro and are attenuated in vivo^47–49. Consequently, neutrophil depletion increases susceptibility to Yp and restores virulence to YopH-mutant bacteria^45,46, indicating that neutrophil defenses are an important target of YopH in vivo. However, whether monocytes combat YopH-dependent blockade of neutrophil function is unknown. As in WT mice, we found that YopH-mutant bacteria induced significantly fewer intestinal lesions in Ccr2^gfp/gfp mice than WT bacteria (Extended Data Fig. 6f). Strikingly, despite the lack of monocytes, YopH^R409A infection of CCR2-deficient mice restored neutrophil recruitment and induced neutrophil-rich lesions without necrosis (Fig. 5b, c, Extended Data Fig. 6g). Intriguingly, CD11b surface expression was restored in PG+ tissue of CCR2-deficient animals infected with YopH^R409A bacteria, indicating that YopH limits neutrophil activation in the absence of monocytes (Fig. 5d). Furthermore, CCR2-deficient mice effectively controlled bacterial burdens in intestinal and systemic tissues following infection with YopH^R409A bacteria (Fig. 5e, f and Extended Data Fig. 6h). Consistently, while CCR2-deficient mice rapidly succumbed to WT Yp, they survived infection with YopH^R409A bacteria at levels similarl to WT mice infected with WT Yp, suggesting that monocytes control Yersinia infection by overcoming the virulence of YopH (Fig. 5g). Altogether, these data indicate that intestinal pyogranulomas form in response to Yersinia disruption of innate immune defense, and that monocytes counteract YopH activity in vivo.

Neutrophils control YopH-deficient Yersinia in absence of monocytes

While monocytes are dispensable for control of YopH-deficient Yp, the cell type that mediates this control in the absence of monocytes is unknown. Neutrophil recruitment to PGs was restored during infection of Ccr2^gfp/gfp mice with YopH-deficient Yp, raising the possibility that neutrophils mediate control of Yersinia in the absence of monocytes. Since Ccr2^gfp/gfp mice largely lack circulating monocytes, anti-Gr-1, which depletes both neutrophils and monocytes and is highly efficient at depleting neutrophils in infectious or inflammatory settings^50,51, allowed us to interrogate the role of neutrophils in control of YopH-mutant Yp (Fig. 6a, Extended Data Fig. 7a). Strikingly, Ccr2^gfp/gfp mice injected with anti-Gr-1 had elevated burdens of YopH-mutant Yp in PG− tissue (Fig. 6b). Anti-Gr-1 did not further reduce blood monocyte frequencies in Ccr2^gfp/gfp mice (Extended Data Fig. 7b), indicating that increased susceptibility of these mice was not attributable to depletion of remaining Ly-6C⁺ monocytes. CCR2-deficient mice infected with YopH^R409A bacteria had very few detectable PG (Fig. 6c), making it difficult to robustly analyze bacterial burdens in this tissue. Intriguingly, Ccr2^gfp/gfp mice injected with anti-Gr-1 had elevated numbers of macroscopic intestinal lesions during infection with YopH^R409A bacteria (Fig. 6c). Moreover, in peripheral organs, Ccr2^gfp/gfp mice injected with anti-Gr-1 had elevated bacterial burdens following YopH-mutant infection (Fig. 6d), demonstrating that in the absence of monocytes, neutrophils play a key role in control of YopH-deficient bacteria. Notably, mice infected with WT bacteria did not exhibit increased systemic bacterial burdens upon neutrophil depletion (Fig. 6d), consistent with YopH blockade of neutrophil functions. Collectively, these data indicate that neutrophils control YopH-deficient Yp in the absence of monocytes.

Figure 6. Neutrophils control YopH-deficient Yersinia in absence of monocytes.

(a) Frequency of neutrophils in blood at day 5 post infection was determined by flow cytometry. Each symbol represents one mouse (n = 10-11). Lines represent median. Pooled data from three independent experiments.

(b) Bacterial burdens in small-intestinal PG− tissue at day 5 post WT or YopH^R409A Yp infection. Each symbol represents one mouse (n = 10-12). Lines represent geometric mean. Pooled data from three independent experiments.

(c) Quantification of total number of intestinal lesions at day 5 post infection. Each symbol represents one mouse (n = 10-12). Lines represent median. Pooled data from three independent experiments.

(d) Bacterial burdens in indicated organs at day 5 post WT or YopH^R409A Yp infection. Each symbol represents one mouse (n = 10-12). Lines represent geometric mean. Pooled data from three independent experiments.

All statistical analyses by Kruskal-Wallis test with Dunn’s post-test. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Discussion

Granulomas are a conserved response to persistent or long-term infectious and non-infectious stimuli^1,2. The natural rodent and human pathogen Yersinia pseudotuberculosis induces formation of neutrophil-rich pyogranulomas (PG) in infected lymphoid tissues^11,13. Here, we report that PG form throughout the intestine during murine enteropathogenic Yersinia infection. Peyer’s patches are considered the primary site of Yersinia intestinal infection. However, intestinal pyogranulomas harbored a similar total bacterial burden as the PPs, suggesting that PGs are a previously unrecognized niche for enteropathogenic Yersinia intestinal colonization. Notably, wild-type mice almost entirely restricted intestinal Yp and inflammation to these granulomatous foci. The transcriptional profile of PGs was similar to Yp-infected Peyer’s patches²⁵, indicating a shared response to intestinal Yp infection driven by recruitment and activation of innate immune cells, which may serve to limit bacterial spread and tissue damage within the gut mucosa.

While initial formation of PGs was similar in CCR2-deficient and WT mice, monocytes were critical for maintaining architectural integrity of PGs and bacterial control in the intestine and deeper tissues during Yersinia infection, as monocyte deficiency was associated with widespread tissue necrosis and elevated bacterial burdens in intestinal and systemic tissues that occurred between day 3 and day 5. Interestingly, Y. pestis, which causes more severe disease than Y. pseudotuberculosis, is commonly associated with necrotic and necrosuppurative lesions⁵², suggesting that Y. pestis has evolved to overcome monocyte-mediated host defense and pyogranuloma formation in order to enhance systemic dissemination and transmission.

Monocytes were previously reported to be dispensable for acute control of Yp following oral inoculation¹⁵. These studies employed a Ccr2^−/− mouse line^17,27 originally generated on the 129 mouse background and backcrossed to C57BL/6¹⁶, raising the possibility that distinct polymorphisms in this mouse line might account for this difference. Ccr2^gfp/gfp mice were generated directly on the C57BL/6J background³⁰, making it unlikely that our findings are due to immunologically impactful polymorphisms that co-segregate with the Ccr2 locus. Infection of co-housed littermates also demonstrated that susceptibility of CCR2-deficient mice is not due to differences in maternally-transmitted or environmentally-acquired microbiota. Additionally, acute depletion of monocytes with anti-CCR2 antibodies phenocopied CCR2-deficient mice in abrogating control of intestinal and systemic bacterial burdens, demonstrating that monocytes are critical for acute control of enteric Yersinia infection.

Lack of both YopE and YopH abrogated PG formation, whereas either effector alone was sufficient to induce PG formation, implying a key role for actin cytoskeleton disruption in PG formation. Notably, ablation of YopH activity alone significantly reduced numbers of intestinal PGs, indicating that YopH was predominantly responsible for PG formation. Precisely how YopH and YopE lead to pyogranuloma formation remains to be determined.

Monocytes and neutrophils comprise a large proportion of Yop-injected cells in vivo⁵³. YopH blocks neutrophil degranulation, ROS production, phagocytosis, and release of neutrophil extracellular traps^{20,45,46,49,54}. The virulence of YopH-deficient bacteria is restored upon neutrophil depletion or ROS-deficiency^45,46, suggesting that YopH potently targets neutrophil function in tissues during infection. In line with these observations, PGs of CCR2-deficient mice exhibited decreased numbers of neutrophils and reduced surface expression of the activation marker CD11b, which were restored in the setting of YopH deficiency. CD11b is expressed basally on the plasma membrane and in a subset of neutrophil granules, which are recruited to the surface upon neutrophil activation^35–37,55. Our findings imply that monocytes enhance neutrophil degranulation to overcome YopH-dependent blockade of their antimicrobial functions. Consistently, while monocytes were required for host survival in response to WT bacteria, they were dispensable for controlling YopH mutant Yp. Furthermore, depletion of neutrophils in YopH^R409A-infected CCR2-deficient mice restored bacterial virulence. Overall, our findings imply that monocytes play an important role in host defense by promoting maintenance of PG architecture and neutrophil function in the face of YopH-dependent blockade. Our study reveals a previously unappreciated site of Yersinia colonization within the intestine and provides insight into granuloma formation and function during Yersinia infection.

Methods

Animals

C57BL/6 wild-type and Ccr2^gfp/gfp mice³⁰ (in which insertion of EGFP into the translation initiation site of Ccr2 disrupts its expression) were acquired from the Jackson Laboratory and bred at the University of Pennsylvania. Ccr2r^−/− mice³¹ were provided by Dr. Sunny Shin (University of Pennsylvania). Unless specifically noted, all animals were bred by homozygous mating and housed separately by genotype. Mice of either sex between 8-12 weeks of age were used for all experiments. All animal studies were performed in accordance with University of Pennsylvania IACUC-approved protocols (protocol #804523).

Bacteria

Wild-type Yp (clinical isolate strain 32777, serogroup O1)⁵⁶ and isogenic mutants were provided by Dr. James Bliska (Dartmouth College). Ye (strain 8081, serogroup O8)⁵⁷ was provided by Dr. Stanley Falkow (Stanford University). Additional mutants lacking YopE (ΔyopE), enzymatic activity of YopH (YopH^R409A), both (denoted yopEH) or YopE/YopT/YopH (YopE^R144AT^C139AH^R409A, denoted yopETH) were generated by two-step allelic recombination as previously described⁵⁸ with plasmids provided by Dr. James Bliska. Fluorescent Yp (mCherry⁺) was generated from plasmids provided by Dr. Kimberly Davis (Johns Hopkins University).

Infections

Yp and Ye were cultured to stationary phase at 28°C and 250 rpm shaking for 16 hours in 2xYT broth supplemented with 2 μg/ml triclosan (Millipore Sigma). Mice were fasted for 16 hours and subsequently inoculated by oral gavage with 100-200 μl phosphate-buffered saline (PBS). All bacterial strains were administered at 2x10⁸ colony-forming units (CFU) per mouse with the exception of pYV-, yopETH, and yopEH which were administered at 2x10⁹ CFU per mouse.

Antibody-mediated depletions

Mice were given daily rat IgG2 isotype control or depletion antibodies in 100 μl PBS by intraperitoneal injections from day −1 to 4 post infection. To deplete monocytes, mice were given 20 μg Rat anti-mouse CCR2⁵⁹ (clone MC-21 AK). To deplete neutrophils, mice were given 200 μg rat anti-mouse Ly-6G (1A8; Bio X Cell) followed by a secondary anti-Rat kappa Ig light chain (MAR 18.5; Bio X Cell) 8 hours later to enhance depletion efficiency⁶⁰. To deplete neutrophils in CCR2-deficient animals, mice were given 200 μg rat anti-mouse Gr-1 (RB6-8C5; Bio X Cell).

Protein quantifications

Cytokines were measured in supernatants from homogenized tissue using Cytometric Bead Array (BD Biosciences) according to manufacturer’s instructions with the following modification: the amount of capture beads, detection reagents, and sample volumes was scaled down tenfold. Data were collected on an LSRFortessa flow cytometer (BD Biosciences) with FACSDiva v9.0 (BD Biosciences) and analyzed with FlowJo v10 (BD Biosciences).

Tissue preparation and cell isolation

Blood was harvested by cardiac puncture upon euthanasia and collected in 250 U/ml Heparin solution (Millipore Sigma) prior to erythrocyte lysis with Red Blood Cell Lysing Buffer (Millipore Sigma).

Lymph nodes and spleens were homogenized through a 70 μm cell strainer (Fisher Scientific), then flushed with R10 buffer consisting of RPMI 1640 (Millipore Sigma) supplemented with 10 mM HEPES (Millipore Sigma), 10% FBS (Omega Scientific), 1 mM sodium pyruvate (Thermo-Fisher Scientific) and 100 U/ml penicillin + 100 μg/ml streptomycin (Thermo Fisher Scientific).

Intestines were excised, flushed luminally with sterile PBS to remove the feces, opened longitudinally along the mesenteric side and placed luminal side down. Small-intestinal tissue containing macroscopically visible pyogranulomas (PG+), adjacent non-granulomatous areas (PG−) and uninfected control tissue (uninf) were excised using a 2 mm-ø dermal punch-biopsy tool (Keyes). Biopsies within each mouse were pooled, suspended in epithelial-dissociation buffer consisting of calcium and magnesium-free HBSS (Thermo Fisher Scientific) supplemented with 15 mM HEPES, 10 mg/ml BSA (Millipore Sigma), 5 mM EDTA (Millipore Sigma) and 100 U/ml penicillin + 100 μg/ml streptomycin, then incubated for 30 minutes at 37°C under continuous agitation. To isolate immune cells from the lamina propria, the tissue was enzymatically digested in R10 buffer, along with 0.5 Wünsch units/ml liberase TM (Roche), 30 μg/ml DNase I (Roche), and 5 mM CaCl₂ for 20 min at 37°C under continuous agitation. The resulting cell suspensions were filtered through 100 μm cell strainers (Fisher Scientific) and subjected to density-gradient centrifugation using Percoll (GE Healthcare). Briefly, cells were suspended in 40% Percoll and centrifuged over a 70% Percoll layer for 20 min at 600 × g with lowest brake at room temperature. Cells collected between the layers were washed with R10 for downstream analysis.

Flow cytometry

Unspecific Fc binding was blocked for 10 minutes on ice with anti-CD16/CD32 (93; Thermo-Fisher Scientific). Cells were subsequently stained for 30 minutes on ice with the following antibodies and reagents: PE-conjugated rat anti-mouse Siglec-F (E50-2440; BD Biosciences), PE-TxR or PE-Cy5-conjugated rat anti-mouse CD11b (M1/70.15; Thermo Fisher Scientific), PE-Cy5-conjugated mouse anti-mouse NK1.1 (PK136; BioLegend), PE-Cy5.5 or PE-Cy7-conjugated rat anti-mouse CD4 (RM4-5; Thermo Fisher Scientific), PE-Cy7-conjugated rat anti-mouse CD3 (17A2; BioLegend), BV510-conjugated rat anti-mouse CD3e (145-2C11; BioLegend), FITC-conjugated Armenian hamster anti-mouse CD11c (N418; BioLegend), PerCP-Cy5.5-conjugated rat anti-mouse Ly-6C (HK1.4; Thermo Fisher Scientific), PB-conjugated rat anti-mouse CD90.2 (53-2.1; BioLegend), BV510-conjugated rat anti-mouse CD19 (1D3; BD Biosciences), BV605-conjugated Armenian hamster anti-mouse TCRβ (H57-597; BD Biosciences), BV650-conjugated rat anti-mouse I-A/I-E (M5/114.15.2; BD Biosciences), BV711-conjugated rat anti-mouse CD8α (53-6.7; BD Biosciences), BV785-conjugated rat anti-mouse Ly-6G (1A8; Thermo Fisher Scientific), AF647-conjugated mouse anti-mouse CD64 (X54-5/7.1; BD Biosciences), AF700-conjugated mouse anti-mouse CD45.2 (104; BioLegend), PE-CF594-conjugated rat anti-mouse CD45R/B220 (RA3-6B2; BD Biosciences) along with eF780 viability dye (BioLegend) diluted in PBS. Antibodies were used at 1:200 dilution and viability dye at 1:1500 dilution.

For intracellular staining, cells were incubated for 3 hours at 37°C with 5% CO₂ in R10 buffer supplemented with 0.33 μl/ml GolgiStop (BD Biosciences) and 15 μg/ml DNase I. Surface proteins were stained as above, then cells were fixed for 20 minutes on ice with Cytofix/Cytoperm Fixation/ Permeabilization solution (BD Biosciences). Lipocalin-2 was stained using biotin-conjugated rat anti-mouse lipocalin-2 (NGAL; BioLegend) on ice for one hour followed by BV711-conjugated streptavidin (BD Biosciences) at 4°C overnight. Intracellular cytokines were stained at 4°C overnight with PerCP-e710-conjugated rat anti-mouse IL-1β (NJTEN3; Thermo Fisher Scientific), eF450-conjugated rat anti-mouse TNF (MP6-XT22; Thermo Fisher Scientific), PE-conjugated Armenian hamster anti-mouse IL-1α (ALF-161; BioLegend). All intracellular antibodies were diluted 1:200 in Perm/Wash Buffer (BD Biosciences). Streptavidin was diluted 1:400 in Perm/Wash Buffer. Cells were acquired on an LSRFortessa flow cytometer with FACSDiva v9.0 and data was analyzed with FlowJo v10. Dead and clustered cells were removed from all analyses.

Histology

Tissues were fixed in 10% neutral-buffered formalin (Fisher Scientific) and stored at 4°C until further processed. Tissue pieces were embedded in paraffin, sectioned by standard histological techniques and stained with hematoxylin and eosin for subsequent histopathological disease scoring by blinded board-certified pathologists. Tissue sections were given scores between 0-4 (healthy-severe) for multiple parameters, including degree of inflammatory cell infiltration, necrosis, and free bacterial colonies along with tissue-specific parameters such as villus blunting and crypt hyperplasia. Healthy mice were characterized by and subsequently scored as having none or low levels of the parameters described, whereas severely afflicted mice presented with high amounts of the respective parameters.

Fluorescence microscopy

Small-intestinal tissue was dissected and flushed with PBS. Intestines were opened longitudinally and ~0.5 cm tissue pieces containing macroscopically visible lesions were excised. Tissues were fixed in 1% paraformaldehyde overnight then blocked for two hours at room temperature in blocking solution containing 10% BSA, 1 μg/ml anti-CD16/32, and 0.5% normal rat IgG in PBS. AF647-conjugated anti-Ly-6G antibody (1A8; BioLegend) was added at 0.01 mg/ml in 100 μl PBS per sample, then whole tissue was stained for 24 hours at 4°C. Samples were washed three times with PBS, mounted whole onto slides in Prolong Glass Antifade Mountant (Thermo Fisher Scientific) and cured for two days at room temperature. Images were acquired on a DMI 6000 laser-scanning confocal microscope (Leica) with a 20x NA 0.75 oil-immersion objective. The center of the sample was determined in the Z direction, then imaged.

Images were analyzed using ImageJ v2.1. Adjustments for brightness and contrast were applied to the entire image. No threshold manipulation was performed. Green and magenta channels were pseudo-colored.

RNA sequencing

Intestinal punch biopsies were collected as described above. Five biopsies per sample type were pooled for each mouse. RNA was extracted using the RNeasy Plus Mini Kit (Qiagen). Sequence-ready libraries were prepared using the Illumina TruSeq Stranded Total RNA kit with Ribo-Zero Gold rRNA depletion (Illumina). Quality assessment and quantification of RNA preparations and libraries were carried out using an Agilent 4200 TapeStation and Qubit 3, respectively. Samples were sequenced on an Illumina NextSeq 500 to produce 150-base pair single end reads with a mean sequencing depth of 9 million reads per sample. Raw reads from this study were mapped to the mouse reference transcriptome (Ensembl; Mus musculus GRCm38) using Kallisto v0.46.2⁶¹. Raw sequence data are available on the Gene Expression Omnibus (GEO; accession no. GSE194334).

All subsequent analyses were carried out using the statistical computing environment R v4.0.3 in RStudio v1.2.5042 and Bioconductor⁶². Briefly, transcript quantification data were summarized to genes using the tximport package⁶³ and normalized using the trimmed mean of M values (TMM) method in edgeR⁶⁴. Genes with <1 CPM in 3 samples were filtered out. Normalized filtered data were variance-stabilized using the voom function in limma⁶⁵, and differentially expressed genes were identified with linear modeling using limma (FDR ≤ 0.05; absolute log₂FC ≥ 1) after correcting for multiple testing using Benjamini-Hochberg.

Statistics

Statistical analyses were performed using Prism v9.0 (GraphPad Software). Independent groups were compared by Mann-Whitney U test or Kruskal-Wallis test with Dunn’s multiple comparisons test. Survival curves were compared by Mantel-Cox test. Statistical significance is denoted * (p<0.05), ** (p<0.01), *** (p<0.001), **** (p<0.0001), NS (not significant).

Extended Data

Extended Data Figure 1. Intestinal pyogranulomas form upon oral Yersinia infection.

(a) Flow cytometry plots displaying the gating strategy employed to identify eosinophils, dendritic cells, B cells and T cells in small-intestinal tissue at day 5 post Yp-infection. Representative from four independent experiments.

(b) Frequency and total number of eosinophils, dendritic cells, B cells, CD4⁺ T cells and CD8⁺ T cells in small-intestinal tissue at day 5 post Yp-infection. Each circle represents one mouse (n = 15-20). Lines represent median. Pooled from four independent experiments.

Wilcoxon test (two-tailed) was performed for paired analyses (PG- vs PG+). Mann-Whitney U test (two-tailed) was performed for all remaining statistical analyses. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Extended Data Figure 2. Non-pyogranuloma tissue does not undergo inflammation.

(a) Heatmap of top 30 significantly upregulated genes in PG- compared to uninfected samples in descending order by fold change. False discovery rate < 0.05 using Benjamini-Hochberg procedure.

(b) Gene ontology analysis of top 30 upregulated genes by fold change only in PG- compared to uninfected samples. Dotted line denotes p = 0.05.

Extended Data Figure 3. CCR2-deficient mice cannot control intestinal Yersinia.

(a) Quantification of total number of intestinal lesions at day 3 post infection. Each circle represents one mouse (n = 9-11). Line represents median. Pooled data from two independent experiments.

(b) H&E-stained paraffin-embedded small-intestinal sections containing Peyer’s patches from WT and Ccr2^gfp/gfp mice at day 5 post-infection. Dashed circle denotes area containing pyogranuloma or necrosuppurative lesion. Scale bars = 500 μm. Representative images of two independent experiments.

(c) Bacterial burdens in small intestinal PG− and PG+ tissue at day 3 post infection. Each circle represents the mean Yp-CFU of 3-5 pooled punch biopsies from one mouse (n = 15-25). Lines represent geometric mean. Pooled data from four independent experiments.

(d) Frequency of monocytes in blood at day 5 post infection. Each circle represents one mouse (n = 29-30). Lines represent median. Pooled data from four independent experiments.

(e) Bacterial burdens in small intestinal PG− and PG+ tissue at day 5 post infection. Each circle represents the mean Yp-CFU of 3-5 pooled punch biopsies from one mouse (n = 21-30). Lines represent geometric mean. Pooled data from four independent experiments.

(f) Bacterial burdens in small intestinal PG− and PG+ tissue at day 3 post infection. Each circle represents the mean Yp-CFU of 3-5 pooled punch biopsies from one mouse (n = 10-13). Lines represent geometric mean. Pooled data from two independent experiments.

(g) H&E-stained paraffin-embedded sections containing pyogranulomas from WT and Ccr2^gfp/gfp mice at day 3 post-infection. Dashed circle denotes area containing pyogranuloma. Scale bars = 100 μm. Representative images of one experiment.

All statistical analyses by Mann-Whitney U test (two-tailed). * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Extended Data Figure 4. Effects of monocyte deficiency on other cell types.

(a) Total numbers and frequencies of CD4⁺ T cells, CD8⁺ T cells, B cells, and DCs in small intestinal PG+ tissue. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 20-22). Lines represent median. Pooled data from three to five independent experiments.

(b) Frequency of neutrophils in MLN and spleen at day 5 post infection. Each circle represents one mouse (n = 8-12). Lines represent median. Pooled data from two independent experiments.

(c) Neutrophil surface CD11b expression in MLN at day 5 post-infection was measured by flow cytometry. Each circle represents one mouse (n = 6). Lines represent median. Representative of two independent experiments.

(d) Intracellular levels of CD11b in neutrophils in small intestinal PG+ tissue at day 5 post-infection were measured by flow cytometry. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 6-7). Lines represent median. Representative of four independent experiments.

(e) Cytokine levels in homogenates of tissue punch biopsies at day 5 post infection were measured by cytometric bead array. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 18-22). Lines represent median. Pooled data from three independent experiments.

(f) Intracellular levels of cytokines and lipocalin in neutrophils in small intestinal PG+ tissue at day 5 post-infection were measured by flow cytometry. Each circle represents the mean of 3-10 pooled punch biopsies from one mouse (n = 18-19). Lines represent median. Pooled data from three independent experiments.

All statistical analyses by Mann-Whitney U test (two-tailed). * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Extended Data Figure 5. CCR2-deficient mice cannot control systemic Yersinia.

(a) Bacterial burdens in indicated organs at day 3 post infection. Each circle represents one mouse (n = 10-14 for PP, MLN, lung; 21-25 for spleen, liver). Lines represent geometric mean. Pooled data from two to four independent experiments.

(b) Bacterial burdens in indicated organs at day 5 post infection. Each circle represents one mouse (n = 29-30). Lines represent geometric mean. Pooled data from four independent experiments.

(c) Bacterial burdens in indicated organs at day 3 post infection. Each circle represents one mouse (n = 10-13). Lines represent geometric mean. Pooled data from two independent experiments.

(d) Survival of infected mice. Pooled data from two independent experiments. Statistical analyses by (a-c) Mann-Whitney U test (two-tailed) or (d) Mantel-Cox test. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Extended Data Figure 6. Yersinia virulence factors induce intestinal pyogranulomas.

(a) Cumulative bacterial burdens in PG- tissue at day 5 post infection. Each symbol represents one mouse (n = 9-11). Lines represent geometric mean. Dotted line represents limit of detection. Pooled from two independent experiments.

(b) Frequency of monocytes in small-intestinal tissue and MLN. Each symbol represents one mouse (n = 3-9). Lines represent median. Pooled and representative data from two independent experiments.

(c) Frequency of neutrophils in small-intestinal tissue and MLN. Each symbol represents one mouse (n = 3-9). Lines represent median. Pooled and representative data from two independent experiments.

(d) Bacterial burdens in indicated organs at day 5 post-infection. Each symbol represents one mouse (n = 10-43). Lines represent geometric mean. Pooled from 2-6 independent experiments.

(e) Total number of intestinal lesions at day 5 post infection. Each symbol represents one mouse (n = 8-10). Lines represent median. Pooled from two independent experiments.

(f) Total number of intestinal lesions at day 5 post infection. Each symbol represents one mouse (n = 11). Lines represent median. Pooled from three independent experiments.

(g) Frequency and total number of monocytes in small-intestinal PG+ tissue at day 5 post WT or YopH^R409A Yp infection. Each circle represents one mouse (n = 8-10). Lines represent median. Pooled from two independent experiments.

(h) Bacterial burdens in indicated organs at day 3 post WT or YopH^R409A Yp infection. Each symbol represents one mouse (n = 10-13). Lines represent geometric mean. Pooled from two independent experiments.

Statistical analyses by (a, d, e) Kruskal-Wallis test with Dunn’s post-test and (b, c, f, g, h) Mann-Whitney U test (two-tailed). * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05).

Extended Data Figure 7. Anti-Gr-1 effectively depletes neutrophils during infection.

(a) Flow cytometry plots displaying the gating strategy employed to identify neutrophils and monocytes upon anti-Gr-1 administration. Due to masking of Ly-6G and Ly-6C epitopes by anti-Gr-1, monocytes were identified as CCR2-GFP⁺ cells (green box) and neutrophils were identified as SSC high cells (pink boxes). Representative images of three independent experiments.

(b) Frequency of monocytes in blood at day 5 post infection was determined by flow cytometry. Each symbol represents one mouse (n = 10-11). Lines represent median. Data from three independent experiments. Statistical analyses by Kruskal-Wallis test with Dunn’s post-test. * (p < 0.05), ** (p < 0.01), *** (p < 0.001), **** (p < 0.0001), NS (not significant, p > 0.05)

Extended Data Table 1.

Gene ontology analysis of PG+ vs PG− samples

p_value	term_id	term_name
2.16113E-06	GO:0009617	response to bacterium
6.32193E-06	GO:0006952	defense response
7.52444E-06	KEGG:04657	IL-17 signaling pathway
2.20616E-05	REAC:R-MMU-6798695	Neutrophil degranulation
2.94164E-05	GO:0006954	inflammatory response
7.19842E-05	GO:0005576	extracellular region
7.26469E-05	REAC:R-MMU-168249	Innate Immune System
0.000111741	GO:0051707	response to other organism
0.000111741	GO:0043207	response to external biotic stimulus
0.000111741	GO:0009607	response to biotic stimulus
0.000111741	GO:0052548	regulation of endopeptidase activity
0.00011503	REAC:R-MMU-6799990	Metal sequestration by antimicrobial proteins
0.000130727	GO:0052547	regulation of peptidase activity
0.000132537	GO:0044419	biological process involved in interspecies interaction between organisms
0.000132537	GO:0060326	cell chemotaxis
0.000190598	GO:0030162	regulation of proteolysis
0.000220253	GO:0006508	proteolysis
0.000288993	GO:0006950	response to stress
0.000291628	GO:0032496	response to lipopolysaccharide
0.000319678	GO:0030595	leukocyte chemotaxis
0.000319678	GO:0097529	myeloid leukocyte migration
0.000319678	GO:0002237	response to molecule of bacterial origin
0.000344762	GO:0030593	neutrophil chemotaxis
0.000395608	GO:0005615	extracellular space
0.000401486	REAC:R-MMU-5621480	Dectin-2 family
0.000438686	GO:0070488	neutrophil aggregation
0.000673081	GO:1990266	neutrophil migration
0.000682844	GO:0071621	granulocyte chemotaxis
0.001021883	REAC:R-MMU-168256	Immune System
0.001021883	REAC:R-MMU-5686938	Regulation of TLR by endogenous ligand
0.001304093	GO:0097530	granulocyte migration
0.001304093	GO:0010727	negative regulation of hydrogen peroxide metabolic process
0.001412884	GO:0035662	Toll-like receptor 4 binding
0.001798466	GO:0009605	response to external stimulus
0.001982425	KEGG:05134	Legionellosis
0.001993156	GO:0035425	autocrine signaling
0.002051465	REAC:R-MMU-5668599	RHO GTPases Activate NADPH Oxidases
0.002194029	GO:0050900	leukocyte migration
0.002194029	GO:0006935	chemotaxis
0.002215455	GO:0042330	taxis
0.00255511	GO:0045861	negative regulation of proteolysis
0.002892713	KEGG:05417	Lipid and atherosclerosis
0.002892713	KEGG:05323	Rheumatoid arthritis
0.003588317	GO:0038094	Fc-gamma receptor signaling pathway
0.004193849	CORUM:2552	Il3rb1-Shc complex, IL-3 stimulated
0.004193849	CORUM:3202	Il3rb1-Shc-Ship complex, IL-3 stimulated
0.004796173	GO:0005102	signaling receptor binding
0.004796173	GO:0048306	calcium-dependent protein binding
0.004796173	GO:0016209	antioxidant activity
0.004796173	GO:0045236	CXCR chemokine receptor binding
0.004796173	GO:0035325	Toll-like receptor binding
0.004831801	GO:0071222	cellular response to lipopolysaccharide
0.00509277	GO:0018119	peptidyl-cysteine S-nitrosylation
0.00509277	GO:0071219	cellular response to molecule of bacterial origin
0.005161487	KEGG:04668	TNF signaling pathway
0.005255007	GO:0098869	cellular oxidant detoxification
0.005255007	GO:0010310	regulation of hydrogen peroxide metabolic process
0.005255007	GO:0017014	protein nitrosylation
0.005516248	KEGG:04060	Cytokine-cytokine receptor interaction
0.005686316	GO:0070486	leukocyte aggregation
0.00618356	GO:0038093	Fc receptor signaling pathway
0.006324423	GO:0071216	cellular response to biotic stimulus
0.006363274	GO:1900048	positive regulation of hemostasis
0.006363274	GO:0030194	positive regulation of blood coagulation
0.006785708	GO:1990748	cellular detoxification
0.006976822	GO:0019538	protein metabolic process
0.006976822	GO:0010951	negative regulation of endopeptidase activity
0.006976822	GO:0050820	positive regulation of coagulation
0.006976822	GO:0061844	antimicrobial humoral immune response mediated by antimicrobial peptide
0.006976822	GO:0051336	regulation of hydrolase activity
0.006976822	GO:0097237	cellular response to toxic substance
0.006976822	GO:0002523	leukocyte migration involved in inflammatory response
0.006976822	GO:0010466	negative regulation of peptidase activity
0.007587123	GO:0030546	signaling receptor activator activity
0.007587123	GO:0048018	receptor ligand activity
0.008029604	GO:0098754	detoxification
0.008029604	GO:0070887	cellular response to chemical stimulus
0.008491456	GO:1901564	organonitrogen compound metabolic process
0.008491456	GO:0043280	positive regulation of cysteine-type endopeptidase activity involved in apoptotic process
0.008871636	REAC:R-MMU-380108	Chemokine receptors bind chemokines
0.009751859	GO:0050729	positive regulation of inflammatory response
0.010252954	GO:0007166	cell surface receptor signaling pathway
0.01137206	GO:2001056	positive regulation of cysteine-type endopeptidase activity
0.011871608	GO:0019730	antimicrobial humoral response
0.012088564	GO:0050727	regulation of inflammatory response
0.01368439	GO:0050896	response to stimulus
0.013729547	GO:0033993	response to lipid
0.013729547	GO:0002376	immune system process
0.014330714	GO:0050790	regulation of catalytic activity
0.015486453	GO:0042060	wound healing
0.015486453	GO:0006955	immune response
0.015667249	GO:0007596	blood coagulation
0.01569326	GO:1902913	positive regulation of neuroepithelial cell differentiation
0.01569326	GO:1901331	positive regulation of odontoblast differentiation
0.01569326	GO:1904843	cellular response to nitroglycerin
0.01569326	GO:0050817	coagulation
0.01569326	GO:0061408	positive regulation of transcription from RNA polymerase II promoter in response to heat stress
0.01569326	GO:0035490	regulation of leukotriene production involved in inflammatory response
0.01569326	GO:2000296	negative regulation of hydrogen peroxide catabolic process
0.01569326	GO:0007599	hemostasis
0.01569326	GO:0035491	positive regulation of leukotriene production involved in inflammatory response
0.01569326	GO:0009405	obsolete pathogenesis
0.01569326	GO:0014002	astrocyte development
0.01569326	GO:0010950	positive regulation of endopeptidase activity
0.017653698	GO:0006953	acute-phase response
0.018417791	GO:0010952	positive regulation of peptidase activity
0.018858959	GO:0032268	regulation of cellular protein metabolic process
0.019282879	GO:0018198	peptidyl-cysteine modification
0.019829956	GO:2000378	negative regulation of reactive oxygen species metabolic process
0.020378866	GO:0042743	hydrogen peroxide metabolic process
0.020384657	GO:0042742	defense response to bacterium
0.021431689	GO:0098542	defense response to other organism
0.021452884	GO:0004857	enzyme inhibitor activity
0.021452884	GO:0061770	translation elongation factor binding
0.021452884	GO:0008009	chemokine activity
0.021651519	GO:0005509	calcium ion binding
0.022590954	GO:0051346	negative regulation of hydrolase activity
0.022590954	GO:0090303	positive regulation of wound healing
0.022999507	GO:0043281	regulation of cysteine-type endopeptidase activity involved in apoptotic process
0.022999507	GO:0044092	negative regulation of molecular function
0.023149342	GO:0005544	calcium-dependent phospholipid binding
0.023526513	GO:0072737	response to diamide
0.023526513	GO:0051246	regulation of protein metabolic process
0.023526513	GO:0062026	negative regulation of SCF-dependent proteasomal ubiquitin-dependent catabolic process
0.023526513	GO:0072738	cellular response to diamide
0.023526513	GO:0062025	regulation of SCF-dependent proteasomal ubiquitin-dependent protein catabolic process
0.023526513	GO:0072593	reactive oxygen species metabolic process
0.023526513	GO:2000295	regulation of hydrogen peroxide catabolic process
0.023526513	GO:1900138	negative regulation of phospholipase A2 activity
0.023526513	GO:1904842	response to nitroglycerin
0.023526513	GO:0043086	negative regulation of catalytic activity
0.023526513	GO:0035606	peptidyl-cysteine S-trans-nitrosylation
0.023526513	GO:2000098	negative regulation of smooth muscle cell-matrix adhesion
0.023526513	GO:0030887	positive regulation of myeloid dendritic cell activation
0.023526513	GO:0010332	response to gamma radiation
0.02358878	REAC:R-MMU-6803157	Antimicrobial peptides
0.02359003	GO:0004866	endopeptidase inhibitor activity
0.023704517	GO:0050789	regulation of biological process
0.023715466	GO:0030414	peptidase inhibitor activity
0.023715466	GO:0046899	nucleoside triphosphate adenylate kinase activity
0.023715466	GO:0061135	endopeptidase regulator activity
0.023891517	HP:0040245	Reduced alpha-2-antiplasmin activity
0.023891517	HP:0040228	Decreased level of plasminogen
0.023891517	HP:0040248	Reduced plasminogen activator inhibitor 1 activity
0.023891517	HP:0040224	Abnormality of fibrinolysis
0.023891517	HP:0040230	Decreased level of tissue plasminogen activator
0.023891517	HP:0040249	Reduced plasminogen activator inhibitor 1 antigen
0.023891517	HP:0100310	Epidural hemorrhage
0.023891517	HP:0032312	Decreased circulating globulin level
0.023891517	HP:0011854	Hemoperitoneum
0.023891517	HP:0040184	Oral bleeding
0.025295615	GO:0030193	regulation of blood coagulation
0.025816626	GO:1900046	regulation of hemostasis
0.025816626	GO:2001244	positive regulation of intrinsic apoptotic signaling pathway
0.025816626	GO:2000116	regulation of cysteine-type endopeptidase activity
0.025816626	GO:0031638	zymogen activation
0.025816626	GO:0009636	response to toxic substance
0.026199211	GO:0050818	regulation of coagulation
0.026878816	GO:0004869	cysteine-type endopeptidase inhibitor activity
0.026878816	GO:0005125	cytokine activity
0.027315059	GO:0009611	response to wounding
0.027561199	HP:0031029	Elevated carcinoembryonic antigen level
0.027561199	HP:0030657	Umbilical cord hematoma
0.027561199	HP:0025391	Crazy paving pattern on pulmonary HRCT
0.028727626	GO:0042379	chemokine receptor binding
0.028727626	GO:0061134	peptidase regulator activity
0.028845751	GO:0042509	regulation of tyrosine phosphorylation of STAT protein
0.029344406	GO:1990911	response to psychosocial stress
0.029344406	GO:1903852	positive regulation of cristae formation
0.029344406	GO:1904845	cellular response to L-glutamine
0.029344406	GO:1990910	response to hypobaric hypoxia
0.029344406	GO:0031349	positive regulation of defense response
0.029344406	GO:0007260	tyrosine phosphorylation of STAT protein
0.029344406	GO:1901329	regulation of odontoblast differentiation
0.029344406	GO:0002292	T cell differentiation involved in immune response
0.029344406	GO:0002540	leukotriene production involved in inflammatory response
0.029344406	GO:1903036	positive regulation of response to wounding
0.029344406	GO:0045204	MAPK export from nucleus
0.030446796	GO:0051716	cellular response to stimulus
0.030446796	GO:0007165	signal transduction
0.030913215	REAC:R-MMU-5621481	C-type lectin receptors (CLRs)
0.031668575	GO:0006919	activation of cysteine-type endopeptidase activity involved in apoptotic process
0.031841638	HP:0030879	Interlobular septal thickening
0.031841638	HP:0033711	Pulmonary interstitial thickening
0.032646273	GO:1990444	F-box domain binding
0.032646273	GO:0030246	carbohydrate binding
0.032646273	GO:0001691	pseudophosphatase activity
0.03353745	GO:0070098	chemokine-mediated signaling pathway
0.03353745	GO:1901700	response to oxygen-containing compound
0.034461909	GO:0031347	regulation of defense response
0.035643318	GO:1903850	regulation of cristae formation
0.035643318	GO:1901491	negative regulation of lymphangiogenesis
0.035643318	GO:1900365	positive regulation of mRNA polyadenylation
0.035643318	GO:0006172	ADP biosynthetic process
0.035643318	GO:1904844	response to L-glutamine
0.035643318	GO:0002538	arachidonic acid metabolite production involved in inflammatory response
0.035643318	GO:0050794	regulation of cellular process
0.035997936	GO:0048708	astrocyte differentiation
0.036002466	REAC:R-MMU-3371511	HSF1 activation
0.036269445	GO:0044267	cellular protein metabolic process
0.037888022	GO:0071396	cellular response to lipid
0.038461896	REAC:R-MMU-168898	Toll-like Receptor Cascades
0.038648573	GO:1900034	regulation of cellular response to heat
0.038648573	GO:1900004	negative regulation of serine-type endopeptidase activity
0.038648573	GO:1900003	regulation of serine-type endopeptidase activity
0.038648573	GO:1901342	regulation of vasculature development
0.038648573	GO:0061044	negative regulation of vascular wound healing
0.038648573	GO:0065009	regulation of molecular function
0.038648573	GO:0009136	purine nucleoside diphosphate biosynthetic process
0.038648573	GO:0009180	purine ribonucleoside diphosphate biosynthetic process
0.038648573	GO:1902572	negative regulation of serine-type peptidase activity
0.038648573	GO:1902571	regulation of serine-type peptidase activity
0.038648573	GO:1990868	response to chemokine
0.038648573	GO:1902512	positive regulation of apoptotic DNA fragmentation
0.038648573	GO:1903936	cellular response to sodium arsenite
0.038648573	GO:1990869	cellular response to chemokine
0.038648573	GO:0045765	regulation of angiogenesis
0.038648573	GO:2000097	regulation of smooth muscle cell-matrix adhesion
0.038648573	GO:0009188	ribonucleoside diphosphate biosynthetic process
0.039982864	GO:0065007	biological regulation
0.041518754	GO:0005126	cytokine receptor binding
0.042452241	GO:0071347	cellular response to interleukin-1
0.043080047	GO:0016477	cell migration
0.043080047	GO:1903626	positive regulation of DNA catabolic process
0.043080047	GO:0097187	dentinogenesis
0.043080047	GO:0010757	negative regulation of plasminogen activation
0.043080047	GO:0071895	odontoblast differentiation
0.043080047	GO:0051918	negative regulation of fibrinolysis
0.043080047	GO:1904528	positive regulation of microtubule binding
0.043080047	GO:0010519	negative regulation of phospholipase activity
0.043080047	GO:0045113	regulation of integrin biosynthetic process
0.043080047	GO:2001033	negative regulation of double-strand break repair via nonhomologous end joining
0.04419202	GO:0023052	signaling
0.04419202	GO:0007159	leukocyte cell-cell adhesion
0.044654435	GO:0032269	negative regulation of cellular protein metabolic process
0.045893648	GO:0009408	response to heat
0.046945217	GO:0002526	acute inflammatory response
0.046945217	GO:0030885	regulation of myeloid dendritic cell activation
0.046945217	GO:0045112	integrin biosynthetic process
0.046945217	GO:1903935	response to sodium arsenite
0.046945217	GO:0061302	smooth muscle cell-matrix adhesion
0.046945217	GO:1901490	regulation of lymphangiogenesis
0.046945217	GO:1902075	cellular response to salt
0.046945217	GO:0007154	cell communication
0.048811625	GO:0045862	positive regulation of proteolysis
0.048811625	GO:0002286	T cell activation involved in immune response
0.049030752	GO:0021782	glial cell development
0.049030752	GO:0050878	regulation of body fluid levels
0.049326198	KEGG:04061	Viral protein interaction with cytokine and cytokine receptor
0.049360301	REAC:R-MMU-3371568	Attenuation phase

Extended Data Table 2.

Gene ontology analysis of PG− vs uninfected samples Supplementary code file

p_value	term_id	term_name
0.00243412	REAC:R-MMU-71387	Metabolism of carbohydrates
0.00243412	REAC:R-MMU-9033658	Blood group systems biosynthesis
0.00243412	REAC:R-MMU-9037629	Lewis blood group biosynthesis
0.00276581	KEGG:00603	Glycosphingolipid biosynthesis - globo and isoglobo series
0.00341096	GO:0005976	polysaccharide metabolic process
0.00372835	KEGG:00601	Glycosphingolipid biosynthesis - lacto and neolacto series
0.00855723	GO:0005975	carbohydrate metabolic process
0.02242855	KEGG:01230	Biosynthesis of amino acids
0.02874227	GO:0000139	Golgi membrane
0.02918355	REAC:R-MMU-8964540	Alanine metabolism
0.03100142	GO:0044264	cellular polysaccharide metabolic process
0.03100142	GO:0006022	aminoglycan metabolic process
0.03100142	GO:0000272	polysaccharide catabolic process
0.03361355	GO:0044262	cellular carbohydrate metabolic process
0.03466317	GO:0102148	N-acetyl-beta-D-galactosaminidase activity
0.03466317	GO:0001632	leukotriene B4 receptor activity
0.03466317	GO:0008107	galactoside 2-alpha-L-fucosyltransferase activity
0.03466317	GO:0051734	ATP-dependent polynucleotide kinase activity
0.03466317	GO:0051733	polydeoxyribonucleotide kinase activity
0.03466317	GO:0051731	polynucleotide 5’-hydroxyl-kinase activity
0.03466317	GO:0047635	alanine-oxo-acid transaminase activity
0.03466317	GO:0046404	ATP-dependent polydeoxyribonucleotide 5’-hydroxyl-kinase activity
0.03466317	GO:0031127	alpha-(1,2)-fucosyltransferase activity
0.03466317	GO:0016740	transferase activity
0.03466317	GO:0003872	6-phosphofructokinase activity
0.03466317	GO:0004021	L-alanine:2-oxoglutarate aminotransferase activity
0.03466317	GO:0015016	[heparan sulfate]-glucosamine N-sulfotransferase activity
0.03466317	GO:0004974	leukotriene receptor activity
0.03500403	REAC:R-MMU-9033807	ABO blood group biosynthesis
0.03932616	KEGG:01200	Carbon metabolism
0.03932616	KEGG:01100	Metabolic pathways
0.04042288	GO:0070095	fructose-6-phosphate binding
0.04128752	GO:0030262	apoptotic nuclear changes
0.04128752	GO:0006921	cellular component disassembly involved in execution phase of apoptosis
0.04167664	GO:0004519	endonuclease activity
0.04167664	GO:0070061	fructose binding
0.04167664	GO:0004563	beta-N-acetylhexosaminidase activity
0.04426833	GO:0098791	Golgi apparatus subcompartment
0.04857165	REAC:R-MMU-391906	Leukotriene receptors

Data Availability

Raw RNA sequencing data are available on the Gene Expression Omnibus (GEO; accession no. GSE194334). All other raw data are available upon request to the corresponding author.