Chronic Trichuris muris infection alters hematopoiesis and causes IFN-γ-expressing T-cell accumulation in the mouse bone marrow

Chronic T. muris infection induces increased LSK cells in the BM

To assess the systemic impact of localized intestinal immune responses on hematopoiesis, we employed the T. muris model of helminth infection. The T. muris model is a well-characterized system that allows the induction of either a Th1 cell-driven persistent infection (low dose) or a Th2 cell-driven acute infection (high dose) in the large intestine of most strains of wild-type mice 13. Using this infection model, we could compare the effects of intestinal Th1 and Th2 cell-dependent immune responses on hematopoiesis. As expected, low dose (LD) T. muris infection of C57BL/6 mice resulted in a persistent worm burden 21 days postinfection and remained chronically infected 35 days postinfection (Fig. 1A). In contrast, high dose (HD) T. muris infection resulted in few or no detectable worms in the cecum, consistent with a protective Th2 cell response. Th1 and Th2 cell-dominant responses, for the LD and HD infections, respectively, were confirmed by analyzing T. muris-specific IgG responses and cytokine expression in the colon (see Supporting Information Fig. 1). Isolation and enumeration of total BM cells from these mice showed that LD, but not HD T. muris infection resulted in a significant reduction in total cellularity in the BM (Fig. 1B). Next, we assessed whether specific changes to the composition of major cell populations in the BM occur in response to LD or HD T. muris infections. Since LD T. muris infection is associated with an IFN-γ-driven Th1 cell responses 13 and that IFN-γ has been shown to induce the expression of Sca-1 on immune cells 14, we assessed Sca-1 expression on hematopoietic cells in the BM of infected mice. The frequency and absolute number of total CD45⁺ BM cells expressing Sca-1 was increased twofold in LD T. muris-infected mice compared to naive controls (Fig. 1C). Conversely, mice that received a HD T. muris infection did not have elevated Sca-1 expression on hematopoietic cells in the BM. We further analyzed lineage-negative cell populations in the BM and found that only LD T. muris infection caused a significant increase in the frequency and absolute number of Sca-1⁺c-Kit⁺ (LSK) cells compared to naive and HD T. muris-infected mice (Fig. 1D). Together, these data show that a persistent LD T. muris infection induces significant and broad changes in the hematopoietic compartment in the BM that are absent after a transient HD T. muris infection.

Chronic T. muris infection alters bone marrow MPP populations

To further characterize the infection-induced changes in the BM during LD T. muris infection, we analyzed progenitor and stem cell populations within the LSK population using the signaling lymphocytic activation molecule family markers CD48 and CD150—multipotent progenitor (MPP) cells are enriched within the CD48⁻CD150⁻ population while HSCs can be identified as being CD48⁻CD150⁺ 15. We found that LD T. muris-infected mice had significant expansion in the number of CD48⁻CD150⁻ MPPs in the BM compared to naive mice (Fig. 2A). No differences were found in terms of the absolute numbers of HSC in BM during infection. Analysis of the BM at various time points of LD T. muris infection indicated that the expansion of MPPs peaked at day 21 and returned to normal levels by day 35 postinfection (Fig. 2B). To test the functional effects of these changes in the BM, we performed in vitro colony forming assays. BM from LD T. muris-infected mice had significantly higher total CFU counts (Fig. 2C). There were no significant differences found in terms of monoblast (CFU-M) or granulocyte (CFU-G) colony counts; however, BM from LD T. muris-infected mice yielded more precursor CFU-GM (granulocyte, monocyte) colonies. Correlating with enhanced CFU-GM formation, analysis of intermediate progenitor populations of the myeloid lineage showed no differences in common myeloid progenitor cells and an increase in the frequency, but not number, of granulocyte-monocyte progenitor (GMP) cells in the BM of infected mice (Fig. 2D). In parallel, analysis of mature cell lineages confirmed a specific increase in the number of neutrophils in the BM of LD T. muris-infected mice (Fig. 2E). This coincided with an elevation of peripheral neutrophils in the blood while other cell populations remained unchanged on day 21 and 35 postinfection (see Supporting Information Fig. 2). Other blood parameters such as red blood cell count and hematocrit were unaffected, suggesting that LD infection does not cause anemia in mice. Therefore, LD T. muris infection alters primitive progenitor populations in the BM that correlate with an increase in myeloid colony forming potential in vitro and an increase in BM and peripheral blood neutrophils.

As LD T. muris infection is associated with a Th1 cell response 16, we next tested whether the effects we saw were indeed dependent on adaptive immunity by analyzing the BM of infected Rag1^−/− mice that lack T and B cells. In contrast to infected WT mice, LD T. muris-infected Rag1^−/− mice did not show an expansion of MPPs in the BM when compared to uninfected controls (Fig. 2F). Similarly, the absolute number of HSCs in infected Rag1^−/− mice was unaffected. Thus, LD T. muris infection alters MPP populations in the hematopoietic compartment of the BM and is dependent on the presence of adaptive immune cells.

Chronic T. muris infection stimulates proliferation in hematopoietic progenitors

IFN-γ has previously been shown to activate HSCs during chronic infection 6. Thus, we measured proliferation of hematopoietic progenitor populations during our model of Th1 cell-associated chronic helminth infection by pulsing mice with 5-ethynyl-2-deoxyuridine (EdU) prior to the peak of the observed BM response on day 21 postinfection with T. muris. LSK cells in the BM of T. muris-infected mice had >twofold increased EdU retention compared to LSK cells from naïve mice (Fig. 3). Among the LSK population, there was a trend for an increase in MPP cells that were EdU⁺ during infection and HSCs had a significant increase in EdU retention compared to naïve controls. Together these data show that chronic T. muris infection broadly stimulates proliferation in hematopoietic progenitor cells with HSCs displaying significantly enhanced activation.

Alterations in hematopoiesis during chronic T. muris infection is IFN-γ-dependent

Since we found that changes in the BM in response to LD T. muris infection depended on adaptive immunity, we sought to determine whether the infection-mediated effects on hematopoiesis were indeed driven by the Th1 cell-associated cytokine IFN-γ. To directly test this, we neutralized IFN-γ systemically during the course of the infection using monoclonal antibody treatment. Blockade of IFN-γ did not cause mice to clear a LD T. muris infection, as infected mice treated with control IgG antibodies or anti-IFN-γ had comparable worm burdens (Fig. 4A). Further, the infection-induced increase in LSK cells, in terms of frequencies and absolute numbers, was completely reversed upon blockade of IFN-γ (Fig. 4B). As shown in Fig. 4C, the expansion of CD48⁻CD150⁻ MPPs in the BM during LD T. muris infection was reversed in infected mice treated with α-IFN-γ that resembled naive MPP numbers. Further, neutralization of IFN-γ reversed the reduction in total BM cellularity caused by chronic T. muris infection (Fig. 4D). However, peripheral blood neutrophilia was still observed in chronically infected mice, independent of IFN-γ signaling (Fig. 4E). Thus, alterations in hematopoietic progenitor populations in the BM, elicited by LD T. muris intestinal infection, is completely dependent on IFN-γ signaling while peripheral blood neutrophilia is not dependent on IFN-γ during infection.

A mixed population of IFN-γ-expressing T cells accumulates in the BM chronic T. muris infection

Given that we found that IFN-γ was playing a major role in influencing hematopoiesis during LD T. muris infection, we next assessed whether IFN-γ-expressing cells are found in the BM. To do this, we infected IFN-γ-YFP reporter (GREAT) mice with LD T. muris and characterized IFN-γ expression in the BM. Compared to naive controls, LD T. muris-infected mice had a small yet significant increase in IFN-γ⁺ cells in the BM on day 21 post-infection (Fig. 5A). The appearance of these IFN-γ⁺ cells coincides with the peak of BM response described earlier as we could not detect a significant increase in these cells as late as day 14 post-infection (see Supporting Information Fig. 3). Additionally, the IFN-γ-producing BM cells from LD T. muris-infected mice also had a significantly higher MFI compared to controls. Upon examination of specific cell types in the BM, we found that the majority of cells with increased IFN-γ expression during infection were CD3⁺ lymphocytes (Fig. 5B). IFN-γ expression was significantly increased in CD4⁺, CD8⁺, and CD4⁻CD8⁻ double negative (DN) T-cell populations in the BM of LD T. muris-infected mice compared to BM from naive mice. These IFN-γ⁺ T cells in the BM were also significantly increased in terms of absolute numbers during LD T. muris infection. In contrast, non-T cell populations such as NK cells showed no significant differences in IFN-γ expression. Therefore, a mixed population of IFN-γ⁺ T cells accumulates in the BM during persistent LD T. muris intestinal infection.

BM IFN-γ⁺ T cells display activated and central memory phenotypes during chronic T. muris infection

We next phenotyped surface marker expression on IFN-γ⁺ T cells in the BM of LD T. muris-infected mice. The majority (> 90%) of IFN-γ-expressing CD4⁺ and CD8⁺ T cells in the BM were predominately TCR-β⁺ but were of comparable frequency between LD T. muris-infected and uninfected mice (Fig. 6A). Conversely, IFN-γ⁺ DN T cells could be split into both TCR-β⁺ and TCR-γδ⁺ populations (2:1 ratio, respectively) and were of similar frequencies between LD T. muris-infected and naive controls. Analysis of CD44 and CD62L expression showed that IFN-γ⁺ CD4⁺ T cells in the BM from both LD T. muris-infected and naive mice were mainly CD44⁺CD62L⁻, consistent with an activated phenotype (Fig. 6B). However, for both IFN-γ⁺ CD8⁺ and DN T cells, there was a significant increase in the frequency of CD44⁺CD62L⁺ cells in the BM of LD T. muris-infected mice, indicative of an increase in central memory type cells. Thus, LD T. muris infection induces a mixed population of T cells expressing IFN-γ that display activated and central memory phenotypes in the BM.

Ethics statement

Experiments were approved by the UBC ACC (Protocol number A13-0010) and were in accordance with the Canadian Guidelines for Animal Research.

Mice and intestinal worm infections

Propagation of T. muris eggs was performed as previously described 30. C57BL/6 GREAT (IFN-γ YFP reporter), and Rag1^−/− mice were originally purchased from Jackson Laboratories (Bar Harbor, ME). Mice were either uninfected or infected with a low dose (30 eggs) or high dose (200 eggs) of T. muris by oral gavage. Infections were carried out for either 7, 14, 21, or 35 days before sacrifice of mice. All mice were bred and housed in specific pathogen-free animal facilities at the UBC Biomedical Research Centre (Vancouver, BC).

Serum ELISA

Serum was collected from mice 21 days post-infection with T. muris. Immulon plates (Thermo Fischer Scientific, NY) were coated with 5 μg/mL of dialyzed T. muris antigen overnight at 4°C. Wash buffer was PBS containing 0.05% Tween 20. Plates were blocked and serum samples were diluted in 3% bovine serum albumin in PBS/0.05% Tween 20. Serum samples were incubated on plates for 1 hour at room temperature. Plates were then incubated with rat anti-mouse IgG1 or IgG2a conjugated to horseradish peroxidase (BD Pharmingen, CA) for 1 hour at room temperature. Plates were developed using 3,3',5,5'-tetramethylbenzidine (TMB) substrate (MandelScientific, ON) and stopped with 1N HCl. Plates were read at 450 nm on a Spectramax 384 (Molecular Devices, CA).

Quantitative real-time PCR

Colon tissue was mechanically homogenized using a TissueLyser (Qiagen, Germany) and RNA was extracted using the TRIzol method according to the manufacturer's instructions (Ambion, TX). Complementary DNA (cDNA) was generated using High Capacity cDNA reverse transcription kits according to the manufacturer's instructions (Applied Biosystems, CA). Quantitative real-time PCR was performed using the SYBR FAST master mix (Kapa Biosystems, MA) and SYBR green-optimized primer sets run on an ABI 7900 real-time PCR machine (Applied Biosystems, CA). Ct values were normalized relative to beta-actin (Actb) gene expression. Mouse primer sequences (5’-3’): Actb forward ACTAATGGCAACGAGCGGTTC; Actb reverse GGATGCCACAGGATTCCATACC; Il13 forward GGTCTTGTGTGATGTTGCTCA; Il13 reverse CCTGGCTCTTGCTTGCCTT; Ifng forward GGATGCATTCATGAGTATTGCC; Ifng reverse CCTTTTCCGCTTCCTGAGG.

Isolation of BM cells

Following sacrifice of mice, hip joints were dislocated and the right hind limbs of each mouse were dissected to isolate single long bones, clean of any muscle and other tissues. The ends of each bone were sliced by a razor blade and a 27 gauge needle was used to flush out the BM with media (DMEM supplemented with 10% FBS, 25 mM HEPES, penicillin/streptomycin, 2 mM l-glutamine, and 50 μM β-mercaptoethanol). The BM was then passed through a 70 μm cell strainer to obtain single-cell suspensions. BM samples were red cell-lysed using ammonium chloride buffer, and suspended to a known volume before being enumerated by a hemocytometer and stained for flow cytometry.

Blood parameter analysis

Whole blood was collected into EDTA-coated Microvette collection tubes (Sarstedt, Germany) by saphenous vein bleed of mice prior to sacrifice. Blood samples were analyzed using an Advia 120 hematology analyzer (Siemens, ON).

EdU labeling and proliferation assay

Mice were infected with T. muris as described above. On days 18 and 20 post-infection, mice were injected i.p. with 3 mg 5-ethynyl-2-deoxyuridine (EdU) dissolved in PBS on each day, prior to sacrifice on day 21. BM cells were harvested as described above and EdU was stained using Click-iT Assay kits for Flow Cytometry (ThermoFisher Scientific, MA) following the manufacturer's protocol. Cells were then surface stained for flow cytometric analysis.

In vivo neutralization of IFN-γ

Mice were infected with T. muris as described above. On day 4 postinfection, mice were injected i.p. with 500 μg of either control IgG1 or anti-IFN-γ (XMG1.2) (produced in-house by AbLab Biologics, UBC (Vancouver, BC)), constituted in sterile PBS. Mice were repeatedly injected thereafter on days 8, 12, and 16 prior to sacrifice on day 21.

Antibodies and flow cytometry

All antibody dilutions and cell staining were done with PBS containing 2% FCS, 1 mM EDTA, and 0.05% sodium azide. Fixable Viability Dye eFluor 506 was purchased from eBioscience to exclude dead cells from analyses. Prior to staining, samples were Fc-blocked with buffer containing anti-CD16/32 (93, eBioscience) and 1% rat serum to prevent nonspecific antibody binding. PE-conjugated anti-CD34 (RAM34) was purchased from BD Biosciences. APC-conjugated anti-CD150 (TC15-12F12.2) was purchased from BioLegend. PE-conjugated anti-CD3e (2C11); PE-Cy7-conjugated anti-Sca1 (D7); eFluor 450-conjugated anti-CD8a (53-6.7), anti-CD48 (HM48-1); APC-conjugated anti-CD4 (GK1.5), anti-CD16/32 (93); and APC-eFluor 780-conjugated anti-c-kit (2B8) were purchased from eBioscience. FITC-conjugated anti-CD19 (1D3), anti-CD11b (M1/70), anti-CD11c (N418), anti-Gr1 (RB6-8C5), anti-NK1.1 (PK136), anti-Ter119 (Ter119); Pacific Blue-conjugated anti-CD45 (I3/2), anti-CD11b (M1/70); and Alexa Fluor 647-conjugated anti-CD11c (N418) were produced in-house (AbLab Biologics, UBC). Data were acquired on an LSR II flow cytometer (BD Biosciences) and analyzed with FlowJo software (TreeStar).

CFU assay

BM cells were isolated as above using proper aseptic technique. Following manufacturer instructions, 1.5 × 10⁴ cells from each sample were plated in duplicate in Methocult Media (STEMCELL M3234, Vancouver, BC) supplemented with 50 ng/mL SCF, 10 ng/mL IL-3, and 10 ng/mL IL-6. Plates were incubated for approximately 10 days prior to morphological analysis and differential counts of CFUs.

Statistics

Data are presented as means ± SEM. Statistical significance between two groups was determined by the Student's t-test while comparisons between 3 or more groups were made by ANOVA with a Bonferroni post-hoc test using GraphPad Prism software. Results were considered statistically significant with a *p < 0.05, **p < 0.01, or ***p < 0.001.