Dysregulated hematopoietic stem and progenitor cell activity promotes interleukin-23-driven chronic intestinal inflammation

Abstract

In interleukin-23 (IL-23)-dependent colitis, there is excessive accumulation of short-lived neutrophils and inflammatory monocytes in the intestine. It is unknown whether this reflects changes in mature cell populations or whether the IL-23-driven colitogenic T cell program regulates upstream hematopoietic stem and progenitor cells (HSPC). Here we have shown dysregulation of hematopoiesis in colitis mediated by inflammatory cytokines. First, there was an interferon-gamma-dependent accumulation of proliferating hematopoietic stem cells in the bone marrow and spleen. Second, there was a strong skew toward granulocyte-monocyte progenitor (GMP) production at the expense of erythroid and lymphoid progenitors. Extramedullary hematopoiesis was also evident, and granulocyte macrophage-colony stimulating factor (GM-CSF) blockade reduced the accumulation of splenic and colonic GMPs, resulting in amelioration of colitis. Importantly, transfer of GMPs exacerbated colitis. These data identify HSPCs as a major target of the IL-23-driven inflammatory axis suggesting therapeutic strategies for the treatment of inflammatory bowel disease.

Graphical abstract

Figure 1

Differential Regulation of Hematopoietic Lineages in Colitis Colitis was induced by transfer of naive CD4⁺CD45RB^hi T cells into C57BL/6.Rag1^−/− mice. Colitic mice were killed 8 wk after transfer (termed colitic) and compared with untransferred Rag1^−/− mice (termed control). (A) Total cell numbers in spleen, bone marrow (one leg, femur and tibia), MLN, and colon (colonic lamina propria leukoctyes, cLPL) (±SEM, n = 5 mice per group). (B) Representative flow cytometry staining of neutrophils (CD11b⁺Gr1^hiSSC^hi) and inflammatory monocytes (CD11b⁺Gr1^intSSC^lo) from spleen, MLN, and colon (cLPL) cell suspensions. (C) Representative staining of neutrophils, monocytes, NK cells (NKp46⁺), erythroid cells (Ter119⁺), and immature B cells (B220^int) from bone marrow cell suspensions. (D) Representative Gr1, Ter119, and NKp46 staining from spleen cell suspensions. Data are representative of three independent experiments.

Figure 2

Increase in Proliferative HSCs in the Bone Marrow of Colitic Mice and Skewed Differentiation toward GMPs Rag1^−/− mice received either naive T cells alone (termed colitic) or in combination with CD4⁺CD25⁺ Treg cells (termed protected) or were untreated (termed control). Transferred mice were killed 8 wk after transfer. BM cells were stained for HSPCs (see Figure S1 for population FACS-definitions). (A) Representative flow cytometry staining and percentages of LSKs (Lin⁻Sca-1⁺c-Kit^hi cells, highly enriched for HSC) among BM Lin⁻ cells. In some experiments, BM cells were stained for the proliferative marker Ki-67 (right panel). (B) Percentages, absolute numbers, and proliferation of LT-HSCs (CD34⁻LSKs) in the BM. (C) Representative flow cytometry staining of CMP, GMP, and MEP population in the BM (left). Absolute numbers of GMPs (right) and percentages among Lin⁻Sca-1⁻c-Kit^hi MP cells (middle). (D) Flow cytometry staining of IL-7R expression among Lin⁻cells in the BM (left) and c-Kit and Sca-1 expression among Lin⁻IL-7R⁺ cells (middle). Percentages of CLP (Lin⁻IL-7R⁺Sca-1^intc-Kit^int) among Lin⁻ cells (right). Numbers in dot plots indicate percentages among gated cells. Each point represents an individual mouse and horizontal bars represent group means (A–D). Data are representative of three independent experiments. Error bars represent SEM.

Figure 3

Accumulation of HSC and GMP in the Spleen of Colitic Mice Rag1^−/− mice received either naive T cells alone (colitic) or in combination with Treg cells (protected) or were untreated (control). Transferred mice were killed 8 wk after transfer. (A) Representative flow cytometry staining, percentages, absolute numbers, and proliferation (Ki-67) of LSK cells in the spleen. (B) Flow cytometry staining of splenic CMPs, GMPs, and MEPs. Percentages (middle) and numbers of splenic GMP (right). (C) Cfu activity. Frequency of clonogenic cfu-GM (cfu-granulocyte and macrophage) among total splenocytes. Data are representative of three independent experiments. Error bars represent SEM.

Figure 4

GMPs Accumulate in the Inflamed Intestinal Mucosa and Trigger Extramedullary cfu Activity in the Colon cLPL were prepared from control, colitic, and protected mice. Transferred mice were killed 8 wk after T cell transfer. (A) Representative CMP, GMP, and MEP staining in cLPL. Percentage and absolute number of colonic GMPs. (B) Percentages of proliferative cells (Ki-67⁺) among MEPs and GMPs from cLPL of colitic mice. (C) G-CSF, GM-CSF, SCF (c-KitL), and IL-23p19 mRNA amounts in total colon homogenate (n = 4). Values are normalized to HPRT and are mean values (±SEM). (D) Cfu activity from plated cLPL. Representative field from photomicrographs of Methocult plate with typical images of cfu (magnification ×10). Frequency of cfu-GM among total cLPL (right). (E) Representative c-Kit and Sca-1 staining gated on Lin⁻ cells from BM and cLPL of colitic mice (left). Number of cfus from 500 plated Lin⁻ cells isolated by cell sorting from BM or cLPL of 5 colitic mice; the various Lin⁻ FACS-sorted populations are identified on the horizontal axis (right). (A–D) Data are representative of two to three independent experiments.

Figure 5

Extramedullary Accumulation of GMPs during Colitis Is Dependent on Local Tissue Inflammation C57BL/6.Rag1^−/− mice were injected daily with G-CSF for 3 days and killed 24 hr after the last G-CSF injection (termed G-CSF treated) and compared with Rag1^−/− mice killed 8 wk after CD4⁺CD45RB^hi T cell transfer (colitic) and with untransferred Rag1^−/− mice (control). (A and B) Representative flow cyometry staining of LSKs (upper panels) and GMPs (lower panels) in the BM (A) and the spleen (B) (individual mice, n = 4). (C) Representative flow cytometry staining of GMPs in colonic LPL is shown on left (n = 4). Colonic cfu-GM activity is shown on right. Data are representative of two independent experiments.

Figure 6

IFN-γ Promotes Exacerbated HSC Activity during IL-23-Driven Colitis Innate colitis was induced by H. hepaticus (Hh) infection of 129SvEv Rag2^−/− mice. Mice were killed 8 wk (colitic) in (A) or 2 wk (Hh) in (B) and (C) after H. hepaticus feeding and compared to untreated Rag2^−/− mice (control) or untreated Rag2^+/+ mice where indicated. (A) Representative Gr1, Ter119, and NKp46 staining in the BM is shown on left. Frequency of clonogenic cfu-GM among total splenocytes and cLPL is shown on right. (B) Infected mice were either injected weekly with anti-IFN-γ (Hh + anti-IFN-γ) or left untreated (Hh). Flow cytometry analysis and percentages of LSKs in BM (upper panels). Percentages and absolute numbers of LT-HSCs in BM (lower panels). (C) FACS analysis, percentages, and absolute numbers of LSKs in spleen. (A–C) Data are representative of two independent experiments.

Figure 7

GM-CSF Promotes Pathogenic Accumulation of Peripheral GMPs and IL-23 Dependent Colitis (A and B) Rag1^−/− mice transferred with naive T cells were injected two times per week with either anti-GM-CSF mAb or isotype control, killed 8 wk after T cell transfer, and compared with untransferred control mice. (A) Representative staining, frequencies, and absolute numbers of GMPs in the spleen. (B) Score of colitis and photomicrographs of proximal-colon sections (upper panels). Original magnification 100×. GMP and neutrophil numbers in the cLPL (lower panels). (A and B) Data are representative of two independent experiments. (C) Left shows that Rag1^−/− mice were cotransferred with 1:1 mixtures of CD45.1⁺ (WT) + CD45.2⁺ (Il23r^−/−) naive CD4 T cells to assess whether IL-23 has a cell intrinsic effect on GM-CSF production by Th cells as described in (Ahern et al., 2010). CD45.1⁺ cells were isolated from congenic B6.SJL-Cd45.1 and CD45.2⁺ cells from Il23r^−/− mice. Mice were killed 4 wk after T cell transfer and percentages of IL-17A⁺ and GM-CSF⁺ cells in colonic CD4⁺ T cells were assessed by intracellular flow cytometry. Right shows that WT and Il23r^−/− C57BL/6 mice were infected with H. hepaticus and treated with anti-IL10R. WT Mice were killed when colitic (week 2) and compared with Il23r^−/− protected mice and unmanipulated WT mice for GM-CSF mRNA amounts in total colon homogenate (n = 4). (D) Rag1^−/− mice that had received naive T cells (“RB^hi”) 14 days before were injected intravenously with PBS or 1 × 10⁴ sorted splenic GMPs or MEPs, where indicated. GMPs and MEPs were sorted from the spleen of colitic mice (8 wk post-T cell transfer). Shown are weight loss and colitis score 14 days after PBS, GMPs, or MEPs injection and frequencies of neutrophils and CD4⁺ T cells in cLPL. Weight loss is representative of two independent experiments, and colitis score is pooled from two independent experiments. Error bars represent SEM.