IDO in human gut Graft-versus-Host Disease

Abstract

While rodent graft-versus-host disease (GVHD) models have suggested that indoleamine 2, 3-dioxygenase (IDO) is a critical regulator of gastrointestinal GVHD parallel human studies on IDO expression have not been reported. IDO expression was assessed in 20 patients who underwent duodenal biopsy. IDO was upregulated in epithelial cells. In situ analyses reveal that macrophages and dendritic cells stain positive for IDO, but that most of the IDO+ cells were a novel population of CD3+CD4+IDO+ cells. The proportion of CD4+IDO+ T-cells was significantly higher in patients with moderate GVHD. In situ regulatory T-cell and Th17 numbers correlated with overall severity. Although needing confirmatory results from larger sample sets, these data are consistent with the hypothesis that IDO is involved in regulating gastrointestinal GVHD.

Introduction

Wider application of allogeneic hematopoietic stem cell transplantation (HSCT) is restricted due to graft-versus-host disease (GVHD). The environmental effects that influence GVHD injury and the natural regulatory mechanisms by which GVHD is controlled at the tissue level are incompletely understood ¹.

Indoleamine 2, 3-dioxygenase (IDO) catalyzes the first and rate-limiting step of tryptophan catabolism. Decreased tryptophan and/or increased metabolite concentrations elicit T-cell anergy or apoptosis [reviewed in ²]. Several aspects of IDO biology make it an intriguing target for transplantation and tolerance [reviewed in ³].

Previous experimental GVHD models in rodents by us ^{4, 5} and others (^6-8) dissected the role of IDO in GVHD. We previously reported^{4, 5} that: 1) IDO is a critical regulator of GVHD, most strikingly in the colon, 2) IDO can act at the site of colon expression to decrease T-cell proliferation and survival, diminishing inflammation and reduce disease severity, and 3) Host IDO expression in both epithelial cells and APCs affected GVHD severity. In humans, only one in vitro study reported IDO expression by RT-PCR of monocytes obtained from acute GVHD patients ⁹.

Herein we report the first study which assayed IDO protein expression in patients with gastro-intestinal (GI) GVHD.

Patients and methods

Patients

Biopsies were performed as diagnostic procedures for digestive symptoms¹⁰. Hospital Saint-Louis ethical review board approved the design of this study. A cohort of allografted patients (total n=2O) was studied after myeloablative conditioning. Their median age was 34 years. Samples were obtained from: 8 patients (7 males, 1 female; median age 30 years) without pathological GI GvHD (grade 0-1); 5 patients (2 males, 3 females; median age 47 years) with pathological GI GvHD grade 2, and 7 patients (5 males, 2 females; median age 34 years) with pathological GI GvHD grade 3-4. All patients underwent biopsy before being treated with steroids and thus only received cyclosporine as GvHD prophylaxis. Five non-transplanted patients, active Helicobacter pylori infection, were used as controls.

Methods

Duodenal biopsies during fiber-optic examination were performed, as described^{10, 11}. All biopsies were performed before any steroid treatment. The histological grading of acute GVHD according to Sale ¹² and Epstein¹³. None had evidence of digestive or systemic viral, bacterial or fungal infections.

Immunostainings were performed with Ventana Discovery reagents (Ventana, Arizona). For IDO staining, polyclonal rabbit antibody (gift from D. M; dilution 1:500) was used and detected using DabMAP kit. Specificity of IDO staining was confirmed by incubation of sections with neutralized antibody (IDO antibody incubated with blocking peptide for 2 hours at 4°C prior to incubation with sections).

Double immunofluorescent staining was performed on frozen sections for IDO/CD4, IDO/CD68, IDO/CD3, IDO/CD123, Foxp3/CD4, IL-17/CD4. Treg and Th17 double stainings were performed as described¹⁴. Additional staining was performed using anti-CD68 (clone KP-1, DAKO), anti-CD3 (Biocare), and anti-CD123 (clone 6H6, Biolegend), antibodies. Endogenous peroxidase inhibition and nonspecific binding sites blocking were systematically performed. Controls with irrelevant isotypic antibodies, and absence of primary antibody were systematically performed.

Antibodies were covalently linked to Alexafluor 488 or Alexafluor 594 using APEX Antibody Labeling Kits (Invitrogen). Sections were incubated in PBS pH 7.4 containing 5% bovine serum albumin for 30 min at room temperature (RT). Double immunofluorescent stainings were performed by incubation with antibodies applied to sections for 1 hour at RT. Sections were finally mounted in Vectashield medium containing DAPI. Cells counts and proportions were expressed as median and inter-quartile range (IQR) of the number or proportion of cell sper field at 400x magnification.

Samples from patients were identified by anonymous 7 digits code corresponding to laboratory identification. Biopsies were independently evaluated by 2 examiners (PR and AJ). In all cases of disagreement between examiners, a common reading was organized to achieve a consensus on count. For automated counts of epithelial cells expressing IDO, Regions of Interest (ROI) corresponding to epithelium and excluding infiltrated cells were independently defined by two pathologists (PR and AJ) using CellSens software (Olympus, Rungis, France). For each patient, the average number of epithelial cells expressing IDO was determined on five different fields at 400x magnification. Automated counts of CD4⁺ lymphocytes expressing IDO was performed using CellSens software (Olympus, Rungis, France) on double immunolabelled sections with CD4 and IDO. For each patient, the average number of CD4⁺ lymphocytes expressing IDO was determined on five different fields at 400x magnification.

Statistical analyses

Reproducibility of counts was assessed through examination of discrepancy levels between the examiners and intraclass correlation estimate with 95% confidence interval. The median level and range of CD4, CD8, CD68, CD4+IDO+, CD68+IDO+, Treg, and Th17 characteristics (cell number or their ratio) are presented. For the comparison on cell count or proportion of Treg, Th17, CD4+IDO+, CD68+IDO+, CD4+, CD68+, IDO+ epithelial cells between the group without GVHD (Grade 0-1) and mild GVHD (Grade 2), between the groups with mild GVHD and severe GVHD (Grade 3-4), or between the groups of GVHD and Helicobacter pylori group, Mann-Whitney non parametric test was used. The correlation between CD4+IDO+ proportion and Th17 or Treg proportion, a comparison was performed using Z-test.

Results and discussion

Figure 1A shows expression of IDO in duodenal intestinal cells with a morphology and location consistent with epithelial cells, which was up-regulated in specimens obtained from patients with GVHD as compared to those without GVHD. IDO expression in the epithelial cells of non-transplanted controls with Helicobacter pylori infection was also detected. These data are consistent with those in rodents and suggest that IDO is upregulated during the GVHD process.

Figure 1. IDO distribution and expression in gut biopsies of transplanted patients without GvHD and with mild and severe GvHD. Control: non-transplanted patients infected by Helicobacter pylori.

A, Upper panel: IDO expressing cells were more numerous in patients without GvHD or with Grade 2 GvHD than in patients with Grade 3 GvHD. Lower panel, specificity control: pre-incubation of the antibody directed against IDO with the corresponding blocking peptide resulted in absence of staining.

B, Double positive IDO (red)/ CD4 (green) cells are absent in biopsies with Helicobacter pylori, present in transplanted patients without GvHD, are more numerous in transplanted patients with Grade 2 GvHD than with Grade 3 GvHD.

In addition to IDO expression in epithelial cells in the duodenum of patients with GVHD, IDO positive mononuclear cells (MNCs) also were readily detected. The specificity of the staining was proven by a blocking peptide (Figure 1A). Double staining surprisingly revealed that most of these IDO+ MNCs were double positive CD4+IDO+ cells. Few MNCs were macrophages [CD68+IDO that were CD4- (Figure 2A) or CD4+ plasmacytoid DCs (CD123+IDO+] (Figure 2B). Instead CD4+IDO+ co-expressed CD3. Thus, the dominant source of IDO in MNCs was CD4+ T-cells (Figure 2B).

Figure 2. IDO expression in gut biopsies in macrophages, dendritic and T cells.

A Double positive IDO (red)/ CD68 (green) cells are absent in biopsies with Helicobacter pylori, present in transplanted patients without GvHD, are more numerous in transplanted patients with histological GvHD.

B Upper panel, double positive IDO (red)/ CD3 (green) cells are more numerous in transplanted patients with histological Grade 2 GvHD than with Grade 3. Lower panel, double positive IDO (red)/ CD123 (green) cells are few in transplanted patients whatever the histological GvHD grade.

We sought to correlate the numbers of CD4+IDO+ T-cells with other T-cell subsets and with the severity of GVHD pathological lesions. As described in Figure 3A and B there was no statistical correlation between CD4+IDO+ T-cell numbers and Treg cell subsets, although there was a statistical correlation between CD4/IL-17 and Treg cells with pathological grade of acute GVHD, as we have previously described¹⁴. Moreover, we found a positive correlation (p= 0.03) between the proportion of CD4/IL-17 and CD4/IDO lymphocytes. The proportion of CD4+IDO+ cells was significantly higher in patients with grade 2 disease, as compared to those with grade 0-1 albeit lower in patients with grade 3-4 vs. grade 2 GVHD (Figure 3A and B).

Figure 3. Counts of inflammatory cells.

A, in control biopsies with Helicobacter pylori, macrophages are prominent and some of them express IDO. In transplanted patients, T lymphocytes are prominent, among them CD4 T lymphocytes expressing IDO are predominant in patients with Grade 2 GvHD whereas Tregs are predominant in patients with Grade 3 GvHD. The highest numbers of Th17-positive cells are found in transplanted patients with Grade 0-1 GvHD.

B, Statistical analysis for CD4/IDO double positive cells in transplanted patients within the indicated histological GvHD grade (n=20); Left, no significant correlation between the proportion of CD4/Foxp3 and CD4/IDO lymphocytes; Right, positive correlation (p < 0.05) between the proportion of CD4/IL-17 and CD4/IDO lymphocytes.

C, Analysis of CD4 cell subpopulations and epithelial cells expressing IDO after automated count; the proportion of CD4/Foxp3 cells is significantly higher in Grade 2 and Grade 3-4 GvHD when compared with Grade 0-1 GvHD. The proportion of CD4/IDO cells is significantly higher in Grade 2 GvHD when compared with Grade 3-4 GvHD. A significant decrease was also found for the proportion of CD4/IL-17 lymphocytes and the histological grade of GVHD. The number of epithelial cells expressing IDO increases significantly with the grade of GVHD (i.e. Grade 0-1 versus Grade 2 versus Grade 3-4 GVHD). Epithelial cells expressing IDO are significantly numerous in Helicobacter pylori group compared with Grade 0-1 GVHD but significantly lower compared with Grade 2 or Grade 3-4 GVHD.

Thus, in this first study ever performed in human we report that IDO was up-regulated in epithelial cells in duodenal biopsy specimens of allogeneic HSCT recipients with acute GVHD at the time of first diagnosis. In situ analyses revealed that macrophages and DCs stain positive for IDO, but that most of the IDO+ cells were a novel population of CD3+CD4+IDO+ cells. The proportion of CD4+IDO+ T-cells was significantly higher in patients with moderate as compared to no acute GVHD. In situ Treg and Th17 numbers correlated with overall severity. However, since numbers of samples analyzed are small and diverse, we would like to point out that these observations would need confirmation with larger sample sets. Furthermore, IDO expression correlated only with moderate but not severe GvHD; we do not have firm conclusion to explain this and we could only hypothesize that these regulatory negative loops are exhausted by unknown mechanisms in severe forms. Duodenal biopsy was used here while most experimental data is for IDO in colonic pathology. This was based on sample availability in clinical material that is often limiting.

Consistent with our rodent GVHD data⁴ and experimental data in inflammatory bowel diseases [reviewed in ¹⁵], we now demonstrate in patients that acute GVHD is associated with IDO upregulation in both epithelial cells and APCs (macrophages, DCs) within GI specimens in the context of acute GVHD. Although previous murine models have mainly assigned IDO imunoregulatory functions through macrophages or DCs², the most preeminent cellular subset in patients with acute GVHD in fact was CD4+IDO+ in the cellular infiltrate. Interestingly, in another model of inflammation, acute simian immunodeficiency infection, a population i.e. IDO+CD3+ T-cells have been described in lymphoid tissue¹⁶. Because the type of mechanistic studies performed in secondary lymphoid organs and the GI tract obtained at sequential time points during GVHD evolution in rodents cannot readily be done in patients, however, we can only speculate that thevarious IDO positive cell types in humans may be beneficial in preventing GVHD progression.

Interestingly, in GI samples obtained from patients with the most severe form of acute GVHD, there was a relative paucity of IDO+CD3+ T-cells. We speculate that the exceedingly high degree of inflammation may have overwhelmed the IDO system, which may explain why GVHD is severe in such patients. Moreover, because IDO alters the balance of Th17 to Treg in human HIV disease¹⁷, this finding sheds new light on our previous ¹⁴, and current, counterintuitive findings that decreased Th17 and increased Treg actually are associated with more severe GvHD. In HIV-seropositive subjects Favre and collaborators, found that progressive disease is associated with the loss of TH17 cells and a reciprocal increase in the fraction of the immunosuppressive Treg cells in rectosigmoid biopsies. The loss of TH17/Treg balance was associated with induction of IDO. Thus Favre et al., postulated that induction of IDO may represent a critical initiating event that results in inversion of the TH17/Treg balance and in the consequent maintenance of a chronic inflammatory state in progressive HIV disease. We thus found similar dysregulation of the Th17/Treg balance in human GVHD. Decreased IDO production in severe GVHD has also been reported in vitro in humans⁹ in stimulated peripheral cells, but to the best of our knowledge this study is the first to study human GVHD in situ.

In conclusion, in human GVHD IDO might be involved in controlling early disease. Since IDO upregulation and the provision of tryptophan catabolites can reduce GVHD in rodents⁵, these data strengthen the possibility that such rodent studies may be translatable into the clinic.