Diverse Activity of IL-17+ Cells in Chronic Skin and Mucosa Graft-Versus-Host Disease

Excessive inflammatory environment in a course of chronic graft-versus-host disease (cGvHD) is associated with T-cell trafficking into inflamed tissues. This study focused on the identification of IL-17-producing cells in the tissue biopsies of cGvHD patients. Forty-one biopsy specimens of cGvHD lesions of the skin (n = 27), gastrointestinal tract (n = 9) and oral mucosa (n = 5), examined in 24 patients, were morphologically defined according to the NIH criteria and analyzed for the presence of cellular infiltrations including: IL-17+, FOXP3+ and CCR6+ cells. IL-17+ cells were identified in 26/27 skin and in all gut and oral mucosa biopsies, being more frequent in mucosa lesions than in the skin (11/14 vs 14/26, respectively; NS: not significant). Double staining documented that CD138+/IL-17+ cells were commonly seen in the gut than in the skin (5/8 vs 3/11, respectively; NS). In the skin, cells expressing trafficking receptor CCR6+ were more frequent than IL-17+ cells compared to the mucosa (23/26 vs 2/13, respectively; p < 0.0001). CCR6 was present on a majority of IL-17+ cells in all examined skin biopsies but only in 6 out of 11 digestive tract biopsies (p = 0.0112). FOXP3+ cells were identified only in five patients (with mild lesions) at least in one biopsy. In this study group, results documented that local expansion of IL-17-producing cells in the digestive tract correlate with moderate and severe clinical symptoms of cGvHD, in contrast to the skin, where IL-17+ cells are rather scarcely present (p = 0.0301) and the course of cGvHD is slowly progressing with final organ deterioration.


Introduction
Graft-versus-host disease (GvHD) is a life-threatening complication post-hematopoietic stem cell transplantation (HSCT) (Arora et al. 2016;Filipovich et al. 2005). This acute disease has a rather rapid course and even when responding well to the therapy may progress into a chronic disease. However, in some cases, chronic GvHD (cGvHD) is not preceded by acute disease. Chronic GvHD independently, whether it constitutes de novo manifestation of alloreactivity or follows acute GvHD (aGvHD), may have a diverse course from mild and limited to progressive and extensive disease. A progress of the disease in clinical staging is associated with deterioration in the immune system function. Therefore, cGvHD appears to be a syndrome of the immune system deregulation which is skewing the differentiation of the naïve CD4 + lymphocytes into effector cell activity at the expense of regulatory T cells (MacDonald et al. 2017;Sakaguchi et al. 2008). Inflammation, which leads to the destruction of affected tissues in a course of GvHD, is associated with pro-inflammatory cytokines production.
It is known that interleukin (IL)-6 is produced at the site of inflammation in the GvHD process and its blood level is high and reflects severity of the disease (Karabon et al. 1995;Liu et al. 2010;Varelias et al. 2015). In the steadystate conditions, IL-6, if present in the milieu of differentiating lymphocytes, favors Th17 lineage. IL-17, secreted by Th17 lymphocytes, is a cytokine with pleiotropic functions important for host defense to bacterial and fungal infections especially in the gut (Geddes et al. 2011). It has been reported that IL-17 is the most potent inflammatory mediator in autoimmune-mediated tissue damage and in tissue injury after transplantation (Chung et al. 2012;Fujino et al. 2003;Gigante et al. 2011). In HSCT patients, IL-17 level increases in blood and at the tissue site involved in acute or cGvHD (Chen et al. 2007;Dander et al. 2009;Hill et al. 2010;Koh et al. 2016;Liu et al. 2013;Ratajczak et al. 2010). Our own study on IL-17 documented that Th17 cells are rather at a low level in the blood at the onset of aGvHD likely being marginalized in the affected tissues (Dlubek et al. 2010). To verify the hypothesis on the local accumulation of Th17 cells in the inflamed tissues, we focused on the identification of IL-17-producing cells in the tissue biopsies of cGvHD patients.

Patients and Methods
Two hundred and twenty patients after allogeneic HSCTs (125 unrelated donors and 95 family matched donors) from our department were enrolled in this study. Clinical records from years 2004-2012 were collected retrospectively according to European Group for Blood and Marrow Transplantation guidelines by transplant physicians from our department.
Diagnosis and treatment of HSCT patients were standardized according to the National Institute of Health (NIH) criteria (Filipovich et al. 2005;Jagasia et al. 2015). In this retrospectively examined group, cGvHD developed in 81 patients. In 39 patients, cGvHD appeared as a de novo disease, but in 42 it was preceded by aGvHD. Forty-two cases had limited/mild and 39 extensive/severe cGvHD. Patients with mild/local disease were usually left untreated or if needed, topical treatment with or without low-dose oral cyclosporine A (CsA) and prednisone was installed.
The standard treatment procedure in extensive/severe disease involved three immunosuppressive drugs: CsA, mycophenolate mofetil (MMF) and prednisone. All patients on immunosuppression or having a low number of CD4 + cells (< 200/µl) received co-trimoxazol (alternatively pentamidine), imidazole and acyclovir prophylaxis.
Fifty-seven patients responded well to the therapy with at least stabilization of the local tissue involvement not progressing to the disability. Twenty-four patients (14 male and 10 female) had the progressive disease and the target organs were biopsied to assess the local inflammatory process activity. The clinical characteristics of this group of patients are summarized in Table 1. Among 24 patients, for whom biopsies were available, 15 received HSCTs from matched sibling (including one family haplotype matched) and 9 patients from unrelated donors (8 patients 10/10 matched at the allele level, 1 mismatched in one allele).
cGvHD was diagnosed from 79 to 1436 days post-HSCT (median: 270 days) and was assessed as mild (n = 6), moderate (n = 11) and severe (n = 7) grade. In 13 cases, cGvHD followed aGvHD, whereas in 11 patients, cGvHD appeared de novo. The clinical overview of this group of patients is summarized in the Table 2.
At the time of biopsy, in 18 out of 24 patients, cGvHD developed in spite of immunosuppressive pharmacotherapy installed either as a continuation of aGvHD treatment or cGvHD prophylaxis; patients received CsA and/or MMF or methylprednisolone (Table 1) according to the best individual response to treatment. In six patients with clinical symptoms of cGvHD, biopsy was taken prior to return/initiation of immunosuppressive therapy.

Specimens and Immunohistochemistry
Altogether 41 biopsies were taken from sites with apparent cGvHD lesions of the skin (n = 27), the gastrointestinal tract (n = 9) and the oral mucosa (n = 5) were fixed in 10% formalin and embedded in paraffin blocks. Hematoxylin and eosin (H&E) staining was performed to classify the degree of morphological lesions according to the criteria proposed by NIH (Shulman et al. 2006(Shulman et al. , 2015. Inflammatory infiltrates were graded as follows: (1) mild when scattered lymphocytes were seen, (2) moderate when 20-30 lymphocytes were present and (3) extensive when lymphocyte clustered.
For immunostaining, 5-µm-thick sections were deparaffinized in xylene and ethanol gradient and placed for 30 min in preheated to 95 °C Target Retrieval Solution  (Dako, Glostrup, Denmark): citrate buffer pH = 6.0 or Tris/EDTA buffer pH = 9.0 according to the recommended procedure.

Two Color Staining
For double immunostaining, two methods, immunoenzymatic and immunofluorescence were used. Paraffin sections were deparaffinized as described above. For the presence of IL-17 + /CD138 + cells, MoAb mouse anti-human CD138 and polyclonal antibody anti-IL-17 (goat IgG) and EnVision TM G/2 Doublestain System (DAB + /Permanent Red) kit (Dako, Glostrup, Denmark) were used according to the manufacturer's recommendation. IL-17 + cells were revealed using DAB chromogen, CD138 + cells were visualized using Permanent Red substrate, counterstained in hematoxylin and mounted in Faramount medium.
For immunofluorescence staining, the tissue sections were incubated with anti-IL-17 polyclonal antibody for 1 h, washed in PBS, and then rabbit anti-goat Ig/FITC (Dako, Glostrup, Denmark) was applied for 30 min. Next, the slides were incubated with MoAb mouse anti-human cytokeratin (clone MNF 116; Dako, Glostrup, Denmark) or CCR6 for 30 min, and the binding of antibodies was revealed by antimouse Ig/TRITC (Dako, Glostrup, Denmark) for 30 min. Slides were mounted in fluorescent mounting medium supplemented with 4′,6-diamidino-2-phenylindole for DNAspecific counterstaining.
Images were analyzed and recorded using immunofluorescence Olympus BX41 microscope, equipped with XC50camera and CellB software (Olympus, Japan).

Statistics
Significances of differences of results were calculated using Fisher's exact test and Statistica StatSoft Version 6.1 program. Differences between groups were considered significant at p < 0.05.

Profile of Mononuclear Cells Infiltrations
In the skin CD4 + prevailed over CD8 + lymphocytes in 10 out of 27 specimens, in contrast, in the oral mucosa and gut biopsies CD8 + predominated over CD4 + lymphocytes in all examined tissues (p = 0.0085) (Figs. 1, 2, 3, 4). Among mononuclear cells, a small proportion of cells of lymphoplasmacytoid morphology with CD138 + phenotype were present. These CD138 + cells were seen frequently in the gastrointestinal tract specimens (8/9 biopsies) than in the skin (11/27 biopsies; p = 0.0198) (Fig. 5A). In oral mucosa biopsies, CD138 + cells were detected in two specimens only as single scattered cells. A majority of lymphocytes infiltrating affected tissue were HLA-DR positive. Moreover, numerous small vessels in the vicinity of lymphocyte infiltrations, especially in the oral mucosa, revealed HLA-DR + staining (Fig. 2).

IL-17-Producing Cells and FOXP3 Cells in Tissue Infiltrates
IL-17 + cells were seen in both gut and oral mucosa (14/14) and skin biopsies (26/27), but they were more frequent in the mucosa biopsies (Fig. 5a). The presence of an increased number of IL-17 + cells (according our grading a total of six or more cells in five HPF) was seen in 11 out of 14 gut and oral mucosa biopsies and in 14 out of 26 skin biopsies (NS: not significant) (Fig. 5a). In the oral mucosa, IL-17 + mononuclear cells were seen in the context of extensive lymphocytic infiltrates (Fig. 2). In the gut specimens, IL-17-producing cells were localized around the damaged crypts (Fig. 3).
The number of IL-17 + cells in the affected tissue was associated with the clinical grade of cGvHD. Skin lesions in the course of cGvHD was rather associated with mild stage of disease, whereas moderate and severe symptoms of disease developed when gastrointestinal tract and oral mucosa were affected (p = 0.0301) (Fig. 5b).
Double staining with anti-cytokeratin or CD45 and anti-IL-17 antibodies documented that IL-17 activity was exclusively associated with cellular infiltrates (Fig. 1). Double staining for CD138/IL-17 revealed that some of CD138 + cells were IL-17 positive and were commonly seen in the gut than in the skin (5/8 vs 3/11, respectively, NS; Fig. 1 and Fig. 3). Double positive CD138 + /IL-17 + cells were not detectable in the oral mucosa biopsies.
In five patients, FOXP3 positive cells were present at least in one examined biopsy. Notably, none of these five patients had extensive lesions which were seen in 7 out of 19 patients lacking FOXP3 + cells at the tissue site.

IL-17-Producing Cells and Cells Expressing CCR6 Trafficking Receptor in the Tissue Infiltrates
IL-17 + lymphocytes respond to the pro-inflammatory signals with the use of chemokine receptor CCR6. In all specimens, a proportion of mononuclear cells infiltrating the affected tissues were CCR6 positive. IL-17 + and CCR6 + cells, counted independently in the parallel sections in the skin, revealed more CCR6 + than IL-17 + cells (23/26 biopsies); in contrast, in the digestive tract IL-17 + cells prevailed over CCR6 + (11/13 biopsies; p < 0.0001) (Figs. 1, 6).
Double immunostaining (performed in 21 skin biopsies) showed that CCR6 + co-expressed with IL-17 + cells in 20 Fig. 3 Rectum biopsy specimen obtained on day 141 post-HSCT from patient with moderate course of cGvHD having digestive tract mucosa lesions and diarrhea. Note, cell infiltration was composed of CD3 + cells and among them the CD8 + prevailed over CD4 + . In the tissue infiltrates, IL-17-producing cells and cells expressing CCR6 were present. A proportion of cells were CD138 + and a double staining documented that some CD138 + cells of lymphoplasmacytoid morphology were also IL-17 positive (arrows); (red staining with Permanent Red, brown staining with diaminobenzidinie-tetrahydrochloride (DAB); double stain with DAB/Permanent Red, magnifications: × 400, except isotype control: × 200) skin biopsies, but only in 6 out of 11 digestive tract biopsies (p = 0.0112) (Fig. 6).

Discussion
According to the NIH criteria, acute and chronic GvHD are distinguishable rather by clinical and histological criteria but not according to the time elapsing after HSCT (Jagasia et al. 2015;Shulman et al. 2015). Severe clinical symptoms of cGvHD developed more often in patients who experienced prior aGvHD compared to these patients when the disease developed de novo. Most of the cGvHD patients examined in this study (92%) received G-CSF mobilized HSCT, which clinically has led to rapid hematopoietic reconstitution, but increased the risk of cGvHD as confirmed in clinical observations (Bensinger et al. 2001) and in experimental studies (Hill et al. 2010). The histopathology of cGvHD lesions seen in our examined group were concordant with prototype pictures introduced by NIH criteria (Shulman et al. 2006). Previous reports, including own studies on the contribution of IL-17-producing cells in aGvHD, documented an increased number of IL17 + cells during immune reconstitution and decreased number of IL-17 + cells in the circulation at the time of clinical manifestation of aGvHD (Broady et al. 2010;Dlubek et al. 2010;van der Waart et al. 2012), and this suggest that they may have migrate to the affected tissues. Very few studies introduced tissue distribution of IL-17 + cells, in limited cases, in a course of cGvHD (Dander et al. 2009;Malard et al. 2014;Ratajczak et al. 2010) and these Fig. 5 a IL-17 + cells were present in all gastrointestinal tract (GIT) and oral mucosa (OM) biopsies and in 26 skin specimens. A small proportion of cells with CD138 + phenotype were present more frequently in the GIT than in the skin and were IL-17 + /CD138 + double positive. b The number of IL-17 + cells in the affected tissue is associated with a clinical grade of cGvHD. Mild stage of disease was diagnosed more frequently when skin lesions were observed, whereas moderate and severe symptoms of cGvHD were developed when GIT and OM were affected (p = 0.0301; Fisher's exact test) studies documented that Th17 cells may contribute to tissue damage. Detailed analysis of the present study documented that IL-17 + -producing cells in a course of cGvHD plays a diverse role in different tissue compartments and were more frequent in the gut and oral mucosa than in the skin lesions. This would suggest that the prevalence of Th17 + cells in the digestive tract reflects a rather physiological representation of these cells in the gut as compared to the skin. Th17 cells are employed in both in physiological and pathological conditions. Physiologically, Th17 cells are seen along the mucosa barrier in the gut and they play an important role in the gut epithelial homeostasis and in the regulation of host immune response against a variety of pathogens (Asigbetse et al. 2010;Blaschitz and Raffatellu 2010). Studies on experimental model of germ-free or antibiotic-treated mice showed that in the absence of a luminal commensal microbiota, the number of Th-17 cells is significantly reduced (Atarashi and Honda 2011). Patients receiving conditioning regimen prior to transplantation are colonized by a number of bacterial species. In the view of this data, we can assume that the presence of IL-17 + cells at the site of cGvHD is due to the response to microorganisms rather than that to alloantigens. The commensal microbiota may have immunomodulatory influence on Th17 cells. Microbial products provide strong stimulation for local IL-6 production (Chung and Kasper 2010) which in the presence of transforming growth factor β facilitate differentiation of antigen-activated CD4 T cells into Th17 lymphocytes. Th17 immune response is associated with secretion of IL-17 and IL-22, neutrophils recruitment, and anti-microbial peptides induction in the early phase of inflammation, and constitutes a link between innate and adaptive immunity (Chung and Kasper 2010;Klimczak and Lange 2012). IL-17 is also considered as an innate cytokine produced by NK cells, γδ T cells, and CD8 + T lymphocytes (Geddes et al. 2011). Our studies also documented prevalence of CD8 + cells within the infiltrateaffected gut mucosa. However, it is also proposed that the expansion of Th17 cells into active phase of cGvHD is primarily due to a reduced frequency of Regulatory T (Treg) cells as confirmed in experimental model and in clinical observations (Chen et al. 2007;Dander et al. 2009;Rieger et al. 2006;Zorn et al. 2005). A defect of regulatory mechanism was also confirmed by the increased ratio of Th17/Treg in the human liver affected by cGvHD (Malard et al. 2014).

B C
Indeed, in the present study, the frequency of FOXP3 + cells in the affected tissues was low and did not follow the expansion of Th17-producing cells. However, the situation is distinct in the lichen planuslike oral mucosa lesions. In contrast to cGvHD of the skin which is slowly deteriorating, in the oral mucosa, IL-17 + cells were often evident, and the intensity of infiltration of the IL-17 + lymphocytes was associated with severity of the clinical symptomatology. Patients with constant pain, hampering normal feeding, had a higher number of IL-17 + cells than those with milder clinical symptomatology of the oral lesions. Our observations documented that IL-17-producing cells participate in the active inflammatory process contributing to the tissue injury with painful symptomatology when oral mucosal lining is affected.
In contrast, the histological picture was more variable in the skin, beginning from scarce or denser papule eruptions to sclerodermic-like lesions. Active inflammatory process recognized as papullo-erythemic lesions is characterized by the presence of lymphocyte infiltrates and some of the cells were IL-17 + , but they were much fewer in proportion than in the mucosa lesions. The profile of lymphocytes concomitant with Th17 also differs when we compare the skin and the gut lesions. In the skin, there are a large proportion of CD4 + lymphocytes, whereas in the gut, evident prevalence of CD8 + cells is visible. This observation may also add new insight to the understanding of the mechanism of cGvHD in different compartments, showing that in the skin and oral mucosa CD4 + lymphocytes contribute to the clinical symptoms by producing pro-inflammatory cytokines including (e.g., IL-17, interferon (IFN)-γ), but in the gut, cytotoxic attack on gut mucosa creates the clinical feature. It has been reported that after HSCT, a proportion of IL-17-producing T cells co-produce IFN-γ and these cells have the ability to differentiate into Th1 phenotype (Annunziato et al. 2007;Bahr et al. 2013;Gartlan et al. 2017). Therefore, IL-17-producing T cells, with or without IFN-γ-secreting properties, may contribute to GvHD without numerical expansion but via differentiation into Th1 phenotype (Bruggen et al. 2014;Dander et al. 2009;van der Waart et al. 2012).
According to the NIH criteria, one of the hallmarks of cGvHD is the presence of lymphoplasmacytoid cells (Shulman et al. 2006). In this study, we added novel information to the clinical role of these cells in a course of cGvHD. We proved that a proportion of CD138 + cells having lymphoplasmacytoid characteristics produce IL-17 and these were seen only in the gut. We believe that the presence of IL-17-producing cells in the gastrointestinal tract in cGvHD reflects the natural characteristic of the gut immune system which frequently employs these cells in the immune response. CD138 + cells were less frequent but still present in the skin lesions and they were IL-17 negative. Labeling for trafficking receptor CCR6 revealed that IL-17 + cells in the skin, but not in the gut, were CCR6 positive. Therefore, we can assume that CCL20 (the only ligand for CCR6) which is produced by macrophages involved in the local inflammatory process, is responsible for the attracting these IL-17 + /CCR6 + cells to the inflammatory sites (Nakayama et al. 2001;van der Waart et al. 2012). This documented that CCR6 expression on the skin IL17 + cells plays a role as a homing receptor helping migration of IL-17-producing cells to the inflamed tissue (Singh et al. 2008;Wang et al. 2009).
Limitation of this study is the inadequate number of biopsies collected from oral mucosa or gut to compare the results with the same number of skin biopsies to identify immunopathological differences between different tissues affected by cGvHD. For this reason, this retrospective study, performed on a limited number of single-center patients, is considered as a pilot study and further studies need to be performed to assess the impact of IL-17-producing cells on tissue damage in the course of cGvHD. However, despite these limitations, the present study adds a novel observation for tissue-specific characteristics of cGvHD and shows that the physiological immune system has influence on the composition of lymphocyte infiltration in the affected organs. In the digestive tract, IL-17-producing cells are strongly represented and their presence is associated with clinical symptoms (pain, diarrhea) of cGvHD. In contrast, in the skin, IL-17 + cells are rather scarcely present and the course of cGvHD is slowly progressing with final organ deterioration.