Tolerogenic IDO1+CD83− Langerhans Cells in Sentinel Lymph Nodes of Patients with Melanoma

Langerhans cells (LCs) are crucial regulators of anti-cancer immune responses. Cancer, however, can alter DCs functions leading to tolerance. The enzyme indoleamine 2,3-dioxygenase (IDO1) plays a crucial role in this process. In sentinel lymph nodes (SLNs) of patients with melanoma, LCs show phenotypical and functional alterations favoring tolerance. Herein we aimed to investigate IDO1 expression in SLN LCs from patients with melanoma. We showed by immunofluorescence analysis that a portion of Langerin+ LCs, located in the SLN T cell-rich area, displayed the typical dendritic morphology and expressed IDO1. There was no significant difference in the expression of IDO between SLN with or without metastases. Double IDO1/CD83 staining identified four LCs subsets: real mature IDO1−CD83+ LCs; real immature IDO1−CD83− LCs; tolerogenic mature IDO1+CD83+ LCs; tolerogenic immature IDO1+CD83− LCs. The latter subset was significantly increased in metastatic SLNs as compared to negative ones (p < 0.05), and in SLN LCs of patients with mitotic rate (MR) > 1 in primary melanoma, as compared to MR ≤ 1 (p < 0.05). Finally, immature SLN LCs, after in vitro stimulation by inflammatory cytokines, acquired a maturation profile by CD83 up-regulation. These results provide new input for immunotherapeutic approaches targeting in vivo LC of patients with melanoma.


Introduction
According to the cancer immune-editing theory, tumor cells are able to evade the immune system by escape mechanisms responsible of tumor tolerance [1]. Among skin tumors, melanoma is considered the most tolerogenic one by its ability to alter immune cells functions, particularly within the sentinel lymph node (SLN), the first draining node from a tumor region, favoring SLN early metastases [2,3]. Melanoma cells create an immunosuppressive microenvironment by cytokines secretion, such as IL-10 and TGF-beta [4,5]. This melanoma-induced milieu is responsible for several morphological, immunophenotypical and functional alterations in SLN dendritic cells (DCs) of patient with melanoma, [6][7][8], as compared to DCs from healthy LN [9]. A particular DCs subset, plasmacytoid DCs (pDCs), displays functional defects in SLN of melanoma patients and accumulates in LNs with melanoma metastases [10]. Even SLN Langerhans cells (LCs), the typical DC subset of epithelia, show an immature immunophenotype, with low levels of CD83 expression, the DC maturation marker, particularly in melanoma bearing SLNs [11]. Additionally, SLN LCs are functionally defective in inducing T cell proliferation and cytokine release [12]. In the past years, the studies on tumor immune escape mechanisms led to the identification of indoleamine 2,3-dioxygenase (IDO) enzymes (IDO1 and 2), responsible of tryptophan (TRP) degradation and kynurenines (KYN) production, as key factors in tumor induced immunosuppression by direct and indirect effects on T cells [13,14]. Particularly, TRP depletion and KYN generation, through aryl hydrocarbon receptor (AhR), induce suppression of effector T cells and generation of T regulatory cells (Treg) by FOXP3 induction in the tumor microenvironment [13,14]. Consistently, Treg counts are significantly increased in peripheral blood of late-stage melanoma patients and positive SLNs [15][16][17], mediating immunosuppression mainly through adenosine generation [17].
In patients with melanoma, IDO1 expression were first observed in a subset of cells with plasmacytoid shape in SLN [18,19]. Some years later, these IDO1 + cells had been identified as pDC [20] by the specific marker BDCA-2 [21] in SLN of patients with melanoma. The presence of numerous IDO1 + pDCs in metastatic SLN suggested a correlation with substantially worse clinical prognosis in these patients [20]. Another skin derived DC subset, SLN FXIII + dermal DCs (dDCs) express IDO1, probably induced by melanoma-derived TGFb-2, and able to transform peripheral DC into tolerogenic cells [5,22]. Consistently, a negative prognostic role of IDO1 expression in SLN has been reported [23], paving the way for the therapeutic use of IDO1 inhibitors to revert tumors tolerance [13,24,25]. In addition, primary and metastatic melanoma cells can express IDO1 [26,27]. IDO1 expression in SLN LC has not been investigated yet. In healthy human skin, immature LCs do not express IDO1, while after in vitro-induced maturation, IDO1 is expressed by a fraction of mature CD83 + LCs, suggesting a potential regulatory/inhibitory function of this subset [28,29]. Since LCs are a promising target for cancer immunotherapy [22,30] we aimed to investigate the expression of IDO1 in SLN LCs of patients with melanoma, correlating it with CD83 and with clinical features of patients. Finally, we investigated the possibility to modulate SLN LC CD83 and IDO1 expression.

Localization of LCs within SLN of Patients with Melanoma
Since DCs in SLNs of patients with melanoma have been shown to have significant morphological defects, we first investigated LCs morphology and localization within the different SLN areas in frozen sections. By the use of specific Abs for LCs, B and T cells, we observed that Langerin + LCs accumulate in CD3 + T cells-areas of SLNs (Figure 1), without differences between negative and positive SLNs. SLN LCs displayed the typical LCs morphology with thin and long dendrites, (Figure 1, inset). Three-color immunofluorescence labelling was performed on frozen SLN sections (n = 4 negative and n = 1 positive SLNs from n = 5 patients) with the indicated Abs and fluorescences (anti-Langerin, red; CD19, green; CD3, blue) in order to localize LCs within specific SLN areas. A representative negative SLN is shown. Original magnification ×200 and ×400 in the inset; scale bars = 20 µm and 40 µm, respectively. SLN LCs are found in the T cells rich areas, as indicated by asterisks; a typical SLN LC, with typical thin and long dendrites, is shown in the inset, as indicated by an arrow.

CD83 Expression in SLN LCs of Patients with Melanoma after In Vitro Stimulation with Inflammatory Cytokines
Subsequently, we investigated whether LCs from melanoma SLNs were still able to reach a mature immunophenotype with CD83 up-regulation by fluorescent immunocytochemistry and flow cytometry (n = 5 negative SLNs from n = 5 patients) ( Figure 4). Fresh SLN, 24 h-unstimulated and -stimulated SLN cells were labelled with anti-Langerin and CD83 Abs with cell membrane permeabilization, in order to detect all SLN LCs (see Material and Methods section for details). Among fresh SLN Langerin + LCs, characterized by both surface and cytoplasmic labelling, only few LCs expressed CD83 molecule, as revealed by fluorescent immunocytochemistry (Figure 4a): co-expression of both markers revealed the strong prevalence of Langerin on CD83 staining (Figure 4a, merge panel). No substantial differences were observed in SLN LCs cultured for 24 h without inflammatory cytokines as compared to t = 0 (data not shown). After 24 h-stimulation with inflammatory cytokines, we observed a strong CD83 up-regulation in SLN LCs (Figure 4b): many SLN Langerin + LCs expressed CD83 with similar intensity (Figure 4b, merge panel). Furthermore, some LCs displayed only cytoplasmic Langerin expression, as further indication of maturation (Figure 4b, merge panel). By flow cytometry, we found that CD83 expression significantly increased in SLN LCs after 24 h-stimulation with inflammatory cytokines (Figure 4f), as compared to both fresh SLN LCs (Figure 4d) and 24 h-culture without inflammatory cytokines (data not shown) (59.15%, range 47.11-77.25%; 19.29%, range 13.26-23.17%; and 19.94%, range 15.47-24.42%, respectively; p < 0.05), suggesting that SLN LCs are not irreversibly impaired by melanoma cells and immunosuppressive soluble factors. Inflammatory cytokines stimulation induced a strong IDO1 upregulation by the majority of all LCs in accordance with previous report [31].

Discussion
In this study, we investigated the expression of the tolerogenic enzyme IDO1 by LCs in SLN of patients with melanoma in order to understand their role in tolerance towards melanoma. We also correlated IDO1 expression with the DC maturation marker CD83, and evaluated SLN LCs functional properties. LCs were located in the T cells-areas of SLN and exhibited the typical dendritic appearance, as observed in healthy LN [9] and in contrast to previous report describing morphological alterations of DC in melanoma SLN [6].
A portion of SLN LCs expressed IDO1, in both negative (3.63%,) and positive (4.18%), SLNs with no significance differences. Approximately 50% of SLN LCs expressed CD83, according to previous reports [11,12]. Based on the expression of IDO1 and/or CD83 by LC, four subsets were identified, in both negative and positive SLN. IDO1 − CD83 + LCs (really mature), IDO1 − CD83 − LCs (real immature), a phenotype corresponding to that of immature skin LCs (Di Gennaro et al., 2014). A further LCs subset co-expressed IDO1 and CD83 (tolerogenic mature), already described in skin LCs and probably with a regulatory role in promoting T cell tolerance [28,29]. Finally, a fourth LCs subset was IDO1 + CD83 − (tolerogenic immature). Notably, this subset was significantly higher in positive SLNs as compared to negative ones and in SLN from melanoma with intermediate/high MR. These findings suggest a role for IDO1 + CD83 − LCs in favoring metastases arrival/homing in LNs and a relation with the proliferation activity of melanoma. The CD83 down-regulation and IDO1 expression might be induced by melanoma and represent a novel immuno-escape mechanism. However, CD83 down regulation on LCs is not definitive. After in vitro 24 h-stimulation with inflammatory cytokines, LCs are able to up-regulate CD83 expression. This finding is particularly important because CD83 down-regulation appears to be crucial for correct T cell activation and proliferation [12] and opens additional perspectives in the treatment of melanoma. Beside inflammatory cytokines, other immune adjuvants might be used to activate LCs and other DC subsets, at tumor site or in the SLN, such as Toll-Like Receptors (TLRs) agonists [30]. In line with this observation, peri-tumoral injection of GM-CSF prior SLN biopsy in melanoma patients reduces SLN immunosuppression [8].
In the last years, numerous reports of immuno-phenotypic and functional defects of SLN DCs in cancer patients have pointed out the concept of "tolerogenic DC", particularly in melanoma [6,7,11,12]. LCs in SLN of patients with melanoma shows an immature immunophenotype, with low CD83 expression [11]. Since CD83 molecule is important in the stimulation of T cell proliferation [32], it is not surprising that LCs resulted to be poor activator of T cells proliferation [12]. Thus, LCs may be ineffective in stimulating melanomaspecific T cells promoting tolerance instead of anti-melanoma immunity. Furthermore, in patients with melanoma, pDCs showed functional defects and accumulated in metastatic SLN [10]. It has been reported that pDCs and dDCs express the tolerogenic enzyme IDO1 in SLN of patients with melanoma (5,19,20), suggesting a crucial role of this enzyme in melanoma tolerance [13]. Here we showed that also LCs expressed IDO1, although IDO1 + LCs were not as numerous as SLN IDO1 + pDCs and no difference in frequency between negative and positive SLNs was detected. Importantly, a negative prognostic role of IDO1 + cells in SLN of patients with melanoma has been reported [13] according to previous hypotheses that IDO1 expression represents a melanoma immune escape mechanism [13,14]. Besides IDO1 induction in peri-tumoral DCs, melanoma cells themselves can express IDO1 [25][26][27]. which is induced in vitro by interferon-gamma [26]. Importantly, the level of IDO1 expression correlates with a poor prognosis. High levels of IDO1 expression have been detected in primary and metastatic LNs of patients with poor survival, while it was not detected or detected at low levels in patients with long survival [26]. Furthermore, IDO1 expression in primary cutaneous melanoma correlates with Breslow thickness, and IDO1 expression in antigen-presenting cells correlates negatively with progression-free survival of patients with melanoma [27]. It is likely that pro-inflammatory cytokines generated by the innate immune system in response to melanoma may indeed favor IDO1 expression in both melanoma cells and peri-tumoral DCs [1,25]. In our study, the observed small-size IDO1 + cells might correspond to dDCs [22] and pDCs [10], while the large-size, observed in positive SLNs, may represent melanoma cells.
Notably, IDO1 inhibitors have already entered clinical trials, combined first with chemotherapeutic drugs and then with immune check-point inhibitors, such as anti-CTLA-4 and/or PD-1, to turn melanoma tolerance into anti-melanoma immunity [24].
Recently, a novel role for IDO1 as an inflammatory modifier has been described, being the key factor of several pro-inflammatory pathways in cancer responsible for disease progression [25]. Furthermore, IDO1 is involved in tumor neo-vascularisation, which is instead inhibited by IFN-gamma [25]. Thus, IFN-gamma shows a conflicting role in tumor immunity, being the main IDO1 inducer but also involved in preventing new blood vessels formation during tumor growth [33]. Furthermore, tumor rejection mediated by CD8 + effector T cells is preceded by the inhibition of tumor-induced angiogenesis by IFN-gamma [33]. Therefore, IDO1 inhibitors might act also as immune adjuvants able to reprogram the correct inflammatory response and block tumor neo-vascularisation [24,25]. Human skin LCs express functional AhR, the well-known target of KYN, responsible for the main IDO1 suppressive actions [34]. Moreover, AhR stimulation by KYN leads to enhance IDO1 expression in LCs, without inducing CD83 expression [34]. Therefore, a possible scenario in melanoma, DC, and IDO1 interaction may be as follows: inflammatory cytokines produced by innate immunity in response to melanoma may modulate CD83 and IDO1 expression on LCs and generate CD83 + LCs, either IDO1 +/− which migrated to SLN (real mature SLN LCs and regulatory/inhibitory SLN LCs). IDO1, expressed by both LCs and melanoma cells, lead to the production of KYN, which directly stimulate LCs to express more IDO1 trough AhR without inducing CD83 expression. Thus, AhR stimulation in addition to immunosuppressive tumor cytokines might be responsible for CD83 downregulation and generation of CD83 − IDO1 + LCs (immature tolerogenic LCs) and/or CD83 − IDO1 − LCs (real immature LCs).
Importantly, SLN IDO1 + DC subsets (LCs, pDCs and dDCs) might be responsible for TRP depletion and KYN production, blocking effector T cells and enhancing T regulatory cells, therefore maintaining a tolerance milieu favoring SLN melanoma metastases [22].
The presence of four SLN LCs subsets reported here might contribute to improving the understanding of LCs biology in melanoma patients. Further studies might elucidate whether CD83 and IDO1 expression identify actually different SLN LCs subsets or different evolution profiles of the same subset, with different roles in immune responses. These results suggest a tolerogenic role of IDO1 + CD83 − SLN LCs in patients with melanoma, being significantly increased in SLN with metastases and in SLN in patients with intermediate/high MR of the primary lesions.
LCs, the first DCs interacting with cutaneous melanoma, are crucial regulators of immune response, representing a promising target for cancer immunotherapy. These findings offer new inputs to strategies targeting LCs directly at the site of primary melanoma or in SLN.

Human Samples
SLN from melanoma patients undergoing SLN biopsy were collected at the Plastic and Reconstructive Surgery Unit, Regional Melanoma Referral Centre and Melanoma & Skin Cancer Unit, Florence, Italy, after obtaining informed written consent. The study was conducted according to the 1964 Helsinki declaration, and its later amendments or comparable ethical standards, and Local Institutional Ethics Committee approval. After pre-operative lymphoscintigraphy and Patent-Bleu injections, SLN was detected and excised using computer-assisted gamma probe with adjustable collimation as previously reported [35]. For research purposes, SLN cell harvesting and processing were performed as previously described [36] with minor modifications [10]. Briefly, SLN was bisected crosswise by the pathologist and only one cutting surface was scraped 10 times with a scalpel blade. SLN cells were collected in a 15 mL tube by blade rinsing with PBS (Eu-roClone, Whetherby, UK) and used without further enzymatic treatments, to avoid DC activation and maturation prior to analysis [32], either for fluorescent immunocytochemistry, deposed on slide by centrifugation (cytospins), in vitro stimulation with inflammatory cytokines, or flow cytometry analyses. Due to this peculiar collecting method, SLN cells were few, thus we performed experiments on subgroups and not on all specimens. SLNs were processed for routine histology with hematoxylin/eosin and immunohistochemistry. For intra-surgery histological analyses, only for n = 5 patients, half SLN was embedded in OCT-like medium (Killik; Bio-Optika, Milan, Italy) and snap-frozen. Cryostat sections (10 µm) and cytospsins were fixed for 10 min in cold acetone (Sigma, Milan, Italy) at 4 • C, air dried at room temperature and stored at −20 • C until labelling.

In Vitro Stimulation with Inflammatory Cytokines
SLN cells were re-suspended in RPMI 1640 medium (EuroClone) supplemented with 1% penicillin/streptomycin (EuroClone) and 2% glutamine (Sigma), without serum, indicate as culture medium. For in vitro stimulation analyses, cells from each SLN were divided into three aliquots: the first was processed immediately, indicated as t = 0; the second and the third were cultured for 24 h in culture medium at 37 • C and 5% CO2 with or without the following inflammatory cytokines: TNF-alpha (10 ng/mL), IL-1beta (10 ng/mL) and IL-6 (1000 U/mL), all from R&D Systems (Minneapolis, MN). Cells were used for fluorescent immunocytochemistry and flow cytometry.

Fluorescent Immunohistochemistry and Immunocytochemistry
Multi-color fluorescent labelling procedures were performed on SLN sections and cell cytospins, at room temperature if not indicated. For LC localization within different SLN areas, three-color fluorescent immunohistochemistry was performed on SLN sections, pre-treated with 20 mg/mL BSA (Sigma-Aldrich, St. Luis, MO, USA) for 1 h, and stained first with anti-Langerin for 2 h andrevealed with AF594-conjugated Abs (red fluorescence); then with anti-CD3 V450-and CD19 FITC-conjugated Abs (blue and green fluorescence, respectively) overnight at 4 • C; next, fluorescein signal was amplified. To study IDO1 and CD83 expression in SLN LCs, two-color fluorescent immunocytochemistry was performed on cytospins, pre-treated with PBS/Triton X-100 0.3% and 20 mg/mL BSA (all from Sigma-Aldrich) for 1 h for cell membrane permeabilization. For Langerin/CD83 labelling, specimens were first stained with anti-Langerin for 2 h, revealed with AF594-conjugated Abs (red fluorescence); then with anti-CD83 FITC-conjugated Abs (green fluorescence), overnight at 4 • C; next fluorescein signal was amplified. For Langerin/IDO1 labelling, specimens were first stained with polyclonal anti-IDO1 overnight at 4 • C, revealed with anti-sheep FITC-conjugated Abs and fluorescein signal was amplified (green fluorescence); then with anti-Langerin for 2 h, revealed with AF594-conjugated Abs (red fluorescence). This labelling order was set up to avoid sheep-goat immune cross-reactions. Nuclei were finally labelled with Hoechst 33,342 (20 µg/mL; Sigma; blue fluorescence). All slides were then mounted with Prolong antiFade (Life Technologies, Thermo Fisher Scientific Inc, Massachusetts, MA, USA.) and observed with Leica DMLB microscope (equipped for epifluorescence; Leica Microsystems GmbH, Wetzlar, Germany). Images were acquired at 0.529 and 0.265 µm per pixel resolution, corresponding to ×200 and ×400 original magnification respectively, using Leica DFC200310 FX microscope digital color camera and LAS software with overlay module (Leica Microsystems GmbH). Adobe Photoshop CS2 software (Adobe Systems Incorporated, San Jose, CA, USA) was used for image processing and figure creating.

Flow Cytometry Analyses
All Abs dilutions and washing steps were done in PBS supplemented with 1% FCS at room temperature if not indicated. For IDO1 and CD83 expression analyses, two-or threecolor cell staining was performed on fresh cells and/or after 24 h-culture with or without inflammatory cytokines cells, as previously reported (Gerlini et al., 2012;Di Gennaro et al., 2014). Briefly, for Langerin/CD83 labelling, cells were first labelled with anti-Langerin-PE and CD83-APC for 15 min; then fixed and permeabilized with Cytofix/Cytoperm solution (PharMingen, San Diego, CA, USA) following manufacturer's instructions, and again labelled with anti-Langerin-PE for 15 min, in order to detect all SLN LCs, because during maturation, surface Langerin is relocated to the cytoplasm (Gerlini et al., 2012). For Langerin/CD83/IDO1 labelling, cells were labelled first with anti-Langerin-PE and CD83-APC as described above; then fixed and permeabilized with Flow Cytometry Fixation Buffer and Permeabilization/Wash Buffer I, all from R&D Systems, following manufacturer's instructions, and labelled with anti-IDO1 AF488-conjugated for 30 min and Langerin-PE, to reveal cytoplasmic Langerin as reported above (Di Gennaro et al., 2014). Cells were then acquired using FACS Canto and FACSDiva software, version 6.0 (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA). Results were expressed as percentage of total events or CD83 + or CD83/IDO1 double positive cells among total Langerin + cells.

Statistical Analysis
Values reported throughout the text are expressed as mean and values range. Data were analyzed using two-sided student's t-test with p < 0.05 adopted as significance level (Origin and Microsoft Excel software).