IL-10-Engineered Dendritic Cells Modulate Allogeneic CD8+ T Cell Responses

Tolerogenic dendritic cells (tolDC) play a central role in regulating immune homeostasis and in promoting peripheral tolerance. These features render tolDC a promising tool for cell-based approaches aimed at inducing tolerance in T-cell mediated diseases and in allogeneic transplantation. We developed a protocol to generate genetically engineered human tolDC overexpressing IL-10 (DCIL-10) by means of a bidirectional lentiviral vector (LV) encoding for IL-10. DCIL-10 promote allo-specific T regulatory type 1 (Tr1) cells, modulate allogeneic CD4+ T cell responses in vitro and in vivo, and are stable in a pro-inflammatory milieu. In the present study, we investigated the ability of DCIL-10 to modulate cytotoxic CD8+ T cell responses. We demonstrate that DCIL-10 reduces allogeneic CD8+ T cell proliferation and activation in primary mixed lymphocyte reactions (MLR). Moreover, long-term stimulation with DCIL-10 induces allo-specific anergic CD8+ T cells without signs of exhaustion. DCIL-10-primed CD8+ T cells display limited cytotoxic activity. These findings indicate that stable over-expression of IL-10 in human DC leads to a population of cells able to modulate cytotoxic allogeneic CD8+ T cell responses, overall indicating that DCIL-10 represent a promising cellular product for clinical applications aimed at inducing tolerance after transplantation.


Introduction
Dendritic cells (DC) are antigen-presenting cells (APC) able to induce naïve T cell activation and differentiation. In homeostatic conditions, DC are also involved in the induction and maintenance of immune tolerance. Indeed, DC at an immature stage or specialized subsets of DC, named tolerogenic DC (tolDC), control T-cell responses via several immunosuppressive mechanisms (reviewed in [1]) and induce the expansion and/or induction of regulatory T cells [2][3][4][5]. Exploiting the immune-regulatory capacities of DC holds great promise for the treatment of autoimmune diseases [6] and the prevention of graft rejection [7]. Several protocols have been established to efficiently differentiate tolDC in vitro starting from human peripheral blood monocytes (reviewed in [8]). A comparative analysis of different populations of in vitro differentiated tolDC suggested that IL-10-modulated DC are the most adequate cells for tolerance-inducing therapies [9]. Two prominent protocols to generate tolDC in the presence of IL-10 have been established. The first involves the exposure of immature DC to IL-10 in the last two days of DC differentiation [10], while the second entails the addition of IL-10 from the beginning of DC differentiation (DC-10) [11]. The latter cells are characterized by a unique cytokine production profile, the ability to spontaneously secrete IL-10 in the absence of IL-12, and by the co-expression of the tolerogenic molecules HLA-G and ILT4 [12]. These are critical features that render DC-10 potent inducers of T-cell anergy and allogeneic (allo)-specific T regulatory type 1 (Tr1) cells [11]. More recently, we developed an efficient protocol to generate IL-10-producing DC by bidirectional lentiviral vector (LV)-mediated IL-10 transduction of monocytes during DC differentiation (DC IL-10 ) [13]. DC IL-10 recapitulate the tolerogenic features of DC-10 as they secrete supraphysiological levels of IL-10, are stable upon exposure to pro-inflammatory signals, modulate allogeneic CD4 + T cells, and induce allo-specific Tr1 cell differentiation in vitro. We previously reported that DC IL-10 inhibit the proliferation of CD8 + T cells in primary cell culture [13], but detailed characterizations of DC IL-10 -conditioned CD8 + T cells is still elusive.
CD8 + T cells, also known as cytotoxic T lymphocytes (CTL), are part of the adaptive immune system and play a crucial role in the clearance of intracellular pathogens and tumors. During the primary response, naive CD8 + T cells recognize antigen(Ag)s in the context of MHC class I molecules and undergo a differentiation process leading to strong clonal expansion to generate large numbers of cytotoxic effector CD8 + T cells. These cells migrate to the site of infection and mount an Ag-specific response (reviewed in [14][15][16]) to directly kill target cells through the release of cytotoxic molecules (e.g., perforin and granzymes) or via Fas-FasL interaction (reviewed in [17]). Moreover, cytotoxic effector CD8 + T cells secrete interferon (IFN)γ and tumor necrosis factor (TNF)-α [18,19], two important mediators for coordinating the innate and adaptive immune response. The CD8 + T cell activation process requires three signals provided by APC, primarily by DC: the engagement of TCR by peptide-MHC-I complexes, which defines the Ag specificity [20]; co-stimulatory signals provided by CD80/CD86 molecules [21,22]; and polarizing signal mediated by DC-derived cytokines, mainly IL-12 [23,24]. Additional stimulatory signals can be provided by CD4 + T helper cells, reinforcing DC-mediated activation and inducing the establishment of long-lasting memory CD8 + T cells [25,26].
In the present study, we investigated the ability of DC IL-10 to modulate allogeneic CD8 + T cell responses using short-term and long-term mixed lymphocyte reactions (MLR). Our findings revealed that DC IL-10 regulate allogeneic CD8 + T cell activation and proliferation in short-term culture, while in long-term co-culture, they induce anergic allo-specific CD8 + T cells, which display limited allo-specific cytotoxicity. Overall, DC IL-10 represent an interesting and effective tool for modulating not only CD4 + , but also CD8 + allo-reactive T cell responses.

DC IL-10 Modulate Allogeneic CD8 + T Cell Priming
We investigated the ability of IL-10-engineered DC to modulate allogeneic CD8 + T cell responses in vitro. We generated immature IL-10-transduced DC (DC IL-10 ) or LPSactivated DC IL-10 (mDC IL-10 ) and compared them with LPS-activated GFP-transduced DC (mDC GFP ). DC IL-10 and mDC IL-10 were efficiently transduced, based on the expression of the marker gene ∆NGFR (91.5 ± 1.1% and 93.5 ± 1.3%, of CD11c + ∆NGFR + cells, n= 22 respectively), and expressed CD14, CD16, CD141, CD163, HLA-G, and ILT4 (Supplementary Figure S1a). Since efficient stimulation of CD8 + T cells requires the expression of MHC class I and costimulatory molecules [27], we evaluated their expression on DC IL-10 , mDC IL-10 , and mDC GFP , which showed that HLA-Class I was expressed at high and comparable levels ( Figure 1a). CD86 was expressed by DC IL-10 and mDC IL-10 (46.6 ± 7.7% and 74.9 ± 4.4%, respectively, n = 22; Figure 1a), but, as previously observed [13], at variable levels among the donors tested. Despite this variability, DC IL-10 and mDC IL-10 expressed CD86 at significantly lower (p = 0.0034 and p < 0.0001, respectively) levels compared with mDC GFP (89.2 ± 2.5%, n = 22; Figure 1a). Figure 1. DC IL-10 modulate allogeneic CD8 + T cell responses. CD14 + cells isolated from peripheral blood of healthy subjects were transduced during DC differentiation with LV-IL-10 and left unstimulated (DC IL-10 ) or activated with LPS (mDC IL-10 ). As a control, DC transduced with LV-GFP and activated with LPS (mDC GFP ) were differentiated from the same donors. (a) At the end of the differentiation, the expression of HLA-class I molecules (HLA-ABC, n = 3-14) and CD86 (n = 17) was evaluated by flow cytometry. Each dot represents a single donor; bars indicate mean ± SEM. (b-d) DC IL-10 inhibited the activation and proliferation of allogenic CD8 + T cells in short-term primary MLR. Allogeneic CD8 + T cells isolated from the peripheral blood of healthy subjects were stained with proliferation dye and stimulated with the indicated DC at a 10:1 ratio for 5 days. (b) After culture, CD8 + T cells were collected, and the expression of the activation markers CD25 (n = 20), CD71 (n = 20), and CD137 (n = 11) was evaluated by flow cytometry. Each dot represents a single donor; bars indicate mean ± SEM. (c) On day 5, the percentage of proliferated CD8 + cells was evaluated by proliferation dye dilution. Dot plots from one representative donor are presented (left panel); numbers indicate the percentage of proliferated cells. Percentage of proliferated CD8 + T cells stimulated with DC IL-10 (light blue dots), mDC IL-10 (blue dots), and mDC GFP (orange dots) are shown (right panel). Each dot represents a single donor (n = 9). (d) IFNγ and GM-CSF were evaluated by ELISA in cell culture supernatants (n = 9). Each dot represents a single donor; bars indicate mean ± SEM. For all the statistical analyses, the Friedman matched-paired test, in association with Dunn's multiple comparison test, was applied. Statistically significant p values are reported.
Overall, these findings indicate the both DC IL-10 at steady state and upon LSP activation modulate allogeneic CD8 + T cell responses.
To further characterize the modulatory activity of IL-10-engineered DC, we limited the comparison between mDC IL-10 and mDC GFP to cells expressing similarly high levels of CD86 (Figure 1a). We first assessed the ability of CD8 + T cells primed with mDC IL-10 (T(mDC IL-10 ) cells) to release cytotoxic granules upon restimulation with mature DC (mDC) generated from the same donor used for priming (Supplementary Figure S2a). We observed an overall lower frequency of cells expressing granzyme (Gz)B and perforin (Prf) in T(mDC IL-10 ) cells (23.3 ± 1.7%, and 6.3 ± 2.5%, respectively, n = 5) compared with T cells stimulated with mDC GFP (T(mDC GFP ) cells) (54.4 ± 6.9% and 24.8 ± 12.6%, respectively, n = 5; Supplementary Figure S2b). Despite differences in the overall granule content, upon restimulation with mDC, the percentage of CD8 + T cells actively degranulating, as indicated by the percentage of GzB + CD107a + or Prf + CD107a + cells [28], was similar to that observed in unstimulated T cells for both T(mDC IL-10 ) and T(mDC GFP ) cells (Supplementary Figure S2c,d). These findings suggest that short-term stimulation with allogeneic mDC IL-10 or mDC GFP is not sufficient to induce fully competent cytotoxic CD8 + T cells.
Functional characterization of T cells generated with mDC IL-10 or mDC GFP was then performed by restimulating cells with the same alloantigen used in priming (Supplementary Figure S3a). As expected, stimulation of allogeneic CD8 + T cells with mDC GFP resulted in the induction of allo-specific IFNγ-producing cells (5.7 ± 2.30% vs. 0.94 ± 0.16%, in T(mDC GFP ) stimulated with mDC vs. T(mDC GFP ) cultured alone, n = 13, p = 0.0007; Figure 2c). Interestingly, although at a lower extent compared with T(mDC GFP ) cells, restimulation of T(mDC IL-10 ) cells with mDC resulted in a significantly higher (p = 0.0344) frequency of IFNγ + cells compared with unstimulated T(mDC IL-10 ) cells (2.2 ± 0.64% vs. 1.4 ± 0.47%, in T(mDC IL-10 ) stimulated with mDC vs. T(mDC IL-10 ) cultured alone, n = 13; Figure 2c), indicating that allogeneic CD8 + T cell priming by mDC IL-10 occurred, but that upon restimulation, T(mDC IL-10 ) cells contained a lower frequency of allo-specific IFNγ + cells compared with T(mDC GFP ) cells (Figure 2c). No major differences were observed at steady state and upon mDC stimulation in the percentages of allo-specific T(mDC IL-10 ) and T(mDC GFP ) cells producing IL-10 ( Figure 2c).
Overall, these findings indicate that mDC IL-10 induce allo-specific anergic CD8 + T cells in vitro.

Discussion
In the present study, we showed that DC IL-10 limit the activation and proliferation of allogeneic CD8 + T cells. Moreover, allogeneic CD8 + T cells primed with DC IL-10 are anergic, secrete low levels of the pro-inflammatory cytokines IFNγ and GM-CSF, and have limited cytotoxic activity upon secondary stimulation, while maintaining responsiveness to unrelated allo-antigens. The anergic phenotype of DC IL-10 -primed CD8 + T cells is not associated with the expression of inhibitory molecules. Overall, these findings indicate that DC IL-10 are effective in modulating allo-specific CD8 + T cell responses in vitro.
In addition to TCR engagement (signal I), optimal CD8 + T cell activation requires costimulatory signaling (signal 2) [29] and pro-inflammatory cytokines (signal 3) [27]. DC IL-10 , at steady state and upon LPS activation, efficiently supplied signal 1 to CD8 + T cells since they expressed HLA class I molecules at high and comparable levels to control DC. Moreover, DC IL-10 expressed CD86, albeit at variable levels, while its expression on LPS activated (m)DC IL-10 was sustained and comparable with control DC. Despite the difference in CD86 expression, both DC IL-10 and mDC IL-10 elicited activation and proliferation of allogeneic CD8 + T cells at comparable levels, indicating that in our culture condition, DC IL-10 provide an efficient signal 2, although to a lower extent compared with control DC. The overall inhibition of allogeneic CD8 + T cell responses induced by DC IL-10 is therefore mediated by the cytokine context during priming, which, in our culture condition, is enriched in IL-10.
The effects of IL-10 on CD8 + T cell responses differ according to the cytokine concentration: IL-10 at low concentration modulates CD8 + T cells in terms of proliferation and cytotoxic activity, while at high concentration, it stimulates CD8 + T cell responses and promotes CD8 + T cell expansion in vitro and in vivo [30][31][32][33][34][35]. It has been previously demonstrated that IL-10 has no direct inhibitory effects on the proliferation of CD8 + T cells activated by anti-CD3 mAb, but it inhibits alloantigen-specific proliferative responses and induces long-lasting anergy in CD8 + T cells indirectly in the presence of antigen-presenting cells in vitro [32]. In line with these findings, IL-10 produced by DC IL-10 was effective in reducing activation, proliferation, and cytotoxic activity in allogeneic CD8 + T cells. We recently reported that the secretion of IL-10 by engineered DC during CD8 + T cell priming is required for modulating T-cell activation and proliferation, while addition of exogenous IL-10 during DC-mediated priming is not effective (Passeri L. et al. [36], under revision). Thus, our findings indicate that inhibition of allogeneic CD8 + T cell responses by DC IL-10 required allo-Ag presentation in the presence of IL-10. In addition to secreting high levels of IL-10, DC IL-10 express the immunomodulatory molecules HLA-G and ILT4, which define their tolerogenic potency [13]. CD8 + T cells express ILT2, a known ligand of HLA-G [37]; thus, it can be speculated that the HLA-G/ILT2 interaction might be also implicated in DC IL-10 -mediated inhibition of CD8 + T cell responses.
The role of the PD1/PDL-1 interaction in modulating T-cell activation and function has been widely characterized, especially in the context of cancer cells [38]. Recently, Li and colleagues [39] demonstrated that the activation of PD-1 signaling during the early phase of Ag recognition disrupts the interaction between the TCR, the MHC/peptide complex and CD8, resulting in the suppression of CD8 + T cell function. DC IL-10 expresses high levels of PDL-1 (Passeri L. et al. [36], under revision); thus, we cannot exclude that the suppression of allogeneic CD8 + T cells observed in our culture conditions might be the results of synergic effects of high levels of IL-10 and a PDL-1/PD-1 interaction during Ag-recognition by CD8 + T cells.
Exogenous or tolerogenic DC-derived IL-10 limit CD4 + T cell responses by inducing T cell hypo-responsiveness and anergy [5,11,40,41]. Moreover, stimulation of allo-specific CD8 + T cells with IL-10-treated DC resulted in hypo-responsiveness and induction of anergic allo-specific cells [41]. We showed that long-term stimulation of allogeneic CD8 + T cells with mDC IL-10 induced allo-specific T cell anergy, which can be reverted by addition of exogenous IL-2, thus confirming the ability of IL-10-secreting DC to promote CD8 + T cell anergy in vitro. In line with previous studies [41], mDC IL-10 -primed anergic CD8 + T cells failed to degranulate upon encountering allo-specific target cells. Gradual functional impairment of cytotoxic CD8 + T cells have been associated with T-cell exhaustion, which is characterized by a loss of pro-inflammatory cytokine production, decreased proliferative capacity, cytotoxic potential [42], and sustained high expression of multiple inhibitory receptors such as PD-1, TIM-3, LAG-3, and TIGIT [43]. Although DC IL-10 -modulated CD8 + T cells secreted lower levels of IFNγ and GM-CSF and proliferated significantly less compared with mDC-primed CD8 + T cells, they did not show up-regulated levels of T-cell exhaustion markers. Overall, our findings support the role of DC IL-10 in actively inducing tolerogenic phenotype and functions in CD8 + T cells.

Vector Production and Titration
VSV-G-pseudotyped third generation bidirectional Lentiviral Vectors (bdLV) encoding human IL-10 and ∆NGFR or green fluorescent protein (GFP) and ∆NGFR were produced by calcium phosphate transfection into 293T cells and concentrated by ultracentrifugation as described previously [44]. Titer was estimated by limiting dilution: vector particles were measured by HIV-1 Gag p24 Ag immune capture (NEN Life Science Products, MA, USA), and vector infectivity was calculated as the ratio between titer and total particles. Titers ranged between 5 × 10 8 and 6 × 10 9 transducing units/mL, while infectivity ranged between 5 × 10 4 and 10 5 transducing units/ng. To produce concentrated Vpx-incorporating viral-like particles (VLPs), 293T cells were co-transfected with a VSV-G expressing plasmid and the Simian Immunodeficiency Virus-derived packaging plasmid SIV3+, as previously described [45].

Dendritic Cell Differentiation
Human peripheral blood was obtained from healthy donors in accordance with local committee approval (TIGET09) and the Declaration of Helsinki. Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation over Lympholyte ® -H (Cederlane, Burlington, Canada). CD14 + cells were isolated from PBMC by positive selection using CD14 MicroBeads (Miltenyi Biotech, Bergish Gladbach, Germany) according to the manufacturer's instructions. Monocytes were exposed for 3 h to Vpx-VLP and then were transduced with bdLV-IL-10 (DC IL-10 ) or with bdLV-GFP (DC GFP ) at a Multiplicity of Infection (MOI) of 10 as previously described [13]. Cells were cultured in RPMI 1640 (Lonza, Verviers, Belgium) with 10% fetal bovine serum (FBS) (Euroclone, Pero, Italy), 100 U/mL penicillin/streptomycin (Euroclone, Pero, Italy) and 2 mM L-glutamine (Euroclone, Pero, Italy) at 10 6 cells/mL in a 1 mL volume in a 24-well culture plate. The media was supplemented with rhGM-CSF (Miltenyi Biotech, Bergish Gladbach, Germany) at 100 ng/mL and with rhIL-4 (Miltenyi Biotech, Bergish Gladbach, Germany) at 10 ng/mL and cultured for 7 days at 37 • C with 5% CO 2 . Matured DC IL-10 (mDC IL-10 ) and mDC GFP were obtained by activation at day 6 with 1 µg/mL of LPS (Sigma-Aldrich, St. Louis, MO, USA). As expected, we reached an average of transduction, evaluated by ∆NGFR expression, above 80% for all the DC IL-10 (matured or not) and mDC GFP . From some donors, un-transduced LPS activated DC (mDC) were differentiated. All DC, transduced or not, were harvested on day 7 for phenotypical and functional analyses.

CD8 + T Cell Isolation and Culture
CD8 + T cells were purified from PBMC by negative selection using a human CD8 T cell isolation kit (Miltenyi Biotech, Bergish Gladbach, Germany) according to the manufacturer's instructions. T cell cultures were performed in X-VIVO 15 (Lonza, Verviers, Belgium) supplemented with 5% human serum (Sigma Aldrich, Burlington, Massachusetts, USA) and 100 U/mL penicillin/streptomycin (Lonza, Verviers, Belgium). T cells were cultured with 10 4 allogeneic DC IL-10 , mDC IL-10 , or mDC GFP at a 10:1 ratio. After 5 days of stimulation (short-term culture), T cells were collected, washed, and phenotypically and functionally analyzed. For long-term culture experiments, T cells were cultured with allogeneic mDC IL-10 (T(mDC IL-10 )) or mDC GFP (T(mDC GFP )) at a 10:1 ratio. At day 3, 1 ng/mL of recombinant human IL-15 (R&D System, Minneapolis, MN, USA) was added. At day 14, cells were collected, washed, and analyzed. In some experiments, CD8 + T cells were labelled with Cell Proliferation Dye eFluor ® 670 (eBioscience, San Diego, CA, USA) according to the manufacturer's instructions and analyzed by flow cytometry for their proliferation at the end of the T:DC co-culture.

CD8 + T Cell Intracytoplasmic Staining
CD8 + T cells were plated for 6 h alone or in the presence of LPS-matured un-transduced DC (mDC) from the same donor used for priming at a 10:1 ratio. A CD107a fluorophoreconjugated antibody was added immediately after cell seeding, and 10 µg/mL of Brefeldin

CD8 + Secondary Stimulation (II MLR)
For the recall response, T cells primed with mDC IL-10 (T(mDC IL-10 )) or with mDC GFP (T(mDC GFP )) were stained with Cell Proliferation Dye eFluor ® 450 (eBioscience, San Diego, CA, USA) and plated for an additional 4 days in the presence of mDC from the same donor used for priming at a T:DC ratio of 10:1 with or without 100 U/mL IL-2 (Proleukin, Novartis, Basel, Switzerland). For some donors, T(mDC IL-10 ) and T(mDC GFP ) cells were restimulated with mDC differentiated from a third-party donor for 4 days. At the end of the culture, the proliferation of CD8 + T cells was analyzed by flow cytometry.

Cytokine Determination
For DC, at the end of the 7 days of differentiation, cells were collected, washed, and plated in RPMI complete medium at a concentration of 1 M/mL. Supernatants were collected after 48 h and levels of IL-10 were evaluated. For CD8 + T cells, IFNγ and GM-CSF production was quantified in co-culture supernatants. Cytokine quantification was performed by standard sandwich ELISA with purified and biotinylated antibody couples (Becton Dickinson, CA, USA). The limits of detection were 31 pg/mL for IFNγ and GM-CSF, and 15 pg/mL for IL-10. Samples were acquired using a CytoFlex LX flow cytometer (Becton Dickinson, Mountain View, CA, USA), and data were analyzed with FCS express 7 (De Novo Software, Glendale, CA, USA). Quadrant markers were set according to the relative fluorescence minus one (FMO) staining.

Statistical Analysis
Friedman matched-pairs test, in association with Dunn's multiple comparison test, was applied in Figure 1. Wilcoxon matched-pairs test (two-tailed) was used for statistical analysis in Figure 2 and in Supplementary Figures S1-S3. Statistically significant p values are reported in all the figures. All results are presented as mean values ± SEM, unless differently specified in the figure legend. Results were analyzed using GraphPad Prism 9.0 (GraphPad Software, San Diego, CA, USA).

Conclusions
In the context of allo-transplantation, recipient T cells can recognize donor allo-Ags and give rise to an inflammatory immune response leading to graft rejection [46]; several studies have highlighted the major role played by cytotoxic CD8 + T cells in allogeneic responses (reviewed in [47]). In the present study, we reported that DC IL-10 effectively inhibit CD8 + T cell effector function and actively promote allogeneic anergic CD8 + T cells in vitro. These findings, in addition to their previously reported ability to modulate allospecific CD4 + T cell responses and to promote Tr1 cells [13], support the conclusion that IL-10-engineered DC represent a promising tool for developing cell-based therapies to promote/restore tolerance in transplantation settings. Further studies in pre-clinical models of allo-transplantation are warranted to better define the feasibility and efficacy of DC IL-10 cell therapy.