Antigen presentation by MART-1 adenovirus-transduced interleukin-10-polarized human monocyte-derived dendritic cells

Introduction

Dendritic cells (DC) are potent and versatile antigen-presenting cells (APC).^1,2 They can present an antigen to generate an immune response as well as to induce tolerance.^3,4 As such, ex vivo cultured DC have also become a useful tool in translational cancer immunotherapy studies, especially for therapeutic immunization with tumour-associated peptide epitopes.^5,6 However, while factors like lipopolysaccharide (LPS), tumour necrosis factor-α (TNF-α), interferon-γ (IFN-γ), CD40 engagement, virus particles, natural adjuvants, etc., induce maturation of DC⁷ other microenvironmental factors such as prostaglandin E₂, transforming growth factor-β (TGF-β), interleukin-10 (IL-10), certain products of nematodes, hyphae, etc., can polarize DC to a non-immunogenic and to an inhibitory or tolerogenic phenotype.⁸

Of the various environmental factors capable of negatively influencing DC ‘maturation’, IL-10 stands out as one of the most potent immunosuppressive cytokines with profound negative effects on dendritic APC.⁹ IL-10 down-regulates the expression of costimulatory molecules, major histocompatibility complex (MHC) class I and class II molecules, and proinflammatory cytokines in DC and thus make them non-immunogenic.^10–12 We have previously shown that two types of DC can be generated in in vitro culture from blood monocytes.¹³ One type is stimulatory in a mixed lymphocyte reaction (MLR) and they synthesize IL-12 on maturation, while the other type secretes large amounts of IL-10 and functions as non-immunogenic or inhibitory accessory cells.¹³ Others have also reported regulatory effects of autocrine and paracrine IL-10 on dendritic cell functions.^14–16 Similarly, resident pulmonary DC synthesize IL-10 and these resident DC have been shown to be ‘tolerogenic’ at their basal state in some models.¹⁷ When taken collectively, while these observations confirm a considerable degree of plasticity within the DC population,¹⁸ the data also suggest that myeloid DC grown in the presence of IL-10 differentiate into an inhibitory or tolerogenic phenotype. It is not, however, clear if IL-10-differentiated DC represent a fixed phenotype. We, therefore addressed how IL-10-polarized DC (referred as inhibitoryDC [inhDC] in this paper), when made to present an epitope internally (i.e. from internally synthesized antigen through an adenoviral vector) behave in a cytotoxic T lymphocyte (CTL) activation assay using the melanoma antigen recognized by T cells (MART-1) antigen as a prototype, in vitro. We show here that the MART-1₂₇₋₃₅ epitope (AAGIGLTV)-loaded inhDC are poor stimulators of the MART-1₂₇₋₃₅ epitope-specific CD8⁺ precursors. However, inhDC, when transduced with the MART-1 adenoviral vector to synthesize the MART-1 protein, process the MART-1 antigen along their MHC class I pathways and efficiently activate the epitope-specific CD8⁺ T cells.

Materials and methods

Results

Discussion

Although distinct classes of DC may exist to generate an immune response or to induce tolerance, it is now quite clear that diverse microenvironmental factors can ‘polarize’ DC precursors to an immunogenic, to a non-immunogenic, or to a tolerogenic phenotype. Of the various microenvironmental factors, IL-10 occupies an interesting place in this schema as monocytes (a class of myeloid DC precursors) and DC themselves can produce this regulatory cytokine. Several investigators have documented a stimulatory role of IL-10.^25–29 However, IL-10 has been widely described as one of the most potent suppressive cytokines. IL-10 exerts suppressive effects on the immune system in multiple ways. For example, IL-10 can be directly suppressive to T cells.³⁰ However, IL-10 exerts its major suppressive effects on APC by down-regulating the expressions of costimulatory molecules, MHC class I and class II molecules, and pro-inflammatory cytokine synthesis.⁹ Thus, IL-10 can exert profound inhibitory effects on the DC-maturation process as well as on the antigen presenting machinery – two critical elements in antigen presentation and T-cell activation. Indeed, the breadth and the magnitude of the inhibitory effect of IL-10 on gene expression by DC seen in the microarray analysis is quite remarkable. Our expression analyses was limited to a small set of genes (Table 1) in an attempt to obtain some insight into the differential expression of immunologically relevant genes in the context to ‘microenvironmental’ regulation of human DC. Although our microarray analysis caught only a single snapshot of gene expression and given that microarray data analysis requires caution, when taken with other published reports of differential gene expression, our data^31–33 clearly show that the inhibitory effect of IL-10 involves virtually all of the important components of DC-function that are critical in the process of generating an effective immune response (i.e. on antigen processing, costimulation, inflammation, and chemotaxis; Table 1). Thus, the immunosuppressive role of IL-10 is quite impressive.

The importance of endogenous IL-10 in physiology is clearly evident in IL-10^−/− mice.³⁴ These mice develop inflammatory bowel disease spontaneously,³⁵ they are prone to developing autoimmune diseases resulting from an experimental IL-10/IL-12 imbalance³⁶ and they succumb to a lethal form of exaggerated immune response to infection.³⁷ Understandably, IL-10 can be therapeutic in inflammatory bowel disease in experimental animal models³⁸ and it is also being tested in several human autoimmune disorders³⁹ where dendritic cells treated with IL-10 have been shown to modulate immune response of naïve and sensitized T cells in vivo. As such, it is not surprising that in a direct peptide presentation system, DC that have been grown in the presence of IL-10 are not efficient stimulators of the cognate precursors (Table 2). This is in agreement with other reports of inefficient or even abortive antigen presentation by IL-10-polarized APC described by others using different types of APC.^12,40,41

Of note, while the IL-10-grown DC that were loaded with exogenous peptide were unable to activate the cognate precursors (Table 2), they were capable of stimulating the preactivated effectors in secondary stimulation (Fig. 4). While this observation essentially confirms the well known difference in the threshold for activation between precursors and fully activated effector cells, our data differ somewhat from those of Steinbrink et al.⁴² who have shown that IL-10-grown DC make cognate T cells ‘anergic’. Presently, the reason for the difference remains unclear. This may be due to the different experimental protocols. Steinbrink et al.⁴² first cocultured the CTL with IL-10-grown DC and then tested their responsiveness against antigen pulsed DC. We have carried out secondary stimulation directly and have tested the responsiveness of the epitope specific CTL in three different assays and the activated CTL responded to the secondary stimulation by IL-10-polarized DC in all three assays (tetramer assay, IFN-γ response, and cytotoxicity).

Finally and importantly, a novel and noteworthy observation that has emerged from our study is the remarkable stimulatory capacity of the AdVMart1-transduced IL-10-grown DCs (Figs 1b and 3; Table 2). The mechanistic basis behind the effectiveness of the AdVMart1-transduced IL-10-grown DC in activating the CTL precursors remains to be fully clarified. It will be tempting to suggest that this might be caused by a maturing effect of the viral vector on the inhDC. Indeed, Lyakh et al. have shown ‘differentiation’ of DC with adenoviral infection⁴³ and we have also observed varying degrees of maturation of DC following adenoviral transduction [see ⁴⁴ and Fig. 1]. However, the effect of AdVMart1 transduction on the IL-10-polarized DCs was marginal, if any (Fig. 1a). The efficiency of the AdVMart1-transduced IL-10-grown DC was therefore unlikely to have resulted from a DC-maturing effect of the virus. Instead, we believe that the source of the antigen (i.e. internally synthesized protein) may have contributed to a more effective presentation by the AdVMart1-transduced inhDC. In this context, it should be mentioned that multiple factors (T-cell receptor (TCR) avidity, degree of costimulation, epitope density on the APC, levels of cytokines available during priming, etc.) determine the nature as well as the magnitude of a primary CD8⁺ T-cell response.^45,46 Thus, costimulation and cytokine(s) provided – or not provided – by the DCs (i.e. as a function of DC maturity) are important. The epitope density on and the duration of antigenic signal provided by the DCs (i.e. as determinants of the magnitude of TCR-driven signalling) are also equally important parameters in the process of antigen presentation.⁴⁷ Our data clearly show that the IL-10-grown DC traffic the MART-1 protein along their MHC class I pathway just as efficiently as the iDCs (Fig. 5a). Admittedly, this does not provides a compelling evidence that the MART-1 protein was processed with comparable efficiency by both cell types. Fluorescence microscopy is not well suited to quantify the degree of antigen processing. A reagent capable of identifying the HLA-A2/MART-1₂₇₋₃₅ complexes inside or on the DC is presently not available. Nonetheless, when the two DC types were analysed for the intensity of the merged signals, both cell types exhibited equal intensity of colocalization of the class I molecules and the MART-1 protein. This piece of data, when taken together with the functional efficiency of the AdVMart1-transduced iDC and inhDC (Table 2; Fig. 3), suggests that both DC types presented the epitope equally well. Further, the AdVMart1-transduced DC present the relevant epitope for a more sustained period of time compared to the peptide-loaded DC (Fig. 5b). It is therefore possible to argue that a more sustained presentation of the epitope from the internally synthesized protein (in contrast to presenting the externally loaded peptide) might have been a reason, among others, for an efficient presentation even by the IL-10-grown DC. Further studies will be needed to fully clarify the mechanistic basis for the efficiency of AdVMart1-transduced DC in activating MART-1₂₇₋₃₅ epitope-specific CTL precursors. Tumour cells have been known to synthesize immunosuppressive cytokines such as TGF-β and IL-10. DC themselves can produce IL-10 even after receiving a ‘maturation signal’.^13,44 Thus, a microenvironment with IL-10 – created by tumour cells or by the DC themselves – is unlikely to be conducive towards an efficient antigen presentation process leading to the effective generation of CTL responses. A mechanism of antigen loading onto DC that could overcome and/or reverse the inhibitory effect of IL-10 would be useful. Our data clearly point out that the phenotype of the IL-10-grown DC is not irreversibly fixed. DC polarized in IL-10 and then loaded with an antigen by adenovector-based transduction activate the antigen specific CTLs fairly efficiently. When taken with our earlier observations on adenovirus-based gene transfer into DC as a tool in the laboratory^{21,22,48–50} or with other reports from the clinic^51,52 the observations presented here have implications on cancer vaccination strategies and support the value of using viral vectors for transferring genes encoding antigens into DC so as to orchestrate a more effective antigen presentation for developing efficient vaccination strategies.^53,54