A Synthetic Peptide Homologous to Functional Domain of Human IL-10 Down-Regulates Expression of MHC Class I and Transporter Associated with Antigen Processing 1/2 in Human Melanoma Cells¹

Interleukin-10 is a protein of 160 aa with a molecular size of 18.5 kDa, and it exists as a 37-kDa homodimer (1). IL-10 was discovered as a T cell growth and differentiation factor (2, 3) and has the ability to suppress cytokine production by monocytes (4), activated macrophages (5), and Th1 cells (6). IL-10 can also have pleiotropic immunosuppressive effects, including the capacity to block monocyte-dependent T cell proliferation (7), and can inhibit class II MHC expression on monocytes (8), dendritic cells (DC)³ (9), and Langerhans cells (10). IL-10 can also inhibit the up-regulation of B7 on monocytes (11), monocyte-associated production of NO, and killing of parasites (12). Overexpression of IL-10 in IL-10 transgenic mice was shown to lead to decreased capacity to reject tumors (13).

Interest in the immune inhibitory effects of IL-10 in relation to antitumor responses was based on the recurrent presence of this cytokine in fresh tumor biopsies or ascites fluid derived from cancer patients (14, 15). Expression of IL-10 mRNA was found in various kinds of solid tumors, such as ovarian carcinomas (16), renal carcinomas (17), nonsmall lung cancer cells (18), and melanomas (19). The preferential expression of IL-10 in metastatic vs primarily lesions of melanomas has been established, indicating an increased metastatic potential of IL-10-secreting melanoma cells (20).

We have previously demonstrated that murine tumor cells treated with rIL-10 or transfected with the IL-10 gene showed a changed phenotype characterized by a down-regulation in MHC class I expression, resistance to lysis by tumor-specific CTL, and increased sensitivity to lysis by NK cells related to an inhibition in the function and expression of TAP molecules (21, 22, 23). In human melanoma, we demonstrated that human rIL-10 (rhIL-10)-pretreated melanoma cell lines became less sensitive to tumor-specific CTL and also showed a significant reduction of MHC class I expression (24). These observations have been confirmed in studies in which IL-10 was described to down-regulate HLA class I and class II and also ICAM-1 molecules in melanoma cell lines (25).

A synthetic nanomeric peptide IT9302 (Ala-Tyr-Met-Thr-Met-Lys-Ile-Arg-Asn), with complete homology to an amino acid sequence located in the C-terminal portion of human IL-10 (residues 152–160), has been shown to be the functional domain of IL-10, sharing several of the functions of this cytokine (26). In this study, we have analyzed the effect of this IT9302 peptide on the expression of MHC class I-related Ag presentation proteins in melanoma cells. We demonstrated that the IL-10 homologous peptide down-regulates the MHC class I surface expression and also the IFN-γ-induced expression of MHC class I, TAP1, and TAP2 molecules, affecting NK sensitivity of treated tumor cells.

The FM55 and FM3D melanoma cell lines were obtained from the Danish Cancer Society through the courtesy of J. Zeuthen. The OCM1 is an ocular melanoma line kindly provided by M. Jager (University of Leiden, Leiden, The Netherlands). These cell lines were negative for IL-10 expression and positive for IL-10R expression, and were maintained in culture in RPMI 1640 medium supplemented with 5% FCS, 1 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin sulfate (Invitrogen Life Technologies, Gaithersburg, MD). The nanomeric peptide IT9302 (AYMTMKIRN) and the control peptide IT6542 (SPGQGTQSE) corresponding to a nonfunctional IL-10 domain (26) were chemically synthesized using an automatic polypeptide synthesizer. The purity of the proteins was confirmed to exceed 95% by HPCL, and the m.w. was controlled by mass spectroscopy. In some experiments, an irrelevant nanomeric peptide derived from influenza matrix protein (GILGFVFTL) was used as a control.

Melanoma cells (6 × 10⁵ cells in 3 ml) were cultured in AIM-V serum-free medium (Invitrogen Life Technologies, Gaithersburg, MD) in the presence of 1, 10, and 100 ng/ml peptide IT9302 or an irrelevant control peptide (100 ng/ml) during 24 h. To investigate the inhibition of IFN-γ-mediated induction of Ag presentation proteins, cells were pretreated during 2 h with different concentrations of IT9302 peptide or control peptide and then incubated in the presence of only medium, or medium containing 100 or 500 U/ml rIFN-γ IMUKIN (Boehringer Ingelheim, Ingelheim, Germany) for additional 22 h for FACS, or Western blot analysis. To specifically block the IT9302-mediated effect, medium containing 100 ng/ml IT9302 was preincubated with 10 μg/ml rabbit polyclonal anti-IT9302 Ab (26) or 10 μg/ml normal rabbit serum as control and then added to cell cultures.

Melanoma cells either untreated or stimulated for 24 h with IT9302 and with or without rIFN-γ were washed in PBS and incubated for 30 min on ice in RPMI 1640 and 5% FBS with mAb W6/32 (anti-MHC class I) (American Type Culture Collection (ATCC), Manassas, VA), mAb HB54 (anti-HLA-A2) (ATCC), or mouse IgG as negative control (BD Pharmingen, San Diego, CA). After two washings, the cells were incubated with a secondary anti-mouse IgG FITC (DakoCytomation, Hamburg, Germany) for 30 min. Cells were analyzed after four washings. For intracellular detection of TAP molecules, anti-TAP1 and anti-TAP2 mAbs (BD Pharmingen, San Diego, CA) were used after cell permeabilization with digitonin (Sigma-Aldrich, St. Louis, MO). Flow cytometry was performed using a FACScan flow cytometer and analyzed by the CellQuest analysis program (BD Biosciences, Heidelberg, Germany).

Cell pellets from harvested melanoma cell lines (1 × 10⁶ cells) were suspended in 100 μl of lysis buffer (65 mM Tris, pH 6.8, 2% SDS, 10% glycerol, 5% 2-ME, 1% bromphenol blue) and maintained for 15 min at room temperature. Then samples were sonicated by 10 μm for 30 s, warmed at 95°C for 5 min, and then centrifuged. A total of 15 μl of each sample was electrophoresed through a 12% SDS-polyacrylamide gel. For immunoblots, proteins were electrotransferred onto polyvinylidene difluoride membrane (Immobilon-P; Millipore, Bedford, MA). Membranes were blocked in PBS containing 5% milk (low fat). All additional immunostaining steps were performed in PBS with 3% milk and washed with PBS 0.05% Tween 20 (PBS-Tween 0.05%). Filters were incubated overnight at 4°C with the corresponding primary Ab: mouse mAb TO-1 anti-TAP1, mouse mAb SY-2 anti-TAP2 (both kindly provided by S. Ferrone (Roswell Park Cancer Institute, Buffalo, NY)), or a rabbit polyclonal Ab against MHC class I H chain (also kindly provided by S. Ferrone). As an internal control, a commercial polyclonal Ab against β-actin was used (Sigma-Aldrich, St. Louis, MO). After washing with PBS/milk, the membranes were incubated with the respective secondary Ab (HRP-conjugated sheep anti-mouse Ig or HRP-conjugated sheep anti-rabbit Ig (Amersham Biosciences, Buckinghamshire, U.K.)) for 2 h at room temperature. Filters were then washed in PBS-0.05% Tween five times and developed with ECL system (Amersham Biosciences).

Lymphokine-activated killer (LAK) cells were obtained from normal donors’ PBMC. After a gradient separation with Lymphoprep (Invitrogen Life Technologies), 25 × 10⁶ PBMC were incubated during 5 days in 15 ml of RPMI 1640 and 10% FBS medium containing 500 U/ml IL-2 (kindly supplied by P. Simon, DuPont Merck Pharmaceutical, Wilmington, DE). A 6-h ⁵¹Cr release assay was used to measure cytotoxic activity of LAK cells against peptide- and/or IFN-γ-treated melanoma cell lines or K562 cells as positive control. All samples were run in triplicate. Specific lysis was calculated according to the formula: percentage of specific lysis = 100 × ((experimental cpm − spontaneous cpm)/(maximum cpm – spontaneous cpm)).

The differences in TAP expression between triplicates were compared by Student′s t test using a computer program, and were considered statistically significant at p < 0.05

The nanomeric peptide IT9302 has been demonstrated to constitute the functional domain of human IL-10 (26). To investigate whether IT9302 could also mimic the capability of IL-10 to down-regulate MHC class I surface expression on melanoma cells, we treated the cutaneous melanoma line FM55 and the ocular melanoma line OCM1 with 0, 1, 10, and 100 ng/ml IT9302 during 24 h. The peptide concentrations were selected considering that 6125 ng of peptide corresponds in molarity to ∼100 ng/ml human IL-10 (26). A dose-dependent inhibition of MHC class I could be observed in both treated melanomas, with a maximal inhibition of 60% for OCM1 (Fig. 1,A) and 50% for FM55 (Fig. 1,B) compared with the controls. These results were in line with previously observed effects using rhIL-10 (24). The effect of the IT9302 peptide and the control peptide on HLA class I expression was measured also on the HLA-A2⁺ FM3D melanoma line. Of interest, while a strong effect on total MHC class I was noted (mean fluorescence intensity (MFI) of 92.5) following IT9302 incubation as compared with MFI of 152 with the control peptide and MFI of 162 for medium only for the FM3D line, using mAb HB54, the effect on HLA-A2 expression was only marginal (MFI of 35.6 following IT9302 incubation as compared with MFI of 47.8 with the control peptide and MFI of 45.5 with medium only) (Fig. 1 C). The reduced effect of IT9302 on HLA-A2 as compared with the effect on total HLA class I could be explained by the known ability of HLA-A2 to present peptides in a TAP-independent way (27, 28, 29).

The IT9302 peptide was previously shown to inhibit the IFN-γ-induced MHC class II expression on human monocytes (26) in line with the known ability of IL-10 to antagonize the effect that IFN-γ has on Ag presentation (30, 31). In this study, we investigated whether the peptide homologue could antagonize the IFN-γ-mediated enhancement of MHC class I H chain expression. OCM1 melanoma cells were first treated with 0, 1, 10, and 100 U/ml IT9302 peptide, medium alone, or 100 U/ml control peptide for 2 h, and thereafter with 100 or 500 U/ml rIFN-γ for an additional 22 h. In this experiment, a 2- to 3-fold induction of MHC class I surface expression was observed after treatment with the two concentrations of IFN-γ. A dose-dependent inhibition of the basal level of MHC class I and, most importantly, of the IFN-γ-induced expression of MHC class I was observed when the cells were pretreated with increasing doses of IT9302 and thereafter IFN-γ (100 U/ml) (Fig. 2). The highest inhibition of MHC class I induction was observed in cells pretreated with 100 ng/ml IT9302 (Fig. 2). The weaker inhibition of MHC class I by IT9302 peptide observed when the cells were treated with a high dose of IFN-γ (500 U/ml) indicated that MHC class I expression is delicately regulated by a balance between both cytokines (Fig. 2).

We and others have previously shown that IL-10 could inhibit the function and expression of TAP1/TAP2 molecules (22, 23, 32). Many tumors, including melanomas, are known to express low levels of TAP molecules, which can be induced by IFN-γ (33, 34, 35, 36). We could confirm that this was also the case for the FM55 melanoma line, as measured by RT-PCR (data not shown), and for OCM1 line measured by Western blot (Fig. 3).

To clarify whether IT9302 could also down-regulate TAP molecules, we induced their expression by treatment with IFN-γ. Using Western blot analysis, we could observe that the constitutive expression of TAP1, TAP2, and LMP2 in the melanoma line OCM1 was very low, at the limit of detection (Fig. 3). In contrast, OCM1 cells treated with 100 U/ml rIFN-γ showed a strong expression of TAP1, TAP2, and LMP2 proteins (Fig. 3). IT9302 markedly inhibited the IFN-γ-induced TAP1 and TAP2 expression in a dose-dependent manner, with a strong inhibition noted already at a concentration of 10 ng/ml peptide. The IFN-γ-induced expression of the LMP2 protein (Fig. 3) or MHC H chain (data not shown) was, however, not affected by the IT9302 peptide. The TAP1 and TAP2 were selectively down-regulated by IT9302, as the expression of the housekeeping protein β-actin used as control in this experiment was unaffected. Also, in the FM55 melanoma line, IT9302 and rhIL-10 did inhibit TAP1 and TAP2 expression (data not shown). Using intracellular staining and analysis by flow cytometry, it was observed that the level of expression of TAP1 (Fig. 4,A) and TAP2 (Fig. 4 B) could be almost completely recovered in the presence of a polyclonal Ab directed against peptide IT9302, but not using a normal rabbit serum as control (data not shown), both in the absence and presence of IFN-γ, confirming the direct effect of this peptide on the expression of TAP1/2 molecules.

We have previously demonstrated that IL-10 could convert murine tumor cells into an NK-sensitive phenotype (21, 22, 23). To investigate whether the IT9302 peptide also could enhance NK sensitivity of human melanoma cells, we performed cytotoxicity assays with LAK cells as effector cells. The melanoma cell line OCM1 pretreated with different concentrations of the IT9302 peptide was found to be more sensitive to LAK cells than melanoma cells treated with a control peptide. The effect of the IT9302 peptide was dose dependent, and the NK sensitivity of the OCM1 line treated with 100 ng/ml IT9302 approached that of the standard NK target cell K562 (Fig. 5,A). A comparable effect was observed using the melanoma line FM55 (Fig. 5 B). These results showed that the changes observed in surface expression of MHC class I translated into enhanced sensitivity of tumor cells to lysis mediated by LAK cells.

It has been demonstrated recently that small molecules such as oligopeptides or other low m.w. nonpeptide structures can act as functional agonists for intact proteins, including some cytokines (26, 37, 38, 39). Those small molecules can interact with cell membrane-specific receptors and mediate intracellular signaling with similar consequences as those induced by the whole protein. They thus offer new and promising avenues to therapeutic interventions based on structurally defined small molecules, which may have several advantages as compared with the corresponding whole protein. In this study, we have analyzed the effect of the synthetic peptide IT9302 on the expression of MHC class I-related Ag presentation proteins in melanoma cells expressing IL-10R. Our observations demonstrate that the IL-10 homologous peptide down-regulates the MHC class I surface expression and also the IFN-γ-induced expression of TAP1 and TAP2 molecules, affecting NK sensitivity of the tumors. These results therefore are the first to show that a peptide homologue of IL-10 markedly can affect MHC class I-related molecules and function in human tumors.

The synthetic nanomeric peptide IT9302, which we have used in this study, has a complete homology to an amino acid sequence located in the C-terminal portion of human IL-10 (residues 152–160), which has been shown to be the functional domain of IL-10 (26). The functions of IL-10 that this peptide previously was demonstrated to mimic are: inhibition of IL-1β-induced IL-8 production by monocytes, inhibition of spontaneous IL-8 expression by in vitro cultured monocytes, down-regulation of TNF-α production by CD8⁺ lymphocytes, induction of IL-4 production by CD4⁺ lymphocytes, and inhibition of IFN-γ induction of MHC class II on monocytes (26). Our results therefore demonstrate that the same domain on IL-10, namely the C-terminal portion, also determines the effect of this molecule on MHC class I expression. In previous studies, we have demonstrated that IL-10 has an inhibitory effect on Ag class I presentation, diminishing the expression of surface MHC class I in murine tumors (21, 22, 23) and human melanoma cells (24). We, in this study, demonstrated a similar inhibitory effect mediated by the IL-10 agonist peptide IT9302. Two melanoma cell lines, FM55 and OCM1, were shown to decrease surface MHC class I expression in a dose-dependent manner when they were treated with IT9302.

Several of the inhibitory effects mediated by IL-10 are opposite to those reported for IFNs, which were shown to up-regulate the expression of several members of the MHC class I Ag presentation pathway (30, 31). This was associated with increased sensitivity for CTL, but protection from NK cell cytotoxicity (40). Therefore, IL-10 is able to block not only the induction of IFN-γ by Th1 cells (4), but also to have an opposite effect from IFN-γ on Ag presentation (31). The net effect on Ag presentation of having both IFN-γ and IL-10 expressed simultaneously, as sometimes seems to be the case in, e.g., ovarian carcinomas (16), was analyzed using our melanoma model. First, we studied whether IT9302 could inhibit the MHC class I induction mediated by IFN-γ. Our results showed that the pretreatment of melanoma cells with the IT9302 peptide could inhibit both the constitutive and the IFN-γ-induced surface MHC class I expression in a dose-dependent manner. IT9302 could more efficiently inhibit MHC class I expression induced by lower concentrations of IFN-γ (10–100 U/ml) than MHC class I induced by the higher concentration (500 U/ml), indicating that the MHC class I regulation responds to a balance between both cytokines. IL-10 is known to inhibit IFN-γ-induced class II Ag expression on mononuclear cells (8). This function appears to be located to the segment of IL-10, which is represented by the IT9302 peptide, and our results therefore demonstrate that also the ability of IL-10 to block induction of MHC class I is confined to the same region of IL-10.

It has been described that several tumors, including melanomas, commonly have a defective TAP expression (35, 36, 41). In confirmation of this, the OCM1 and FM55 melanomas tested in this study expressed very low levels of these proteins in a constitutive form (Figs. 3 and 4, A and B). Only after treatment with IFN-γ was it possible to efficiently detect the expression of TAP1 and TAP2 by Western blot. Both proteins were substantially reduced when the tumor cells were pretreated with IT9302, while MHC class I H chain protein and LMP2, also involved in MHC class I Ag presentation, were not affected. The reduced effect of IT9302 on HLA-A2 (<20%), as compared with the effect on total HLA class I (∼60%) (Fig. 1 C), could be explained by the known ability of HLA-A2 to present peptides in a TAP-independent way (27, 28, 29). The close relation between TAP inhibition and low surface expression of MHC class I has also been clearly established in various studies (33, 34, 35, 36), including our previous works (21, 22, 23).

The effect of IL-10 on MHC class I Ag presentation has been observed also in normal, nonmalignant cell types, including B cells and DC (32, 42), and in tumors and infected cells (43, 44, 45). Thus, cellular and viral IL-10 were shown to affect Ag presentation and MHC class I expression in EBV-infected human B lymphocytes through their ability to reduce TAP1 expression (32). These authors speculate that IL-10 could therefore affect TAP1 levels in vivo, contributing to the persistence of EBV-infected cells in the human host. Furthermore, autocrine IL-10 secretion was shown to inhibit MHC class I in DC, affecting their Ag presentation capability (42). Therefore, it would be important to investigate whether the observed effects of IT9302 can be generalized also to B cells and DC, which would have a direct implication on the ability of this peptide to modulate Ag presentation in vivo.

As cells with defective TAP1/2 function, due to mutations (34, 35) or to IL-10 expression (22, 23, 32), are more susceptible to NK cell-mediated cytotoxicity, we investigated whether the pretreatment of melanoma with IT9302 could also induce a similar effect. Our results showed that IT9302 peptide pretreatment made the melanoma lines OCM1 and FM55 more sensitive to NK-mediated lysis (Fig. 5). NK cells may therefore more readily reject IL-10-expressing tumors. Several lines of evidence indicate that the dose and the mode of administration of IL-10 may affect the in vivo tumor rejection (46). High doses of IL-10 may have a stimulatory effect on the immune response, activating T cells, NK cells, and also macrophages and monocytes. An IL-10 transfectant of the murine mammary tumor cell line (TSA) producing high levels of IL-10, is more rapidly rejected than the control line, an effect explained as an activation of CTL and NK cells (47). Similar effects have been observed in several strains of mice and with a variety of IL-10-producing transfected tumors, such as the murine melanoma CL8–1 (48), the murine melanoma B16 (49, 50), and the mammary carcinoma 410-4 (51). The role of NK cells in these systems is supported by experiments showing that Chinese hamster ovary cells transfected with murine IL-10 (52), and the murine lymphoma cell line (RMA) transfected with the IL-10 gene (our unpublished results) are more readily rejected in T cell-deficient SCID mice. Taken together, these observations demonstrate that IL-10 has profound effects on the NK cell-mediated tumor rejection in vivo, possibly mediated via the effect of IL-10 on TAP1/2 function, as described in this work. Recently, it has been demonstrated that the IT9302 peptide in vivo stimulates the production of proinflammatory cytokines in the serum of rabbits after experimentally induced acute pancreatitis (53), indicating that the peptide can exert its effect across the species barrier. In view of these results, it would be of particular importance to investigate the effects on host antitumor responses following in vivo administration of the IT9302 peptide to mice with NK-sensitive experimental tumors. This may lead to new therapeutic strategies based on coadministration of the IT9302 peptide and NK-activating cytokines to cancer patients.

We thank Dr. Borbala Gesser for providing valuable information about IT9302 peptide and controls. We also thank Felix Gonzalez, Manuel Salazar, and Marisol Briones for technical help.