Constitutive IL-10 Production Accounts for the High NK Sensitivity, Low MHC Class I Expression, and Poor Transporter Associated with Antigen Processing (TAP)-1/2 Function in the Prototype NK Target YAC-1¹

Interleukin-10 was discovered as a T cell growth and differentiation factor (1, 2) and has the ability to suppress cytokine production by Th1 and NK cells (3). IL-10 can also have pleiotropic immunosuppressive effects, which include the capacity to block monocyte-dependent T cell proliferation (4), inhibit monocyte class II MHC expression (5), inhibit the up-regulation B7 on monocytes (6), and influence the monocyte-associated production of nitric oxides and killing of parasites (7). Furthermore, we have recently demonstrated a novel escape mechanism from MHC class I-restricted CTLs in tumor as well as allotransplant rejection mediated by IL-10 (8, 9). Human melanoma and EBV-transformed B cells that had been preincubated in medium containing rIL-10 were protected from lysis by tumor-specific as well as allospecific CD8⁺ CTLs, paralleled by a partial but not total decrease in MHC class I expression (8). Also, transfection with the gene for IL-10 transformed murine RMA lymphoma cells (9, 10) to a phenotype resembling that of the Ag-presentation defective RMA-S mutant (11, 12, 13). Phenotypic changes in IL-10-transfected cells included a resistance to CTL-mediated lysis that was concomitant with enhanced NK sensitivity and also low but peptide-inducible MHC class I expression (9). Thus, we postulated that IL-10 inhibits CTL-mediated immunorecognition by a mechanism that involves the down-regulation of or alterations in MHC class I-restricted Ag processing (9). Since the completion of the study presented here, we have also shown that IL-10 inhibits the expression and function of TAP (10).

IL-10 mRNA has been found in a variety of freshly excised human tumors, including ovarian (14), renal cell (15), squamous cell, and basal cell carcinomas (16); it has also been found in metastatic melanomas (17, 18). The majority of these studies employed RT-PCR methodology, but the presence of IL-10 protein in squamous cell carcinoma lesions (16) and in melanomas (17, 18) was confirmed by immunohistochemistry. In addition, Gotlieb et al. (19) found high serum and ascites fluid IL-10 titers in patients with gynecologic cancer. A variety of human solid tumor lines, including melanomas, colon carcinomas, lung carcinomas, skin carcinomas, and lymphomas (16, 20, 21, 22, 23), reportedly produce IL-10. Taken together, it appears that IL-10, which is secreted by tumor-infiltrating mononuclear cells or tumor cells, is a common constituent in tumor milieu.

The possibility that the constitutive production of IL-10 in tumors will affect their MHC class I-associated Ag presentation to CTLs and their susceptibility to NK cells needs to be tested experimentally. To this end, we have screened a variety of mouse tumor lines for their constitutive expression and production of IL-10. In vitro-passaged YAC-1 mouse T cell lymphoma cells, which are commonly used as the prototype for an NK-sensitive tumor line, produced high levels of IL-10 and had low levels of MHC class I expression and decreased functional activity of TAP. Our experiments demonstrate a direct correlation between constitutive IL-10 production by YAC-1 cells, high NK sensitivity, and low MHC class I expression. Furthermore, passaging YAC-1 in vivo reverts its phenotype to a low producer of IL-10 with high MHC class I expression and low NK sensitivity.

YAC is a lymphoma that is induced by the Moloney leukemia virus in A/Sn mice (H-2^a) and was propagated in the peritoneal cavity as ascites. The in vitro passaged YAC-1 was a subline that was originally adapted to grow as a stationary suspension culture (24).

RMA is a mutagenized, nonselected subline, whereas RMA-S is a mutagenized and selected (anti-H-2^b Abs and complement) subline; both of these sublines were derived from the Rauscher leukemia virus-induced lymphoma RBL-5 of C57BL/6 origin (H-2^b) (11).

All other cell lines were obtained directly or indirectly from the American Type Culture Collection (Manassas, VA).

All cells were grown in RPMI 1640 medium supplemented with 5% FCS, 200 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Life Technologies, Paisley, U.K.).

The mouse strains used were bred and maintained at the Animal Department of the Microbiology and Tumor Biology Center at the Karolinska Institute.

The retrovirus HyTk-IL10AS containing the IL-10 gene in antisense orientation will be published elsewhere. Briefly, the IL-10 cDNA-coding region has been cloned in reverse orientation into the vector HyTk-EF1α behind the internal elongation factor-1α (EF1α)³ promoter. In this vector, IL-10 antisense sequences are contained in both the full-length viral transcript that is driven by the long terminal repeat promoter and a smaller transcript that is driven by the EF1α-promoter (our unpublished observations). In addition, the vector contains a hygromycin-thymidine kinase fusion gene that allows for the selection of transduced cells (25). A virus-producer cell line derived from ψ2 was generated (ψ2-HyTk-IL10AS) that produces a viral titer of 6 × 10⁴ hygromycin-resistant colonies/ml on NIH3T3 cells. For mock infection, the same virus without IL-10 sequences was used. To generate the IL-10 transcription-inhibited cell line of YAC-1, Y-IL-10 AS, we infected the cells with ψ2-HyTk-IL10AS virus as described previously (26) and selected them with 0.5 mg/ml of hygromycin B. The control Y-IL-10 Mock cells were made with the control ψ2-HyTk virus.

Cells were grown for 48 h in RPMI 1640 medium with 5% FCS to a maximal concentration of 1.5 to 2 × 10⁶ cells/ml. Next, 100 μl of the supernatant was analyzed in duplicate by ELISA using the anti-IL-10 mAb JES5-2A5 and biotinylated SXC-1 (PharMingen, San Diego, CA). The assays were performed according to the recommendations of the manufacturer. The lower limit of IL-10 detection by the assay used was <1 U/ml or 50 pg/ml.

Total RNA was isolated from 1 × 10⁷ tissue culture cells using the RNA isolation kit (Stratagene, La Jolla, CA) and prepared according to the specifications of the manufacturer. A total of 1 μg of total RNA was reverse transcribed by the addition of 1× first-strand buffer (Life Technologies, Eggenstein, Germany), 10 μM of DTT (Life Technologies), 1 μM of random hexamers (Life Technologies), 250 μM of deoxynucleotides (Boehringer Mannheim, Mannheim, Germany), 20 U/μl of RNAsin (Life Technologies), 200 U/μl of superscript Moloney leukemia virus reverse transcriptase (Life Technologies), and diethylpyrocarbonate-treated H₂O in a 20 μl volume. Reactions were incubated at room temperature for 10 min and at 37°C for 60 min; they were subsequently heat-inactivated at 72°C for 10 min.

The primers for murine IL-10 were: sense 5′-CCTGGCTCAGCACTGCTAT-3′ (exon 1) and antisense 5′-GCAGGATCCTTAGCTTTTCATTTTGATCATC-3′ (exon 5); these primers yielded an RT-PCR product of 517 bp. PCR amplification was performed using 1 μl of the cDNA reaction, 1× PCR buffer (Perkin Elmer, Weiterstadt, Germany), 200 μM of deoxynucleotides (Boehringer Mannheim), 800 nM of each primer, and 1 U of Taq polymerase (Perkin Elmer) in a total volume of 25 μl. PCR reactions were overlayed with mineral oil and subjected to denaturation at 94°C for 2 min, annealing at 55°C for 3 min, and extension at 72°C for 3 min for 40 cycles using a Perkin Elmer DNA Thermal Cycler 480. The reaction product was visualized by subjecting 10 μl of the reaction mix to electrophoresis in 1% agarose in 1 × Tris/acetate/EDTA buffer with ethidium bromide. The specificity of the amplified target sequences was confirmed by including reaction mix amplifications without cDNA and repeating the RT-PCR once with the same result. Amplifications with β-actin-specific primers were run in parallel as a control.

In vivo-activated NK cells were obtained from fresh spleens of CBA mice that had been treated orally 24 h before splenectomy with tilorone (Sigma, St. Louis, MO) at 200 μl/mouse (10 mg/ml) in PBS.

A 4-h standard ⁵¹Cr release assay was used to measure NK cell susceptibility. Specific lysis was calculated according to the following formula: percentage of specific lysis = 100 × ([experimental cpm − spontaneous cpm]/[maximum cpm − spontaneous cpm]). All assays were performed in triplicate; spontaneous release and the percent error never exceeded 20% and 15%, respectively.

The mAbs used for H-2 detection were 36-7-5, which is specific for H-2K^k, and 34-5–8S, which is specific for H-2 D^d (PharMingen). The mAbs were used at a concentration of 2 μg/ml. Cells that had been treated with the anti-H-2 mAbs were stained with a secondary FITC-coupled rabbit anti-mouse Ig F(ab′)₂ (Dakopats, Glostrup, Denmark). Viable cells (5,000–10,000) were analyzed using a FACScan flow cytometer (Becton Dickinson, Frankin Lake, NJ).

The COOH fixed terminal peptide library sequence is T/(V, D, K), Y, N, R/A, T, R/(V, D, T), A/(V, K), L/(T, D, K), I (27) (kindly provided by Dr. H. L. Ploegh, Massachusetts Institute of Technology, Cambridge, MA). It was labeled by the chloramine-T-catalyzed iodination method with 1 mCi of Na¹²⁵I (28). Free iodine was separated from iodinated peptide by passage through a Dowex (OH⁻) column (Dow, Midland, MI). The specific activity was ∼30 μCi/μg of peptide.

Peptide translocation assays were performed according to Neefjes et al. (29). Briefly, 3 × 10⁶ cells were harvested and washed with incubation buffer (130 mM KCl, 10 mM NaCl, 1 mM CaCl, 2 mM EGTA, 2 mM MgCl₂, and 5 mM HEPES (pH 7.3)). The cells were permeabilized using 4 hemolytic units of streptolysin O (BioMerieux, Lyon, France) at 37°C for 10 min. Approximately 75 ng of iodinated peptide library was added to cells in a total volume of 100 μl incubation buffer and maintained for 10 min at 37°C in the presence or absence of 10 mM ATP (Sigma). The reaction was stopped by adding 1 ml of lysis buffer (1% Nonidet P-40, 150 mM NaCl, 5 mM MgCl₂, and 50 mM Tris-HCl (pH 7.5)). Nuclei were removed by centrifugation, and the cleared lysate was incubated with 100 μl of packed Con A-Sepharose (Pharmacia, Uppsala, Sweden) for 1 h. The Con A-Sepharose was washed five times with the lysis buffer, and the amount of the bound-labeled peptide was quantitated by gamma-counting (1282 CompuGamma, LKB Wallac, Turku, Finland). All assays were performed in triplicate and repeated at least once.

The synthetic peptides used were HIV glycoprotein 160 318-327, which was presented by H-2D^d, and the influenza matrix protein 58–66, which was used as a control (30, 31). YAC-1 cells were cultured overnight at 26°C in RPMI 1640 supplemented with 10% FCS. Subsequently, 0.5 × 10⁶ cells were preincubated with or without synthetic peptides (5 μg/ml) at 26°C for 6 h and then at 37°C for 1 h. Indirect staining using mAb 34-5–8S, which is specific for H-2 D^d, was performed and followed by FACS analysis to detect the effect of peptide binding on the surface expression of H-2D^d.

A total of 26 mouse tumor lines were tested for IL-10 mRNA expression by RT-PCR and for IL-10 production by ELISA (Table I). Of the 9 plasmocytoma lines analyzed, 4 expressed IL-10 mRNA, as did a hybridoma comprised of one of the plasmocytomas. All of the IL-10 mRNA-positive plasmocytomas examined for IL-10 production were also found to produce this cytokine. None of the fibrosarcomas (n = 3); renal cell carcinomas, bladder carcinomas, or adenocarcinomas (n = 3); melanomas (n = 2); or mastocytomas (n = 1) were found to express or produce IL-10. A total of 3 of 6 T cell lymphomas expressed IL-10 mRNA, but only YAC-1 cells also produced detectable levels of IL-10.

As in vitro YAC-1 cells are the prototype NK-sensitive target and have relatively low levels of cell surface MHC class I proteins, we investigated the possible relationship between constitutive IL-10 production and this phenotype (Fig. 1, A–C). High IL-10 production was only observed when long-term in vitro-established YAC-1 cells were examined, whereas propagation as an ascites tumor in vivo resulted in a loss of IL-10 production (Fig. 1,A). Upon in vitro explantation, IL-10 production was not observed until after 1 to 2 wk in culture, and IL-10 levels were one-third of the original amounts. IL-10 production did not return to levels that were typical of normal in vitro-maintained YAC-1 cells until after more than 3 wk of in vitro culture (Fig. 1 A).

It has been described previously that YAC lymphoma gradually loses H-2 expression as a result of in vitro culture, whereas the reverse was shown when in vitro-cultured cells were reinoculated into syngeneic or semisyngeneic animals (24, 32, 33, 34). In this study, we have confirmed this decrease in H-2 expression following in vitro culture of the in vivo-passed YAC (Fig. 1,C) and found that such a decrease is in parallel with a gradual increase in IL-10 production (Fig. 1 A). Thus, 1 to 2 wk of in vitro culture resulted in a 20 to 50% loss of H-2 (D^d) cell surface expression; following 3 wk of culture, H-2 levels were had decreased to match those of the long-term-established YAC-1 line. Similar results were seen with regard to H-2 K^k expression (data not shown).

Although in vitro-maintained YAC-1 cells were used for the definition and naming of NK cells (35) when grown in vivo as ascites cells in syngeneic mice and tested as freshly explanted target cells by a standard ⁵¹Cr release assay, these cells are relatively NK-resistant (Fig. 1 B). In vitro culture of in vivo-grown YAC-1 cells resulted in a gradual recovery of NK susceptibility that reached levels close to that of long-term-cultured YAC-1 cells within a period of 3 wk; this finding is in line with previously published results (24, 32, 33, 34).

To investigate whether differences in MHC class I expression on the in vitro-cultured vs the in vivo-grown YAC-1 would be paralleled by alterations in the functional activity of the TAP-1/2 proteins, we performed peptide-translocation assays according to the method of Neefjes et al. (29). TAP-2 mutant RMA-S cells showed a total inability to transport peptides in an ATP-dependent manner (Fig. 2); this observation is in contrast to the high TAP activity of the wild-type RMA line. Also, freshly explanted in vivo-grown YAC-1 showed high TAP functional activity. The high TAP activity that is characteristic of in vivo-passaged YAC-1 cells diminished to levels that are characteristic of in vitro-passaged cells after 1 mo (Fig. 2). The original YAC lymphoma from which YAC-1 was first established (24) showed a similar in vitro culture-dependent decrease in TAP function, with high TAP activity of the in vivo-passaged line that returned to low levels after 1 mo of in vitro culture (Fig. 2).

To analyze whether constitutive IL-10 production was causally related to low H-2 expression and high NK sensitivity, YAC-1 cells were infected with a retroviral construct that contained the IL-10 gene in an antisense orientation driven by the long terminal repeat and EF1α promoters; stable transfectants were then obtained by selection. The YAC-1 cells infected with the IL-10 antisense construct (Y-IL-10 AS) produced only 14 U/ml of IL-10, as compared with the 46 U/ml produced by YAC-1 infected with the control construct (Y-IL-10 Mock) or the 44 U/ml produced by parental YAC-1 (Fig. 3 A).

The decreased IL-10 production in Y-IL-10 AS cells coincided with an increased expression of H-2, with a mean fluorescence intensity that was in between that of recently in vivo-passed YAC-1 and the long-term in vitro-cultured YAC-1 or the Y-IL-10 Mock-transduced cells (Fig. 3,C). Y-IL-10 AS cells were also significantly less NK-sensitive as compared with Y-IL-10 Mock and YAC-1 (Fig. 3 B), with a percentage of lysis value that was intermediate (44% reduction) to that of the in vitro-passaged YAC-1.

There was a highly significant correlation between the levels of H-2 expression and the production of IL-10 among the different transfectants and the in vitro/in vivo-passed sublines of YAC-1 (Fig. 4,B). A highly significant correlation among these YAC lines was also found between their levels of NK sensitivity, which were expressed as the percentage of lysis at a 50:1 or 25:1 E:T ratio, and the amount of IL-10 produced (Fig. 4,A, r² = 0.986 at an E:T ratio of 50:1; and r² = 0.989 at an E:T ratio of 25:1). In addition, the levels of H-2 expression strongly correlated with the levels of NK sensitivity among the same YAC lines (Fig. 4 C, r² = 0.961; and 0.943, respectively).

By incubating YAC-1 at 26°C overnight, a 56% increase in cell surface H-2 expression was observed as compared with incubation at 37°C (data not shown). The addition of H-2 D^d binding peptides following this incubation period resulted a 90% increase in cell surface expression, while a non-H-2 D^d binding peptide yielded an increase in H-2 expression (60%) that was similar to that seen with incubation at 26°C without the addition of peptide. These observations are in line with a deficient TAP-1/2 function and are similar to what was shown previously for IL-10-transfected RMA and TAP-2 mutant RMA-S (9, 36).

Taken together, we have been able to establish a casual relationship between constitutive IL-10 production and the low H-2 expression/high NK sensitivity of YAC-1 lymphoma in this study; we have also shown that this phenotype is associated with poorly functioning TAP-1/2 complexes.

It is possible that the IL-10-mediated TAP-1/2 dysfunction is sufficient to explain the enhanced NK susceptibility of in vitro-established YAC-1, analogous to how a mutated TAP-2 gene is sufficient to explain the NK-susceptible phenotype of RMA-S (37). Because TAP-1/2 function in the IL-10-producing YAC-1 was not as low as in the RMA-S mutant, decreased TAP-1/2 function might not fully account for the enhanced NK susceptibility; the additional effects of IL-10 on other members of the MHC class I Ag presentation machinery cannot be excluded. To what extent the high NK sensitivity of YAC-1 and RMA-S depends upon a decrease in the total MHC class I content or, alternatively, on the appearance of “empty” MHC class I molecules or MHC class I molecules containing an altered TAP-independent peptide repertoire remains to be shown.

It has been shown previously that the mechanism favoring the in vivo outgrowth of the NK-resistant, H-2^high-expressing phenotype of YAC is T cell-independent but requires a mature immune system (34). The NK-mediated cytotoxic selection and/or IFN-γ-mediated induction of an NK-resistant H-2^high phenotype (38, 39, 40) may be the active in vivo mechanisms. However, we were never able to detect IFN-γ (<0.5 U/ml by ELISA) in the ascites fluid or in the medium of freshly explanted YAC ascites cells (our unpublished observations), which argues against a role for this cytokine. Furthermore, while we have shown that the NK-resistant H-2^high phenotype of in vivo-passaged YAC-1 persists for several weeks, the effects of IFN-γ treatment are not as long-lasting (data not shown). The possibility that the H-2^high phenotype of the Y-IL-10 AS is an in vitro artifact mediated by the antisense construct via the induction of IFN-γ can also be excluded, since IL-10 antisense-transduced cells do not express IFN-γ mRNA or produce this cytokine (our unpublished observations).

IL-10 is the first described example of a cytokine with a suppressive effect on the MHC class I Ag-presentation pathway. Our results underline the opposing effect between IL-10 and IFN-γ, a cytokine that is known to protect tumor cells from NK cytotoxicity (39, 40) and enhance MHC class I-restricted Ag presentation, including TAP-1/2 function (41, 42, 43). In tumors or virus-infected cells, the induction of IL-10 production might represent an important mechanism of escape from attack by specific T cells, as we have suggested previously (8, 9). Viral genes and proteins with a capacity to inactivate TAP-1/2 have also been described previously (44, 45, 46, 47). Intervention with these mechanisms by antisense technology as shown here, leading to enhanced MHC class I Ag presentation and possibly to improved host immunity, may prove to be an efficient new therapeutic modality. As the same approaches may have an opposing effect on NK-mediated host resistance, their net effect on the rejection of transformed or virus-infected cells in vivo has to be analyzed for each virus or tumor type.

Note. Additional proof of an effect by IL-10 on TAP function has been published by us and others since the original submission of this manuscript (10, 48). Furthermore, in a recent paper (49) confirming our original observation of the effect of IL-10 on MHC class I in melanomas (8), Yue et al. showed that long-term culture in neutralizing anti-IL-10 Ab will provide the same effect as the antisense approach that we used.

We thank Dr. Anne O’Garra at DNAX Research Institute (Palo Alto, CA) for kindly providing rIL-10 and S. Lupton for plasmid tgLS⁺HyTk. The technical assistance of Maj-Lis Solberg, Margareta Hagelin, and Marcelo Toro is gratefully acknowledged. We are also grateful to Dr. Ken Wasserman, Microbiology and Tumor Biology Center, for correcting the language and commenting on this manuscript.