Mouse tumor models studied were D122 metastatic clone of the Lewis lung carcinoma (Eisenbach et al., 1984), F3II mammary adenocarcinoma (Alonso et al., 1998), CT26 colon carcinoma (ATCC CRL-2638), 4T1 mammary carcinoma (ATCC CRL-2539), and B16F10 metastatic clone of B16 melanoma (Poste et al., 1980). For human tumor models we used A431 (epidermoid carcinoma; ATCC CRL-1555), H125 (lung adenocarcinoma) (Carney et al., 1985), and U1906 (small cell lung carcinoma) (Bergh et al., 1985) cell lines. Cells were grown in RPMI-1640 medium supplemented with 10% fetal calf serum (FCS) and antibiotics (Life Technologies, Gaithersburg, MD, United States). All cells were grown at 37°C in a humidified 5% CO₂ incubator. 7A7 and nimotuzumab were obtained at the Center of Molecular Immunology (Havana, Cuba). Cetuximab was manufactured by Merck Pharma GmbH (Deutschland, Germany). AG1478 was purchased from LC Laboratories (Woburn, MA, United States). AG1478 was dissolved in DMSO for in vitro studies and in 0.05% (v/v) Tween 80 solution for in vivo use. Mouse and human IFN-γ were purchased from Sigma (St. Louis, MO, United States). Mouse EGF that stimulates efficiently mouse and human EGFR-positive cells was obtained from R&D Systems (Minneapolis, MN, United States). Antibodies specific for H-2K^b (HB-11), H-2D^b (HB-27), H-2K^d (K9-18), H-2D^d (34-5-8), and HLA-ABC (W6/32) molecules were obtained from ATCC mouse hybridomas (Rockville, MD, United States) and used for flow cytometry analysis. Antibodies specific for HLA-A (108265) (Lozano et al., 1989) and HLA-B (42-IB5) (Lozano et al., 1990) molecules were also used for flow cytometry analysis.

Cells were plated (1.25 × 10⁵) in 10% FCS RPMI-1640 medium in six-well plates (Costar, Cambridge, MA, United States). To measure the effect of EGFR activation on MHC-I HC, β₂-m and APM components expression, 12 h later the tissue culture medium was replaced with 1% FCS RPMI-1640 containing EGF (0.5 nM for mouse cells or 0.08 nM for human cells) and cells were incubated for 48 h. Control cells were cultured in absence of EGF. To determine the effect of EGFRIs on MHC-I HC, β₂-m and APM components expression, 12 h later AG1478 (5 μM), 7A7 (1 μg/ml), nimotuzumab or cetuximab (10 μg/ml) were added in 1% FCS RPMI-1640 medium supplemented with EGF and the cells were incubated for an additional 12, 24, 48, 72, 96, or 120 h. Control cells were grown without EGFRI treatment. Maximum induction of MHC-I expression was obtained by treatment with mouse (100 UI/ml) or human (800 UI/ml) IFN-γ during 48 h.

EGFR membrane expression was analyzed in mouse cell lines using 7A7 as primary antibody. MHC-I membrane expression was detected in mouse and human cells in basal conditions and after indicated treatments using specific antibodies described above. FITC-labeled rabbit anti-mouse Ig (Fab₂) (Sigma) was used as a secondary antibody. Flow cytometry experiments were done according to standard methods: 5 × 10⁵ cells were washed twice in PBS supplemented with 2% FCS and 0.1% sodium azide. Cells were incubated with the primary antibodies at saturating concentration for 30 min at 4°C. The secondary antibody was used at a 1:100 dilution and incubated with cells for 30 min at 4°C in the dark. Isotype-matched non-immune mouse IgG and cells labeled with only the FITC-conjugated antibody were used as controls. Analyses were performed on a BD FACScan flow cytometer and the data were processed using WinMDI 2.8 software package (The Scripps Research Institute, La Jolla, CA, United States).

All animal studies were carried out in accordance with the recommendations of the Institutional Animal Care and Use Committee of the Center of Molecular Immunology which approved the protocols. Six- to eight-week-old female C57BL/6 mice (Center for Laboratory Animal Production, Havana, Cuba) were challenged with D122 cells (5 × 10⁵) subcutaneously (s.c.) into dorso-lumbar regions and the development of primary tumors was monitored. The largest perpendicular diameters of the resulting tumors were measured with caliper and tumor volume was calculating using the formula: π/6 × length × width². Twelve days later when tumor reached ∼0.25 cm³, the animals were randomly separated into six groups: intratumoral 7A7 (56 μg), intratumoral AG1478 (1 mg), intratumoral control (PBS), systemic 7A7 [56 μg intraperitoneal (i.p.)], AG1478 systemic (1 mg orally) and systemic control (PBS i.p.). EGFRIs administration began at day 12 and continued every 72 h until day 21. On day 22 mice were euthanized and primary tumors were removed, digested and prepared as a single-cell suspension. MHC-I surface expression was analyzed in cell suspensions by flow cytometry as described above.

D122-specific CD8⁺ T cells were generated as described previously (Garrido et al., 2011a) and co-cultured with D122 or B16F10 cells pretreated with EGFRIs (96 h) or IFN-γ (48 h). Cytotoxic activity was determined in 4 h using an in vitro lactate dehydrogenase assay. Percentage of specific lysis was calculated as: ([experimental release – effector cell release – spontaneous release]/[maximum release – spontaneous release]) × 100. Maximum release was obtained by adding 1% Triton X-100 to target cells, and spontaneous release was determined by incubating target cells with medium alone. Culture supernatants were collected after 48 h and IFN-γ concentration was determined by mouse IFN-γ ELISA kit (eBioscience, San Diego, CA, United States) according to manufacturer’s instructions. To measure the relative contribution of H-2K^b versus H-2D^b alleles to antitumor CD8⁺ T cell response, antibodies specific for these molecules (10 μg/ml) were added during the 48 h incubation period.

Total RNA from cells was isolated by TRIzol method and was reverse transcribed using a High Capacity Reverse Transcription Kit (Applied Biosystems, Foster City, CA, United States). SYBR Green-based quantitative real-time PCR for several mouse and human genes was performed with the Applied Biosystem 7500 Real Time PCR System. Samples for each experimental condition were run in quadruplicate using mouse or human GADPH as housekeeping genes. PCR conditions were 40 cycles of 15 s of denaturation at 95°C and 60 s at 60°C. Primer sequences were manually designed and listed in Supplementary Table S1. The expression of each target gene is presented as the fold change relative to the expression in untreated cells.

Data were expressed as means ± SD. The Student’s t-test was used to compare mean values. For comparisons between three or more groups, an analysis of variance (ANOVA) was performed, followed by multiple comparisons of means by Dunnet or Bonferroni post-test. When the variables were not normally distributed, Mann–Whitney U-test was used to compare variables between two groups. When three or more groups were compared, the Kruskall–Wallis test with Dunn post-test was used. A significance level of P < 0.05 was assumed for all statistical tests. SPSS software (IBM, Armonk, NY, United States) was used for the data analyses. All statistical tests were two-sided.

We determined whether nimotuzumab mediates upregulation of HLA class I molecules and APM components in tumors. Previous results from Ferris and colleagues have suggested a dependence on EGFR density for cetuximab-mediated HLA class I modulation (Srivastava et al., 2015). Consequently, we selected two human tumor cell lines having high to intermedium EGFR expression levels, A431 that express 2–3 × 10⁶ EGFR molecules per cell (Haigler et al., 1978) and H125 that express 2.1 × 10⁵ EGFR molecules per cell (Suarez Pestana et al., 1997). U1906 cell line which does not express the EGFR at all (Suarez Pestana et al., 1997) was used as negative control. First, we examined how EGFR signaling impacts HLA class I membrane expression in these tumor cell lines. Previous findings showed that A431 and H125 cells have a maximum increase of cell viability after stimulation with 0.08 nM of EGF (Garrido et al., 2011b). EGF-mediated EGFR activation reduced significantly the expression levels of HLA-ABC molecules in A431 and H125 cells when compared with cells cultured in the absence of EGF (Figure 1A). A ≈4-fold decrease of HLA class I molecules expression was obtained for both EGF-cultured tumors. As expected, this effect was not confirmed in EGF-treated U1906 cells (Figure 1A). Subsequently, we explored the effect of nimotuzumab on HLA class I membrane expression, and included others EGFR inhibitors (cetuximab and AG1478) and IFN-γ as positive controls. Nimotuzumab at 10 μg/ml provoking a suboptimal reduction in cell viability for A431 and H125 cells (Supplementary Figure S1A), produced a significant increase in HLA-ABC membrane expression (Figure 1B). A similar HLA-I enhancement was achieved in cetuximab and AG1478-treated A431 and H125 tumor cells when compared with untreated cells. Notably, in EGFR inhibitor-treated cells, we found an increase index for HLA class I expression comparable to that produced by IFN-γ treatment. As expected, we did not detect increase in HLA class I expression in EGFR-inhibitor-treated U1906 cells (Figure 1B). In experiments with A431 and H125 cells we observed a dose-dependent effect of nimotuzumab on upregulation of HLA-ABC (Supplementary Figure S1A) with a maximum effect at 96 h of incubation (Supplementary Figure S1B). When we evaluated the effect of nimotuzumab on HLA-I locus-specific expression, an increase of membrane levels of HLA-A and HLA-B was verified in both EGFR-positive tumors when compared to untreated cells (Figure 1C). For A431 cells a similar increase was detected for both alleles, whereas for H125 cells upregulation of HLA-A alleles was more pronounced than HLA-B alleles.

Next, we analyzed the effect of nimotuzumab on HLA class I molecules at transcriptional levels in comparison with cetuximab and AG1478. Similar to these EGFR inhibitors, a significant upregulation of mRNA levels of HLA-A, HLA-B, and HLA-C was seen in nimotuzumab-treated A431 and H125 cells when compared to untreated cells (Figure 2A). In contrast, a transcriptional induction of HLA class I HC genes was only detected in IFN-γ-treated U1906 cells (Figure 2A). Subsequently, to determine whether nimotuzumab-induced EGFR inhibition influences other molecules of HLA class I and APM machinery, we expanded our real-time quantitative PCR (qPCR) analyses to β₂-microglobulin (β₂-m), the complex for peptide generation/transport and the chaperones necessary for MHC-I assembly (Figure 2A). A significant increase in mRNA levels for β₂-m, LMP subunits, TAPs, as well as tapasin was demonstrated in nimotuzumab-treated A431 and H125 cells when compared to untreated cells. Similar effects were induced by cetuximab and AG1478 in the selected tumor cells. In agreement with our previous findings, a relevant reduction of mRNA levels for HLA class I HC, β₂-m and APM components was found in EGF-treated A431 and H125 cells when compared with cells cultured in the absence of this growth factor (Figure 2B). Addition of the EGF to U1906 cells did not reduce the levels of mRNA for HLA class I HC, β₂-m and APM components (Figure 2B).

Examples of anti-EGFR antibodies augmenting HLA class I expression in tumors mostly come from in vitro studies. In order to combine the preclinical evidence using immunocompetent mice, we used 7A7 a surrogate antibody that recognizes murine EGFR. In these experiments, the effect of AG1478 on MHC-I expression was also determined. We selected four EGFR-positive murine tumors from different histological origins (CT26, colon; 4T1 and F3II, mama; D122, lung) and included melanoma B16F10 cells as negative control for EGFR expression. Results from flow cytometry showed that F3II and D122 cells have similar membrane levels of EGFR which were ≈2-fold higher than in CT26 and 4T1 cells with comparable EGFR membrane levels (Supplementary Figure S2A). Using 0.5 nM of EGF, a concentration that induces an increase in cell viability for all EGFR-positive cells (Supplementary Figure S2B), a significant reduction in MHC-I expression levels was achieved for these tumors as compared to cells cultured without EGF (Figure 3A). It is noteworthy that MHC-I repression was slightly affected by EGFR expression levels. A ≈3.5-fold decrease of H-2K and H-2D expression was detected for F3II and D122 cells, whereas MHC-I HC reduction for CT26 and 4T1 cells was of ≈1.5-fold. EGF presence in culture medium did not affect MHC-I surface expression for B16F10 cells (Figure 3A). As expected, EGFR inhibition by 7A7, even in suboptimal cell viability reducing concentration (7A7: 1 μg/ml, Supplementary Figure S2C), had an opposite effect on MHC-I membrane expression in EGFR-positive tumors (Figure 3B). Treatment with 7A7 caused upregulation of MHC-I expression, effect which also depended on EGFR expression levels. In tumor cells with higher EGFR expression (F3II and D122), we observed a strong MHC-I upregulation after 7A7 treatment. In contrast, MHC-I induction mediated by 7A7 in CT26 and 4T1 cells with lower EGFR density, was less potent. We did not detect MHC-I increase in EGFRIs-treated B16F10 cells- in which only the IFN-γ treatment was able to induce MHC-I increase. For all EGFR-positive tumors, the greatest H-2D and H-2K induction by EGFR inhibitors treatment was detected at 96 h (Supplementary Figure S2D), and was dose dependent (Supplementary Figure S2C).

To determine whether the results obtained with human tumor cell lines could be reproduced in murine tumors, we conducted transcriptional studies using F3II and D122 tumor cells (mouse cells with higher EGFR expression). Similar to IFN-γ-cultured cells, a significant augment in mRNA levels for H-2K and H-2D was detected in 7A7 and AG1478-treated F3II and D122 cells when compared with untreated cells (Figure 3C). In contrast, a transcriptional induction of MHC-I HC genes was only detected in IFN-γ-treated B16F10 cells (Figure 3C). In addition, a significant increase in mRNA levels for β₂-m, LMP subunits, PA28α, BLH, TAPs, ERAAP1, tapasin, calnexin, calreticulin, PDI, as well as ERp57, was found in EGFRI-treated F3II and D122 cells when compared with untreated cells. Of note, EGFR blockade changed the expression of several genes also inducible by IFN-γ, but some genes were exclusively influenced by EGFRIs (BLH, calnexin, calreticulin, PDI, and ERp57) (Figure 3D). In agreement with our previous findings, a relevant reduction of mRNA levels for MHC-I HC, β₂-m and APM components was found in EGF-treated F3II and D122 cells when compared to cells cultured in absence of this growth factor (Supplementary Figures S3A,B). EGF addition to B16F10 cells did not decrease the levels of mRNA for MHC-I APM (Supplementary Figure S3C).

Next, we studied the capacity of 7A7 to alter the MHC-I expression in vivo. To conduct these experiments we selected D122 cells since they demonstrated a higher MHC-I induction after treatment with EGFRIs as compared to other EGFR-positive tumor cells (Figure 3B). Moreover, our previous experiments showed that 7A7 and AG1478, after 21 days of tumor inoculation, decrease both primary tumor as well as the appearance of experimental and spontaneous D122 metastases (Garrido et al., 2007, 2011a). D122 tumor-bearing mice were treated with 7A7 or AG1478 in two modalities as described in Figure 4A. Consistent with our previous studies, administration of 7A7 and AG1478 significantly reduced tumor volume when compared to control mice (Figure 4B). As expected, animals receiving intratumoral injections of anti-EGFR agents achieved a higher antitumor effect than those injected systemically. MHC-I surface expression was studied by flow cytometry of cells isolated from primary tumors. Intratumoral inoculation of 7A7 and AG1478 promoted a strong induction of H-2K^b and H-2D^b expression when compared with control mice (Figure 4C). Notably, this effect was also achieved in systemic studies (Figure 4D). The effect of 7A7 antibody administration on MHC cell surface tumor expression was even higher than in case of AG1478 administration.

We also investigated whether MHC-I increase induced by 7A7 treatment in D122 cells leads to enhanced antitumor immunity. To this end, the ability of D122-reactive CD8⁺ T cells to secrete IFN-γ in response to tumor recognition was studied by co-incubation of these T cells with D122 pretreated or not with 7A7 or AG1478 (Figure 5A). As shown in Figure 5B, D122 treatment with EGFRIs increased significantly the IFN-γ production by D122-specific CD8⁺ T cells when compared to untreated cells. Remarkably, the concentrations of IFN-γ detected in supernatants from CD8⁺ T cells: EGFRI-treated D122 co-cultures were similar to those obtained in CD8⁺ T cells: IFN-γ-treated D122 co-cultures. B16F10 cells (that expressed matching MHC-I but did not present appropriate antigens) failed to stimulate D122-reactive CD8⁺ T cells to release IFN-γ (Figure 5B). Consistent with a predominant role for H-2K^b molecules in the T cell-based cytotoxic response to D122 tumor (Eisenbach et al., 1984; Mandelboim et al., 1992; Plaksin et al., 1992), the capacity of 7A7 and AG1478-treated D122 to stimulate D122-reactive CD8⁺ T cells was strongly blocked by the addition of an anti-H-2K^b antibody, whereas the presence in the culture of an anti-H-2D^b antibody did not reduce the tumor recognition (Figure 5B). Finally, cytotoxic activity of D122-specific CD8⁺ T cells co-cultured with EGFRIs-treated D122 cells was examined (Figure 5A). Our results demonstrated that 7A7 and AG1478 treatment of D122 cells enhance their susceptibility to be lysed by CD8⁺ T cells (Figure 5C). T cells eliminated a high percentage of EGFRIs-treated D122 cells (∼48% using an E:T ratio of 50:1). This response was not present when B16F10 cells were used as target cells (Figure 5C).