Engagement of immune effector cells by trastuzumab induces HER2/ERBB2 downregulation in cancer cells through STAT1 activation

Cancer cell lines BT474, SKBr-3, and SKOV-3 were obtained from the American Tissue Culture Collection (ATCC, Manassas, VA, USA). MCF-7/HER2 is a stable MCF-7 cell line overexpressing HER2 [26]. RPMI1640 media and fetal bovine serum (FBS) were from Invitrogen (Carlsbad, CA, USA) and penicillin streptomycin was from Sigma-Aldrich (St. Louis, MO, USA). Reagent antibodies used for assays were purchased from commercial sources as indicated. Trastuzumab was purchased from a specialty pharmacy and single hinge cleaved trastuzumab (scIgG-T) and isotype control monoclonal antibody (human immunoglobulin G1 (IgG1)) were prepared as previously described [27]. Small molecule inhibitors, MG132, chloroquine, and fludarabine were from Sigma-Aldrich.

Cancer cells (1 × 10⁵) were pre-seeded in a 24-well plate (Corning, Tewksbury, MA, USA) overnight in RPMI1640 media supplemented with 10% FBS at 37°C. Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood of healthy donors (Gulf Coast Blood Center, Houston, TX, USA) using lymphocyte separation media (Cellgro, Manassas, VA, USA). The isolated PBMCs (effector cells, E) were added into cancer cells (target cells, T) at a ratio of E:T = 10:1 and cultured at 37°C, 5% CO₂ as indicated for all treatment conditions. Conditioned media (CM) were collected after 48 h co-culture and used for study with CMs or analysis of cytokines and chemokines. PBMCs were removed from cancer cells and washed with phosphate-buffered saline (PBS) twice. Cancer cells were lysed with RIPA buffer containing proteinase inhibitor cocktails (Calbiochem, San Diego, CA, USA) for Western blotting (WB). For the Fc gamma receptor (FcγR) blocking study, PBMCs were pretreated with human isotype IgG (10 μg/ml) for 1 h and the FcγR-blocked PBMCs were then added to cancer cells in the presence or absence of trastuzumab (5 μg/ml) as described above.

Cancer cells were cultured in the lower chamber in the presence or absence of trastuzumab and PBMCs were added to the upper chamber using the 0.4 μm transwell insert (Corning). After 48 h co-culturing, PBMCs were removed from top chamber and cancer cells were prepared for WB analysis.

Cancer cell lysates or tumor tissue lysates were subjected to SDS-PAGE separation on 10% gels (Bio-Rad Laboratories, Hercules, CA, USA) and proteins were transferred to a nitrocellulose membrane and immunoblotted with primary antibodies for anti-HER2 (Epitomics, Burlingame, CA, USA), anti-signal transducer and activator of transcription 1 (STAT1), anti-phosphorylated pSTAT1 (Cell Signaling Technology, Danvers, MA, USA), and anti-β-actin antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and detected with horseradish peroxidase (HRP)-conjugated secondary antibody. All WB images were captured and quantified using a FluorChem M imager (Cell BioSciences, Santa Clara, CA, USA) after adding HyGLO™ Quick Spray Chemiluminescent HRP substrate (Denville Scientific Inc., South Plainfield, NJ, USA).

Cancer cells were detached from cell culture plates using a nonenzymatic solution with EDTA (Invitrogen) and stained with FITC-anti-HER2 antibody (BD Biosciences, San Jose, CA, USA). Briefly, 5 × 10⁵ cells were dispensed in 100 μl aliquots and stained with FITC-anti-HER2 antibodies for 30 min at 4°C. After washing with PBS buffer, cells were analyzed for mean fluorescence intensity (MFI) using a guava easyCyte HT instrument based on the manufacturer’s instructions (Millipore, Danvers, MA, USA). Data were analyzed by the FlowJo software (Tree Star Inc., Ashland, OR, USA) and light scatter characteristics were used to gate tumor cells for the analysis.

Total RNA was extracted using Trizol. Two micrograms of total RNA was transcribed into cDNA using the SuperScript™ III First-Strand Synthesis System (Invitrogen). qPCR was performed on a CFX96 Touch™ real-time PCR detection system (Bio-Rad Laboratories) using the SYBR green method based on the manufacturer’s manual. Gene-specific primers are listed in Table S1 in Additional file 1. Gene expression was normalized to GAPDH and the relative gene expression was calculated using the 2^-ΔΔCt method [28].

Human PBMCs were stained with PerCP-Cy5.5-conjugated CD56, Alexa Fluor 700-conjugated CD14, and APC-Cy7-conjugated CD3 (BD Biosciences). The stained cells were then sorted by the BD FACSAria™ II flow cytometer. NK cells were collected as CD3⁻ CD14⁻ CD56⁺ cells, monocytes as CD3⁻ CD56⁻ CD14⁺, and T cells as CD3⁺ CD14⁻ CD56⁻. Similar to the co-culture study with PBMCs, the sorted immune effector cells were co-cultured with SKOV-3 cancer cells at E:T = 5:1 in the presence or absence of trastuzumab (5 μg/ml) for 48 h. Immune cells were removed and cancer cells were lysed for detecting HER2 expression by WB.

Mouse tumor xenograft studies were carried out in accordance with the animal care and use guidelines, and the protocol was approved by the Animal Welfare Committee of the University of Texas Health Science Center at Houston (HSC-AWC-10-128). HER2 overexpressing BT474 and MCF7/HER2 breast cancer cells were implanted into immunodeficient nu/nu mice obtained from Charles River Laboratories (Wilmington, MA, USA) and treated with trastuzumab as previously described [25]. Tumor size was measured using a Vernier scale caliper and tumor growth inhibition was calculated as ‘(1 - mean tumor volume of the test group/mean tumor volume of the control group) × 100’. Fresh tumor tissues were collected by snap freeze in liquid nitrogen for WB analysis and 10% formaldehyde-fixed tumor tissues were used for HER2 detection by immunohistochemistry (IHC).

HER2 levels in xenograft tumor tissues after trastuzumab treatment were detected using a primary anti-HER2 antibody from Epitomics. An HRP polymer system (Dako, Glostrup, Denmark) was used for tissue staining according to the manufacturer’s procedures.

Interferon gamma (IFN-γ) expression in the cell supernatants was detected by an enzyme-linked immunosorbent assay (ELISA) and both capture and detection antibodies were from R&D Systems (Minneapolis, MN, USA). Briefly, the ELISA plate was coated overnight with anti-IFN-γ capture antibody (1 μg/ml). After blocking with 5% bovine serum albumin (BSA)-PBS, diluted standard IFN-γ (ProSpec, East Brunswick, NJ, USA) and the cell lysates were added and incubated overnight, followed by a secondary anti-IFN-γ biotinylated capture antibody (0.5 μg/ml) for 2 h. After washing, the biotinylated antibody was detected by streptavidin-labeled HRP for 30 min. After washing, TMB was used as substrate to develop color and plates were read at 450 nm on a SpectraMax M4 microplate reader (Molecular Devices, Sunnyvale, CA, USA). IFN-γ concentration in the samples was derived using a standard curve.

Where appropriate, statistical analysis was performed using Student’s t test. A P value <0.05 between treatment groups is considered significantly different. Experiments were repeated at least three times.

We previously observed that HER2 level in high HER2-expressing BT474 breast cancer cells was not affected by trastuzumab treatment in vitro, but HER2 was downregulated in BT474 mouse xenograft tumors treated with trastuzumab [25]. To test whether immune cells play a role in HER2 downregulation in response to trastuzumab treatment, four high HER2-expressing cancer cell lines were treated with trastuzumab in the presence or absence of human PBMCs. HER2 levels in cancer cells were detected by WB after 48 h of co-culture treatment. Treatments of cancer cells with trastuzumab or PBMCs alone had no effects on HER2 levels in cancer cells (Figure 1A). In contrast, HER2 levels in cancer cells co-cultured with PBMCs in the presence of trastuzumab were significantly reduced in all four high HER2-expressing cancer cell lines in comparison with cancer cell controls or with trastuzumab or PBMCs single treatment (Figure 1A). To test if the low HER2 levels detected in the whole cell lysates by WB in cancer cells treated with PBMCs and trastuzumab reflects HER2 downregulation on the cell surface, we measured surface HER2 levels by a flow cytometer. Consistent with HER2 levels detected in the whole cell lysates by WB, HER2 levels determined by flow cytometry were lower on the cancer cell surface with co-treatment of trastuzumab and PBMCs than that on control cells as indicated by the decreased MFI in the flow histograms (Figure 1B). To understand if HER2 downregulation mediated by trastuzumab in the presence of PBMCs is a result of reduced HER2 gene transcription, we determined HER2 mRNA levels by qPCR after different time points of co-treatment of cancer cells with human PBMCs and trastuzumab. As shown in Figure 1C, HER2 mRNA level in BT474 cells was significantly reduced when co-cultured with PBMCs in the presence of trastuzumab, while treatment of cancer cells with trastuzumab or PBMCs alone had no effects on HER2 mRNA levels. The decrease of HER2 mRNA level in BT474 cells treated with PBMCs and trastuzumab was detectable after 8 h, and the reductions reached 50% and 75% after 24 h and 48 h of treatment, respectively (Figure 1C). To study if protein stability played a role in the PBMCs and trastuzumab-mediated HER2 downregulation, we measured HER2 levels in cancer cells treated with PBMCs and trastuzumab in the presence of a proteasome inhibitor (MG132) or a lysosome inhibitor (chloroquine). As shown in Figure 1D, inhibitors of the protein degradation pathways (MG132 or chloroquine) had no effects on the HER2 downregulation mediated by PBMCs and trastuzumab co-treatment. Taken together, these results demonstrated that co-treatment of high HER2 cancer cells with PBMCs and trastuzumab led to HER2 downregulation at the transcriptional level.

To test whether the interaction between trastuzumab Fc and FcγRs on immune cells is required for HER2 downregulation in cancer cells, we used three variants of trastuzumab with compromised or no Fc functions [25, 29, 30]: the scIgG-T variant has a single proteolytic cleavage at the hinge region of trastuzumab; the N297A-T has one amino acid mutation at the position 297 (from N to A, European numbering) and lacks N-glycosylation and Fc function; and the F(ab’)₂-T was generated by cleavage of trastuzumab Fc with the protease pepsin. Unlike the cancer cells treated with trastuzumab and PBMCs, cancer cells treated with the scIgG-T, N297A-T, or F(ab’)₂-T in the presence of PBMCs showed no HER2 downregulation (Figure 2A). Since immune cell subtypes have different expression profiles of FcγRs (Figure S1 in Additional file 1), we isolated NK cells, monocytes, and T cells (no detectable FcγRs) from PBMCs by a flow cytometer with a cell sorter and HER2 downregulation mediated by the co-treatment of trastuzumab and different immune cell subtypes were evaluated. Similar to the cancer cells treated with PBMCs and trastuzumab, cancer cells showed HER2 downregulation after co-treatment with trastuzumab and NK cells or monocytes, but cancer cells treated with T cells and trastuzumab did not show HER2 downregulation (Figure 2B). Furthermore, we blocked trastuzumab Fc binding with FcγRs on immune cells by preincubating PBMCs with human isotype IgGs before co-culturing with cancer cells. The preblocking of FcγRs on PBMCs with isotype IgGs abolished the HER2 downregulation mediated by PBMCs in the presence of trastuzumab (Figure 2C). These data suggest that engagement of FcγRs on immune cells by trastuzumab Fc is required for HER2 downregulation in cancer cells.

To further test whether direct engagement between cancer cells and immune cells is necessary for trastuzumab to trigger HER2 downregulation, we employed a transwell co-culturing system that retained immune cells in the upper chamber without direct contact with cancer cells at the bottom of the culture system. As shown in Figure 2D, cancer cells in the transwell co-culturing condition showed no HER2 downregulation, suggesting that direct engagement of cancer cells and immune cells is important for HER2 downregulation mediated by PBMCs and trastuzumab. Collectively, these results indicate that HER2 downregulation in cancer cells requires direct interaction of the trastuzumab Fab region with the HER2 antigen on cancer cells and the Fc region with FcγRs on immune cells, which brings cancer cells and immune cells in proximity for activating the immune effector function.

Activated immune cells secrete an array of soluble effectors. To determine if cancer cell HER2 downregulation is a function of soluble effectors secreted from the activated immune cells in the presence of trastuzumab, we tested HER2 downregulation in cancer cells after culturing with the CM collected from co-cultures of cancer cells and PBMCs in the presence or absence of trastuzumab. Cancer cells were cultured in 50% (1:1 diluted with fresh media) CM that were collected from four culturing conditions: cancer cell only, cancer cells treated with trastuzumab alone, cancer cells plus PBMCs, and cancer cells plus PBMCs and trastuzumab. Similar to the results shown in Figure 1A using the cells from co-culturing conditions, the CM from the co-culture of cancer cells and PBMCs in the presence of trastuzumab also showed significant HER2 downregulation (Figure 3A, B). HER2 mRNA levels were significantly reduced in both BT474 and SKOV-3 high HER2 cancer cell lines when cultured in the CM collected from the co-cultures of the cancer cells and PBMCs plus trastuzumab, but not in CMs collected from the other conditions (Figure 3C, D). These results suggest that HER2 downregulation in cancer cells is induced by certain soluble effectors released in the CM during the co-culture of cancer cells and PBMCs in the presence of trastuzumab.

To investigate the soluble effectors in the CM that resulted in the HER2 downregulation in cancer cells, we profiled cytokines and chemokines in the CMs using a cytokine array. Levels of cytokines and chemokines in the CM from the co-culture of BT474 cells with PBMCs in the presence of trastuzumab were compared to that in the CM from the co-culture of BT474 cells with PBMCs only. A panel of cytokines and chemokines including IFN-γ showed higher levels in the CM from the co-culture of BT474 cells with PBMCs in the presence of trastuzumab than that in the absence of trastuzumab (Figure S2 in Additional file 1). Since IFN-γ has been shown to exhibit inhibitory effects on HER2 expression in cancer cells in vitro[31], we investigated whether the increased IFN-γ in the CM from the co-treatment of cancer cells with PBMCs and trastuzumab has a role in the cancer cell HER2 downregulation. IFN- γ production under the different treatment conditions was quantified by an ELISA. As shown in Figure 4A, the co-culture of BT474 cells with PBMCs in the presence of trastuzumab produced significantly more IFN-γ than the other culture conditions. Culture media from BT474 cells only or BT474 cells with trastuzumab had no detectable IFN-γ.

To determine if the increased IFN-γ expression was a result of activation of PBMCs after their engagement with cancer cells through trastuzumab Fc, we first inactivated cancer cells with 4% paraformaldehyde (PFA); the inactivated BT474 cells were then co-cultured with PBMCs in the presence or absence of trastuzumab for 24 h; and IFN-γ mRNA levels in live PBMCs were determined by qPCR. As shown in Figure 4B, the PBMCs in the presence of both trastuzumab and the inactivated cancer cells (trastuzumab-coated cancer cells) had significantly higher IFN-γ mRNA than that in the other three conditions: PBMCs only, PBMCs with trastuzumab alone, or PBMCs with the inactivated cancer cell alone (Figure 4B). The results indicate that PBMCs secrete more IFN-γ upon activation in the presence of trastuzumab-coated cancer cells. To determine if the elevated IFN-γ contributes to HER2 downregulation in cancer cells, we used an IFN-γ-neutralizing antibody to block IFN-γ function. As shown in Figure 4C, HER2 downregulation was rescued by the IFN-γ-neutralizing antibody, while addition of exogenous IFN-γ in the cancer cell cultures induced HER2 downregulation at both the protein level and the mRNA level (Figure 4D).

Among the multiple signaling pathways regulated by IFN-γ, signal transducer and activator of transcription 1 (STAT1) has been reported to be associated with HER2 downregulation [31]. We measured STAT1 expression and activation in BT474 cells cultured in the CM collected from the co-culture of BT474 and PBMCs in the presence of trastuzumab. Total STAT1 protein and pSTAT1 (Tyr701) were significantly increased in BT474 cells treated with the CM collected from co-treatment with PBMCs and trastuzumab, when compared with the cells cultured with CMs from the other treatments (Figure 4E). Addition of fludarabine, an inhibitor of STAT1, reduced levels of both pSTAT1 and total STAT1 and blocked HER2 downregulation in cancer cells cultured with the CM from the co-treatment with PBMCs and trastuzumab (Figure 4F). Collectively, these results suggest that elevated IFN-γ expression by activated PBMCs and STAT1 activation in cancer cells play an important role in HER2 downregulation in cancer cells by the co-treatment of PBMCs and trastuzumab.

As expected, HER2 downregulation in cancer cells by trastuzumab engagement of immune cells was correlated with the inhibition of cancer cell proliferation (Figure S3 in Additional file 1). To understand whether active engagement of immune cells in vivo by trastuzumab can also mediate HER2 downregulation, we used a mouse xenograft tumor model that we showed immune cell engagement by trastuzumab previously [25]. Tumor-bearing mice were treated with trastuzumab or scIgG-T (with a compromised Fc) weekly at 5 mg/kg for four weeks and an isotype IgG was used as nontreatment control. IHC detection of HER2 level in the BT474 tumor tissue treated with trastuzumab showed significant downregulation in comparison with that in tumor tissue treated with isotype control antibody (Figure 5A). In the Western blotting studies, HER2 levels in the residue tumor tissues after trastuzumab treatment were barely detectable by WB, while HER2 levels remained high in tumors treated with scIgG-T or isotype control antibodies in both BT474 and MCF7/HER2 tumor models (Figure 5B). As scIgG-T has less capability of FcγRs engagement in comparison with trastuzumab, the reduced HER2 downregulation by scIgG-T supports the notion that HER2 downregulation in cancer cells requires active engagement of immune cells by the antibody Fc. Treatment with scIgG-T antibody also showed reduced antitumor efficacy than trastuzumab in both BT474 and MCF7/HER2 xenograft tumor models (Figure 5C). The correlation between less HER2 downregulation and decreased anticancer efficacy by scIgG-T treatment in vivo suggests that HER2 downregulation directly contributes to the antitumor efficacy of trastuzumab.