Therapeutic targeting of HER2-CB2R heteromers in HER2-positive breast cancer

Abstract

Although human epidermal growth factor receptor 2 (HER2)-targeted therapies have dramatically improved the clinical outcome of HER2-positive breast cancer patients, innate and acquired resistance remains an important clinical challenge. New therapeutic approaches and diagnostic tools for identification, stratification, and treatment of patients at higher risk of resistance and recurrence are therefore warranted. Here, we unveil a mechanism controlling the oncogenic activity of HER2: heteromerization with the cannabinoid receptor CB₂R. We show that HER2 physically interacts with CB₂R in breast cancer cells, and that the expression of these heteromers correlates with poor patient prognosis. The cannabinoid Δ⁹-tetrahydrocannabinol (THC) disrupts HER2-CB₂R complexes by selectively binding to CB₂R, which leads to (i) the inactivation of HER2 through disruption of HER2-HER2 homodimers, and (ii) the subsequent degradation of HER2 by the proteasome via the E3 ligase c-CBL. This in turn triggers antitumor responses in vitro and in vivo. Selective targeting of CB₂R transmembrane region 5 mimicked THC effects. Together, these findings define HER2-CB₂R heteromers as new potential targets for antitumor therapies and biomarkers with prognostic value in HER2-positive breast cancer.

Keywords: CB2R; HER2; breast cancer; cannabinoids; receptor heteromers.

Fig. 1.

HER2–CB₂R heteromer expression correlates with poor patient prognosis. Proximity ligation assays were performed in tissue microarrays and patient-derived xenografts. (Scale bars, 25 µm.) For the TMAs, samples were ranked based on HER2–CB₂R heteromer expression (i.e., PLA signal), and the best cutoff was manually selected. (A) Representative confocal microscopy images of a low– and a high–heteromer-expressing sample in TMA 1. The red dotted signal corresponds to the heteromers, and the blue staining corresponds to cell nuclei. (B–D) Kaplan–Meier curves for disease-free survival [from samples included in TMA 1 (n = 57) (B) or TMA 2 (n = 39) (C)] and overall patient survival [from the HER2+ samples included in TMA 2 (n = 33) (D)]. Curves were statistically compared by the log-rank test (P < 0.05). (E and F, Upper) Representative images of HER2–CB₂R heteromer expression in two pairs of PDXs, consisting of a PDX established from the patient’s primary tumor and a sample derived from a metastasis in the same patient [in the liver in one case (E), and in a lymph node in the other (F)]. (E and F, Lower) Quantification of HER2–CB₂R heteromer expression in the PDX samples. Results are expressed as PLA ratio (number of red dots per cell), and error bars represent SEM (n = 7 technical replicates in primary tumor samples; n = 5 in metastatic samples). HR, hazard ratio.

Fig. 2.

THC decreases HER2–CB₂R complexes. (A) Schematic representation of bioluminescence resonance energy transfer experiments. (B) BRET saturation curve in HEK293 cells transfected with a fixed concentration of HER2-Rluc and increasing concentrations of CB₂R-YFP. HER2-Rluc/GHS-R1a-YFP and D44R-Rluc/YFP were used as negative controls for the interaction (n = 8). (C) Effect of THC (4 h), alone or in combination with the CB₂R-selective antagonist SR144528 (SR2; 1 µM), on HER2-Rluc/CB₂R-YFP BRET_max signal in HEK293 cells (n = 3). (D and E) Viability of CB₂R- and HER2-transfected HEK293 cells after 24-h treatment with increasing concentrations of THC (n = 5) (D), or THC in combination with SR2 (1 μM) (n = 4) (E). (F and G) Viability of BT474 (n = 6) and HCC1954 (n = 3) cells in response to increasing concentrations of THC (F), or in combination with the CB₂R-selective antagonist SR144528 (1 μM) (G). Results (n = 3 to 6 independent experiments) are expressed as percent vs. vehicle-treated cells, set at 100%, and error bars represent SEM. (H) Coimmunoprecipitation of HER2 with CB₂R after THC treatment (4 h), in BT474 and HCC1954 cells transfected with an HA-tagged CB₂R plasmid. IP, immunoprecipitation. (I) Representative PLA confocal microscopy images of HER2–CB₂R heteromers (in red) in BT474 (Upper) and HCC1954 cells (Lower), treated with THC (4 h) alone or in combination with SR2 (1 μM). Cell nuclei are stained in blue. (Scale bars, 25 µm.) (J) Quantification of HER2–CB₂R PLA signal (number of red dots per cell) (n = 3). Results are expressed as percent vs. vehicle-treated cells, set at 100%, and error bars represent SEM. Multigroup comparisons were analyzed by one-way ANOVA with Tukey’s post hoc test. *P < 0.05, **P < 0.01 vs. vehicle-treated cells; ^# P < 0.05, ^## P < 0.01 vs. THC.

Fig. 3.

HER2–CB₂R heteromer disruption by THC hampers HER2 activation. (A) HER1, HER2, HER3, and HER4 expression, as determined by Western blot analysis, in the indicated breast cancer cell lines. (B) Representative PLA confocal microscopy images of the effect of THC (4 h) on HER2–HER1 (n = 4), HER2–HER2 (n = 5), and HER2–HER3 (n = 3) dimers (in red) in HCC1954 cells (B), with the corresponding quantification (C), or on HER2–HER2 expression after THC treatment, alone or in combination with the CB₂R-selective antagonist SR144528 (1 μM) (n = 3) (D), with the corresponding quantification (E). Cell nuclei are in blue. (Scale bars, 20 µm.) (F and G, Left) Schematic representation of the BRET experiments conducted in HEK293 cells. CoH, coelenterazine H. (F and G, Right) Quantification of HER2-Rluc/HER2-YFP BRET_max after THC treatment (4 h) alone or in combination with SR2 (1 µM) where indicated, in cells cotransfected with HER2-Rluc, HER2-YFP, and a CB₂R untagged receptor (n = 3) (F), or an empty vector (n = 4) (G) (used as a negative control for THC activation). In C and E–G, results are expressed as percent vs. vehicle-treated cells, set as 100%, and graph bars represent SEM. (H) Expression of pHER2¹²⁴⁸ in BT474 and HCC1954 cells, as determined by Western blot, upon THC treatment at the indicated times. (I) Quantification. Results are normalized vs. the corresponding total HER2 levels at each individual time point, and expressed as fold increase vs. time 0, set at 1 (n = 4 in BT474; n = 7 in HCC1954). Error bars represent SEM. Unpaired independent groups of two were analyzed by two-tailed Student’s t test. When multigroup comparison was required, data were analyzed by one-way ANOVA with Tukey’s post hoc test. *P < 0.05, **P < 0.01 vs. vehicle-treated cells; ^## P < 0.01 vs. THC. n.s., not significant.

Fig. 4.

HER2–CB₂R heteromer disruption by THC induces HER2 degradation in vitro and in vivo. (A, B, and D) Effect of THC on HER2 protein (A and B) and mRNA levels (D) at the indicated times, as determined by Western blot and qPCR, respectively, in BT474 and HCC1954 cells. For quantification, HER2 expression was normalized with the loading control [β-actin in B; β-actin and β-glucuronidase in D], and results (n = 4 in B; n = 3 in D) are expressed as fold increase vs. time 0, set at 1. Data were analyzed by one-way ANOVA. (C) Western blot analysis of the effect of the CB₂R-selective antagonist SR144528 (1 μM) on THC-induced HER2 protein decrease (n = 4 in BT474; n = 7 in HCC1954). (E) Growth of orthotopic tumors generated in NOD-SCID mice by injection of HCC1954 cells into the mammary fat pad. Animals were treated with vehicle (sesame oil) (n = 10) or THC (1.5 mg per dose) (n = 9) thrice a week. Results were analyzed by two-way ANOVA. (F) Representative Western blot of HER2 in the animal tumor samples. (G) Corresponding quantification. (H) Representative PLA confocal microscopy images of HER2–CB₂R and HER2–HER2 heteromers (red signal). Cell nuclei are in blue. (Scale bar, 50 µm.) (I) Quantification. Error bars in B, D, E, and I represent SEM. Unpaired, two-tailed Student’s t test. *P < 0.05, **P < 0.01 vs. time 0 (B) or vehicle-treated animals in E, G, and I.

Fig. 5.

HER2–CB₂R heteromer disruption by THC induces HER2 degradation via the c-CBL E3 ligase. Western blot-based analyses of the effect of different pharmacological and genetic tools on THC-induced HER2 degradation. (A and B) Effect of lactacystin (LAC; 1 µM) on BT474 cells (n = 4). (C–F) Effect of THC (4 h) on ubiquitinated HER2 (UB) (C), on c-CBL and CHIP levels (D and E), or on HER2 phosphorylation at Tyr1112 (F) in the indicated breast cancer cell lines. (G and H) HER2 protein expression after genetic silencing of c-CBL with selective siRNAs (siCBL). A nontargeted siRNA was used as a control (siC). The densitometric analyses of HER2 immunoblots were normalized to β-actin (n = 4 in B; n = 6 in E; n = 4 in H). Results are expressed as fold increase vs. vehicle-treated cells, set at 1, and graph bars represent SEM. Unpaired, independent groups of two were analyzed by two-tailed Student’s t test. When multigroup comparison was required, data were analyzed by one-way ANOVA with Tukey’s post hoc test. *P < 0.05, **P < 0.01 vs. vehicle-treated group; ^# P < 0.05, ^## P < 0.01 vs. THC-treated group.

Fig. 6.

CB₂R transmembrane domain 5 is involved in HER2–CB₂R heterodimerization. (A) Schematic representation of the HA-tagged truncated forms of CB₂R used in this study. Each construct contains the HA tag, followed by the N-terminal domain of the receptor, one of its seven transmembrane domains, and the C-terminal end. (B) Each of the seven CB₂R constructs (named HA-TMX, where X is the corresponding transmembrane domain) and a pcDNA3-HER2 plasmid was coexpressed in HEK293 cells. Immunoprecipitation of HER2 with an anti-HER2 antibody was followed by Western blot analysis with an anti-HA antibody. Full-length pcDNA3-HA-CB₂R was also coexpressed with HER2 as a positive control of interaction. (C) Schematic representation of the bimolecular fluorescence complementation experiments between HER2-cYFP and CB₂R-nYFP in the absence (Upper) or presence of the CB₂R transmembrane constructs (Lower). (D and E) Complementation signal (i.e., fluorescence at 530 nm) of HEK293 cells transfected with CB₂R-cYFP, HER2-nYFP, and the indicated CB₂R TM constructs (n = 3) (D), or after 4 h of incubation with the indicated TAT-TM peptides (4 µM) (n = 3) (E). Results were analyzed by one-way ANOVA with Tukey’s post hoc test. Error bars represent SEM. **P < 0.01 vs. pcDNA3 (D) or vehicle-treated group (E).

Fig. 7.

HER2–CB₂R heteromer disruption by targeting CB₂R TM5 mimics THC effects. (A–D) Effect of TM peptides on HER2–CB₂R and HER2–HER2 heteromer expression as determined by PLA. (A and C) Representative PLA images in the indicated breast cancer cell lines, after treatment for 4 h with vehicle (DMSO), a TAT-TM peptide targeting CB₂R TM5 (4 μM), or a TAT-TM peptide targeting dopamine receptor D44 TM5 (4 μM), used as a negative control. Dimer signal is in red, and cell nuclei are in blue. (Scale bars, 25 µm.) (B and D) Results (n = 7 technical replicates) are expressed as percent of PLA (red dots per cell) vs. vehicle-treated cells, set as 100%. (E) pHER2¹²⁴⁸ and HER2 protein levels, as determined by Western blot, after treatment with vehicle, CB₂R TAT-TM5, or D44R TAT-TM5 peptides for 24 h in BT474 and HCC1954 cells. (F) Densitometric analysis of HER2 normalized to β-actin (n = 3). Results are represented as fold increase vs. vehicle-treated cells, set as 1. (G) Viability of HCC1954, BT474, and HEK293 cells in response to the indicated treatments for 24 h. Data (n = 4) are represented as percent vs. vehicle-treated cells, set as 100%, and graph bars represent SEM. One-way ANOVA with Tukey’s post hoc test. **P < 0.01 vs. vehicle-treated cells.

Fig. 8.

Schematic drawing of the proposed mechanism of control of HER2 activity by CB₂R. (A) HER2 forms heteromers with CB₂R at the plasma membrane of HER2+ breast cancer cells, protecting it from degradation and favoring its prooncogenic signaling. (B) Disruption of HER2–CB₂R heteromers, either by THC or by specific tools targeting CB₂R transmembrane domain 5, triggers inactivation of HER2 by inducing the separation of HER2–HER2 homodimers and increasing HER2 susceptibility to degradation by the E3 ligase c-CBL. HER2 degradation and CB₂R activation result in antitumor responses.