The Vitamin D Receptor-BIM Axis Overcomes Cisplatin Resistance in Head and Neck Cancer

Abstract

Treatment success of head and neck squamous cell carcinoma (HNSCC) is often hindered by cisplatin resistance. As inherent and acquired therapy resistance counteracts improvement in long-term survival, novel multi-targeting strategies triggering cancer cell apoptosis are urgently required. Here, we identify the vitamin D receptor (VDR) as being significantly overexpressed in tumors of HNSCC patients (n = 604; p = 0.0059), correlating with tumor differentiation (p = 0.0002), HPV status (p = 0.00026), and perineural invasion (p = 0.0087). The VDR, a member of the nuclear receptor superfamily, is activated by its ligand vitamin D (VitD) and analogs, triggering multiple cellular responses. As we found that the VDR was also upregulated in our cisplatin-resistant HNSCC models, we investigated its effect on overcoming cisplatin resistance. We discovered that VitD/cisplatin combinations synergistically killed even cisplatin-resistant cells at clinically achievable levels. Similar results were obtained for the clinically used VitD analog Maxacalcitol. Moreover, VitD/cisplatin combinations inhibited tumor cell migration by E-cadherin upregulation. Signaling pathway analyses revealed that VitD co-treatments triggered cancer cell death by increasing the expression of the pro-apoptotic BCL-2 family protein BIM. BIM's pro-apoptotic activity in HNSCC cells was confirmed by ectopic overexpression studies. Importantly, BIM expression is positively associated with HNSCC patients' (n = 539) prognosis, as high expression correlated with improved survival (p = 0.0111), improved therapy response (p = 0.0026), and remission (p = 0.004). Collectively, by identifying, for the first time, the VDR/BIM axis, we here provide a molecular rationale for the reported anti-cancer activity of VitD/analogs in combination therapies. Our data also suggest its exploitation as a potential strategy to overcome cisplatin resistance in HNSCC and other malignancies by inducing additional pro-apoptotic pathways.

Keywords: HPV; calcitriol; combination therapy; nuclear receptors; platinum-based drugs; pro-apoptotic pathways; therapy resistance.

Figure 1

Clinical relevance of VDR in HNSCC patients. VDR is significantly overexpressed in well-differentiated, invasive HNSCCs. Bioinformatic analysis of the TCGA HNSCC cohort (n = 604). Overexpression of VDR was found in (a) primary tumors versus normal tissue and correlates with (b) tumor differentiation, (c) HPV status, (d) perineural invasion, and (e) tumor localization. Significant p values and sample size (n) are indicated. *, p < 0.05; **, p < 0.01, ***, p < 0.005.

Figure 2

VDR overexpression contributes to cisplatin resistance. (a) Generation of VDR-overexpressing HNSCC cell models. Fadu cells were transfected with VDR C-terminally fused to GFP (green). Positive clones were selected using puromycin, and sorted by FACS into high, medium, and low subpopulations. Low-expressing subpopulations were used for experiments. (b) Fluorescence-microscopy-visualized VDR expression levels in stable HNSCC (Fadu) cell line. Scale bar, 10 µm. (c,d) Viability assay revealed that VDR-overexpressing Fadu cells (c) and HNSCCUM-02T cells (d) are more resistant to cisplatin treatment. Cells were treated with 20 µM cisplatin for 72 h, and viability was measured and normalized to untreated controls (Cont). **** p < 0.0001. (e) Overexpression of VDR in resistant Pica_CIS versus sensitive (Pica_WT) cells was demonstrated by immunoblot analysis. GAPDH served as loading control. (f,g) Immunofluorescent staining confirmed overexpression of endogenous VDR protein in cisplatin-resistant Pica cells (Pica_cis). VDR expression was automatically quantified using the high-content screening platform Array Scan VTI. **** p < 0.0001. (g) Fluorescence-microscopy-visualized VDR overexpression in Pica_cis cells. Cells were stained with specific fluorescent VDR Ab (green), and nuclei marked with Hoechst (blue). Scale bar, 10 µm.

Figure 3

VitD co-treatment allows HNSCC to overcome cisplatin resistance. (a) Scheme illustrating the VitD pre-treatment schedule, which mimics physiological high versus low VitD serum levels. Cells were seeded in the presence or absence of 100 nM VitD; after 24 h, cells were treated with 20 µM cisplatin with and without VitD. (b,c) Combinational treatment of VitD/cisplatin synergistically triggered cell death of not only the cisplatin-sensitive (b) but also the highly resistant Pica cis cells (c). Cells were treated as described in (a). The viability of the cells was measured and normalized to untreated controls. (d) Fluorescence microscopy was used to visualize DNA damage. Compared to the VitD/cisplatin combination treatment, even cisplatin-treated cells showed a lower number of DNA damage events (quantified as γH₂AX damage foci/per cell). Cells were treated for 24 h (10 µM VitD; 20 µM cisplatin), and γH₂AX foci were detected by specific antibodies. Scale bar, 10 µm. (e) DNA damage events were automatically quantified by high-throughput microscopy and normalized to untreated controls. ***, p < 0.005, **** p < 0.0001 (f) Immunoblot analysis confers significantly increased γH₂AX levels in VitD/cisplatin co-treated HNSCCUM-02T cells. Actin served as the loading control.

Figure 4

VitD reduces the migration ability of HNSCC cells. (a) Wound-healing assays were performed to evaluate the migration of HNSCCUM-02T cells. Cells were treated with VitD (100 nM) or cisplatin (10µM) alone, and with the combination. Wound size is shown for 0 h (upper panel), and after 72 h (lower panel). Scale bar, 100µm. (b) Relative wound size after 72 h was significantly increased after combination treatment of VitD/cisplatin. ***, p < 0.005, **** p < 0.0001. (c) Immunoblot analysis shows upregulation of VDR and E-cadherin in response to VitD, counteracting the cisplatin-induced downregulation of E-cadherin. Actin and GAPDH served as the loading control. Proteins were detected by specific Ab.

Figure 5

The VitD–BIM axis aids in overcoming cisplatin resistance in head and neck cancer. (a,b) Immunoblot analyses reveal significant increase in VDR and BIM in VitD and VitD/cisplatin co-treated HNSCCUM-02T cells. GAPDH served as the loading control. (c) Fluorescence microscopy was used to visualize cytoplasmic BIM-GFP (BIM_EL/extra-long isoform) expression in Pica cells. Cells were analyzed 16 h post-transfection. Scale bar, 10 µm. (d) Ectopic BIM-GFP expression triggers apoptosis of HNSCC cells. GFP was used as control. Cells were treated for 72 h (100 µM VitD; 20 µM cisplatin, and combination). Cell viability was normalized to untreated controls. (e) Survival analysis demonstrates that high BIM expression correlates with improved disease-specific survival of HNSCC patients. n = 539; p = 0.0111*. Hazard ratio (Mantel–Haenszel) = 1.528. BIM low < 9.458 (median), and BIM high ≥ 9.458. (f) High BIM expression significantly correlates with improved therapy success, as shown by the status after primary therapy (disease after curative treatment). p values and sample size (n) are indicated. *, p < 0.05; **, p < 0.01.

Figure 6

Model summarizing how the VitD/VDR/BIM axis overcomes cisplatin resistance in head and neck cancer. VitD treatment is able to sensitize HNSCC cells to cisplatin-induced cell death via induction of DNA damage and activation of pro-apoptotic pathways. After cellular uptake, VitD binds to VDR, which then translocates to the nucleus and binds to specific promoter sites, thereby inducing multiple pathobiological pathways. In HNSCC, the VitD/VDR complex also induces the expression of the pro-apoptotic protein BIM. BIM efficiently induces the activation of mitochondria-dependent apoptosis, and together with cisplatin-induced DNA-damage, efficiently triggers cancer cell death, including that of cisplatin-resistant cancer cells. Simplified schematic illustration of pathways and players, not drawn to scale.