Regulation of submaxillary gland androgen-regulated protein 3A via estrogen receptor 2 in radioresistant head and neck squamous cell carcinoma cells

Human cell lines (FaDu and Cal27) were purchased from ATCC (http://​www.​lgcstandards-atcc.​org/​) and were maintained in Dulbecco’s Modified Eagle’s Medium (Sigma-Aldrich, Germany) supplemented with 10% fetal bovine serum (Invitrogen, Germany), 2 mM L-Glutamine (Invitrogen, Germany) and 50 μg/ml Penicillin-Streptomycin (Invitrogen, Germany) in a humidified atmosphere of 6% CO₂ at 37 °C. Authentication of both cell lines was confirmed by the Multiplex Human Cell Line Authentication Test (Multiplexion, Germany, latest update August 11th, 2016). To generate single cell clones, FaDu cells were transfected with either pcDNA3.1 (Invitrogen, USA) or a plasmid encoding a Myc-DDK-tagged ORF of Homo sapiens SMR3A (Origene, USA) using FuGene HD Transfection Reagent (Promega, Germany) according to the manufacturer’s instruction. Following selection with 0.5–0.8 mg/ml G418 (Calbiochem Merck, Germany) for two weeks, single clones were isolated and ectopic SMR3A-Myc/DDK expression was confirmed on transcript and protein level. Two independent mock controls and FaDu-SMR3A clones were selected for further analysis. Parental FaDu and Cal27 cells were treated with the indicated concentrations of estradiol (E2) or 4-Hydroxytamoxifen (TAM) dissolved in ethanol (EtOH) or Fulvestrant dissolved in DMSO. All compounds were purchased from Sigma-Aldrich, Germany.

RNA extraction, cDNA synthesis and real-time quantitative polymerase chain reaction (RQ-PCR) were done as described previously [7]. Primer sequences were selected using Primer-Blast (http://​www.​ncbi.​nlm.​nih.​gov/​tools/​primer-blast) and are listed in Additional file 1. To quantify relative SMR3A transcript levels by RQ-PCR analysis cDNA samples were used undiluted. The cycle of threshold (CT) of the gene of interest was standardized to the CT value of the reference gene (LMNB1) using the ∆∆CT method.

Western blot analysis was performed with whole cell lysates as described previously [8]. Membranes were incubated in enhanced chemiluminiscence solution (Thermo Scientific, Germany) and developed with the ImageQuant LAS500 system (GE Healthcare, Germany). Immunofluorescence staining was done as described elsewhere [7], and pictures were taken with the Fluorescence Microscope BX50F, Olympus XC30 Camera and cellSens Entry imaging software (Olympus, Germany). Antibodies and dilutions that were used for Western blot analysis and immunofluorescence staining are listed in Additional file 2.

20–30,000 cells per well were seeded on glass coverslips in 12-well plates. 48 h after seeding, cells were treated for 2 h with BrdU according to the manufacture’s instruction (BD Pharmingen, Germany). Staining was done as described previously [7, 8] and relative amount of proliferative cells was calculated as the ratio of BrdU-positive versus total cell counts.

Cell migration assays were performed using ibidi Culture Inserts (Ibidi, Germany) according to the instruction of the manufacturer. After removal of the inserts, pictures were taken at the indicated time points using the Keyence BZ-9000 microscope system to calculate the relative gap closure over time.

To determine the clonal expansion of tumor cells upon fractionated irradiation, 100, 300 and 1,000 cells were seeded per well in 6-well plates and irradiated on four consecutive days with a daily dose of 2 gray (Gy) using X-RAD 320 (Precision X-Ray, North Branford, CT USA) or kept untreated as controls. Half of the cells were treated daily with 1 μM 4-Hydroxytamoxifen (TAM, Sigma-Adrich, Germany) or every second day with 30 nM Fulvestrant (Sigma-Adrich, Germany). After 10–14 days in culture, clones were stained with crystal violet and total amount of colonies was quantified as described in [9]. The survival fraction was computed according to [10].

The retrospective study cohort, generation of tissue microarrays, immunohistochemical staining and assessment of the immunoreactivity score for SMR3A was described before [6]. Paraffin-embedded tissue specimens were provided by the tissue bank of the National Center for Tumor Disease (Institute of Pathology, University Hospital Heidelberg) after approval by the local institutional review board (ethic votes: 176/2002 and 206/2005). The study was performed according to the ethical standards of the Declaration of Helsinki. For all tumor samples, clinical and follow-up data were available from the Department of Otolaryngology, Head and Neck Surgery at the University Hospital Heidelberg. Immunohistochemical staining for ESR2 was done using the 3, 3’-diaminobenzidine (DAB) peroxidase substrate kit (Perkin Elmer, Germany) according to the manufacturer’s instruction and the assessment of the immunoreactivity score for ESR2 was according to the procedure described before [6]. Clinical and histopathological features of samples for which informative values for SMR3A and ESR2 were available and which were included in the final analysis are listed in Additional file 3.

Statistical analysis was done using SPSS (version 21) and GraphPad software (http://​www.​graphpad.​com). Differences between patient subgroups were assessed using Chi square test. Disease-specific survival (DSS) was calculated as the time from the date of primary tumor diagnosis to the date of cancer-related death within the follow-up interval (events). Survival times of patients, who were alive or were dead due to causes other than cancer were censored. Progression-free survival (PFS) was calculated from the date of primary tumor diagnosis to the date of the first local recurrence, lymph node or distant metastasis, second primary carcinoma, or date of cancer-related death within the follow-up period (events). Patients without progression (no event) or cancer-unrelated death were censored. The method of Kaplan–Meier was used to estimate survival distributions and differences between subgroups were determined by log-rank tests. To adjust for possible confounders, multivariable Cox proportional hazard regression models were fitted. Models included the covariates gender, age, clinical staging, alcohol and tobacco consumption, HPV status and therapy. Based on availability of complete clinical data sets, 66 cases were included for the analysis of subgroups with ESR^highSMR3A^low versus ESR^highSMR3A^high staining patterns and 103 cases were included for the analysis of subgroups with ESR^highSMR3A^low versus all other staining patterns.

SMR3A expression was assessed by RQ-PCR analysis with cDNA from human HNSCC cell lines. Transcript levels were close to the detection limit in all cell lines tested, indicating no or a rather low SMR3A expression under normal growth conditions (data not shown). FaDu cells were selected to generate stable clones (FaDu-SMR3A) with ectopic SMR3A expression in order to address its impact on tumor-relevant processes in vitro. Ectopic expression in FaDu-SMR3A clones was confirmed on transcript and protein levels (Additional file 4a-c). However, we did not observe any significant difference in tumor cell proliferation or migration between FaDu-SMR3A clones and mock controls (Additional file 4d-e). In summary, these data suggested that SMR3A has no major impact on tumor cell physiology under normal growth conditions but raised the attractive question, whether it serves as a marker for a distinct subpopulation of tumor cells with higher resistance against well-established treatment options.

To support this assumption, FaDu cells were treated with fractionated irradiation (IR, 4× 2Gy), which revealed a prominent SMR3A staining in all vital cells after treatment as determined by IF analysis (Fig. 1a-b). It is worth noting that we observed a gradual increase in the relative amount of positive cells with increasing cycles of fractionated IR (Additional file 5a and data not shown). An induction after fractionated IR was also detected in Cal27 cells, though the staining was more heterogeneous and not all vital cells were SMR3A-positive (Fig. 1b). In both cell lines induced SMR3A expression after fractionated IR was confirmed on transcript level (Fig. 1c). However, the induction as compared to control-treated cells was highly significant for FaDu cells, while in Cal27 cells a clear trend was found without reaching statistical significance.

In the past, several studies in rodents provided experimental evidence for the regulation of opiorphin family members by steroid hormones [11‐13]. To assess whether hormone signaling also contributes to SMR3A regulation during fractionated IR, expression of the androgen receptor (AR) as well as estrogen receptor 1 and 2 (ESR1 and ESR2) was analyzed in FaDu and Cal27 cells on transcript and protein levels. No AR and ESR1 expression was detected under normal growth conditions and only few cells exhibited a minor staining for ESR1 after fractionated IR (Fig. 2a, Additional file 5b and data not shown). In contrast, ESR2 was expressed in both cell lines and was strongly induced after fractionated IR (Fig. 2a-c). Again, the amount of ESR2-positive cells was more heterogeneous in Cal27 as compared to FaDu cells and accumulated with increasing cycles of fractionated IR (Additional file 5a), suggesting regulation of SMR3A expression by ESR2-dependent signaling. In line with this assumption, stimulation of FaDu cells with estradiol (E2) revealed a significant and concentration dependent increase of SMR3A transcript levels (Fig. 2d), which was impaired by co-treatment with 4-Hydroxytamoxifen (TAM) or pre-treatment with Fulvestrant (Fig. 2e-f). In summary, these data suggested that SMR3A expression serves as a potential surrogate marker for active ESR2 signaling in a subpopulation of radioresistant tumor cells, raising the question, whether pharmacological interference sensitizes these cells to fractionated IR.

As a proof of concept, fractionated IR was applied to FaDu cells with or without administration of TAM or Fulvestrant, respectively (Additional file 6). Fulvestrant inhibited irradiation-induced ESR2 expression as determined by IF staining and Western blot analysis (Fig. 3a-b), and augmented apoptosis, which was monitored by elevated caspase three and PARP cleavage (Fig. 3c). An increase in apoptosis was also detected for administration of TAM in combination with fractionated IR (Fig. 3c). A higher radiosensitivity of FaDu cells by inhibition of ESR2 signaling was further confirmed in a colony-forming assay after fractionated IR alone or in combination with Fulvestrant or TAM treatment (Fig. 3d-e). In line with the less efficient induction of ESR2 and SMR3A, Cal27 exhibited a reduced relative survival fraction after fractionated irradiation as compared to FaDu cells, suggesting a positive correlation between ESR2 induction and radioresistance. Furthermore, in Cal27 cells only administration of TAM but not Fulvestrant revealed a significant decrease in the relative survival fraction upon fractionated IR.

So far, experimental data support a model in which ESR2 and SMR3A co-expression after fractionated IR is a characteristic feature for a subpopulation of treatment resistant tumor cells. In line with this model the detection of SMR3A expression in tumor cells prior to therapy might serve as a surrogate marker for active ESR2 signaling during malignant progression and the presence of tumor cells with intrinsic radioresistance. To address the clinical relevance of our in vitro findings, we assessed ESR2 expression by IHC staining on tissue microarrays containing tumor specimens of OPSCC patients, which were treated with either definitive or post-surgical radiotherapy with or without adjuvant chemotherapy. Data on SMR3A staining on serial sections were already available from a previous retrospective study [6]. Evaluable staining patterns for both proteins were obtained for n = 109 OPSCC patients (Fig. 4a), and clinical features of the study cohort are summarized in Additional file 3. Positive staining for ESR2 (ESR2^pos) in tumor cells was detected in 65.1% of tumors and a high staining pattern correlated significantly with a higher SMR3A immunoreactivity score as compared to samples without detectable ESR2 staining (Fig. 4b, p = 0.044). We did not observe a statistically significant correlation between ESR2^neg and ESR2^pos subgroups and patient characteristics tested, except for age, T status and tobacco consumption. Significant associations were due to an older age, smaller tumor size and an enrichment of never/former smokers in the subgroup with ESR2^pos staining (Additional file 7). Concerning progression-free (PFS) and disease-specific survival (DSS), univariate analysis revealed an unfavorable clinical outcome in the absence of ESR2 staining as compared to ESR2^pos tumors (Fig. 4c-d). However, patients with ESR2^pos tumors had a significantly shorter PFS and DSS in the presence of high SMR3A expression, similar to the ESR2^neg subgroup, while ESR2^posSMR3A^low tumors exhibited the most favorable clinical outcome (Fig. 4c-d, Additional file 8). Multivariable Cox proportional hazard regression models adjusted for the covariates gender, age, clinical staging, alcohol and tobacco consumption, HPV status and therapy supported a favorable PFS and DSS of the subgroup with an ESR2^highSMR3A^low staining as compared to either ESR2^highSMR3A^high or all other staining patterns (Additional file 9).