Antiestrogen Resistant Cell Lines Expressing Estrogen Receptor α Mutations Upregulate the Unfolded Protein Response and are Killed by BHPI

Outgrowth of metastases expressing ERα mutations Y537S and D538G is common after endocrine therapy for estrogen receptor α (ERα) positive breast cancer. The effect of replacing wild type ERα in breast cancer cells with these mutations was unclear. We used the CRISPR-Cas9 genome editing system and homology directed repair to isolate and characterize 14 T47D cell lines in which ERαY537S or ERαD538G replace one or both wild-type ERα genes. In 2-dimensional, and in quantitative anchorage-independent 3-dimensional cell culture, ERαY537S and ERαD538G cells exhibited estrogen-independent growth. A progestin further increased their already substantial proliferation in micromolar 4-hydroxytamoxifen and fulvestrant/ICI 182,780 (ICI). Our recently described ERα biomodulator, BHPI, which hyperactivates the unfolded protein response (UPR), completely blocked proliferation. In ERαY537S and ERαD538G cells, estrogen-ERα target genes were constitutively active and partially antiestrogen resistant. The UPR marker sp-XBP1 was constitutively activated in ERαY537S cells and further induced by progesterone in both cell lines. UPR-regulated genes associated with tamoxifen resistance, including the oncogenic chaperone BiP/GRP78, were upregulated. ICI displayed a greater than 2 fold reduction in its ability to induce ERαY537S and ERαD538G degradation. Progestins, UPR activation and perhaps reduced ICI-stimulated ERα degradation likely contribute to antiestrogen resistance seen in ERαY537S and ERαD538G cells.

into the vector pSpCas9(BB)-2A-Puro(PX459) (Addgene, MA) digested with BbsI, as described in the Zhang lab protocol. 46 Transfection and Selection. To deliver the plasmid expressing guide sequences, the Cas9 gene and homology-directed repair template to T47D cells, low toxicity Lipofectamine 3000 (Fisher, NH) was used following the manufacturer's instructions. Briefly, one day before transfection, 1.5 million cells were plated in a 100 mm plate. The next day, 5 µg of plasmid DNA for each guide sequence, and 10 µg of linearized HDR template plasmid containing either the ERαY537S or ERαD538G mutation was co-transfected into each plate. The ratio of DNA to Lipofectamine 3000 was 1:3. The cells were incubated with DNA-Lipofectamine 3000 complex for 24 hours, after which the medium was replaced with normal growth medium. To select transfected cells, after 24 hours for recovery, the cells were maintained in regular growth medium plus 2.5 µg/ml puromycin. Puromycin selection was carried out for 3 days, with the medium changed daily.
After selection, the cells were allowed to recover for 1 day in regular growth medium. The cells from each 100 mm plate were then harvested by trypsin-EDTA and split into 4X 100 mm plates in MEM supplemented with 10% CD-FBS and pen-strep, plus 30% conditioned medium.
Conditioned medium was collected from wild-type T47D cells; when the T47D cells reached ~30% confluence, standard medium supplemented with 10% CD-FBS was added. After 2 days, the medium was collected and filtered through a 0.2 µM filtration unit. The resulting conditioned medium can be stored up to 2 weeks at 4 °C. The medium was changed every 4 days until colonies were visible to the naked eye. The colonies were washed twice with 1X PBS, and were kept in 1X PBS for picking under a microscope with a p20 micropipette. The picked colonies were digested in 20 µl of trypsin in a 96-well plate and transferred to a 24-well plate. After a week of growth, half of the cells were split from the 24-well plate to a 12-well plate and another half to a T25 flask. The cells from the 12-well plate were used for genomic DNA preparation using the Blood & Tissue DNeasy kit (Qiagen, CA).
Genotyping. The primers for genomic genotyping are located just outside of the HDR template arms, their sequences are: 5'-CATGGCAAGTCTCCAACTTGAGCTG and 5'-AGTCAGCCTGAGTAGTGACAGCAAC. Genomic DNA from each individual clone was used as a template to amplify a fragment that covers the whole HDR template. To verify insertion of the restriction sites by HDR, the PCR product was digested with AclI or SpeI and bands were visualized by agarose gel electrophoresis followed by staining with ethidium bromide.
To confirm homologous recombination in these positive clones at the mRNA level, total mRNA was prepared. PCR following reverse transcription was carried out to generate an ERα cDNA fragment using a set of ERα-specific primers: 5'-AATCCTCACGCTTAGTAACATA as the RT primer, and 5'-ACAGGGAGCTGGTTCACATGATCAAC and 5'-GAGAGCTGTTACAAAGATTTAGCCTTGG primer set for PCR. The resulting PCR product was then digested with AclI or SpeI to verify HDR at the mRNA level. To confirm the presence of the mutations, the digested or undigested DNA bands were recovered from the agarose gels and sequenced.

Supplementary Figures
Supplementary Figure S1. CRISPR-Cas9 replacement of wild-type ER with ERY537S and  Table 1 summarizes data for all the clonal cell lines genotyped. Despite the low frequency of HDR for both the ERY537S and ERD538G mutations, almost half as many assayed cell lines had both copies of the wild-type ER replaced with the mutation as cell lines that had one copy replaced.
Although we did not observe a substantial growth advantage for the double replacement cell lines in standard cell culture, it remains possible that they exhibited a somewhat larger growth advantage during outgrowth of the clones. The reason for the small number of false positive cell lines in the ERD538G experiments is unknown. (b) Western blot analysis of ER expression.
To ensure detection of deletions and truncations in the C-terminal region of ER, we used an antibody recognizing epitopes in the N-terminal region of ER. The Western blot shows that the pattern of ER gene expression in the cell lines replacing one copy of the ER gene matches that predicted from the DNA sequencing. For example: In Supplementary Figure S1c, clones ERD538G-4 and ERD538G-5 show a deletion of 17 amino acids and a reading frame shift after the deletion. This results in a protein, containing 588 amino acids with almost 60 amino acids after a reading frame shift. This protein's size is very close to the size of wild-type ERα as seen by two bands in the western (595 aa). For clones D538G-6, Y537S-5 and Y537S-8, a single nt insertion causes a reading frame shift and premature termination, resulting in truncation of the protein. For these 3 clones, this insertion results in a protein with 537 total amino acids seen as a second, lower molecular weight band than wild-type ERα. For clones Y537S-6 and Y537S-7,