Archives

  • 2018-07
  • 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • br Cancer cells that lack components of the

    2020-03-24


    Cancer cells that lack components of the mismatch repair (MMR) pathway, such as MLH1 and MSH2, are resistant to cisplatin in vitro [3,4]. Ovarian tumors with low MLH1 expression associate with poorer survival in platinum-treated patients [5–7]. Similarly, a study of 17 ovarian cancer patients treated with cisplatin-based chemotherapy found that patients with a poor response had significantly lower MSH2 protein levels [8]. Interestingly, low MMR protein expression does not always correlate with resistance to platinum-based therapies. In contrast to ovarian cancer, non–small cell lung cancer (NSCLC) patients with low MSH2 had improved survival when treated with platinum [9,10]. While the MMR pathway has not been associated with platinum resistance in 3X FLAG Peptide cancer, a subset of bladder cancers have reduced or absent protein expression of MSH2 as determined by immunohistochemistry (IHC) [11–15].
    Here, we take an unbiased approach to investigate mediators of cisplatin resistance by performing, to the best of our knowledge, the first genome-wide CRISPR-Cas9 cisplatin resistance screen in bladder cancer cells. Our screen results show that cells with loss of MSH2 are more resistant to cisplatin. We validated this finding by showing that bladder cancer cell lines with knockdown of MSH2 had a reduction in cisplatin-mediated apoptosis. Consistent with our in vitro results, we found that MIBC tumors with low levels of MSH2 protein expression had poorer survival when treated with platinum-based chemotherapy compared with those with higher MSH2 expression. MSH2 levels did not associate with survival in patients who did not receive a pharmacologic or radiation treatment, suggesting that the association with survival is specific to platinum treatment.
    2. Patients and methods
    2.1. Cell culture, shRNA knockdown, and drug treatments
    MGHU4 and 253J bladder cancer cell lines are representative of alterations found in MIBC [16,17]. Cells were cultured in Minimum Essential Medium media (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (MilliporeSigma, Burlington, MA, USA). PLKO.1 TRC shRNAs were transduced using psPAX2 (Addgene plasmid #12260) and pMD2.G (Addgene plasmid #12259; see Supplementary Table 1 for shRNAs) [18]. Transduced cells were selected with 2–5 mg/ml puromycin (Milli-poreSigma). Cisplatin, gemcitabine (MilliporeSigma), methotrexate, 
    doxorubicin, vinblastine, and oxaliplatin (ApexBio, Houston, TX, USA) were solubilized in water or dimethyl sulfoxide. For dose response experiments, cell viability was measured using the CellTiter-Glo lumines-cent assay (Promega, Madison, WI, USA).
    2.2. Performing the CRISPR resistance screen
    To generate sgRNA lentivirus, 12 mg of the human GeCKO (Genome-Scale CRISPR Knock-Out) lentiviral A library was transfected into 30 million HEK293 T cells for 24 h [19,20]. Viral supernatant was harvested, and MGHU4 cells were transduced at a calculated multiplicity of infection of 0.3 followed by selection with 2 mg/ml puromycin. Two million MGHU4 cells were plated in quadruplicate for each condition. Cells were treated with 3 mM cisplatin or vehicle for 30 h. Treatment media were removed and cells were allowed to grow to confluency prior to harvesting genomic DNA (Machery-Nagel, Bethlehem, PA, USA).
    2.3. Sequencing of sgRNA
    The sgRNA sequences of each sample were polymerase chain reaction (PCR) amplified from 4 mg of genomic DNA using primers containing adaptor and barcoding sequences. DNA fragments were size selected using agarose gel and sequenced using a 1 125 bp run on the HiSeq2500 (Illumina Inc., San Diego, CA, USA). Reads generated from each sample were aligned to the indexed sgRNA sequences with the Bowtie2 sequence aligner using the “very-sensitive-local” option [21]. The counts of sgRNA were summarized using htseq [22]; sgRNA counts across all samples were compiled, and differential sgRNA abundance was calculated using DESeq2 [23]. To map the sgRNA results to the gene level (approximately three gRNAs per gene), we calculated the mean fold change and combined individual sgRNA p values from DESeq2 using Fisher's method followed by multiple hypothesis testing correction using the Benjamini-Hochberg procedure [24].