Functional Genomic and Immune Response Characterization of PTEN Loss: Therapeutic Implications for Myeloma

PTEN (phosphatase and tensin homolog) is a known tumor suppressor gene (TSG) in solid tumors and hematologic malignancies, including multiple myeloma (MM), yet several aspects of its biology and clinical implications remain incompletely understood in MM. For instance, PTEN mutations/deletions are infrequent in newly diagnosed MM in the CoMMpass study (6 and 9 cases with heterozygous mutations or loss of one PTEN gene copy, respectively in 957 patients, IA16; no evidence of bi-allelic loss), and there is no enrichment for PTEN transcript suppression or biallelic PTEN loss at relapse. Moreover, while MM cell lines are considered to reflect advanced relapsed/refractory MM, nearly all are PTEN-proficient (with notable exceptions e.g. biallelic loss in OPM2). These considerations raise a question whether PTEN loss is indeed functionally associated with a competitive advantage for MM cells compared to their PTEN-proficient counterparts. In genome-scale CRISPR-Cas9 gene editing (knockout, KO) screens we observed that 16/19 MM cell lines exhibited CERES or CHRONOS scores >0.4, consistent with the proposed role of PTEN as a negative regulator of survival and proliferation: the 3 outlier cell lines included OPM2 (biallelic PTEN deletion) and 2 other lines (SKMM2 and AMO1) with no obvious genetic events suppressing PTEN function or causing constitutive activation of PI3K/Akt/mTOR signaling. We performed subgenome-scale CRISPR KO screens with focused libraries of single guide RNAs (sgRNAs) for several known (including PTEN) or candidate TSGs and control sgRNAs from MM and other tumors in vitro and confirmed that in vitro enrichment of PTEN KO in several PTEN-proficient (but not -deficient lines), consistent with the genome-scale CRISPR study results. These observations were further confirmed with in vitro competition assays where isogenic Cas9+ cells transduced with sgRNAs against PTEN or olfactory receptor (OR) genes (as DNA-cutting negative controls) were mixed, cultured for 4-12 weeks and DNA sequenced for the distribution of their sgRNAs, and determined a significant relative enrichment of PTEN KOs compared to their OR KO counterparts. Importantly, when cells harboring focused sub-genome scale sgRNA library were implanted in vivo in a bicalcium phosphate scaffold-based bone marrow (BM)-like model with a "humanized" mesenchymal bone stromal cell compartment, we observed that CRISPR KO of PTEN represented the top enriched gene perturbation in the PTEN-proficient MM cell lines MM.1S and KMS11 (not the AMO1 line) in vivo, consistent with the in vitro results. Collectively, these observations indicate that while PTEN deficiency is infrequent in newly diagnosed MM patient samples and in preclinical cell line models that reflect advanced disease, PTEN KO is associated with major "fitness" advantage for MM cells in preclinical in vitro and in vivo xenograft models, i.e. systems in which tumor cell "fitness" does not account for potential immunogenicity of human tumor cells. We hypothesized that PTEN KO remain highly immunogenic which curtails their "fitness" advantage: indeed, we observed in a series of CRISPR KO screens in PTEN-proficient cell lines from MM, other hematologic malignancies or solid tumors, PTEN loss was not associated with increased resistance to allogeneic donor-derived NK cells (in-house studies) or T-cell treatment (e.g. 41T, RENCA cells; GSE149933). In fact, for some MM cell lines (e.g. MM1S, KMS11), PTEN KO led to more pronounced sensitivity to NK cells (compared to control OR KOs) in genome-scale or focused CRISPR screens, and in validation experiments with individual sgRNAs for PTEN (vs. OR gene sgRNAs). While our functional genomic studies document that PTEN loss confers a generalizable "fitness advantage" for MM lines in vitro and in immunocompromised in vivo models, the limited, if any, association of advanced MM with PTEN loss raises intriguing questions about the potentially complex role of PTEN as a TSG. While alternative hypotheses are plausible, our results point to the immune responses of PTEN-deficient tumor cells as an additional factor that can reconcile our integrated observations in molecular annotated clinical samples and preclinical functional genomic studies.