Functional Oncogenomic and Immune Response Landscape for Genes Recurrently Mutated in Myeloma

For many genes that are recurrently mutated in patient-derived samples or cell lines from multiple myeloma (MM), the functional roles of these genes have remained incompletely understood. We have sought to address the functional implications of these genes through extensive CRISPR-Cas9-based functional genomics studies for loss-of-function (LOF, e.g. CRISPR-based gene editing) and gain of function (GOF, e.g. CRISPR-based gene activation) in MM cell lines in vitro and in vivo. We focused on 96 genes mutated in >2% of newly diagnosed MM patients in the CoMMpass (IA17) study (after excluding non-expressed genes [<1 FPKM in 95% of samples] and immunoglobulin/MHC genes) and specifically examined the performance of these genes in a series of in vitro genome-scale studies for CRISPR gene editing or activation (n=19 and n=5 MM lines, respectively); in vitro validation studies with subgenome-scale focused sgRNA libraries of individual sgRNAs; and in vivo focused subgenome-scale CRISPR screens (in 3 cell lines) in conventional subcutaneous (s.c.) xenograft or in bone marrow (BM)-like scaffolds with "humanized" mesenchymal BM stromal cell compartment (seeking a more translationally-relevant simulation of the BM microenvironment in MM patients).

In these LOF and GOF functional studies, there were several examples of recurrently mutated genes which performed in vitro and in vivo in a manner consistent with prior knowledge on these genes as essential for MM survival/growth (e.g. NRAS or KRAS in MM lines harboring mutations for those genes, FGFR3 in t(4;14) lines) or, conversely, as tumor suppressor genes (e.g. TENT5C/FAM46C, TRAF3, TP53). Beyond these genes, which served as "positive controls", other genes with notable functional roles identified in these studies include recurrently essential genes such as SETD2, the E3 ligases HUWE1 and HERC2, or the RNA binding protein genes RBMX and ELAVL1. Interestingly, however, the large majority of genes recurrently mutated in MM do not exhibit a recurrent role as either essential genes/drivers of MM cell growth or tumor suppressor genes in these in vitro or in vivo functional genomics studies. For instance, genes such as LTB, ATM, DUSP2, the phosphatases PTPRD, PTPRF, PTPRM do not exhibit recurrent functional role as major positive or negative regulators of MM cell growth in these studies, regardless of the mutational or copy number status of these genes in the cell lines tested.

The fact that these observations were concordant in both in vitro and in vivo CRISPR studies suggests that the majority of genes recurrently mutated in MM patients do not seem to impact, either when lost or gained, the MM cell survival/proliferation in not only cell autonomous in vitro conditions, but also in the context of interaction with the local microenvironment of a "humanized" BM-like milieu. Given the immunocompromised nature of these in vivo systems, we also examined whether these recurrently mutated genes could influence the ability of MM cells to escape immune surveillance. However, in our in vitro genome-scale LOF or GOF CRISPR screens in three MM cell lines exposed to allogeneic donor-derived NK cells, perturbations of genes recurrently mutated in MM patient samples also did not emerge as "hits" associated with NK cell resistance.

The fact that most recurrently mutated genes in MM did not elicit a clear phenotype in either loss- or gain-of-function studies with endpoints focusing on "cell fitness" (survival/proliferation) or in the context of exposure to immune effector cells such as NK cells, does not preclude alternative functional roles that may conceivably involve e.g. responses to other types of immune effector cells; cooperative interactions with other genes; roles in DNA repair (e.g. a plausible scenario in the case of ATM); impact on MM self-renewal; or potential role in early stages of myelomagenesis, but not in its late/advanced forms (which are generally considered to be represented by cell lines).

More broadly, this study highlights the significance of identifying candidate drivers of MM cell biology by integrating conventional mutational analyses with functional characterization of the respective genes of interest, ideally through assessment of diverse endpoints related to both cell-autonomous and non-autonomous tumor cell behavior.