Whole genome doubling confers unique genetic vulnerabilities on tumor cells

Abstract

47Whole genome doubling (WGD) occurs early in tumorigenesis and generates genetically 48 unstable tetraploid cells that fuel tumor development. Cells that undergo WGD (WGD + ) 49 must adapt to accommodate their abnormal tetraploid state; however, the nature of these 50 adaptations, and whether they confer vulnerabilities that can subsequently be exploited 51 therapeutically, is unclear. Using sequencing data from ~10,000 primary human cancer 52 samples and essentiality data from ~600 cancer cell lines, we show that WGD gives rise to 53 common genetic traits that are accompanied by unique vulnerabilities. We reveal that 54 WGD + cells are more dependent on spindle assembly checkpoint signaling, DNA 55 replication factors, and proteasome function than WGDcells. We also identify KIF18A, 56 which encodes for a mitotic kinesin, as being specifically required for the viability of 57 WGD + cells. While loss of KIF18A is largely dispensable for accurate chromosome 58 segregation during mitosis in WGDcells, its loss induces dramatic mitotic errors in 59 WGD + cells, ultimately impairing cell viability. Collectively, our results reveal new 60 strategies to specifically target WGD + cancer cells while sparing the normal, non-61 transformed WGDcells that comprise human tissue. 62The vast majority of human cells are diploid and numerous cell cycle controls exist to help 63 ensure that this state is maintained across successive cell divisions 1 . Despite these controls, 64 errors can occur that result in a whole genome doubling (WGD), in which a natively diploid cell 65 transitions to a tetraploid state 1-3 . It has been demonstrated that cells that have experienced a 66 WGD event (hereafter WGD + ) are oncogenic and can facilitate tumorigenesis 4,5 . WGD promotes 67 tumorigenesis in at least two ways: first, proliferating WGD + cells are genomically unstable and 68 rapidly accumulate both numerical and structural chromosomal abnormalities 5 , and second, 69 WGD + cells are better able to buffer against the negative effects of deleterious mutations and 70 ongoing chromosome instability 6-10 . Such traits enable nascent WGD + tumor cells to proliferate 71 in the presence of otherwise lethal genomic alterations while simultaneously sampling increased 72 genetic permutations, ultimately enabling phenotypic leaps that give rise to tumors 8,11 . WGD also 73 carries important clinical implications, with recent reports showing its correlation with advanced 74 metastatic disease and a worse overall prognosis 12,13 . 75 76 Given the oncogenic potential associated with WGD, tumor suppression mechanisms exist to limit 77 the proliferation of these unstable cells. WGD + cells activate both the p53 and Hippo tumor 78 suppressor pathways and are prone to apoptosis, senescence, and immune clearance [14][15][16] . WGD also 79gives rise to numerous abnormalities in cellular physiology that impair fitness 6,14,17 . Therefore, in 80 order to promote tumorigenesis, WGD + cells must adapt to overcome these barriers 5,14,18,19 . Thus, 81 while ...

Chromosomally unstable tumor cells specifically require KIF18A for proliferation

“…Our tests of other kinesins that control spindle microtubule dynamics and chromosome movements suggest the observed dependence of CIN cells on KIF18A is unique. In agreement, two recent, large-scale bioinformatics studies identified Kif18A , but not other kinesins, as a gene specifically required for the growth of cells displaying aneuploidy or whole genome duplication 47,48 . Thus, KIF18A represents a potential target for exploiting vulnerabilities specific to a significant fraction of tumor cells displaying CIN or aneuploidy.…”

“…KIF18A KD increases the frequency of micronucleus formation as a result of chromosome alignment defects, and these micronucleated cells display reduced proliferation 20 . In support of this idea, recent work indicates that the frequency of KIF18A KD-dependent micronuclei is increased in cells with elevated chromosome number and correlates with reduced proliferation in aneuploid cells 47,48 .…”

Spindle Architectural Features Must Be Considered Along With Cell Size to Explain the Timing of Mitotic Checkpoint Silencing

“…Previous studies have shown that, in some experimental systems, tetraploidy can lengthen mitosis, but it is unclear if cell and spindle scaling contribute to this delay (Kuznetsova et al, 2015 ; Viganó et al, 2018 ; Paim and FitzHarris, 2019 ; Cohen-Sharir et al, 2020 ; Quinton et al, 2020 ). To investigate whether the differences in cell size and spindle geometry among the 4N clones affect their progression through mitosis, we measured the timing from cell rounding to anaphase onset (referred to as mitotic duration henceforth) using live-cell phase contrast microscopy ( Figure 3A and Supplementary Videos 11–15 ).…”

“…Our findings suggest that rather than multipolarity resulting from extra centrosomes, relatively subtle variations in spindle geometry, spindle pole size, microtubule abundance, and cell volume can affect mitotic progression and SAC silencing in tetraploid cells. Thus, even though spindle bipolarity may be preserved, other changes in cell size and spindle architecture—its geometry and composition—may promote mitotic dysfunction and contribute to the genomic instability caused by tetraploidy (Storchova and Kuffer, 2008 ; Dewhurst et al, 2014 ; Cohen-Sharir et al, 2020 ; Quinton et al, 2020 ).…”

“…For instance, cell cycle and mitotic progression were not perturbed in tetraploid yeast cells (Storchová et al, 2006 ), which may be due to the lack of spindle scaling in those strains. In contrast, tetraploid mammalian cells often progress through mitosis more slowly than their diploid counterparts (Kuznetsova et al, 2015 ; Paim and FitzHarris, 2019 ; Cohen-Sharir et al, 2020 ; Quinton et al, 2020 ), suggesting a possible delay in SAC silencing. Several recent studies, which explored the link between cell and spindle sizes and SAC function more closely (Gerhold et al, 2015 , 2018 ; Galli and Morgan, 2016 ; Kyogoku and Kitajima, 2017 ), also produced conflicting results.…”