Patterns of Aneuploidy and Signaling Consequences in Cancer

Abstract

Aneuploidy, or a change in the number of whole chromosomes or chromosome arms, is a near-universal feature of cancer. Chromosomes affected by aneuploidy are not random, with observed cancer-specific and tissue-specific patterns. Recent advances in genome engineering methods have allowed the creation of models with targeted aneuploidy events. These models can be used to uncover the downstream effects of individual aneuploidies on cancer phenotypes including proliferation, apoptosis, metabolism, and immune signaling. Here, we review the current state of research into the patterns of aneuploidy in cancer and their impact on signaling pathways and biological processes.

Introduction

Aneuploidy is the state of an incorrect (or unbalanced) chromosome number. Traditionally, aneuploidy definitions primarily included changes of whole chromosomes. Now, particularly in cancer, the field considers both numerical and structural events to be aneuploidy (1, 2). Here, we include both whole chromosome and chromosome arm alterations as we review the cancer signaling consequences attributed to these events in human cells.

Aneuploidy: Causes and Consequences

Aneuploidy is relatively rare in normal tissues (3); however, it has been a known feature of cancer cells for more than 100 years (4) and is the most prevalent somatic alteration in the cancer genome (5–7). Unlike aneuploidies in germline cells that typically arise from chromosome segregation errors during meiosis, somatic aneuploidies occur due to mitotic segregation errors (8). The majority are thought to arise from nondisjunction caused by merotelic kinetochore attachments, defects in the spindle-assembly checkpoint or chromosome cohesion, and cytokinesis failure (9, 10).

Aneuploidy is caused by and contributes to chromosomal instability (CIN), a state of ongoing chromosome segregation errors, resulting in high rates of chromosomal gains and losses (11). The relationship between aneuploidy and CIN, as well as its role in cancer, has been extensively reviewed (9, 11–13). Aneuploidy and CIN are induced by processes that are often impaired in cancer cells such as a defective DNA damage response (DDR), degenerating telomeres, chromosome missegregation (nondisjunction) and dysregulated cell cycle control (14–17). Mutations in common oncogenes and tumor suppressors, including RAS, can promote genome instability and precipitate oncogene-induced mitotic stress and aneuploidy (18–20). Importantly, not all aneuploid cells have ongoing CIN—a cell can be aneuploid with a uniform and stable karyotype, as observed in Down syndrome (DS) and certain hematological cancers (21). Aneuploidy without CIN can occur due to low missegregation rates or fitness benefits and tolerance of specific aneuploidy events (11, 12).

Aneuploidy is generally poorly tolerated and associated with fitness disadvantages in untransformed human primary cells as well as yeast and mouse models (22–25). Aneuploidies in the germline result in embryonic lethality or can cause miscarriages and developmental phenotypes (8, 26), likely the consequence of stress response pathways activated by aneuploidy (8, 27, 28). The stoichiometric imbalance caused by aneuploidy induces proteasomal stress (8, 28, 29) and is sufficient to promote an immune response (30, 31). Aneuploidy also activates p53, triggering cell cycle arrest and apoptosis (32, 33).

In mammalian cells, dosage compensation at both the RNA and protein level has been observed for both gains and losses of autosomes (34–37) though likely not through mechanisms observed for the sex chromosomes (i.e., X-inactivation). Although thought to be pathway specific (38), mechanisms of transcriptional compensation are still unknown (39). At the protein level, compensation mostly occurs for components of large protein complexes (29, 40, 41). Initial studies suggest that proteins affected by aneuploidy are compensated by changes of protein turnover rates (42, 43), which may explain how mutations that improve protein turnover can aid in aneuploidy tolerance (44). Individual monosomies are also sufficient to downregulate ribosome biogenesis, protein synthesis, and translation (45), another possible mechanism of compensation in deletions.

Aneuploidy in Cancer

Nearly 90% of solid tumors and 50% of hematopoietic cancers display gains or losses of at least once chromosome arm, with the average cancer having 10 chromosome arms gained and/or lost (5–7). Both in healthy cells (3, 46) and in the tumor context (5, 47–49), certain aneuploidies occur more frequently than others. Solid cancers generally show higher levels of aneuploidy compared with leukemias and lymphomas as well as a higher frequency of deletion, but clear patterns of individual aneuploidy events emerge in each tumor type (50) just as for copy number signatures (51, 52). Patterns of chromosome arm deletions and amplifications clustered into squamous, gynecological, or gastrointestinal cancers (6, 7). Some events occur across most epithelial cancers, such as deletion of chr8p and gain of chr8q (6). Conversely, other patterns are quite specific, such as chr1p/chr19p co-deletion in low-grade glioma (53) and chr7 gain/chr10 del in glioblastoma (47). In some cases, particular chromosomes are differentially lost or gained depending on cancer type; for example, chr21 is frequently gained in hematological cancers but preferentially lost in solid tumors (54, 55). Similarly, DS individuals with constitutional trisomy 21 have an increased risk of hematological malignancies but are protected against solid tumors (54). In addition, co-occurrence (or mutual exclusivity) of individual aneuploidy events suggests that patterns can be driven by genetic interactions (56, 57). Unsurprisingly, the most frequently mutated gene is p53 (6, 58), which allows tolerance of aneuploidy as well as whole-genome doubling (WGD; ref. 59). Tumors that have undergone WGD show increased aneuploidy (particularly deletions; refs. 6, 60), likely because an individual gain or loss causes a smaller DNA dosage change and is more tolerable.

Aneuploidy is observed at all stages of tumor development, from premalignant lesions to metastases. The stepwise acquisition of aneuploidy has been particularly well-characterized in the transition to esophageal carcinoma from Barrett’s esophagus, a premalignant condition; here, p53 mutation and whole-genome doubling occur early, along with chr3p/chr9p/chr17p deletion (61, 62). Ongoing CIN contributes to increased aneuploidy accumulation throughout cancer progression and subsequent tumor evolution (2, 57, 63). The extent of CIN and aneuploidy impact oncogenic potential (64). Some mouse models have suggested that only moderate CIN and aneuploidy levels promote tumorigenesis but extreme levels may be unviable or tumor suppressive (13, 65). Other studies have suggested that aneuploidy serves as a tumor-initiating event, introducing replication stress, genome instability, and eventually CIN to induce transformation (66).

Methods to Study Individual Aneuploidies

Although the consequences of random aneuploidy have been extensively studied, it has only recently become possible to assess the mechanisms through which specific events contribute to cancer signaling and progression. Computational methods to identify aneuploidy from patient copy number data (6, 67) and panel sequencing data (68) have allowed for increased correlation between individual aneuploidy events and clinical/genomic phenotypes (6, 69). In parallel, in vitro studies to perform gene knockdown screens of all genes in encompassed regions have been performed (70, 71); although this has been quite informative, it cannot uncover the consequences of hundreds of genes affected at the same time. Trisomy 21 was among the first targeted aneuploidy models to be explored through a combination of approaches, including individual gene screening (72) and mouse models (Robertsonian translocation or deletions of chromosomal regions with synteny to human chr21; refs. 73, 74). Many DS phenotypes, including cancer predisposition and cancer protection, have been attributed to individual chr21 genes (75). Cre/loxP systems have been applied to create mouse models of megabase deletions (76). Advances in genome engineering technologies, including TALENs and CRISPR editing, have resulted in the development of new methods and approaches to create chromosome specific models of aneuploidy as well as smaller megabase deletions (77–81). Although each of these methods have their limitations, leveraging both patient data and experimental model systems has taught us a lot about the consequences of individual aneuploidies.

Karyotype-Specific Consequences on Signaling

Aneuploidy-induced dosage alterations of fitness-enhancing and -reducing genes likely contribute to tumor development. In cancer, individual aneuploidy events provide benefits under selection (25, 82–84), likely due to the change of individual genes (or gene combinations) for each chromosome alteration (84). In human tumors, chromosome arms tend to either be gained or deleted, at least partially attributed to prevalence of pro-proliferative (GO) and antiproliferative (STOP) genes located in each arm (7, 85). Breakpoint analysis of aneuploidy events in tumors also shows evidence of both positive and negative selection (49). Gene expression analyses of patient tumors find that different aneuploidy alterations correlate with changes in different pathways (6, 69, 86). Tumor cells can become dependent on individual aneuploidies, as recently demonstrated with chr1q gain (87). Now specific chromosome arm aneuploidies can be associated with fitness effects of key cancer phenotypes including proliferation, apoptosis, DNA damage, nuclear architecture, metabolism, immune signaling, differentiation, drug response, and metastasis (summarized in Fig. 1; Table 1). Here, we will discuss the impacts of specific aneuploidy events for each of these cancer phenotypes.

Figure 1.

Summary of pathways affected by aneuploidy in cancer and the implicated chromosomes. Studies validated in human tumors are in bold. (Created with BioRender.com.)

Table 1.

Summary of pathways affected by individual aneuploidy alterations in cancer and targeted models.

Pathway	Implicated chromosomes	Mechanism	Cancer/tumor Type
Proliferation and cell cycle	Chr8p loss	Loss of tumor suppressors DLC1, FGL1, FBXO25, and TRIM35	Hepatocellular carcinoma (71, 88)
	Chr9p loss	Loss of CDKN2A/B causes premature G1/S transition and centrosome amplification	Pan-cancer (89)
	Chr8q gain	Proliferation advantage attributed to increased MYC activity	Pan-cancer (90–94)
	Trisomy 7	Survival advantage in serum-free media attributed to EGFR	Colon epithelial cells (83), Colon cancer cell lines (82)
	Chr4p loss	Associated with increased proliferation, proposed to be mediated by C4orf19	Triple negative breast cancer (95)
p53	Chr17p loss	TP53 loss	Pan-cancer (76, 96, 97)
		TP53 loss with codeletion of Eifa and Alox15b or Nf1	Lymphoma, AML (76)
	Chr1q gain	MDM4-mediated TP53 suppression and reduction of cell-cycle arrest and apoptosis	Melanoma, gastric, and ovarian cancer (87), Burkitt Lymphoma (98)
	Chr9p loss	Loss of CDKN2A, which stabilizes p53 via HDM2	Primary melanoma (99)
	Monosomy 3	Associated with downregulated expression of p53 effector PERP is associated	Uveal melanomas (100)
DNA damage	Chr8p loss	Loss of WRN is sufficient to decrease cell death	Immortalized epithelial cells (49)
	Chr11q loss	Loss of chr11q DDR genes (ATM, MRE11A, H2AFX, CHEK1) conferred PARPi sensitivity	Neuroblastoma (101)
		Sensitized cells to the CHK1 inhibitor prexasertib	Neuroblastoma (102)
	Trisomy 8	Reduced replication stress due to associated co-expression of RAD21 and MYC	Ewing sarcoma (103), APL (104)
	Trisomy 12	Trisomy 12 causes sensitivity to several replication inhibitors including etoposide, cytarabine hydrochloride, and gemcitabine hydrochloride	Human-induced pluripotent stem cells (105)
Epigenetics	Trisomy 7	Trisomy 7 increases interchromosomal contacts and induces alterations in chromatin organization involving A/B compartmentalization and TAD boundaries	Human colon epithelial cells (106)
	Trisomy 21	Trisomy 21 induced disruption of lamina-associated domains and alters nuclear architecture and chromatin accessibility	Human induced pluripotent stem cell–derived neural progenitor cells (107)
	ChrY loss	Loss of chrY causes downregulation of epigenetic modifier KDM5D, which affects epigenetic cell state	Clear cell renal cell carcinoma (108)
Metabolism	Chr8p loss	Alterations in fatty acid and ceramide metabolism, resistance to HMGCR inhibitors	Immortalized mammary epithelial cells (80)
	Chr10q loss	Loss of PTEN	Oligodendroglioma (109), Glioblastoma (6)
	Chr9p loss	Loss of MTAP, resulting in blockade of methionine salvage pathway	Pan-cancer (110–114)
	Chr1q gain	Amplification of chr1q phosphoinositide signaling enzymes	Breast cancer (115)
	Trisomy 21	Trisomy 21 is associated with mitochondrial defects, oxidative stress, differential sensitivity to insulin, and altered sphingolipid biosynthesis, among other metabolic phenotypes	Human fibroblasts (116, 117), human brain tissue (118), mouse models (119, 120)
	Chr21 monosomy	Partial monosomy for a region syntenic for chr21q alters the regulation of fat deposition	Mouse model (121)
Immune signaling	Chr3p loss	Negative correlation with interferon gamma response and TNFα signaling	Immortalized lung epithelial cells (6)
	Chr9p loss	Loss of chr9p interferon gene cluster, CD274 (PD-L1), and CDKN2A establishes immune-cold tumor microenvironment, immune evasion, reduced T-cell infiltrate, resistance to immunotherapies	Melanoma (122, 123), ALL (124), HPV− head and neck SCC (114, 125, 126), EBV+ nasopharyngeal sarcoma (127), NSCLC (128), Melanoma (99, 122)
	Chr6p gain	Anti-correlation with T cell abundance, HLA loss of heterozygosity	Lung adenocarcinoma (129)
	Chr9q gain	Gain of PD-L1, chemokines and cytokines, immunoregulatory genes	Multiple cancer types (130)
Differentiation and cell fate	Chr3q gain	Chr3q squamous-specific transcription factor inhibit epithelial differentiation	SCC (131, 132)
	Trisomy 21	Increased megakaryopoiesis and erythropoiesis and impaired B cell differentiation	Hematological malignancies (133–135)
	Chr5q loss	Hematopoietic defects due to RPS14, HSPA9, CSNK1A1 haploinsufficiency	Myelodysplastic syndromes (136)
	Chr18q loss	Resistance to TGF-β-induced growth-inhibitory signals	Colon epithelial cells (137)
	Chr1q gain	Chr1q gene BCL9 promotes WNT signaling and proliferation	Melanoma (87), Gastric (87), Ovarian (87)
		Chr1q-resident γ-secretase genes APH1A, NCSTN, and PSEN2 mediate Notch signaling	Breast cancer (138)
	Trisomy 12	Trisomy 12 increases tumorigenicity of human pluripotent stem cells in vivo and induces transcriptionally distinct teratomas	Teratomas (105)
Drug response and metastasis	Chr5q loss	Metastasis suppressor gene KIBRA encoded on chr5q	Triple-negative breast cancer (139)
	Chr8p loss	Increased resistance to microtubule inhibitors, vinblastine and docetaxel; increased sensitivity to autophagy inhibitors	Breast cancer (80)
		Increased metastatic potential partially due to resident metastasis-associated genes	Liver cancer cell lines (140)
	Trisomy 7 & 13	Increased colony size in serum-free media and resistance to 5-fluorouracil treatment	Colon cancer cell lines (82)
	Chr3p loss	Increased expression of epithelial–mesenchymal transition genes	Immortalized lung epithelial cells (6)
	Chr19p gain	Increased sensitivity to a TRAIL receptor agonist	Diffuse large B-cell lymphoma cells (57)
	Chr16p loss	Increased sensitivity to anticancer agent disulfiram due to the chr16q-resident metallothionein-encoding genes MT1E and MT2A	SF295 human glioblastoma cell line (141)
	Chr1q gain	Sensitized cells to nucleoside analogs RX-3117 and 3-deazuridine due to gain of the uridine-cytidine kinase UCK2	Ovarian, colorectal, and breast cancer cells (87)
	Chr7p gain	Sensitized cells to the AHR inhibitor CGS-15943 due to gain of AHR	Melanoma (87)
	ChrY loss	Loss of chrY correlates with increased epithelial-mesenchymal signaling and metastasis	Uveal melanoma (142)

Abbreviations: ALL, acute lymphocytic leukemia; APL, acute promyelocytic leukemia; AML, acute myeloid leukemia; SCC, squamous cell carcinoma; NSCLC, non–small cell lung cancer.

Proliferation and cell cycle

As described above, aneuploidy is generally detrimental to cell proliferation when initially induced (8, 23, 30, 143), but this is an advantage for adaptability (144, 145). Fitness benefits are observed in vitro in challenging culture conditions such as serum-free media, hypoxia, or drug treatment (82, 84, 146). Higher aneuploidy correlates with a higher mitotic index in cancer, consistent with aneuploidy in general being pro-proliferative in the cancer setting (2).

For individual aneuploidies, initial acquisition also slows proliferation in vitro; this has been observed for most trisomies (143, 147) and deletions (6, 45, 143, 148). Again, individual aneuploidy events are also positively correlated with cell cycle gene expression in cancer (6), suggesting that cells must quickly adapt to this fitness loss. A trisomy 8 model in human fibroblasts exhibited initial decreased proliferation but developed loss of contact inhibition and subsequent transformation (148). For some aneuploidy events, there is no initial decreased fitness observed, including in two independently engineered cell line models of chr8p deletion (49, 80). Integration of patient expression data and experimental studies identified putative tumor suppressors in chr8p in hepatocellular carcinoma, including the Rho-GTPase-activating protein DLC1 (88); fitness benefits could be attributed to haploinsufficiency of multiple chr8p genes in parallel (71).

Chr8q gain is associated with increased expression of cell cycle genes (6) as well as tumor progression and poor prognosis (90, 91). The MYC oncogene, located on chr8q, has a well-established role in promoting cancer cell proliferation and immune evasion, and it is frequently overexpressed in more aggressive and metastatic tumors (92, 93). MYC overexpression induces chromosomal abnormalities as well as aneuploidy, promoting tumorigenesis by further exacerbating the CIN state (94). MYC is likely not the only contributor on chr8q; increased MYC activity can also promote proliferation and tumorigenesis in collaboration with other chr8q genes in relation to the DDR (described under the subsection DNA Damage).

Aneuploidy can also affect genes critical for proper cell cycle progression. Loss of chr9p is a marker of poor prognosis and overall survival across many cancer types, reviewed by Spiliopoulou and colleagues (89). Chr9p contains tumor suppressor genes CDKN2A–encoding CDK inhibitors p16 and p14 and CDKN2B–encoding p15. In melanoma, reduced gene dosage of CDKN2A via chr9p deletion is predictive of relapse and underlies increased mitotic rate and invasion (149). CDKN2A deletion due to chr9p loss is a predictor of prognosis in low-grade glioma, particularly in IDH1/2-mutated tumors (150, 151). Chr9p deletion is also implicated in immune signaling and metabolism, described below.

Several models have identified roles for individual aneuploidies in tumor growth phenotypes. Recent studies in triple-negative breast cancer found that chr4p deletion promotes proliferation (95). Engineered removal of chr1q gain in vitro results in decreased tumorigenic capability, attributed to mediation of p53 and WNT pathways (described in more detail below; ref. 87). Deletion of large regions of chr11q in vitro is associated with increased colony-forming capability (152). Two independent studies (human colon epithelial cells and colon cancer cell lines) found that trisomy 7 provides a survival advantage in serum-free conditions (82, 83); this is at least partially attributed to amplification of EGFR, located on chr7p, affecting proliferation (83). Cell migration is also promoted by trisomy 7 in one of the models (83) and growth in hypoxia in the other (82). These studies may indicate potential vulnerabilities associated with specific aneuploidy events in a tumor context.

p53

Aneuploidy triggers activation of the p53 pathway, regulating apoptosis, senescence, and cell cycle arrest (153). The p53 pathway limits viability of aneuploid cells and in turn, p53 pathway inhibition is likely required for cells to tolerate aneuploidy. Some studies suggest that p53 is not always activated by whole-chromosome aneuploidies. However, propagation of chromosome arm aneuploidies seems to be only tolerated in p53-deficient backgrounds (154). These findings are consistent with p53 mutation being an early event in tumorigenesis (155, 156) and being the number one correlate with aneuploidy in cancer (6, 157, 158).

Tumor cells also inhibit p53 via individual arm aneuploidies. TP53 is located on chr17p, and deletion of this chromosome arm has been associated with promoting TP53 loss across cancers (96, 97). TP53 is often considered the primary driver of chr17p deletion. However, heterozygous deletion of mouse chr11B3, syntenic to human chr17p13, had a larger impact on lymphomagenesis and acute myeloid leukemia compared to Trp53 deletion, due to co-deletion of certain genes (Eifa and Alox15b in lymphoma and Nf1 in leukemia; ref. 76).

Other non-chr17 aneuploidy events have also been implicated in p53 inhibition. Chr1q is gained across many cancer types, particularly breast cancer (159). Chr1q contains the p53 inhibitor MDM4, and chr1q trisomy (i) induced MDM4-mediated TP53 suppression in cancer cells, (ii) was mutually exclusive with TP53 mutations, and (iii) was required for malignant growth in engineered melanoma, gastric cancer, and ovarian cancer cell lines (87). Work in cancer cell lines found that MDM4 is targeted by chr1q gain, increasing p53 degradation to reduce cell-cycle arrest and apoptosis and promote proliferation in Burkitt lymphoma (98). These studies highlight a relationship between chr1q and p53 regulation. Aneuploidies affecting other p53 inhibitors, like MDM2 on chr12q and PPM1D on chr17q, should also be explored. Similarly, chr9p deletion has been shown to increase primary melanoma relapse due to loss of CDKN2A, which otherwise encodes P14ARF, which then interacts with HDM2 to stabilize TP53 and facilitate cell cycle arrest at the G1/G2 phase (99).

Other studies have indirectly associated p53 with specific aneuploidies via changes in its effectors. p53-induced apoptosis effector PERP is located on chr6q in a region for which loss of heterozygosity has been implicated in ovarian, breast, and cervical cancers as well as melanoma, reviewed by Attardi and colleagues (160). Downregulated expression of PERP is associated with monosomy 3 in aggressive subtypes of uveal melanomas (100). These studies implicate specific aneuploidies or potential p53-associated targets for deletion or mutation in cancer.

DNA damage

As described above, aneuploidy is associated with DNA damage, compromised DDR and repair pathways, and replication stress that in turn lead to increased CIN and genomic instability (24, 161). Individual chromosome missegregation, including chrY, is sufficient to induce subsequent structural aberrations of many types, including micronucleation and chromothripsis (162, 163). Replication stress is a hallmark of trisomic cells (30, 45, 164), inducing sensitivity to replication inhibitors in human and yeast studies (24, 161). Even a single chromosome gain can cause replication-associated DNA damage, impaired S-phase progression, increased susceptibility to chromosome fragility and structural aberrations, and sensitivity to the replication inhibitor aphidicolin; these phenotypes were attributed to a decrease in expression of the DNA helicase MCM2-7 (161). For example, trisomy 12 in a human pluripotent stem cell induces sensitivity to replication inhibitors including the DNA topoisomerase II poison etoposide and the nucleoside analogues cytarabine hydrochloride and gemcitabine hydrochloride (105).

Recent work showed an association between DNA damage pathways and chr8p loss, which occurs frequently across epithelial cancers (49, 71, 80, 88). Chr8p deletion cell line models do not show decreased fitness in vitro (49, 80) and have decreased rates of cell death (49). This was attributed, at least in part, to the WRN helicase located on chr8p. WRN inactivation is synthetic lethal with microsatellite instability (165), and chr8p deletion also does not co-occur with microsatellite instability (49). The relationship between WRN loss, chr8p loss, and cell fitness highlights a connection between chr8p aneuploidy, DDR, and repair pathways.

Trisomy 8, while initially detrimental to wild-type cells (148), was found to dampen replication stress in the context of EWS-FLI1 fusion oncogene-driven Ewing sarcomas (103). EWS-FLI1 induces premature S phase entry and inhibits BRCA1-mediated DNA repair that ultimately promotes replication stress and cellular senescence (103). The dampening by trisomy 8 was attributed to copy number gain of the chr8q gene RAD21—a cohesin complex subunit that regulates replication fork stabilization and is a rate-limiting factor for DNA repair (103). Interestingly, copy number gain of both RAD21 and MYC is necessary to confer the oncogenic effects of trisomy 8 in EWS-FLI1 expressing cells. In mice, Myc and Rad21 are co-located on chr15, which is frequently gained in T-cell lymphoma models. Mice with chr15 gain but ablated endogenous Myc and subsequent exogenous human Myc expression on murine chr6 showed chr6 gains following clonal selection (104). Furthermore, they retained the chr15 gain except following deletion of a single Rad21 copy (104), further suggesting an association between MYC and RAD21 in driving this aneuploidy event.

Aneuploidy events also affect DDR genes. Recurrent deletions of chr11q involve the loss or genetic imbalance of DDR genes including ATM, MRE11A, H2AFX, and CHK1, which mediate cell cycle checkpoints, apoptosis, and DNA repair pathways (101). Chr11q loss sensitized neuroblastoma cells to the CHK1 inhibitor prexasertib, and concurrent chr11q loss and MYCN amplification caused synthetic lethality with CHK1 inhibition (102). Furthermore, as ATM is a known regulator of homologous recombination repair, defects in this pathway induced by this event conferred sensitivity to the PARP inhibitors olaparib; whether this is also true for 11q aneuploidy is still unknown (166). Elucidating the context-specific relationship between aneuploidy and DNA damage pathways will aid in identifying therapeutic targets.

Nuclear architecture and epigenetics

Altered epigenetic status and chromosome dynamics has been associated with aneuploidy and CIN (167). The impairment of DNA methyltransferases promotes demethylation at pericentromeric regions, and this is sufficient to induce aneuploidy (168–173). Hypomethylation of repetitive elements correlated with nuclear size and aneuploidy in ovarian cancers (174). Hypomethylation of individual regions correlates with copy number changes at other locations on the chromosome, such as chr1 heterochromatin DNA and chr1q copy gain (169). Aneuploidy can impact nuclear architecture and three-dimensional chromatin state and topological domains, leading to loss of regulatory elements, dysregulated gene expression, and further copy number changes (175). Nuclear chromosome locations in the interphase nucleus prior to the onset of mitosis can directly influence missegregation frequencies (176). The loss or reduction of lamins can impair cytokinesis and induce formation of tetraploid and aneuploid cells (177–179).

Specific chromosome arm aneuploidies have also been associated with affecting nuclear architecture and epigenetics. Interestingly, trisomy 7 impacts nuclear organization and gene expression by increasing interchromosomal contacts between triploid chromosomes and altering A/B chromatin compartmentalization and topologically associated domain (TAD) boundaries (106). In another study, trisomy 21 in iPSC-derived neural progenitor cells induced disruption of lamina-associated domains, altered global chromatin accessibility, and impacted transcriptional and nuclear architecture of senescent cells, which suggested a relationship between senescence-associated phenotypes and neurodevelopmental pathogenesis of Down syndrome (107). Lastly, loss of chrY has been shown to affect epigenetic cell state in clear cell renal cell carcinoma, due to downregulation of the resident epigenetic factor KDM5D (108). Further study is required to elucidate the relationship between chromosomal aneuploidy and nuclear organization and epigenetic state.

Metabolism

Compromised metabolic pathways and stress can drive aneuploidy and tumorigenesis; similarly, aneuploidy and CIN can induce alterations in cellular metabolism involving lipid biosynthesis, proteotoxicity, mitochondrial activity, and reactive oxygen species (23, 28, 40, 180, 181). Aneuploidy-induced stress alters mitochondrial reactive oxygen species (182–184), glucose uptake (23), and nucleotide and carbohydrate metabolism (23). Highly aneuploid colorectal cancer cells have increased intracellular ceramide, potentially attributed to dysregulated sphingolipid metabolism (185–187). Inhibiting sphingolipid synthesis or increasing ceramide levels slowed the proliferation of aneuploid cells (186) whereas inhibiting ceramide synthesis improved proliferation (116). Trisomic and tetrasomic human HCT116 cells have also been shown to have upregulated mitochondrial, carbohydrate, and membrane metabolism pathways (40). Trisomic mouse embryonic fibroblasts, aneuploid primary cells, and aneuploid cancer cells have increased production of metabolites involved in the citric acid cycle and glycolysis such as lactate and glucose (143, 188).

Metabolism is also influenced by individual aneuploidy events, particularly assessed in the context of trisomy 21 (119, 120, 189). Trisomy 21 has been associated with mitochondrial defects, increased oxidative stress, reduced energy production, altered glucose metabolism, and insulin resistance in DS (118, 190). In a transchromosomic mouse model carrying a near-complete copy of human chr21, the mice were hypermetabolic and insulin sensitive, attributed to changes in mitochondrial respiration (119). Human fibroblasts with trisomy 21 have elevated sphingolipid levels and increased serine-driven lipid biosynthesis (189) as well as mitochondrial dysfunction and altered carbon metabolism (117). When overexpressed, the chr21 gene RCAN1 was found sufficient to reduce β-cell mitochondrial function and ATP availability as well as impairing glucose-stimulated insulin secretion (120). In another study, partial monosomy for chr21q identified a locus relevant to regulation of fat deposition (121). Altogether, these models demonstrate diverse consequences of specific trisomies in relation to metabolic phenotypes.

The PI3K/AKT/mTOR signaling pathway can reprogram cellular metabolism to promote growth (191) and is often affected by aneuploidy. PIK3CA is located on chr3q, frequently gained in squamous cancers (6). Loss of the PI3K inhibitor PTEN by deletion of chr10q is observed at high frequencies in many cancers including oligodendrogliomas and glioblastomas (6, 109). Amplification of chr1q-resident genes encoding phosphoinositide signaling enzymes PI4KB, AKT3, PIP5K1A, and PI3KC2B have been associated with breast cancer (115). AKT1 and AKT2 are located on chr14q and chr19q, respectively, and these chromosome arms are gained in some cancer types (6). Some studies have also implicated PI3K-Akt in mediating oncogenic hepatocyte growth factor receptor (Met)-induced centrosome amplification, aneuploidy, and CIN (192).

Additional arm deletions have also shown effects on metabolism. The chr9p enzyme MTAP is involved in polyamine metabolism and the salvage pathway of adenine and methionine (110). Frequent loss of MTAP via chr9p loss has been shown to result in blockade of the methionine salvage pathway in gliomas (111). MTAP deficiency has been shown to confer sensitivity to antifolate therapy in chr9p21-loss cancers (112). MTAP depletion also confers metabolic vulnerabilities like sensitivities to purine biosynthesis inhibitors (113); to our knowledge, it still remains an open question whether chr9p deletion has the same effect. Chr8p deletion alters fatty acid and ceramide metabolism in immortalized mammary epithelial cells, and this can promote invasiveness and tumor growth under hypoxic conditions due to increased autophagy (80). This altered lipid metabolism also conferred resistance to statins (HMG-CoA reductase inhibitors) and sensitivity to autophagy inhibitors (80). These studies identify several metabolic pathways and phenotypes that can potentially be exploited therapeutically.

Immune signaling

Aneuploidy, when induced in cells, activates immune responses (30, 31,193) with CIN demonstrated to promote CGAS-STING activity (194). Congenital trisomies have been associated with dysregulated immune response, which has been reviewed elsewhere (164). In contrast, aneuploidy in cancer anticorrelates with immune gene expression and infiltrate (6, 69, 195), and similarly anticorrelates with immunotherapy response (69, 196). The cause of this paradox is still up for debate but is currently attributed to acute versus chronic immune activation and/or adaptation during tumorigenesis (2, 8).

Several pieces of evidence suggest that specific aneuploidy patterns may also affect immune signaling. For example, immune gene expression signatures positively correlate with deletion of chr3p, chr8p, chr13q, and chr17p (6), suggesting that genes within these aneuploidies may underlie differential immune signaling responses and pathways. The correlation with chr3p deletion was experimentally validated; in vitro chr3p loss was correlated with interferon gamma response and TNFα signaling (6).

Chr9p deletion in non–small cell lung cancer and head and neck squamous cell carcinoma is a predictor of immunotherapy resistance and low immune infiltrate (197) attributed to several loci (69, 114, 125). In computational analyses of patient tumor data, the effects of these two regions were separated, with deletion of chr9p24 associated with CD8⁺ T-cell immune-cold microenvironments driven by CD274 (PD-L1) at this locus (125) and chr9p21 deletion associated with cell-intrinsic senescence suppression (126). The interferon gene cluster is also located at chr9p21, implicating chr9p deletion as a mechanism through which cancer cells can evade type I interferon induction (114, 122–124). Chr9p loss is associated with depletion of cytotoxic T-cell infiltration via reductions in IFNγ-related chemokines, like CXCL9 and other genes as well as JAK-STAT, NF-κB, and cytokine-related pathways (125, 126, 128). CDKN2A is also located on chr9p at chr9p21; beyond its role in proliferation, CDK2NA has been identified as a biomarker for increased infiltration of activated CD4⁺/CD8⁺ T cells and NK cells (89, 198). The associated loss of MTAP has also been suggested to inhibit interferon-mediated STAT signaling pathways (89, 150), reprogramming the tumor microenvironment, and in other studies conferred sensitivity to MAT2A inhibition (127). Overall, this suggests that chr9p loss confers cancer cells with an ability to evade aneuploidy-induced immune checkpoints and resistance to immune checkpoint therapies like anti-PD-1/PD-L1 therapy (126).

Chr9p deletion is not the only aneuploidy event to correlate with immune changes. Deletion of chr6p, where the HLA locus is located, is observed frequently during tumor evolution (199). Genomic and immune profiling of preinvasive lung adenocarcinoma found chr6p gain anticorrelated with T-cell abundance (129). Chr9p gain has been associated with gain of PD-L1 and specific immune-modulating gene expression profiles across multiple cancer types (130). Pan-cancer analyses found that deletion of chr1p or chr19q correlates with reduced leukocyte fraction (200); other amplifications (chr2, chr20q, chr22q) and deletions (chr5q, chr9p, chr19) are associated with variations in macrophage polarity (200).

Differentiation and cell fate

Stem cells are thought to better tolerate aneuploidy (201), and generation of induced pluripotent stem cells (iPSC) often comes with acquisition of aneuploidy (202). Mitotic error rates correlate with pluripotency, with fewer errors in more differentiated states (201). Cancers of different tissues are characterized by different aneuploidy levels; in general, epithelial cancers and sarcomas have high aneuploidy compared to leukemias and lymphomas (5, 6, 60). The tissue-specific patterns of aneuploidy observed in both cancer and healthy tissues also suggest a role for specific aneuploidies in cell fate. Breakpoint distributions of telomere- and centromere-bound copy number events also show lineage specificity (49), further suggesting that tumor types specify aneuploidy selection biases. Gene expression analyses of healthy and cancer tissue suggests that tissue-specific expressed genes contribute to the observed cancer aneuploidy patterns (203). Chromosome evolution screens in different cell types found that in vitro patterns mirrored what is observed in human cancers (138).

The clearest demonstration of individual aneuploidy events affecting differentiation is in blood cancers. Individuals with DS have an increased risk of some leukemias and lymphomas, attributed to enhanced megakaryopoiesis and erythropoiesis and problems with proper B-cell differentiation (133, 134). These hematopoietic phenotypes are thought to at least in part be due to two hematopoietic transcription factors located on chr21 (135). Trisomies in general are associated with leukemias, including mosaic trisomy of chr3, chr8, or chr12 (204). Chr18 is gained in follicular lymphoma (among other blood cancers) and is associated with misregulation of lymphocyte balance (205).

A subset of patients with myelodysplastic syndrome (MDS) have a large region deleted on chr5q. Patients with 5q– MDS have defects in hematopoietic differentiation that have now been attributed to chr5q genes (136). Specifically, RPS14 haploinsufficiency contributes to macrocytic anemia, CD74 and HSPA9 haploinsufficiency to neutropenia, and CSNK1A1 contributes to clonal expansion. WNT pathway regulators CSNK1A1 and PP2A are also located in the chr5q minimal deleted region; their haploinsufficiency causes sensitivity to lenalidomide in 5q– MDS (136).

Solid tumors also show stem cell and differentiation pathways affected by aneuploidy. Gain of chr1q includes amplification of BCL9, a member of the WNT signaling pathway; BCL9 contributes to the growth advantage of chr1q trisomy (87). In breast cancer, chr1q trisomy is thought to promote tumor development through Notch activation, as genes for three gamma-secretase components reside on chr1q (138). In colon epithelial cells, chr18q loss—which recurs in gastrointestinal cancers—results in resistance of growth-inhibitory signals induced by the cytokine TGFβ via an unknown mechanism; TGFβ inhibition is required for proliferation and expansion of colon epithelial cells (137, 206). Squamous cancers of many tissue types (including lung, head and neck, esophagus, and cervix) are all characterized by gain of chr3q (6, 47, 48). Chr3q contains multiple squamous-specific transcription factors thought to contribute to this specificity, including SOX2 and TP63 (131, 132). Overexpression of these genes is sufficient to inhibit proper epithelial stratification and differentiation, instead promoting squamous metaplasia. As more targeted aneuploidy models are generated, the field will surely identify additional roles in cell fate.

Specific aneuploidy patterns have also been observed in stem cells. iPSCs are most likely to gain an extra copy of chr12, which has been shown to provide a proliferative benefit (105). In the same study, trisomy 12 increased the tumorigenicity of iPSCs in vivo and induced transcriptionally distinct teratomas (105). In human ES cell lines, other frequently observed aneuploidies include gain of chr1, chr17, or chr20, with a particularly common amplification of 20q11.21 (207). BCL2L1 at this locus is thought to contribute to embryonic stem cell adaptation mechanisms (207); whether this has similar implications in cancer, where 20q is frequently gained across many cancer types (6), is still unknown.

Drug response and metastasis

Aneuploidy evolution within the tumor is also thought to contribute to drug response (57, 78), resistance (57, 208), epithelial–mesenchymal transition (209), and metastasis (13, 210, 211). Aneuploidy and CIN have been associated with driving metastasis (13, 211–213) in prostate, pancreatic, breast, colorectal, and renal cell carcinoma (63, 214) as well as relapse (13, 63, 215). Aneuploidy correlates with increased sensitivity to several drug classes including protein folding inhibitors (216), autophagy inhibitors (80), chemotherapeutics (208, 217), and inhibitors of KIF18A (218–220). Compounds targeting these pathways have been identified across multiple studies and experimental approaches (57, 185). Higher aneuploidy load has also been shown to correlate with worse response to immunotherapy (69, 196).

Frequently occurring aneuploidies can shape therapeutic response, which may be relevant to exploit as therapeutic targets or biomarkers (57). In a systematic chromosome arm aneuploidy study of a panel of cancer cell lines, cells with chr19p gain were found to be more sensitive to treatment with a TRAIL receptor agonist, among other differential sensitivities (57). Another cancer cell line screen found that deletion of chr16p correlates with sensitivity to the anticancer agent and acetaldehyde dehydrogenase inhibitor disulfiram, attributed to two metallothionein-encoding genes on chr16q (141). Trisomy 7 and trisomy 13 colon cell lines both confer selective advantages of increased colony size (in soft agar) when grown in serum-free media or with 5-fluorouracil chemotherapy (82). Chr1q gain sensitizes ovarian, colorectal, and breast cancer cells to the nucleoside analogs RX-3117 and 3-deazuridine due to copy number gain of UCK2, and chr7p gain sensitizes melanoma cells to the AHR inhibitor CGS-15943 due to gain of AHR (143). Engineered cells with chr8p deletion are more resistant to some chemotherapeutics, including DNA damaging agents and microtubule inhibitors (80).

Related to metastasis, a triple-negative breast cancer mouse model of chr5q loss (via deletion of a region syntenic with human 5q33.2-35.3) found that the metastasis suppressor gene KIBRA contributes to the effects of chr5q loss on tumor growth and metastasis (139). Specifically, as KIBRA inhibits nuclear localization of YAP/TAZ, reduced dosage of KIBRA due to chr5q loss promotes YAP/TAZ oncogenic functioning (139). Gene expression studies suggest that chr3p deletion could also affect the epithelial–mesenchymal transition, though mechanistic validation is still ongoing (6). In analyses of paired primary and metastatic samples, loss of chr9p has also been found to promote metastasis (221). In uveal melanoma, loss of chrY correlates with increased epithelial–mesenchymal transition signaling and metastasis (142). Recently, Huth and colleagues engineered chr8p deletion in liver cancer cell lines and found increased migration, invasion, and dysregulation of pathways related to migration and metastasis; this is attributed at least in part to a handful of metastasis suppressors on chr8p (140). Further study of the relationship between aneuploidy and metastasis will help determine the role that aneuploidy plays along cancer progression.

Summary and Future Impact

Studying the downstream consequences of aneuploidy patterns not only provides insight into their convergent roles in cancer development, but also insight into their prognostic value and aneuploidy-induced vulnerabilities. Aneuploidy often correlates with worse prognosis, but chr1p/19q deletion correlates with improved prognosis in low-grade glioma for any treatments assessed (53). As described above, chr5q deletion in MDS is a biomarker for lenalidomide response (136). Chr9p deletion in HPV⁻ head and neck squamous cell carcinoma is a biomarker for immunotherapy response based on studies showing its effect on immune infiltrate (126); one could imagine that other aneuploidies that affect immune signaling may also predict response to this treatment modality. Another possibility is to target aneuploid cells by identifying synthetic lethalities. For example, cells with chr1p deletion are dependent on inhibition of MAGOHB, as its paralog MAGOH is located on chr1p (222). Cells with chr8p deletion were recently found to be sensitive to inhibition of NUDT17, as its paralog NUDT18 is located on chr8p (140). Computational analyses have identified additional aneuploidy-pattern targets, including tumor type–specific sensitivities correlating with gains of chr1p, chr10p, chr12p, chr18p, or chr19p (57); validating these experimentally could identify useful targets for preclinical studies. With the increased generation of targeted aneuploidy models in human cells, we expect these events to be a new class of precision medicine targets in oncology.

Authors’ Disclosures

A.M. Taylor reports grants from NCI, NIGMS, and American Cancer Society during the conduct of the study and is a co-founder of KaryoVerse Therapeutics, Inc. No disclosures were reported by the other author.