**5. BRCA1/2 mutations and homologous repair defects**

Though few SNVs in 'classic' tumor genes are found in SOC relative to other cancer types, BRCA1/2 mutations are among the most frequent at ~10% [12]. BRCA genes work in coordination with dozens of other proteins to perform genome maintenance via homologous recombination [4]. The double-stranded break repair pathway begins with PARylation of the break site by PARP1, megabases of phosphorylation of H2AX and subsequent formation of Rad51 filaments. Brca1 & Brca2 bind Rad51 to stimulate strand invasion of sister chromatids during homology directed repair. While only ~10% of SOC are mutated in BRCA1 or BRCA2, it is noteworthy that 75% of patients have lost one of two alleles of BRCA1 and 57% have lost an allele of BRCA2. Very few tumors (~1.5%) have homozygous deletions in BRCA1 [12], suggesting a system of compromised (but not lost) function. In fact, mRNA expression level does not track linearly with such deletions. It remains somewhat unclear if these monoallelic deletions do have a phenotype under genotoxic stress in human cells.

amplification, but SCNA instability, which may explain why it is observed in only a small minority of ovarian cancer patients, and is not linked to BRCA mutation status. Nonetheless, BRCA genes are inactivated through allelic deletions and expression modulation in ovarian cancer. Inactivation of *TP53* also suppresses BRCA1 expression [58]. BRCA1 interacts with ATM, directing it to phosphorylate p53 to enable p21 induction and G1/S phase cell cycle arrest [59, 60]. Without these complementary functions, tumors spontaneously form in BRCA1−/− p53+/− mice [61]. Centrosome duplication correlates with BRCA1 deletion [60]. This complex is thus a critical factor determining proper chromosome segregation during mitosis, and upon centrosome duplication aneuploidy is assured to form upon cell division [62, 63]. This is facilitated by prior mutations of p53 that abolish its checkpoint control func-

Genomic Copy Number Alterations in Serous Ovarian Cancer

http://dx.doi.org/10.5772/intechopen.72695

119

Coincident SCNA events enable subsequent SCNA catastrophe. BRCA1 is located within kilobases of the neighboring autophagy gene BECN1. Autophagy is a critical catabolic infrastructure that enables cellular survival, requiring only 10% or less of normal autophagy gene dose [64]. Monoallelic loss of *BECN1* promotes centrosome amplification and aneuploidy among apoptosis-resistant cells [65]. An activator of the BECN1–PIK3C3 autophagy initiating complex, UVRAG, displays a similar phenotype [66]. Remarkably, this centrosome amplification promotes cell migration independent of its function in aneuploidy via indirect, hyperactivation of Rac1 [67]. LC3B, a ubiquitin-related protein which marks autophagosomal membranes, also acts in microtubule quality control [68]. One allele of the LC3B gene, *MAP1LC3B*, is genetically deleted in over 75% of SOC, independent of changes to *BRCA1* and *BECN1*. These three aneuploidy-accelerating lesions are likely to play key roles in serous ovarian cancer tumor initiation: mutant p53 expression, along with *BRCA1* and *BECN1* loss. Additional tumor suppressors may synergize to foster genomic instability: *NF1* is on the same chromosome arm as *BRCA1* and *RB1* lies nearby to *BRCA2*. Cumulative haploinsufficiency associated with prime

targets and nearby neighbors certainly contribute to SOC aneuploidy [9] (**Figure 3**).

**Figure 3.** A model consistent with the ubiquitous mutation of p53 (SNV, indicated by green text) together with early SNV or haploid loss (indicated by light blue text) of repair enzymes, permitting a self-sustaining cascade of SCNAs and

subsequent *in situ* selection for the observed SOC phenotypes.

tion, again stressing the role for early mutation of TP53 in SOC.

Mutations in homologous repair coordinating factors are often found in serous, clear cell, endometrioid, and carcinosarcoma ovarian cancers [47, 48]. Specific mutation patterns are found within BRCA1/2 or otherwise homologous repair deficient cells. Without functional homologous repair, cells default to non-homologous end joining (NHEJ) to repair doublestranded DNA (dsDNA) lesions. NHEJ does not perfectly repair DNA, but rather often introduces small insertions or deletions along with single-nucleotide variants at the break site. These mutational marks are frequently found in non-serous ovarian cancers, yet are unlikely the main drivers of SCNA instability in serous ovarian cancer. However, NHEJ factors involved in repairing unresolved dsDNA breaks across different chromosomes, or creating translocations and other complex rearrangements, are compromised in 40% *ex vivo* ovarian cancer isolates [13, 49]. This can lead to resistance to PARP inhibitors and promote inappropriate translocation or "repair" events [50]. Such defects in NHEJ may explain why the majority of SOC initially respond to cisplatin-based chemotherapy. Complex dsDNA lesions incurred by cisplatin target NHEJ and promote mitotic catastrophe [51].

Genetic and epigenetic changes alter *BRCA1* and the homologous repair pathways in SOC. Gene breakage is commonly observed within *RB1*, *NF1*, *RAD51B*, and *PTEN* [13]. While suppression of homologous recombination may lead to initial disease formation, there is evidence of *BRCA1/2* reversion mutations in tumors which become chemoresistant. This phenomenon follows the strong selective effects of carboplatin and taxanes which require cellular DNA repair pathways to enable cell division. Clinically, these findings should be considered in the context of the search for patient populations for PARP inhibitors. A primary hypothesis for how PARP inhibitors like Olaparib and Niraparib work is by targeting a DNA repair pathway which compensates for homologous repair, thereby presenting synthetic lethality specifically in cancer cells [52]. In a Phase III trial of Niraparib, clinicians treated both BRCA1 or BRCA2 mutant tumors as well as patients who were not found to have mutations in homologous repair genes in their tumors. Unexpectedly, all groups responded to Niraparib therapy [53], although patients with mutated BRCA1/2 or otherwise were defective in homologous repair were further delayed in cancer progression. While this is certainly an exciting development in the treatment possibilities for SOC patients, some caution is warranted. Reversion mutations enabling resistance to Niraparib may actually confer resistance to subsequent chemotherapeutics normally used upon disease recurrence [54].

Loss in BRCA1 enables microsatellite instability in mouse models and in colorectal cancer [55], though not in ovarian cancer [56, 57]. Microsatellite instability directly leads to centrosome amplification, but SCNA instability, which may explain why it is observed in only a small minority of ovarian cancer patients, and is not linked to BRCA mutation status. Nonetheless, BRCA genes are inactivated through allelic deletions and expression modulation in ovarian cancer. Inactivation of *TP53* also suppresses BRCA1 expression [58]. BRCA1 interacts with ATM, directing it to phosphorylate p53 to enable p21 induction and G1/S phase cell cycle arrest [59, 60]. Without these complementary functions, tumors spontaneously form in BRCA1−/− p53+/− mice [61]. Centrosome duplication correlates with BRCA1 deletion [60]. This complex is thus a critical factor determining proper chromosome segregation during mitosis, and upon centrosome duplication aneuploidy is assured to form upon cell division [62, 63]. This is facilitated by prior mutations of p53 that abolish its checkpoint control function, again stressing the role for early mutation of TP53 in SOC.

coordination with dozens of other proteins to perform genome maintenance via homologous recombination [4]. The double-stranded break repair pathway begins with PARylation of the break site by PARP1, megabases of phosphorylation of H2AX and subsequent formation of Rad51 filaments. Brca1 & Brca2 bind Rad51 to stimulate strand invasion of sister chromatids during homology directed repair. While only ~10% of SOC are mutated in BRCA1 or BRCA2, it is noteworthy that 75% of patients have lost one of two alleles of BRCA1 and 57% have lost an allele of BRCA2. Very few tumors (~1.5%) have homozygous deletions in BRCA1 [12], suggesting a system of compromised (but not lost) function. In fact, mRNA expression level does not track linearly with such deletions. It remains somewhat unclear if these monoallelic

Mutations in homologous repair coordinating factors are often found in serous, clear cell, endometrioid, and carcinosarcoma ovarian cancers [47, 48]. Specific mutation patterns are found within BRCA1/2 or otherwise homologous repair deficient cells. Without functional homologous repair, cells default to non-homologous end joining (NHEJ) to repair doublestranded DNA (dsDNA) lesions. NHEJ does not perfectly repair DNA, but rather often introduces small insertions or deletions along with single-nucleotide variants at the break site. These mutational marks are frequently found in non-serous ovarian cancers, yet are unlikely the main drivers of SCNA instability in serous ovarian cancer. However, NHEJ factors involved in repairing unresolved dsDNA breaks across different chromosomes, or creating translocations and other complex rearrangements, are compromised in 40% *ex vivo* ovarian cancer isolates [13, 49]. This can lead to resistance to PARP inhibitors and promote inappropriate translocation or "repair" events [50]. Such defects in NHEJ may explain why the majority of SOC initially respond to cisplatin-based chemotherapy. Complex dsDNA lesions incurred

Genetic and epigenetic changes alter *BRCA1* and the homologous repair pathways in SOC. Gene breakage is commonly observed within *RB1*, *NF1*, *RAD51B*, and *PTEN* [13]. While suppression of homologous recombination may lead to initial disease formation, there is evidence of *BRCA1/2* reversion mutations in tumors which become chemoresistant. This phenomenon follows the strong selective effects of carboplatin and taxanes which require cellular DNA repair pathways to enable cell division. Clinically, these findings should be considered in the context of the search for patient populations for PARP inhibitors. A primary hypothesis for how PARP inhibitors like Olaparib and Niraparib work is by targeting a DNA repair pathway which compensates for homologous repair, thereby presenting synthetic lethality specifically in cancer cells [52]. In a Phase III trial of Niraparib, clinicians treated both BRCA1 or BRCA2 mutant tumors as well as patients who were not found to have mutations in homologous repair genes in their tumors. Unexpectedly, all groups responded to Niraparib therapy [53], although patients with mutated BRCA1/2 or otherwise were defective in homologous repair were further delayed in cancer progression. While this is certainly an exciting development in the treatment possibilities for SOC patients, some caution is warranted. Reversion mutations enabling resistance to Niraparib may actually confer resistance to subsequent che-

Loss in BRCA1 enables microsatellite instability in mouse models and in colorectal cancer [55], though not in ovarian cancer [56, 57]. Microsatellite instability directly leads to centrosome

deletions do have a phenotype under genotoxic stress in human cells.

118 Ovarian Cancer - From Pathogenesis to Treatment

by cisplatin target NHEJ and promote mitotic catastrophe [51].

motherapeutics normally used upon disease recurrence [54].

Coincident SCNA events enable subsequent SCNA catastrophe. BRCA1 is located within kilobases of the neighboring autophagy gene BECN1. Autophagy is a critical catabolic infrastructure that enables cellular survival, requiring only 10% or less of normal autophagy gene dose [64]. Monoallelic loss of *BECN1* promotes centrosome amplification and aneuploidy among apoptosis-resistant cells [65]. An activator of the BECN1–PIK3C3 autophagy initiating complex, UVRAG, displays a similar phenotype [66]. Remarkably, this centrosome amplification promotes cell migration independent of its function in aneuploidy via indirect, hyperactivation of Rac1 [67]. LC3B, a ubiquitin-related protein which marks autophagosomal membranes, also acts in microtubule quality control [68]. One allele of the LC3B gene, *MAP1LC3B*, is genetically deleted in over 75% of SOC, independent of changes to *BRCA1* and *BECN1*. These three aneuploidy-accelerating lesions are likely to play key roles in serous ovarian cancer tumor initiation: mutant p53 expression, along with *BRCA1* and *BECN1* loss. Additional tumor suppressors may synergize to foster genomic instability: *NF1* is on the same chromosome arm as *BRCA1* and *RB1* lies nearby to *BRCA2*. Cumulative haploinsufficiency associated with prime targets and nearby neighbors certainly contribute to SOC aneuploidy [9] (**Figure 3**).

**Figure 3.** A model consistent with the ubiquitous mutation of p53 (SNV, indicated by green text) together with early SNV or haploid loss (indicated by light blue text) of repair enzymes, permitting a self-sustaining cascade of SCNAs and subsequent *in situ* selection for the observed SOC phenotypes.

In summary, *BRCA1/2* and homologous repair components are often suppressed by genetic deletions in SOC. This leads to further increases in SCNA formation and potentially independent metastatic phenotypes. However, SOC patients may benefit from the cancer's reliance on DNA repair pathways, as inhibition of PARPs prolongs progression free survival.

known, simply, as autophagy. The term autophagy ("*self-eating*") appropriately defines the process that cells use to recycle various macromolecular components, such as protein aggregates, lipids, and even entire organelles [75]. Autophagy is a primary method for the cellular catabolism, complementing turnover of proteins by the ubiquitin-proteosome system. Autophagy is described in terms of flux, which is the throughput of cellular '*detritus'* into autophagosomes, their transport to lysosomes, and subsequent enzymatic digestion. Ovarian cancer autophagy deletions impact the process primarily through deletions in *BECN1* (>75% of serous ovarian cancers) and in *MAP1LC3B* (>80% of serous ovarian cancers), though other genes are frequently affected. The average SOC tumor is 1 N across at least five different alleles; 95% of all serous ovarian cancers are deleted in BECN1 or *MAP1LC3B* and two others. The deficiency is as characteristic of SOC as p53 mutation. In addition to their roles in regulating chromosomal instability outlined above, *MAP1LC3B* and BECN1 (with the class three PI3 Kinase VPS34) play key roles in the formation of the early autophagosome, the phagophore,

Genomic Copy Number Alterations in Serous Ovarian Cancer

http://dx.doi.org/10.5772/intechopen.72695

121

Given this critical cellular function, we considered it counter-intuitive that cancer cells would delete a wide array of autophagy genes. In fact, KRAS mutant cancers have been described as "addicted" to autophagy, particularly in hypoxic or otherwise nutrient-stressed microenvironments [78]. This interpretation has been debated [79, 80], but the fact that autophagy has been established as a tumor suppressor system [81, 82], it is not exclusive of the possibility that specific tumor genotypes can promote addiction to autophagy [78]. Mono-allelic losses in the autophagy gene *BECN1* (homozygous deletions occur in only 0.9% of SOC cases) potentiate early development of tumors in mouse models [83, 84]. In this context, it is not at all counterintuitive to consider that these gene losses likely synergize with defects in the BRCA1/2 pathway, the p53 pathway, and other initial SCNAs, thus producing the unique extreme level of aneuploidy associated with SOC. Moreover, the loss of gene copies does not completely "turnoff" autophagy. In fact, ovarian tumor cells, like other cells, require autophagy to provide clearing of protein aggregates, metabolic byproducts (especially in hypoxic environments), and possibly even to permit cell division, given their aneuploid state and relative chromosomal instability. This, in turn, may provide a second selection criteria for depressed autophagy. Autophagy is induced by missegregating chromosomes, chromosomal instability is a hallmark of SOC, and extreme induction of autophagy can promote cell death [85]. Therefore, it may be more appropriate to use the term "disrupted" rather than "suppressed" to define how ovarian cancer autophagy varies from that seen in normal somatic cells. The state renders SOC sensitive to agents that perturb autophagy by inhibiting the autophagic flux, or via the creation of proteotoxic stresses which must be resolved by autophagy (as discussed below).

Interestingly, a number of other proteostasis control pathways were suppressed in serous ovarian cancer, and foremost among these is complementary to autophagy, the ubiquitinproteasome system. The core subunits, encoded by *PSMA1*, *PSMB1*, and *PSMC1*, are monoallelically deleted in 49, 62, and 41% of patients, respectively. Interestingly, the most interactive and deleted components of the proteasomal degradation pathway in ovarian cancer are enriched for cell cycle control related E3 ligases, including Park2, Fzr1, and Ube2d3. This suggests that not only is the core recycling process partly compromised by the core component

and recruitment autophagosome expansion proteins [76, 77].

**6.2. Proteosome**
