Mutant Allele Specific Imbalance in Oncogenes with Copy Number Alterations: Occurrence, Mechanisms, and Potential Clinical Implications

Abstract

Mutant allele specific imbalance (MASI) was initially coined to describe copy number alterations associated with the mutant allele of an oncogene. The copy number gain (CNG) specific to the mutant allele can be readily observed in electropherograms. With the development of genome-wide analyses at base-pair resolution with copy number counts, we can now further differentiate MASI into those with CNG, with copy neutral alteration (also termed acquired uniparental disomy; UPD), or with loss of heterozygosity (LOH) due to the loss of the wild-type (WT) allele. Here we summarize the occurrence of MASI with CNG, aUPD, or MASI with LOH in some major oncogenes (such as EGFR, KRAS, PIK3CA, and BRAF). We also discuss how these various classifications of MASI have been demonstrated to impact tumorigenesis, progression, metastasis, prognosis, and potentially therapeutic responses in cancer, notably in lung, colorectal, and pancreatic cancers.

1. INTRODUCTION

An oncogene is a gene that, when activated by mutation, increases the selective growth advantages of the cell in which it resides. Oncogenes often encode proteins that control cell processes such as proliferation and survival. These proteins include transcription factors, chromatin remodelers, growth factor receptors, signal transducers, and apoptosis regulators. Mutation, structural rearrangement (chromosomal rearrangement, gene fusion), and gene amplification (or copy number gains, CNGs) are common mechanisms that activate an oncogene, and result in an increased or a deregulated expression and/or function of the gene. Therefore, cells with such alterations in oncogenes possess a growth advantage or an increased survival rate. Translocations and mutations typically occur early in tumorigenesis, whereas copy number changes usually occur during late tumor stages.

Until recently somatic mutation and gene amplification were considered two independent and largely mutually exclusive mechanisms of oncogene activation. In general, a single copy of the mutated allele is sufficient to convert a proto-oncogene into an activated oncogene, while amplification often involves the non-mutated, wild-type (WT) allele. For example, KRAS is activated via point-mutations while HER-2/neu and MYC family genes gain their oncogenic properties largely through chromosomal aberrations and amplification [1; 2; 3; 4]. With the advances in genome-wide analysis, through single nucleotide polymorphism (SNP)-based array techniques and sequencing of the exome or whole genome, identifying genes altered in cancer at base-pair resolution with copy number counts has become routine. These advances have led to growing recognition of mutant allele specific imbalance (MASI) with CNG [5], acquired uniparental disomy (aUPD) (equivalent to MASI without CNG) [6; 7; 8], and MASI with loss of the WT allele in oncogenes [9], and their significance in tumorigenesis. While it remains true that activating somatic mutations in one allele of an oncogene is sufficient to confer a selective growth advantage on the cell, MASI with CNG, aUPD, and MASI with loss of the WT allele in oncogenes have now been reported to contribute to enhanced malignancies of tumor cells and therapeutic susceptibilities (Figure 1 and Table 1).

Figure 1. Mutant allele specific imbalance (MASI) with or without copy number alteration.

The diagram depicts three commonly observed MASI classifications that involve copy number gain (CNG), copy number neutral alteration (equivalent to uniparental disomy; UPD), or loss of heterozygosity (LOH) due to the loss of the wild-type (WT) allele.

Table 1.

Copy number alterations in oncogene activation

Classifications	Mechanisms
Amplification	CNG of WT, no Mutation
Balanced	Mut: WT = 1: 1, no CNG
MASI or MASI with CNG	Relative increase of the Mut allele; Mut: WT > 1 due to CNG of the mutant allele
aUPD, MASI without CNG, or copy-neutral LOH	MASI with copy neutral changes, Mut: WT= 2:0
Reverse MASI	Relative increase of the WT allele; Mut: WT < 1 due to CNG of the WT allele
MASI with LOH or MASI with loss of the WT allele	Relative increase of the Mut allele; Mut: WT > 1 due to the loss of the WT allele

The EGFR pathway contains numerous well-investigated oncogenes, such as the EGFR, KRAS, BRAF, and PIK3CA genes, with activating alterations in various tumors including lung, colorectal (CRC), and pancreatic ductal adenocarcinoma (PDAC) [10; 11; 12; 13; 14; 15]. These three cancer types are responsible for >410,000 cases (~24%) of cancer incidence and ~250,000 (42%) cancer deaths in the USA estimated for 2016 [16]. KRAS is a major oncogene that is frequently activated in PDAC (>90%) [17; 18], CRC (>40%) [19; 20], and lung cancer (predominantly in lung adenocarcinoma where it is the most commonly mutated oncogene at >20%) [21; 22; 23]. BRAF and PIK3CA are also often activated by mutations in CRC (18% and 32% respectively) [11; 12; 24; 25], occasionally in lung cancers (3% and 4%) [11; 25], and in a subtype of PDAC (those associated with IPMN; intraductal papillary mucinous neoplasm) [13; 26]. Activating mutations of the EGFR gene are present in 15–30% of NSCLC (non-small cell lung cancer), more frequently in adenocarcinomas (>30%), women (~37%), and non-smokers (~50%), while they have been rarely detected in other types of human cancers [27; 28; 29]. EGFR CNGs have also been reported in NSCLC and may play a role in survival and in response to tyrosine kinase inhibitor therapy [10; 30; 31], while KRAS CNGs have not been investigated in-depth in clinical tumors including NSCLCs. Here we discuss how activating alterations of these oncogenes via mutation, mitotic error, and/or copy number alteration, resulting in MASI with CNG, aUPD, or MASI with loss of the WT allele, may impact tumorigenesis, progression, and potentially therapeutic responses in cancers.

2. MUTANT ALLELE SPECIFIC IMBALANCE (MASI) WITH CNG

Activating somatic mutations with nucleotide changes (Figure 1 & Table 1, balanced) and gene CNGs due to focal amplification or chromosomal polysomy (Figure 1 & Table1, amplification) are two major categories of oncogene activation, which occur via independent mechanisms, and are largely but not completely mutually exclusive in tumor cells [32].

EGFR is a tyrosine kinase (TK) receptor of the ErbB family that is commonly altered in epithelial tumors. The EGFR can induce cancer via at least three major mechanisms: overexpression of EGFR ligands, amplification of EGFR, and mutational activation of EGFR. Mutations that have been described to activate EGFR include many variants of small mutations, insertions, and deletions, leading to enhanced dimerization/enhanced ATP binding/pathway activation [27; 28; 29]. Notably, mutations in the tyrosine kinase (TK) domain of the EGFR are unique to NSCLC and rarely detected in other tumor types including CRC and PDAC [27; 28; 29]. While most tumors are heterozygous for EGFR mutations, studies suggest that these genetic alterations are often coupled with gene amplification [27; 33; 34]. For example, analysis of electropherograms discovered ~40% of tumors with mutations have the mutant allele equal to or greater than the WT allele, indicating gene amplification of the mutant allele [27]. In some instances, the mutations were presented as homozygous on the electropherograms, with no detectable WT sequence, which implied the amplification of the mutant allele or the loss of the WT allele [33; 34].

KRAS is a member of the Ras family GTPases that regulates cell growth, differentiation, and survival [35]. KRAS is a major oncogene for PDAC, CRC, and lung cancer and is activated mainly via somatic mutations [17; 18; 19; 20; 21; 22; 23]. KRAS amplification is rare in comparison, although it has been detected in gastric, lung, and uterine cancers [32]. Conversely, unlike in other cancer types, 11% of ovarian cancers harbor KRAS amplification, a much higher rate than KRAS mutations in this cancer type [32], suggesting that amplification of WT KRAS may be an independent cancer driver in disease subtypes [36]. KRAS mutation and WT KRAS amplifications are largely mutually exclusive in these cancer types [32]. Amplifications of the mutant KRAS allele have also been reported decades ago [37; 38; 39; 40], but the term mutant allele specific imbalance (MASI) was only recently coined by Soh et al. to describe these preferential copy number alterations specific to the mutant allele in oncogenes [5] (Table 1).

Using both SNP array analysis and gene specific assays, Soh et al demonstrated that complete MASI (homozygous mutations, with or without CNG) is a frequent occurrence in KRAS (36%) and EGFR (29%), rare for PIK3CA (6%), and intermediate at the BRAF locus (17%) in a panel of 833 cell lines of 12 tumor types [5]. Among the cancer cell lines with mutations, MASI is a frequent event in mutant EGFR (75%) and was due mainly to CNGs, while MASI in mutant KRAS (58%) was due to UPD [5]. MASI is less frequent in cancer cells with mutant BRAF (38%) and was more associated with CNGs although not conclusively. PIK3CA mutation is rare and MASI frequency among those with mutant PIK3CA is equally rare (8%). For PIK3CA, CNGs without mutation were the most frequent change [5]. The cause for the observed variations in MASI frequency among the oncogenes is unknown.

Subsequent studies confirmed that KRAS MASI is associated with selective amplification of the KRAS mutant allele at 4–18% lung cancer cases examined and correlates with poor prognosis [41; 42; 43]. Recently it has been demonstrated in vivo that mutant Kras copy gain can result in the metabolic reprogramming and increased malignancy in lung tumor cells [44]. KRAS MASI has also been reported in 18.4% of PDAC and is associated with the progression to undifferentiated carcinoma of the pancreas [45]. In CRC, KRAS MASI has been detected at 12.8–55%, is more frequently associated with G13D mutation [46; 47], and has been demonstrated to be an independent adverse prognostic factor in one study [47]. In in vitro studies, simulated KRAS MASI (with increasing amount of the mutant KRAS plasmid vector) was shown to reduced treatment responses to cetuximab treatment [46].

Preferential amplification of the mutant EGFR allele in lung cancer has also been confirmed by independent studies [48; 49]. In a report on lung adenocarcinoma, EGFR MASI was detected in 26% of the cases examined, more commonly associated with exon 19 mutations than with exon 21 mutation, and also associated with poor disease-specific survival [48]. In another study on NSCLC, EGFR MASI was detected in 37% of the cases and associated with exon 19 deletion [49]. The different techniques employed to evaluate MASI, intertumor and intratumor heterogeneity, and variations in sample collection (such as the portion of early stage vs. advanced tumors in a given study) may all account for the disparities among the reported EGFR or KRAS MASI frequencies. However, EGFR MASI does not appear to play a role in therapeutic responses to first-generation EGFR small molecular inhibitors [49], TTP (time to progress), or OS (overall survival) [50]. It’s possible that such a difference is difficult to be discerned statistically because either EGFR mutation or EGFR CNG (most studies refer to toal CNG without making a distinction between WT and mutant EGFR) alone is already associated significantly with better clinical outcomes to treatment of EGFR inhibitors [30; 50; 51; 52]. Given the rarity of the EGFR mutation, non-exclusivity of KRAS and EGFR CNGs, the lack of genomic data from large clinical trials, to thoroughly unravel the predictive value of EGFR MASI on therapeutic responsiveness and patient prognosis may require a global collaboration.

BRAF mutation is not a frequent event for lung, colorectal, or pancreatic cancers, therefore there is no BRAF MASI-oriented report in these cancer types to-date. BRAF is a major oncogene for melanoma. Recently Pfarr et al. reported that even though BRAF was mutated at high frequency in the melanoma examined (36%), only 3% harbored co-occurring mutations and amplifications of BRAF [43]. This is consistent with the previous report of Soh et al that BRAF MASI frequency is far lower comparing to KRAS and EGFR MASI, and may not be a significant mechanism for regulating BRAF signaling in cancer.

It’s important to note that in the majority of the studies on MASI, there were no vested efforts made to differentiate those MASI with CNGs from MASI without CNGs (also called aUPD) or MASI with loss of the WT allele, which will be discussed below. Given that MASI may correlate with prognosis and/or treatment responses, extra care to discern various types of MASI in future research will be highly desirable to appreciate the significance of MASI in clinical correlative studies.

3. ACQUIRED UNIPARENTAL DISOMY (aUPD) or MASI WITHOUT CNG

The development of genome-wide approaches, especially high resolution SNP arrays, enables evaluation of dynamic chromosomal as well as focal changes of copy number alterations and loss of heterozygosity (LOH) with high resolution, and thus allowing the identification of aUPD in cancer cells. Whereas LOH has been most commonly referred to gross deletions of chromosomal material encountered in cancer, aUPD is equivalent to copy-neutral LOH or MASI without CNG (Figure 1 and Table 1). Both are somatic events, with LOH due to deletion results in hemizygosity, while aUPD results in homozygosity.

UPD was initially defined to describe the inheritance of a pair of chromosomes/segments from only one parent with regard to development and developmental disorders. UPD was first reported by Engels et al in 1980 proposing that constitutional UPD was likely to occur due to the high rates of meiotic error [53]. Recent advances in molecular genetics have permitted the precise mapping and frequency of UPD to be assessed. It is now known that the extent of UPD can range from a small chromosomal segment to an entire chromosome, as proposed by Engel. The severity of the resulting phenotype of each germline UPD varies greatly. UPDs that are expected to lead to early lethality in zygotic development would go undetected. UPDs without discernable phenotypes will also go undetected. As a result, constitutional UPDs are often identified through studies of inherited diseases.

Acquired UPDs have been revealed in a variety of cancer types comparing constitutional and tumor DNA genotypes, with most studies being performed on hematological malignancies [6; 7; 8], but also in some solid tumors including CRC [54; 55], lung cancer [5], PDAC [56], pancreatic endocrine tumors [57], breast cancer [58; 59; 60], basal cell carcinoma [61], retinoblastoma [62], neuroblastomas [63], and clear cell renal carcinomas [64]. These studies have revealed that aUPD can occur in almost any chromosome, but it is becoming evident that aUPDs are non-randomly distributed with cooperation occurring between the aUPD and gene mutations. These genes can be homozygous for mutations, and mutated genes can be inactivated tumor-suppressor genes or activated oncogenes. Affected genes in regions of aUPD to-date include but not restricted to KRAS, EGFR, BRAF, MPL, MSH2, MAP2, MLH1, FHIT, TET2, APC, HLA-A/B/C, A20, EZH2, JAK2, CDKN2A, PTCH, WT1, H19, IGF2, H19, HRAS, CDKN1C, CBL, FLT3, miR-15a, miR-16-1, RB1, BRCA2, p53, NF1, BRCA1, CEBPA, RUNX1,JAK, IRF8,TNFRSF14, VPS39,PPM1D, PPM1E, C17orf71, SLCA3R1, TRIM37, PIK3CA, PTEN, CDH1, TPM3, MUC1, THBS3, CBLB, MAF, and FBXW7 [5; 6; 7; 8; 59; 60; 65; 66; 67; 68].

The underlying mechanism for aUPD is not completely understood. It has been proposed that aUPD can result from mitotic homologous recombination events, or it may represent an attempt to correct for the unbalanced loss of chromosomal material by using the remaining allele as a template. It is thought that if aUPD occurs in only a segment of the chromosome, it does so probably through mitotic recombination between identical low copy repeats in the G2 phase of the cell cycle. If aUPD involves the entire chromosome, the aUPD probably arises from a chromosomal segregation error in mitosis, in which one allele is lost in anaphase lag and the remaining allele is reduplicated [7; 8].

Correlations between somatic UPDs and clinical outcomes have been reported. In breast cancer, aUPD regions are often associated with poor overall survival, although they are not consistent adverse predictors for metastasis-free survival [60]. Higher frequency of aUPD has also been reported in triple-negative breast cancer than Her2/neu-positive and/or ER or PR-positive cases [59]. In serous ovarian cancer, aUPD is a common event and some recurrent loci are associated with a poor outcome [69]. Both better and inferior clinical outcomes have been reported to be associated with different regions of aUPD in acute myeloid leukemia [70; 71].

Despite the recognition of aUPD in oncogenes, aUPD in the EGFR pathway has been scarcely described to-date. In the study of lung cancer, CRC, and PDAC by Soh et al [5], it was demonstrated that KRAS aUPD contributed to 55% of all KRAS MASI observed, while MASI at the EGFR, BRAF, and PIK3CA were largely caused by CNG (78%, 67%, and 100%, respectively). The cause for the disparities among the oncogenes remains unexplored. Chiosea et al also reported that 47% of KRAS MASI in lung adenocarcinoma was due to CNGs, implying that the remaining cases (53%) might have harbored KRAS aUPD [41] and therefore is consistent with the previous finding. In juvenile myelomonocytic leukemia (JMML), KRAS aUPD was thought to facilitate aggressive transformation from an indolent course to fatal malignancy in a case study [67]. aUPD at the PIK3CA, FGFR3, and CDKN2A genes were thought to contribute to bladder tumorigenesis [72]. As mentioned previously, although MASI at KRAS or EGFR have been reported to be adverse prognostic markers in lung, CRC, and PDAC [41; 42; 45; 47; 48; 49], those studies failed to make a distinction between MASI with or without CNG, therefore the clinical significance of aUPD at KRAS and EGFR remains to be determined in those cancer types. Evaluation of oncogene aUPD frequency and their potential impacts on prognosis, survival, and treatment responses is much anticipated and should be addressed in future studies.

4. MASI with loss of wild-type allele or MASI with LOH

We have recently detected loss of the WT Kras allele in a genetically-engineered mouse model (GEMM) that we have developed for PDAC [9]. The GEMM, with the genotype of p16^lox/lox;LSL-Kras^G12D;Pdx1-Cre (referred to as PKP mice), uniformly develops precursor lesions (pancreatic intraepithelial neoplasia, PanIN) that progress to PDAC and eventually metastasis, mimicking human PanIN/PDA development and progression to metastasis at both genetic and histologic levels. From the clonal cell lines derived from both primary pancreatic tumors and metastases developed in the PKP GEMM, we detected loss of the WT Kras allele (MASI with LOH at Kras). Intriguingly the frequency of MASI with LOH at Kras is higher among those cancer cell lines derived from metastases then those from primary pancreatic tumors [9]. MASI with LOH at Kras did not appear to be a random event because it resulted in discernible functional advantages shown by colony formation, cell proliferation, and motility assays [9]. Using real-time PCR, we eliminated the possibility that differential amplification of the mutant or WT Kras allele contributed to the differences observed [9]. We further confirmed the functional advantages resulted from the loss of WT KRAS allele in isogenic colorectal cancer cell lines [73; 74] (KRAS^G12D/WT vs. KRAS^G12D/−genotypes) using colony formation, cell proliferation, and motility assays (data not shown). To ascertain that these results are not artifacts of tissue culture adaption, we microdissected liver metastases from both our mouse models and human specimens to confirm that MASI with LOH at Kras indeed occurred in vivo [9].

Non-biased LOH profiling comparing cancer cell lines derived from human primary pancreatic tumors (n=19) and metastases (n=10) was performed using SNP chip analyses. LOH at chromosome 12p, which contains KRAS, was observed in 37% of primary and 80% of metastatic cancer cell lines (p<0.02) and is the singular chromosomal arm that showed statistical difference between the two LOH profiles in this whole-genome scanning study [9]. The lack of significant difference in allelic loss on other chromosomes, indicates that this event at chromosome 12p likely occurs selectively, and is not a random manifestation of increased genomic stability during progression [9]. These data corroborate the observations made in our PKP mice and indicate that MASI with LOH at Kras is a selective event that occurs in vivo, confers growth advantage to tumor cells, and may promote metastasis.

Prior to our discovery, LOH at chromosome 12p (where KRAS resides) had been reported to correlate with the presence of KRAS mutations in human lung cancer, CRC, PDAC, and prostate cancer [75; 76; 77; 78; 79; 80]. Loss of the WT Kras allele was also found associated with Kras activation in a spontaneous lung cancer mouse model [81; 82]. Increased loss of the WT Kras allele and elevated Ras signaling also correlated with high-grade tumors in a lung cancer mouse model [83], suggesting that the loss of the WT Kras allele may facilitate tumor progression. Heterozygous deletion of the WT Kras allele was shown to be sufficient to promote tumorigenesis in vivo [84; 85]. In vitro, the reintroduction of the WT KRAS allele to colonic cancer cells harboring oncogenic KRAS resulted in growth inhibition and altered expression profiles in cell proliferation, metabolism, and transcriptional control [86; 87]. Recently, spontaneous loss of the WT Kras allele has also been reported in a T-ALL GEMM, in which the authors also demonstrated that restoration of the WT Kras protein expression in vivo has tumor-suppressive effects [88]. These results are independently verified by another related study on oncogenic Kras-driven leukemia, which demonstrated that genetic or epigenetic loss of WT Kras expression promoted tumor growth and shortened survival [89]. Our data echo these observations, but take them a step further and offer the first potential explanation for the selective loss of the WT allele- the loss of the WT KRAS allele in the context of mutant KRAS, which is akin to MASI with LOH (Table 1, Figure 1), may promote metastasis in mice and humans [9]. Our data also is supported by the findings of Martins and colleagues that the loss of the WT Kras is a selected event in tumorigenesis and that mutant Kras lung tumors are not a single disease but rather a heterogeneous group comprising two classes of tumors based on their mutant Kras allelic counts (or the WT Kras status) which contribute to their distinct metabolic profiles, prognosis, and therapeutic susceptibility [44; 83]. Together, these reports provide evidence in favor of the tumor-suppressive role of WT KRAS.

Over the decades, there has been a considerable amount of speculations on the mechanisms underlying the tumor-suppressive function of WT KRAS [90; 91; 92]. One potential mechanism is that WT RAS can bring about growth suppression by contesting for the same targets as oncogenic Ras or by interacting with an unexplored downstream target. Alternatively, normal RAS may compete with the oncogenic form, either for unknown regulatory events, or for downstream effectors. Presently, there is no consensus supporting a particular mechanism or downstream pathway from published literatures. For example, an inverse correlation between WT Kras expression and ERK activity was reported in one study and was offered as a possible molecular mechanism for the inhibitory effect of WT Kras on cellular transformation [84]. Other studies revealed that the disruption of WT RAS isoforms not cognate to the mutant RAS would compromise MAP/ERK signaling, suggesting WT RAS isoforms serve as tumor promoters in this scenario [93; 94; 95]. It has also been proposed that WT RAS may exert growth inhibition by binding to tumor suppressive RASSF1A-Nore heterodimers to activate downstream pro-apoptotic genes [96]. However, conflicting data revealed that WT KRAS can antagonize mutant KRAS-induced apoptosis via the RASSF1A-MST2-LATS1 pathway [97; 98]. This area of research remains to be further elucidated and is essential to secure the tumor-suppressive role for WT KRAS.

As discussed previously, it has been reported that MASI at KRAS is associated with worse prognosis in pancreatic and lung cancers and the observed MASI is presumably due to amplification of the mutant allele [42; 45]. Acquired UPD has been considered to be associated with tumor-suppressor genes, and until recently, rarely with oncogenes [5; 6; 99; 100]. However, we now have evidence indicating that in addition to MASI with CNGs and aUPD, MASI with LOH at KRAS also contribute to the observed MASI at the KRAS locus. Using a combination of direct genomic sequencing (mutation and MASI are determined by electrophergram), Q-PCR (provides total copy number of KRAS alleles), and Taqman Mutation Analysis Assay (provides ratio of mutant to WT KRAS alleles), we are able to differentiate between MASI resulting from amplification of the mutant allele alone (MASI with CNG) and MASI with LOH at KRAS (loss of the WT KRAS allele with or without amplification of the mutant allele) (Figure 2A). Applying this two-step analysis in a pilot study to 64 lung cancer samples with known KRAS mutations at the Columbia University Medical Center, we found that 29.4% of primary tumors and 76.9% of metastases displayed MASI with LOH at KRAS (p=0.003) (Figure 2B), further confirming that MASI with LOH at KRAS is a significant event in tumor progression and is associated with metastasis. The impacts of MASI with LOH at KRAS on other clinical parameters, such as survival or therapeutic responses, remain to be investigated.

Figure 2. Copy number analyses for mutant vs. wild-type alleles of KRAS at hot-spot G12D.

MASI with LOH at KRAS was determined by a combination of the direct genomic sequencing (electropherogram), Taqman mutation analysis assay, and Q-PCR. A) Pancreatic cancer cell lines with G12D mutation with no LOH, G12V mutation (negative control), and G12D mutation with LOH (positive control) were used to develop the assay. The LOH status for each formalin-fixed, paraffin-embedded (FFPE) patient sample was then assessed using this combinative assay. Three representative cases are presented here: patient #1 had MASI with CNG, and #2 and #3 had MASI with LOH. B) Breakdowns of MASI with or without LOH present in primary lung tumors vs. metastases. MASI with LOH at KRAS is a significant event in lung cancer and is associated with metastasis (p=0.003).

5. CONCLUSION

The recognition of MASI and its classifications allows a common language to describe genomic changes pertaining to the allele specific (mutant and/or WT) alterations in oncogenes. While it is now indisputable that MASI, aUPD, and MASI with loss of the WT allele are common occurrence in cancers, their functional or clinical implications are less certain. In general, MASI is an adverse prognostic marker in numerous cancer types examined. We and colleagues have demonstrated that MASI with loss of the WT allele can enhance or promote malignant growth. We have also demonstrated that MASI with LOH at KRAS is associated with metastasis. aUPD is not unique to oncogenes; in fact, it is more consistently offered as a mechanism to inactivate tumor-suppressor genes, such as p16, p53, Rb. Therefore, the frequency of aUPD in oncogenes and their implication in tumorigenesis are scarcely described. Overall, most of the MASI studies to-date do not make a distinction on MASI with CNGs, MASI without CNGs/aUPD, or MASI with loss of the WT allele. The value of discerning various types of MASI in clinicopathological correlative studies may seem uncertain, however, given that genetic mutation profiles alone do not always predict treatment responses of target therapies based on genetic mutations, and MASI has been shown to associate with treatment responses in some early studies, additional efforts to distinguish various types of MASI may be warranted to generate more precise data in future clinical correlative analyses.

Investigations on the mechanisms and functional impacts of MASI would be equally anticipated in the future research. Whereas it may be conceivable to image amplification of the mutant allele of an oncogene would promote tumorigenesis and correlate to worse prognosis, it has been more challenging to convince skeptics that the WT allele of an oncogene may possess tumor-suppressive function in the presence of the mutant allele. The identification of the underlying mechanism responsible for the profound effect of the WT KRAS allele on mutant KRAS-driven tumorigenesis would undoubtedly further cement its putative role in tumor suppression. The finding that WT KRAS does function as a tumor suppressor may come with significant implications for the development of therapeutics that target KRAS activity- it may be essential to design future target therapies to inactivate mutant form of KRAS specifically without affecting WT KRAS functions, because KRAS target therapies that do not distinguish between WT and mutant KRAS may inadvertently promote tumor progression and/or metastasis, and would have unintended devastating impacts on cancer patients.

Over the past decade, comprehensive sequencing efforts have revealed the genomic landscapes of common forms of human cancer. To date, these studies have revealed ~140 genes that, when altered by intragenic mutations can promote or “drive” tumorigenesis. As these efforts have matured, it would only be appropriate to consider other mechanisms of gene regulation that may contribute to tumor progression and metastasis in addition to intragenic mutation. While the concept of gene amplification is not new, the significance of secondary copy number alterations that could further regulate an activated oncogene and thus provide growth advantages to tumor cells has not yet been fully realized. The profound impacts of MASI on tumor progression and metastasis, patient prognosis and survival, therapeutic responses, and tumor recurrence would be worthy in-depth examinations in the near future.

HIGHLIGHTS.

• This review describes and explains various classes of mutant allele specific imbalance (MASI) pertaining to oncogenes and their significance in tumorigenesis.

• This review highlights the occurrence of MASI at the EGFR and KRAS loci and their impacts on tumor growth, progression, and metastasis, as well as their implications in patient prognosis and treatment responses.

• We review and discuss the putative tumor-suppressive role of the wild-type KRAS allele in the context of the activated mutant KRAS allele in cancer, notably in pancreatic and lung cancers.