The Cancer Genome Atlas: Clinical Applications for Breast Cancer

Publication
Article
OncologyOncology Vol 27 No 12
Volume 27
Issue 12

This article aims to provide an overview of The Cancer Genome Atlas findings, with a particular focus on their potential biological relevance and therapeutic implications.

Table 1: Six Platforms Used in TCGA's Breast Cancer Project

Table 2: Mutation Types

Table 3: Significantly Mutated Genes by Molecular Subtypes: Data From TCGA

Table 4: Highlights of TCGA Findings, Biologic Relevance, and Therapeutic Implications

Figure 1: Mutations in Estrogen Receptor (ER) Pathway Genes in Luminal Breast Cancer

Figure 2: Correlation of Mutations With Luminal A vs B Subtype and Aromatase Inhibitor Response

Figure 3: Simplified Diagram of Key Genetic Abnormalities in Breast Cancer

Figure 4: Epigenetic Regulation by Histone Acetylation or Methylation

Advances in next-generation sequencing technologies in recent years have allowed in-depth study of somatic mutations in over 1,000 breast cancer samples. The Cancer Genome Atlas (TCGA) is the largest single genome-characterization effort to date. It is remarkable for the integration of DNA sequencing with genome-wide profiling of the epigenome, microRNAome, transcriptome, and proteome for more than 500 diverse primary untreated breast cancers. This article aims to provide an overview of TCGA findings, with a particular focus on their potential biological relevance and therapeutic implications.

Introduction

It is well established that cancer is the result of genomic alterations that drive uncontrolled cell proliferation and metastatic spread. Understanding the relationships between the genomic landscape of cancer and the clinical characteristics of the disease has therefore become the foremost priority in cancer research. A key assumption is that this understanding will lead to new and successful therapeutic hypotheses. The completion of the Human Genome Project in 2003 and the rapid advances in high-throughput sequencing technologies have made it possible to comprehensively characterize cancer genomes in a rapid and affordable manner, with findings published for more than 1,000 breast cancers to date.[1-8] The Cancer Genome Atlas (TCGA) project, a large-scale collaboration funded by the National Cancer Institute (NCI) and National Human Genome Research Institute (NHGRI), generated “omics” data for over 500 diverse breast cancers at the levels of DNA (point mutation, copy number change, epigenetic modification), RNA (messenger RNA [mRNA] and microRNA [miRNA]), and protein (both protein and phosphoprotein analysis) (Table 1).[6] The integrated analysis of these data allowed identification of subtype-specific genetic, epigenetic, and proteomic alterations, and it provided a potential functional interpretation of the underlying biology. A therapeutic roadmap is emerging that is setting priorities for clinical trial development.[9] In this review, we will summarize key findings from TCGA and other similar studies and discuss challenges and strategies in the effort toward the goal of “individualized medicine” in the era of cancer genomics.

Findings From TCGA: Divergence of the Four Main Breast Cancer Subtypes

The molecular classification of breast cancer originates from the unsupervised hierarchical clustering analysis of complementary DNA (cDNA) microarray data by Perou et al, which revealed the existence of four major intrinsic subtypes: luminal A, luminal B, human epidermal growth factor receptor 2 (HER2)-enriched (HER2-E), and basal-like.[10] Luminal breast cancers have a gene expression signature that includes estrogen receptor 1 (ESR1), GATA-binding protein 3 (GATA3), forkhead box protein A1 (FOXA1), B-cell chronic lymphocytic leukemia (CLL)/lymphoma 2 (BCL-2), X-box binding protein 1 (XBP1), and the myeloblastosis gene (MYB), which are highly characteristic of luminal epithelial cells in the inner layer of a normal breast duct. Luminal B cancers differ from luminal A in their lower levels of luminal gene expression, higher levels of proliferation genes, and a worse clinical outcome.[11,12] Basal-like breast cancer is characterized by the expression of the basal gene signature that contains keratins 5, 6, and 17, and by high-level expression of cell proliferation–related genes.[10] In the absence of treatment, patients with basal-like breast cancer experience poor clinical outcome, particularly in the first 5 years after diagnosis.[11,12] Although basal-like breast cancers are often triple-negative for the estrogen receptor (ER), progesterone receptor, and HER2, basal-like disease and triple-negative breast cancer (TNBC) do not completely overlap.[13,14]

HER2-E breast cancers usually express high levels of HER2 and growth factor receptor–bound protein 7 (GRB7), the latter of which is also located in the HER2 amplicon on 17q21. While HER2-E breast cancers are often HER2-amplified, a small group of TNBC is classified as HER2-E without HER2 amplification, and some HER2-positive breast cancers are luminal or basal-like. Subsequent studies of larger sample sizes of breast cancer have indicated the presence of less common subtypes, including claudin-low,[15] immunomodulatory, and lumimal androgen receptor subtypes.[16]

To further explore this molecular heterogeneity, TCGA analyzed approximately 500 breast cancers on six platforms, including the Agilent mRNA expression microarrays (for mRNA expression), Illumina Infinium DNA methylation chips (for DNA methylation), Affymetrix 6.0 single nucleotide polymorphism (SNP) arrays (for DNA copy number), miRNA sequencing, whole-exome sequencing (for DNA sequencing of all coding regions of the genome), and reverse-phase protein array (RPPA). An integrated analysis of data from these different platforms demonstrated the divergence in genomic and epigenomic alterations of the four previously defined intrinsic subtypes of breast cancer. A glossary of different mutation types is provided in Table 2.

The four main breast cancer subtypes exhibited a striking difference in mutation spectra (Table 3). Luminal type breast cancers harbored the most diverse and recurrent significantly mutated genes (SMGs, genes with mutations occurring more frequently than the background mutation rate), despite a lower mutation rate overall compared with the basal-like and HER2-E subtypes; this finding suggests a causative role of these mutations in luminal breast cancers. Luminal A breast cancers were characterized by a high frequency of mutation in the phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha (PIK3CA) gene (45%); multiple SMGs, including mitogen-activated protein kinase kinase kinase 1 (MAP3K1), GATA3, cadherin 1 (CDH1), and mitogen-activated protein kinase kinase 4 (MAP2K4), which occur almost exclusively in this subtype; and a low frequency of TP53 mutation (12%). Compared with luminal A breast cancers, the luminal B subtype was associated with a higher rate of TP53 mutation (29%) and a slightly lower rate of PIK3CA mutation (29%). In contrast, 80% of basal-like cancers carried mutations in TP53 and were lacking other SMGs, except for PIK3CA (9%). The HER2-E subtype was characterized by HER2 amplification (80%), a high frequency of mutated TP53 (72%) and PIK3CA (39%), and a much lower frequency of other SMGs.

Luminal/ER+ breast cancer

Overall, luminal/ER+ breast cancer has the highest inter-tumor heterogeneity in terms of gene expression, mutation spectrum, and copy number changes. Several signaling pathways are characteristically altered in this subtype of breast cancer. These include the cellular apoptosis pathway, ER signaling, phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) signaling, mitogen-activated protein kinase (MAPK)/c-Jun N-terminal kinase (JNK) signaling, cyclin D–cyclin-dependent kinase (CDK)4/6-retinoblastoma (RB) pathway, tumor protein p53 (TP53)/mouse double minute 2 homolog (MDM2) pathway, and growth factor–receptor signaling pathways (Table 4).

ER pathway. Estrogen is the prime growth regulator of ER-positive breast cancer. Not surprisingly, unique to luminal subtypes are mutations in a set of genes, including GATA3 (14% in luminal A and 15% in luminal B), FOXA1 (2% each in luminal A and B), and the runt-related transcription factor 1 (RUNX1) gene (5% in luminal A and 2% in luminal B), that are important for the genomic action of the ER. FOXA1 and GATA3 were found to be mutated in a mutually exclusive manner. Mutations in these genes could potentially mark disease that is sensitive to endocrine therapy (GATA3 and FOXA1 mutations) or resistant to it (RUNX1 mutations).[4]

GATA3 and FOXA1 are both ER-regulated genes and play key roles in ER-mediated target gene regulation, luminal cell fate, and differentiation[17-20] (Figure 1). Together with ER, GATA3 and FOXA1 comprise the minimum gene set that defines luminal type breast cancer. These two transcription factors have their DNA binding sites within the ESR1-binding regions and facilitate ER binding to DNA.[21,22] Depletion of FOXA1 abolishes ER binding capacity and ER-mediated transcriptional activity,[21] while over-expression of FOXA1 and GATA3 in ER-negative cells leads to cellular reprogramming to establish an estrogen-dependent cell phenotype.[23]

GATA3 is the third most commonly mutated gene in luminal breast cancer, following PIK3CA and TP53. These are inactivating mutations that disrupt its nuclear localization and transactivation function, leading to redistribution of ER to target genes that are less dependent on GATA3.[7,8] GATA3 mutation did not appear to influence the baseline tumor Ki67 level but was associated with a significantly greater Ki67 reduction in response to neoadjuvant aromatase inhibitor (AI) therapy,[4] suggesting that GATA3 mutation could serve as a predictor of endocrine sensitive disease. Since FOXA1 and GATA3 mutations are mutually exclusive, mutation of FOXA1 could also be a marker of endocrine sensitivity.

RUNX transcription factors are also critical regulators of cell growth and differentiation but serve as cell context–dependent tumor suppressors or oncogenes.[24,25] RUNX1 and its dimerization partner, core binding factor beta (CBFB), regulate ER-mediated transcription by tethering ER to target genes without an estrogen response element (ERE)[26] (Figure 2). In MCF10A cells, RUNX1 silencing led to the formation of abnormal, hyperproliferative acinar structures in 3D cultures,[27] indicating the tumor-suppressing role of RUNX1 in breast cancer tumorigenesis. This was further supported by the finding of decreased RUNX1 expression in more aggressive breast cancer cell lines.[28] In breast cancer, mutations in RUNX1 are often frame shift insertions, deletions, or missense mutations that likely lead to loss of function.[29] RUNX1 mutation appeared to be associated with less responsiveness to AI therapy in neoadjuvant endocrine studies[4] (Figure 2).

Interestingly, translocations and mutations in RUNX1 have frequently been observed in hematopoietic malignancies.[30] Another example of a luminal SMG believed to play a role in leukemogenesis is SF3B1,[4,6] a splicing factor commonly mutated in CLL.[31] These examples reinforce the concept of luminal breast cancer as a stem cell disorder driven by defects in terminal differentiation and programmed cell death.

PI3K pathway. The PI3K/AKT/mTOR pathway plays a key role in mediating cell growth, proliferation, survival, migration, and angiogenesis (Figure 3). Its importance in ER+ breast cancer has been well established based on both preclinical and clinical evidence. In preclinical studies, dual inhibition of the PI3K pathway and ER led to synergistic cell killing of ER+ breast cancer.[32,33] In clinical trials, mTOR inhibition improved the progression-free survival of patients with advanced AI-resistant ER+ breast cancer,[34] leading to the US Food and Drug Administration (FDA) approval of everolimus in this setting. Inhibitors of PI3K, AKT, or mTOR kinase are being actively investigated in clinical trials.

TCGA identified PIK3CA, the alpha catalytic subunit of PI3K, as being the most common SMG in luminal breast cancer, occurring at a frequency of 45% and 29% in luminal A and luminal B subtypes, respectively. These are often missense mutations that cluster in the helical domain (HD) and the kinase domain (KD) of PI3-kinase and are capable of inducing cellular transformation when introduced into mammary epithelial cells.[35] Although mutations in PIK3CA were not particularly associated with increased activity of PI3K pathway signaling as assessed by the level of pAKT, tumor cells with mutations in PIK3CA have been shown to be highly dependent on p110 alpha for cell survival.[33] Therefore, direct inhibitors of PI3K, particularly the alpha-specific inhibitors, are of great interest. The alpha-specific inhibitors are still in early stages of clinical development, but promising data have been reported in a phase I study of BYL719 in PIK3CA-mutated tumors.[36] Pan-PI3K inhibitors have now entered phase II and III randomized clinical trials in combination with endocrine agents for the treatment of advanced ER+ breast cancer,[37] and several of these trials prospectively test the mutation statuses of the PI3K pathway components as stratification factors or eligibility criteria for patient enrollment. A neoadjuvant phase II trial of the AKT inhibitor MK-2206 in combination with anastrozole in newly diagnosed clinical stage II or III PIK3CA-mutant ER+ breast cancer, with the primary objective of pathologic complete response (pCR) to study treatment, is also ongoing (National Cancer Institute [NCI] ClinicalTrials.gov Identifier NCT01776008). The results of these studies will provide insight into whether PIK3CA mutation status predicts response to these agents.

In addition to mutations in PIK3CA, mutations have also been identified in PIK3R1, the gene that encodes the regulatory subunit of PI3K (0.4% in luminal A, 2% in luminal B), which clustered in the PI3-kinase interaction domain; AKT1 (4% in luminal A, 2% in luminal B); and PTEN (4% in luminal A, 4% in luminal B). Mutations in PIK3R1, PIK3CA, PTEN, and AKT1 were mutually exclusive (P = .025), suggesting that these mutations are functionally important. Correlative studies of the ongoing large phase II and III trials of PI3K inhibitors may provide further insight into the contribution of PI3K inhibitors to treatment response.

MAP3K1, MAP2K4, and the JNK and ERK pathways. MAP3K1, also known as MEKK1, is a serine/threonine kinase that regulates the ERK and JNK kinase pathways as well as nuclear factor–kappa-B signaling[38] (see Figure 3). In a genome-wide association study, a germ-line MAP3K1 polymorphism was associated significantly with familial breast cancer.[39] In TCGA’s report, about 13% of luminal A and 5% of luminal B breast cancers have mutations in MAP3K1, occurring mutually exclusively with mutations in MAP2K4 (7% in luminal A and 2% in luminal B), a serine/threonine kinase immediately downstream of MAP3K1 that activates JNK. Mutations in MAP3K1 and MAP2K4 include frame-shift deletions or insertions, and missense or nonsense mutations that are predicted to cause truncation or loss of function of the protein.[4] Almost all mutations in MAP3K1 and MAP2K4 occur in luminal breast cancer, indicating that they are likely driver events specific for this subtype of breast cancer. Interestingly, mutations in MAP3K1 were found to be associated with lower tumor grade and Ki67 level at baseline, indicating that mutations in this gene are associated with indolent clinical features[4] (see Figure 2).

The implications of MAP3K1 and MAP2K4 mutations in the management of ER+ breast cancer are not fully understood and are an important area of future research. In experimental models, disruption of MAP3K1 function was shown to protect cells from specific stress-induced apoptosis. Mouse embryonic stem cells deficient in MAP3K1 were found to have lost the JNK-mediated apoptotic response to a set of stimuli, including microtubule disruption and cold stress, but they retained a response to heat shock, anisomycin, and ultraviolet radiation.[40] However, cells deficient in MAP3K1 could have an enhanced apoptotic effect to oxidative stress. Minamino et al showed that MAP3K1-deficient mouse cardiac myocytes had heightened sensitivity to hydrogen peroxide–induced apoptosis due to enhanced tumor necrosis factor (TNF)-alpha production.[41] Further investigations are needed to elucidate (1) the mechanism by which MAP3K1/MAP2K4 contribute to tumorigenesis of luminal breast cancer; and (2) whether MAP3K1/MAP2K4 mutations determine sensitivities to treatments such as chemotherapy or targeted agents-and importantly, whether therapeutic opportunities exist for the MAP3K1/MAP2K4-mutated tumors.

Cyclin D-CDK4/6-Rb pathway. CDKs 4 and 6 regulate cell cycle progression by activating the Gap 1 (G1) to DNA synthesis (S) phase transition (see Figure 3). Binding to cyclin D1 activates CDK4/6, leading to phosphorylation of the retinoblastoma susceptibility (RB1) gene product Rb, which releases E2F transcription factors to activate cell cycle progression genes. Data from TCGA point to an association between dysregulation of the cyclin D-CDK4/6-Rb pathway and luminal B type breast cancer. CCND1 (cyclin D1) gene amplification occurs frequently in luminal B breast cancers (58%) compared with luminal A breast cancers (29%). In addition, luminal B tumors were associated more frequently with gain of CDK4 (25% in luminal B vs 14% in luminal A) and loss of negative regulators, including CDKN2A (p16) and CDKN2C (p18). In contrast to basal-like breast cancers, Rb is intact in most luminal breast cancers. Because functional Rb is a prerequisite for the efficacy of CDK4/6 inhibitors, patients with luminal B breast cancers are ideal candidates for treatment with these agents, and early success has been observed in clinical trials of CDK4/6 inhibitors.

Palbociclib, a first-in-class, oral, highly selective inhibitor of CDK4/6 kinase (IC50 [half maximal inhibitory concentration] = 11 nM; Ki = 2 nM) was found to be preferentially effective in ER+ cancer cells in preclinical studies.[42] In a randomized phase II study of letrozole with or without palbociclib as first-line therapy for metastatic ER+ HER2− breast cancer, a significant improvement in progression-free survival from 7.5 to 26.1 months (hazard ratio [HR] = 0.37; 95% confidence interval [CI], 0.21–0.63; P < .001) was observed.[43] This led to the FDA fast-track listing of palbociclib to accelerate the review and approval process. However, although preclinical studies indicated that cyclin D1 (CCND1) amplification, loss of p16, and intact RB1 predicted response to palbociclib, these results have not been confirmed in clinical trials. Two neoadjuvant studies of palbociclib in combination with AIs in patients with clinical stage II or III ER+ HER2− breast cancer are ongoing (NCT01723774 and NCT01709370), and their results will provide important information on predictive markers of response to these agents.

Epigenetic pathways. An interesting finding from TCGA is the identification of mutations in genes that regulate histone and DNA modifications (Figure 4). Mutations in the myeloid/lymphoid or mixed lineage leukemia gene (MLL3) occur at a rate of 7% overall in breast cancer, and are not specific to luminal breast cancer (8% in luminal A, 6% in luminal B, 7% in HER2-E, 5% in basal-like). The MLL3 gene encodes a DNA-binding protein that methylates histone H3 lys4 (H3K4), which leads to an open chromatin structure and therefore activation of target gene expression. In addition, an array of coding mutations and structural variations was discovered in other methyltransferases (MLL2, MLL4, and MLL5), demethyltransferases (KDM6A, KDM4A, KDM5B, and KDM5C), and acetyltransferases (MYST1, MYST3, and MYST4), and in several adenine-thymine (AT)-rich interactive domain–containing protein genes (ARID1A, ARID2, ARID3B, and ARID4B). Although the importance of these mutations in breast cancer remains to be proven, the promising data from early phase clinical trials of histone deacetylase (HDAC) inhibitors in patients with resistant ER+ breast cancer are encouraging.[44] Undoubtedly, additional studies are needed.

MDM2 TP53 pathway. The TP53 gene product is a potent tumor suppressor that induces apoptosis or cell cycle arrest in response to cellular stress. Mutation in the TP53 gene is a prominent feature of basal-like breast cancer (discussed in the section below). In ER+ breast cancer, mutation in TP53 is less frequent (30% in luminal B and 12% in luminal A). However, a significant number of luminal B breast cancers have amplification of MDM2 (30% in luminal B and 14% in luminal A), and rarely, mutations in ataxia telangiectasia mutated (ATM) and cell cycle checkpoint checkpoint kinase 2 (Chk2) checkpoint homolog (CHEK2), which could inactivate wild-type p53. In the neoadjuvant AI study, mutations in TP53 were prominently correlated with a significantly higher Ki67 level both at baseline and at surgery despite 4 months of neoadjuvant AI therapy, indicating an endocrine-resistant phenotype[4] (see Figure 2).

The high incidence of MDM2 amplification in the presence of wild type p53 in luminal B breast cancer presents a unique therapeutic opportunity for inhibitors of MDM2, which could potentially arrest cell proliferation by reactivation of p53.[45] MDM2 negatively regulates p53 through a direct interaction that inhibits its transactivation activity, exports p53 out of the nucleus, and acts as an E3 ubiquitin ligase to promote the proteasome-mediated degradation of p53.[45] Small-molecule inhibitors (MDM2 inhibitors) that interrupt the MDM2-p53 interaction are being developed as cancer therapeutics.[45] A phase I study of RO5503781 (Hoffmann-La Roche) is ongoing in patients with advanced malignancies (NCT01462175).

Basal-like breast cancer

Basal-like breast cancer is characterized by significant genomic instability, dysregulated cell cycle checkpoints, and resistance to apoptosis, which are explained at the genomic level by the loss of the three key tumor suppressor genes-TP53, BReast CAncer gene 1/2 (BRCA1/2), and PTEN-and gain of cell proliferation genes, including cyclin E1 (CCNE1) and c-Myc (see Figure 3) (see Table 4). Although there are more mutations identified in basal-like breast cancer, only three SMGs-TP53 (80%), PIK3CA (9%), and RB (4%)-have been recurrent in basal-like breast cancer.

Cell cycle checkpoint. The G1 to S transition during interphase is tightly controlled by the CDK4/6-cyclin D (mid-to-late G1) and CDK2-cyclin E (late G1) complexes, which inactivate Rb to release E2F transcription factors important for cell cycle progression. In the presence of genotoxic stress, the G1 checkpoint constituted by the ATM-Chk2-TP53 pathway is activated, which induces p21 to inhibit CDKs and arrest the cell cycle at G1 to allow time for DNA repair or to induce apoptosis if the DNA damage is too severe. The S and G2 to M checkpoints are controlled by the ATR-Chk1 pathway. BRCA1 also contributes to the S and G2 to M checkpoints,[46] in addition to its function in mediating DNA repair by homologous recombination.

A key feature of basal-like breast cancer is mutations in key components of the cell cycle control mechanism discussed above, including mutations in TP53 (80%), loss of Rb (30%), and germline and/or somatic mutation in BRCA1 or BRCA2 (20%). Mutations in these genes provide an underlying mechanism for the genomic instability of basal-like breast cancer.

Because CDK4/6 inhibition relies on a functional Rb to induce G1 arrest, it is unlikely to be effective in a significant proportion of basal-like breast cancers. Careful selection of patients in clinical trials of these agents in basal-like breast cancer is needed. Whether inhibitors against CDK2 (the partner of cyclin E) are effective in this setting remains to be tested.

Since cells defective in the ATM-TP53 pathway rely on the S and G2 to M checkpoint controlled by the ATR-Chk1 pathway, Chk1 inhibition in combination with chemotherapy has been tested as a therapeutic strategy for the treatment of TP53-mutated tumor in preclinical studies and has been shown to induce “mitotic cell death” in patient-derived xenograft models of basal-like breast cancer.[47,48] The results remain to be seen in clinical trials, as selective Chk1 inhibitors are still in the early phase of clinical development.

Hyperactivated forkhead box M1 (FOXM1) transcription. FOXM1 is an oncogenic transcription factor that is expressed only in proliferating cells and has important roles in activating genes that promote cell cycle progression, cell migration, invasion, angiogenesis, and metastasis. Analysis of TCGA gene expression data indicated a proliferation signature that is driven by FOXM1.[49] Interestingly, several chemotherapy agents have been shown to inhibit FOXM1. Direct inhibitors of FOXM1 are in preclinical development.

Myc amplification. Myc is a transcription factor and a known oncogene that promotes cell cycle progression. Amplification of Myc was observed in about one-third of basal-like breast cancers, and is another characteristic feature of this subtype. Development of Myc inhibitors is challenging. Recent preclinical studies have indicated that Myc and mTOR signaling pathways converge on mTOR-dependent phosphorylation of a common node in protein synthesis control, eukaryotic translation initation factor 4E binding protein–1 (4EBP1), with a remarkable anti-tumor effect in inducing apoptosis in Myc-driven tumors.[50] Again, clinical evaluation is needed.

DNA damage repair mechanisms. Up to 20% of basal-like breast cancers were found to have mutations in BRCA1/BRCA2 due to germline or somatic changes, which potentially could be targeted by poly–ADP ribose polymerase (PARP) inhibitors. Early trials of PARP inhibition in sporadic cases of TNBC, without patient pre-selection, have not shown convincing activity.[51] However, preclinical and clinical trials have reported success with PARP inhibition or use of platinum agents, which induce further DNA damage and synthetic lethality in germline BRCA1/BRCA2-related breast cancers.[52,53] It remains to be seen whether a subgroup of patients with sporadic basal-like breast cancer without BRCA mutations could respond to these agents.

Recent studies have shown platinum agents to be promising in the treatment of early-stage breast cancer; for example, a neoadjuvant study of cisplatin in TNBC demonstrated a pCR rate of 21%,[54] and the Gepar Sixto study reported that carboplatin improved the pCR rate compared with a taxane/anthracycline regimen in the neoadjuvant setting.[55] Additional data are expected from the Cancer and Leukemia Group B (CALGB) 40603 neoadjuvant trial. There is a pressing need for development of biomarkers that predict platinum sensitivity in TNBC.

The PI3K pathway in basal-like breast cancer. Among the breast cancer subtypes, basal-like breast cancer was associated with the highest activity of PI3K pathway signaling as demonstrated by the phosphoproteomic analysis and by gene expression signature. Genetically, PIK3CA is one of the three SMGs (with a 9% mutation rate), in addition to TP53 and RB, in basal-like breast cancer. However, a greater loss of negative regulators, including loss of phosphatase and tensin homolog (PTEN) (35%) and inositol polyphosphate-4-phosphatase type II (INPP4B) (30%), occurs more frequently in this subtype. Furthermore, inactivation of PTEN drives the development of basal-like breast cancer in animal models.[56,57] In preclinical studies, in a subpopulation of basal-like breast cancer, growth was dependent on the PI3K pathway activity.[58] The value of targeting PI3K in TNBC is currently being evaluated in clinical trials.

Other potential druggable targets identified in basal-like breast cancers. These include amplifications of components of the PI3K and rat sarcoma (Ras)–proto-oncogene serine/threonine-protein kinase (Raf)–MAPK kinase (MEK) pathway. These include, for example, PIK3CA (occurring in 49% of cases of basal-like disease), the Kirsten rat sarcoma viral oncogene homolog (KRAS) (32%), the proto-oncogene B-Raf (BRAF) (amplified in 30% of cases), and epidermal growth factor receptor (EGFR) (amplified in 23% of cases). Basal-like breast cancers are also characterized by amplifications of receptor tyrosine kinases such as fibroblast growth factor receptors 1 and 2 (FGFR1, FGFR2), insulin-like growth factor receptor 1 (IGFR-1), KIT, mesenchymal-epithelial transition factor (MET), and platelet-derived growth factor receptor, alpha polypeptide (PDGFRA). Finally, the high activity of the hypoxia-inducible factor 1-alpha (HIF1A)/aryl hydrocarbon receptor nuclear translocator (ARNT) pathway, as evaluated by gene expression analysis, suggests that basal-like breast cancer might be susceptible to angiogenesis inhibitors and/or bioreductive drugs that become activated under hypoxic conditions. These hypotheses remain to be tested in clinical trials.

HER2+ breast cancer

In TCGA, HER2+ breast cancer fell into two major subgroups by mRNA and proteomic analysis: HER2-E (about 50%) and luminal. The HER2-E subgroup showed strong EGFR and HER2 signaling activation by RPPA, features likely associated with sensitivity to HER2-targeted therapy. Genetically, the HER2-E subtype has a high frequency of TP53 mutation (75%), high aneuploidy, and amplifications of MDM2 (30%); CCND1 amplification (38%) and CDK4 gain (24%); and amplifications of FGFR and EGFR (see Table 4). The luminal subtype was enriched for the expression of luminal cluster genes, including GATA3, BCL2, and ESR1, and GATA3 mutation occurred in the luminal subtype but not in the HER2-E subtype.

In addition, there was a high frequency of PIK3CA mutation (39%), low frequency of PTEN and PIK3R1 mutation, and loss of PTEN (19%) and INPP4B (30%). Rarely observed mutations in HER2, EGFR, and HER3 were reported, providing additional therapeutic targets. Agents targeting the PI3K pathway and growth factor receptor signaling pathways are in clinical trials in the treatment of HER2+ breast cancer. However, prospective patient selection based on genomic or proteomic characteristics is needed in future trials.

Distinguishing Driver Mutations From Passenger Mutations

TCGA’s breast cancer project identified a striking 30,626 somatic mutations by whole exome sequencing of 510 tumors, including 28,319 point mutations, 4 dinucleotide mutations, and 2,302 insertions/deletions (indels) (ranging from 1 to 53 nucleotides). The complexity and multitude of the genetic abnormalities make it a daunting task to identify the few drivers from a large majority of passenger mutations in a given tumor.

TCGA employed the approach of SMGs-which, as previously noted, are recurrent mutations observed at a higher frequency than expected from background mutation across the tumors-to identify likely driver mutations. In addition to focusing on individual mutations, an integrated pathway approach called Pathway Recognition Algorithm Using Data Integration on Genomic Models (PARADIGM) was developed and applied to TCGA data to develop graphic models for multi-platform data analysis.[59] The integration of information from multiple platforms-including mRNA, miRNA expression, phosphoproteomics, and patterns of methylation-in addition to changes in DNA in the same tumor set-allows identification of recurrent pathway abnormalities.

In TCGA’s breast cancer project, only 35 genes met the criteria of SMGs. However, very-low-recurrence mutations that did not qualify as SMGs may still have clinical importance because breast cancer is so common. An example would be activating mutations in HER2 identified in HER2–nonamplified/nonoverexpressed disease. Despite their low frequency in primary breast cancers (1.6%), the mutation pattern and subsequent functional studies indicated that they are likely driver mutations.[60] A phase II trial of neratinib, an irreversible pan-HER inhibitor, in patients with advanced HER2 nonamplified but HER2-mutated breast cancer is ongoing (NCT 01670877). While these tumors are rare, breast cancer is so common that a 2% population still represents thousands of patients, comparable to the incidence of leukemias for which targeted therapy has been successful.

Several computational models are available to predict the functional consequences of mutations occurring in the coding regions. The PolyPhen (for “polymorphism phenotyping”) tool predicts the possible impact of an amino acid substitution on the structure and function of a protein based on structural and comparative (evolutionary conservation) considerations.[61,62] Mutation Assessor assesses the mutation impact by calculating a “functional impact score,” based on the evolutionary conservation patterns of the amino acid affected.[63,64] Other predictive algorithms include SIFT, CanPredict, and CHASM.[65-68]

In addition to statistical tools, functional validation of candidate driver mutations in preclinical studies is necessary to identify therapeutic opportunities. To this end, researchers commonly employ a candidate gene approach, large-scale screening by gain-of-function studies, or RNA interference (RNAi)-mediated loss-of-function investigations. These studies are often performed using existing cancer cell lines, genetically engineered cells, and patient-derived primary or xenograft cell lines; therefore, the conclusions are limited by the specific cellular context. Ideally, preclinical models that encompass a variety of genetic aberrations are tested.

Ultimately, prospective clinical trials are needed for definitive validation of TCGA’s findings. The design of clinical trials for low-frequency driver mutations is challenging, since a large number of patients need to be screened. As functional genomic data accumulate, more potentially actionable mutations will be added to the list of the screening genetic tests, so enrollment in several therapeutic trials, with each targeting a specific mutation, is possible.

Clinical Application of Genomic Information

TCGA’s breast cancer project has generated a list of key mutations and suggested signaling pathways likely to be drivers. However, further validation is needed before this information can be applied to routine clinical practice. The medical community awaits informative genomic data from prospective clinical trials. At the same time, affordable clinical assays are increasingly available, and physicians are now offering patients high-throughput genomic assays in the hope that they will facilitate more rational decision making in clinical trial enrollment for advanced disease. Multiplexed gene mutation assays for whole exome sequencing are now available using small amount of clinical material with a turnaround time that is feasible for real-time clinical decision making.

New bioinformatics tools such as dGene (collection of Druggable Genes) have been developed to provide annotations of potential druggable mutations based on high throughput sequencing data.[69] Success of this approach has been reported in the care of individual patients,[70] but clinical evidence has not yet risen above the level of anecdotes. Some have proposed the molecular tumor board approach,[71] with disease- or pathway-specific experts-including bioinformaticians, genomics scientists, cancer biologists, pathologists, and medical oncologists-to interpret the findings and provide recommendations on treatment. Geneticists and bioethicists are also present to provide patients and their families with insight regarding the ethics of reporting incidental germline mutations associated with hereditary cancer or other diseases. Guidelines are being developed to address ethical, legal, and counseling challenges of sharing genetic results of high-throughput sequencing studies with patients.[72] Among the many challenges of this approach is the availability of mutation-matched drugs. In the absence of a clinical trial that guarantees drug availability, drug reimbursement will always be a problem when prescribing treatment outside of an FDA-approved indication.

Conclusion

Information from TCGA and other studies has provided an in-depth catalogue of the molecular alterations occurring in breast cancer and has generated a wide spectrum of therapeutic hypotheses for further investigation. The rapid development of clinical assays based on next-generation sequencing has made it possible to conduct clinical trials in molecularly defined patient populations. Clearly the next step will involve deep collaborations among biologists, pharmacologists, clinical laboratories, the pharmaceutical industry, the National Cancer Institute, and federal regulatory agencies, with a focus on the development of clinical trials that, informed by genomic analysis, triage patients to the appropriate targeted therapeutic trials.

Financial Disclosure:Dr. Ellis is named on the patent for PAM50, used in the Prosigna Breast Cancer Prognostic Gene Signature Assay, which is licensed to NanoString Technologies by Bioclassifier LLC. He also holds shares in Bioclassifier LLC. Dr. Ma has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.

References:

1. Shah SP, Roth A, Goya R, et al. The clonal and mutational evolution spectrum of primary triple-negative breast cancers. Nature. 2012;486:395-9.

2. Stephens PJ, Tarpey PS, Davies H, et al. The landscape of cancer genes and mutational processes in breast cancer. Nature. 2012;486:400-4.

3. Banerji S, Cibulskis K, Rangel-Escareno C, et al. Sequence analysis of mutations and translocations across breast cancer subtypes. Nature. 2012;486:405-9.

4. Ellis MJ, Ding L, Shen D, et al. Whole-genome analysis informs breast cancer response to aromatase inhibition. Nature. 2012;486:353-60.

5. Curtis C, Shah SP, Chin SF, et al. The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature. 2012;486:346-52.

6. Network TCGA. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490:61-70.

7. Nik-Zainal S, Van Loo P, Wedge DC, et al. The life history of 21 breast cancers. Cell. 2012;149:994-1007.

8. Craig DW, O’Shaughnessy JA, Kiefer JA, et al. Genome and transcriptome sequencing in prospective metastatic triple-negative breast cancer uncovers therapeutic vulnerabilities. Mol Cancer Ther. 2013;12:104-16.

9. Ellis MJ, Perou CM. The genomic landscape of breast cancer as a therapeutic roadmap. Cancer Discov. 2013;3:27-34.

10. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature. 2000;406:747-52.

11. Sorlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001;98:10869-74.

12. Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci USA. 2003;100:8418-23.

13. Kreike B, van Kouwenhove M, Horlings H, et al. Gene expression profiling and histopathological characterization of triple-negative/basal-like breast carcinomas. Breast Cancer Res. 2007;9:R65.

14. Atchley DP, Albarracin CT, Lopez A, et al. Clinical and pathologic characteristics of patients with BRCA-positive and BRCA-negative breast cancer. J Clin Oncol. 2008;26:4282-8.

15. Prat A, Parker JS, Karginova O, et al. Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer. Breast Cancer Res. 2010;12:R68.

16. Lehmann BD, Bauer JA, Chen X, et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J Clin Invest. 2011;121:2750-67.

17. Kong SL, Li G, Loh SL, et al. Cellular reprogramming by the conjoint action of ER[alpha], FOXA1, and GATA3 to a ligand-inducible growth state. Mol Syst Biol. 2011;7:526.

18. Cirillo LA, Lin FR, Cuesta I, et al. Opening of compacted chromatin by early developmental transcription factors HNF3 (FoxA) and GATA-4. Mol Cell. 2002;9:279-89.

19. Kouros-Mehr H, Slorach EM, Sternlicht MD, Werb Z. GATA-3 maintains the differentiation of the luminal cell fate in the mammary gland. Cell. 2006;127:1041-55.

20. Hurtado A, Holmes KA, Ross-Innes CS, et al. FOXA1 is a key determinant of estrogen receptor function and endocrine response. Nat Genet. 2011;43:27-33.

21. Carroll JS, Liu XS, Brodsky AS, et al. Chromosome-wide mapping of estrogen receptor binding reveals long-range regulation requiring the forkhead protein FoxA1. Cell. 2005;122:33-43.

22. Lin CY, Vega VB, Thomsen JS, et al. Whole-genome cartography of estrogen receptor alpha binding sites. PLoS Genet. 2007;3:e87.

23. Kong SL, Li G, Loh SL, et al. Cellular reprogramming by the conjoint action of ERalpha, FOXA1, and GATA3 to a ligand-inducible growth state. Mol Syst Biol. 2011;7:526.

24. Friedman AD. Cell cycle and developmental control of hematopoiesis by Runx1. J Cell Physiol. 2009;219:520-4.

25. Ito Y. RUNX genes in development and cancer: regulation of viral gene expression and the discovery of RUNX family genes. Adv Cancer Res. 2008;99:33-76.

26. Stender JD, Kim K, Charn TH, et al. Genome-wide analysis of estrogen receptor alpha DNA binding and tethering mechanisms identifies Runx1 as a novel tethering factor in receptor-mediated transcriptional activation. Mol Cell Biol. 2010;30:3943-55.

27. Wang L, Brugge JS, Janes KA. Intersection of FOXO- and RUNX1-mediated gene expression programs in single breast epithelial cells during morphogenesis and tumor progression. Proc Natl Acad Sci U S A. 2011;108:E803-12.

28. Kadota M, Yang HH, Gomez B, et al. Delineating genetic alterations for tumor progression in the MCF10A series of breast cancer cell lines. PLoS One. 2010;5:e9201.

29. Chimge NO, Frenkel B. The RUNX family in breast cancer: relationships with estrogen signaling. Oncogene. 2013;32:2121-30.

30. Okuda T, van Deursen J, Hiebert SW, et al. AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell. 1996;84:321-30.

31. Wang L, Lawrence MS, Wan Y, et al. SF3B1 and Other Novel Cancer Genes in Chronic Lymphocytic Leukemia. N Engl J Med. 2011;365:2497-506.

32. Sanchez CG, Ma CX, Crowder RJ, et al. Preclinical modeling of combined phosphatidylinositol-3-kinase inhibition with endocrine therapy for estrogen receptor-positive breast cancer. Breast Cancer Res. 2011;13:R21.

33. Crowder RJ, Phommaly C, Tao Y, et al. PIK3CA and PIK3CB inhibition produce synthetic lethality when combined with estrogen deprivation in estrogen receptor-positive breast cancer. Cancer Res. 2009;69:3955-62.

34. Baselga J, Campone M, Piccart M, et al. Everolimus in postmenopausal hormone-receptor-positive advanced breast cancer. N Engl J Med. 2012;366:520-9.

35. Zhang H, Liu G, Dziubinski M, et al. Comprehensive analysis of oncogenic effects of PIK3CA mutations in human mammary epithelial cells. Breast Cancer Res Treat. 2008;112:217-27.

36. Juric D, Argiles G, Burris HA, et al. Phase I study of BYL719, an alpha-specific PI3K inhibitor, in patients with PIK3CA mutant advanced solid tumors: preliminary efficacy and safety in patients with PIK3CA mutant ER-positive (ER+) metastatic breast cancer (MBC). Cancer Res. 2012;72(24 suppl):abstr P6-10-07.

37. Pacheco JM, Ma CX. PI3K inhibitors in breast cancer treatment. Int J Targeted Ther Cancer. 2013 May 7. [Epub ahead of print]

38. Roy SK, Hu J, Meng Q, et al. MEKK1 plays a critical role in activating the transcription factor C/EBP-beta-dependent gene expression in response to IFN-gamma. Proc Natl Acad Sci USA. 2002;99:7945-50.

39. Easton DF, Pooley KA, Dunning AM, et al. Genome-wide association study identifies novel breast cancer susceptibility loci. Nature. 2007;447:1087-93.

40. Yujiri T, Sather S, Fanger GR, Johnson GL. Role of MEKK1 in cell survival and activation of JNK and ERK pathways defined by targeted gene disruption. Science. 1998;282:1911-14.

41. Minamino T, Yujiri T, Papst PJ, et al. MEKK1 suppresses oxidative stress-induced apoptosis of embryonic stem cell-derived cardiac myocytes. Proc Natl Acad Sci USA. 1999;96:15127-32.

42. Finn RS, Dering J, Conklin D, et al. PD 0332991, a selective cyclin D kinase 4/6 inhibitor, preferentially inhibits proliferation of luminal estrogen receptor-positive human breast cancer cell lines in vitro. Breast Cancer Res. 2009;11:R77.

43. Finn RS, Crown JP, Lang I, et al. Results of a randomized phase 2 study of PD 0332991, a cyclin-dependent kinase (CDK) 4/6 inhibitor, in combination with letrozole vs letrozole alone for first-line treatment of ER+/HER2– advanced breast cancer (BC). Cancer Res. 2012;72(24 suppl):abstr S1-6.

44. Yardley DA, Ismail-Khan RR, Melichar B, et al. Randomized phase II, double-blind, placebo-controlled study of exemestane with or without entinostat in postmenopausal women with locally recurrent or metastatic estrogen receptor-positive breast cancer progressing on treatment with a nonsteroidal aromatase inhibitor. J Clin Oncol. 2013;31:2128-35.

45. Shangary S, Wang S. Targeting the MDM2-p53 interaction for cancer therapy. Clin Cancer Res. 2008;14:5318-24.

46. Deng CX. BRCA1: cell cycle checkpoint, genetic instability, DNA damage response and cancer evolution. Nucleic Acids Res. 2006;34:1416-26.

47. Ma CX, Cai S, Li S, et al. Targeting Chk1 in p53-deficient triple-negative breast cancer is therapeutically beneficial in human-in-mouse tumor models. J Clin Invest. 2012;122:1541-52.

48. Ma CX, Janetka JW, Piwnica-Worms H. Death by releasing the breaks: CHK1 inhibitors as cancer therapeutics. Trends Mol Med. 2011;17:88-96.

49. Halasi M, Gartel AL. Targeting FOXM1 in cancer. Biochem Pharmacol. 2013;85:644-52.

50. Pourdehnad M, Truitt ML, Siddiqi IN, et al. Myc and mTOR converge on a common node in protein synthesis control that confers synthetic lethality in Myc-driven cancers. Proc Natl Acad Sci USA. 2013;110:11988-93.

51. Gelmon KA, Tischkowitz M, Mackay H, et al. Olaparib in patients with recurrent high-grade serous or poorly differentiated ovarian carcinoma or triple-negative breast cancer: a phase 2, multicentre, open-label, non-randomised study. Lancet Oncol. 2011;12:852-61.

52. Farmer H, McCabe N, Lord CJ, et al. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature. 2005;434:917-21.

53. Tutt A, Robson M, Garber JE, et al. Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and advanced breast cancer: a proof-of-concept trial. Lancet. 2010;376:235-44.

54. Silver DP, Richardson AL, Eklund AC, et al. Efficacy of neoadjuvant cisplatin in triple-negative breast cancer. J Clin Oncol. 2010;28:1145-53.

55. German Breast Group: Addition of carboplatin to neoadjuvant therapy for triple-negative and HER2-positive early breast cancer (Gepar Sixto). National Cancer Institute ClinicalTrials.gov identifier NCT01426880.

56. Dourdin N, Schade B, Lesurf R, et al. Phosphatase and tensin homologue deleted on chromosome 10 deficiency accelerates tumor induction in a mouse model of ErbB-2 mammary tumorigenesis. Cancer Res. 2008;68:2122-31.

57. Saal LH, Gruvberger-Saal SK, Persson C, et al. Recurrent gross mutations of the PTEN tumor suppressor gene in breast cancers with deficient DSB repair. Nat Genet. 2008;40:102-7.

58. Xu S, Li S, Guo Z, et al. Combined targeting of mTOR and AKT is an effective strategy for basal-like breast cancer in patient-derived xenograft models. Mol Cancer Ther. 2013;12:1665-75.

59. Vaske CJ, Benz SC, Sanborn JZ, et al. Inference of patient-specific pathway activities from multi-dimensional cancer genomics data using PARADIGM. Bioinformatics. 2010;26:i237-45.

60. Bose R, Kavuri SM, Searleman AC, et al. Activating HER2 mutations in HER2 gene amplification negative breast cancer. Cancer Discov. 2013;3:224-37.

61. Sunyaev S, Ramensky V, Koch I, et al. Prediction of deleterious human alleles. Hum Mol Genet. 2001;10:591-7.

62. Ramensky V, Bork P, Sunyaev S. Human non-synonymous SNPs: server and survey. Nucleic Acids Res. 2002;30:3894-900.

63. Reva B. Revealing selection in cancer using the predicted functional impact of cancer mutations. Application to nomination of cancer drivers. BMC Genomics. 2013;14(suppl 3):S8.

64. Reva B, Antipin Y, Sander C. Determinants of protein function revealed by combinatorial entropy optimization. Genome Biol. 2007;8:R232.

65. Ng PC, Henikoff S. Predicting deleterious amino acid substitutions. Genome Res. 2001;11:863-74.

66. Kaminker JS, Zhang Y, Watanabe C, Zhang Z. CanPredict: a computational tool for predicting cancer-associated missense mutations. Nucleic Acids Res. 2007;35:W595-8.

67. Carter H, Chen S, Isik L, et al. Cancer-specific high-throughput annotation of somatic mutations: computational prediction of driver missense mutations. Cancer Res. 2009;69:6660-7.

68. Adzhubei IA, Schmidt S, Peshkin L, et al. A method and server for predicting damaging missense mutations. Nat Methods. 2010;7:248-9.


69. Kumar RD, Chang LW, Ellis MJ, Bose R. Prioritizing potentially druggable mutations with dGene: an annotation tool for cancer genome sequencing data. PLoS One. 2013;8:e67980.

70. Welch JS, Westervelt P, Ding L, et al. Use of whole-genome sequencing to diagnose a cryptic fusion oncogene. JAMA. 2011;305:1577-84.

71. Roychowdhury S, Iyer MK, Robinson DR, et al. Personalized oncology through integrative high-throughput sequencing: a pilot study. Sci Transl Med. 2011;3:111ra21.

72. Lolkema MP, Gadellaa-van Hooijdonk CG, Bredenoord AL, et al. Ethical, legal, and counseling challenges surrounding the return of genetic results in oncology. J Clin Oncol. 2013;31:1842-8.

73. Goldstein TC, Paull EO, Ellis MJ, Stuart JM. Molecular pathways: extracting medical knowledge from high-throughput genomic data. Clin Cancer Res. 2013;19:3114-20.

Recent Videos
Heather Zinkin, MD, states that reflexology improved pain from chemotherapy-induced neuropathy in patients undergoing radiotherapy for breast cancer.
Study findings reveal that patients with breast cancer reported overall improvement in their experience when receiving reflexology plus radiotherapy.
Patients undergoing radiotherapy for breast cancer were offered 15-minute nurse-led reflexology sessions to increase energy and reduce stress and pain.
Whole or accelerated partial breast ultra-hypofractionated radiation in older patients with early breast cancer may reduce recurrence with low toxicity.
Ultra-hypofractionated radiation in those 65 years or older with early breast cancer yielded no ipsilateral recurrence after a 10-month follow-up.
The unclear role of hypofractionated radiation in older patients with early breast cancer in prior trials incentivized research for this group.
Patients with HR-positive, HER2-positive breast cancer and high-risk features may derive benefit from ovarian function suppression plus endocrine therapy.
Paolo Tarantino, MD discusses updated breast cancer trial findings presented at ESMO 2024 supporting the use of agents such as T-DXd and ribociclib.
Paolo Tarantino, MD, discusses the potential utility of agents such as datopotamab deruxtecan and enfortumab vedotin in patients with breast cancer.
Paolo Tarantino, MD, highlights strategies related to screening and multidisciplinary collaboration for managing ILD in patients who receive T-DXd.
Related Content