Challenging Genomic Blueprint of Head and Neck Cancers Emerges

Despite better treatment options, outcomes for patients with head and neck cancers have not substantially improved over the past several decades. Historically a tobacco- and alcohol-associated cancer, the epidemiology of head and neck cancer has shifted in recent years to reflect an increase in human papillomavirus (HPV)-driven tumors.

Resisting Improved Outcomes

Next-generation sequencing efforts focusing on the most common form, head and neck squamous cell carcinoma (HNSCC), are illuminating the hidden complexities of its genome. Multiple molecular subtypes and key differences between tumors with and without HPV infection are beginning to emerge. The challenge ahead is to find ways to use this molecular blueprint to guide more personalized and effective therapies for individual patient populations.Head and neck cancers are a group of heterogeneous malignancies that develop in a common anatomic region, with an incidence 600,000 cases per year worldwide including 50,000 annually in the United States. The vast majority are HNSCCs arising in the epithelial cells that line the mucosal surfaces of the head and neck.

Outcomes for the substantial number of patients who present with advanced disease have remained stubbornly low for the past several decades, despite improvements in treatment options, including the development of molecularly targeted therapies. In particular, drugs targeting the epidermal growth factor receptor (EGFR) have been a central focus of research and development, given that this protein is overexpressed in more than 90% of HNSCCs.

Turning to Next-Generation Sequencing

Although research efforts culminated with the FDA’s approval of cetuximab (Erbitux) in 2006, other EGFR-targeting therapies have proved clinically disappointing and cetuximab remains the only targeted agent approved for HNSCC. Cetuximab has demonstrated improvements in overall survival compared with standard therapies; it is indicated as a single agent and in combination with radiation or platinum-based therapy in initial and recurrent HNSCC treatment settings.The success of targeted therapies in HNSCC has been tempered by a poor understanding of the genomic background of this disease but with the advent of next-generation sequencing technologies, that is beginning to change.

Landmark studies published in 2011 identified common somatic mutations, some linked to HNSCC for the first time, providing valuable insight into HNSCC carcinogenesis.

Commonly mutated genes were TP53, FAT1, CDKN2A, PIK3CA, NOTCH1, and CASP8. The tumor suppressor gene TP53, frequently described as the “guardian of the genome” since its primary function is maintaining genomic integrity, is a transcription factor that regulates the expression of a whole host of target genes. It is the most frequently mutated gene in cancer and had long been suspected to be involved in the development of HNSCC.

Ligand binding to the Notch receptor

Tough-to-Target Tumor Suppressors

FAT1 and CDKN2A are also tumor suppressor genes. FAT1 is involved in the Wnt signaling pathway, binding to Wnt and limiting its ability to translocate into the nucleus. CDKN2A encodes the p16 protein that is an inhibitor of cyclin-dependent kinases (CDKs), which play a vital role in the normal progression of the cell cycle. CASP8 encodes a caspase enzyme involved in programmed cell death. A key novel finding was frequent mutations in the NOTCH1 gene, which encodes a receptor involved in embryonic development and cellular differentiation.drives a cascade of cleavage events that ultimately release a portion of the receptor that is inside the cell membrane, known as the notch intracellular domain (NICD). The NICD then moves into the nucleus where it activates numerous genes. Depending upon the cellular context, Notch1 can act to either promote cancer cell growth or to suppress it, but in the case of HNSCC it is thought to predominantly function as a tumor suppressor.The picture that began to emerge was that these tumors are characterized by frequent loss of tumor suppressor genes and a paucity of readily targetable driver mutations, making them much less amenable to current therapeutic strategies. Researchers have made some strides in working around the difficulties in targeting tumor suppressor proteins.

Since p16 is upstream of CDK4/6, its loss can drive overexpression of these proteins. As kinases, they present a much more readily druggable target and, indeed, CDK4/6 inhibitors, led by palbociclib (Ibrance), have been developed and are being evaluated in HNSCC (NCT02499120).

Studies are also being conducted to seek out other genes that are specifically required for the survival of cancerous cells with mutations in tumor suppressor genes such as TP53, a phenomenon known as synthetic lethality. One gene that was identified in this manner is the Wee1 kinase, which plays an important role in the cell cycle. Phase I trials of a Wee1 inhibitor, AZD1775, are ongoing in HNSCC (NCT02508246).

The only known oncogene found to be mutated at the level of statistical significance in HNSCC was PIK3CA. Phosphatidylinositol-3-kinase (PI3K) is an enzyme that regulates the production of a lipid found in the cell membrane (PIP3), which activates a number of proteins in the cell and conducts a variety of cell-signaling pathways involved in key cellular process including cell growth and survival. The PI3K pathway is one of the most frequently perturbed in cancer and is an integral node in many other signaling pathways, notably the EGFR pathway, and may be partly responsible for the resistance that often develops during treatment with EGFR inhibitors.

Rise of HPV-Associated HNSCC

Numerous components of the pathway are being targeted in HNSCC, including PI3K itself, Akt, and mammalian target of rapamycin (mTOR). PI3K inhibitors being evaluated in HNSCC include BYL719 (NCT02145312) and SF1126 (NCT02644122). SF1126, a pro-drug that targets both PI3K and mTOR, is composed of the drug LY294002 conjugated to a peptide that binds to specific integrins in the tumor microenvironment to increase the target specificity of the inhibitor.Historically, HNSCC has been associated with tobacco and alcohol consumption, but the incidence of disease associated with these risk factors has been steadily declining. Infection with high-risk strains of HPV was first recognized as a major independent risk factor for HNSCC in 2007, following a rapid rise in the incidence of HPV-positive tumors that has continued over the past decade.

HPV is thought to be the most common etiologic agent in virally associated cancers; of the 40 types of HPV that are transmitted by sexual contact, 12 of them are believed to have a causative role in cancer, and these are collectively known as the “high-risk” types. HPV is most widely known for its role in cervical cancer, in which the HPV16 and HPV18 subtypes are responsible for more than 70% of tumors. Subsequently, it was shown that HPV16 also accounts for more than 85% of head and neck cancers.

Although our understanding of precisely how HPV causes cancer continues to evolve, the oncogenic activity of these viruses is thought to result predominantly from the action of two HPV genes, E6 and E7. Once HPV has infected the host’s epithelial cells, it hijacks the cellular machinery to express the eight genes contained in its genome. The proteins encoded by the E6 and E7 genes act as oncogenes, driving carcinogenesis mainly through the suppression of p53 and pRb. Additionally, the HPV genome is thought to directly insert itself into the host genome and this is also suspected to play an important role in cancer development.

HPV-associated tumors are histologically and epidemiologically distinct from those in which the virus is not present. HPV-positive HNSCC is rapidly emerging in the developed world and many patients present with no smoking history, while in developing countries, where smoking remains the major risk factor, the incidence is much lower. The rise in HPV-positive HNSCC has been particularly evident among patients younger than age 45.

Differences Extend Beyond Epidemiology

While HPV-negative tumors are found at all sites of the head and neck, the increase in HPV-positive tumors has been primarily concentrated in cancers of the oropharynx. HPV-positive tumors are also associated with better clinical outcomes.Earlier sequencing studies were limited in their enrollment of HPV-positive patients, but more recent studies have examined a significantly larger proportion of such patients and some have focused specifically on HPV-positive samples. These studies have revealed that the genomic background of HPV-negative and HPV-positive tumors are quite distinct. In 2014, a comparative genomic analysis of HPV-positive and —negative HNSCC was published.

The following year, The Cancer Genome Atlas (TCGA) research network published its own comprehensive, multiplatform analysis of HNSCC tumors, and also directly compared the genomic background of HPV-positive and HPV-negative tumors. A subsequent study went on to examine just the HPV-positive samples from the TCGA to evaluate the impact of HPV integration on the host genome.

While earlier sequencing studies had suggested that the mutation rate in HPV-positive and —negative tumors may differ, this was not seen in the TCGA study or others. In general, there were similar mutation and copy number change rates between the two subtypes and the types of copy number changes were also broadly concordant.

Shared copy number amplifications were observed in chromosomes 1q, 3q, 5p, and 8q and shared deletions in chromosomes 3p, 5q, and 11q, among others. The region of chromosome 3q that was frequently amplified contains the transcription factors TP63 and SOX2 and the oncogene PIK3CA.

There were some distinct alterations, notably in HPV-positive tumors, where there were focal deletions in chromosomal regions associated with the TRAF3 gene and amplifications in the region containing the E2F1 gene. TRAF3 (TNF receptor- associated factor 3) is a tumor suppressor protein involved in regulating the immune response to viruses and also acts as a negative regulator of nuclear factor kappa B (NFκB) signaling. E2F1 encodes a transcription factor involved in cell cycle regulation. Meanwhile, HPV-negative tumors had novel coamplifications of the 11q13 and 11q22 chromosomal regions. The authors suggested that this likely promotes the interaction of two genes involved in cell death, BIRC2 and FADD.

Another key difference, with relevance to the use of targeted therapies in HNSCC, was that HPV-positive tumors were lacking in chromosome 7 amplifications that were prominent in HPV-negative tumors, specifically in the location where the EGFR gene can be found.

In contrast to the broad concordance between copy number changes, the mutation profiles of HPV-positive and —negative tumors were distinctly different. While HPV-negative tumors displayed somatic mutations that were similar to lung squamous cell carcinomas, HPV-positive tumors were unique. The most frequent mutations in tumors that were HPV-negative were in the tumor suppressor genes TP53 and CDKN2A. Conversely, HPV-positive tumors rare harbored CDKN2A mutations and also showed predominantly wild-type TP53. This may, in part, explain the better prognosis of patients with HPV-positive tumors. In fact, the TCGA researchers suggested that wild-type TP53 could act as a universal marker of favorable outcome, regardless of HPV status, as they also identified a subset of HPV-negative oral cavity cancers that had wild-type TP53, in combination with other genomic alterations, such as CASP8 and HRAS mutations, that had significantly better outcomes.

In HPV-positive tumors, the most commonly mutated gene was PIK3CA but the type of PIK3CA mutation was unique. HPV-negative tumors also display mutations in PIK3CA, but these are found in a hotspot within the kinase domain, similar to many other cancer types. In HPV-positive cancers, PIK3CA mutations occur almost exclusively at one of two hotspots in the helical domain.

Some of the differences between virally associated HNSCC and HPV-negative disease may be explained by the fact that molecular alterations induced by HPV often “phenocopy” the effects of mutations in HPV-negative tumors; thus, HPV achieves the same goals by different ends. An example of this is the near-universal presence of TP53 mutations in HPV-negative tumors, while the same mutation is virtually nonexistent in HPV-positive tumors. Nevertheless, the function of the TP53 gene is often lost because the HPV E6 protein acts to suppress it. Therefore, the loss of p53 function removes the selective pressure for TP53 mutations. Viral Integration Drives Unique Genomic Events In addition to hijacking the cellular machinery to meet its own needs, the HPV genome is able to insert itself into the host genome in HNSCC cells.

The exact rate at which this occurs and the effects of viral integration on oncogenesis are the subject of ongoing studies.

Emerging research suggests that viral integration in HNSCC is similar to that observed in cervical cancer. The E2 gene, which normally suppresses the activity of E6 and E7, is often broken apart during the integration process, which suggests one mechanism through which integration promotes cancer cell growth.

Studies have also looked at where the HPV genome inserts itself into the host genome and the process appears to be nonrandom, with insertion often occurring in sites that cause disrupted expression or rearrangement of genes, which could also contribute to the genomic instability of HPV-positive tumors. The fact that some tumors contain integrated DNA and others do not, however, suggests that integration is not required for tumors to develop.

Divide and Conquer

Interestingly, a recent study that evaluated 35 HPV-positive tumors from among the TCGA study samples suggested that HNSCCs with integrated HPV had a unique gene expression and DNA methylation profile to those that did not have integrated viral DNA, suggesting that these two subtypes may be vulnerable to different types of therapy.Several studies have also looked at the patterns of gene expression across HNSCCs, to determine if distinct subtypes with differing expression profiles emerged. Four key groups have consistently been observed: classic, atypical, mesenchymal, and basal. The classical subtype is the most closely associated with heavy smoking history, and is defined by altered expression of genes involved in oxidative stress (such as KEAP1, NFE2L2, and CUL3), and frequently displays TP53 mutations, CDKN2A loss of function, and chromosome 3q amplification.

The basal subtype, by contrast, has intact oxidative stress signaling pathways and fewer alterations of chromosome 3q. There is also often wild-type TP53, with co-mutations in HRASCASP8, and co-amplification of chromosome 11q, as well as frequent inactivation of NOTCH1. One particularly interesting finding was that no HPV-positive tumors were observed in the basal subtype and EGFR amplification was also common in this group, again suggesting that EGFR inhibitors would likely play a limited role in the treatment of HPV-positive HNSCC.

The atypical subgroup is highly enriched for HPV-positive tumors, and correspondingly demonstrates few chromosome 7 amplifications (on which the EGFR gene is located), few TP53 mutations, and a high rate of activating mutations in exon 9, which contains the PIK3CA helical domain. Finally, the mesenchymal subtype displayed a gene expression profile that suggested highly activated T cells. These tumors could prove particularly susceptible to novel immunotherapies such as immune checkpoint inhibitors, which are currently being evaluated in the treatment of HNSCC.

Jane de Lartigue, PhD, is a freelance medical writer and editor based in New Haven, Connecticut.

Key Research

• Birkeland AC, Ludwig ML, Meraj TS, et al. The tip of the iceberg: clinical implications of genomic sequencing projects in head and neck cancer. Cancers (Basel). 2015;7(4):2094-2109.
• Hayes DN, Van Waes C, Seiwert TY. Genetic landscape of human papillomavirus- associated head and neck cancer and comparison to tobacco-related tumors. J Clin Oncol. 2015;33(29):3227-3334.
• Lechner M, Fenton TR. The genomics, epigenomics, and transcriptomics of HPV-associated oropharyngeal cancer—understanding the basis of a rapidly evolving disease. Adv Genet. 2016;93:1-56.
• Parfenov M, Pedamallu CS, Gehlenborg N, et al. Characterization of HPV and host genome interactions in primary head and neck cancers. Proc Natl Acad Sci. 2014;111(43):15544-15549.
• Riaz N, Morris LG, Lee W, Chan TA. Unraveling the molecular genetics of head and neck cancer through genome-wide approaches. Genes Dis. 2014;1(1):75-86.
• Sethi N, MacLennan K, Wood HM, Rabbitts P. Past and future impact of next-generation sequencing in head and neck cancer [published online July 20, 2015]. Head Neck. doi: 10.1002/hed.24085.
• Thariat J, Vignot S, Lapierre A, et al. Integrating genomics in head and neck cancer treatment: promises and pitfalls. Crit Rev Oncol Hematol. 2015;95(3)397-406. The Cancer Genome Atlas Network. Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature. 2015;517(7536):576-582.