Genomic Diversity Clouds Outlook for Neuroendocrine Tumors

Despite a spate of therapeutic advances in recent years, neuroendocrine (NET) tumors remain a poorly investigated collection of cancers whose rarity has limited clinical trial enrollment. Now, genome sequencing studies are providing clues for therapeutic avenues of inquiry for the most common types of NETs but are raising as many questions as they answer.

The Treatment Landscape

What has emerged thus far is a picture of a group of molecularly heterogeneous malignancies, with a paucity of readily targetable, classic oncogenic drivers. Although clinical applications are a long way off, important insights have been gained that suggest the need for distinct treatment strategies.NETs arise from neuroendocrine cells of the endocrine and nervous systems. Although they are most commonly found in the gastrointestinal (GI) tract, they can arise throughout the body.

In addition to being classified by their anatomical location, NETs are divided according to whether they are functional or nonfunctional— that is, whether or not they secrete hormones such as insulinoma, gastrinoma, or glucagonoma— and by how well differentiated they are.

Surgery is currently the only potentially curative treatment option, but tumors frequently return and metastasize. Chemotherapy is considered the standard of care for patients with poorly differentiated tumors, which tend to be highly aggressive, but the options have been expanding for well-differentiated, relatively indolent tumors.

Somatostatin receptors are highly expressed on the surface of many NETs and receptor analogs have a central place in the treatment of NETs. In December 2014, the FDA approved the somatostatin analog lanreotide (Somatuline Depot) for patients with unresectable, well- or moderately differentiated locally advanced or mestastatic gastroenteropancreatic NETS after clinical trial data demonstrated an improvement in progression-free survival (PFS).

The identification of alterations in the mammalian target of rapamycin (mTOR) pathway, as well as aberrant expression of proangiogenic molecules including vascular endothelial growth factor (VEGF) and the highly vascularized nature of NETs, has driven the development of two alternative therapeutic strategies.

Several mTOR and VEGF receptor inhibitors have shown clinical activity in NETs. In 2011, the FDA approved sunitinib (Sutent), which inhibits VEGF receptor and other angiogenic kinases, and everolimus (Afinitor), which targets mTOR, for the treatment of patients with pancreatic NETs (pNETs).

A Clinical Challenge

The indication for everolimus was expanded to include nonfunctional gastrointestinal and lung NETs earlier this year, based on data from the phase III RADIANT-4 trial in which everolimus reduced the risk of progression by 52% (HR = 0.48; P <0.001) compared with placebo and increased median PFS by 7.1 months as assessed by central radiology review (11.0 months vs 3.9 months, respectively; P <.001).Despite such notable advances, these drugs are not effective in many patients, and there is a pressing need for new therapeutic options. NETs present a significant clinical challenge, and survival rates have remained stubbornly unchanged for decades, for the most part because these rare tumors are difficult to study.

While they may look identical under the microscope, these tumors often respond very differently to treatment and have varied prognoses. Relatively few genome sequencing studies have been performed to date, but they have begun to reveal important molecular distinctions between the different types of NETs.

On the whole, the mutation rate of NETs is significantly lower than for other types of cancer from the same primary location, suggesting they are more genetically stable. In general, NETs have a paucity of somatic mutations, though the frequency of mutations has been shown to increase with higher tumor grade, which may partly explain why poorly differentiated tumors tend to be more aggressive in clinical course than well-differentiated ones.

Distinct Tumor Types Revealed

Small Intestine NETs

The classical tumor suppressors and oncogenes implicated in the development of other tumors do not seem to play a significant role in the pathogenesis of NETs. Even within groups of NETs from the same primary site, there are genomic distinctions. A picture has begun to emerge of biologically unique entities that may require very different treatment strategies.Most sequencing studies to date have been performed in patients with pNET and small intestine NETs (SI-NET).In addition to being the most common malignancy of the small intestine, SI-NETs make up the largest group of NETs by primary site. These tumors have been analyzed in two separate genome sequencing studies.

In the first, an analysis of mutations and copy number changes in 48 tumors was performed. The rate of somatic mutations was low and although a number of genes were mutated, none were recurrently altered. Both the frequency and type of mutations differed significantly from other GI cancers as well as other types of NETs.

Although no recurrently mutated genes were observed, many clustered into the same pathway. Almost half of the tumor samples had mutated or deleted SMAD genes, which encode proteins involved in the transforming growth factor beta (TGFβ) pathway, suggesting it may be an important regulator of SI-NET growth.

Drugs targeting this pathway might be an effective treatment option in SI-NET and some are in clinical development for other tumor types. The study authors also suggested that patients with SMAD mutations may be less likely to respond to 5-fluorouracil, as has previously been observed in patients with colorectal cancer.

Genes involved in the phosphatidylinositol- 3-kinase (PI3K)/Akt/mTOR pathway, a key regulator of numerous hallmark cellular processes involved in the development of cancer, were mutated in around a third of tumors, particularly amplification of the AKT1 or AKT2 genes, which were mutually exclusive events.

Amplification of the SRC gene was also frequently observed in SI-NET. SRC encodes a nonreceptor tyrosine kinase, and a number of multitargeted tyrosine kinases include Src among their targets. Src also has been shown to activate mTOR activity in neuroendocrine cells, offering the possibility of synergistic activity between mTOR and Src-targeting therapies.

Pancreatic NETs

In the second study, 55 tumors were profiled with whole-exome and whole-genome sequencing. Again, the vast majority of mutations were found in only a single patient. Statistically significant mutations were observed in a single gene, CDKN1B, in 10% of patients. This gene encodes a cyclin-dependent kinase inhibitor p27, which regulates the cell cycle.pNETs are the second most common NET after SI-NETs. The first whole-exome sequencing study of pNET was published in 2011 and analyzed 58 tumor samples. In addition to being distinct from SI-NET, the pNETs had a mutation profile that was distinct from other types of pancreatic cancer; the genes most frequently mutated in pancreatic ductal adenocarcinoma were rarely altered in pNET.

Three commonly mutated genes emerged: MEN1, DAXX, and ATRX. DAXX and ATRX mutations, observed in 43% of cases collectively, were mutually exclusive, but were found in combination with MEN1 mutations in around one-fifth of tumors. MEN1, found in 44% of cases, encodes a nuclear protein called menin that interacts with a complex of other proteins that remodel chromatin by methylating histone proteins, ultimately regulating gene expression.

In particular, menin seems to be involved in the activation of genes involved in the cell cycle and the repair of damaged DNA. Thus, MEN1 acts as a tumor suppressor and its loss may serve a dual blow in pNET cells, promoting excessive cell growth and genome instability. What the ATRX and DAXX genes do and how they lead to pNET is much less clear.

Like menin, both are nuclear proteins. ATRX is a member of the SWItch/Sucrose Non-Fermentable (SWI/SNF) complex, and both it and DAXX are involved in chromatin remodeling. It is suggested that they might play an important role in a process called alternative lengthening of telomeres (ALT).

Telomeres are specialized structures that form a protective cap on the end of the chromosomes. They are progressively eroded over time and this limits the replicative potential of a cell. In order to become immortal, cancer cells have to overcome telomere shortening and they usually achieve this goal by reactivating the telomerase enzyme, which lengthens telomeres. However, in a minority of cases they can use an alternative mechanism—ALT&mdash;which is much more poorly understood. ALT was noted in 60% of cases of pNET and these correlated with mutations in ATRX and DAXX.

Genes involved in the mTOR pathway, including tuberous sclerosis gene (TSC2) and phosphatase and tensin homolog gene (PTEN), were also mutated (8.8% and 7.3%, respectively). Both regulate the activity of mTOR, perhaps explaining why mTOR inhibitors have proved so successful in pNET in particular.

A second study of pNETs also identified mutations in MEN1, ATRX, and DAXX as well as TSC2, but also noted alterations in TP53, KRAS, and VHL, which were infrequently mutated in the previous study. It was suggested that differences in the mutational profiles may have resulted from the different patient populations, which were 85% Caucasian in the former and 100% Chinese in the latter.

In a separate study, researchers specifically focused on poorly differentiated neuroendocrine carcinomas of the pancreas (PNEC), which have a particularly poor prognosis and are highly aggressive. This group includes small cell and large cell histologies and is generally very rare. The molecular alterations found in these tumors was distinct from other, well-differentiated pNETs.

The most commonly mutated genes were TP53 and RB1 and abnormal immunolabeling of these two proteins by immunohistochemistry was also almost universal. TP53 and RB1 mutations are rare, by contrast, in well-differentiated pNETs; however, they may be inactivated by other mechanisms, such as extra copies of p53- and pRb-related genes including MDM2, MDM4, and WIP1 (p53) and CDK4/6, CCND1 (pRb).

Pulmonary NETs

Finally, a small sequencing study of insulinomas, the most common type of functional NET, was also performed in 10 tumor samples and matched normal pairs. The most notable mutations were in the YY1 transcription factor (30%). Mutations in other potential cancer-related genes included MLL3, H3F3A, and LMO2.Around a quarter of lung tumors are neuroendocrine in nature, with small cell lung cancer (SCLC) being the most common and the most malignant type. A limited number of sequencing studies have been performed and have mostly focused on carcinoid tumors.

Altered genes involved in chromatin remodeling were frequently identified in this tumor type, including those encoding proteins involved in histone modification (in 40% of tumors), such as MEN1, PSIP1, HDAC5, and BRWD3, and in the SWI/SNF complex (in 22%) such as the ARID1A and SMARCA genes. The MEN1, PSIP1, and ARID1A genes were recurrently mutated. TP53 and RB1 mutations were rare.

Another study sequenced 70 pulmonary NETs, including typical carcinoids, atypical carcinoids, large cell neuroendocrine carcinomas (LCNECs) and SCLCs. Mutations in JAK3, NRAS, RB1, and VHL1 were exclusively found in SCLCs, FGFR2 mutations were unique to LCNEC. KIT, PTEN, HNF1A, and SMO were altered in atypical carcinoids and SMAD4 was found in typical carcinoids. The frequency of mutations observed overall increased as the malignancy of the subtype increased.

Other NETs

The most frequently mutated gene in pulmonary NETs overall was TP53 (22%), but this mutation was not found in carcinoids and was most prevalent in SCLC, highlighting the genomic diversity among different subtypes. Mutations common to all pulmonary NET subtypes included RET and EGFR, though the latter are not the usual activating mutations observed in other tumor types.There are numerous other types of less common NETs. Among prostate cancers, true NETs are rare, but often prostate adenocarcinoma can evolve a neuroendocrine phenotype. A growing body of evidence suggests that the evolution of the neuroendocrine phenotype in castrationresistant prostate cancer (CRPC) is one of leading causes of resistance to androgen receptor (AR)-targeted therapies.

A recent study published in Nature Medicine suggested that the selective pressure of AR-targeted therapy causes genomic aberrations in CRPC that drive clonal evolution of AR-positive CRPC adenocarcinoma into an AR-negative CRPC NET. The study further concluded that these genomic alterations were most likely epigenetic rather than genetic in nature.

Medullary thyroid cancers (MTC) are also a form of NET. Whole-exome sequencing of 17 tumors found that 90% tumors had mutually exclusive mutations in RET, HRAS, and KRAS.

An Epigenetic Malignancy?

There were relatively few other mutations and no common recurrent mutations. RET is a proto-oncogene that encodes a receptor tyrosine kinase. Mutations and gene rearrangements in RET have been recognized in MTC for more than a decade, and drugs that inhibit RET including vandetanib (Caprelsa) and cabozantinib (Cometriq) have been successfully developed for its treatment.It is clear from genome sequencing studies that somatic mutations are rare in NETs compared with other tumor types, and those that are present are often not recurrently altered and many may represent passenger mutations rather than oncogenic drivers. What makes the management of NETs more challenging still is that most recurrently mutated genes encode tumor suppressor proteins that are particularly difficult to target therapeutically.

Epigenetic changes, which impact gene expression without affecting the DNA sequence, have also been evaluated in genome-wide studies and are much more common. In addition, when genes are mutated in NETs they are often epigenetic modifiers involved in processes like chromatin remodeling. This has led many researchers to suggest that epigenetic regulation may play a more important role in the development of NETs than the acquisition of mutations.

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

Key Research

• Agrawal N, Jiao Y, Sausen M, et al. Exomic sequencing of medullary thyroid cancer reveals dominant and mutually exclusive oncogenic mutations in RET and RAS. J Clin Endocrinol Metab. 2013;98(2):E364-E369.
• Banck MS, Kanwar R, Kulkarni AA, et al. The genomic landscape of small intestine neuroendocrine tumors. J Clin Invest. 2013;123(6):2502-2508.
• Beltran H, Prandi D, Mosquera JM, et al. Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer. Nature Med. 2016;22(3):298-305.
• Capdevila J, Meeker A, García-Carbonero R, et al. Molecular biology of neuroendocrine tumors: from pathways to biomarkers and targets. Cancer Metastasis Rev. 2014;33(1):345-351.
• Feldman R, Astsaturov I, Millis S, et al. Molecular profiling in small cell lung cancer and lung neuroendocrine tumors. Int J Radiat Oncol. 2014;90(5S):suppl;pS7.
• Francis JM, Kiezun A, Ramos AH, et al. Somatic mutation of CDKN1B in small intestine neuroendocrine tumors. Nat Genet. 2013;45(12):1483-1486.
• Fernandez-Cuesta L, Peifer M, Lu X, et al. Frequent mutations in chromatin-remodeling genes in pulmonary carcinoids. Nat Commun. 2014;5:3518-3535.
• Jiao Y, Shi C, Edil BH, et al. DAXX/ATRX, MEN1 and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors. Science. 2011;331(6021):1199-1203.
• Karpathakis A, Dibra H, Thirlwell C. Neuroendocrine tumors: cracking the epigenetic code. Endocr Relat Cancer 2013;20(3):R65-R82.
• Vollbrecht C, Werner R, Walter RFH, et al. Mutational analysis of pulmonary tumors with neuroendocrine features using targeted massive parallel sequencing: a comparison of a neglected tumor group. Br J Cancer 2015;113(12):1704-1711.
• Yachida S, Vakiani E, White CM, et al. Small cell and large cell neuroendocrine carcinomas of the pancreas are genetically similar and distinct from well-differentiated pancreatic neuroendocrine tumors. Am J Surg Pathol. 2012;36(2):173-184.