Molecular characterization of sessile serrated adenoma/polyps with dysplasia/carcinoma based on immunohistochemistry, next-generation sequencing, and microsatellite instability testing: a case series study

In 2003, Torlakovic et al. [1] reported evidence of abnormal proliferation in colorectal serrated polyps that superficially resembled hyperplastic polyps (but that could be distinguished histologically based on their abnormal architectural features) and introduced the terms “sessile serrated polyp” and “sessile serrated adenoma” to describe their observations. Currently, this category is designated as “sessile serrated adenoma/polyp (SSA/P),” as recommended by the World Health Organization [2]. SSA/P is considered as an early precursor lesion in the serrated neoplasia pathway, which largely results in colorectal carcinomas with high levels of microsatellite instability (MSI-high) [3‐5]. Recent studies have shown associations of SSA/Ps and those with dysplasia/carcinoma with DNA methylation or lost protein expression of DNA-repair genes (i.e., MLH1) [1, 4, 6‐9], a CpG island methylator phenotype [3, 4, 6, 7], and BRAF mutations [3, 4, 6‐13]. This pathway is thought to be distinct from the conventional adenoma-carcinoma pathway, where adenomas progress to invasive colorectal carcinomas through the influence of several genetic alterations including adenomatous polyposis coli (APC) and KRAS mutations [4, 6, 10, 11, 14, 15].

Recently, targeted next-generation sequencing (NGS) has shown unprecedented potential for detecting underlying changes in the genetic architecture of cancer in a comprehensive and economically feasible manner, and the development of its platforms has enabled comprehensive analysis of genetic alterations in tumors [16, 17]. A comprehensive understanding of the genetic alterations associated with cancer could improve the molecular biological classifications of tumors and identify effective molecularly targeted therapies.

At present, the molecular biological features underlying the development of colorectal serrated neoplasia remain unclear. Hence, the aim of this study was to employ NGS to elucidate the molecular biological features of SSA/Ps with dysplasia/carcinoma, representing relatively early stages of the serrated neoplasia pathway.

Rare occurrences of BRAF mutations have been documented for conventional colorectal carcinomas, although they are frequent in dysplasia/carcinoma arising from SSA/Ps (50–90%) [3, 4, 9, 10, 12, 13]. In this study, BRAF mutations were detected in 7 of 8 SSA/P lesions, in accordance with previous reports [3, 4, 9, 10, 12, 13]. BRAF mutations result in activation of the RAS-RAF-MAPK pathway, and therefore our findings and previous reports [3, 4, 9, 10, 12, 13] suggest that those signaling pathways might be activated in serrated neoplasia.

We investigated the core protein-expression levels of MUC2, MUC5AC, MUC6, and CD10 because altered expressions of these proteins may be significantly correlated with the biological behavior of colorectal carcinoma and, possibly its prognosis. In this study, MUC2, MUC5AC, and MUC6 expression was frequently positive in the serrated lesions. MUC2 is a goblet cell-type mucin predominantly expressed in the colon. In contrast, MUC5AC and MUC6 are two gastric type mucins that are expressed in the surface foveolar epithelium and deep antral/pyloric glands, respectively, but not normally in colonic mucosa. MUC5AC and MUC6 expression have been strongly associated with tumorigenesis via the serrated neoplasia pathway [18‐20], in agreement with our current findings.

The WNT signaling pathway involves β-catenin and plays a crucial role in the development of colorectal carcinomas through a conventional adenoma-carcinoma progression [21]. Several previous reports including our own [8, 9, 12, 13, 22] have also demonstrated that nuclear β-catenin accumulation is common in SSA/Ps with dysplasia and invasive carcinoma, but not in those without dysplasia. This observation implies that activation of the WNT/β-catenin signaling pathway is involved in the development of dysplasia in SSA/Ps and progression to carcinoma. Although APC and CTNNB1 mutations are major causes of WNT/β-catenin signaling activation in conventional-type adenomas, these genetic alterations are absent in SSA/Ps [13]. In this study, all serrated lesions examined also displayed nuclear β-catenin immunoreactivity, whereas no mutations in WNT/β-catenin signaling-associated genes (such as APC and CTNNB1) were found in the studied lesions. Previously, we reported that activation of the WNT/β-catenin signaling pathway might be associated with methylation of the associated genes, including SFRP4, MCC, and AXIN2 [9]. This association could explain at least some of the discrepancies between β-catenin immunoreactivity and the lack of associated gene mutations.

The MSI phenotype has been regarded as a main subtype of colorectal cancers. Sporadic colorectal cancers with the MSI-high phenotype account for approximately 3–15% of all colorectal cancers [23]. MSI is a unique molecular alteration induced by deficiencies in the DNA-mismatch repair system and is characterized by unstable (variable length) microsatellites, a type of simple DNA sequence repeat. Some previous studies have shown that serrated neoplasia is associated with MSI-high colorectal carcinomas, which is accompanied with DNA methylation or loss of protein expression of DNA-repair genes such as MLH1 [3‐5]. In our study, all 4 MSI-high SSA/P lesions showed a loss of MLH1 expression, in accordance with previous reports [3‐5].

A recent report by the Cancer Genome Atlas Project elucidated the molecular landscape of colorectal cancers and revealed that the hypermutated phenotype mainly overlaps with MSI status in colorectal cancers [24]. Many types of genetic mutations can occur in MSI-high colorectal cancers. Indeed, a previous report showed that mutations in various genes, including the tumor-suppressor gene PTEN and the oncogene PIK3CA, were caused by the instability of microsatellites in MSI-high colorectal cancers [25]. In this study, no MSI-high lesions harbored PTEN or PIK3CA mutations. However, it is interesting that 3 of 4 MSI-high lesions harbored an FBXW7 mutation, whereas no MSS lesions harbored an FBXW7 mutation. FBXW7 is a tumor-suppressor gene located on human chromosome 4q that encodes a substrate-recognition component of SKP1–Cullin1–F-box protein-ubiquitin E3 ligase complexes [26]. These specific E3 ligase complexes negatively regulate the intracellular abundance of an expanding list of key oncogenic proteins such as cyclin E [27], c-JUN [28, 29], c-MYC [30, 31], MCL1 (myeloid cell leukemia 1) [32, 33], NOTCH [34‐36], AURKA (aurora kinase A) [37, 38], KLF5 (Krüppel-like factor 5) [39], mTOR [40], and TGIF1 [41]. Therefore, the loss of FBXW7 function results in accumulation of its substrates, which leads to oncogenesis and progression of multiple cancers including colorectal cancers [42, 43]. A study of over 500 primary tumors of diverse tissue origins suggested that FBXW7 mutations occurred in approximately 6% of all evaluated tumors. Of these, the most commonly affected tumors were cholangiocarcinoma (35%), T-cell acute lymphocytic leukemia (31%), endometrial cancer (9%), and gastric cancer (6%) [43]. FBXW7 has also consistently been identified as one of the most commonly mutated genes in colorectal cancer, being observed in 6 to 10% of all cases [43, 44]. Furthermore Chang et al. reported that FBXW7-mutant colorectal cancers had significant associations with MSI-high tumors in a large-scale study of 1519 cases [45]. Our findings and those previous reports indicated that FBXW7 mutations might potentially be involved in the progression of MSI-high serrated lesions.

Activation of the second arm of the serrated neoplasia pathway, also driven by CpG island methylation of unspecified tumor-suppressor genes but not DNA-repair genes, may indicate progression to BRAF-mutant MSS carcinoma [46, 47]. Furthermore, as reported in earlier works [48‐50], a strong inverse correlation was found between TP53 alterations and the MSI phenotype. In this study, 2 of 4 lesions with MSS harbored a TP53 mutation, whereas no MSI-high lesions harbored a TP53 mutation, similar to previous observations [48‐50]. The TP53 gene encodes a tumor-suppressor protein containing transcriptional-activation, DNA-binding, and oligomerization domains. The encoded protein responds to diverse cellular stresses to regulate the expression of target genes, thereby inducing cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism [51]. Mutations in the TP53 gene are reportedly associated with a variety of human cancers, including colon, breast, lung, and brain cancers [51]. Our findings and those previous reports indicated that TP53 mutations are potentially involved in the progression of MSS serrated lesions. A schematic depiction of differences in the expression of key proteins and genetic alterations in the serrated neoplasia pathway is shown in Fig. 4.

Our study had a major limitation; the sample size was very small. A major impediment to investigating dysplasia/carcinoma arising in SSA/P is the rarity of this lesion. Additionally, a sufficient amount of specimen is necessary for NGS analysis, which limited the lesions that could be studied to those of 10 mm in diameter or more. These limitations lowered the statistical power of this study. Particularly, although the high rate of FBXW7 or TP53 mutations in the studied lesions is important, the possibility that this result was a chance occurrence cannot be denied, considering that only 2 or 3 cases of the studied lesions harbored TP53 or FBXW7 mutations, respectively.

To the best of our knowledge, few comprehensive genetic studies have been conducted in SSA/Ps with dysplasia and invasive carcinoma. Our results revealed mutations in BRAF, FBXW7, TP53, KIT, PTEN, SMAD4, and SMARCB1. Furthermore, it is very interesting that all MSI-high lesions might be associated with MLH1 expression loss and mutation of FBXW7, but not TP53. Our results may help clarify the detailed mechanism of serrated neoplasia development. A comprehensive understanding of genetic alterations associated with the serrated neoplasia pathway could help identify effective molecularly targeted therapies.

In conclusion, colorectal SSA/Ps with dysplasia and invasive carcinoma frequently harbored BRAF mutations and showed nuclear β-catenin expression. Furthermore, these lesions might not only be associated with MSI-high colorectal cancer, but also MSS, and MSI-high and MSS serrated lesions might have distinct genetic features (such as FBXW7 and TP53 mutations). BRAF-mutant MSS colon carcinomas are particularly important because they have a dismal prognosis and an aggressive clinical course with adverse histologic features, such as lymphatic and perineural invasion and high tumor budding [52, 53]. Further investigations are required to elucidate molecular biological characteristics in the serrated neoplasia pathway in greater detail.