|Home | About | Journals | Submit | Contact Us | Français|
Traditional serrated adenomas (TSAs) are a type of colorectal polyp with neoplastic potential. Immunohistochemical analysis and sequencing were performed on 24 TSAs from 23 patients to characterize the molecular genetics of TSAs. Abnormal Ki-67 and p53 labeling were observed in 7 (29%) of 24 and 6 (25%) of 24 TSAs, respectively; both types were significantly associated with the presence of conventional epithelial dysplasia (P = .0005 and P = .0001, respectively). Activating KRAS mutation was identified in 11 TSAs (46%) and was mutually exclusive with activating BRAF mutations, which were seen in 7 TSAs (29%). Abnormal p53 nuclear labeling in a TSA was significantly associated with BRAF mutation status (P = .04), whereas no relationship was found for β-catenin labeling patterns. The overall morphologic features of TSA do not correlate with the genetic status of the KRAS and BRAF genes. However, conventional epithelial dysplasia and abnormal p53 labeling in a TSA are seen more often in the setting of BRAF mutation.
Serrated colorectal polyps consist of hyperplastic polyps, sessile serrated adenomas (SSAs), and traditional serrated adenomas (TSAs).1,2 Whereas hyperplastic polyps are benign “metaplastic” polyps, SSAs and TSAs have neoplastic potential.3,4 This notion is supported by the finding of high-grade dysplasia or cancer arising in SSAs and TSAs, and the association of preexisting SSAs or TSAs with sporadic colon cancer.5-9 Mixed hyperplastic polyptubular adenomas likely represent SSAs or TSAs that have developed conventional epithelial neoplasia. The 2010 World Health Organization classification defines such lesions as “SSAs/polyps with cytological dysplasia.”10,11 Studies by several investigators have clarified the molecular genetics of SSAs, and indicate that activating mutations in the BRAF oncogene are an early event in the genetic progression of SSA that is followed by Wnt pathway activation by epigenetic mechanisms.5,10,12,13 Microsatellite instability is also a defining feature of SSAs that occurs in association with the development of low- or high-grade dysplasia.5,7
Although the molecular features of SSAs are perhaps better understood, TSAs were a form of serrated polyp with neoplastic potential initially recognized by Longacre and Fenoglio-Preiser in 1990.14 In that study, what is currently regarded as a TSA was originally defined as a polyp with serrated architecture in association with adenomatous change. However, because of overlapping morphologic descriptions among SSAs and TSAs in published literature, the distinction between these entities has remained unclear. This has been confounded by studies suggesting that TSAs contain features of both sporadic and microsatellite instability–associated neoplasia.15 In 2008, Torlakovic et al3 performed a comprehensive study of TSA and noted that morphologic features that distinguish TSA from SSA include the presence of ectopic crypt formations (ECFs) and, though less specific, cytoplasmic eosinophilia. Moreover, whereas SSAs are typically right-sided polyps, TSAs are more commonly seen in the left colon.3,4,14,16,17 The neoplastic nature of TSA has also been questioned, because TSA may be nonproliferative on Ki67 labeling.3
Given these controversies, and the newfound understanding of the distinct morphologic features of TSAs,3 we sought to clarify the molecular genetics of this form of serrated polyp. This study builds on current understanding of the molecular features of TSAs.3,4,14,16,17 From a practical standpoint, an improved understanding of the features of TSA, including their relationship to other forms of serrated neoplasia and their relative risk of neoplastic progression, remains an important issue for guiding the treatment of patients diagnosed with serrated forms of colorectal polyps.
We performed a search of the Johns Hopkins University (Baltimore, MD) pathology archives using the terms traditional serrated adenoma or left colon and serrated adenoma spanning the period January 1999 to December 2010. This identified 57 potential TSAs for study. Slides of each were reviewed and TSAs were identified based on the criteria of any polypoid lesion with serrated epithelial architecture in association with ECFs and/or eosinophilic cytoplasm.3 ECFs were recognized based on their loss of orientation of the crypt base toward the muscularis mucosa.18 Features of conventional-type epithelial dysplasia in a TSA were considered present when extensive nuclear crowding, nuclear enlargement, complete loss of nuclear polarity, pseudostratification extending to the upper half of the neoplastic cell, and increased mitoses were seen.11,16 After slide review, 24 TSAs that were well oriented on routinely stained sections, lacked cautery artifacts, and had available paraffin blocks were selected for further study. Clinicopathologic data of all patients whose polyps were used for the study were collected from the corresponding pathology reports. The project was approved by the institutional review board before collection of samples for study.
Immunohistochemical labeling was performed using standard methods. Unstained 5-μm sections were cut from paraffin blocks and the slides were deparaffinized using routine techniques followed by incubation in ×1 sodium citrate buffer (diluted from ×10 heat-induced epitope retrieval buffer, Ventana-Bio Tek Solutions, Tucson, AZ) before steaming for 20 minutes at 80°C. Slides were cooled 5 minutes and incubated with a panel of antibodies consisting of prediluted Ki67 (catalog 790-2910, Ventana Medical Systems, Tucson), 1:1000 dilution β-catenin (catalog 610154, BD Transduction Laboratories, Lexington, KY), and prediluted p53 (catalog 760-2542, Ventana). Immunolabeling was detected using kit instructions (catalog 760091, Ventana IVIEW detection kits).
β-Catenin labeling was evaluated with respect to membranous and/or nuclear localization as previously described in detail.10 Briefly, membranous labeling was considered normal, whereas the labeling pattern for β-catenin was considered abnormal when nuclear labeling was accompanied by a reduction or loss of membranous labeling outside the crypt bases where the β-catenin positive progenitor population normally resides. Nuclear accumulation of p53 in 30% or more of the adenomatous epithelium was considered an abnormal labeling pattern.19 Ki67 labeling was evaluated for both percentage of positive nuclei per 1,000 epithelial cells and the distribution of positive labeling nuclei within ectopic crypt foci vs intervening surface epithelium as described.3 Slides were scored using a 2-headed microscope by 2 of the authors (B.F. and C.A.I.-D.) blinded to the genetic status of each polyp.
Regions of TSAs from unstained sections were dissected from unstained 10-μm-thick sections. Genomic DNA was extracted from each sample by phenol-chloroform and 20 ng used for polymerase chain reaction (PCR) amplification of KRAS exon 2 and BRAF exon 15 using intronic primers flanking these exons.10 PCR products were sequenced in both directions with an M13F primer (5'-GTAAAACGACGGCCAGT-3') and an M13R primer (5'-CAGGAAACAGCTATGACC-3') that were incorporated into the forward and reverse primer of each primer pair, respectively (Agencourt Bioscience, Beverly, MA). Sequence data were analyzed with Sequencher 4.8 software (Gene Codes, Ann Arbor, MI). All mutations were verified by means of bidirectional sequencing of a second PCR product derived independently from the original template.
For comparing parametric distributions a 1-sided Student t test was used, and for frequency distributions, a χ2 test or Fisher exact test was used for values less than 5. P values less than or equal to .05 were considered statistically significant.
Clinicopathologic features of all 23 patients are shown in Table 1. Among all patients, 15 (65%) were male and 22 (96%) were white. The mean age was 61.0 ± 15.3 years (range, 40-86 years). Twenty-four TSAs were studied, of which 23 (96%) were located in the left colon. The mean size of all 24 TSAs was 0.9 ± 0.6 cm (range, 0.1 to 2.5 cm). Of these 24 polyps, 20 (83%) were originally diagnosed as a serrated adenoma or TSA, and 4 (17%) were diagnosed as a mixed serrated adenoma/tubular adenoma. Of interest, 12 (50%) of 24 patients had a past or subsequent finding of a tubular adenoma, and 3 (13%) of 24 had a past or subsequent finding of an SSA or TSA.
The salient features of TSAs studied are shown in Image 1 and Image 2. Histologic review of each TSA also indicated that 3 (13%) were associated with an adjacent sessile lesion with morphologic features similar to SSA, seen as a flat growth pattern with associated broad-based crypts and basilar crypt branching (Image 1E).
All TSAs are considered dysplastic lesions.14 However, TSAs in the current study showed a range of nuclear and cytologic atypia, from uniform pencillate nuclei with slightly hyperchromatic chromatin in 19 (79%) of 24 TSAs (Image 1B and Image 1D) to high nuclear/cytoplasmic ratio, nuclear crowding, extensive pseudostratification, and a relatively higher mitotic rate in 5 (21%) of 24 TSAs (Image 2). Based on criteria for evaluating dysplasia in colorectal polyps,11,16 these latter changes are suggestive of conventional epithelial dysplasia arising in a TSA. Changes consistent with conventional epithelial dysplasia were seen focally in 3 TSAs (Image 2A) and were predominant in 2 (Image 2B, Image 2C, and Table 1).
To better understand the significance of conventional epithelial dysplasia in a TSA (“TSAD”), we first determined the relationship of conventional dysplasia to clinicopathologic findings in the same patients. No significant differences were seen in the mean age, race or gender distribution, or location between TSAs and TSADs (Table 1). However, compared with TSAs, TSADs were more likely to be larger than 1.0 cm (P = .002).
To determine the extent of cellular proliferation in TSAs and TSADs, each polyp was immunolabeled for Ki67. All polyps showed positive labeling for Ki67 in the adenomatous epithelium, though labeling corresponded to 2 patterns Image 3. In the first pattern (crypt predominant pattern), Ki67 labeling was confined to the ECFs and/or deep crypts in the polyp fronds. Labeling of the surface epithelium was either absent or limited to scattered epithelial cell nuclei (Image 3B). By contrast, in other cases, the Ki67 labeling was noted throughout the epithelium when both the ECFs and surface epithelial cells intervening between the ECFs showed diffuse Ki67 labeling (diffuse pattern, Image 3D). No difference was found in the overall percentage of positive labeling Ki67 nuclei among TSAs and TSADs (7.2% vs 7.6%, P = .44) Table 2. However, when analyzed based on Ki67 labeling pattern, labeling of surface epithelium was seen in only 2 (11%) of 19 TSAs but in all 5 (100%) TSADs (P = .0005) (Image 3F).
To determine the molecular features characteristic of TSAs with and without conventional epithelial dysplasia, genomic DNA was extracted from all 24 polyps for subsequent sequencing for mutations in exon 2 of KRAS and exon 15 of BRAF. Mutations of KRAS were found in 11 (46%) of 24 TSAs and mutations of BRAF (V600E) were found in 7 (29%) of 24 TSAs Image 4 (Table 2). The presence of a KRAS mutation was mutually exclusive with the presence of a BRAF mutation; in total KRAS or BRAF mutations were encountered in 18 (75%) of 24 TSAs. No difference was noted in the clinicopathologic features of KRAS-mutant vs BRAF-mutant TSAs, including polyp size, location, presence or absence of conventional epithelial dysplasia, or Ki67 labeling patterns Image 5 and Table 3. We also determined if KRAS or BRAF mutation status correlated with the presence of an adjoining SSA-like lesion seen for 3 TSAs. However, no relationships were found because 1 TSA was KRAS mutant, 1 was BRAF mutant, and in the final TSA, no mutations were found.
To clarify the molecular features associated with development of a TSAD, each polyp was immunolabeled for β-catenin and p53 protein. Normal labeling for β-catenin was noted in 20 (83%) of 24 TSAs Image 6A and Image 6E, seen as positive membranous labeling within the bases of ECFs and deep crypts in association with scattered positive labeling nuclei. Of interest, 4 (20%) of these 20 TSAs showed a unique tigroid-like pattern of β-catenin labeling categorized by uniformly distributed patches of intense membranous labeling throughout the adenomatous epithelium Image 6B. By contrast, weak nuclear labeling of β-catenin that extended outside the crypt bases, in association with loss of membranous labeling, was identified in 4 (17%) of 24 TSAs Image 6C. No relationship was found between β-catenin labeling patterns and the presence or absence of conventional epithelial dysplasia (Table 2).
Nuclear accumulation of p53 was noted in 6 (25%) of 24 TSAs Image 6D and Image 6F, consistent with reports of p53 inactivating mutations in serrated adenomas of the colorectum.20 When further analyzed in relation to conventional epithelial dysplasia, abnormal p53 labeling was present in 1 (5%) of 19 TSAs and in 5 (100%) of 5 TSADs (P = .0001, Table 2). Three of 5 TSADs with abnormal p53 labeling had focal areas of conventional epithelial dysplasia, indicating that the conventional dysplasia developed within a preexisting TSA. A comparison of the p53 labeling patterns in the regions of classic TSA vs conventional epithelial dysplasia indicated that p53 nuclear accumulation had either preceded or coincided with the development of conventional dysplasia in all 3 cases. Finally, we compared β-catenin and p53 labeling patterns to the genetic status of KRAS and BRAF in the same TSA. Although no relationship was found for β-catenin, abnormal nuclear accumulation of p53 was significantly more common in TSAs with BRAF mutations (P = .04, Table 3).
This study provides a clinicopathologic and genetic analysis of TSAs since the introduction of the refined diagnostic criteria by Torlakovic et al.3 In agreement with prior reports, TSAs are serrated lesions with neoplastic potential, and a subset of TSAs shows focal or extensive features of conventional epithelial dysplasia in the same lesion.4,14 Thus, the presence of conventional epithelial dysplasia in a TSA has biologic significance because it likely represents a morphologic change associated with genetic progression to a TSA-derived infiltrating carcinoma. This notion is supported by our observation that conventional epithelial dysplasia is significantly more common in larger TSAs, and abnormal nuclear accumulation of p53 protein is significantly more common in TSAs with conventional epithelial dysplasia.
Although TSAs may be represented by those with KRAS vs those with BRAF mutations, it is also conceivable that a subset of so-called TSAs with BRAF mutations includes SSAs with prominent exophytic growth, perhaps as a consequence of their left-sided location. Overall, this observation was likely not appreciated until now because many studies stratified polyps based on morphologic features rather than their genetic alterations, and morphologic differences do not reliably distinguish TSAs with a KRAS mutation from those with a BRAF mutation.4,5,8,21-23 This is indirectly supported by the presence of microsatellite instability in some reported TSAs,4,5 the identification of BRAF mutations as an almost universal feature of SSAs,5,10 and the development of conventional epithelial dysplasia or infiltrating carcinoma in SSAs as well.7 Moreover, genetic progression in a TSA, characterized by the development of conventional epithelial dysplasia in association with abnormal nuclear accumulation of p53 protein, is significantly more common in TSAs that contain BRAF mutations than in those with KRAS mutations; p53 accumulation has also been described as a feature of SSAs with high-grade dysplasia.24,25 We did not find a relationship of β-catenin nuclear labeling with genetic progression of TSAs as previously reported for SSAs.10,12 However, our sample size was not sufficient to fully exclude this possibility, and intriguingly 2 (40%) of 5 TSADs did show nuclear labeling for β-catenin compared with only 2 (12%) of 17 TSAs. Of note, nuclear labeling for β-catenin was also reported in a subset of filiform TSAs, though the relationship of this finding to genetic features of the same polyps was not explored.4 Overall, the finding of conventional epithelial dysplasia may signify the presence of a BRAF mutation in an actual left-sided SSA, but in the absence of this feature, the neoplastic potential of a left-sided TSA remains unknown. Until clarified, TSAs should nonetheless be managed similar to other adenomatous polyps in this location.2,26
Interestingly, despite their overlapping morphologic features, TSAs with KRAS mutations may represent a different entity from those with BRAF mutations.23 In support of this notion, mouse models of KRAS activation in the colorectal mucosa led to crypt hyperplasia and serrated features analogous to those seen in TSAs.27,28 Moreover, KRAS mutations have been described as a distinguishing feature of goblet cell hyperplastic polyps,5,22 and it is conceivable that goblet cell hyperplastic polyps, as opposed to the more common microvesicular type, represent the earliest phase of development of a KRAS-mutant TSA.22 Colorectal polyps that arise because of KRAS mutations deviate from the typical adenoma-carcinoma sequence that is initiated by inactivating mutations in the adenomatous polyposis coli (APC) tumor suppressor gene, leading to activation of Wnt signaling. KRAS mutations subsequently occur after APC inactivation, leading to formation of early tubular adenoma.29 By contrast, KRAS mutations in the absence of APC genetic inactivation may lead to formation of a TSA that recapitulates normal crypt formation and maturation. This does not rule out the possibility that KRAS-mutant TSAs may also have neoplastic potential, as seen in a series of polyps from Korean individuals.8,9
In summary, we found that polyps with features with TSA are represented by 2 distinct genetic variants, and that TSAs may undergo a genetic progression as described for other colorectal polyps with neoplastic potential.
Upon completion of this activity you will be able to:
Supported by National Institutes of Health grants CA140599 and P50 CA62924 and The Uehara Memorial Foundation.
The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.