Abstract
Background
Intraductal papillary mucinous neoplasms (IPMN) are being increasingly recognized as important precursors to pancreatic adenocarcinoma. Elucidation of the genetic changes underlying IPMN carcinogenesis may improve the diagnosis and management of IPMN. We sought to determine whether different histologic subtypes of IPMN would exhibit different frequencies of specific genetic mutations.
Study Design
Patients with resected IPMN-associated invasive carcinoma (IPMN-INV) between 1997 and 2012 were reviewed. Areas of carcinoma, high-grade dysplasia and low-grade dysplasia were micro-dissected from each pathologic specimen. Targeted, massively-parallel sequencing was then performed on a panel of 275 genes (including KRAS, GNAS and RNF43).
Results
38 patients with resected IPMN-INV and sufficient tissue for microdissection were identified. Median follow-up was 2.6 years. GNAS mutations were more prevalent in colloid-type IPMN-INV than tubular-type IPMN-INV, 89% vs 32% respectively (p=0.0003). Conversely, KRAS mutations were more prevalent in tubular-type than colloid-type IPMN-INV, 89% vs 52%, respectively (p=0.01). For non-invasive IPMN subtypes, GNAS mutations were more prevalent in intestinal (74%) compared with pancreatobiliary (31%) and gastric (50%) subtypes (p=0.02). The presence of these mutations did not vary according to the degree of dysplasia (GNAS: invasive 61%, high-grade 59%, low-grade 53%, KRAS: invasive 71%, high-grade 62%, low-grade 74%), suggesting that mutations in these genes occur early in IPMN carcinogenesis.
Conclusions
IPMN-associated colloid carcinoma and its intestinal-type pre-invasive precursor are associated with high frequencies of GNAS mutations. The mutation profile of tubular carcinoma resembles conventional pancreatic adenocarcinoma. Preoperative determination of mutational status may assist with clinical treatment decisions.
Intraductal papillary mucinous neoplasms of the pancreas (IPMN) are mucin-producing neoplasms of the pancreas first categorized as a distinct entity by the World Health Organization in 1996. Since this first categorization, pathologists have recognized that this disease exhibits significant heterogeneity. Histopathologically, four subtypes of non-invasive IPMN have been described: gastric, pancreatobiliary, intestinal and oncocytic(1,2). Macroscopic lesions often contain several of these histopathologic sub-types, as well as varying levels of dysplasia. Invasive IPMN carcinoma (IPMN-INV) is found in approximately 40-60% of resected IPMN lesions, and there are two distinct histopathologic subtypes of IPMN-INV(3). Colloid carcinoma typically arises in the setting of intestinal-type IPMN, and tubular carcinoma typically arises in the setting of pancreatobiliary-type IPMN. The survival outcomes following resection of IPMNINV have been found to be associated with the type of invasive cancer: patients resected for colloid carcinoma have five-year survival rates of almost 75%, compared with 20-40% for those resected for tubular carcinoma.
This heterogeneity, and in particular the inability to determine degree of dysplasia or pathologic subtype preoperatively, has made IPMN a very challenging clinical problem. Operative treatment recommendations are currently based on radiographic and endoscopic findings(4). In the absence of a mass lesion suspicious for IPMN-INV, the strongest factor associated with high-grade dysplasia is dilation of the main pancreatic duct(5,6). There are currently no radiographic or endoscopic means of differentiating IPMN-INV subtype (colloid vs. tubular).
Recently, mutations in KRAS, GNAS, and RNF43 have been identified in the tissue and cyst fluid of patients with IPMN(7,8). Importantly, these mutations were not identified in the cyst fluid of patients with serous cystadenoma, and other common types of benign pancreatic cysts. The association of these mutations with the pathway of progression to IPMN-INV has not been clearly defined. While mutation in KRAS has been shown to be an early genetic event(9), recent studies identifying mutations in GNAS and RNF43 suggest that IPMN carcinogenesis may be distinct from that of conventional pancreatic adenocarcinoma(7,8,10–12).
In this study, we hypothesized that different histologic subtypes of IPMN would exhibit different mutation patterns of KRAS, GNAS and RNF43. This study was also designed to explore the timing of these mutations within the progression pathway. Our results confirm that mutations in these genes occur early in IPMN carcinogenesis, that GNAS mutations predominate in colloid carcinoma, and KRAS mutations predominate in tubular carcinoma. Importantly, these findings may allow us to identify those with colloid carcinoma from those with tubular carcinoma prior to operation.
Methods
Patients
With approval of the Institutional Review Board and in accordance with Health Insurance Portability and Accountability Act regulations, a prospectively maintained pancreatic database was used to identify all patients who underwent resection of IPMN-associated carcinoma (IPMN-INV) at Memorial Sloan-Kettering Cancer Center between 1997 and 2012. A retrospective review of each patient's medical record recorded demographic information, preoperative variables, operative therapy, long-term outcome, and recurrence patterns.
Histopathology
Tissue blocks from these patients were retrieved from the Department of Pathology archive and reviewed independently by two expert pancreatic pathologists (OB, DK). Areas of low- and high-grade dysplasia, as well as carcinoma, were identified and marked within each pathologic specimen. Dysplastic areas were also sub-classified by IPMN subtype: gastric, intestinal and pancreatobiliary. Carcinomas were classified as colloid or tubular type, as previously described(3). Briefly, an IPMN-INV was classified as tubular if it exhibited conventional ductal morphology, similar to that of pancreatic ductal adenocarcinoma with neoplastic cells arranged in small tubular glands with an infiltrative desmoplastic stroma. An IPMN-INV was classified as colloid if more than 80% of the invasive component consisted of extensive stromal pools of acellular mucin either lined by neoplastic epithelial cells or containing floating neoplastic epithelial cells. None of the tumors had a mixed histologic subtype. Ten-micron sections were then cut from tissue blocks containing areas of interest. From these sections, areas of dysplasia or carcinoma were needle micro-dissected. DNA extraction was performed on dissected tissue, as well as normal, non-pancreatic tissue (stomach, spleen or duodenum) for each patient (Figure 1).
FIGURE 1.
Study workflow.
High Throughput, Targeted Deep Sequencing
Targeted, massively-parallel sequencing was then performed on a panel of 275 genes (including KRAS, GNAS and RNF43; see Appendix (available online) known to undergo somatic genomic alterations in cancer, as previously described(13,14). Briefly, massively parallel sequencing libraries (Kapa Biosystems, New England Biolabs) that contain barcoded universal primers were generated from 78-250 ng genomic DNA of the micro-dissected, formalin-fixed, paraffin-embedded tumor material and matched normal tissue. After library amplification and DNA quantification, equimolar pools were generated consisting of up to 24 barcoded libraries. These DNA pools were subjected to solution-phase hybrid capture with synthetic biotinylated DNA probes (Nimblegen SeqCap) targeting all protein-coding exons from the selected 275 cancer genes. Genes were selected to include commonly implicated oncogenes, tumor suppressor genes, and members of pathways deemed actionable by targeted therapies. Each hybrid capture pool was sequenced to deep coverage in a single paired-end lane of an Illumina flow cell. Subsequently, the sequencing data were deconvoluted to match all high-quality barcoded reads with the corresponding tumor samples, and genomic alterations (single-nucleotide sequence variants, small insertions / deletions, and DNA copy-number alterations) were identified. Somatic single nucleotide variants and indels were called using MuTect and the SomaticIndelDetector tools in GATK, respectively(15,16). All candidate mutations and indels were reviewed manually using the Integrative Genomics Viewer(17). Because the tumor purity of the invasive samples was typically lower than the high-grade and low-grade samples, sites of mutations called in high-grade and low-grade samples but not in invasive samples were inspected and called positive if present in at least 10 unique reads and 2% of reads.
Statistical Analyses
Statistical analyses were performed with the software package JMP (JMP, Cary, NC, USA). Disease-specific survival (DSS) was calculated from the time of pancreatic resection until cancer-related death. Survival curves were generated using the Kaplan-Meier method, with patients censored when lost to follow-up or upon death from non-IPMN causes. Kaplan-Meier estimates of survival and chi-square tests were used to identify variables associated with DSS, histologic subtypes and presence of specific genetic mutations.
RESULTS
There were 74 patients who underwent resection of IPMN-INV between 1997 and 2012, and 61 of these patients had sufficient tissue for microdissection. Sufficient DNA for targeted sequencing of IPMN-INV was present in 38 of the 61 patients and these 38 patients comprise the final study population (Table 1). The majority of these patients were male and over 70 years of age. Median follow-up was 2.6 years. The vast majority (90%) of patients had main duct involvement. Pancreaticoduodenectomy was performed in 79% of the patients. Colloid carcinoma was present in 50%, and the other 50% had tubular type carcinoma. No patients received neoadjuvant chemotherapy or chemoradiation. All patients received adjuvant chemotherapy, but not chemoradiation. There was no difference in disease-specific survival according to macroscopic type (main-duct, branch-duct, or mixed).
TABLE 1.
Patient Demographics
| Patient Characteristic | n | % |
|---|---|---|
| Sex | ||
| Male | 24 | 63 |
| Female | 14 | 37 |
| Age, y | ||
| ≤70 | 15 | 39 |
| >70 | 23 | 61 |
| Location of tumor | ||
| Head | 30 | 79 |
| Other | 8 | 21 |
| Size of invasive component | ||
| Microscopic, <0.5cm | 6 | 16 |
| Macroscopic, ≥0.5cm | 32 | 84 |
| Macroscopic type | ||
| Main-duct | 20 | 53 |
| Mixed | 14 | 37 |
| Branch-duct | 4 | 10 |
| Type of resection | ||
| Standard Whipple | 28 | 74 |
| Distal pancreatectomy | 5 | 13 |
| Total pancreatectomy | 5 | 13 |
| Resection margin | ||
| Positive | 7 | 18 |
| Negative | 31 | 82 |
| Lymph node status | ||
| Positive | 14 | 37 |
| Negative | 24 | 63 |
| Histologic subtype | ||
| Colloid | 19 | 50 |
| Tubular | 19 | 50 |
| GNAS status | ||
| Wild-type | 15 | 39 |
| Mutant | 23 | 61 |
| KRAS status | ||
| Wild-type | 11 | 29 |
| Mutant | 27 | 71 |
| RNF43 status | ||
| Wild-type | 31 | 82 |
| Mutant | 7 | 18 |
Within this population of patients with IPMN-INV, adjacent areas of dysplasia were micro-dissected (Table 2). Low-grade dysplasia was identified in 19 patients: gastric subtype in 10 (53%), intestinal subtype in 9 (47%). There were no areas of low-grade pancreatobiliary subtype IPMN. High-grade dysplasia was identified in 34 patients: intestinal subtype in 18 (53%), pancreatobiliary subtype in 16 (47%). There were no areas of high-grade gastric subtype IPMN.
TABLE 2.
IPMN Subtypes and Histologies
| Grade and subtype | n (%) |
|---|---|
| Low (n=19) | |
| Gastric | 10 (53) |
| Intestinal | 9 (47) |
| Pancreato-biliary | 0 |
| High (n=34) | |
| Gastric | 0 |
| Intestinal | 18 (53) |
| Pancreato-biliary | 16 (47) |
Spectrum of genetic mutations
A total of 91 tumor samples from the 38 patients underwent mutational profiling alongside matched normal tissue from each patient to filter out inherited variants. Low-grade dysplasia was present in 19 samples, high-grade dysplasia in 34 samples, and carcinoma in 38 samples (Figure 1).
Figure 2 shows all the genes mutated in ≥5% of specimens, stratified by level of dysplasia. In the 18 low grade specimens, there were a total of 49 mutations. The median number of mutations per specimen was 3, range 0-5.
FIGURE 2.
Genetic mutations and degree of dysplasia. Genes were included if mutation occurred in ≥5% of the population.
In the 34 high grade specimens, there were a total of 175 mutations. The median number of mutations per specimen was 4.5, range 0-40. One patient (patient 46) had 40 mutations in his high grade specimen – this patient had a 1.3cm, node-positive tubular carcinoma on a background of pancreatobiliary high-grade dysplasia that recurred 2.5 years after distal pancreatectomy. If this patient is excluded, then the remainder of the high grade specimens had 0-8 genetic mutations.
In the 38 invasive specimens, there were a total of 162 mutations. The median number of mutations per specimen was 4, range 0-40. Just as for the high grade specimen, patient 46's invasive specimen also had 40 mutations; when this patient is excluded, the remainder of the invasive specimens had 0-8 genetic mutations. The number of mutations was not proportional to the size of the invasive IPMN component.
KRAS was the most commonly mutated gene, being mutated in 14 of 18 (78%), 21 of 34 (62%) and 23 of 38 (61%) of low grade, high grade and invasive specimens, respectively. GNAS was the next most commonly mutated gene, being mutated in 10 (56%), 20 (59%) and 19 (50%) of low grade, high grade and invasive specimens, respectively. Two (11%) low grade specimens had an RNF43 mutation, compared with 9 (26%) high grade specimens and seven (18%) invasive specimens. TP53 was mutated in 5% of low grade specimens, and in 21% of both high grade and invasive specimens (NS). Frequencies of mutation in KRAS, GNAS and RNF43 did not vary according to tumor size, nodal status or macroscopic IPMN type.
Distribution of mutations in KRAS, GNAS and RNF43
KRAS mutations were predominantly at codon 12 (G12D/R/V; 47 specimens (81%)). There were 2 instances of G13D, 6 instances of Q61R and 2 of A146G mutations.
Of the 49 detected GNAS mutations, 45 (92%) were at codon 201 (R201H or R201C), the remaining four mutations were Q227L mutations. Three of these Q227L mutations were identified in the low-grade (intestinal-type), high-grade (intestinal-type) and colloid carcinoma specimens of a single patient.
Unlike GNAS and KRAS, there was no hotspot mutation for RNF43. RNF43 mutations were identified in 19 samples: eight were frame-shift deletions, two were frame-shift insertions, six were missense mutations and three were nonsense mutations (Supplemental Table 1, online only).
There was no statistical difference in the frequency of mutations of these KRAS, GNAS andRNF43 by level of dysplasia. For non-invasive IPMN (low-grade and high-grade combined, n=53), KRAS, GNAS and RNF43 mutations were detected in 66%, 56% and 21%, respectively. For IPMN-INV (n=38), KRAS, GNAS and RNF43 mutations were detected in 61%, 50% and 18%, respectively. 13 of 23 (56%) of GNAS-mutated IPMN-INV also harbored a KRAS mutation. Three invasive tumors contained mutations for all three genes (but none had recurred at an average follow-up of 3 years).
Analyzing the distribution of mutations by histologic type of invasive and non-invasive IPMN revealed that GNAS mutations were more prevalent in colloid-type IPMN-INV than tubular-type IPMN-INV, 89% vs 32% respectively (p=0.0003). Conversely, KRAS mutations were more prevalent in tubular-type than colloid-type IPMN-INV, 89% vs 52%, respectively (p=0.01, see Figure 3). Of the 17 GNAS-mutant colloid carcinomas, eight (47%) were also KRAS mutant. Of the 17 KRAS-mutant tubular carcinomas, five (29%) also harbored a GNAS mutation. For noninvasive IPMN subtypes (Table 3), GNAS mutations were more prevalent in intestinal (74%) compared with pancreatobiliary (31%) and gastric (50%) subtypes (p=0.02). No differences between the subtypes were observed with respect to KRAS or RNF43 mutations.
FIGURE 3.
GNAS and KRAS mutations in IPMN-associated carcinoma. Blue bar, colloid (n=19); red bar, tubular (n=19). *p=0.0003; **p=0.01.
TABLE 3.
Distribution of Common Mutations by IPMN Subtype
| Gastric (n=10) | Intestinal (n=27) | Pancreato-biliary (n=16) | p Value | |
|---|---|---|---|---|
| KRAS mutant | 8 (80%) | 17 (63%) | 10 (63%) | 0.6 |
| GNAS mutant | 5 (50%) | 20 (74%) | 5 (31%) | 0.02 |
| RNF43 mutant | 1 (10%) | 5 (19%) | 5 (31%) | 0.4 |
On univariate analysis, factors associated with disease-specific survival were histologic subtype, GNAS status, nodal status and positive resection margin (Table 4). Small sample size precluded multivariate analysis. Figure 4 illustrates outcome of patients stratified by histology (colloid or tubular) and GNAS status. When considering both KRAS and GNAS mutational status, patients with GNAS mutant, KRAS wild type tumors showed a non-significant trend towards improved survival compared with patients with KRAS mutant tumors (Figure 4C).
TABLE 4.
Factors Associated with Disease-Specific Survival
| Factor | 3-year DSS, % | Univariate p Value |
|---|---|---|
| Resection margin | 33 | 0.05 |
| Positive (n=7) | 69 | |
| Negative (n=31) | ||
| Lymph node status | 39 | 0.04 |
| Positive (n=14) | 72 | |
| Negative (n=24) | ||
| Histologic subtype | 81 | 0.007 |
| Colloid (n=19) | 42 | |
| Tubular (n=19) | ||
| GNAS status | 45 | 0.03 |
| Wild-type (n=15) | 71 | |
| Mutant (n=23) | ||
| KRAS status | 31 | 0.2 |
| Wild-type (n=11) | 68 | |
| Mutant (n=27) | ||
| RNF43 status | 59 | 0.5 |
| Wild-type (n=31) | 33 | |
| Mutant (n=7) |
FIGURE 4.
Kaplan-Meier estimates of DSS in IPMN-associated carcinoma. The effect of (A) histologic subtype (colloid versus tubular, p=0.007); (B) GNAS mutational status (p=0.03); and (C) considering mutational status of both KRAS and GNAS (p=0.3).
Polyclonality
In 36 of 38 patients, at least one gene was found to be mutated in two specimens of varying dysplasia from that same patient. For ten patients (28%), the site and/or type of these mutations were different between specimens from the same patient (Table 5). For example, in patient 11, the low grade specimen had an R844H GNAS mutation and a G12V KRAS mutation, but the paired high grade specimen had a Q870L GNAS mutation and a G13D KRAS mutation. For 7 of these ten patients, one of the differentially mutated genes was KRAS.
TABLE 5.
Distinct Mutations in Genes across Varying Levels of Dysplasia but from the Same Macroscopic IPMN Lesion
| Pt No. | Level of dysplasia | Histologic subtype | Gene | Type of mutation | Protein change |
|---|---|---|---|---|---|
| 8 | High grade | Intestinal | RNF43 | Frame Shift Insertion | p.R117fs |
| High grade | Intestinal | RNF43 | Frame Shift Deletion | p.K560fs | |
| Invasive | Tubular | RNF43 | Frame Shift Deletion | p.L7fs | |
| 10 | High grade | Pancreatobiliary | KRAS | Missense Mutation | p.G12D |
| Invasive | Tubular | KRAS | Missense Mutation | p.G12V | |
| 11 | High grade | Intestinal | GNAS | Missense Mutation | p.Q870L |
| Low grade | Gastric | GNAS | Missense Mutation | p.R844H | |
| High grade | Intestinal | KRAS | Missense Mutation | p.G13D | |
| Low grade | Gastric | KRAS | Missense Mutation | p.G12V | |
| 13 | Invasive | Tubular | KRAS | Missense Mutation | p.G12D |
| Low grade | Gastric | KRAS | Missense Mutation | p.G12R | |
| 17 | High grade | Intestinal | KRAS | Missense Mutation | p.G12D |
| Low grade | Intestinal | KRAS | Missense Mutation | p.G12V | |
| 20 | High grade | Pancreatobiliary | KRAS | Missense Mutation | p.G12R |
| Invasive | Tubular | KRAS | Missense Mutation | p.G12D | |
| 44 | Invasive | Tubular | KRAS | Missense Mutation | p.G12D |
| Low grade | Gastric | KRAS | Missense Mutation | p.G12V | |
| 46 | High grade | Pancreatobiliary | MLL2 | Frame Shift Deletion | p.P2354fs |
| Invasive | Tubular | MLL2 | Frame Shift Insertion | p.P647fs | |
| High grade | Pancreatobiliary | PTEN | Frame Shift Deletion | p.L265fs | |
| Invasive | Tubular | PTEN | Frame Shift Deletion | p.T321fs | |
There were 18 patients where a low grade specimen could be paired with a higher grade specimen (either high grade dysplasia or invasive). In 12 of these 18 patients (67%), there was at least one mutation (median 2, range 1-4) present in the low grade specimen that was NOT present in the higher grade specimen.
Similarly, there were 32 patients where a high grade specimen could be paired with an invasive specimen. In 25 of these 32 patients (78%), there was at least one mutation (median 3, range 1-7) present in the high grade specimen that was NOT present in the invasive specimen.
In summary, the presence of different mutations in the same gene between specimens from the same patient, as well as the absence of mutations in specimens of higher levels of dysplasia that were present in specimens of lesser dysplasia, is consistent with a multifocal neoplastic process within the pancreatic parenchyma.
Discussion
The results of this study provide two important findings regarding IPMN carcinogenesis. First, KRAS and GNAS likely represent markers for two distinct progression pathways. GNAS was predominant in colloid carcinoma, as well as within its intestinal subtype precursor lesion. Similarly, KRAS was predominant in tubular carcinoma, as well as within the pancreatobiliary precursor lesion. This finding may allow us to identify those with colloid carcinoma from those with tubular carcinoma, and make inferences regarding disease specific outcome, prior to operation. Second, these events occur very early within the progression pathway, and likely will not be useful for differentiating between low-grade and high-grade dysplasia. KRAS and GNAS mutations were identified equally in the early gastric subtype precursor lesions, which suggests that although these lesions appear similar pathologically, their future progression pathway may have already been determined.
Our work extends the studies of Wu et al(7) investigating the polyclonality of IPMN. In their study of IPMN cyst fluid, they quantified the levels of mutations in cyst samples containing more than one mutation of the same gene (KRAS or GNAS). There was a predominant mutation in four of the 11 IPMNs that had more than one mutation in either KRAS or GNAS. Since IPMN cysts, by definition, communicate with the remainder of the pancreatic ductal network, this finding may be explained by the presence of multiple, macroscopically-separated foci of IPMN arising independently, or by the presence of distinct clones within a single macroscopic focus of IPMN.
In our study, for each patient, microdissected lesions of all levels of dysplasia were all obtained from a single macroscopic focus of IPMN. We found that in 28% of gross tumor lesions, the site of mutation of a specific gene (most commonly KRAS) was different across varying levels of dysplasia. Combined with our finding that the prevalence of GNAS and KRAS mutations did not increase with increasing dysplasia, it would appear that these disparate mutations occur very early in IPMN tumorigenesis. Thus, these data suggest that the microscopic regions of differing dysplasia within a single macroscopic region may either diverge from a single clone at a very primitive stage, or arise completely independently. Thus, a gross tumor lesion may represent a “focal point” of genomic instability within an at-risk pancreas. Further work remains to be done to elucidate the underlying molecular mechanisms triggering this field instability.
The focus of the published literature studying mutational events in IPMN has been on noninvasive IPMN(7,8,10,11,18–20). The current study differs from these previous reports by focusing on a cohort of resected patients with invasive IPMN and by evaluating both the invasive and non-invasive components of these complex lesions.
Our finding of a higher frequency of GNAS mutations in colloid IPMN-INV, compared to tubular IPMN-INV, is novel. Dal Molin et al(19) found GNAS mutations more frequently in intestinal type IPMN, compared with gastric or pancreatobiliary subtypes. Thus, our observation is consistent with our understanding of IPMN carcinogenesis, as colloid carcinoma typically arises from intestinal IPMN, whereas tubular carcinoma typically arises from pancreatobiliary IPMN(21).
The higher frequency of GNAS mutations in colloid IPMN-INV and KRAS mutations in tubular IPMN-INV may have important clinical implications. For the first time we may have a marker that can define colloid from tubular carcinoma pre-operatively. In this study, within the subgroup of patients with GNAS mutation only (colloid carcinoma), not a single patient died of disease within four years of follow-up. Within the KRAS mutation-only group (tubular carcinoma) only 30% of patients were alive after four years of follow-up. Although the numbers are very small, if these findings can be validated, this information may be very helpful in clinical decision making. Many patients who present with IPMN are at an advanced age, and in the setting of multiple comorbidities, one may choose to not proceed with resection in the setting of a GNAS-only mutant. Conversely, in a young patient, one may choose more extensive resection in the setting of GNAS only as the long-term survival outcomes are likely excellent. From a practical standpoint, mutations in GNAS (as well as other genes) can be identified in DNA extracted from small volumes (0.25ml) of pancreatic cyst fluid(7).
Yamada et al(22) recently identified GNAS-activating mutations in 83% of colonic villous adenomas, but only rarely (3%) from tubulovillous adenomas or adenocarcinomas. The same group also reported GNAS-activating mutations in 63% of pyloric gland adenomas of the stomach and duodenum, but none in foveolar- or intestinal-type adenomas or gastric adenocarcinomas(23). These two studies suggest that intestinal-type IPMN and colloid carcinoma share not only morphologic features with tubulovillous adenomas and pyloric gland adenomas, but a distinct genetic background.
In tumors with GNAS-activating mutations, codon 201 mutations predominate(10,12,22–28). Although all series of GNAS-mutated tumors are small (<25 patients), in all but one study, codon 201 mutations comprise >90%. Therefore, while codon 227 mutations have not previously been described in IPMN, our finding of Q227L mutations in 8% of GNAS mutations is consistent with mutation patterns in other tumors.
KRAS is mutated in >90% of conventional pancreatic adenocarcinoma, with the vast majority of mutations occurring in codon 12. Mutations in GNAS and RNF43 have not been found. Thus, the presence of KRAS mutations in 89% of IPMN-associated tubular carcinoma, all of which occurred at codon 12, along with low frequencies of GNAS (32%) and RNF43 (11%) mutations, is consistent with previous work demonstrating that this type of IPMN-INV has a similar biologic behavior to conventional PDAC(3).
There are several limitations to this study. Only patients who had undergone resection for IPMNINV were included due to the need for sufficient tissue for precise histologic sub-classification, microdissection and DNA extraction. Therefore, our findings may not reflect the frequency and distribution of gene mutations in all patients with IPMN. The number of patients in this study is relatively small (38), but due to separate microdissection of distinct areas of dysplasia from each patient's tumor, there were a total of 91 sequenced specimens. Of these, 38 were of invasive tumors, which represents the largest number of IPMN-associated carcinomas sequenced thus far. This small sample size, especially of low grade lesions, limits further statistical analysis and identification of patterns of genetic mutations between histologic subtypes and across varying levels of dysplasia. The inherent limitation of this type of study on IPMN is the possibility of contamination during microdissection, and that in higher grade lesions there is an abundance of stroma which may make identification of low-level mutations difficult.
In conclusion, this study of 38 patients with invasive IPMN demonstrated that mutations in KRAS and GNAS occur early in IPMN carcinogenesis. GNAS mutations predominated in colloid IPMN-INV, and KRAS mutations predominated in tubular IPMN-INV. GNAS status was associated with outcome, and this finding suggests that KRAS and GNAS mutations may represent prognostic biomarkers that could provide valuable information for clinical decision making. These preliminary findings will require validation with larger patient- and tissue-sample sets.
Supplementary Material
Précis.
In this mutational analysis of patients who underwent resection of carcinoma arising from intraductal papillary mucinous neoplasms (IPMN), we found high frequencies of GNAS mutations in colloid carcinoma and its intestinal-type pre-invasive precursor. Mutations in GNAS and KRAS appear to occur early in IPMN carcinogenesis.
Acknowledgments
Support: National Cancer Institute of the National Institutes of Health under award number 5T32CA160001 (ARB).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure Information: Nothing to disclose.
REFERENCES
- 1.Furukawa T, Klöppel G, Adsay N. Classification of types of intraductal papillary-mucinous neoplasm of the pancreas: a consensus study. Virchows Arch. 2005;447:794–9. doi: 10.1007/s00428-005-0039-7. [DOI] [PubMed] [Google Scholar]
- 2.Adsay NV, Merati K, Basturk O, et al. Pathologically and biologically distinct types of epithelium in intraductal papillary mucinous neoplasms: delineation of an “intestinal” pathway of carcinogenesis in the pancreas. Am J Surg Pathol. 2004;28:839–48. doi: 10.1097/00000478-200407000-00001. [DOI] [PubMed] [Google Scholar]
- 3.Yopp AC, Katabi N, Janakos M, et al. Invasive carcinoma arising in intraductal papillary mucinous neoplasms of the pancreas: a matched control study with conventional pancreatic ductal adenocarcinoma. Ann Surg. 2011;253:968–74. doi: 10.1097/SLA.0b013e318214bcb4. [DOI] [PubMed] [Google Scholar]
- 4.Tanaka M, Fernández-del Castillo C, Adsay V, et al. International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas. Pancreatology. 12:183–97. doi: 10.1016/j.pan.2012.04.004. [DOI] [PubMed] [Google Scholar]
- 5.Sohn TA, Yeo CJ, Cameron JL, et al. Intraductal papillary mucinous neoplasms of the pancreas: an updated experience. Ann Surg. 2004;239:788–97. doi: 10.1097/01.sla.0000128306.90650.aa. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.D'Angelica M, Brennan MF, Suriawinata AA, et al. Intraductal Papillary Mucinous Neoplasms of the Pancreas. Ann Surg. 2004;239:400–8. doi: 10.1097/01.sla.0000114132.47816.dd. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wu J, Matthaei H, Maitra A, et al. Recurrent GNAS mutations define an unexpected pathway for pancreatic cyst development. Sci Transl Med. 2011;3:92ra66. doi: 10.1126/scitranslmed.3002543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wu J, Jiao Y, Dal Molin M, Maitra A, et al. Whole-exome sequencing of neoplastic cysts of the pancreas reveals recurrent mutations in components of ubiquitin-dependent pathways. Proc Natl Acad Sci U S A. 2011;108:21188–93. doi: 10.1073/pnas.1118046108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chadwick B, Willmore-Payne C, Tripp S, et al. Histologic, immunohistochemical, and molecular classification of 52 IPMNs of the pancreas. Appl Immunohistochem Mol Morphol. 2009;17:31–9. doi: 10.1097/PAI.0b013e31817c02c6. [DOI] [PubMed] [Google Scholar]
- 10.Amato E, Molin MD, Mafficini A, et al. Targeted next-generation sequencing of cancer genes dissects the molecular profiles of intraductal papillary neoplasms of the pancreas. J Pathol. 2014 doi: 10.1002/path.4344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Furukawa T, Kuboki Y, Tanji E, et al. Whole-exome sequencing uncovers frequent GNAS mutations in intraductal papillary mucinous neoplasms of the pancreas. Sci Rep. 2011;1:161. doi: 10.1038/srep00161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sasaki M, Matsubara T, Nitta T, et al. GNAS and KRAS mutations are common in intraductal papillary neoplasms of the bile duct. PLoS One. 2013;8:e81706. doi: 10.1371/journal.pone.0081706. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wagle N, Berger MF, Davis MJ, et al. High-throughput detection of actionable genomic alterations in clinical tumor samples by targeted, massively parallel sequencing. Cancer Discov. 2012;2:82–93. doi: 10.1158/2159-8290.CD-11-0184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Won HH, Scott SN, Brannon AR, et al. Detecting somatic genetic alterations in tumor specimens by exon capture and massively parallel sequencing. J Vis Exp. 2013:e50710. doi: 10.3791/50710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cibulskis K, Lawrence MS, Carter SL, et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnol. 2013;31:213–9. doi: 10.1038/nbt.2514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.DePristo MA, Banks E, Poplin R, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43:491–8. doi: 10.1038/ng.806. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Robinson JT, Thorvaldsdóttir H, Winckler W, et al. Integrative genomics viewer. Nat Biotechnol. 2011;29:24–6. doi: 10.1038/nbt.1754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Komatsu H, Tanji E, Sakata N, et al. A GNAS Mutation Found in Pancreatic Intraductal Papillary Mucinous Neoplasms Induces Drastic Alterations of Gene Expression Profiles with Upregulation of Mucin Genes. PLoS One. 2014;9:e87875. doi: 10.1371/journal.pone.0087875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Dal Molin M, Matthaei H, Wu J, et al. Clinicopathological correlates of activating GNAS mutations in intraductal papillary mucinous neoplasm (IPMN) of the pancreas. Ann Surg Oncol. 2013;20:3802–8. doi: 10.1245/s10434-013-3096-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Matthaei H, Wu J, Dal Molin M, et al. GNAS sequencing identifies IPMN-specific mutations in a subgroup of diminutive pancreatic cysts referred to as “incipient IPMNs”. Am J Surg Pathol. 2014;38:360–3. doi: 10.1097/PAS.0000000000000117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hruban RH, Maitra A, Kern SE, Goggins M. Precursors to pancreatic cancer. Gastroenterol Clin North Am. 2007;36:831–49. vi. doi: 10.1016/j.gtc.2007.08.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Yamada M, Sekine S, Ogawa R, et al. Frequent activating GNAS mutations in villous adenoma of the colorectum. J Pathol. 2012;228:113–8. doi: 10.1002/path.4012. [DOI] [PubMed] [Google Scholar]
- 23.Matsubara A, Sekine S, Kushima R, et al. Frequent GNAS and KRAS mutations in pyloric gland adenoma of the stomach and duodenum. J Pathol. 2013;229:579–87. doi: 10.1002/path.4153. [DOI] [PubMed] [Google Scholar]
- 24.Carter JM, Inwards CY, Jin L, et al. Activating GNAS Mutations in Parosteal Osteosarcoma. Am J Surg Pathol. 2014;38:402–9. doi: 10.1097/PAS.0000000000000144. [DOI] [PubMed] [Google Scholar]
- 25.Collins MT, Sarlis NJ, Merino MJ, et al. Thyroid carcinoma in the McCune-Albright syndrome: contributory role of activating Gs alpha mutations. J Clin Endocrinol Metab. 2003;88:4413–7. doi: 10.1210/jc.2002-021642. [DOI] [PubMed] [Google Scholar]
- 26.Freda PU, Chung WK, Matsuoka N, et al. Analysis of GNAS mutations in 60 growth hormone secreting pituitary tumors: correlation with clinical and pathological characteristics and surgical outcome based on highly sensitive GH and IGF-I criteria for remission. Pituitary. 2007;10:275–82. doi: 10.1007/s11102-007-0058-2. [DOI] [PubMed] [Google Scholar]
- 27.Kalfa N, Lumbroso S, Boulle N, et al. Activating mutations of Gsalpha in kidney cancer. J Urol. 2006;176:891–5. doi: 10.1016/j.juro.2006.04.023. [DOI] [PubMed] [Google Scholar]
- 28.Nault JC, Fabre M, Couchy G, et al. GNAS-activating mutations define a rare subgroup of inflammatory liver tumors characterized by STAT3 activation. J Hepatol. 2012;56:184–91. doi: 10.1016/j.jhep.2011.07.018. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.