PIK3CA Mutations in Intraductal Papillary Mucinous Neoplasm/Carcinoma (IPMN/IPMC) of the Pancreas

Abstract

Background

Recent studies have reported high frequencies of somatic mutations in the phosphoinositide-3-kinase, catalytic, alpha (PIK3CA) gene in various human solid tumors. More than 75% of those somatic mutations are clustered in the helical (exon 9) and kinase domains (exon 20). The three hot-spot mutations, E542K, E545K, and H1047R, have been proven to elevate the lipid kinase activity of PIK3CA and activate the Akt signaling pathway. The mutational status of PIK3CA in intraductal papillary mucinous neoplasm/carcinoma (IPMN/IPMNC) has not been evaluated previously.

Methods

To evaluate a possible role for PIK3CA in the tumorigenesis of IPMN and IPMNC, exons 1, 4, 5, 6, 7, 9, 12, 18, and 20 were analyzed in 36 IPMN/IPMC and two mucinous cystadenoma specimens by direct genomic DNA sequencing.

Results

We identified four missense mutations in the nine screened exons of PIK3CA from 36 IPMN/IPMC specimens (11%). One of the four mutations, H1047R, has been previously reported as a hot-spot mutation. The remaining three mutations, T324I, W551G, and S1015F, were novel and somatic.

Conclusion

This is the first report of PIK3CA mutation in pancreatic cancer. Our data provide evidence that oncogenic properties of PIK3CA contribute to the tumorigenesis of IPMN/IPMC.

Introduction

Intraductal papillary mucinous neoplasm of the pancreas (IPMN) is an increasingly recognized noninvasive cystic neoplasm of the pancreas. It is characterized by unique clinical, pathologic, and molecular features (1–7). These neoplasms are subdivided into three groups based on increasing nuclear and architectural atypia: adenoma, borderline, and intraductal papillary mucinous carcinoma (IPMC) (5). According to the absence or presence of neoplastic cells invading the pancreatic tissue surrounding the involved ducts, IPMC are separated into invasive and noninvasive types. The overall incidence of invasive carcinoma associated with an IPMN is 20% to 40% (8).

Recently, there has been an increase in the number of IPMN cases reported, although it is not clear if this represents a true increase in incidence or a manifestation of increased recognition and detection of these lesions (9). Most IPMN are slow growing and less aggressive compared with conventional ductal adenocarcinoma. An infiltrating adenocarcinoma, however, is frequently identified in pancreata affected by IPMN, suggesting that IPMN can also evolve into invasive ductal adenocarcinomas (1, 2, 7, 10). Although the overall outcome for IPMN is good, even with associated invasive carcinoma, a significant proportion of the patients with completely resected noninvasive IPMN may develop pancreatic adenocarcinoma in the pancreatic remnant and die of disseminated disease. A 4-year survival of only 64% for noninvasive IPMN was reported in one large series- a survival rate almost equal to that of IPMN with invasive carcinoma (11). Other studies have also reported recurrences of invasive carcinoma in completely resected noninvasive IPMN (1, 10), some of which demonstrated only moderate dysplasia (borderline IPMN). Although the majority of invasive carcinomas are associated with IPMC (intraductal carcinoma), invasive carcinoma coexisting with adenoma and borderline IPMN can occur (12). In addition invasive carcinoma is sometimes found distant from an IPMN, and small IPMNs have been detected incidentally in pancreata resected for conventional ductal pancreatic cancer (9).

Reported genetic alterations in IPMNs include mutations in the KRAS (13), TP53 (14), and STK11/LKB1 genes (15, 16), as well as loss of heterozygosity (LOH) of several chromosomal loci (15, 17). Recent evidence suggests that in addition to these genetic alterations, aberrant DNA methylation may contribute to the inactivation of a subset of tumor-suppressor genes in IPMNs (18, 19). Furthermore, two recent studies have evaluated gene expressing profiling in IPMNs mainly focusing on genes that are preferentially expressed in IPMNs (20, 21). To-date, no study has evaluated the mutational status of the PIK3CA gene in IPMNs/IPMCs.

Phosphatidylinositol-3 kinases (PI3Ks) constitute a large and complex family of lipid kinases encompassing three classes with multiple subunits and isoforms (22–24). PI3Ks play an important role in several cellular functions, such as proliferation, differentiation, chemotaxis, survival, trafficking, and glucose homeostasis (22). Class IA PI3Ks are heterodimeric proteins composed of a p110 catalytic subunit and a p85 regulatory subunit (25), which can be activated through interaction with phosphotyrosine residues of receptor tyrosine kinases (RTK) (26, 27) or through the binding of active RAS to the p110 catalytic subunit (24, 27–29). P85 lacks kinase activity and acts as an adaptor, coupling with the p110 subunit to activate protein tyrosine kinases (30). Activated PI3Ks phosphorylate the inositol ring 3’-OH group in inositol phospholipids to generate the second messenger phosphatidylinositol-3,4,5-triphosphate (PIP₃) (31), which in turn activates diverse cellular target proteins such as the survival signaling kinase AKT/PKB (22, 23, 32). A tumorigenic role has been proposed for the PIK3CA gene that encodes the catalytic p110alpha subunit of phosphatidylinositol 3-kinase belonging to the class IA of PI3Ks (22, 24). One recent study reported mutations in PIK3CA in different tumor types, namely colorectal cancer, gastric cancer, glioblastoma, breast and lung cancer (33). Several other independent studies in hepatocellular carcinomas, breast carcinomas, lung cancers, ovarian carcinomas, brain tumors, acute leukemias, and head and neck squamous cell carcinomas have since supported and emphasized the oncogenic potential of PIK3CA in the development of cancer (34–38).

In the study by Samuels et al. (33), two PIK3CA mutational hot-spots were described and found to affect the helical (exon 9) and catalytic (exon 20) protein domains. In addition, exons 9 and 20 of PIK3CA were preferentially mutated in colon carcinomas (33). Mutations were also described in exons 1, 2, 4, 7, 12, 14 and 18 of PIK3CA, but only in a minority of cases (33, 34). Similar to colon tumors, PIK3CA mutations also clustered in the two hotspot regions (exons 9 and 20) in gastric carcinomas (33, 35, 39). No PIK3CA mutations have been previously reported in IPMN, IPMC, or conventional pancreatic ductal adenocarcinoma (33).

Materials and Methods

Patients and Tissue Samples

Surgical paraffin-embedded IPMN/IPMC and mucinous cystadenoma samples from 38 patients (female n=14, male n= 24, median age 68.1 years, range 41–84 years) were obtained from the archival tissue collection of the Columbia University Medical Center. Acquisition of the tissue specimens was approved by the Institutional Review Board of Columbia University Medical Center and performed in accordance with Health Insurance Portability and Accountability Act (HIPAA) regulations. In detail, we analyzed three IPMN, adenoma (female n= 1, male n = 2, median age 62.7 years, range 53–77 years); four IPMN, borderline (female n= 1, male n= 3, median age 66.3 years, range 62–72 years), five IPMC without invasion (male n= 5, median age 69.2 years, range 59–81), 24 IPMC with invasive carcinoma (male n= 14, female= 10, median age 68.9 years, range 41–84 years), and two mucinous cystadenomas (female n=2, median age 57.5 years, range 53–62 years). Thirty-two of these lesions arose in the pancreatic head, one in the uncinate process, four within the transition from pancreatic head to the body, one within the body, and one diffusely involving the entire gland. The maximum diameter of the lesions ranged from 0.4 to 7cm (mean: 4.2 cm). For a more detailed register, see Table 1.

Table 1.

Summary reports of the 38 patient samples

Case No.	Sex	Age	Lesion analyzed	IPMN Nuclear Grade	Differentiation of invasive carcinoma, if present in resection	Location within pancreas	Maximum Dimension** (cm)
1	M	62	IPMN, borderline	2	N/A	body	-
2	M	73	IPMC with invasion	3	moderate	head and body	4
3	M	67	IPMC with invasion	3	moderate	head	-
4	M	69	IPMC	3	N/A	head	3.5
5	F	75	IPMC with invasion	3	moderate	head	5.5
6	F	68	IPMC with invasion	3	moderate	head	5
7	M	65	IPMN, borderline	2	poor (*)	head	5
8	F	66	IPMN, borderline	2	N/A	uncinate process	2
9	M	84	IPMC with invasion	3	moderate	head	5
10	M	53	IPMN, adenoma	1	N/A	head	3.5
11	M	71	IPMC with invasion	2–3	N/A	head	-
12	M	81	IPMC	3	N/A	head	2.5
13	M	63	IPMC	3	moderate to poor (*)	head	2.3
14	M	66	IPMC with invasion	3	moderate to poor	head	6
15	F	70	IPMC with invasion	3	moderate to poor	head and body	7
16	F	70	IPMC with invasion	3	moderate	head	1.5
17	M	72	IPMN, borderline	2	N/A	head	0.4
18	F	53	mucinous cystadenoma	1	N/A	head	3
19	M	79	IPMC with invasion	3	moderate to poor	head	6
20	M	63	IPMC with invasion	3	moderate to poor	head	3.5
21	M	77	IPMN, adenoma	1	N/A	head and body	2.2
22	F	62	mucinous cystadenoma	1	N/A	head	2
23	M	41	IPMC with invasion	3	moderate to poor	head	5
24	M	71	IPMC with invasion	3	moderate to poor	head	1.5
25	F	58	IPMN, adenoma	1	N/A	head	1.5
26	M	49	IPMC with invasion	3	moderate	head	4.5
27	M	71	IPMC with invasion	3	moderate to poor	head	5.5
28	M	74	IPMC	3	well (*)	head and body	-
29	M	59	IPMC	3	poor (*)	head	7
30	M	81	IPMC with invasion	3	moderate to poor	head	3
31	F	80	IPMC with invasion	3	moderate to poor	head	5
32	F	66	IPMC with invasion	3	poor	head	3
33	F	77	IPMC with invasion	3	poor	head	3
34	M	73	IPMC with invasion	3	poor	head	5.5
35	F	77	IPMC with invasion	3	well	head	3.2
36	F	61	IPMC with invasion	3	well	head	1
37	M	62	IPMC with invasion	3	moderate	head	2.2
38	F	59	IPMC with invasion	3	moderate	head	3.4

N/A, not applicable.

^*Invasive carcinoma was associated with IPMN/IPMC in pancreatic resection, but lesion analyzed did not sample invasive carcinoma.

^**Maximum tumor size includes both invasive and non-invasive components of tumor; -, not available.

DNA Samples for Mutation Analysis

All pre-polymerase chain reaction (PCR) tissue samples were handled in an environment free of PCR products. All samples were coded, and the investigator was unaware of all patients’ clinical data. Paraffin-embedded tumor samples were micro-dissected by hand to ensure the highest possible amount of tumor cells. Surrounding non-tumorous tissue or tissue derived from a tumor-free block of each case served as the corresponding normal control. Genomic DNA was extracted using QIAmp DNA Mini Kit (Qiagen, CA). The procedures were performed according to the manufacturer’s instructions for paraffin-embedded tissues.

Exons 1, 4, 5, 6, 7, 9, 12, 18, and 20 of PIK3CA were analyzed by PCR amplification of genomic DNA, and the purified PCR products were directly sequenced. Genomic DNA (40ng per sample) was amplified with primers that had been designed to amplify each exon and its exon/intron boundaries (33, 38). All PCR products were purified using QIAquick PCR Purification Kit according to the manufacturer’s instructions prior to sequencing. Sequencing was performed with ABI’s 3100 capillary automated sequencer at the DNA Core Facility of Columbia University Medical Center. Each sample found to have a genetic alteration in the target gene was subsequently sequenced in the reverse direction to confirm the mutation. The mutation was then further verified by sequencing of a second PCR product derived independently from the original template.

Results and Discussion

In the present study, four of the 36 specimens contained a somatic mutation of the PIK3CA gene (Figure 1 and Table 2)- one in exon 4 (T324I), one in exon 9 (W551G), and two in exon 20 (S1015F, H1047R). None of these mutations were detected in the corresponding normal tissues. One of the missense mutations in exon 20 of PIK3CA, H1047R, is a previously described hot-spot mutation (33). The other mutations in exons 4 and 9 are novel mutations.

Figure 1. Somatic PIK3CA mutations found in IPMN/IPMC.

One of the four mutations (H1047R) was a hot-spot mutation. The other three mutations were novel. All four mutations were confirmed to be somatic.

Table 2.

Nucleotide alterations within the coding exons of PIK3CA identified in the 36 IPMN/IPMC

Case No.	Lesion analyzed	Exon	Nucleotide	Amino acid	Present in normal Tissue
4	IPMC	20	C3044T	S1015F	No
14	IPMC with invasion	20	A3140G	H1047R	No
1	IPMN, bordernline	9	T1654G	W551G	No
5	IPMC with invasion	4	C0971T	T324I	No
3, 9, 19, 23		6	A1173G	I391M	yes
4, 8, 24, 31, 33, 35		12	T1929C	Y644H	yes

The Nucleotide Alterations are described according to the cDNA sequence with GenBank accession number NM_006218. Case Numbers correspond to those described in Table 1.

Recently, much attention has been given to the significance of the PIK3CA gene mutations identified in several human tumors. Mutational analysis of the PIK3CA gene has revealed that genetic alterations at its locus occur in a wide spectrum of human neoplasms (33–38). PIK3CA mutations preferentially occur in exons 9 and 20, affecting the functionally important helical and kinase domains of the protein (33–35, 37, 39). Functional studies have showed that PI3Ks carrying any one of the three hot-spot mutations is able to induce transformation in cultures of chicken embryo fibroblasts, and that transforming activity of the mutant is correlated with increased lipid kinase activity and activation of the Akt signaling pathway (33, 40). Although two of our mutations in exons 9 and 20 are not hot-spot mutations, the mutations are likely to have affected the kinase activity of the PIK3CA. The mutation within exon 4, nucleotide 971 C → T, which leads to an alteration of codon 324 ACA (T) → ATA (I), has not been described before. Although the significance of the novel mutation T324I, which belongs to the C2 domain, is unclear, a recent study found that the C2 domain of PKC d could be a phosphotyrosine binding domain (41). Since 7% of PIK3CA mutations have been detected within the C2 domain (33), it might be of value to study whether the C2 domain also plays a critical role in PIK3CA activity in the future.

Two more nucleotide alterations were detected in the exons of the PIK3CA gene. One is located at exon 6 nucleotide 1173 A → G, leading to a change at codon 391 ATA (I) → ATG (M). This alteration, observed in four tumor specimens and their matching normal tissues, was subsequently identified as a known single-nucleotide polymorphism (SNP, rs2230461). The other is located at exon 12 nucleotide 1929 T → C, leading to a change at codon 644 TAT (Y) → CAT (H). This alteration was observed in six tumor cases and also present within their corresponding normal tissues. Although a search of the SNP database found no match, this alteration is most likely a polymorphism without carcinogenic pathologic value. This alteration has not been reported by previous publications probably because the majority of the PIK3CA studies focused only on exon 9 and 20, and polymorphisms are often omitted from the reports. No mutations were detected in the two mucinous cystadenoma cases.

The frequency of PIK3CA mutations has been reported to be 32% in colon cancer, 4–25% in gastric cancer, 8–40% in breast cancer, 5–27% in brain tumors, 4% in lung cancer, and 4–7% in ovarian cancer (33, 36, 39, 42). Samuels et al screened 11 pancreatic ductal adenocarcinoma cell lines and found no mutation in the entire coding region of the PIK3CA gene (33). A negative finding was also reported by Gallmeier et al who examined the exons 9 and 20 of PIK3CA for mutation using direct genomic sequencing on the genomic DNA from 91 pancreatic cancer xenografts (43). In the present study we report 11% (4/36) of IPMN/IPMC to have PIK3CA mutations. Two of these mutations (W551G and S1015F) were found in IPMN with nuclear grade 3 (IPMC) and nuclear grade 2 (IPMN, borderline), respectively, without associated invasive carcinoma. The other two (T324I and H1047R) were observed in IPMC with associated invasive carcinoma. The findings in colorectal cancers indicate that PIK3CA mutations generally arise just before or coincident with invasion (33). Our data show that in IPMN, mutations of the PIK3CA gene seem to be a rather late event on the transition of these lesions to malignancy. So far, genetic analyses of IPMN have disclosed abnormalities in many of the same genes altered in conventional ductal adenocarcinoma, including mutations of KRAS (13), TP53/p53 (44) and CDKN2A/p16 genes (45). In addition, as is true for pancreatic ductal carcinomas, a number of genes, including CDKN2A/p16, may be epigenetically inactivated in IPMNs through aberrant DNA methylation (18, 19, 46, 47). The Peutz-Jeghers gene STK11/LKB1 is inactivated more frequently in IPMN (up to one third) than in ductal adenocarcinoma (4%) (16, 48), and some patients with the Peutz-Jeghers Syndrome develop IPMNs (15). In contrast to ductal adenocarcinomas and PanIN-3 (pancreatic intraepithelial neoplasia-3) lesions, abnormalities in the MADH4/SMAD4/DPC4 gene seem to be rare in IPMNs (6). PIK3CA is the first gene to be found mutated in IPMN that had not been reported in ductal adenocarcinoma.

In summary, this is the first report of missense mutations of the PIK3CA gene in IPMN/IPMC (4/36, 11%). All four mutations were found to be somatic. Our data suggest that PIK3CA is important to IPMN/IPMC tumorigenesis. The knowledge of PIK3CA’s involvement in IPMC is important because specific kinase inhibitors could be considered as a future additional therapeutic option for more advanced IPMCs with PIK3CA mutations. Recently kinase inhibitors, such as Gleevec (Imatinib), Herceptin (Trastzumab), and Iressa (Gefitinib) have been successfully developed for therapies in some cancer types (49). Our finding may provide a potential target in IPMC for pathway-specific or kinase-inhibitor-based therapies, in addition to surgery.