Background and aim Gastric carcinoma is the second most frequent cause of cancer-related death worldwide. As PTEN is a potential modifier of tumour response to trastuzumab, a recently approved therapy in metastatic HER2 positive gastric cancer, the existence of PTEN deletions in primary gastric cancer was investigated.
Methods 230 primary gastric cancers were analysed in a tissue microarray format by dual labelling fluorescence in situ hybridisation for PTEN deletion. HER2 analysis was also performed. To study PTEN deletion heterogeneity, all available large tissue sections from primary cancer and corresponding metastases were analysed in seven patients.
Results Eight of 180 interpretable primary gastric cancer spots showed PTEN deletions (4.4%), including seven hemizygous and one homozygous deletion. PTEN deletion was correlated with nodal (8 of 122 cases (6.6%); p=0.041) and distant metastases (4 of 19 (21.1%); p<0.001). Large section validation showed a homogeneous distribution of PTEN deletion. HER2 positivity was seen in one PTEN deleted case.
Conclusion Genomic PTEN deletion is a rare event in gastric adenocarcinoma but correlates with metastatic disease. The homogeneous distribution pattern indicates that this alteration occurs early in tumour development.
- PTEN protein
- gene deletion
- stomach neoplasms
- microarray analysis
- in Situ hybridisation
- cancer research
- gut pathology
Statistics from Altmetric.com
- PTEN protein
- gene deletion
- stomach neoplasms
- microarray analysis
- in Situ hybridisation
- cancer research
- gut pathology
Gastric adenocarcinoma is the second most frequent cause of cancer death worldwide, despite an ongoing decline of incidence and mortality.1 Adequate surgery with and without neoadjuvant and adjuvant radiotherapy and/or chemotherapy are the potentially curative therapeutic options.2 More recently, the spectrum of therapeutic options was complemented by trastuzumab and potentially also other anti-HER2 (human epidermal growth factor receptor 2) therapies.3 The ToGA-trial demonstrated an extended median overall survival in the trastuzumab plus chemotherapy group compared with the chemotherapy alone group in 594 patients with advanced or gastro-oesophageal junction cancer. The adverse events were similar in the two groups and so the authors considered trastuzumab as a new standard option for advanced HER2 positive gastric cancer.4
The tumour suppressor gene PTEN (phosphatase and tensin homologue) is located on chromosome 10q23.3 and acts as a plasma membrane lipid phophatase.5 Its primary target for dephosphorylation is the second messenger phosphatidylinositol-3,4,5-triphosphate, the product of phosphatidylinositol-3-kinase. Thereby PTEN negatively regulates the phosphatidylinositol-3-kinase/ATP-dependant tyrosine kinase (Akt) pathway.6 ATP-dependent tyrosine kinase activation suppresses apoptosis, induced by a number of different stimuli, and regulates cellular functions including cell proliferation, cell growth and metabolism, migration and resistance to hypoxia.7 Alterations of PTEN by inactivating mutations and/or chromosomal deletions have been described in many different tumour types including gastric cancer. Furthermore, reduced PTEN function has been associated with aggressive tumour phenotype and unfavourable disease course in various cancer types.8
Moreover, PTEN aberration has been associated with failure of anti-HER2 therapy in breast cancer.9 PTEN alterations may thus also become clinically relevant in gastric cancer. Most studies evaluating PTEN alterations are not very recent, and there are no studies using fluorescence in situ hybridisation (FISH)—the gold standard for deletion analysis. Given the potential relevance of PTEN alterations as a predictive tool for trastuzumab therapy, we designed this study to evaluate the frequency of chromosomal deletions in primary gastric cancers, and also to determine whether or not such alterations may be represented homogeneously or heterogeneously in affected cancers.
Materials and methods
Tissues from a total of 382 consecutive patients who had undergone either endoscopic treatment at the Interdisciplinary Endoscopy Department and Clinic of the University Medical Center Hamburg–Eppendorf or surgical therapy at the Department of General, Visceral and Thoracic Surgery between 1994 and 2006 were taken for tissue microarray (TMA) construction. Written informed consent for the use of resected samples was obtained from all patients and approval was obtained from the Ethics Committee of the Chamber of Physicians in Hamburg, Germany. Age, sex, tumour localisation, tumour size, lymphatic invasion, pTNM stage,10 grading, and tumour type according to Lauren were evaluated by a review of medical charts and pathological records. Glass slides were reviewed for determining the histological type. Clinical follow-up data were available from 143 patients. The median follow-up was 24.3±26.8 months (range 1–145 months).
TMA construction and large section validation
Tissue samples were fixed in buffered 4% formalin, paraffin embedded, and used for TMA construction as previously described.11 H&E stained histological sections from the gastrectomy specimens were used to mark representative tumour areas. A 0.6-mm tissue core was punched out from the tumour area of each specimen using a homemade semi-automated tissue arrayer and transferred in a TMA recipient block. The TMA contained 770 tissue samples, including 230 primary gastric tumours, 79 matched lymph node metastases, four distant metastases (ovary, skin, liver), 140 dysplasias (10 low-grade, 77 high-grade, 53 intramucosal cancers), and 317 normal gastric tissue samples from patients with gastric cancer as well as from healthy patients. To study a potential heterogeneous distribution pattern of PTEN deletion, all available large tissue sections (4 μm) from primary cancer and corresponding metastases were analysed in seven patients.
Fluorescence in situ hybridisation
SpectrumGreen-labelled PTEN probes were manufactured by using two bacterial artificial chromosome clones (RP11-38065 and RP11-81303). Bacterial artificial chromosome DNA was extracted and labelled with SpectrumGreen-dUTP (Vysis-Abbott, Chicago, Illinois, USA) using a commercial nick-translation kit (Vysis-Abbott). These homemade PTEN probes were used together with a SpectrumOrange-labelled centromere 10 (CEP10) reference probe (Vysis-Abbott). For proteolytic slide pretreatment, a commercial kit was used (paraffin pretreatment reagent kit, Vysis-Abbott). Before hybridisation, sections were deparaffinised, air dried and dehydrated in 70%, 85% and 100% ethanol, and then denatured for 15 min at 80°C in 70% formamide-2× SSC solution. Following overnight hybridisation at 37°C in a humidified chamber, slides were washed, counterstained with 0.2 μmol/l 4′,6-diamidino-2-phenylindole, and mounted in antifade solution (Vector Labs, Vectashield, California, USA). Each spot was evaluated and the predominant signal numbers were recorded for each FISH probe. Homozygous deletion of PTEN was defined as unequivocal complete absence of PTEN signals in all tumour cell nuclei of the tissue spot, but presence of centromere signals in tumour cells and presence of PTEN and CEP10 signals in adjacent normal cells. Hemizygous deletion of PTEN was defined as presence of fewer PTEN signals than CEP10 probe signals (ratio PTEN:CEP10 <1) in ≥60% tumour cell nuclei. Tissue spots with lack of PTEN signals in all (tumour and normal cells) or lack of any normal cells as an internal control for successful hybridisation of the PTEN probe were excluded from analysis.
Immunohistochemical (IHC) staining was performed with the HercepTest (Dako, Glostrup, Denmark) according to the manufacturer's protocol. Sections were deparaffinised and antigen retrieval was performed in a water bath at 95–99°C for 50 min. After peroxidase blocking the sections were incubated with the prediluted primary antibody. Immunostaining was scored following a standardised four-step scoring system.12 Tumours graded as 2+ and 3+ were evaluated as positive.
Continuous data are given as median with range or mean with SD. Cross-table analysis (χ2 test) was used to study the relationship between FISH and IHC results and clinicopathological features. Categorical variables are shown as numbers and percentage. Overall survival was estimated by the Kaplan–Meier method and evaluated by log-rank testing. Significance statements refer to p values of 2-tailed tests that are <0.05. Statistical calculation was performed with the PASW V.18.0 (SPSS).
A total of 520 of 770 (67.5%) tissue spots were interpretable for PTEN copy number estimation by FISH. Samples were from 180 primary cancers, 54 lymph nodes, two distant metastases (ovary, liver), 82 dysplasias (6 low-grade, 45 high-grade and 31 intramucosal carcinomas) and 202 normal gastric tissues. The remaining 250 (32.5%) spots were non-informative due to lack of tissue, absence of unequivocal dysplasia or cancer tissue, or insufficient hybridisation.
PTEN deletion in gastric cancer by FISH
Representative images of PTEN FISH findings are shown in figure 1. PTEN gene deletions were found in 8 of 180 (4.4%) evaluable primary gastric adenocarcinomas, including seven tumours with a hemizygous deletion and one tumour with a homozygous deletion. Additionally, one lymph node metastasis and one high-grade dysplasia showed a hemizygous PTEN deletion. Clinicopathological associations are given in table 1. We were able to show a significant correlation between PTEN deletion and lymph node metastases (p=0.041) as well as distant metastases (p<0.001). Another significant correlation was seen for PTEN deletion and tumour localisation (p=0.036). Furthermore, we found a tendency towards higher rates of PTEN deletion in intestinal type cancers. Seven of eight deletions were found in intestinal type cancer, yet without statistical significance. We were unable to show a relation between PTEN status and clinical outcome (figure 2).
HER2 expression in gastric cancer
HER2 positivity was seen in 19 of 216 (8.8%) analysable primary gastric cancers. We were able to show a significant correlation between HER2 positivity and WHO grading (p<0.001) as well as Lauren's classification (p=0.016). Another significant correlation was seen for HER2 positivity and intestinal metaplasia (p=0.029) and gastritis (p=0.027). One HER2 positive case showed a PTEN deletion.
PTEN status in primary tumour, corresponding metastases and associated dysplasia
Two deleted cases showed interpretable FISH results in corresponding metastases and associated dysplasia. One case showed an identical PTEN status in the primary tumour, associated dysplasia and corresponding metastasis. In the other case a discordant PTEN result occurred. The primary tumour showed a PTEN deletion, while a normal PTEN status was recorded in the corresponding lymph node metastasis. This case (no. 3) was included in our large section heterogeneity analysis.
Analysis of PTEN heterogeneity
To assess potential heterogeneity of the PTEN status in gastric cancers, all available tumour blocks were analysed by FISH on large tissue sections in seven cases (table 2). This analysis showed a homogeneous distribution of PTEN deletions in six of these seven cases. Case 3 showed a homogeneous distribution in all sections (4/4) of the primary tumour and in 5/8 evaluable nodal metastases. The three remaining nodal metastases showed an aberrant finding with increased copy number of both PTEN and CEP10 in all tumour cells. PTEN and also CEP10 signals were often co-located in a cluster-like arrangement in these tumour cells (figure 3).
The results of our study show that PTEN deletions occur rarely in gastric cancer. Only 8 of 180 evaluable primary gastric cancers (4.4%) showed a PTEN deletion by FISH. This rate of deletion is markedly lower than described in earlier studies using loss of heterozygosity (LOH) analysis for assessing 10q23 deletions. LOH studies had reported 10q23 deletions in 16–47% of gastric cancers (summarised in table 3 13–17). It is possible that a high rate of polysomic cases (unequal numbers of paternal and maternal chromosomes mimic LOH) might have contributed to the higher number of deletions found by LOH. Substantial aneuploidy has been described in gastric cancer in earlier studies by cytometry. For example, Furuya et al described aneuploidy in 98 of 186 analysed gastric cancers18 and Osterheld et al have seen aneuploidy in 15 of 16 analysed gastric cancers.19 It is unlikely that FISH would underestimate the true rates of deletion substantially as FISH enables a cell by cell analysis which is not disturbed by an admixture of inflammatory cells.
That direct PTEN alterations are rather rare in gastric cancer is also supported from earlier studies finding PTEN mutations by direct sequencing in 2–10% of cases (summarised in table 4 15 ,17 ,20–25). A markedly higher mutation rate (18.8%, 28.3%) was only suggested in two studies using the single-strand conformation polymorphism technique and subsequent DNA sequencing of positive cases.20 ,23 A validation of positive cases by a second independent PCR was not reported in these studies, however. One study reported a promoter hypermethylation in 39% of examined cases with a correlation to advanced tumour stage.25
High rates of PTEN alterations have been suggested by IHC studies describing PTEN expression loss in 18–77% of cases (summarised in table 5 21 ,25–33). Despite considerable variability in antibody selection, scoring criteria and frequency of PTEN loss, multiple IHC studies proposed a link between reduced PTEN expression and unfavourable tumour phenotype and disease progression.26 ,27 ,31–33 It can be speculated that post-transcriptional modifications of PTEN protein expression have a stronger impact on the cellular protein levels than copy number variations and are thus better distinguishable by IHC. Several mechanisms of post-transcriptional PTEN down-regulation have been suggested in the literature. For example, human protein NEDD4-1 (neural precursor cell expressed, developmentally down-regulated-4-1) was identified as a negative regulator for PTEN stability by catalysing PTEN polyubiquitination.34 Tumour transforming growth factor-β1 (TGF-β1) and the necrosis factor-α/nuclear factor-κB (NF-κB)-inducing kinase/NF-κB pathway were also shown to suppress PTEN transcription.35 ,36
It is noteworthy that a negative prognostic impact of PTEN alterations has been described for various other tumour types. Studies using FISH for PTEN deletion detection demonstrated the presence of PTEN genomic losses and its impact on development of advanced disease, for example in prostate cancer,37 gliomas38 and endometrial cancer.39
Perhaps due to a shortage of sufficiently large clinical studies, it is not entirely clear whether PTEN alterations have clinical significance in gastric cancer. Although only eight deleted cancers were found, our data seem to suggest a tendency towards more deletions in intestinal type and advanced stage gastric cancer. Similar observations were described in several IHC studies.21 ,25 ,28 ,31 ,32 A critical clinical significance of PTEN changes in gastric cancer may arise from its interactions with the HER2 pathway, however. Since studies have demonstrated a significant survival benefit of trastuzumab treated HER2 positive gastric cancers,4 there is substantial interest in biomarkers that will predict response to trastuzumab in individual patients. PTEN inactivation is considered a potential mechanism of trastuzumab resistance in breast cancer.9 ,40 In our subset, one of eight tumours with a PTEN deletion was HER2 positive. This patient might be affected in case of PTEN deletion representing a resistance mechanism to trastuzumab therapy.
Though PTEN deletion is rare in gastric cancer it seems to be of clinical and biological relevance. Our findings of a higher prevalence of lymphatic and haematogeneous metastases in PTEN deleted cases argue for a more aggressive behaviour of these tumours. The fact that PTEN deleted gastric cancer occurs more often in the cardia than in other parts of the stomach indicates a biological difference between cancers of the cardia and the distal stomach as, for example, suggested by Xue et al.41
However, in case of a predictive role of molecular features, it is important to determine whether a pivotal alteration is present in all or only a fraction of the tumour cells. Our finding of a homogeneous distribution of PTEN deletions in six of seven evaluated cases and also in corresponding dysplasia in one case makes it very likely that PTEN deletions occur early in gastric cancer. The highly homogeneous distribution makes it also unlikely that a significant number of heterogeneous deletions were missed in this analysis. Therefore it validates the TMA technique as a useful tool in this assay.
In another patient analysed for PTEN heterogeneity, a surprising finding of a cluster-like copy number increase was seen in three metastases, while the remaining cancer and metastases showed clear-cut PTEN deletion. It is possible that this unusual FISH finding reflects a structural rearrangement, potentially including a duplication of genomic PTEN material. Recent studies using next generation sequencing methods have demonstrated recurrent genomic disruption of the PTEN gene in prostate cancer.42
In summary, a thorough examination of 180 gastric cancers using FISH analysis showed PTEN gene deletion in 4.4% of carcinomas and a tendency towards tumour progression in such cases. Homogeneous deletion in six of seven cancers evaluated on large sections may suggest that PTEN deletions represent a critical event occurring early in a small subset of gastric cancers.
PTEN deletions occur rarely in gastric cancer. Only eight of 180 evaluable primary gastric cancers (4.4%) showed a PTEN deletion by fluorescence in-situ hybridisation.
The higher prevalence of lymphatic and haematogeneous metastases in PTEN deleted gastric cancers suggests a more aggressive behaviour of these tumours.
PTEN deletions may represent a critical event occurring early in a small subset of gastric cancers.
The authors thank Werner Otto Stiftung, Hamburg for their financial support. We thank Sascha Eghtessadi and Sylvia Schnöger for their support with the FISH technique.
SM and BAB have contributed equally to this work.
Funding Financial support was provided by Werner Otto Stiftung, Hamburg.
Competing interests None.
Ethics approval Ethics approval was provided by the Ethics Committee of the Chamber of Physicians in Hamburg, Germany.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.