Prognostic and Putative Predictive Biomarkers of Gastric Cancer for Personalized Medicine : Diagnostic Molecular Pathology

Journal Logo

00019606-201309000-00001ArticleDiagnostic Molecular PathologyDiagnostic Molecular Pathology© 2013 by Lippincott Williams & Wilkins.22September 2013 p 127–137Prognostic and Putative Predictive Biomarkers of Gastric Cancer for Personalized MedicineOriginal ArticlesWarneke, Viktoria S. MD*; Behrens, Hans-Michael MSc*,†; Haag, Jochen PhD*; Balschun, Katharina MD*; Böger, Christine MD*; Becker, Thomas MD, PhD‡; Ebert, Matthias P.A. MD, PhD§; Lordick, Florian MD, PhD∥; Röcken, Christoph MD, PhD*Departments of *Pathology‡General Surgery and Thoracic Surgery, Christian-Albrechts-University, Kiel†Department of Pathology, Charité University Hospital, Berlin§Department of Medicine II, Faculty of Clinical Medicine Mannheim, University of Heidelberg, Mannheim∥University Cancer Center Leipzig (UCCL), University of Leipzig, Leipzig, GermanySupplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website, and H.-M.B. share first authorship.C.R. is supported by grants of the Deutsche Forschungsgemeinschaft (Grant No. Ro 1173/11 and Ro 1173/12).The authors declare no conflict of interest.Reprints: Christoph Röcken, MD, PhD, Department of Pathology, Christian-Albrechts University, Arnold-Heller-Str. 3, Haus 14, D-24105 Kiel, Germany (e-mail: [email protected]).AbstractWe investigated various phenotypic and genotypic biomarkers of gastric cancer (GC) testing the following hypotheses: are these biomarkers suitable for the identification of GC subtypes, are they of prognostic significance, and should any of these biomarkers be considered to tailor patient treatment in the future. The study cohort consisted of 482 patients. pTNM-stage was based on surgical pathologic examination. The Laurén and mucin phenotype was assessed. Helicobacter pylori and Epstein-Barr virus infections were documented. The following biomarkers were determined: BRAF, KRAS, NRAS, and PIK3CA genotype, microsatellite instability, mucin 1, mucin 2, mucin 5, and mucin 6, CD10, E-cadherin, β-catenin, and lysozyme. The histologic phenotype correlated with 10/13 (77%) clinicopathologic patient characteristics and 6/13 (46%) immunohistochemical/molecular biological biomarkers. Inversely, immunohistochemical biomarkers (mucin phenotype, E-cadherin, β-catenin, and lysozyme) were unsuitable for subclassification of GC. It showed too much overlap between the different subtypes. Among the genotypes, only microsatellite instability correlated with tumor type being more prevalent in intestinal and unclassified GCs. Patient survival correlated significantly with 8 (62%) clinicopathologic and 5 (36%) immunohistochemical/molecular biomarkers. Interestingly, in proximal GCs, KRAS mutation was associated with worse prognosis, as was persistent H. pylori infection in unclassified GCs. Mucin 2 (all patients, proximal GCs) and PIK3CA (exon 20; intestinal type GC) prognosticated independently patient survival. The biomarkers examined herein are unsuitable to aid histologic classification of GC. However, several of them show a correlation with either phenotype and/or prognosis and may be considered to tailor patient treatment in the future, such as KRAS, PIK3CA, MSI, and H. pylori status.Gastric cancer (GC) is one of the most common cancers worldwide. In recent decades we witnessed major advancements in the understanding of the epidemiology, pathology, and pathogenesis of GC. Infection with Helicobacter pylori or Epstein-Barr virus, dietary and lifestyle factors contribute to the risk of developing GC. These advancements were accompanied by the introduction of chemotherapy for the treatment of GC, which is evolving continuously and improves patient survival.1–3 However, there is overwhelming evidence from a variety of cancers that patient prognosis and treatment responses do not only depend on tumor stage but also on phenotypic and genotypic tumor characteristics. These are increasingly used to individualize patient management. With regard to GC, and with the exception of trastuzumab, chemotherapy of localized and advanced GC still does not consider phenotypic and genotypic tumor characteristics. This implies that part of the patients, if not the majority, receives medical treatment with suboptimal or even lacking efficacy. Compared with colon and lung cancer, phenotypic and genotypic characterization of GC is still in its infancies and cannot yet be applied to tailor patient treatment. A more in-depth view is urgently needed. In this study, we carried out a comprehensive analysis of phenotypic and genotypic biomarkers for GC on a patient cohort, testing the following hypotheses: are these biomarkers suitable for the identification of GC subtypes, are they of prognostic significance, and should any of these biomarkers be considered to tailor patient treatment in the future.MATERIALS AND METHODSEthics StatementThis project was approved by the local ethics committee of the University Hospital in Kiel, Germany (reference number D 453/10). All patient data were pseudonymized before study inclusion.Study PopulationFrom the archive of the Institute of Pathology, University Hospital Kiel, we identified all Caucasian patients who had undergone either total or partial gastrectomy for adenocarcinomas of the stomach or esophagogastric junction between 1997 and 2009. The following patient characteristics were retrieved: type of surgery, age at diagnosis, sex, tumor localization, tumor size, tumor type, tumor grade, depth of invasion, number of lymph nodes resected, and number of lymph nodes with metastases. Date of patient death was obtained from the Epidemiological Cancer Registry of the state of Schleswig-Holstein, Germany. Follow-up data of patients that are still alive were retrieved from hospital records and general practitioners.Study Inclusion and Exclusion CriteriaInclusion and exclusion criteria were defined as follows: patients were included when: (1) histology confirmed an adenocarcinoma of the stomach or esophagogastric junction, and (2) the date of death or survival data were available. Patients were excluded when: (1) histology identified a tumor type other than adenocarcinoma, (2) histopathologic data were incomplete, (3) patients had previously undergone a resection of a Billroth-II stomach with cancer in the gastric remnant, and (4) date of patient death or survival data had not been recorded. Patients who received perioperative chemotherapy were also excluded.Histology and TNM ClassificationTissue specimens were fixed in formalin and embedded in paraffin. Deparaffinized sections were stained with hematoxylin and eosin. Tumors were classified according to the Laurén classification4 and the mucin phenotype.5 All cases included in this study were reexamined by 2 surgical pathologists (V.S.W. and C.R.). pTNM-stage of all study patients was determined according to the seventh edition of the UICC guidelines6 and our recent proposal (Kiel-stage),7 and was based solely on surgical pathologic examination including classification of distant metastases (pM-category). In the seventh edition, all tumors of the esophagogastric junction as well as tumors of the proximal 5 cm of the stomach with extension into the esophagus are classified as esophageal tumors.6 Patients were recategorized accordingly.Tissue Micro Array (TMA) ConstructionFormalin-fixed and paraffin-embedded tissue samples were used to generate TMAs as described previously.8 Three morphologically representative regions of the paraffin “donor” blocks were chosen. Tissue cylinders of 1.5 mm diameter were punched from these areas and precisely arrayed into a new “recipient” paraffin block. Sections (4 µm) of the TMA blocks were cut for further analysis.ImmunohistochemistryImmunohistochemistry was carried out with monoclonal antibodies directed against mucin 1 (clone MA695, dilution 1:100), mucin 2 (clone Ccp58, 1:100; both Novocastra, Leica Microsystems GmbH, Wetzlar, Germany), mucin 5 (clone 45M1, 1:100; Thermo Scientific, Schwerte, Germany), mucin 6 (clone CLH5, 1:100), CD10 (clone 56C6, 1:10; both Novocastra), E-cadherin (clone SPM471; 1:400; ZYTOMED Systems GmbH, Berlin, Germany), β-catenin (clone Cat-5H10, 1:300; Life Technologies GmbH, Darmstadt, Germany), MLH1 (clone G168-15, 1:50; BD Biosciences, Heidelberg, Germany), PMS2 (clone MRQ-28, 1:20; Cell Marque Corporation, Rocklin, CA), MSH2 (clone FE11, 1:30; Calbiochem, Merck KGaA, Darmstadt, Germany, Germany), MSH6 (clone 44, 1:30; BD Biosciences), and a polyclonal antibody directed against lysozyme (1:3000; DAKO, Glostrup, Denmark).Antigen retrieval was done with ER1 (citrate buffer bond pH 6.0; mucin 1, CD10), ER2 (EDTA-buffer bond pH 8.9; mucin 2, mucin 6, E-cadherin, β-catenin, PMS2), Enz1 (protease bond; BerEP4), and TEC (Tris-EDTA-citrate pH 7.8; MLH1, MSH2, MSH6) according to the manufacturer’s instructions and using the autostainer Bond Max System (Menarini, Berlin, Germany). Immunostaining was done with the BondMax (Leica Microsystems GmbH, Wetzlar, Germany) and the UltraView Universal Alkaline Phosphatase Red Detection Kit (antilysozyme antibody; both Ventana Medical Systems Inc., Tucson, AZ) or the Ultravision Detection system (all other antibodies; Thermo ScientificBond, LabVision Corp., Fremont, CA). The immunoreaction was visualized with either the Bond Polymer Refine Detection Kit (DS 9800, brown labeling) or the Bond Polymer Refine Red Detection Kit (DS 9390, red labeling; both Novocastra). Epstein-Barr virus–encoded RNA was detected using the EBER probe (Novocastra) and the BondMax detection system according to the manufacturer’s instructions (Leica Microsystems GmbH).Evaluation of ImmunostainingImmunostaining of the TMAs was evaluated by applying an immunoreactivity scoring system. Briefly, category A documented the intensity of immunostaining as 0 (no immunostaining), 1 (weak), 2 (moderate), and 3 (strong). Category B documented the percentage of immunoreactive cells as 0 (no immunoreactive cells), 1 (few scattered immunoreactive cells, <1%), 2 (1% to 10%), 3 (11% to 50%), 4 (51% to 80%), and 5 (>80%). The addition of category A and B resulted in an immunoreactivity scoring system ranging from 0 to 8 for each individual case.DNA IsolationGenomic DNA was extracted from formalin-fixed and paraffin-embedded tissue using the QIAamp DNA mini kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. The integrity and amplifiability of the isolated DNA was evaluated by a qualitative size range PCR assay.9 Tissue sections were manually microdissected before DNA isolation to enrich for tumor cells (>40%).Mutational AnalysisMutational analyses of codons 12 and 13 of the KRAS and NRAS genes, codon 600 of the BRAF gene, and mutational hotspots in exons 9 and 20 of the PIK3CA gene were performed by pyrosequencing on a PyroMark Q24 instrument (Qiagen). Fragments of the different genes were amplified by PCR with following primers: KRAS (codon 12/13), forward 5′-NNNGGCCTGCTGAAAATGACTGAA-3′ and reverse 5′-TTAGCTGTATCGTCAAGGCACTCT-3′10; NRAS (codon 12/13), forward 5′-CTTGCTGGTGTGAAATGACTG-3′ and reverse 5′-TTCTGGATTAGCTGGATTGTCAGT-3′; BRAF (codon 600), forward 5′-TGAAGACCTCACAGTAAAAATAGG-3′ and reverse 5′-TCCAGACAACTGTTCAAACTGAT-3′; PIK3CA exon 9, forward 5′-AACAGCTCAAAGCAATTTCTACACG-3′ and reverse 5′-ACCTGTGACTCCATAGAAAATCTTT-3′; and PIK3CA exon 20, forward 5′-CAAGAGGCTTTGGAGTATTTCA-3′ and reverse 5′-CAATCCATTTTTGTTGTCCA-3′.11 The resulting PCR products were analyzed by agarose electrophoresis and sequenced using following sequencing primers: KRAS (codon 12/13), 5′-TGTGGTAGTTGGAGCT-3′; NRAS (codon12/13), 5′-GTGGTGGTTGGAGCA-3′; BRAF (codon 600), 5′-GTAAAAATAGGTGATTTTGG-3′; PIK3CA exon 9, 5′-TAGAAAATCTTTCTCCTGCT-3′ and 5′-TTCTCCTGCTCAGTGATTT-3′; and PIK3CA exon 20, 5′-GTTGTCCAGCCACCA-3′.Microsatellite Instability (MSI) AssayMSI was determined by comparison of the allelic profiles of the mononucleotide repeat markers BAT-25, BAT-26, NR-21, NR-24, and NR-27 in tumor and corresponding normal tissue.12 All markers were coamplified in a pentaplex PCR assay with the QIAGEN Multiplex PCR Master Mix (Qiagen) following the manufacturer’s recommendations for amplification of microsatellite loci. The amplified loci were analyzed on an ABI Prism 310 Genetic Analyzer (Applied Biosystems, Darmstadt, Germany). Samples were judged as microsatellite unstable (MSI-H) when the tumor showed instability in at least 2 of the 5 (40%) microsatellites analyzed.External Quality AssuranceThe KRAS mutational assay, the MSI assay, and the immunohistochemical evaluation of DNA mismatch repair proteins (MSH2, MSH6, MLH1, and PMS2) were certified successfully by the quality assurance program of the German Society of Pathology and the Bundesverband Deutscher Pathologen e.V.Detection of H. pylori by Histology and Polymerase Chain ReactionInfection with H. pylori was evaluated histologically using the modified Giemsa staining and PCR.13H. pylori-specific DNA sequences were detected by a PCR-based assay targeting the 16S rRNA gene of H. pylori.14 In short, a 109 bp fragment of the 16S rRNA gene was amplified with the primers Hp1 (5′-CTGGAGAGACTAAGCCCTCC-3′) and Hp2 (5′-ATTACTGACGCTGATTGTGC-3′) following the PCR parameters described by Chisholm et al.14 Amplification products were assessed by agarose gel electrophoresis.StatisticsStatistical analyses were performed using SPSS 20.0 (IBM Corporation). For continuous variables, cases were divided into 2 groups by splitting at the median value. Median overall survival was determined using the Kaplan-Meier method, and the log-rank test was used to determine significance. For comparison purposes, the median survival time, its SD, and 95% confidence interval were calculated. To investigate prognostic relevance, we included all variables having P<0.10 into a Cox regression model and used the backward LR method (Pin and Pout=0.05) to reduce the model to the independent variables. The significance of correlation between clinicopathologic parameters and biomarker expressions was tested using the Fisher exact test. For parameters of ordinal scale (T-category, N-category, and tumor stage) we applied Kendall τ test instead. A P≤0.05 was considered statistically significant. To account for the effects of multiple testing, we applied the explorative Simes (Benjamini-Hochberg) procedure within each group of tests (correlations, log-rank tests, and Cox regressions). The P values are given unadjusted but are marked where they lose significance under the explorative Simes procedure.RESULTSPhenotypic and Genotypic Characterization of the Study PopulationA total of 482 patients fulfilled all study criteria (Table 1). Non-neoplastic mucosa was available from 411 patients and was screened for H. pylori. Sixty-two (15.1%) patients had a persistent infection with H. pylori. EBV-RNA was found in 15 (4.0%) GCs (Table 2; Fig. 1).JOURNAL/dimp/04.03/00019606-201309000-00001/table1-1/v/2021-02-17T200100Z/r/image-tiffClinicopathologic Patient CharacteristicsJOURNAL/dimp/04.03/00019606-201309000-00001/table2-1/v/2021-02-17T200100Z/r/image-tiffPhenotype-Genotype Correlation of Gastric CancerJOURNAL/dimp/04.03/00019606-201309000-00001/figure1-1/v/2021-02-17T200100Z/r/image-jpegPhenotypes of gastric cancer. The mucin phenotype was explored by immunostaining using antibodies directed against mucin 1 (A), mucin 2 (B), mucin 5 (C), mucin 6 (D), and CD10 (E). Note the membranous and cytoplasmic immunoreactions. Epstein-Barr virus–encoded RNA was detected using the EBER probe showing strong nuclear staining (F). E-cadherin (G) and β-catenin (H) showed membranous staining (G and H, left) or loss of expression (G and H, right). Lysozyme was found in tumor cells of intestinal (I, left) and diffuse (I, right) type gastric cancer. MLH1 (J) in microsatellite-stable gastric cancer showed nuclear expression. Note loss of nuclear expression of MLH1 (K) and PMS2 (L) in a MSI-H gastric cancer. Original magnifications ×400 (A-L).According to Laurén, an intestinal type GC was found in 247 (50.9%), a diffuse type in 152 (31.3%), a mixed type in 30 (6.2%), and an unclassifiable type in 53 (10.9%) patients. According to the mucin phenotype, 169 (39.9%) GCs were of the mixed, 123 (29.0%) of the intestinal, 69 (16.3%) of the unclassified, and 63 (14.9%) of the gastric type (Table 2). Two hundred twenty-six (50.8%) GCs were categorized as lysozyme positive, 196 (43.9%) as β-catenin positive, and 119 (26.9%) as E-cadherin positive. The immunostaining characteristics for β-catenin, CD10, E-cadherin, lysozyme, mucin 1, mucin 2, mucin 5, and mucin 6 are detailed in Supplementary Table 1 (Supplemental Digital Content 1, KRAS, NRAS, BRAF, and PIK3CA genotype was determined in 475 primary GCs. Seventeen GCs (3.6%) showed a KRAS mutation, 12 (2.5%) a PIK3CA (exon 9) mutation, and 9 (1.9%) a PIK3CA (exon 20) mutation. NRAS and BRAF mutations were not found in any of the patients screened (Table 2; Supplementary Table 2, Supplemental Digital Content 2, was assessed by immunohistochemistry and subsequent molecular pathologic MSI analysis in 451 patients. One hundred fifty-eight GCs showed a decreased (at least mildly) or missing immunoreaction of any of the 4 DNA mismatch repair proteins studied, that is, MSH2, MSH6, MLH1, or PMS2 and were selected for subsequent molecular pathologic analysis, which identified 33 (7.3%) highly microsatellite-unstable GCs (MSI-H). Thirty-one (93.9%) of the MSI-H GCs were associated with reduced or loss of expression of MLH1 and PMS2, and 2 (6.1%) showed reduced or loss of expression of MSH2 and MSH6 (Fig. 1).Statistical AnalysesCorrelation of Phenotype With GenotypeFirst we compared the phenotype of GC according to Laurén with the prevalence of infectious agents (H. pylori and EBV), E-cadherin, β-catenin, and lysozyme status, the mucin phenotype and the genotype (MSI, KRAS, PIK3CA, NRAS, and BRAF). The phenotype according to Laurén correlated with E-cadherin (P<0.001), β-catenin (P<0.001), and lysozyme status (P=0.007; Table 2). It also correlated with the MSI status (P<0.001). Microsatellite-unstable GCs were more commonly of intestinal (63.3%) or unclassified (33.3%) type than diffuse (3.0%) or mixed type (0%; P<0.001; Table 2). No correlation was found between mucin phenotype and MSI (Supplementary Table 3, Supplemental Digital Content 3, infection with H. pylori and EBV did not correlate with the Laurén phenotype, it was interesting to note that EBV was more prevalent in diffuse and unclassified type GC (Suppl. Table 3).JOURNAL/dimp/04.03/00019606-201309000-00001/table3-1/v/2021-02-17T200100Z/r/image-tiffPrognostic Markers of Gastric Cancer (Univariate Analysis)The analysis of the genotype showed that a KRAS mutation is rare in diffuse type GC (11.8%) and more prevalent in unclassified GCs (23.5%). PIK3CA exon 9 mutations were commonly found in intestinal type GCs (75%). However, neither observation reached statistical significance due to the overall small number of KRAS and PIK3CA mutations (Table 2).Correlation of Phenotype/Genotype With Clinicopathologic Patient CharacteristicsNext we tested the hypothesis that the phenotype and genotype of GC correlates with clinicopathologic patient characteristics (Supplementary Table 3, Supplemental Digital Content 3, correlated with Laurén, mucin 1 and CD10 (P<0.001 or P<0.004, respectively). Patient age correlated with Laurén phenotype (P<0.001). Patients aged 68 years and older commonly harbored microsatellite-unstable GCs than patients younger than 68 years (P=0.025; insignificant after correction for multiple testing). Local tumor growth (T-category) and nodal spread (N-category) correlated with tumor type according to Laurén, CD10, and β-catenin. The R-status correlated with Laurén phenotype (P=0.002). Microsatellite-unstable tumors less commonly metastasized into lymph nodes and showed a lower lymph node ratio (Supplementary Table 3, Supplemental Digital Content 3, In addition, KRAS mutations (P=0.005) and PIK3CA exon 20 mutations (P=0.016; insignificant after correction for multiple testing) were more common in MSI-H GCs, whereas the PIK3CA exon 9 mutations were slightly more often found in men and were virtually absent in the mixed type of the mucin phenotype (Supplementary Table 3, Supplemental Digital Content 3, However, the latter findings did not reach statistical significance due to small sample numbers.Prognostic Markers of GCNext we explored the prognostic significance of phenotypic and genotypic characteristics of GC. Patient prognosis significantly depended on Laurén-phenotype, tumor grade, T-category, N-category, LNR, R-status, as well as UICC stage and “Kiel-stage” (Supplementary Table 3, Supplemental Digital Content 3, difference in patient survival was also evident for microsatellite unstable (35.0±16.6 mo) versus stable tumors (14.1±1.1 mo; Table 3). No correlation was found between patient survival and KRAS or PIK3CA status in the entire study cohort.The subgroup analyses (intestinal vs. diffuse, proximal vs. distal) showed that the T-category, N-category, LNR, R-status, UICC stage, and “Kiel-stage” correlated highly significantly with patient survival in every subgroup (data not shown). Interestingly, the expression of CD10, mucin 2, and mucin 5 was prognostically significant in different combinations for unclassified, intestinal, and mixed type GCs (Table 3). With regard to genotype, KRAS mutant proximal GCs had a worse prognosis compared with their KRAS wild-type counterparts (3.5±3.1 mo vs. 12.7±0.7 mo; Table 3). The PIK3CA exon 20 mutation was associated with opposite prognostic significances for intestinal and diffuse type GC (Table 3). However, the number of patients with PIK3CA mutation was low (4 and 2 patients, respectively).The Kaplan-Meier plots showed significant differences for the Laurén phenotype (P=0.012), R-status (P<0.001), lysozyme status (P=0.035), and MSI (P=0.036; Fig. 2) in the entire cohort, and for the Laurén phenotype (P=0.003; not shown) and KRAS genotype (P=0.021; not shown) in proximal GCs. Interestingly, unclassified GCs with persistent H. pylori infection had a significant worse prognosis compared with patients without H. pylori infection (P=0.003; Fig. 2; Table 3). Expression of CD10 was also associated with a significantly worse prognosis in unclassified GCs (P>0.003; Fig. 2). KRAS mutant intestinal type GCs had a worse prognosis compared with nonmutant intestinal type GC (Fig. 2). However, this difference did not reach significance on univariate analysis (P=0.098). Similarly, on univariate analysis, mucin 2-positive had a more favorable prognosis than mucin 2-negative GCs (P=0.067).JOURNAL/dimp/04.03/00019606-201309000-00001/figure2-1/v/2021-02-17T200100Z/r/image-jpegPatient survival. Kaplan-Meier curves depicting patient survival according to Laurén phenotype (all cases, P=0.012), resection margin (all cases; P=0.001), mucin 2 (all cases, P=0.067), CD10 (unclassified GCs; P=0.003), lysozyme status (all cases, P=0.035), Helicobacter pylori infection (unclassified GCs, P=0.003), MSI status (all cases, P=0.036), and KRAS genotype (intestinal, P=0.098). Mucin 2 (all cases) and R-status (all cases) were independent prognosticators of patient survival in multivariate analyses. All cases indicates entire study population; GC, gastric cancer; MSI, microsatellite instability; unclassified, unclassified GCs according to Laurén.Explorative Multivariate AnalysisExplorative multivariate survival analysis was done with parameters that had shown a P-value of <0.100 on univariate analysis. For the entire study population, patient age, T-category, LNR, R-status, and mucin 2 were found to be highly significantly independent prognosticators of patient survival (Table 4). The subgroup analysis confirmed the independent prognostic significance of age group, LNR, tumor grade, R-status, mucin phenotype, and PIK3CA exon 20 genotype for intestinal and of T-category, N-category, and R-status for diffuse type GCs.JOURNAL/dimp/04.03/00019606-201309000-00001/table4-1/v/2021-02-17T200100Z/r/image-tiffExplorative Multivariate Survival AnalysisWith regard to tumor localization, the N-category, R-status, and mucin 2 status independently prognosticated patient survival for proximal GCs, whereas patient age, T-category, LNR, and R-status were independent prognosticators of patient survival in distal GCs (Table 4).DISCUSSIONUntil few years ago, oncologic treatment of malignant epithelial tumors largely depended on the anatomic tumor site. However, with the advancements of targeted therapy it became increasingly evident that cancers of the same anatomic origin, for example, lung or colon, show great variability in their response rates to chemotherapies necessitating a more in depth phenotypic/genotypic classification before treatment. Although this has lead to major improvements in lung and colon cancer, it is still in its infancies in GC and, currently, only Her2/neu is used to tailor patient treatment. Recently, Tan et al15 examined the influence of the intrinsic gene expression profile on chemotherapy response in GC cell lines. GC cell lines with the intestinal gene profile were significantly more sensitive to 5-fluorouracil and oxaliplatin, but more resistant to cisplatin, than the cell lines with a gene expression profile of the diffuse type. Thus, the phenotypic/genotypic classification of GC may be increasingly required to tailor patient treatment. Biomarkers aiding subclassification of GC are urgently needed.In this study we aimed to fill this gap by carrying out a comprehensive and comparative analysis of various etiologic, phenotypic, and genotypic variables of GC. In line with many previous studies, we confirmed that patient prognosis significantly depends on the Laurén phenotype, tumor grade, tumor growth (T-category), nodal spread (N-category and LNR), resection status (R-status), and tumor stage, thereby validating our patient cohort. In addition to that we also show that microsatellite-unstable GCs have a much more favorable prognosis, compared with microsatellite-stable GCs, and that PIK3CA (exon 20) might be an independent prognosticator of patient survival in intestinal type GC.Occasionally, histologic classification of tumors (particularly in biopsy specimens) can be cumbersome and immunohistochemistry is applied to aid subclassification, for example, in breast,16,17 lung,18,19 or kidney cancer.20 Interestingly, although E-cadherin, β-catenin, and lysozyme are significantly differentially expressed among intestinal and diffuse type GCs, neither can be used to reliably differentiate between any of the two. This finding is in accordance with the recent study published by Tan et al,15 who screened 37 primary GC cell lines on the translational level. A total of 171 genes separated intestinal from diffuse type GC. The separating gene expression profiles were validated on an independent set of primary GC specimens15 and demonstrated to be of prognostic significance. Lysozyme was among the genes highly significantly differentially expressed in the study reported by Tan et al15 and was selected by us as another putative immunohistochemical marker. Interestingly, Tan et al15 reported an overall concordance between the intrinsic genomic subtypes and Lauren’s histopathologic classification only in 64% of the cases.15 In our study, lysozyme was expressed in the tumor cells of 262 GCs, being significantly more commonly found in diffuse type (62.2%) compared with intestinal type (46.8%) GCs. Like for E-cadherin and β-catenin, the differential expression of lysozyme was highly significant among the intestinal and diffuse type GC and was even of prognostic significance. However, the expression still showed substantial overlap. The mucin phenotype has also been recommended to classify GC.5 Although we were able to show that the Laurén and mucin phenotype correlated with each other, immunohistochemistry cannot be recommended to aid histologic subclassification of GC. Immunostaining characteristics show too much overlap between the different histologic subtypes.Infection with H. pylori and EBV are known risk factors for the development of GC. H. pylori was linked to the intestinal phenotype and EBV to the diffuse phenotype of GC.21 The low prevalence of persistent H. pylori infections in our cohort (15.1%) may explain the lack of correlation with Laurén phenotype. However, with regard to EBV, we noticed the same prevalence between intestinal (40%) and diffuse (40%) type GCs. Thus, tumor histology cannot be used for the consideration (or exclusion) of an EBV infection. More interestingly, neither infection with H. pylori nor with EBV correlated with any other molecular biomarker analyzed in our study.Novel technologies can help to advance the molecular classification of malignant tumors, including GC, and to tailor oncologic treatment.22–24 Zang et al24 recently carried out exome sequencing on 15 GCs (11 intestinal type, 3 diffuse, and 1 mixed type according to Laurén). Fifty nonsynonymous somatic mutations were found on average with genes coding for cell adhesion, chromatin remodeling, and epigenetic modification being most commonly altered. Lee et al23 applied high-throughput mutation profiling on 237 GCs and found mutations of PIK3CA, p53, APC, STK11, CTNNB1, and CDKN2A. Deng et al22 recently studied 193 primary GCs and 40 GC cell lines. They found 150 different genetic alterations including gene amplifications and deletions. Thirty-seven percent of the GCs harbored receptor tyrosin kinase gene amplifications affecting FGFR2, EGFR, ERBB2, and cMET. An amplification of the KRAS gene was found in 8.8% of the patients. Applying conventional sequencing technologies, we analyzed the mutational status of 4 separate genes (BRAF, KRAS, NRAS, and PIK3CA) and we assessed the MSI status in our patient cohort aiming for a more in-depth phenotypic/genotypic correlation and to assess their putative prognostic significance. Although the mutational status of KRAS and PIK3CA was not randomly distributed among the histologic subtypes, their overall prevalence was too low to aid histologic classification. Inversely, should testing of any of these biomarkers be considered to tailor patient treatment in the future at all? Subgroup analyses showed that the median survival of KRAS mutant proximal GCs was 3.5±3.1 months compared with 12.7±0.7 months for KRAS wild-type GCs. Thus, KRAS testing may be recommended in proximal GCs to identify patients with a very poor prognosis (Table 3).MSI correlates with patient prognosis and may also influence response to 5-fluoruracil-based chemotherapy.25,26 In part, this may also apply to GC. Several previous studies have shown that microsatellite-unstable GCs have a much more favorable prognosis than microsatellite-stable GCs. However, the range of MSI varies considerably (9.5% to 23.1%).27–34 A shortcoming of previous investigations was the methodology. The number of markers used to assess MSI ranged from 1 to 40. There is currently no standard for MSI analysis of GC and markers applied for CRC may be unsuitable for GC. In our study, we used a pentaplex panel of mononucleotide markers. The NCI-recommended panel of markers applied for CRC uses 2 mononucleotide and 3 dinucleotide markers. Dinucleotide markers are usually less sensitive and less specific than mononucleotide markers and the NCI-recommended panel carries the risk of underestimating the rate of MSI-H cancers.35 A pentaplex panel of mononucleotide repeats performs better than the NCI-recommended panel for the detection of mismatch repair–deficient tumors.36 Immunohistochemical analyses of GCs for MSI usually investigated two mismatch repair proteins27,28,30–33 and rarely four.29,34 In our series we identified 33 (7.0%) MSI-H GCs. Patients with MSI-H GCs were older and lived longer than patients with MSS GCs, confirming previous findings.28,31,33,37 On the basis of these observations one may consider avoiding chemotherapy in this small subgroup of elderly GC patients,38 and testing MSI is meaningful in GC.Several studies underline the importance and necessity of a genotypic and phenotypic classification of GC with regard to patient treatment. Our comprehensive study of 13 different biomarkers showed that they are unsuitable to aid histologic classification of GC. However, several of them correlated with either phenotype and/or patient prognosis and may be considered to tailor patient treatment in the future, such as KRAS, PIK3CA, MSI, and H. pylori status.ACKNOWLEDGMENTThe authors thank Sandra Krüger for her excellent technical assistance.REFERENCES1. Alberts SR, Cervantes A, van de Velde CJ.Gastric cancer: epidemiology, pathology and treatment.Ann Oncol.2003;14suppl 2ii31–ii36.[Context Link]2. Cunningham D, Allum WH, Stenning SP, et al..Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer.N Engl J Med.2006;355:11–20.[Context Link]3. Paoletti X, Oba K, Burzykowski T, et al..Benefit of adjuvant chemotherapy for resectable gastric cancer: a meta-analysis.JAMA.2010;303:1729–1737.[Context Link]4. Lauren T.The two histologic main types of gastric carcinoma: Diffuse and so-called intestinal-type carcinoma.Acta Pathol Microbiol Scand.1965;64:31–49.[Context Link][CrossRef][Medline Link]5. Namikawa T, Hanazaki K.Mucin phenotype of gastric cancer and clinicopathology of gastric-type differentiated adenocarcinoma.World J Gastroenterol.2010;16:4634–4639.[Context Link][CrossRef][Medline Link]6. Sobin LH, Gospodarowicz M, Wittekind C.TNM Classification of Malignant Tumours.2009.Oxford, England:Wiley-Blackwell.[Context Link]7. Warneke VS, Behrens HM, Hartmann JT, et al..Cohort study based on the seventh edition of the TNM classification for gastric cancer: proposal of a new staging system.J Clin Oncol.2011;29:2364–2371.[Context Link]8. Weichert W, Röske A, Gekeler V, et al..Association of patterns of class I histone deacetylase expression with patient prognosis in gastric cancer: a retrospective analysis.Lancet Oncol.2008;9:139–148.[Context Link]9. van Dongen JJ, Langerak AW, Bruggemann M, et al..Design and standardization of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936.Leukemia.2003;17:2257–2317.[Context Link]10. Ogino S, Kawasaki T, Brahmandam M, et al..Sensitive sequencing method for KRAS mutation detection by pyrosequencing.J Mol Diagn.2005;7:413–421.[Context Link]11. Nosho K, Kawasaki T, Ohnishi M, et al..PIK3CA mutation in colorectal cancer: relationship with genetic and epigenetic alterations.Neoplasia.2008;10:534–541.[Context Link]12. Buhard O, Cattaneo F, Wong YF, et al..Multipopulation analysis of polymorphisms in five mononucleotide repeats used to determine the microsatellite instability status of human tumors.J Clin Oncol.2006;24:241–251.[Context Link]13. Fischbach W, Malfertheiner P, Hoffmann JC, et al..S3-guideline “helicobacter pylori and gastroduodenal ulcer disease” of the German society for digestive and metabolic diseases (DGVS) in cooperation with the German society for hygiene and microbiology, society for pediatric gastroenterology and nutrition e. V., German society for rheumatology, AWMF-registration-no. 021/001.Z Gastroenterol.2009;47:1230–1263.[Context Link]14. Chisholm SA, Owen RJ, Teare EL, et al..PCR-based diagnosis of Helicobacter pylori infection and real-time determination of clarithromycin resistance directly from human gastric biopsy samples.J Clin Microbiol.2001;39:1217–1220.[Context Link]15. Tan IB, Ivanova T, Lim KH, et al..Intrinsic subtypes of gastric cancer, based on gene expression pattern, predict survival and respond differently to chemotherapy.Gastroenterology.2011;141:476–485485.[Context Link]16. Qureshi HS, Linden MD, Divine G, et al..E-cadherin status in breast cancer correlates with histologic type but does not correlate with established prognostic parameters.Am J Clin Pathol.2006;125:377–385.[Context Link]17. Singhai R, Patil VW, Jaiswal SR, et al..E-Cadherin as a diagnostic biomarker in breast cancer.N Am J Med Sci.2011;3:227–233.[Context Link]18. Travis WD, Brambilla E, Noguchi M, et al..International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society: international multidisciplinary classification of lung adenocarcinoma: executive summary.Proc Am Thorac Soc.2011;8:381–385.[Context Link]19. Travis WD, Rekhtman N.Pathological diagnosis and classification of lung cancer in small biopsies and cytology: strategic management of tissue for molecular testing.Semin Respir Crit Care Med.2011;32:22–31.[Context Link][Full Text][Medline Link]20. Lhermitte B, de LL.Interpretation of needle biopsies of the kidney for investigation of renal masses.Virchows Arch.2012;461:13–26.[Context Link]21. Chen JN, He D, Tang F, et al..Epstein-Barr virus-associated gastric carcinoma: a newly defined entity.J Clin Gastroenterol.2012;46:262–271.[Context Link]22. Deng N, Goh LK, Wang H, et al..A comprehensive survey of genomic alterations in gastric cancer reveals systematic patterns of molecular exclusivity and co-occurrence among distinct therapeutic targets.Gut.2012;61:673–684.[Context Link]23. Lee J, van HP, Go C, et al..High-throughput mutation profiling identifies frequent somatic mutations in advanced gastric adenocarcinoma.PLoS One.2012;7:e38892.[Context Link]24. Zang ZJ, Cutcutache I, Poon SL, et al..Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes.Nature Genet.2012;44:570–574.[Context Link]25. Damia G, D’Incalci M.Genetic instability influences drug response in cancer cells.Curr Drug Targets.2010;11:1317–1324.[Context Link][CrossRef][Medline Link]26. Des GG, Schischmanoff O, Nicolas P, et al..Does microsatellite instability predict the efficacy of adjuvant chemotherapy in colorectal cancer? A systematic review with meta-analysis.Eur J Cancer.2009;45:1890–1896.[Context Link]27. Bacani J, Zwingerman R, Di NN, et al..Tumor microsatellite instability in early onset gastric cancer.J Mol Diagn.2005;7:465–477.[Context Link]28. Beghelli S, de MG, Barbi S, et al..Microsatellite instability in gastric cancer is associated with better prognosis in only stage II cancers.Surgery.2006;139:347–356.[Context Link]29. Carvalho R, Milne AN, Van Rees BP, et al..Early-onset gastric carcinomas display molecular characteristics distinct from gastric carcinomas occurring at a later age.J Pathol.2004;204:75–83.[Context Link]30. Chiaravalli AM, Furlan D, Facco C, et al..Immunohistochemical pattern of hMSH2/hMLH1 in familial and sporadic colorectal, gastric, endometrial and ovarian carcinomas with instability in microsatellite sequences.Virchows Arch.2001;438:39–48.[Context Link]31. Falchetti M, Saieva C, Lupi R, et al..Gastric cancer with high-level microsatellite instability: target gene mutations, clinicopathologic features, and long-term survival.Hum Pathol.2008;39:925–932.[Context Link]32. Kulke MH, Thakore KS, Thomas G, et al..Microsatellite instability and hMLH1/hMSH2 expression in Barrett esophagus-associated adenocarcinoma.Cancer.2001;91:1451–1457.[Context Link]33. Lee HS, Choi SI, Lee HK, et al..Distinct clinical features and outcomes of gastric cancers with microsatellite instability.Mod Pathol.2002;15:632–640.[Context Link]34. Leite M, Corso G, Sousa S, et al..MSI phenotype and MMR alterations in familial and sporadic gastric cancer.Int J Cancer.2011;128:1606–1613.[Context Link]35. Pino MS, Chung DC.Application of molecular diagnostics for the detection of Lynch syndrome.Expert Rev Mol Diagn.2010;10:651–665.[Context Link][Full Text][CrossRef][Medline Link]36. Xicola RM, Llor X, Pons E, et al..Performance of different microsatellite marker panels for detection of mismatch repair-deficient colorectal tumors.J Natl Cancer Inst.2007;99:244–252.[Context Link]37. Oki E, Kakeji Y, Zhao Y, et al..Chemosensitivity and survival in gastric cancer patients with microsatellite instability.Ann Surg Oncol.2009;16:2510–2515.[Context Link]38. An JY, Kim H, Cheong JH, et al..Microsatellite instability in sporadic gastric cancer: its prognostic role and guidance for 5-FU based chemotherapy after R0 resection.Int J Cancer.2012;131:505–511.[Context Link]gastric cancer; phenotype; genotype; KRAS; MSI00019606-201309000-0000100019574_2011_32_22_travis_classification_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1316_citationRF_FLOATING))|11065404||ovftdb|SL000195742011322211065404citation_FROM_JRF_ID_d7e1316_citationRF_FLOATING[Full Text]00019574-201132010-0000400019606-201309000-0000100019574_2011_32_22_travis_classification_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1316_citationRF_FLOATING))|11065405||ovftdb|SL000195742011322211065405citation_FROM_JRF_ID_d7e1316_citationRF_FLOATING[Medline Link]2150012100019606-201309000-0000100130550_2010_11_1317_damia_instability_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1369_citationRF_FLOATING))|11065213||ovftdb|SL00130550201011131711065213citation_FROM_JRF_ID_d7e1369_citationRF_FLOATING[CrossRef]10.2174%2F138945011100701131700019606-201309000-0000100130550_2010_11_1317_damia_instability_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1369_citationRF_FLOATING))|11065405||ovftdb|SL00130550201011131711065405citation_FROM_JRF_ID_d7e1369_citationRF_FLOATING[Medline Link]2084007400019606-201309000-0000100134338_2010_10_651_pino_application_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1443_citationRF_FLOATING))|11065404||ovftdb|SL0013433820101065111065404citation_FROM_JRF_ID_d7e1443_citationRF_FLOATING[Full Text]00134338-201007000-0002000019606-201309000-0000100134338_2010_10_651_pino_application_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1443_citationRF_FLOATING))|11065213||ovftdb|SL0013433820101065111065213citation_FROM_JRF_ID_d7e1443_citationRF_FLOATING[CrossRef]10.1586%2Ferm.10.4500019606-201309000-0000100134338_2010_10_651_pino_application_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1443_citationRF_FLOATING))|11065405||ovftdb|SL0013433820101065111065405citation_FROM_JRF_ID_d7e1443_citationRF_FLOATING[Medline Link]00019606-201309000-0000100000167_1965_64_31_lauren_classification_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1193_citationRF_FLOATING))|11065213||ovftdb|SL000001671965643111065213citation_FROM_JRF_ID_d7e1193_citationRF_FLOATING[CrossRef]10.1111%2Fapm.1965.64.1.3100019606-201309000-0000100000167_1965_64_31_lauren_classification_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1193_citationRF_FLOATING))|11065405||ovftdb|SL000001671965643111065405citation_FROM_JRF_ID_d7e1193_citationRF_FLOATING[Medline Link]1432067500019606-201309000-0000100128364_2010_16_4634_namikawa_clinicopathology_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1216_citationRF_FLOATING))|11065213||ovftdb|SL00128364201016463411065213citation_FROM_JRF_ID_d7e1216_citationRF_FLOATING[CrossRef]10.3748%2Fwjg.v16.i37.463400019606-201309000-0000100128364_2010_16_4634_namikawa_clinicopathology_|00019606-201309000-00001#xpointer(id(citation_FROM_JRF_ID_d7e1216_citationRF_FLOATING))|11065405||ovftdb|SL00128364201016463411065405citation_FROM_JRF_ID_d7e1216_citationRF_FLOATING[Medline Link]20872962Clinicopathologic Patient CharacteristicsPhenotype-Genotype Correlation of Gastric CancerPhenotypes of gastric cancer. The mucin phenotype was explored by immunostaining using antibodies directed against mucin 1 (A), mucin 2 (B), mucin 5 (C), mucin 6 (D), and CD10 (E). Note the membranous and cytoplasmic immunoreactions. Epstein-Barr virus–encoded RNA was detected using the EBER probe showing strong nuclear staining (F). E-cadherin (G) and β-catenin (H) showed membranous staining (G and H, left) or loss of expression (G and H, right). Lysozyme was found in tumor cells of intestinal (I, left) and diffuse (I, right) type gastric cancer. MLH1 (J) in microsatellite-stable gastric cancer showed nuclear expression. Note loss of nuclear expression of MLH1 (K) and PMS2 (L) in a MSI-H gastric cancer. Original magnifications ×400 (A-L).Prognostic Markers of Gastric Cancer (Univariate Analysis)Patient survival. Kaplan-Meier curves depicting patient survival according to Laurén phenotype (all cases, P=0.012), resection margin (all cases; P=0.001), mucin 2 (all cases, P=0.067), CD10 (unclassified GCs; P=0.003), lysozyme status (all cases, P=0.035), Helicobacter pylori infection (unclassified GCs, P=0.003), MSI status (all cases, P=0.036), and KRAS genotype (intestinal, P=0.098). Mucin 2 (all cases) and R-status (all cases) were independent prognosticators of patient survival in multivariate analyses. All cases indicates entire study population; GC, gastric cancer; MSI, microsatellite instability; unclassified, unclassified GCs according to Laurén.Explorative Multivariate Survival AnalysisPrognostic and Putative Predictive Biomarkers of Gastric Cancer for Personalized MedicineWarneke Viktoria S. MD; Behrens, Hans-Michael MSc; Haag, Jochen PhD; Balschun, Katharina MD; Böger, Christine MD; Becker, Thomas MD, PhD; Ebert, Matthias P.A. MD, PhD; Lordick, Florian MD, PhD; Röcken, Christoph MD, PhDOriginal ArticlesOriginal Articles322p 127-137

You can read the full text of this article if you:

Access through Ovid