Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A rapid, sensitive, reproducible and cost-effective method for mutation profiling of colon cancer and metastatic lymph nodes

A rapid, sensitive, reproducible and cost-effective method for mutation profiling of colon cancer... Background: An increasing number of studies show that genetic markers can aid in refining prognostic information and predicting the benefit from systemic therapy. Our goal was to develop a high throughput, cost- effective and simple methodology for the detection of clinically relevant hot spot mutations in colon cancer. Methods: The Maldi-Tof mass spectrometry platform and OncoCarta panel from Sequenom were used to profile 239 colon cancers and 39 metastatic lymph nodes from NSABP clinical trial C-07 utilizing routinely processed FFPET (formalin-fixed paraffin-embedded tissue). Results: Among the 238 common hot-spot cancer mutations in 19 genes interrogated by the OncoCarta panel, mutations were detected in 7 different genes at 26 different nucleotide positions in our colon cancer samples. Twenty-four assays that detected mutations in more than 1% of the samples were reconfigured into a new multiplexed panel, termed here as ColoCarta. Mutation profiling was repeated on 32 mutant samples using ColoCarta and the results were identical to results with OncoCarta, demonstrating that this methodology was reproducible. Further evidence demonstrating the validity of the data was the fact that the mutation frequencies of the most common colon cancer mutations were similar to the COSMIC (Catalog of Somatic Mutations in Cancer) database. The frequencies were 43.5% for KRAS, 20.1% for PIK3CA, and 12.1% for BRAF. In addition, infrequent mutations in NRAS, AKT1, ABL1, and MET were detected. Mutation profiling of metastatic lymph nodes and their corresponding primary tumors showed that they were 89.7% concordant. All mutations found in the lymph nodes were also found in the corresponding primary tumors, but in 4 cases a mutation was present in the primary tumor only. Conclusions: This study describes a high throughput technology that can be used to interrogate DNAs isolated from routinely processed FFPET and identifies the specific mutations that are common to colon cancer. The development of this technology and the ColoCarta panel may provide a mechanism for rapid screening of mutations in clinically relevant genes like KRAS, PIK3CA, and BRAF. Trial Registration: ClinicalTrials.gov: NSABP C-07: NCT00004931 Background prognosis [1-6]. However, these results remain contro- Recent evidence suggests that mutation profiling can versial because other studies have shown that mutations assist in the prognosis and prediction for colon cancer. in these genes are not prognostic. A large study, a meta KRAS, PIK3CA and BRAF mutations are frequent in analysis, of KRAS mutations, found that only tumors of the colon and have been associated with poor KRASG12V was a bad prognostic marker; other KRAS mutations were not associated with bad prognosis [2]. Evidence has also demonstrated that KRAS mutations * Correspondence: kay.pogue@nsabp.org are potential markers for prediction because tumors † Contributed equally Department of Pathology, National Surgical Adjuvant Breast and Bowel with KRAS mutations are significantly associated with Project (NSABP), 1307 Federal St, Pittsburgh, PA 15212, USA © 2010 Fumagalli et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Fumagalli et al. BMC Cancer 2010, 10:101 Page 2 of 14 http://www.biomedcentral.com/1471-2407/10/101 resistance to EGFR antibody based therapies [7-11]. intersect with the same pathways as that of KRAS, Publications have reported the same phenomenon with BRAF, and PIK3CA, such as AKT1, EGFR, HRAS, NRAS, BRAF and PIK3CA mutations, although these observa- MET and others. The frequency of KRAS, PIK3CA,and tions are still not well established [12,13]. The published BRAF mutations in the National Surgical Adjuvant study suggesting that BRAF mutant tumors were resis- Breast and Bowel Project (NSABP) trial C-07 were simi- tant to EGFR therapies was a small study [13]. Predic- lar to the frequencies for colon samples listed in the tive value of PIK3CA mutations remains controversial in COSMIC data base. This observation provides evidence that other publications have shown that these mutations that the mutation data obtained with the Sequenom have no predictive value [14,15]. These inconsistencies, platform is accurate. Our results also demonstrate that a together with two other factors, have limited the impact majority of colon cancer samples have aberrant PIK(3) - of mutation profiling for prognosis and prediction in RAS/RAF network; similar results have been seen pre- standard care of colon cancer. A large sample size is viously [6]. required to establish that a gene mutation has a signifi- Mutations in ABL1 and MET, not previously identified cant impact for patient prediction or prognosis. Another in colon cancer, were identified, and 13 other genes limitation is that until recently conducting such large were screened and found not to be mutated in hot spot studies with the standard sequencing technologies was locations. Furthermore, this study identified the most too time consuming and too expensive to be practical frequent colon cancer mutations from OncoCarta, pro- for clinical studies. viding the necessary information to reduce the number Moreover, while the high frequency of KRAS, BRAF of assays from 187 to 24, creating a smaller, more speci- and PIK3CA mutations in colon cancer is well docu- fic and economical panel, requiring less DNA, and thus mented, other potentially important mutations have not conserving precious clinical samples. been profiled with a large number of clinical samples. Whole genome sequencing of a small number of colon Methods samples demonstrated that somatic cancer mutations Clinical samples and histological evaluation consist of a few genes that occur frequently and many Samples used in this study were from NSABP clinical more mutations that occur very infrequently in many trial C-07. This trial enrolled patients between 02/2000 different genes [16,17]. Mutations in these infrequently and 11/2002 to compare oxaliplatin and bolus 5-FU/LV mutated genes could have a similar effect or synergize to bolus 5FU/LV alone for resected stage II and III with mutations in KRAS, PIK3CA, and BRAF. colon cancer [20]. Tissue samples were obtained at sur- Given these considerations, our goal was to find a gery before the patients had received any treatment and cost-effective and high throughput methodology that were routinely processed with FFPE. C-07 was approved would detect frequent and infrequent cancer mutations by an institutional review board, and informed consent genes in a large number of samples. Furthermore, it was was obtained from each subject. A pathologist categor- essential that the methodology would work with ized tumors into poor, moderate, well differentiated and degraded DNAs isolated from FFPET. signet cells carcinoma, according to the World Health The mass spectrometric SNP genotyping technology Organization (WHO) criteria. Samples were graded for based on the Sequenom platform provided an ideal mucinous character, based on the amount of mucin choice for mutation profiling to address these several retained within the tumor (1 = no mucin, 2 = < 50% criteria. It has been shown to work with small amounts mucinous volume/total tumor volume, 3 = >50% muci- of degraded DNAs (5 ng), and the high multiplexing nous volume/total tumor volume). Only grade 3 tumors capacity minimizes the use of irreplaceable clinical sam- were considered mucinous tumors carcinoma, which is ples. In addition, a variety of studies have demonstrated in accordance with WHO criteria. that the sensitivity of mass spectrometric methods exceeds that of traditional Sanger sequencing and is DNA isolation highly concordant with Sanger sequencing, Pyrosequen- DNA was isolated from FFPE tumor blocks collected cing, and allele-specific PCR [16,18,19]. Furthermore, from patients participating in NSABP clinical trial C-07. Sequenom has recently developed the OncoCarta Panel, FFPE tumor blocks were cut, and sections of the slide an oncogene panel that offers a rapid and parallel analy- containing the most tumor cells were defined by a sis of 238 simple and complex cancer mutations across pathologist and isolated by macrodissection. Genomic 19 genes. DNA was extracted from 4 five μm unstained sections. The OncoCarta panel includes assays for most colon After attempting several extraction procedures from a cancer mutations in the clinically relevant genes, BRAF variety of manufacturers (Machery Negel, Qiagen, (99%), KRAS (98%), and PIK3CA (78%), and in addition Ambion), it was determined that the Ambion Recover- contains assays for other cancer mutations in genes that All™ Total Nucleic Acid isolation kit (Applied Biosystem, Fumagalli et al. BMC Cancer 2010, 10:101 Page 3 of 14 http://www.biomedcentral.com/1471-2407/10/101 Foster City, CA) yielded the best DNA based on the NCI-H1299, NCI-H1395, UACC-893) were purchased quality, quantity, and the performance on the mass from American Type Culture Collection (ATCC, Mana- spectrometer (data not shown). The extraction was per- ssas, VA, US). Two cell lines, SKBR3 and MCF-7, were formed as recommended by the manufacturer, with two grown in culture and cell pellets were fixed in formalin, exceptions; the protease digestion was extended over- andembedded in paraffin and DNAs were isolated as night and the elution volume was increased to 150 ul to described for the clinical samples. maximize the total amount of DNA recovered. Addi- tional protease was added to samples incompletely Mass Spec Type Plex Technology and the digested after the overnight treatment. DNA was mea- OncoCarta Panel sured with fluorescence, using the Quant-iT ™ Pico- For mutation detection, the Sequenom platform and Green® dsDNA Assay Kit (Invitrogen, Carlsbad, CA) and the OncoCarta mutation panel were used and the pro- the InfiniteF200 fluorometer (Tecan, Mannedorf, tocol provided by Sequenom (San Diego, CA) was fol- Switzerland). lowed with minor modifications. A schematic of the As a positive control for known mutations and to test procedure is shown in Fig. 1. A Tecan Evo liquid the performance of the platform, annotated cell line handler was used to normalize the DNA samples and DNAs (A2058, HS578T, HL60, MCF7, MDAMB231, to set up the PCR reactions. The amount of DNA Figure 1 Methodology for mutation detection. Genomic DNA from the samples is amplified by PCR, resulting in copies of both mutant and wildtype alleles. Shrimp Alkaline Phosphatase removed excess nucleotides from the sample wells. Primer extension was performed using terminator nucleotides A, C, T, G, each with distinct masses. This linear amplification results in sequences proportional to the alleles that can be distinguished by mass spec (Maldi-Tof Separation). Fumagalli et al. BMC Cancer 2010, 10:101 Page 4 of 14 http://www.biomedcentral.com/1471-2407/10/101 added to the PCR was reduced to 15 ng or less. DNAs Results were amplified using the OncoCarta PCR primer pools, Mutations were detected in control DNAs from intact and unincorporated nucleotides were inactivated by shrimp FFPETsamples alkaline phosphatase (SAP), and a single base extension Previously described mutations in control cell lines were reaction was performed using extension primers that detected. BRAF_V600E, HRAS_G12D, NRAS_Q61L, hybridize immediately adjacent to the mutations and a PIK3CA_E545K, KRAS_G13D, NRAS_Q61K, custom mixture of nucleotides. Salts were removed by EGFR1_S125L, and PIK3CA_H1047R were detected in theaddition of acationexchangeresin.Multiplexed the appropriate cell lines (A2058, HS578T, HL60, reactions were spotted onto the SpectroChipII, and MCF7, MDAMB231, NCI-H1299, NCI-H1395, and mutations, if present, were resolved by MALDI-TOF UACC-893, respectively). The appropriate mutation was on the Compact Mass Spectrometer (Sequenom, San found in MCF-7 (PIK3CA_E545K) from both intact Diego, CA). DNA and DNA isolated from FFPET. DNAs from clini- The OncoCarta™ Panel v1.0 (Sequenom, San Diego, cal samples, control cell lines, and cell lines formalin- CA) consists of 24 pools of primer pairs and 24 pools of fixed, paraffin-embedded cell lines showed the same extension primers, and has the capacity to detect 238 rates of primer extension and performance on mass mutations in 19 genes, listed in Table 1. Each pool con- spectrometer. sists of 5-9 primer pairs in the PCR reaction. Two types The proportion of the mutated alleles in each cell line, of assays have been designed in the OncoCarta panel, as observed from the area under the mutant peak on referred to as simple and complex. The simple assays the spectra, ranged from 0.4-0.6, as expected for a pure arethose in whichasingle assayisabletodetectthe clonal population with a heterozygote mutation. Spectra amino acid changes at that codon. The complex assays for cell line UACC-893 had equal fractions of mutant arethose that requiremorethan one assaytoidentify and wt alleles (Fig. 2A). One exception to this distribu- codon changes or deletions and insertion, and thus are tion among cell lines was seen in A2058, which showed able to detect multiple different amino acid substitutions spectra consistent with 2 copies of the WT allele and or deletions. An example of a complex assay is KRAS_1 one mutant BRAF mutant allele (Fig. 2B). The 3 alleles and KRAS_2, which interrogate 2 different nucleotide of BRAF in A2058 are consistent with the observation positions within codon 12 and together identify all that there are 3 copies of chromosome 7 in this cell line codon 12 amino acid changes. Much more complex (COSMIC in the SNP Array Based LOH and Copy assays are included in OncoCarta, which interrogate Number Analysis data base) [21]. insertions and deletions within the EGFR gene. The Sequenom platform was sensitive and quantitative Data analysis Pilot studies demonstrated that the assays worked with Data analysis was performed using MassArray Typer Ana- as little as 1 ng of DNA (Fig. 3). The fraction of unex- lyzer software 4.0.4.20 (Sequenom), which facilitates visua- tended primer was .09 even when the input DNA was lization of data patterns as well as the raw spectra. between 1-3 ng, When concentrations of the amount of Mutations were identified in two different ways. Typer DNA was between 3-14 ng, the fraction of unextended automates the identification of mutants by comparing primer was similar, .07. Thus, the assays worked well ratios of the wild type peak to that of all suspected even when only 1 ng of DNA was used. mutants and generates an Onco Mutation report detailing In clinical samples with some assays it was possible to specific mutations and the ratios of wild type and muta- detect mutations that only represented 5% of the total 2 tion peaks. In addition, raw data was exported to Excel peak areas. The spectra in Fig. 4 show a small but clear and an in-house macro was used to duplicate the analysis. peak at the expected size for a PIK3CA 1047R mutation Theareaunder thepeaks allows forquantificationfor in a lymph node. We also were able to demonstrate the each allele, giving a direct evaluation of the proportion of sensitivity of the platform by performing a cell mixing mutated and wildtype (wt) allele in the sample [18]. experiment. Mutation analysis was done using MCF-7 All mutations from both the Onco mutation report cell line DNA alone or mixed with SKBR3 at various and the in-house Excel report were reviewed manually percentages. MCF-7 cells contain a PIK3CA mutation, by 3 investigators (DF, PGG, KPG). Manual review of and SKBR3 cells do not. Fig. 5 demonstrates that the mutations was necessary to identify “real” mutant peaks mutation was detectable even when the MCF-7 cells from salt peaks or other background peaks. Selected represented only 5 to 10% of the total DNA and only 5 reviewed mutations from the Onco Mutation Report to 2.5% of the alleles. This sensitivity is important for and from the in-house macro were compared and were mutation detection in clinical cancer samples, which concordant. usually contain some amount of normal tissue, which Fumagalli et al. BMC Cancer 2010, 10:101 Page 5 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 1 Mutations detected with OncoCarta ABL1-G250E EGFR-L747_E749del, A750P KIT-P585P ABL1-Q252H EGFR-E746_A750del KIT-D579del ABL1-Y253H EGFR-L747_E749del, A750P KIT-K642E ABL1-Y253F EGFR-L747_S752del, P753S KIT-D816V ABL1-E255K EGFR-E746_T751del, V ins KIT-D816H/D816Y ABL1-E255V EGFR-L747_S752del, Q ins KIT-V825A ABL1-D276G EGFR-L747_S752del, Q ins KIT-E839K ABL1-F311L EGFR-E746_T751del, S752D/SNP C2255T KIT-M552L ABL1-T315I EGFR-D770_N771>AGG/V769_D770insASV/V769_D770insASV KIT-Y568D ABL1-F317L EGFR-D770_N771insG KIT-F584S ABL1-M351T EGFR-L747_T750del, P ins KIT-P551_V555del ABL1-E355G EGFR-E746_A750del KIT-P551_V555del ABL1-F359V EGFR-E746_T751del, I ins KIT-Y553_Q556del ABL1-H396R EGFR-L747_T751del KIT-Y553_Q556del AKT1-rs11555435 EGFR-L747_T751del KRAS-G12V/A/D/C/S/R/F AKT1-rs11555431 EGFR-E746_A750del, V ins KRAS-G13C/S/V/D AKT1-rs11555432 EGFR-E746_A750del, V ins KRAS-A59T AKT1-rs12881616 EGFR-S752_I759del KRAS-Q61E/K/L/R/P/H AKT1-rs11555433 ERBB2-L755P MET-R970C AKT1-rs11555436 ERBB2-G776S/G776LC MET-T992I AKT1-rs34409589 ERBB2-G776VC MET-Y1230C AKT2-S302G ERBB2-G776VC/G776VC MET-Y1235D AKT2-R371H ERBB2-M774_A775insYVMA MET-M1250T BRAF-G464R ERBB2-A775_G776insYVMA NRAS-G12V/G12A/G12D BRAF-G464V/G464E ERBB2-P780_Y781insGSP NRAS-G12C/G12R/G12S BRAF-G466V/G466G/G466E ERBB2-P780_Y781insGSP NRAS-G13V/G13A/G13D BRAF-G466R ERBB2-S779_P780insVGS NRAS-G13C/G13R/G13S BRAF-F468C FGFR1-S125L NRAS-A18T BRAF-G469S/E/A/V/R FGFR1-P252T NRAS-Q61L/Q61R/Q61P BRAF-D594V| G FGFR3-R248C NRAS-Q61H BRAF-F595L FGFR3-S249C NRAS-Q61E/Q61K BRAF-G596R FGFR3-G370C PDGFRA-V561D BRAF-L597S/R/Q/V FGFR3-Y373C PDGFRA-T674I BRAF-T599I FGFR3-A391E PDGFRA-F808L BRAF-V600E/K/R/L FGFR3-K650Q/E PDGFRA-D846Y BRAF-K601N/E FGFR3-K650T/M PDGFRA-N870S CDK-R24C/H FLT3-I836del PDGFRA-D1071N EGFR-R108K FLT3_2 PDGFRA-D842_H845del EGFR-T263P FLT3_3 PDGFRA-I843_D846del EGFR-A289V FLT3-D835H/D835Y PDGFRA-S566_E571>K EGFR-G598V HRAS-G12V/D PDGFRA-I843_S847>T EGFR-E709K/E709H HRAS-G13C/R/S PDGFRA-D842V EGFR-E709A/E709G/E709V HRAS-G13V/D PIK3CA-R88Q EGFR-G719S/G719C HRAS-Q61H PIK3CA-N345K EGFR-G719A HRAS-Q61H/L/R/P/K PIK3CA-C420R EGFR-M766_A767insAI JAK2-V617F PIK3CA-P539R EGFR-S768I KIT-D52N PIK3CA-E542K EGFR-V769_D770insASV KIT-Y503_F504insAY PIK3CA-E545K EGFR-V769_D770insCV KIT-W557R/W557R/W557G PIK3CA-Q546K EGFR-D770_N771>AGG/V769_D770insASV/V769_D770insASV KIT-V559D/V559A/V559G PIK3CA-H701P EGFR-D770_N771insG KIT-V559I PIK3CA-H1047R/H1047L EGFR-N771_P772>SVDNR KIT-V560D/V560G PIK3CA-H1047Y Fumagalli et al. BMC Cancer 2010, 10:101 Page 6 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 1: Mutations detected with OncoCarta (Continued) EGFR-P772_H773insV KIT-K550_K558del PIK3CA-G1049R EGFR-H773>NPY KIT-K558_V560del PIK3CA-R38H EGFR-H773_V774insNPH/H773_V774insPH/H773_V774insH KIT-K558_E562del PIK3CA-C901F EGFR-V774_C775insHV KIT-V559del PIK3CA-M1043I/M1043I EGFR-T790 M KIT-V559_V560del RET-C634R EGFR-L858R KIT-V560del RET-C634W/Y EGFR-L861Q KIT-Y570_L576del RET-E632_L633del EGFR-L747_T750del, P ins/E746_A750del, T751A KIT-E561K RET-M918T EGFR-E746_T751del, I ins/S752_I759del KIT-L576P RET-A664D dilutes the number of tumor cells. This is of particular database [21]. The COSMIC frequencies seen in Table 2 concern when profiling lymph nodes, which may con- are based only on those mutations that are detectable tain a minority of tumor cells. with OncoCarta. OncoCarta assays interrogate 99%, 98%, and 78% of the known colon cancer mutations in BRAF, Frequencies of C-07 mutations in KRAS, NRAS, PIK3CA, KRAS,and PIK3CA, respectively, based on a large num- and BRAF detected with OncoCarta and the Sequenom ber of colon cancer samples that have been sequenced in platform were similar to previous reports BRAF (n = 4628), KRAS (n = 858) and PIK3CA (n = 247). In this preliminary assessment of the feasibility of using The OncoCarta panel found that the most frequent the Sequenom platform to do large-scale mutation profil- mutations in C-07 were KRAS (43.5%), PIK3CA (20.1%), ing of colon cancer samples isolated from FFPET, it was and BRAF (12.1%), which are similar to what is seen in essential to determine if our data yielded frequencies COSMIC. NRAS mutations, while infrequent, were typical of what has been seen previously. Table 2 shows detected in codons 12, 13 and 61 and represent a sizable the mutation frequencies obtained here and from the minority of the C0-7 samples (3.8%). These data suggest COSMIC (Catalog of Somatic Mutations in Cancer) Figure 2 Spectra for cell lines UACC-893 and A2058. The expected positions for the unexteneded primer (UEP), and the extension products (Mutant and WT) from assays PIK3CA_9 and BRAF_15 in cell lines UACC-893 and A2058, respectively, are indicated with red dashed lines. The proportion of peak areas and the specific base is also shown. Assays PIK3CA_9 and BRAF_15 detected mutations in PIK3CA at amino acid position 1047 and in BRAF at amino acid position 600, respectively. Other peaks included in these spectra as result of multiplexing but not part of the designated assays are indicated as grey dashed lines. Fumagalli et al. BMC Cancer 2010, 10:101 Page 7 of 14 http://www.biomedcentral.com/1471-2407/10/101 Figure 3 Fractional unextended primer versus input DNA. The range and the average for the percent of unextended primer for different amounts of input DNA into the PCR reactions are shown. The number of samples used in each category was 4 for 1-3 ng, 9 for 3-9 ng, 13 for 9-13 and 210 for 14 ng. Figure 4 Sensitive detection of mutations in clinical FFPE samples with the Sequenom platform. Small mutant but definitive peak illustrating a PIK3CA-1047R mutation in approximately 5% of the sample DNA is shown. Fumagalli et al. BMC Cancer 2010, 10:101 Page 8 of 14 http://www.biomedcentral.com/1471-2407/10/101 Figure 5 Quantification of the sensitivity with a cell line mixing experiment. Spectra of MCF-7 cells (mutant) alone or mixed with SKBR3 cells (WT) are shown. Percents are based on the ng amounts of DNA. This assay detects an E545K mutation in PIK3CA. that FFPET samples can be interrogated with the tech- Sequenom data was reproducible nology described here and yield accurate data. Most of the assays in the OncoCarta panel did not While most of the specific amino acid mutations mir- detect mutations or the frequency of mutations was ror what is seen on the COSMIC database, some unique very low (below 1%) in our colon cancer samples. colon cancer gene mutations were found, which include OncoCarta assays interrogate mutations in these 19 ABL1-F359V, AKT1-E17K, MET-R970C, and MET- genes listed in Table 1. To reduce the cost, time and T992I. Other amino acid changes that were not in the the amount of DNA required for profiling, only 24 COSMIC database were amino acid changes R88Q, assays, which detected mutations at a frequency of 1% H701P, and C420R in PIK3CA, BRAF-594V/G, and or greater in C-07, were selected, resorted in 6 pools KRAS-Q61R, and several in NRAS, including G12C, and included in a new panel, termed ColoCarta (Table G12D, G13R, G13V, Q61H and Q61K (Table 2). 3). Mutation profiles of 32 mutant samples with 41 mutations were repeated with the ColoCarta. The mutations detected by the 2 panels (OncoCarta and MET mutations were found in C0-7 and amplified in ColoCarta) were identical, demonstrating the reprodu- sometumors cibility of the methodology. MET mutations were found in 3.3% of C-07 samples. Interestingly, these mutations were not only unexpected Multiple mutation frequencies suggest an order to the in their appearance within the colon cancer population but also the frequency within the samples was unex- acquisition of different mutations pected. In four of the eight samples with MET muta- A majority of the tumors (64%) contained at least one tions, the mutant alleles were present at 58-70%, or more mutations in the following genes: BRAF, suggesting an amplification of the mutant allele or a loss KRAS, NRAS, MET,or PIK3CA, and 18% had 2 or of the wt gene (Fig. 6). Amplification may represent the more mutations. The most common double mutation best explanation, in that amplification of the MET geno- was in KRAS and PIK3CA,followedby PIK3CA and mic region, 7q31, has been observed in the Progenetix BRAF (Table 4). Most samples with PIK3CA mutations CGH Database in 23% of colorectal cancers [22] (80%) also had mutations in other genes, the most Fumagalli et al. BMC Cancer 2010, 10:101 Page 9 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 2 Frequency of colon cancer mutations † ‡ Mutation No. Mutated Samples Frequency in Primary Tumor* in COSMIC Multiple Mutations ABL1-F359V 1 0.40% 0/66 ABL1 Total 1 0.40% NF(0/66) 100% AKT1-E17K 1 0.40% 0/31 AKT1 Total 1 0.40% NF(0/31) 100% BRAF-D594V| G 1 0.40% NF (0/3179) BRAF-V600E 28 11.70% 14.60% BRAF Total 29 12.10% 14.70% 24% KRAS-G12A 2 0.80% 1.80% KRAS-G12C 10 4.20% 3.60% KRAS-G12D 40 16.70% 13.20% KRAS-G12R 3 1.30% 0.40% KRAS-G12S 3 1.30% 4.20% KRAS-G12V 20 8.40% 7.40% KRAS-G13D 23 9.60% 5.20% KRAS-A59T 1 0.40% 0.10% KRAS-Q61L 1 0.40% 0.20% KRAS-Q61R 1 0.40% NF (0/1927) KRAS Total 104 43.50% 36.10% 34% MET-R970C 2 0.80% NF (0/77) MET-T992I 6 2.50% NF (0/77) MET Total 8 3.30% 0% 50% NRAS-G12C 1 0.40% NF (0/46) NRAS-G12D 4 1.70% NF (0/46) NRAS-G13R 1 0.40% NF (0/46) NRAS-G13V 1 0.40% NF (0/46) NRAS-Q61H 1 0.40% NF (0/46) NRAS-Q61K 1 0.40% NF (0/46) NRAS Total 9 3.80% 2.2% 38% PIK3CA-R88Q 5 2.10% NF (0/171) PIK3CA-C420R 2 0.80% NF (0/171) PIK3CA-E542K 9 3.80% 4.10% PIK3CA-E545K 12 5.00% 4.10% PIK3CA-Q546K 4 1.70% 1.20% PIK3CA-H701P 1 0.40% NF (0/171) PIK3CA-H1047L 1 0.40% 1.80% PIK3CA-H1047R 14 5.90% 5.30% PIK3CA Total 48 20.10% 16.40% 80% *% of C-07 samples with this mutation. Data from COSMIC for colon adenocarcinoma limited to the same mutations interrogated with OncoCarta. The mutations listed are only the ones found in C-07. Some COSMIC amino acid changes are not shown here if they were not mutated in C-07. % of samples with a mutation in the gene shown and at least one other mutation in C-07 samples. COSMIC data are from large intestine, not specific to colon. frequent of which was KRAS; other mutated genes of colon cancer [23,24]. The multiple mutation fre- were BRAF, MET, NRAS,and a second PIK3CA muta- quencies for tumors with KRAS and PIK3CA or with tion (Table 2, last column). Tumors with MET and PIK3CA and BRAF were slightly higher and lower, NRAS mutations also have an unexpectedly high fre- respectively, than expected based on their individual quency of co-occurring mutations, which suggests that frequencies (Table 4). Conversely, the expected double they occur as a second mutation and perhaps later in mutation frequency of BRAF and KRAS would be 5.1%, the etiology of the tumor. Many tumors contain only a based on our data, but this combination was not KRAS or BRAF mutation, which is consistent with pre- found, also in agreement with previous reports [24] vious reports finding these mutations in earlier stages (Table 4). Fumagalli et al. BMC Cancer 2010, 10:101 Page 10 of 14 http://www.biomedcentral.com/1471-2407/10/101 Primary tumors with KRAS and PIK3CA mutations vary with respect to the frequency of these mutant alleles In the samples with co-occurring mutations, the ratios of KRAS mutation ratio (KRAS mutation peak area/total peak area) to the PIK3CA mutation ratio (PIK3CA mutation peak area/total peak area) was determined. Twenty-two out of 31 samples (71%) had KRAS/PIK3CA ratios above 1.25 (Table 5). PIK3CA mutations were more prevalent in only 2 out of 31 samples. These dif- ferences demonstrate that in a majority of primary tumors with double mutations in KRAS and PIK3CA, the KRAS mutations are more prevalent than the PIK3CA. This unequal distribution of mutant alleles within a tumor may be due to the fact that a majority of thetumor cellshaveonlythe KRAS mutation, and cells with a PIK3CA mutation are in the minority, or it could be due to copy number variations in the KRAS and PIK3CA loci. BRAF mutations were correlated with poorly differentiated tumors and with mucinous tumors Figure 6 MET mutation is amplified. The proportion and position The frequency of mutations for KRAS, PIK3CA,and (blue dashed lines) of mutant and wt alleles are shown. The MET_1 BRAF were tested for correlation to the degree of differ- assay detects R970C mutations in MET. entiation and to the prevalence of mucin in the tumor. BRAF mutations were found in 26.2% of the poorly dif- ferentiated tumors and in 8.2% of the moderate and well differentiated. These frequencies were significantly dif- Table 3 ColoCarta panel ferent by Chi square test (p value = 0.001). BRAF muta- Sequenom’s Assay Name Amino Acid Change tions were also associated with mucinous tumors: BRAF BRAF_15 &16 V600E/K/R/L mutations occurred in 28% of grade 3 mucinous tumors BRAF_9 BRAF-D594V (>50% mucinous tumor cells) but in only 9.4% of the HRAS_6* HRAS-Q61L non-mucinous tumors (grade 1 and 2). This was signifi- KRAS_1 & 2 G12V/A/D/C/S/R/F cant by the Chi square test at p value = 0.006. Similar KRAS_4 KRAS-G13D data have been reported previously [25,26]. KRAS and KRAS_5 KRAS-A59T PIK3CA mutations did not correlate with either the KRAS_7 KRAS-Q61L degree of differentiation or with prevalence of mucinous KRAS_8 Q61H/Q61H cells. MET_1 MET-R970C MET_2 MET-T992I Mutation profiling demonstrated a majority of primary NRAS_1 NRAS-G12V and lymph node samples were concordant but NRAS_2 NRAS-G12C differences were detected NRAS_3 NRAS-G13V Lymph node metastases were not routinely collected in NRAS_4 NRAS-G13C C-07 but as a pilot study to determine the feasibility of NRAS_7 NRAS-Q61H using lymph nodes for mutation profiling was conducted. NRAS_8 NRAS-Q61E We isolated DNA from 39 lymph nodes containing PIK3CA_1 PIK3CA-R88Q tumor cells and their corresponding primary tumors. PIK3CA_3 PIK3CA-C420R These primary and lymph nodes samples were profiled PIK3CA_5 PIK3CA-E542K with the entire OncoCarta panel. The majority of lymph PIK3CA_6 PIK3CA-E545K nodes and their corresponding primary tumors (89.7%) PIK3CA_7 PIK3CA-Q546K were concordant. A total of 26 mutations were detected PIK3CA_8 PIK3CA-H701P in lymph nodes, including KRAS, BRAF, PIK3CA,and PIK3CA_9 PIK3CA-H1047R NRAS. Thirty-five out of 39 lymph nodes had identical *HRAS_6 was included in panel but occurred in < 0.1% of samples. mutation profiles, but in 4 cases mutations in the primary Fumagalli et al. BMC Cancer 2010, 10:101 Page 11 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 4 Single and double mutations in C-07 Double Mutation Frequencies KRAS PIK3CA All other Single Actual Expected Actual Expected Actual Expected KRAS 43.70% NA NA 10.40% 8.70% 14.60% 7.21% PIK3CA 20.10% 10.40% 8.70% NA NA 15.50% 8.06% BRAF 11.80% 0 5.10% 1.80% 2.40% 2.50% 5.71% MET 3.30% 1.67% 1.44% 0 0.66% 1.67% 2% NRAS 3.80% 0 1.66% 0.42% 0.40% 1.30% 2.14% All Mutations 60.20% primary tumor in 3 out of 4 samples. In sample 0940, Table 5 KRAS/PIK3CA ratio mutation frequencies within the KRAS/PIK3CA mutation decreased by almost 1/2 in primary tumors the lymph node tumor compared to the primary. Thus, No of Samples KRAS/PIK3CA in these samples there is either a loss of KRAS muta- 22 1.25-3.22 tions or an accumulation of PIK3CA mutations, suggest- 7 0.93-1.13 ing that PIK3CA mutations may impart a selective 2 0.81-.42 advantage in the lymph node. Average 1.67 In contrast, two other samples have a less frequent Median 1.6 occurrence of their PIK3CA mutation in the lymph node than in the primary tumor. In sample 2244, the PIK3CA tumors were not found in the corresponding lymph mutation was undetectable in the lymph node (Table 6). nodes (BRAF [2], PIK3CA [1] and KRAS [1]). In fact, if there was a selection for both mutations in the lymph node, then the PIK3CA mutation frequency would Mutation profiles demonstrate that tumor cell have been thesameasthat ofthe KRAS mutation (0.15). populations may be different in lymph nodes and in the On the other hand, if the PIK3CA/KRAS ratio were the primary tumors same in the primary and lymph node tumor, then the Peak area evaluation of tumors that had 2 mutations PIK3CA mutation frequency would have been .08, which and for which a metastatic lymph node was available is still detectable with this technology (Fig. 3). Thus, in demonstrated differences between the primary and sample 2244 there were fewer PIK3CA mutant alleles in lymph node samples. Table 6 details the frequency of the lymph node than in the primary tumor. In sample mutant and wt alleles based on the peak areas for 5 1837, mutations in both BRAF and PIK3CA were such samples. detected and the BRAF/PIK3CA ratio was 1.67, but KRAS to PIK3CA ratios demonstrated that there were increased to 4.4 in the metastatic lymph node. more KRAS mutations than PIK3CA mutations in 4 of 4 primary samples, and in 3 of the 4 lymph node samples. Discussion However, it is also notable that the ratio of KRAS/ The Sequenom platform provides a superior technology PIK3A was lower in the lymph node compared to their for the screening of many hot spot mutations in cancer Table 6 Allele frequencies for primary tumors with two mutations and metastatic lymph nodes Sample Mutations Mutant Allele Frequency Mutation Ratios Primary Lymph Node Primary M1/M2 Lymph node M1/M2 C07-0388 M1. KRAS-G12D 0.45 0.44 1.88 1.63 M2. PIK3CA-H1047R 0.24 0.27 C07-0717 M1. KRAS-G13D 0.08 0.09 1.14 0.90 M2. PIK3CA-H1047R 0.07 0.1 C07-0940 M1. KRAS-G12D 0.42 0.34 1.91 1.10 M2. PIK3CA-E542K 0.22 0.31 C07-2244 M1. KRAS-G12C 0.37 0.15 1.76 _ M2. PIK3CA-H1047R 0.21 0 C07-1837 M1. BRAF-V600E 0.2 0.22 1.67 4.40 M2. PIK3CA-H1047R 0.12 0.05 Fumagalli et al. BMC Cancer 2010, 10:101 Page 12 of 14 http://www.biomedcentral.com/1471-2407/10/101 samples. Sanger sequencing would require amplification reported for KRAS, PIK3CA and BRAF. For example, of at least 60 different fragments per sample, and many AKT1 and NRAS are molecules that are downstream reactions would require optimization, thus adding con- mediators of the EGFR signaling pathway, and muta- siderable time and expense. Multiplexing and the use of tions in these genes are likely to affect the response to the OncoCarta panel allowed us to skip this time con- drugs that target EGFR. suming step. Thus, conservatively, Sanger sequencing Mutations in ABL, AKT1, and MET were detected here would be 40 times more expensive, and require at least but were not listed in COSMIC, probably due to the 2 times more DNA. Other sequencing technologies, small number of samples analyzed. The AKT1-E17K which employ differential melting of mutant and wt mutation was initially identified as a SNP, rs34409589, sequences, such as HRMA, still require that the PCR but in a recent publication it was found to be a somatic product be sequenced. This would add significant cost mutation and was found in 3 of 51 colon cancers [27]. and time to the procedure because 60% of the colon The frequency of these mutations in this small study (51 samples contained one or more mutations. In addition samples) was 6% and is much greater than in the C0-7 the Sequenom platform is more sensitive than Sanger samples (0.4%). This difference in frequencies may be sequencing in that it was able to detect mutations that because the Carpten et al [27] samples were from more represented only 5% of the DNA. Pyrosequencing repre- advanced stages than those from the C-07 trial. More- sented a potential alternative to the Sequenom platform, over, they selected large tumors (>100 mg) and contain- but in our hands assays needed to be optimized, and the ing more than 60% tumor cells. No such selection was lack of multiplexing made the procedure more time done for our study, and samples were from stages II and consuming and demanded more DNA. The Sequenom III exclusively. The significance of ABL1 and AKT1 muta- methodology also focuses on only those nucleotides that tions for patient prediction and prognosis in our study is are known to be cancer mutations and thus makes questionable given that they each were found in only in review of the sequence information considerably faster one sample and represented only 0.4% of the cases. than Sanger. Next-Generation sequencing was cost pro- To ourknowledge,thisisthe firstreportof MET hibitive and has not been shown to work with DNAs mutations in the primary colon cancer, but a different isolated from FFPET. Thus, the Sequenom platform and MET mutation (N1118Y) was found in a lung metastasis the OncoCarta Panel provided the simplest, most rapid, of the large intestine [28]. The MET mutations, R970C sensitive and cost-effective method for detecting hot and T992I, were detected in 8 out of 239 C-07 colon spot cancer mutations in degraded DNAs isolated from cancers. These mutations correspond to MET-R988C archival and routinely processed FFPET. The ColoCarta and MET-T1010I, respectively, in the long form of MET panel provides a more specific panel for colon cancer which is the isoform referred to in the COSMIC data- mutation detection and greatly reduces the amount of base [29]. The R970C and T992I mutations are located DNA needed for mutation profiling. in the juxtamembrane segment of the protein and were The frequencies and specific amino acid mutations detected in lung carcinoma [30]. These mutations, when detected here were similar to the COSMIC database and introduced into a lung cell line, increased focus forma- other publications [6]. The small variation in frequency tion, formation of colonies in soft agar, cell motility, and between our data and other reports may be attributed migration. These mutations also resulted in constitutive to differences in the stage of the samples analyzed, the tyrosine phosphorylation on several cellular proteins number of samples considered, and the sensitivity of the including paxillin at key tyrosine residues and may technology [18]. These observations, combined with the account for the increased motility of cells with this perfect match that we obtained between the expected mutation. Another critical amino acid in this location is and the detected mutations in our control cell lines, a Ser 985, which, when phosphorylated, has been found both fresh and FFPE, and the fact that mutations to diminish MET signaling [31]. If phosphorylation at detected with OncoCarta and ColoCarta were identical, Thr residue 992 (1010) reduces signaling, then the suggest that the technology is reliable and reproducible R992I mutation would inhibit this negative feedback and in DNAs isolated from FFPE samples. may result in constitutive signaling [30]. In our study, the majority of tumors (60.3%) had one If MET mutations confer an alternative activated sig- or more mutations in KRAS, PIK3CA,and BRAF. Muta- naling pathway, then these mutations could also confer tions in these genes are likely to perturb many different resistance to anti-EGFR-based therapies or provide a and overlapping signaling pathways, including PI3K/ new target for directed therapies. Therapeutic drugs AKT, ERK/MAPK, SAPK/JNK, NFKb,and others. We have been developed to specifically target MET,includ- were also able to detect other less frequent mutations ing small molecule kinase inhibitors, anti-MET mono- that are likely to perturb the same pathways and these clonal antibodies, and inhibitors of HGF, the MET may cause resistance to EGFR-targeted therapies, as ligand. Invitro assays have demonstrated that a number Fumagalli et al. BMC Cancer 2010, 10:101 Page 13 of 14 http://www.biomedcentral.com/1471-2407/10/101 of MET targeted therapies were able to prevent MET correspond to SNPs rs34589476 and rs56391007 in the signaling, decrease cell viability, and limit cell motility NCBI SNP data base, respectively. The frequency for and migration in vitro[32].The smallmoleculeARQ these SNPs is unknown so whether these nucleotide 197, a kinase inhibitor, has entered phase II clinical changes are associated with cancer is unknown. trials so may represent a possible therapeutic strategy for some colon tumors. Acknowledgements To our knowledge, this is also one of the most The authors would like to thank Melanie Finnigan, Bill Hiller, and Theresa exhaustive analyses of mutation profiling of metastatic Oeler for help in cutting and cataloging slides, and Hema Liyanage from Sequenom for replexing the assays. This study was supported by Public lymph nodes and their corresponding primary colon Health Service grants U10-CA-37377, U10-CA-69974, U10-CA-12027, and U10- tumors. Our analysis showed that a majority of samples CA-69651 from the National Cancer Institute, National Institutes of Health, were concordant (89.7%) but in a few samples mutations and Department of Health and Human Services. This project is funded, in part, under a grant with the Pennsylvania Department of Health. The were detected only in the primary tumor and not in the Department specifically disclaims responsibility for any analyses, metastatic lymph node. Also in samples with 2 co- interpretations or conclusions. The authors retain the right to provide a copy occurring mutations, the ratio of the double mutations of the final manuscript to the NIH upon acceptance for journal publication, for public archiving in PubMed Central as soon as possible but no later than varied in primary and lymph node tumors. Discordance 12 months after publication by the journal. in the genetic profile between primary tumors and the metastatic lymph nodes has been observed [33]. Such Authors’ contributions DF and PGG designed and carried out the experiments and participated in data may indicate that tumor cell migration selects dif- the drafting of the manuscript. YT graded tumors with regard to the degree ferent cell populations from the one in the primary of differentiation, mucin content, and defined tumor regions, S-IK carried out tumor. However, it is also possible that these mutational experiments. H-JC defined the tumor regions; SP participated in the coordination of the study; KLP-G designed and coordinated the study and differences between the lymph node and the primary drafted the manuscript. All authors have given final approval of the version tumor are a result of tumor heterogeneity. to be published. Another interesting observation in our study was that Competing interests BRAF mutations were significantly correlated with The authors declare that they have no competing interests. poorly differentiated tumors and the prevalence of mucin; similar observations have been reported [25,26]. Received: 12 August 2009 Accepted: 16 March 2010 Published: 16 March 2010 These characteristics are both associated with a worse prognosis and are consistent with other reports associat- References ing BRAF mutations with a bad prognosis [3]. However, 1. Ogino S, Nosho K, Kirkner GJ, Shima K, Irahara N, Kure S, Chan AT, in our study we found that there were 2 metastatic Engelman JA, Kraft P, Cantley LC, et al: PIK3CA Mutation Is Associated With Poor Prognosis Among Patients With Curatively Resected Colon lymph nodes that did not maintain the BRAF mutation Cancer. J Clin Oncol 2009. present in the corresponding primary tumor, suggesting 2. Andreyev HJ, Norman AR, Cunningham D, Oates J, Dix BR, Iacopetta BJ, that BRAF mutations are not essential for metastatic Young J, Walsh T, Ward R, Hawkins N, et al: Kirsten ras mutations in patients with colorectal cancer: the ‘RASCAL II’ study. Br J Cancer 2001, spread to the lymph node in all tumors. Clearly, addi- 85(5):692-696. tional studies would be required to understand these 3. Samowitz WS, Sweeney C, Herrick J, Albertsen H, Levin TR, Murtaugh MA, apparent inconsistencies; additional lymph node samples Wolff RK, Slattery ML: Poor Survival Associated with the BRAF V600E Mutation in Microsatellite-Stable Colon Cancers. Cancer Res 2005, are not currently available but could be the subject of 65(14):6063-6069. further studies when samples become available [3]. 4. Ogino S, Nosho K, Kirkner GJ, Kawasaki T, Meyerhardt JA, Loda M, Giovannucci EL, Fuchs CS: CpG island methylator phenotype, microsatellite instability, BRAF mutation and clinical outcome in colon Conclusions cancer. Gut 2009, 58(1):90-96. The Sequenom platform provided a superior technology 5. Conlin A, Smith G, Carey FA, Wolf CR, Steele RJC: The prognostic for the screening of 238 common hot spot cancer muta- significance of K-ras, p53, and APC mutations in colorectal carcinoma. Gut 2005, 54(9):1283-1286. tions in 19 genes. The frequent occurrence of KRAS, 6. Barault L, Veyrie N, Jooste V, Lecorre D, Chapusot C, Ferraz JM, Lievre A, PIK3CA,and BRAF was confirmed, and mutations not Cortet M, Bouvier AM, Rat P, et al: Mutations in the RAS-MAPK, PI(3)K detected before in colon cancer were found in MET and (phosphatidylinositol-3-OH kinase) signaling network correlate with poor survival in a population-based series of colon cancers. Int J Cancer 2008, ABL1. Twenty-five assays from the OncoCarta were 122(10):2255-2259. replexed to form a new panel, termed ColoCarta, which 7. Khambata-Ford S, Garrett CR, Meropol NJ, Basik M, Harbison CT, Wu S, will be used to screen an additional 800 tumors from Wong TW, Huang X, Takimoto CH, Godwin AK, et al: Expression of epiregulin and amphiregulin and K-ras mutation status predict disease NSABP clinical trial C0-7 with the purpose of identify- control in metastatic colorectal cancer patients treated with cetuximab. ing prognostic or predictive markers for stage II and III J Clin Oncol 2007, 25(22):3230-3237. colon cancer. 8. Linardou H, Dahabreh IJ, Kanaloupiti D, Siannis F, Bafaloukos D, Kosmidis P, Papadimitriou CA, Murray S: Assessment of somatic k-RAS mutations as a Note Added in Proof: Although MET-R988C and mechanism associated with resistance to EGFR-targeted agents: a MET-T1010I mutations were listed in COSMIC as systematic review and meta-analysis of studies in advanced non-small- somatic cancer mutations, these nucleotide changes Fumagalli et al. BMC Cancer 2010, 10:101 Page 14 of 14 http://www.biomedcentral.com/1471-2407/10/101 cell lung cancer and metastatic colorectal cancer. The Lancet Oncology molecular features of colorectal cancer independently of microsatellite 2008, 9(10):962-972. instability status. Mol Cancer 2006, 5:2. 9. Lievre A, Bachet JB, Boige V, Cayre A, Le Corre D, Buc E, Ychou M, 27. Carpten JD, Faber AL, Horn C, Donoho GP, Briggs SL, Robbins CM, Bouche O, Landi B, Louvet C, et al: KRAS mutations as an independent Hostetter G, Boguslawski S, Moses TY, Savage S, et al: A transforming prognostic factor in patients with advanced colorectal cancer treated mutation in the pleckstrin homology domain of AKT1 in cancer. Nature with cetuximab. J Clin Oncol 2008, 26(3):374-379. 2007, 448(7152):439-444. 10. Amado RG, Wolf M, Peeters M, Van Cutsem E, Siena S, Freeman DJ, Juan T, 28. Lorenzato A, Olivero M, Patane S, Rosso E, Oliaro A, Comoglio PM, Di Sikorski R, Suggs S, Radinsky R, et al: Wild-type KRAS is required for Renzo MF: Novel Somatic Mutations of the MET Oncogene in Human panitumumab efficacy in patients with metastatic colorectal cancer. J Carcinoma Metastases Activating Cell Motility and Invasion. Cancer Res Clin Oncol 2008, 26(10):1626-1634. 2002, 62(23):7025-7030. 11. Baselga J, Rosen N: Determinants of RASistance to anti-epidermal growth 29. Loriaux MM, Levine RL, Tyner JW, Frohling S, Scholl C, Stoffregen EP, factor receptor agents. J Clin Oncol 2008, 26(10):1582-1584. Wernig G, Erickson H, Eide CA, Berger R, et al: High-throughput sequence 12. Sartore-Bianchi A, Martini M, Molinari F, Veronese S, Nichelatti M, Artale S, analysis of the tyrosine kinome in acute myeloid leukemia. Blood 2008, Di Nicolantonio F, Saletti P, De Dosso S, Mazzucchelli L, et al: PIK3CA 111(9):4788-4796. Mutations in Colorectal Cancer Are Associated with Clinical Resistance 30. Ma PC, Kijima T, Maulik G, Fox EA, Sattler M, Griffin JD, Johnson BE, Salgia R: to EGFR-Targeted Monoclonal Antibodies. Cancer Res 2009, c-MET Mutational Analysis in Small Cell Lung Cancer: Novel 69(5):1851-1857. Juxtamembrane Domain Mutations Regulating Cytoskeletal Functions. 13. Di Nicolantonio F, Martini M, Molinari F, Sartore-Bianchi A, Arena S, Saletti P, Cancer Res 2003, 63(19):6272-6281. De Dosso S, Mazzucchelli L, Frattini M, Siena S, et al: Wild-type BRAF is 31. Hashigasako A, Machide M, Nakamura T, Matsumoto K, Nakamura T: Bi- required for response to panitumumab or cetuximab in metastatic directional Regulation of Ser-985 Phosphorylation of c-Met via Protein colorectal cancer. J Clin Oncol 2008, 26(35):5705-5712. Kinase C and Protein Phosphatase 2A Involves c-Met Activation and 14. Lambrechts D, De Roock W, Prenen H, De Schutter J, Jacobs B, Biesmans B, Cellular Responsiveness to Hepatocyte Growth Factor. J Biol Chem 2004, Claes B, De Hertogh G, Van Cutsem E, Tejpar S: The role of KRAS, BRAF, 279(25):26445-26452. NRAS, and PIK3CA mutations as markers of resistance to cetuximab in 32. Seiwert TY, Jagadeeswaran R, Faoro L, Janamanchi V, Nallasura V, El chemorefractory metastatic colorectal cancer. J Clin Oncol (Meeting Dinali M, Yala S, Kanteti R, Cohen EE, Lingen MW, et al: The MET receptor Abstracts) 2009, 27(15S):4020. tyrosine kinase is a potential novel therapeutic target for head and neck 15. Prenen H, De Schutter J, Jacobs B, De Roock W, Biesmans B, Claes B, squamous cell carcinoma. Cancer Res 2009, 69(7):3021-3031. Lambrechts D, Van Cutsem E, Tejpar S: PIK3CA Mutations Are Not a Major 33. Gamblin TC, Finkelstein SD, Upsal N, Kaye JD, Blumberg D: Microdissection- Determinant of Resistance to the Epidermal Growth Factor Receptor based allelotyping: a novel technique to determine the temporal Inhibitor Cetuximab in Metastatic Colorectal Cancer. Clin Cancer Res 2009. sequence and biological aggressiveness of colorectal cancer. Am Surg 16. Thomas RK, Baker AC, Debiasi RM, Winckler W, Laframboise T, Lin WM, 2006, 72(5):445-453. Wang M, Feng W, Zander T, MacConaill L, et al: High-throughput oncogene mutation profiling in human cancer. Nat Genet 2007, Pre-publication history 39(3):347-351. The pre-publication history for this paper can be accessed here: 17. Wood LD, Parsons DW, Jones S, Lin J, Sjoblom T, Leary RJ, Shen D, [http://www.biomedcentral.com/1471-2407/10/101/prepub] Boca SM, Barber T, Ptak J, et al: The genomic landscapes of human breast doi:10.1186/1471-2407-10-101 and colorectal cancers. Science 2007, 318(5853):1108-1113. Cite this article as: Fumagalli et al.: A rapid, sensitive, reproducible and 18. Vivante A, Amariglio N, Koren-Michowitz M, Ashur-Fabian O, Nagler A, cost-effective method for mutation profiling of colon cancer and Rechavi G, Cohen Y: High-throughput, sensitive and quantitative assay metastatic lymph nodes. BMC Cancer 2010 10:101. for the detection of BCR-ABL kinase domain mutations. Leukemia 2007, 21(6):1318-1321. 19. van Puijenbroek M, Dierssen JW, Stanssens P, van Eijk R, Cleton-Jansen AM, van Wezel T, Morreau H: Mass spectrometry-based loss of heterozygosity analysis of single-nucleotide polymorphism loci in paraffin embedded tumors using the MassEXTEND assay: single-nucleotide polymorphism loss of heterozygosity analysis of the protein tyrosine phosphatase receptor type J in familial colorectal cancer. J Mol Diagn 2005, 7(5):623-630. 20. Kuebler JP, Wieand HS, O’Connell MJ, Smith RE, Colangelo LH, Yothers G, Petrelli NJ, Findlay MP, Seay TE, Atkins JN, et al: Oxaliplatin combined with weekly bolus fluorouracil and leucovorin as surgical adjuvant chemotherapy for stage II and III colon cancer: results from NSABP C-07. J Clin Oncol 2007, 25(16):2198-2204. 21. Forbes SA, Bhamra G, Bamford S, Dawson E, Kok C, Clements J, Menzies A, Teague JW, Futreal PA, Stratton MR: The Catalogue of Somatic Mutations in Cancer (COSMIC). Curr Protoc Hum Genet 2008, Chapter 10(Unit 10):11. 22. Baudis M, Cleary ML: Progenetix.net: an online repository for molecular cytogenetic aberration data. Bioinformatics 2001, 17(12):1228-1229. 23. Pratilas C, Solit D: Therapeutic strategies for targeting BRAF in human Submit your next manuscript to BioMed Central cancer. Rev Recent Clin Trials 2007, 2(2):121-134. and take full advantage of: 24. Yuen ST, Davies H, Chan TL, Ho JW, Bignell GR, Cox C, Stephens P, Edkins S, Tsui WW, Chan AS, et al: Similarity of the phenotypic patterns associated • Convenient online submission with BRAF and KRAS mutations in colorectal neoplasia. Cancer Res 2002, 62(22):6451-6455. • Thorough peer review 25. Tanaka H, Deng G, Matsuzaki K, Kakar S, Kim GE, Miura S, Sleisenger MH, • No space constraints or color figure charges Kim YS: BRAF mutation, CpG island methylator phenotype and • Immediate publication on acceptance microsatellite instability occur more frequently and concordantly in mucinous than non-mucinous colorectal cancer. Int J Cancer 2006, • Inclusion in PubMed, CAS, Scopus and Google Scholar 118(11):2765-2771. • Research which is freely available for redistribution 26. Li WQ, Kawakami K, Ruszkiewicz A, Bennett G, Moore J, Iacopetta B: BRAF mutations are associated with distinctive clinical, pathological and Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Cancer Springer Journals

A rapid, sensitive, reproducible and cost-effective method for mutation profiling of colon cancer and metastatic lymph nodes

Loading next page...
 
/lp/springer-journals/a-rapid-sensitive-reproducible-and-cost-effective-method-for-mutation-UAZjbeoSCc

References (69)

Publisher
Springer Journals
Copyright
Copyright © 2010 by Fumagalli et al; licensee BioMed Central Ltd.
Subject
Biomedicine; Cancer Research; Oncology; Surgical Oncology; Health Promotion and Disease Prevention; Biomedicine general; Medicine/Public Health, general
eISSN
1471-2407
DOI
10.1186/1471-2407-10-101
pmid
20233444
Publisher site
See Article on Publisher Site

Abstract

Background: An increasing number of studies show that genetic markers can aid in refining prognostic information and predicting the benefit from systemic therapy. Our goal was to develop a high throughput, cost- effective and simple methodology for the detection of clinically relevant hot spot mutations in colon cancer. Methods: The Maldi-Tof mass spectrometry platform and OncoCarta panel from Sequenom were used to profile 239 colon cancers and 39 metastatic lymph nodes from NSABP clinical trial C-07 utilizing routinely processed FFPET (formalin-fixed paraffin-embedded tissue). Results: Among the 238 common hot-spot cancer mutations in 19 genes interrogated by the OncoCarta panel, mutations were detected in 7 different genes at 26 different nucleotide positions in our colon cancer samples. Twenty-four assays that detected mutations in more than 1% of the samples were reconfigured into a new multiplexed panel, termed here as ColoCarta. Mutation profiling was repeated on 32 mutant samples using ColoCarta and the results were identical to results with OncoCarta, demonstrating that this methodology was reproducible. Further evidence demonstrating the validity of the data was the fact that the mutation frequencies of the most common colon cancer mutations were similar to the COSMIC (Catalog of Somatic Mutations in Cancer) database. The frequencies were 43.5% for KRAS, 20.1% for PIK3CA, and 12.1% for BRAF. In addition, infrequent mutations in NRAS, AKT1, ABL1, and MET were detected. Mutation profiling of metastatic lymph nodes and their corresponding primary tumors showed that they were 89.7% concordant. All mutations found in the lymph nodes were also found in the corresponding primary tumors, but in 4 cases a mutation was present in the primary tumor only. Conclusions: This study describes a high throughput technology that can be used to interrogate DNAs isolated from routinely processed FFPET and identifies the specific mutations that are common to colon cancer. The development of this technology and the ColoCarta panel may provide a mechanism for rapid screening of mutations in clinically relevant genes like KRAS, PIK3CA, and BRAF. Trial Registration: ClinicalTrials.gov: NSABP C-07: NCT00004931 Background prognosis [1-6]. However, these results remain contro- Recent evidence suggests that mutation profiling can versial because other studies have shown that mutations assist in the prognosis and prediction for colon cancer. in these genes are not prognostic. A large study, a meta KRAS, PIK3CA and BRAF mutations are frequent in analysis, of KRAS mutations, found that only tumors of the colon and have been associated with poor KRASG12V was a bad prognostic marker; other KRAS mutations were not associated with bad prognosis [2]. Evidence has also demonstrated that KRAS mutations * Correspondence: kay.pogue@nsabp.org are potential markers for prediction because tumors † Contributed equally Department of Pathology, National Surgical Adjuvant Breast and Bowel with KRAS mutations are significantly associated with Project (NSABP), 1307 Federal St, Pittsburgh, PA 15212, USA © 2010 Fumagalli et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Fumagalli et al. BMC Cancer 2010, 10:101 Page 2 of 14 http://www.biomedcentral.com/1471-2407/10/101 resistance to EGFR antibody based therapies [7-11]. intersect with the same pathways as that of KRAS, Publications have reported the same phenomenon with BRAF, and PIK3CA, such as AKT1, EGFR, HRAS, NRAS, BRAF and PIK3CA mutations, although these observa- MET and others. The frequency of KRAS, PIK3CA,and tions are still not well established [12,13]. The published BRAF mutations in the National Surgical Adjuvant study suggesting that BRAF mutant tumors were resis- Breast and Bowel Project (NSABP) trial C-07 were simi- tant to EGFR therapies was a small study [13]. Predic- lar to the frequencies for colon samples listed in the tive value of PIK3CA mutations remains controversial in COSMIC data base. This observation provides evidence that other publications have shown that these mutations that the mutation data obtained with the Sequenom have no predictive value [14,15]. These inconsistencies, platform is accurate. Our results also demonstrate that a together with two other factors, have limited the impact majority of colon cancer samples have aberrant PIK(3) - of mutation profiling for prognosis and prediction in RAS/RAF network; similar results have been seen pre- standard care of colon cancer. A large sample size is viously [6]. required to establish that a gene mutation has a signifi- Mutations in ABL1 and MET, not previously identified cant impact for patient prediction or prognosis. Another in colon cancer, were identified, and 13 other genes limitation is that until recently conducting such large were screened and found not to be mutated in hot spot studies with the standard sequencing technologies was locations. Furthermore, this study identified the most too time consuming and too expensive to be practical frequent colon cancer mutations from OncoCarta, pro- for clinical studies. viding the necessary information to reduce the number Moreover, while the high frequency of KRAS, BRAF of assays from 187 to 24, creating a smaller, more speci- and PIK3CA mutations in colon cancer is well docu- fic and economical panel, requiring less DNA, and thus mented, other potentially important mutations have not conserving precious clinical samples. been profiled with a large number of clinical samples. Whole genome sequencing of a small number of colon Methods samples demonstrated that somatic cancer mutations Clinical samples and histological evaluation consist of a few genes that occur frequently and many Samples used in this study were from NSABP clinical more mutations that occur very infrequently in many trial C-07. This trial enrolled patients between 02/2000 different genes [16,17]. Mutations in these infrequently and 11/2002 to compare oxaliplatin and bolus 5-FU/LV mutated genes could have a similar effect or synergize to bolus 5FU/LV alone for resected stage II and III with mutations in KRAS, PIK3CA, and BRAF. colon cancer [20]. Tissue samples were obtained at sur- Given these considerations, our goal was to find a gery before the patients had received any treatment and cost-effective and high throughput methodology that were routinely processed with FFPE. C-07 was approved would detect frequent and infrequent cancer mutations by an institutional review board, and informed consent genes in a large number of samples. Furthermore, it was was obtained from each subject. A pathologist categor- essential that the methodology would work with ized tumors into poor, moderate, well differentiated and degraded DNAs isolated from FFPET. signet cells carcinoma, according to the World Health The mass spectrometric SNP genotyping technology Organization (WHO) criteria. Samples were graded for based on the Sequenom platform provided an ideal mucinous character, based on the amount of mucin choice for mutation profiling to address these several retained within the tumor (1 = no mucin, 2 = < 50% criteria. It has been shown to work with small amounts mucinous volume/total tumor volume, 3 = >50% muci- of degraded DNAs (5 ng), and the high multiplexing nous volume/total tumor volume). Only grade 3 tumors capacity minimizes the use of irreplaceable clinical sam- were considered mucinous tumors carcinoma, which is ples. In addition, a variety of studies have demonstrated in accordance with WHO criteria. that the sensitivity of mass spectrometric methods exceeds that of traditional Sanger sequencing and is DNA isolation highly concordant with Sanger sequencing, Pyrosequen- DNA was isolated from FFPE tumor blocks collected cing, and allele-specific PCR [16,18,19]. Furthermore, from patients participating in NSABP clinical trial C-07. Sequenom has recently developed the OncoCarta Panel, FFPE tumor blocks were cut, and sections of the slide an oncogene panel that offers a rapid and parallel analy- containing the most tumor cells were defined by a sis of 238 simple and complex cancer mutations across pathologist and isolated by macrodissection. Genomic 19 genes. DNA was extracted from 4 five μm unstained sections. The OncoCarta panel includes assays for most colon After attempting several extraction procedures from a cancer mutations in the clinically relevant genes, BRAF variety of manufacturers (Machery Negel, Qiagen, (99%), KRAS (98%), and PIK3CA (78%), and in addition Ambion), it was determined that the Ambion Recover- contains assays for other cancer mutations in genes that All™ Total Nucleic Acid isolation kit (Applied Biosystem, Fumagalli et al. BMC Cancer 2010, 10:101 Page 3 of 14 http://www.biomedcentral.com/1471-2407/10/101 Foster City, CA) yielded the best DNA based on the NCI-H1299, NCI-H1395, UACC-893) were purchased quality, quantity, and the performance on the mass from American Type Culture Collection (ATCC, Mana- spectrometer (data not shown). The extraction was per- ssas, VA, US). Two cell lines, SKBR3 and MCF-7, were formed as recommended by the manufacturer, with two grown in culture and cell pellets were fixed in formalin, exceptions; the protease digestion was extended over- andembedded in paraffin and DNAs were isolated as night and the elution volume was increased to 150 ul to described for the clinical samples. maximize the total amount of DNA recovered. Addi- tional protease was added to samples incompletely Mass Spec Type Plex Technology and the digested after the overnight treatment. DNA was mea- OncoCarta Panel sured with fluorescence, using the Quant-iT ™ Pico- For mutation detection, the Sequenom platform and Green® dsDNA Assay Kit (Invitrogen, Carlsbad, CA) and the OncoCarta mutation panel were used and the pro- the InfiniteF200 fluorometer (Tecan, Mannedorf, tocol provided by Sequenom (San Diego, CA) was fol- Switzerland). lowed with minor modifications. A schematic of the As a positive control for known mutations and to test procedure is shown in Fig. 1. A Tecan Evo liquid the performance of the platform, annotated cell line handler was used to normalize the DNA samples and DNAs (A2058, HS578T, HL60, MCF7, MDAMB231, to set up the PCR reactions. The amount of DNA Figure 1 Methodology for mutation detection. Genomic DNA from the samples is amplified by PCR, resulting in copies of both mutant and wildtype alleles. Shrimp Alkaline Phosphatase removed excess nucleotides from the sample wells. Primer extension was performed using terminator nucleotides A, C, T, G, each with distinct masses. This linear amplification results in sequences proportional to the alleles that can be distinguished by mass spec (Maldi-Tof Separation). Fumagalli et al. BMC Cancer 2010, 10:101 Page 4 of 14 http://www.biomedcentral.com/1471-2407/10/101 added to the PCR was reduced to 15 ng or less. DNAs Results were amplified using the OncoCarta PCR primer pools, Mutations were detected in control DNAs from intact and unincorporated nucleotides were inactivated by shrimp FFPETsamples alkaline phosphatase (SAP), and a single base extension Previously described mutations in control cell lines were reaction was performed using extension primers that detected. BRAF_V600E, HRAS_G12D, NRAS_Q61L, hybridize immediately adjacent to the mutations and a PIK3CA_E545K, KRAS_G13D, NRAS_Q61K, custom mixture of nucleotides. Salts were removed by EGFR1_S125L, and PIK3CA_H1047R were detected in theaddition of acationexchangeresin.Multiplexed the appropriate cell lines (A2058, HS578T, HL60, reactions were spotted onto the SpectroChipII, and MCF7, MDAMB231, NCI-H1299, NCI-H1395, and mutations, if present, were resolved by MALDI-TOF UACC-893, respectively). The appropriate mutation was on the Compact Mass Spectrometer (Sequenom, San found in MCF-7 (PIK3CA_E545K) from both intact Diego, CA). DNA and DNA isolated from FFPET. DNAs from clini- The OncoCarta™ Panel v1.0 (Sequenom, San Diego, cal samples, control cell lines, and cell lines formalin- CA) consists of 24 pools of primer pairs and 24 pools of fixed, paraffin-embedded cell lines showed the same extension primers, and has the capacity to detect 238 rates of primer extension and performance on mass mutations in 19 genes, listed in Table 1. Each pool con- spectrometer. sists of 5-9 primer pairs in the PCR reaction. Two types The proportion of the mutated alleles in each cell line, of assays have been designed in the OncoCarta panel, as observed from the area under the mutant peak on referred to as simple and complex. The simple assays the spectra, ranged from 0.4-0.6, as expected for a pure arethose in whichasingle assayisabletodetectthe clonal population with a heterozygote mutation. Spectra amino acid changes at that codon. The complex assays for cell line UACC-893 had equal fractions of mutant arethose that requiremorethan one assaytoidentify and wt alleles (Fig. 2A). One exception to this distribu- codon changes or deletions and insertion, and thus are tion among cell lines was seen in A2058, which showed able to detect multiple different amino acid substitutions spectra consistent with 2 copies of the WT allele and or deletions. An example of a complex assay is KRAS_1 one mutant BRAF mutant allele (Fig. 2B). The 3 alleles and KRAS_2, which interrogate 2 different nucleotide of BRAF in A2058 are consistent with the observation positions within codon 12 and together identify all that there are 3 copies of chromosome 7 in this cell line codon 12 amino acid changes. Much more complex (COSMIC in the SNP Array Based LOH and Copy assays are included in OncoCarta, which interrogate Number Analysis data base) [21]. insertions and deletions within the EGFR gene. The Sequenom platform was sensitive and quantitative Data analysis Pilot studies demonstrated that the assays worked with Data analysis was performed using MassArray Typer Ana- as little as 1 ng of DNA (Fig. 3). The fraction of unex- lyzer software 4.0.4.20 (Sequenom), which facilitates visua- tended primer was .09 even when the input DNA was lization of data patterns as well as the raw spectra. between 1-3 ng, When concentrations of the amount of Mutations were identified in two different ways. Typer DNA was between 3-14 ng, the fraction of unextended automates the identification of mutants by comparing primer was similar, .07. Thus, the assays worked well ratios of the wild type peak to that of all suspected even when only 1 ng of DNA was used. mutants and generates an Onco Mutation report detailing In clinical samples with some assays it was possible to specific mutations and the ratios of wild type and muta- detect mutations that only represented 5% of the total 2 tion peaks. In addition, raw data was exported to Excel peak areas. The spectra in Fig. 4 show a small but clear and an in-house macro was used to duplicate the analysis. peak at the expected size for a PIK3CA 1047R mutation Theareaunder thepeaks allows forquantificationfor in a lymph node. We also were able to demonstrate the each allele, giving a direct evaluation of the proportion of sensitivity of the platform by performing a cell mixing mutated and wildtype (wt) allele in the sample [18]. experiment. Mutation analysis was done using MCF-7 All mutations from both the Onco mutation report cell line DNA alone or mixed with SKBR3 at various and the in-house Excel report were reviewed manually percentages. MCF-7 cells contain a PIK3CA mutation, by 3 investigators (DF, PGG, KPG). Manual review of and SKBR3 cells do not. Fig. 5 demonstrates that the mutations was necessary to identify “real” mutant peaks mutation was detectable even when the MCF-7 cells from salt peaks or other background peaks. Selected represented only 5 to 10% of the total DNA and only 5 reviewed mutations from the Onco Mutation Report to 2.5% of the alleles. This sensitivity is important for and from the in-house macro were compared and were mutation detection in clinical cancer samples, which concordant. usually contain some amount of normal tissue, which Fumagalli et al. BMC Cancer 2010, 10:101 Page 5 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 1 Mutations detected with OncoCarta ABL1-G250E EGFR-L747_E749del, A750P KIT-P585P ABL1-Q252H EGFR-E746_A750del KIT-D579del ABL1-Y253H EGFR-L747_E749del, A750P KIT-K642E ABL1-Y253F EGFR-L747_S752del, P753S KIT-D816V ABL1-E255K EGFR-E746_T751del, V ins KIT-D816H/D816Y ABL1-E255V EGFR-L747_S752del, Q ins KIT-V825A ABL1-D276G EGFR-L747_S752del, Q ins KIT-E839K ABL1-F311L EGFR-E746_T751del, S752D/SNP C2255T KIT-M552L ABL1-T315I EGFR-D770_N771>AGG/V769_D770insASV/V769_D770insASV KIT-Y568D ABL1-F317L EGFR-D770_N771insG KIT-F584S ABL1-M351T EGFR-L747_T750del, P ins KIT-P551_V555del ABL1-E355G EGFR-E746_A750del KIT-P551_V555del ABL1-F359V EGFR-E746_T751del, I ins KIT-Y553_Q556del ABL1-H396R EGFR-L747_T751del KIT-Y553_Q556del AKT1-rs11555435 EGFR-L747_T751del KRAS-G12V/A/D/C/S/R/F AKT1-rs11555431 EGFR-E746_A750del, V ins KRAS-G13C/S/V/D AKT1-rs11555432 EGFR-E746_A750del, V ins KRAS-A59T AKT1-rs12881616 EGFR-S752_I759del KRAS-Q61E/K/L/R/P/H AKT1-rs11555433 ERBB2-L755P MET-R970C AKT1-rs11555436 ERBB2-G776S/G776LC MET-T992I AKT1-rs34409589 ERBB2-G776VC MET-Y1230C AKT2-S302G ERBB2-G776VC/G776VC MET-Y1235D AKT2-R371H ERBB2-M774_A775insYVMA MET-M1250T BRAF-G464R ERBB2-A775_G776insYVMA NRAS-G12V/G12A/G12D BRAF-G464V/G464E ERBB2-P780_Y781insGSP NRAS-G12C/G12R/G12S BRAF-G466V/G466G/G466E ERBB2-P780_Y781insGSP NRAS-G13V/G13A/G13D BRAF-G466R ERBB2-S779_P780insVGS NRAS-G13C/G13R/G13S BRAF-F468C FGFR1-S125L NRAS-A18T BRAF-G469S/E/A/V/R FGFR1-P252T NRAS-Q61L/Q61R/Q61P BRAF-D594V| G FGFR3-R248C NRAS-Q61H BRAF-F595L FGFR3-S249C NRAS-Q61E/Q61K BRAF-G596R FGFR3-G370C PDGFRA-V561D BRAF-L597S/R/Q/V FGFR3-Y373C PDGFRA-T674I BRAF-T599I FGFR3-A391E PDGFRA-F808L BRAF-V600E/K/R/L FGFR3-K650Q/E PDGFRA-D846Y BRAF-K601N/E FGFR3-K650T/M PDGFRA-N870S CDK-R24C/H FLT3-I836del PDGFRA-D1071N EGFR-R108K FLT3_2 PDGFRA-D842_H845del EGFR-T263P FLT3_3 PDGFRA-I843_D846del EGFR-A289V FLT3-D835H/D835Y PDGFRA-S566_E571>K EGFR-G598V HRAS-G12V/D PDGFRA-I843_S847>T EGFR-E709K/E709H HRAS-G13C/R/S PDGFRA-D842V EGFR-E709A/E709G/E709V HRAS-G13V/D PIK3CA-R88Q EGFR-G719S/G719C HRAS-Q61H PIK3CA-N345K EGFR-G719A HRAS-Q61H/L/R/P/K PIK3CA-C420R EGFR-M766_A767insAI JAK2-V617F PIK3CA-P539R EGFR-S768I KIT-D52N PIK3CA-E542K EGFR-V769_D770insASV KIT-Y503_F504insAY PIK3CA-E545K EGFR-V769_D770insCV KIT-W557R/W557R/W557G PIK3CA-Q546K EGFR-D770_N771>AGG/V769_D770insASV/V769_D770insASV KIT-V559D/V559A/V559G PIK3CA-H701P EGFR-D770_N771insG KIT-V559I PIK3CA-H1047R/H1047L EGFR-N771_P772>SVDNR KIT-V560D/V560G PIK3CA-H1047Y Fumagalli et al. BMC Cancer 2010, 10:101 Page 6 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 1: Mutations detected with OncoCarta (Continued) EGFR-P772_H773insV KIT-K550_K558del PIK3CA-G1049R EGFR-H773>NPY KIT-K558_V560del PIK3CA-R38H EGFR-H773_V774insNPH/H773_V774insPH/H773_V774insH KIT-K558_E562del PIK3CA-C901F EGFR-V774_C775insHV KIT-V559del PIK3CA-M1043I/M1043I EGFR-T790 M KIT-V559_V560del RET-C634R EGFR-L858R KIT-V560del RET-C634W/Y EGFR-L861Q KIT-Y570_L576del RET-E632_L633del EGFR-L747_T750del, P ins/E746_A750del, T751A KIT-E561K RET-M918T EGFR-E746_T751del, I ins/S752_I759del KIT-L576P RET-A664D dilutes the number of tumor cells. This is of particular database [21]. The COSMIC frequencies seen in Table 2 concern when profiling lymph nodes, which may con- are based only on those mutations that are detectable tain a minority of tumor cells. with OncoCarta. OncoCarta assays interrogate 99%, 98%, and 78% of the known colon cancer mutations in BRAF, Frequencies of C-07 mutations in KRAS, NRAS, PIK3CA, KRAS,and PIK3CA, respectively, based on a large num- and BRAF detected with OncoCarta and the Sequenom ber of colon cancer samples that have been sequenced in platform were similar to previous reports BRAF (n = 4628), KRAS (n = 858) and PIK3CA (n = 247). In this preliminary assessment of the feasibility of using The OncoCarta panel found that the most frequent the Sequenom platform to do large-scale mutation profil- mutations in C-07 were KRAS (43.5%), PIK3CA (20.1%), ing of colon cancer samples isolated from FFPET, it was and BRAF (12.1%), which are similar to what is seen in essential to determine if our data yielded frequencies COSMIC. NRAS mutations, while infrequent, were typical of what has been seen previously. Table 2 shows detected in codons 12, 13 and 61 and represent a sizable the mutation frequencies obtained here and from the minority of the C0-7 samples (3.8%). These data suggest COSMIC (Catalog of Somatic Mutations in Cancer) Figure 2 Spectra for cell lines UACC-893 and A2058. The expected positions for the unexteneded primer (UEP), and the extension products (Mutant and WT) from assays PIK3CA_9 and BRAF_15 in cell lines UACC-893 and A2058, respectively, are indicated with red dashed lines. The proportion of peak areas and the specific base is also shown. Assays PIK3CA_9 and BRAF_15 detected mutations in PIK3CA at amino acid position 1047 and in BRAF at amino acid position 600, respectively. Other peaks included in these spectra as result of multiplexing but not part of the designated assays are indicated as grey dashed lines. Fumagalli et al. BMC Cancer 2010, 10:101 Page 7 of 14 http://www.biomedcentral.com/1471-2407/10/101 Figure 3 Fractional unextended primer versus input DNA. The range and the average for the percent of unextended primer for different amounts of input DNA into the PCR reactions are shown. The number of samples used in each category was 4 for 1-3 ng, 9 for 3-9 ng, 13 for 9-13 and 210 for 14 ng. Figure 4 Sensitive detection of mutations in clinical FFPE samples with the Sequenom platform. Small mutant but definitive peak illustrating a PIK3CA-1047R mutation in approximately 5% of the sample DNA is shown. Fumagalli et al. BMC Cancer 2010, 10:101 Page 8 of 14 http://www.biomedcentral.com/1471-2407/10/101 Figure 5 Quantification of the sensitivity with a cell line mixing experiment. Spectra of MCF-7 cells (mutant) alone or mixed with SKBR3 cells (WT) are shown. Percents are based on the ng amounts of DNA. This assay detects an E545K mutation in PIK3CA. that FFPET samples can be interrogated with the tech- Sequenom data was reproducible nology described here and yield accurate data. Most of the assays in the OncoCarta panel did not While most of the specific amino acid mutations mir- detect mutations or the frequency of mutations was ror what is seen on the COSMIC database, some unique very low (below 1%) in our colon cancer samples. colon cancer gene mutations were found, which include OncoCarta assays interrogate mutations in these 19 ABL1-F359V, AKT1-E17K, MET-R970C, and MET- genes listed in Table 1. To reduce the cost, time and T992I. Other amino acid changes that were not in the the amount of DNA required for profiling, only 24 COSMIC database were amino acid changes R88Q, assays, which detected mutations at a frequency of 1% H701P, and C420R in PIK3CA, BRAF-594V/G, and or greater in C-07, were selected, resorted in 6 pools KRAS-Q61R, and several in NRAS, including G12C, and included in a new panel, termed ColoCarta (Table G12D, G13R, G13V, Q61H and Q61K (Table 2). 3). Mutation profiles of 32 mutant samples with 41 mutations were repeated with the ColoCarta. The mutations detected by the 2 panels (OncoCarta and MET mutations were found in C0-7 and amplified in ColoCarta) were identical, demonstrating the reprodu- sometumors cibility of the methodology. MET mutations were found in 3.3% of C-07 samples. Interestingly, these mutations were not only unexpected Multiple mutation frequencies suggest an order to the in their appearance within the colon cancer population but also the frequency within the samples was unex- acquisition of different mutations pected. In four of the eight samples with MET muta- A majority of the tumors (64%) contained at least one tions, the mutant alleles were present at 58-70%, or more mutations in the following genes: BRAF, suggesting an amplification of the mutant allele or a loss KRAS, NRAS, MET,or PIK3CA, and 18% had 2 or of the wt gene (Fig. 6). Amplification may represent the more mutations. The most common double mutation best explanation, in that amplification of the MET geno- was in KRAS and PIK3CA,followedby PIK3CA and mic region, 7q31, has been observed in the Progenetix BRAF (Table 4). Most samples with PIK3CA mutations CGH Database in 23% of colorectal cancers [22] (80%) also had mutations in other genes, the most Fumagalli et al. BMC Cancer 2010, 10:101 Page 9 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 2 Frequency of colon cancer mutations † ‡ Mutation No. Mutated Samples Frequency in Primary Tumor* in COSMIC Multiple Mutations ABL1-F359V 1 0.40% 0/66 ABL1 Total 1 0.40% NF(0/66) 100% AKT1-E17K 1 0.40% 0/31 AKT1 Total 1 0.40% NF(0/31) 100% BRAF-D594V| G 1 0.40% NF (0/3179) BRAF-V600E 28 11.70% 14.60% BRAF Total 29 12.10% 14.70% 24% KRAS-G12A 2 0.80% 1.80% KRAS-G12C 10 4.20% 3.60% KRAS-G12D 40 16.70% 13.20% KRAS-G12R 3 1.30% 0.40% KRAS-G12S 3 1.30% 4.20% KRAS-G12V 20 8.40% 7.40% KRAS-G13D 23 9.60% 5.20% KRAS-A59T 1 0.40% 0.10% KRAS-Q61L 1 0.40% 0.20% KRAS-Q61R 1 0.40% NF (0/1927) KRAS Total 104 43.50% 36.10% 34% MET-R970C 2 0.80% NF (0/77) MET-T992I 6 2.50% NF (0/77) MET Total 8 3.30% 0% 50% NRAS-G12C 1 0.40% NF (0/46) NRAS-G12D 4 1.70% NF (0/46) NRAS-G13R 1 0.40% NF (0/46) NRAS-G13V 1 0.40% NF (0/46) NRAS-Q61H 1 0.40% NF (0/46) NRAS-Q61K 1 0.40% NF (0/46) NRAS Total 9 3.80% 2.2% 38% PIK3CA-R88Q 5 2.10% NF (0/171) PIK3CA-C420R 2 0.80% NF (0/171) PIK3CA-E542K 9 3.80% 4.10% PIK3CA-E545K 12 5.00% 4.10% PIK3CA-Q546K 4 1.70% 1.20% PIK3CA-H701P 1 0.40% NF (0/171) PIK3CA-H1047L 1 0.40% 1.80% PIK3CA-H1047R 14 5.90% 5.30% PIK3CA Total 48 20.10% 16.40% 80% *% of C-07 samples with this mutation. Data from COSMIC for colon adenocarcinoma limited to the same mutations interrogated with OncoCarta. The mutations listed are only the ones found in C-07. Some COSMIC amino acid changes are not shown here if they were not mutated in C-07. % of samples with a mutation in the gene shown and at least one other mutation in C-07 samples. COSMIC data are from large intestine, not specific to colon. frequent of which was KRAS; other mutated genes of colon cancer [23,24]. The multiple mutation fre- were BRAF, MET, NRAS,and a second PIK3CA muta- quencies for tumors with KRAS and PIK3CA or with tion (Table 2, last column). Tumors with MET and PIK3CA and BRAF were slightly higher and lower, NRAS mutations also have an unexpectedly high fre- respectively, than expected based on their individual quency of co-occurring mutations, which suggests that frequencies (Table 4). Conversely, the expected double they occur as a second mutation and perhaps later in mutation frequency of BRAF and KRAS would be 5.1%, the etiology of the tumor. Many tumors contain only a based on our data, but this combination was not KRAS or BRAF mutation, which is consistent with pre- found, also in agreement with previous reports [24] vious reports finding these mutations in earlier stages (Table 4). Fumagalli et al. BMC Cancer 2010, 10:101 Page 10 of 14 http://www.biomedcentral.com/1471-2407/10/101 Primary tumors with KRAS and PIK3CA mutations vary with respect to the frequency of these mutant alleles In the samples with co-occurring mutations, the ratios of KRAS mutation ratio (KRAS mutation peak area/total peak area) to the PIK3CA mutation ratio (PIK3CA mutation peak area/total peak area) was determined. Twenty-two out of 31 samples (71%) had KRAS/PIK3CA ratios above 1.25 (Table 5). PIK3CA mutations were more prevalent in only 2 out of 31 samples. These dif- ferences demonstrate that in a majority of primary tumors with double mutations in KRAS and PIK3CA, the KRAS mutations are more prevalent than the PIK3CA. This unequal distribution of mutant alleles within a tumor may be due to the fact that a majority of thetumor cellshaveonlythe KRAS mutation, and cells with a PIK3CA mutation are in the minority, or it could be due to copy number variations in the KRAS and PIK3CA loci. BRAF mutations were correlated with poorly differentiated tumors and with mucinous tumors Figure 6 MET mutation is amplified. The proportion and position The frequency of mutations for KRAS, PIK3CA,and (blue dashed lines) of mutant and wt alleles are shown. The MET_1 BRAF were tested for correlation to the degree of differ- assay detects R970C mutations in MET. entiation and to the prevalence of mucin in the tumor. BRAF mutations were found in 26.2% of the poorly dif- ferentiated tumors and in 8.2% of the moderate and well differentiated. These frequencies were significantly dif- Table 3 ColoCarta panel ferent by Chi square test (p value = 0.001). BRAF muta- Sequenom’s Assay Name Amino Acid Change tions were also associated with mucinous tumors: BRAF BRAF_15 &16 V600E/K/R/L mutations occurred in 28% of grade 3 mucinous tumors BRAF_9 BRAF-D594V (>50% mucinous tumor cells) but in only 9.4% of the HRAS_6* HRAS-Q61L non-mucinous tumors (grade 1 and 2). This was signifi- KRAS_1 & 2 G12V/A/D/C/S/R/F cant by the Chi square test at p value = 0.006. Similar KRAS_4 KRAS-G13D data have been reported previously [25,26]. KRAS and KRAS_5 KRAS-A59T PIK3CA mutations did not correlate with either the KRAS_7 KRAS-Q61L degree of differentiation or with prevalence of mucinous KRAS_8 Q61H/Q61H cells. MET_1 MET-R970C MET_2 MET-T992I Mutation profiling demonstrated a majority of primary NRAS_1 NRAS-G12V and lymph node samples were concordant but NRAS_2 NRAS-G12C differences were detected NRAS_3 NRAS-G13V Lymph node metastases were not routinely collected in NRAS_4 NRAS-G13C C-07 but as a pilot study to determine the feasibility of NRAS_7 NRAS-Q61H using lymph nodes for mutation profiling was conducted. NRAS_8 NRAS-Q61E We isolated DNA from 39 lymph nodes containing PIK3CA_1 PIK3CA-R88Q tumor cells and their corresponding primary tumors. PIK3CA_3 PIK3CA-C420R These primary and lymph nodes samples were profiled PIK3CA_5 PIK3CA-E542K with the entire OncoCarta panel. The majority of lymph PIK3CA_6 PIK3CA-E545K nodes and their corresponding primary tumors (89.7%) PIK3CA_7 PIK3CA-Q546K were concordant. A total of 26 mutations were detected PIK3CA_8 PIK3CA-H701P in lymph nodes, including KRAS, BRAF, PIK3CA,and PIK3CA_9 PIK3CA-H1047R NRAS. Thirty-five out of 39 lymph nodes had identical *HRAS_6 was included in panel but occurred in < 0.1% of samples. mutation profiles, but in 4 cases mutations in the primary Fumagalli et al. BMC Cancer 2010, 10:101 Page 11 of 14 http://www.biomedcentral.com/1471-2407/10/101 Table 4 Single and double mutations in C-07 Double Mutation Frequencies KRAS PIK3CA All other Single Actual Expected Actual Expected Actual Expected KRAS 43.70% NA NA 10.40% 8.70% 14.60% 7.21% PIK3CA 20.10% 10.40% 8.70% NA NA 15.50% 8.06% BRAF 11.80% 0 5.10% 1.80% 2.40% 2.50% 5.71% MET 3.30% 1.67% 1.44% 0 0.66% 1.67% 2% NRAS 3.80% 0 1.66% 0.42% 0.40% 1.30% 2.14% All Mutations 60.20% primary tumor in 3 out of 4 samples. In sample 0940, Table 5 KRAS/PIK3CA ratio mutation frequencies within the KRAS/PIK3CA mutation decreased by almost 1/2 in primary tumors the lymph node tumor compared to the primary. Thus, No of Samples KRAS/PIK3CA in these samples there is either a loss of KRAS muta- 22 1.25-3.22 tions or an accumulation of PIK3CA mutations, suggest- 7 0.93-1.13 ing that PIK3CA mutations may impart a selective 2 0.81-.42 advantage in the lymph node. Average 1.67 In contrast, two other samples have a less frequent Median 1.6 occurrence of their PIK3CA mutation in the lymph node than in the primary tumor. In sample 2244, the PIK3CA tumors were not found in the corresponding lymph mutation was undetectable in the lymph node (Table 6). nodes (BRAF [2], PIK3CA [1] and KRAS [1]). In fact, if there was a selection for both mutations in the lymph node, then the PIK3CA mutation frequency would Mutation profiles demonstrate that tumor cell have been thesameasthat ofthe KRAS mutation (0.15). populations may be different in lymph nodes and in the On the other hand, if the PIK3CA/KRAS ratio were the primary tumors same in the primary and lymph node tumor, then the Peak area evaluation of tumors that had 2 mutations PIK3CA mutation frequency would have been .08, which and for which a metastatic lymph node was available is still detectable with this technology (Fig. 3). Thus, in demonstrated differences between the primary and sample 2244 there were fewer PIK3CA mutant alleles in lymph node samples. Table 6 details the frequency of the lymph node than in the primary tumor. In sample mutant and wt alleles based on the peak areas for 5 1837, mutations in both BRAF and PIK3CA were such samples. detected and the BRAF/PIK3CA ratio was 1.67, but KRAS to PIK3CA ratios demonstrated that there were increased to 4.4 in the metastatic lymph node. more KRAS mutations than PIK3CA mutations in 4 of 4 primary samples, and in 3 of the 4 lymph node samples. Discussion However, it is also notable that the ratio of KRAS/ The Sequenom platform provides a superior technology PIK3A was lower in the lymph node compared to their for the screening of many hot spot mutations in cancer Table 6 Allele frequencies for primary tumors with two mutations and metastatic lymph nodes Sample Mutations Mutant Allele Frequency Mutation Ratios Primary Lymph Node Primary M1/M2 Lymph node M1/M2 C07-0388 M1. KRAS-G12D 0.45 0.44 1.88 1.63 M2. PIK3CA-H1047R 0.24 0.27 C07-0717 M1. KRAS-G13D 0.08 0.09 1.14 0.90 M2. PIK3CA-H1047R 0.07 0.1 C07-0940 M1. KRAS-G12D 0.42 0.34 1.91 1.10 M2. PIK3CA-E542K 0.22 0.31 C07-2244 M1. KRAS-G12C 0.37 0.15 1.76 _ M2. PIK3CA-H1047R 0.21 0 C07-1837 M1. BRAF-V600E 0.2 0.22 1.67 4.40 M2. PIK3CA-H1047R 0.12 0.05 Fumagalli et al. BMC Cancer 2010, 10:101 Page 12 of 14 http://www.biomedcentral.com/1471-2407/10/101 samples. Sanger sequencing would require amplification reported for KRAS, PIK3CA and BRAF. For example, of at least 60 different fragments per sample, and many AKT1 and NRAS are molecules that are downstream reactions would require optimization, thus adding con- mediators of the EGFR signaling pathway, and muta- siderable time and expense. Multiplexing and the use of tions in these genes are likely to affect the response to the OncoCarta panel allowed us to skip this time con- drugs that target EGFR. suming step. Thus, conservatively, Sanger sequencing Mutations in ABL, AKT1, and MET were detected here would be 40 times more expensive, and require at least but were not listed in COSMIC, probably due to the 2 times more DNA. Other sequencing technologies, small number of samples analyzed. The AKT1-E17K which employ differential melting of mutant and wt mutation was initially identified as a SNP, rs34409589, sequences, such as HRMA, still require that the PCR but in a recent publication it was found to be a somatic product be sequenced. This would add significant cost mutation and was found in 3 of 51 colon cancers [27]. and time to the procedure because 60% of the colon The frequency of these mutations in this small study (51 samples contained one or more mutations. In addition samples) was 6% and is much greater than in the C0-7 the Sequenom platform is more sensitive than Sanger samples (0.4%). This difference in frequencies may be sequencing in that it was able to detect mutations that because the Carpten et al [27] samples were from more represented only 5% of the DNA. Pyrosequencing repre- advanced stages than those from the C-07 trial. More- sented a potential alternative to the Sequenom platform, over, they selected large tumors (>100 mg) and contain- but in our hands assays needed to be optimized, and the ing more than 60% tumor cells. No such selection was lack of multiplexing made the procedure more time done for our study, and samples were from stages II and consuming and demanded more DNA. The Sequenom III exclusively. The significance of ABL1 and AKT1 muta- methodology also focuses on only those nucleotides that tions for patient prediction and prognosis in our study is are known to be cancer mutations and thus makes questionable given that they each were found in only in review of the sequence information considerably faster one sample and represented only 0.4% of the cases. than Sanger. Next-Generation sequencing was cost pro- To ourknowledge,thisisthe firstreportof MET hibitive and has not been shown to work with DNAs mutations in the primary colon cancer, but a different isolated from FFPET. Thus, the Sequenom platform and MET mutation (N1118Y) was found in a lung metastasis the OncoCarta Panel provided the simplest, most rapid, of the large intestine [28]. The MET mutations, R970C sensitive and cost-effective method for detecting hot and T992I, were detected in 8 out of 239 C-07 colon spot cancer mutations in degraded DNAs isolated from cancers. These mutations correspond to MET-R988C archival and routinely processed FFPET. The ColoCarta and MET-T1010I, respectively, in the long form of MET panel provides a more specific panel for colon cancer which is the isoform referred to in the COSMIC data- mutation detection and greatly reduces the amount of base [29]. The R970C and T992I mutations are located DNA needed for mutation profiling. in the juxtamembrane segment of the protein and were The frequencies and specific amino acid mutations detected in lung carcinoma [30]. These mutations, when detected here were similar to the COSMIC database and introduced into a lung cell line, increased focus forma- other publications [6]. The small variation in frequency tion, formation of colonies in soft agar, cell motility, and between our data and other reports may be attributed migration. These mutations also resulted in constitutive to differences in the stage of the samples analyzed, the tyrosine phosphorylation on several cellular proteins number of samples considered, and the sensitivity of the including paxillin at key tyrosine residues and may technology [18]. These observations, combined with the account for the increased motility of cells with this perfect match that we obtained between the expected mutation. Another critical amino acid in this location is and the detected mutations in our control cell lines, a Ser 985, which, when phosphorylated, has been found both fresh and FFPE, and the fact that mutations to diminish MET signaling [31]. If phosphorylation at detected with OncoCarta and ColoCarta were identical, Thr residue 992 (1010) reduces signaling, then the suggest that the technology is reliable and reproducible R992I mutation would inhibit this negative feedback and in DNAs isolated from FFPE samples. may result in constitutive signaling [30]. In our study, the majority of tumors (60.3%) had one If MET mutations confer an alternative activated sig- or more mutations in KRAS, PIK3CA,and BRAF. Muta- naling pathway, then these mutations could also confer tions in these genes are likely to perturb many different resistance to anti-EGFR-based therapies or provide a and overlapping signaling pathways, including PI3K/ new target for directed therapies. Therapeutic drugs AKT, ERK/MAPK, SAPK/JNK, NFKb,and others. We have been developed to specifically target MET,includ- were also able to detect other less frequent mutations ing small molecule kinase inhibitors, anti-MET mono- that are likely to perturb the same pathways and these clonal antibodies, and inhibitors of HGF, the MET may cause resistance to EGFR-targeted therapies, as ligand. Invitro assays have demonstrated that a number Fumagalli et al. BMC Cancer 2010, 10:101 Page 13 of 14 http://www.biomedcentral.com/1471-2407/10/101 of MET targeted therapies were able to prevent MET correspond to SNPs rs34589476 and rs56391007 in the signaling, decrease cell viability, and limit cell motility NCBI SNP data base, respectively. The frequency for and migration in vitro[32].The smallmoleculeARQ these SNPs is unknown so whether these nucleotide 197, a kinase inhibitor, has entered phase II clinical changes are associated with cancer is unknown. trials so may represent a possible therapeutic strategy for some colon tumors. Acknowledgements To our knowledge, this is also one of the most The authors would like to thank Melanie Finnigan, Bill Hiller, and Theresa exhaustive analyses of mutation profiling of metastatic Oeler for help in cutting and cataloging slides, and Hema Liyanage from Sequenom for replexing the assays. This study was supported by Public lymph nodes and their corresponding primary colon Health Service grants U10-CA-37377, U10-CA-69974, U10-CA-12027, and U10- tumors. Our analysis showed that a majority of samples CA-69651 from the National Cancer Institute, National Institutes of Health, were concordant (89.7%) but in a few samples mutations and Department of Health and Human Services. This project is funded, in part, under a grant with the Pennsylvania Department of Health. The were detected only in the primary tumor and not in the Department specifically disclaims responsibility for any analyses, metastatic lymph node. Also in samples with 2 co- interpretations or conclusions. The authors retain the right to provide a copy occurring mutations, the ratio of the double mutations of the final manuscript to the NIH upon acceptance for journal publication, for public archiving in PubMed Central as soon as possible but no later than varied in primary and lymph node tumors. Discordance 12 months after publication by the journal. in the genetic profile between primary tumors and the metastatic lymph nodes has been observed [33]. Such Authors’ contributions DF and PGG designed and carried out the experiments and participated in data may indicate that tumor cell migration selects dif- the drafting of the manuscript. YT graded tumors with regard to the degree ferent cell populations from the one in the primary of differentiation, mucin content, and defined tumor regions, S-IK carried out tumor. However, it is also possible that these mutational experiments. H-JC defined the tumor regions; SP participated in the coordination of the study; KLP-G designed and coordinated the study and differences between the lymph node and the primary drafted the manuscript. All authors have given final approval of the version tumor are a result of tumor heterogeneity. to be published. Another interesting observation in our study was that Competing interests BRAF mutations were significantly correlated with The authors declare that they have no competing interests. poorly differentiated tumors and the prevalence of mucin; similar observations have been reported [25,26]. Received: 12 August 2009 Accepted: 16 March 2010 Published: 16 March 2010 These characteristics are both associated with a worse prognosis and are consistent with other reports associat- References ing BRAF mutations with a bad prognosis [3]. However, 1. Ogino S, Nosho K, Kirkner GJ, Shima K, Irahara N, Kure S, Chan AT, in our study we found that there were 2 metastatic Engelman JA, Kraft P, Cantley LC, et al: PIK3CA Mutation Is Associated With Poor Prognosis Among Patients With Curatively Resected Colon lymph nodes that did not maintain the BRAF mutation Cancer. J Clin Oncol 2009. present in the corresponding primary tumor, suggesting 2. Andreyev HJ, Norman AR, Cunningham D, Oates J, Dix BR, Iacopetta BJ, that BRAF mutations are not essential for metastatic Young J, Walsh T, Ward R, Hawkins N, et al: Kirsten ras mutations in patients with colorectal cancer: the ‘RASCAL II’ study. Br J Cancer 2001, spread to the lymph node in all tumors. Clearly, addi- 85(5):692-696. tional studies would be required to understand these 3. Samowitz WS, Sweeney C, Herrick J, Albertsen H, Levin TR, Murtaugh MA, apparent inconsistencies; additional lymph node samples Wolff RK, Slattery ML: Poor Survival Associated with the BRAF V600E Mutation in Microsatellite-Stable Colon Cancers. Cancer Res 2005, are not currently available but could be the subject of 65(14):6063-6069. further studies when samples become available [3]. 4. Ogino S, Nosho K, Kirkner GJ, Kawasaki T, Meyerhardt JA, Loda M, Giovannucci EL, Fuchs CS: CpG island methylator phenotype, microsatellite instability, BRAF mutation and clinical outcome in colon Conclusions cancer. Gut 2009, 58(1):90-96. The Sequenom platform provided a superior technology 5. Conlin A, Smith G, Carey FA, Wolf CR, Steele RJC: The prognostic for the screening of 238 common hot spot cancer muta- significance of K-ras, p53, and APC mutations in colorectal carcinoma. Gut 2005, 54(9):1283-1286. tions in 19 genes. The frequent occurrence of KRAS, 6. Barault L, Veyrie N, Jooste V, Lecorre D, Chapusot C, Ferraz JM, Lievre A, PIK3CA,and BRAF was confirmed, and mutations not Cortet M, Bouvier AM, Rat P, et al: Mutations in the RAS-MAPK, PI(3)K detected before in colon cancer were found in MET and (phosphatidylinositol-3-OH kinase) signaling network correlate with poor survival in a population-based series of colon cancers. Int J Cancer 2008, ABL1. Twenty-five assays from the OncoCarta were 122(10):2255-2259. replexed to form a new panel, termed ColoCarta, which 7. Khambata-Ford S, Garrett CR, Meropol NJ, Basik M, Harbison CT, Wu S, will be used to screen an additional 800 tumors from Wong TW, Huang X, Takimoto CH, Godwin AK, et al: Expression of epiregulin and amphiregulin and K-ras mutation status predict disease NSABP clinical trial C0-7 with the purpose of identify- control in metastatic colorectal cancer patients treated with cetuximab. ing prognostic or predictive markers for stage II and III J Clin Oncol 2007, 25(22):3230-3237. colon cancer. 8. Linardou H, Dahabreh IJ, Kanaloupiti D, Siannis F, Bafaloukos D, Kosmidis P, Papadimitriou CA, Murray S: Assessment of somatic k-RAS mutations as a Note Added in Proof: Although MET-R988C and mechanism associated with resistance to EGFR-targeted agents: a MET-T1010I mutations were listed in COSMIC as systematic review and meta-analysis of studies in advanced non-small- somatic cancer mutations, these nucleotide changes Fumagalli et al. BMC Cancer 2010, 10:101 Page 14 of 14 http://www.biomedcentral.com/1471-2407/10/101 cell lung cancer and metastatic colorectal cancer. The Lancet Oncology molecular features of colorectal cancer independently of microsatellite 2008, 9(10):962-972. instability status. Mol Cancer 2006, 5:2. 9. Lievre A, Bachet JB, Boige V, Cayre A, Le Corre D, Buc E, Ychou M, 27. Carpten JD, Faber AL, Horn C, Donoho GP, Briggs SL, Robbins CM, Bouche O, Landi B, Louvet C, et al: KRAS mutations as an independent Hostetter G, Boguslawski S, Moses TY, Savage S, et al: A transforming prognostic factor in patients with advanced colorectal cancer treated mutation in the pleckstrin homology domain of AKT1 in cancer. Nature with cetuximab. J Clin Oncol 2008, 26(3):374-379. 2007, 448(7152):439-444. 10. Amado RG, Wolf M, Peeters M, Van Cutsem E, Siena S, Freeman DJ, Juan T, 28. Lorenzato A, Olivero M, Patane S, Rosso E, Oliaro A, Comoglio PM, Di Sikorski R, Suggs S, Radinsky R, et al: Wild-type KRAS is required for Renzo MF: Novel Somatic Mutations of the MET Oncogene in Human panitumumab efficacy in patients with metastatic colorectal cancer. J Carcinoma Metastases Activating Cell Motility and Invasion. Cancer Res Clin Oncol 2008, 26(10):1626-1634. 2002, 62(23):7025-7030. 11. Baselga J, Rosen N: Determinants of RASistance to anti-epidermal growth 29. Loriaux MM, Levine RL, Tyner JW, Frohling S, Scholl C, Stoffregen EP, factor receptor agents. J Clin Oncol 2008, 26(10):1582-1584. Wernig G, Erickson H, Eide CA, Berger R, et al: High-throughput sequence 12. Sartore-Bianchi A, Martini M, Molinari F, Veronese S, Nichelatti M, Artale S, analysis of the tyrosine kinome in acute myeloid leukemia. Blood 2008, Di Nicolantonio F, Saletti P, De Dosso S, Mazzucchelli L, et al: PIK3CA 111(9):4788-4796. Mutations in Colorectal Cancer Are Associated with Clinical Resistance 30. Ma PC, Kijima T, Maulik G, Fox EA, Sattler M, Griffin JD, Johnson BE, Salgia R: to EGFR-Targeted Monoclonal Antibodies. Cancer Res 2009, c-MET Mutational Analysis in Small Cell Lung Cancer: Novel 69(5):1851-1857. Juxtamembrane Domain Mutations Regulating Cytoskeletal Functions. 13. Di Nicolantonio F, Martini M, Molinari F, Sartore-Bianchi A, Arena S, Saletti P, Cancer Res 2003, 63(19):6272-6281. De Dosso S, Mazzucchelli L, Frattini M, Siena S, et al: Wild-type BRAF is 31. Hashigasako A, Machide M, Nakamura T, Matsumoto K, Nakamura T: Bi- required for response to panitumumab or cetuximab in metastatic directional Regulation of Ser-985 Phosphorylation of c-Met via Protein colorectal cancer. J Clin Oncol 2008, 26(35):5705-5712. Kinase C and Protein Phosphatase 2A Involves c-Met Activation and 14. Lambrechts D, De Roock W, Prenen H, De Schutter J, Jacobs B, Biesmans B, Cellular Responsiveness to Hepatocyte Growth Factor. J Biol Chem 2004, Claes B, De Hertogh G, Van Cutsem E, Tejpar S: The role of KRAS, BRAF, 279(25):26445-26452. NRAS, and PIK3CA mutations as markers of resistance to cetuximab in 32. Seiwert TY, Jagadeeswaran R, Faoro L, Janamanchi V, Nallasura V, El chemorefractory metastatic colorectal cancer. J Clin Oncol (Meeting Dinali M, Yala S, Kanteti R, Cohen EE, Lingen MW, et al: The MET receptor Abstracts) 2009, 27(15S):4020. tyrosine kinase is a potential novel therapeutic target for head and neck 15. Prenen H, De Schutter J, Jacobs B, De Roock W, Biesmans B, Claes B, squamous cell carcinoma. Cancer Res 2009, 69(7):3021-3031. Lambrechts D, Van Cutsem E, Tejpar S: PIK3CA Mutations Are Not a Major 33. Gamblin TC, Finkelstein SD, Upsal N, Kaye JD, Blumberg D: Microdissection- Determinant of Resistance to the Epidermal Growth Factor Receptor based allelotyping: a novel technique to determine the temporal Inhibitor Cetuximab in Metastatic Colorectal Cancer. Clin Cancer Res 2009. sequence and biological aggressiveness of colorectal cancer. Am Surg 16. Thomas RK, Baker AC, Debiasi RM, Winckler W, Laframboise T, Lin WM, 2006, 72(5):445-453. Wang M, Feng W, Zander T, MacConaill L, et al: High-throughput oncogene mutation profiling in human cancer. Nat Genet 2007, Pre-publication history 39(3):347-351. The pre-publication history for this paper can be accessed here: 17. Wood LD, Parsons DW, Jones S, Lin J, Sjoblom T, Leary RJ, Shen D, [http://www.biomedcentral.com/1471-2407/10/101/prepub] Boca SM, Barber T, Ptak J, et al: The genomic landscapes of human breast doi:10.1186/1471-2407-10-101 and colorectal cancers. Science 2007, 318(5853):1108-1113. Cite this article as: Fumagalli et al.: A rapid, sensitive, reproducible and 18. Vivante A, Amariglio N, Koren-Michowitz M, Ashur-Fabian O, Nagler A, cost-effective method for mutation profiling of colon cancer and Rechavi G, Cohen Y: High-throughput, sensitive and quantitative assay metastatic lymph nodes. BMC Cancer 2010 10:101. for the detection of BCR-ABL kinase domain mutations. Leukemia 2007, 21(6):1318-1321. 19. van Puijenbroek M, Dierssen JW, Stanssens P, van Eijk R, Cleton-Jansen AM, van Wezel T, Morreau H: Mass spectrometry-based loss of heterozygosity analysis of single-nucleotide polymorphism loci in paraffin embedded tumors using the MassEXTEND assay: single-nucleotide polymorphism loss of heterozygosity analysis of the protein tyrosine phosphatase receptor type J in familial colorectal cancer. J Mol Diagn 2005, 7(5):623-630. 20. Kuebler JP, Wieand HS, O’Connell MJ, Smith RE, Colangelo LH, Yothers G, Petrelli NJ, Findlay MP, Seay TE, Atkins JN, et al: Oxaliplatin combined with weekly bolus fluorouracil and leucovorin as surgical adjuvant chemotherapy for stage II and III colon cancer: results from NSABP C-07. J Clin Oncol 2007, 25(16):2198-2204. 21. Forbes SA, Bhamra G, Bamford S, Dawson E, Kok C, Clements J, Menzies A, Teague JW, Futreal PA, Stratton MR: The Catalogue of Somatic Mutations in Cancer (COSMIC). Curr Protoc Hum Genet 2008, Chapter 10(Unit 10):11. 22. Baudis M, Cleary ML: Progenetix.net: an online repository for molecular cytogenetic aberration data. Bioinformatics 2001, 17(12):1228-1229. 23. Pratilas C, Solit D: Therapeutic strategies for targeting BRAF in human Submit your next manuscript to BioMed Central cancer. Rev Recent Clin Trials 2007, 2(2):121-134. and take full advantage of: 24. Yuen ST, Davies H, Chan TL, Ho JW, Bignell GR, Cox C, Stephens P, Edkins S, Tsui WW, Chan AS, et al: Similarity of the phenotypic patterns associated • Convenient online submission with BRAF and KRAS mutations in colorectal neoplasia. Cancer Res 2002, 62(22):6451-6455. • Thorough peer review 25. Tanaka H, Deng G, Matsuzaki K, Kakar S, Kim GE, Miura S, Sleisenger MH, • No space constraints or color figure charges Kim YS: BRAF mutation, CpG island methylator phenotype and • Immediate publication on acceptance microsatellite instability occur more frequently and concordantly in mucinous than non-mucinous colorectal cancer. Int J Cancer 2006, • Inclusion in PubMed, CAS, Scopus and Google Scholar 118(11):2765-2771. • Research which is freely available for redistribution 26. Li WQ, Kawakami K, Ruszkiewicz A, Bennett G, Moore J, Iacopetta B: BRAF mutations are associated with distinctive clinical, pathological and Submit your manuscript at www.biomedcentral.com/submit

Journal

BMC CancerSpringer Journals

Published: Mar 16, 2010

There are no references for this article.