Clin Transl Oncol DOI 10.1007/s12094-014-1233-3
RESEARCH ARTICLE
Gene expression differences in primary colorectal tumors and matched liver metastases: chemotherapy related or tumoral heterogeneity? M. Lo´pez-Go´mez • J. Moreno-Rubio • I. Sua´rez-Garcı´a • P. Cejas • R. Madero • E. Casado • A. M. Jime´nez • M. Sereno • C. Go´mez-Raposo F. Zambrana • M. Merino • D. Ferna´ndez-Luengas • J. Feliu
•
Received: 26 May 2014 / Accepted: 20 September 2014 Ó Federacio´n de Sociedades Espan˜olas de Oncologı´a (FESEO) 2014
Abstract Background Treatment of metastatic colorectal cancer (mCRC) is generally based on genetic testing performed in primary tumor biopsies, but whether the genomic status of primary tumors is identical to that of metastases is not well known. We compared the gene expression profiles of formalin-fixed paraffin-embedded (FFPE) biopsies of colorectal primary tumors and matched liver metastases. Patients and methods We compared the expression of 18 genes in FFPE CRC tumors and their matched liver All authors have agreed to the submission and have participated in the study. M. Lo´pez-Go´mez (&) E. Casado A. M. Jime´nez M. Sereno C. Go´mez-Raposo F. Zambrana M. Merino Medical Oncology Department, Infanta Sofı´a University Hospital, Paseo de Europa 34, San Sebastia´n de los Reyes, 28702 Madrid, Spain e-mail: [email protected] E. Casado e-mail: [email protected] A. M. Jime´nez e-mail: [email protected] M. Sereno e-mail: [email protected] C. Go´mez-Raposo e-mail: [email protected]
metastases from 32 patients. The expression of each gene in CRC primary tumors and their matched liver metastases was tested using Student’s t test for paired samples. Pairwise correlations of each gene in the primary tumors and matched liver metastases were evaluated by Pearson’s correlation coefficient. Results The expression of six genes was significantly different in primary tumors compared with their matched liver metastases [CXCR4 (p \ 0.001), THBS1 (p = 0.007), MMP 9 (p = 0.048), GST Pi (p = 0.050), TYMP (p = 0.042) and DPYD (p \ 0.001)]. For the remaining genes, where no significant differences were observed, only SMAD4 (rs = 0.447, p = 0.010), ERCC1 (rs = 0.423, I. Sua´rez-Garcı´a Internal Medicine Department, Infanta Sofı´a University Hospital, Paseo de Europa 34, San Sebastia´n de los Reyes, 28702 Madrid, Spain e-mail: [email protected] P. Cejas Translational Oncology Department, La Paz University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected] R. Madero Statistics Department, La Paz University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected]
M. Merino e-mail: [email protected]
D. Ferna´ndez-Luengas General Surgery Department, Infanta Sofı´a University Hospital, Paseo de Europa 34, San Sebastia´n de los Reyes, 28702 Madrid, Spain e-mail: [email protected]
J. Moreno-Rubio Precision Oncology Laboratory (POL), Infanta Sofı´a University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected]
J. Feliu Medical Oncology Department, La Paz University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected]
F. Zambrana e-mail: [email protected]
123
Clin Transl Oncol
p = 0.016) and VEGF A (rs = 0.453, p = 0.009) showed significant correlation in expression between the two tissues. Therefore, we only detected similar gene expression levels between the tumor and the metastases in these three markers. Conclusions We only found similar gene expression levels between the tumor and the metastases in three genes (SMAD4, ERCC1, and VEGF A). However, our study could not assess whether the differences in gene expression were secondary to tumoral heterogeneity or to molecular changes induced by previous chemotherapy.
TLDA TOP 1 TYMP TYMS THBS1 VEGF A
Keywords Colon cancer Liver metastasis Predictive biomarkers Prognostic factors
Basing treatment strategies for metastatic colorectal cancer (mCRC) on the results of biomarker assays have been widely incorporated into daily practice [1, 2]. Information not only on KRAS but also in NRAS mutational status is now considered an essential prerequisite for the selection of a targeted therapy based on anti-EGFR drugs. [3]. Moreover, several studies have suggested that mutations in other genes such as BRAF, PIK3CA and p53, are also associated with resistance to anti-EGFR therapies [4–6] The genomic status of these predictive genes is usually tested in primary tumor tissues because material from metastatic sites is not routinely collected and because the latter is generally of poorer quality and harbors fewer tumor cells as a result of tumor necrosis or chemotherapyinduced changes. However, it is unclear whether the genomic status of a given gene will be the same in the primary tumor and its associated metastases. Some studies performed in mCRC patients have suggested that the mutational status of genes such as KRAS, BRAF and PIK3CA in primary tumors serves as an adequate surrogate of the gene’s status in the metastases, whereas other genes, such as EGFR or PTEN show a higher degree of discordance [7]. The development of distant metastases is a dynamic and a multistep process involving several genetic and epigenetic changes [8] in which tumor cells acquire a more aggressive phenotype [9]. For this reason, the importance of genetic heterogeneity across different tumor foci or even within the same tumor site has been widely evaluated in recent years. Genetic intratumor heterogeneity can contribute to treatment failure and drug resistance. For instance, in kidney cancer, tumor heterogeneity occurs not only across distant tumor foci but also within the same tumor site [10]. Furthermore, we often observe that different metastasis sites show different clinical responses to chemotherapy. However, the mechanism by which tumor cells acquire differential chemotherapeutic sensitivity during metastasis is not well understood. Previous studies have reported discordance between the genetic expression profiles of colorectal primary tumors and their corresponding metastases and have consequently
Abbreviations 5FU 5-Fluorouracil B2M Beta 2 microglobulin BAX Bcl-2 associated X protein BRAF v-Raf murine sarcoma viral oncogene homolog B1 CCR 6 Chemokine receptor 6 CXCR4 Chemokine receptor 4 DNA Deoxyribonucleic acid DPYD Dihydropyrimidine dehydrogenase E-cadherin Cell adhesion promoter EGFR Epidermal growth factor receptor ERCC1 Excision repair cross-complementing factor 1 FAS/CD95 TNF factor superfamily member 6 FFPE Formalin fixed paraffin embedded FOLFIRI Chemotherapy scheme including 5-FU leucovorin and irinotecan FOLFOX Chemotherapy scheme including 5-FU leucovorin and oxaliplatin GAPDH Glyceraldehyde 3-phosphate dehydrogenase reference gene GST Pi Glutathione S-Transferase Pi h TERT Human telomerase reverse transcriptase IHC Immunohistochemistry KRAS v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog mCRC Metastatic colorectal cancer MMP-9 Matrix metalloproteinase 9 p p value pAKT Phosphorylated v-akt murine thymoma viral oncogene homolog 1 PCR Polymerase chain reaction PIK3CA Phosphoinositide-3-Kinase, catalytic, alpha polypeptide PLAU Urokinase-type plasminogen PTEN Phosphatase and tensin homolog qPCR quantitative Polymerase chain reaction RNA Ribonucleic Acid
123
Taq low density arrays Topoisomerase I Thymidine phosphorylase Thymidylate synthase Thrombospondin 1 Vascular endothelial growth factor
Introduction
Clin Transl Oncol
suggested that metastatic tissues should be tested [11]. In other tumors, such as breast cancer, the expression of certain biomarkers (e.g., estrogen receptor, progesterone receptor and human epidermal growth factor receptor 2) has been reported to change with disease progression. Because of this discordance, some researchers have proposed mandatory evaluation of biomarkers in breast metastases prior to the selection of a targeted therapy [12]. A variety of chemotherapy regimens have been used for the treatment of mCRC, with the most common agents being fluoropyrimidines [5-fluorouracil (5-FU) and capecitabine] used in combination with either oxaliplatin (FOLFOX) or irinotecan (FOLFIRI). However, chemotherapy resistance is often observed, limiting the effectiveness of these drugs. Overcoming drug resistance has become one of the main challenges of current research [13]. Information on the expression profiles of genes related to response to chemotherapy at different disease stages might be particularly useful for the selection of treatments to avoid resistance. For example, TYMS (Thymidylate Synthase), TYMP (Thymidine Phosphorylase) and DPYD (Dihydropyrimidine Dehydrogenase) are enzymes involved in 5-FU metabolism [14]. Oxaliplatin, in turn, is detoxified by glutathione-related enzymes, such as glutathione S-transferase Pi (GST Pi), and forms platinum DNA adduct lesions that are repaired by enzymes of the nucleotide excision repair system (NER), including ERCC1 (Excision Repair Cross-Complementing factor 1) [15]. Finally, irinotecan targets DNA topoisomerase I, thereby inhibiting DNA replication [16]. Tumor metastasis is a multistage process involving proteolysis, motility, migration, proliferation, inhibition of apoptosis and neoangiogenesis. Determining the expression levels of specific tumor metastasis-associated genes or proteins responsible for colorectal metastasis at different stages of the disease might help elucidate the molecular mechanisms involved in the development of distant metastases. Chemokines (e.g., CXCR4 and CCR6), cell adhesion molecules (e.g., matrix metalloproteases and E-cadherin), plasminogen activator system-related genes (e.g., uPAR), apoptosis promoters (e.g., BAX, FAS and SMAD4) and angiogenic factors (e.g., VEGF A and thrombospondin) all play important roles in the development of distant metastases [17, 18]. The question of whether primary colorectal tumors and their associated liver metastases in a specific patient have similar genetic profiles, or they behave as biologically distinct entities, is still unsolved. Research that investigates this problem might help inform the reliability of the strategy of using information from markers evaluated in primary tumors for determining a course of treatment in metastatic disease. The aim of our study was to compare
the gene expression profiles of primary colorectal tumors and associated liver metastases to determine whether they show similar gene expression patterns.
Materials and methods Patient selection We performed a retrospective study in the Oncology Department of La Paz University Hospital in Madrid, Spain. We included patients diagnosed with CRC resected synchronous liver metastases (diagnosed at the same time of primary tumors) or metachronous liver metastases (diagnosed weeks or months after the diagnoses of primary tumors), between the years 2003 and 2007. We included all diagnosed patients with available clinical data who had undergone surgery for both primary and liver disease and for whom formalin-fixed paraffin-embedded (FFPE) tissues from both primary and metastatic tumor specimens were available. Patients who had received locoregional treatment for liver metastases (radiofrequency or chemoembolization) instead of surgery were excluded. No unresectable patients were included. Clinical information was collected retrospectively from the patients’ medical records. Both in the case of synchronous and metachronous liver metastases, liver specimens were collected several months after the patients had undergone liver surgery. In the case of patients who had received treatment prior to liver resection—neoadjuvant treatment—tissue collection was made once they had finished chemotherapy treatment as well. Institutional approval from the ethics committee was obtained for the study. Gene selection We performed a literature review of all the published studies of genes involved in both the development of liver metastases and chemotherapy resistance. An electronic search was performed on the MEDLINE database from 1980 to 2007 using the MeSH headings ‘‘colorectal cancer’’, ‘‘liver metastases’’, ‘‘liver resection’’, ‘‘primary tumor’’, ‘‘chemotherapy resistance’’, ‘‘carcinogenesis’’, ‘‘metastases spread’’ and ‘‘tumoral markers’’. The search was limited to English-language publications. All titles and abstracts were reviewed, and appropriate papers were further assessed. Case reports, editorials, abstracts and reviews were excluded. 11 genes related to carcinogenesis and metastases spread and 7 genes related to mCRC chemotherapy resistance were selected. The selected genes are shown in Table 1.
123
Clin Transl Oncol Table 1 Genes determined by qPCR Genes related to carcinogenesis and metastases spread
Genes related to chemotherapy resistance
Description
Geneunigene
BAXHs.624291
BCL2-associated X protein
DPYDHs.335034
Dihydropyrimidine dehydrogenase
CCR 6Hs.34526
Chemokine (C-X-C motif) receptor 6
CDH1Hs.461086
Cadherin 1, type 1, E-cadherin (epithelial)
ERCC1Hs.435981
Excision repair cross-complementing rodent repair deficiency, complementation group 1 Glutathione S-transferase Pi
Geneunigene
Number
Number
Description
CXCR4Hs.593413
Chemokine (C-X.C motif) receptor 4
GST PiHs.523836
FAS/CD95Hs.244139
TNF factor superfamily, member 6
KRASHs.505033
V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog Topoisomerase I
h TERTHs.492203
Human Telomerase Reverse Transcriptase
TOP1Hs.
MMP 9Hs.297413
Matrix Metalloproteinase
TYMSHs.369762
Thymidylate synthase
PLAUHs.77274
Plasminogen activator, urokinase
TYMPHs.180903
Thymidine phosphorylase
SMAD4Hs.75862
SMAD family member 4
THBS1Hs.164226
Thrombospondin-1
VEGF AHs.73793
Vascular Endothelial Growth Factor A
RNA extraction and qPCR FFPE tumor sections from the initial biopsies of the 50 patients were reviewed by an expert pathologist. Thirtytwo were selected after excluding cases with insufficient tumor cells, incomplete clinical data, or unsuccessful amplification results. More than 80 % tumor cells enrichment was ensured, when necessary, by subsequent macrodissection with the use of a safety blade. Tissue sections were deparaffinated with xylol and rehydrated with downgraded alcohols. Tissues were digested with Proteinase K. RNA was extracted from 10 to 15 five-micrometer sections using the Masterpure RNA Purification Kit (Epicentre Biotechnologies) according to the manufacturer’s instructions. One microgram of total RNA was used for cDNA synthesis according to the protocol provided with the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA). Reverse transcription reactions were primed with random hexamers. qPCR was done using the ABI Prims 7900 HT Sequence Detection System (Applied Biosystems), according to manufacturer’s instructions. TLDAs were designed to include genes from published literature (18 genes). Three control genes were included in the TLDAs (GAPDH, B2M and PMSB4). Taqman assays with small amplicon sizes were used for optimal qPCR on FFPE material. The expression of each gene was measured in duplicate. Statistical analyzes Testing for differences between the gene expression profiles of primary tumors and their matched liver metastases
123
472737
was performed using Student’s t test for paired samples. Correlations between the mRNA levels of each gene in primary tumors and liver metastases were assessed using Pearson’s correlation coefficient. Statistical significance was set at a 95 % confidence level. qPCR data were normalized against the three most stable genes (GAPDH, B2M and PMSB4) prior to analysis. The stability of the expression of the candidate control genes was determined with geNorm [19].
Results We identified 50 patients with available tumor specimens and complete clinical data. In this group, qPCR analysis of both primary tumors and matched liver metastases was performed for 32 patients. This reduction in the sample size was secondary to problems in qPCR quantification and in the mRNA extraction method, which was less successful than expected. Patient’s clinical characteristics are shown in Table 2. Twelve patients (37.5 %) received 5-FU-based chemotherapy prior to liver surgery (‘‘neoadjuvant’’ treatment). 75 % patients received chemotherapy after resection of primary tumors. Chemotherapy adjuvant treatment was either a fluoropirimidin-based infusion or a combination of fluoropirimidin and oxaliplatin. The expression of 6 of the 18 genes (CXCR4, THBS1, GST Pi, TYMP, DPYD and MMP9) was significantly different in primary tumor specimens compared with their corresponding matched liver metastases. For CXCR4, TYMP and DPYD, expression was higher in the primary tumor relative to the liver metastases. In contrast, higher levels of THBS1, GST Pi and MMP 9 were detected in the metastatic liver lesions (Table 3). Two genes involved in
Clin Transl Oncol Table 2 Clinicopathological characteristics Variable
Number of patients (percentage) (n = 32)
Table 3 Difference of gene expression of genes in liver metastasis and in primary colorectal tumors Standard deviation
95 % CI
p
-1.028
1.650
-1.623 to -0.433
0.001
0.576 0.553
1.121 1.537
0.172 to 0.980 -0.008 to 1.107
0.007 0.050
D. TYMP
-0.612
1.636
-1.202 to 0.0226
0.042
E. DPYD
-0.851
1.379
-1.348 to 0.354
0.001
0.710
1.947
0.007 to 1.412
0.048
Gene
22 (68.8 %) 10 (31.3 %)
Difference in gene expression between metastasis and primary tumora
A. CXCR4
Colon
21 (65.6 %)
B. THBS1 C. GST Pi
Rectum
11 (34.4 %)
Age, years
Median 63 (range 29–83)
Gender Male Female Tumor
Clinical stage (AJCC TNM)a I
2 (6.3 %)
F. MMP 9
II
4 (12.5 %)
a
III
6 (18.8 %)
IV
20 (62.5 %)
Difference has been calculated as mRNA gene levels in the metastasis minus mRNA gene levels in the primary tumor. Only the genes showing statistically significant differences are shown
Pathological stage (AJCC TNM) I
1 (3.1 %)
II
28 (87.5 %)
III
3 (9.4 %)
Gene
Correlation coefficient (r)
p
8 (25.0 %) 17 (53.1 %)
A. SMAD4
0.447
0.010
7 (21.9 %)
B. CXCR4
0.386
0.029
C. THBS1 D. ERCC1
0.547 0.423
0.001 0.016
Adjuvant chemotherapy No Fluoropyrimidines Oxaliplatin based
Table 4 Correlation coefficients of mRNA gene levels of primary colorectal cancer and corresponding liver metastasis
ECOG 0
85 (70.2 %)
1
35 (28.9 %)
E. VEGF A
0.453
0.009
2 (0.8 %)
F. DPYD
0.473
0.006
12 (37.5 %)
Only genes with a significant correlation are shown
2 Chemotherapy prior to liver resection (‘‘neoadjuvant’’) a
Discussion
apoptosis (BAX and FAS) showed a non-significant trend towards higher expression in liver samples. No significant differences in expression were observed for the rest of the genes. Pairwise correlations were determined for the primary tumors and matched liver metastases. Correlations can be used to measure the degree of concordance between increases or decreases in gene expression in primary tumors and metastases, regardless of expression levels. Significant correlations were found for six genes: SMAD4, CXCR4, THBS1, ERCC1, VEGF A and DPYD (Table 4), CXCR4, THBS1 and DPYD showed both a significant correlation and a significant difference in expression between the two tissues. Conversely, three other genes, SMAD4, ERCC1 and VEGF A, showed significant correlations in gene expression between the primary tumors and matched liver metastases but no significant differences in expression. Therefore, only these three genes exhibited a trend towards similar expression levels in the two tissues.
Our study compared the expression of 18 genes in primary colorectal tumors and their associated liver metastases in 32 patients. We found that only SMAD4, ERCC1 and VEGF A showed similar expression levels and were positively correlated between primary tumors and their corresponding liver metastases. Among all the other genes analyzed, only CXCR4, THBS1, and DPYD (and with a marginal significance also GST Pi, TYMP and MMP9) showed a difference of mRNA expression between both tissues; however, CXCR4, THBS1 and DPYD also showed a significant correlation between both tissues. For the rest of the genes, we did not find a significant difference of gene expression between primary tumors and metastases but we could not find a significant correlation between the expressions in both tissues either. Therefore, only SMAD4, ERCC1 and VEGF A showed a similar expression in colorectal primary tumors and related colorectal liver metastases. Our results suggest that the expression of several genes might be different in liver metastases than in primary colorectal tumors, and that using only the genomic results
Clinical stage at diagnosis of primary tumors considering patients who developed metachronous liver metastases
123
Clin Transl Oncol Table 5 Series of studies of comparative analysis of genes measured in primary tumors and liver metastases Study
Genes
n
Year
Technique
Differences among primary tumors and liver metastases
Guichard et al. [29]
Topo I
Backus et al. [30]
TS, p53, Rb, FasR, FasL, bcl-2, mcl-1, bax, bcl-x, Ki-67
8
1999
Activity assay
Reduced activity (p \ 0.001)
8
2002
IHQ
Increase of TS expression (p = 0.004) and Ki-67 (p = 0.016)
TS, DPD, TP, UP, PRT
23
Decrease of Rb expression (p = 0.024) Lack of expression of Fas Different expression patterns in Bax, mcl-1 y bcl-xl
Inokuchi et al. [31]
2004
qPCR
Increase of DPD (p = 0.005) OPRT (p = 0.016), TP (p = 0.000), UP (p = 0.003) No difference of expression of TS (p = 0.280) Lack of correlation of TS (p = 0.48), DPD (p = 0.94), OPRT (p = 0.19), TP (p = 0.81), UP (p = 0.90)
Kim et al. [32]
CXCR4
29
2006
qPCR
Increase of expression (p value not found)
Ghadjar et al. [33]
CCR6
16
2006
IHQ
Reduction of expression (p = 0.02)
Kuramochi et al. [34]
VEGF
31
2006
qPCR
No expression difference (p = 0.989)
Illemann et al. [35]
MMP9
15
2006
IHQ
Different growth pattern in primary tumors and liver metastases
Kobayashi et al. [36]
TS, ERCC1
31
2008
qPCR
Correlation of TS (rs = 0.875, p = 0.0024) and ERCC1 (rs = 0.835, p = 0.0038) in primary tumors with synchronous liver metastases
Koh et al. [37]
80 genes
12
2008
Micro-arrays
Presence of correlation (rs = 0.663, p \ 0.0001)
Similar gene expression profile Different expression in individual genes
Illemann et al. [38]
uPAR, uPA, PAI-1,
14
2009
IHQ
Kuramochi et al. [39]
EGFR
31
2010
qPCR
Chen et al. [40]
c-erbB2, VEGF
44
2010
IHQ
Different growth pattern in primary tumors and liver metastases No expression difference (p = 0.99) Presence of correlation (rs = 0.78, p \ 0.001).
from primary tumor biopsies to select a treatment regime might not be the optimal approach. Previous studies have suggested that the measurement of biomarkers in metastatic tissue in addition to primary tumors should provide more information for predicting treatment response [20], and the finding of different genetic profiles in liver metastases and their primary CRC tumors has led some authors to propose the selection of targeted therapies based on the genetic properties of the metastases [21]. The study of the differences detected in gene expression during the process of carcinogenesis and metastases could also provide a better understanding of the process of progression of colorectal cancer [22, 23]. Gene expression in metastatic tumor cells might be affected by the use of chemotherapy, and it is possible that our results could have been influenced by this as 75 % of the patients had received adjuvant chemotherapy after resection of primary tumors. We detected significant differences in the expression of six genes between primary tumors and liver metastases: TYMP and DPYD—involved in 5-FU metabolism- and GST Pi—polymorphisms in this gene have been associated with increased survival of
123
No expression difference in c-erbB2 (p = 0.503) and VEGF (p = 0.285)
patients with advanced CRC receiving 5-FU/oxaliplatin chemotherapy [24]—We observed higher expression of TYMP and DPYD in primary tumors relative to liver metastases, which might suggest that primary cells would be more resistant to 5-FU chemotherapy leading subsequently to clone selection, lower expression levels in metastatic lesions, and undifferentiated liver metastases. In contrast, we found significantly higher levels of GST Pi expression in liver metastases. Considering that GST Pi has been associated with the acquisition of cisplatin resistance [25–27], liver metastases might show an increased oxaliplatin resistance compared to primary tumors. With respect to genes involved in carcinogenesis and distant metastasis, we observed significant differences in the expression of MMP 9, THBS1 and CXCR4 which would support the hypotheses that different genetic profiles are shown among primary tumors and related metastases. Among the remaining genes, where significant differences in expression could not be demonstrated, we found significant correlations between expression levels in the primary tumor and liver metastases for only SMAD4, ERCC1 and VEGF A. Therefore, only these three genes showed
Clin Transl Oncol
similar levels of expression between primary tumors and liver metastases. Our study was not designed to establish if this was secondary to changes in expression which might occur at earlier stages of CRC pathogenesis and persist during the metastatic process. Our study has several limitations. It is a retrospective single-center study, which has to be taken into account when extrapolating the results to other settings. Also, our sample was reduced from 50 to 32 patients after performing RNA extraction and qPCR, due to technical issues with qPCR amplification. These issues were related to the low quality of the tumor cDNA and the fact that we only had access to FFPE specimens: although qPCR can be performed on FFPE samples, the use of fresh or frozen tumor material is generally preferred for molecular analyzes, as this type of processing yields higher-quality tumor DNA and formalin can fragment and induce modifications in genomic DNA during the storage and fixation process [28]. As a result of the bad quality of the tissues, we did not have enough left material to perform IHC for testing protein expression and making a validation set of specimens. Seventy-five percent of our patients received adjuvant therapy after surgery of primary tumors, which might have affected the gene expression profiles, as metastatic tumor cells can be affected by chemotherapy treatment. Moreover, almost 40 % of patients received chemotherapy prior to liver resection. Our small sample size did not allow us to perform subgroup analyzes to evaluate the influence of the presence of synchronous or metachronous liver disease, the chemotherapy treatment received (fluorouracil and leucovorin versus fluorouracil, leucovorin and oxaliplatin— FOLFOX), or neoadjuvant chemotherapy on gene expression. However, all of the reviewed studies analyzing gene expression in primary CRC and metastases have included fairly small samples (Table 5) [29–40]: only one of them [40] included more patients than ours: this study analyzed two genes in a sample of 44 patients. Our results suggest that colorectal primary tumors and liver metastases can show differences in gene expression and might have the potential to behave as biologically different entities. We could not assess if this difference was secondary to chemotherapy-induced molecular changes, tumoral heterogeneity or most likely both of them. This distinction acquires special relevance with respect to potentially predictive biomarkers, and greater effort should be made to analyze biomarkers in metastases prior to therapy selection. Similar to other studies [41], our study might suggest that clinicians cannot assume that gene expression is correlated among primary tumors and liver metastases. In fact, it has been shown that although detectable genetic modifications in metastases may be retained from primary tumors, epigenetic changes in
metastases can be acquired de novo [42]. The high degree of concordance in KRAS and BRAF mutations observed between primary tumors and metastatic lesions makes genotyping of primary tumors sufficient prior to the use of anti-EGFR therapies. Nevertheless, for new mCRC therapies directed against other genes, genotyping of metastatic lesions might also be necessary to account for the heterogeneity between primary tumors and liver metastases.
Conclusion The expression levels of the genes CXCR4, THBS1, GST Pi, TYMP, DPYD and MMP9 were significantly different between primary colorectal tumors and their matched liver metastases. Among the remaining genes, where significant differences in expression could not be demonstrated, we found significant correlations between expression levels in primary tumors and liver metastases for only SMAD4, ERCC1 and VEGF A. Given that some of our patients had received adjuvant chemotherapy after resection of primary tumors and also prior to liver resection, these differences of gene expression profiles might not only be due to tumor heterogeneity but also due to the administration of chemotherapy. Nevertheless, our results suggest that that gene expression might not be necessarily correlated among primary CRC tumors and liver metastases and that genotyping potential chemotherapy resistance markers in both metastatic lesions and primary tumors could be beneficial. The development of new therapeutic agents against specific targets might therefore require gene expression profiling in liver metastases, as many genes may exhibit different expression levels in primary tumors relative to metastases. Our study could not assess whether these differences were secondary to chemotherapy-induced changes, to tumoral heterogeneity or to epigenetic changes; this could be investigated in further prospective studies. Conflict of interest All listed authors have no conflicts of interest regarding the material presented in this article.
References 1. Amado RG, Wolf M, Peeters M, Cutsem E, Siena S, Freeman DJ, et al. Wildtype KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol. 2008;26:1626–34. 2. Lie`vre A, Bachet JB, Boige V, Cayre A, Le Corre D, Buc E, et al. KRAS mutations as an independent prognostic factor in patients with advanced colorectal cancer treated with cetuximab. J Clin Oncol. 2008;26:374–9. 3. Douillard JY, Oliner KS, Siena S, Tabernero J, Burkes R, Barugel M, et al. Panitumumab-Folfox 4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013;369(11):1023–34. 4. De Roock W, Claes B, Bernasconi D, De Schutter J, Biesmans B, Fountzilas G, et al. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol. 2010;11:753–62.
123
Clin Transl Oncol 5. Sartore-Bianchi A, Martini M, Molinari F, Veronese S, Nichelatti M, Artale S, et al. PIK3CA mutations in colorectal cancer are associated with clinical resistance to EGFR-targeted monoclonal antibodies. Cancer Res. 2009;69:1851–7. 6. Oden-Gangloff A, Di Fiore F, Bibeau F, Lamy A, Bougeard G, Charbonnier F, et al. TYMP53 mutations predict disease control in metastatic colorectal cancer treated with cetuximab-based chemotherapy. Br J Cancer. 2009;100(8):1330–5. 7. Cejas P, Lo´pez-Go´mez M, Aguayo C, Madero R, Moreno-Rubio J, de Castro Carpen˜o J, et al. Analysis of the concordance in the EGFR pathway status between primary tumors and related metastases of colorectal cancer patients: implications for cancer therapy. Curr Cancer Drug Targets. 2012;12(2):124–31. 8. Miranda E, Bianchi P, Destro A, Morenghi E, Malesci A, Santoro A, et al. Genetic and epigenetic alterations in primary colorectal cancers and related lymph node and liver metastases. Cancer. 2013;119(2):266–76. 9. Zeitoun G. Cellular and molecular deregulations driving the metastatic phenotype. Med Sci (Paris) 2009, Spec No 1, p. 29–32. 10. Gerlinger M, Rowan AJ. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med. 2012;366(10):883–92. 11. Albanese I, Scibetta AG, Migliavacca M, Russo A, Bazan V, Tomasino RM, et al. Heterogeneity within and between primary colorectal carcinomas and matched metastases as revealed by analysis of Ki-ras and p53 mutations. Biochem Biophys Res Commun. 2004;325:784–91. 12. Arapantoni-Dadioti P, Valavanis C, Gavressea T, Tzaida O, Trihia H, Lekka I. Discordant expression of hormone receptors and HER2 in breast cancer. A retrospective comparison of primary tumors with paired metachronous recurrences or metastases. J BUON. 2012;17(2):277–83. 13. Jensen NF, Smith DH, Nygard SB. Predictive biomarkers with potential of converting conventional chemotherapy to targeted therapy in patients with metastatic colorectal cancer. Scand J Gastroenterol. 2012;47(3):340–55. 14. Soong RC, Sha N, Salto-Tellez M, Han HC, Ng SS, Zeps N, et al. Prognostic and predictive significance of 5-fluorouracil metabolic enzymes in colorectal cancer. J Clin Oncol, ASCO Annual Meeting Proceedings Part I, 2006, vol 24, No. 18S (June 20 Supplement). 15. Arnould S, Hennebelle I, Canal P, Bugat R, Guichard S. Cellular determinants of oxaliplatin sensitivity in colon cancer cell lines. Eur J Cancer. 2003;39(1):112–9. 16. Mathijssen RH, Loos WJ, Verweij J, Sparreboom A. Pharmacology of TOPBP1isomerase I inhibitors irinotecan (CPT-11) and TOPBP1tecan. Curr Cancer Drug Targets. 2002;2(2):103–23. 17. Schimanski C, Schwald S, Simiantonaki N, Jayasinghe C, Go¨nner U, Wilsberg V, et al. Effect of chemokine receptors CXCR4 and CCR7 on the metastatic behavior of human colorectal cancer. Clin Cancer Res. 2005;11(5):1743–50. 18. Zhang YY, Chen B, Ding YQ. Metastasis-associated factors facilitating the progression of colorectal cancer. Asian Pac J Cancer Prev. 2012;13(6):2437–44. 19. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3(7):RESEARCH0034. 20. Vermaat JS, Nijman IJ, Koudijs MJ, Gerritse FL, Scherer SF, Mokry M, et al. Primary colorectal cancers and their subsequent hepatic metastases are genetically different: implications for selection of patients for targeted treatment. Clin Cancer Res. 2012;18(3):688–99. 21. Ganepola GA, Mazziotta RM, Weeresinghe D, Corner GA, Parish CJ, Chang DH, et al. Gene expression profiling of primary and metastatic colon cancer identifies a reduced proliferative rate in metastatic tumors. Clin Exp Metastasis. 2010;27(1):1–9. 22. Iqbal S, Lenz HJ. Determinants of prognosis and response to therapy in colorectal cancer. Curr Oncol Rep. 2001;3(2):102–8. 23. Gmeiner WH, Hellmann GM, Shen Pj. Tissue-dependent and independent gene expression changes in metastatic colon cancer. Oncol Rep. 2008;19(1):245–51. 24. Stoehlmacher J, Park DJ, Zhang W. Association between glutathione S-transferase P1, T1 and M1 genetic polymorphism and survival of patients with metastatic colorectal cancer. J Natl Cancer Inst. 2002;94(12):936–42.
123
25. Ban N, Takahashi Y, Takayama T, Kura T, Katahira T, Sakamaki S, et al. Transfection of glutathione S-transferase (GST-Pi) antisense complementary DNA increases the sensitivity of a colon cancer cell line to adriamycin, cisplatin, melphalan and eTOPBP1side. Cancer Res. 1996;56(15):3577–82. 26. Nishimura T, Newkirk K, Sessions RB, Andrews PA, Trock BJ, Rasmussen AA, et al. Immunohistochemical staining for gluthatione S-transferase predicts response to platinum-based chemotherapy in head and neck cancer. Clin Cancer Res. 1996;2(11):1859–65. 27. Goto S, Iida T, Cho S, Oka M, Kohno S, Kondo T. Overexpression of gluthatione S-transferase pi enhances the adduct formation of cisplatin with gluthatione in human cancer cells. Free Radic Res. 1996;31(6):549–58. 28. Garraway Levi A. Concordance and discordance in tumor genomic profiling. J Clin Oncol. 2012;30(24):2937–9. 29. Guichard S, Terret C, Hennebelle I, Lochon I, Chevreau P, Fre´tigny E, et al. CPT-11 converting carboxylesterase and TOPBP1isomerase activities in tumor and normal colon and liver tissues. Br J Cancer. 1996;80(3–4):364–70. 30. Backus HH, Van Groeningen CJ, Vos W, Dukers DF, Bloemena E, Wouters D, et al. Differential expression of cell cycle and apoptosis related proteins in colorectal mucosa, primary colon tumors, and liver metastases. J Clin Pathol. 2002;55(3):206–11. 31. Inokuchi M, Uetake H, Shirota Y, Yamada H, Tajima M, Sugihara K. Gene expression of 5 fluorouracil metabolic enzymes in primary colorectal cancer and corresponding liver metastases. Cancer Chemother Pharmacol. 2004;53(5):391–6. 32. Kim J, Mori T, Chen SL, Amersi FF, Martinez SR, Kuo C, et al. Chemokine receptor CXCR4 expression in patients with melanoma and colorectal cancer liver metastases and the association with disease outcome. Ann Surg. 2006;244(1):113–20. 33. Ghadjar P, Coupland SE, Na IK. Chemokine receptor CCR6 expression level and liver metastases in colorectal cancer. J Clin Oncol. 2006;24(12):1910–6. 34. Kuramochi H, Hayashi K, Uchida K, Miyakura S, Shimizu D, Vallbo¨hmer D, et al. Vascular endothelial growth factor messenger RNA expression level is preserved in liver metastases compared with corresponding primary colorectal cancer. Clin Cancer Res. 2006;12(1):29–33. 35. Illemann M, Bird N, Majeed A, Sehested M, Laerum OD, Lund LR, et al. MMP9 is differentially expressed in primary human colorectal adenocarcinomas and their metastases. Mol Cancer Res. 2009;4(5):293–302. 36. Kobayashi H, Sugihara K, Uetake H, Higuchi T, Yasuno M, Enomoto M, et al. Messenger RNA expression of TS and ERCC1 in colorectal cancer and matched liver metastasis. Int J Oncol. 2008;33(6):1257–62. 37. Koh KH, Rhee H, Kang HJ, Yang E, You KT, Lee H, et al. Differential gene expression profiles of metastases in paired primary and metastatic colorectal carcinomas. Oncology. 2008;75(1–2):92–101. 38. Illemann M, Bird N, Majeed A, Laerum OD, Lund LR, Danø K, et al. Two distinct expression patterns of urokinase, urokinase receptor and plasminogen activator inhibitor-1 in colon cancer liver metastases. Int J Cancer. 2009;124(8):1860–70. 39. Kuramochi H, Hayashi K, Nakajima G, Kamikozuru H, Yamamoto M, Danenberg KD, et al. Epidermal growth factor receptor (EGFR) mRNA levels and protein expression levels in primary colorectal cancer and corresponding liver metastases. Cancer Chemother Pharmacol. 2010;65(5):825–31. 40. Chen J, Li Q, Wang C, Wu J, Zhao G. Prognostic significance of c-erbB-2 and vascular endothelial growth factor in colorectal liver metastases. Ann Surg Oncol. 2010;17(6):1555–63. 41. Vakiani E, Janakiraman M, Shen R, Sinha R, Zeng Z, Shia J, et al. Comparative genomic analysis of primary versus metastatic colorectal carcinomas. J Clin Oncol. 2012;30(24):2956–62. 42. Miranda E, Destro A, Malesci A, Balladore E, Bianchi P, Baryshnikova E, et al. Genetic and epigenetic changes in primary metastatic and nonmetastatic colorectal cancer. Br J Cancer. 2006;95(8):1101–7.
RESEARCH ARTICLE
Gene expression differences in primary colorectal tumors and matched liver metastases: chemotherapy related or tumoral heterogeneity? M. Lo´pez-Go´mez • J. Moreno-Rubio • I. Sua´rez-Garcı´a • P. Cejas • R. Madero • E. Casado • A. M. Jime´nez • M. Sereno • C. Go´mez-Raposo F. Zambrana • M. Merino • D. Ferna´ndez-Luengas • J. Feliu
•
Received: 26 May 2014 / Accepted: 20 September 2014 Ó Federacio´n de Sociedades Espan˜olas de Oncologı´a (FESEO) 2014
Abstract Background Treatment of metastatic colorectal cancer (mCRC) is generally based on genetic testing performed in primary tumor biopsies, but whether the genomic status of primary tumors is identical to that of metastases is not well known. We compared the gene expression profiles of formalin-fixed paraffin-embedded (FFPE) biopsies of colorectal primary tumors and matched liver metastases. Patients and methods We compared the expression of 18 genes in FFPE CRC tumors and their matched liver All authors have agreed to the submission and have participated in the study. M. Lo´pez-Go´mez (&) E. Casado A. M. Jime´nez M. Sereno C. Go´mez-Raposo F. Zambrana M. Merino Medical Oncology Department, Infanta Sofı´a University Hospital, Paseo de Europa 34, San Sebastia´n de los Reyes, 28702 Madrid, Spain e-mail: [email protected] E. Casado e-mail: [email protected] A. M. Jime´nez e-mail: [email protected] M. Sereno e-mail: [email protected] C. Go´mez-Raposo e-mail: [email protected]
metastases from 32 patients. The expression of each gene in CRC primary tumors and their matched liver metastases was tested using Student’s t test for paired samples. Pairwise correlations of each gene in the primary tumors and matched liver metastases were evaluated by Pearson’s correlation coefficient. Results The expression of six genes was significantly different in primary tumors compared with their matched liver metastases [CXCR4 (p \ 0.001), THBS1 (p = 0.007), MMP 9 (p = 0.048), GST Pi (p = 0.050), TYMP (p = 0.042) and DPYD (p \ 0.001)]. For the remaining genes, where no significant differences were observed, only SMAD4 (rs = 0.447, p = 0.010), ERCC1 (rs = 0.423, I. Sua´rez-Garcı´a Internal Medicine Department, Infanta Sofı´a University Hospital, Paseo de Europa 34, San Sebastia´n de los Reyes, 28702 Madrid, Spain e-mail: [email protected] P. Cejas Translational Oncology Department, La Paz University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected] R. Madero Statistics Department, La Paz University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected]
M. Merino e-mail: [email protected]
D. Ferna´ndez-Luengas General Surgery Department, Infanta Sofı´a University Hospital, Paseo de Europa 34, San Sebastia´n de los Reyes, 28702 Madrid, Spain e-mail: [email protected]
J. Moreno-Rubio Precision Oncology Laboratory (POL), Infanta Sofı´a University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected]
J. Feliu Medical Oncology Department, La Paz University Hospital, P8 de la Castellana 261, 28046 Madrid, Spain e-mail: [email protected]
F. Zambrana e-mail: [email protected]
123
Clin Transl Oncol
p = 0.016) and VEGF A (rs = 0.453, p = 0.009) showed significant correlation in expression between the two tissues. Therefore, we only detected similar gene expression levels between the tumor and the metastases in these three markers. Conclusions We only found similar gene expression levels between the tumor and the metastases in three genes (SMAD4, ERCC1, and VEGF A). However, our study could not assess whether the differences in gene expression were secondary to tumoral heterogeneity or to molecular changes induced by previous chemotherapy.
TLDA TOP 1 TYMP TYMS THBS1 VEGF A
Keywords Colon cancer Liver metastasis Predictive biomarkers Prognostic factors
Basing treatment strategies for metastatic colorectal cancer (mCRC) on the results of biomarker assays have been widely incorporated into daily practice [1, 2]. Information not only on KRAS but also in NRAS mutational status is now considered an essential prerequisite for the selection of a targeted therapy based on anti-EGFR drugs. [3]. Moreover, several studies have suggested that mutations in other genes such as BRAF, PIK3CA and p53, are also associated with resistance to anti-EGFR therapies [4–6] The genomic status of these predictive genes is usually tested in primary tumor tissues because material from metastatic sites is not routinely collected and because the latter is generally of poorer quality and harbors fewer tumor cells as a result of tumor necrosis or chemotherapyinduced changes. However, it is unclear whether the genomic status of a given gene will be the same in the primary tumor and its associated metastases. Some studies performed in mCRC patients have suggested that the mutational status of genes such as KRAS, BRAF and PIK3CA in primary tumors serves as an adequate surrogate of the gene’s status in the metastases, whereas other genes, such as EGFR or PTEN show a higher degree of discordance [7]. The development of distant metastases is a dynamic and a multistep process involving several genetic and epigenetic changes [8] in which tumor cells acquire a more aggressive phenotype [9]. For this reason, the importance of genetic heterogeneity across different tumor foci or even within the same tumor site has been widely evaluated in recent years. Genetic intratumor heterogeneity can contribute to treatment failure and drug resistance. For instance, in kidney cancer, tumor heterogeneity occurs not only across distant tumor foci but also within the same tumor site [10]. Furthermore, we often observe that different metastasis sites show different clinical responses to chemotherapy. However, the mechanism by which tumor cells acquire differential chemotherapeutic sensitivity during metastasis is not well understood. Previous studies have reported discordance between the genetic expression profiles of colorectal primary tumors and their corresponding metastases and have consequently
Abbreviations 5FU 5-Fluorouracil B2M Beta 2 microglobulin BAX Bcl-2 associated X protein BRAF v-Raf murine sarcoma viral oncogene homolog B1 CCR 6 Chemokine receptor 6 CXCR4 Chemokine receptor 4 DNA Deoxyribonucleic acid DPYD Dihydropyrimidine dehydrogenase E-cadherin Cell adhesion promoter EGFR Epidermal growth factor receptor ERCC1 Excision repair cross-complementing factor 1 FAS/CD95 TNF factor superfamily member 6 FFPE Formalin fixed paraffin embedded FOLFIRI Chemotherapy scheme including 5-FU leucovorin and irinotecan FOLFOX Chemotherapy scheme including 5-FU leucovorin and oxaliplatin GAPDH Glyceraldehyde 3-phosphate dehydrogenase reference gene GST Pi Glutathione S-Transferase Pi h TERT Human telomerase reverse transcriptase IHC Immunohistochemistry KRAS v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog mCRC Metastatic colorectal cancer MMP-9 Matrix metalloproteinase 9 p p value pAKT Phosphorylated v-akt murine thymoma viral oncogene homolog 1 PCR Polymerase chain reaction PIK3CA Phosphoinositide-3-Kinase, catalytic, alpha polypeptide PLAU Urokinase-type plasminogen PTEN Phosphatase and tensin homolog qPCR quantitative Polymerase chain reaction RNA Ribonucleic Acid
123
Taq low density arrays Topoisomerase I Thymidine phosphorylase Thymidylate synthase Thrombospondin 1 Vascular endothelial growth factor
Introduction
Clin Transl Oncol
suggested that metastatic tissues should be tested [11]. In other tumors, such as breast cancer, the expression of certain biomarkers (e.g., estrogen receptor, progesterone receptor and human epidermal growth factor receptor 2) has been reported to change with disease progression. Because of this discordance, some researchers have proposed mandatory evaluation of biomarkers in breast metastases prior to the selection of a targeted therapy [12]. A variety of chemotherapy regimens have been used for the treatment of mCRC, with the most common agents being fluoropyrimidines [5-fluorouracil (5-FU) and capecitabine] used in combination with either oxaliplatin (FOLFOX) or irinotecan (FOLFIRI). However, chemotherapy resistance is often observed, limiting the effectiveness of these drugs. Overcoming drug resistance has become one of the main challenges of current research [13]. Information on the expression profiles of genes related to response to chemotherapy at different disease stages might be particularly useful for the selection of treatments to avoid resistance. For example, TYMS (Thymidylate Synthase), TYMP (Thymidine Phosphorylase) and DPYD (Dihydropyrimidine Dehydrogenase) are enzymes involved in 5-FU metabolism [14]. Oxaliplatin, in turn, is detoxified by glutathione-related enzymes, such as glutathione S-transferase Pi (GST Pi), and forms platinum DNA adduct lesions that are repaired by enzymes of the nucleotide excision repair system (NER), including ERCC1 (Excision Repair Cross-Complementing factor 1) [15]. Finally, irinotecan targets DNA topoisomerase I, thereby inhibiting DNA replication [16]. Tumor metastasis is a multistage process involving proteolysis, motility, migration, proliferation, inhibition of apoptosis and neoangiogenesis. Determining the expression levels of specific tumor metastasis-associated genes or proteins responsible for colorectal metastasis at different stages of the disease might help elucidate the molecular mechanisms involved in the development of distant metastases. Chemokines (e.g., CXCR4 and CCR6), cell adhesion molecules (e.g., matrix metalloproteases and E-cadherin), plasminogen activator system-related genes (e.g., uPAR), apoptosis promoters (e.g., BAX, FAS and SMAD4) and angiogenic factors (e.g., VEGF A and thrombospondin) all play important roles in the development of distant metastases [17, 18]. The question of whether primary colorectal tumors and their associated liver metastases in a specific patient have similar genetic profiles, or they behave as biologically distinct entities, is still unsolved. Research that investigates this problem might help inform the reliability of the strategy of using information from markers evaluated in primary tumors for determining a course of treatment in metastatic disease. The aim of our study was to compare
the gene expression profiles of primary colorectal tumors and associated liver metastases to determine whether they show similar gene expression patterns.
Materials and methods Patient selection We performed a retrospective study in the Oncology Department of La Paz University Hospital in Madrid, Spain. We included patients diagnosed with CRC resected synchronous liver metastases (diagnosed at the same time of primary tumors) or metachronous liver metastases (diagnosed weeks or months after the diagnoses of primary tumors), between the years 2003 and 2007. We included all diagnosed patients with available clinical data who had undergone surgery for both primary and liver disease and for whom formalin-fixed paraffin-embedded (FFPE) tissues from both primary and metastatic tumor specimens were available. Patients who had received locoregional treatment for liver metastases (radiofrequency or chemoembolization) instead of surgery were excluded. No unresectable patients were included. Clinical information was collected retrospectively from the patients’ medical records. Both in the case of synchronous and metachronous liver metastases, liver specimens were collected several months after the patients had undergone liver surgery. In the case of patients who had received treatment prior to liver resection—neoadjuvant treatment—tissue collection was made once they had finished chemotherapy treatment as well. Institutional approval from the ethics committee was obtained for the study. Gene selection We performed a literature review of all the published studies of genes involved in both the development of liver metastases and chemotherapy resistance. An electronic search was performed on the MEDLINE database from 1980 to 2007 using the MeSH headings ‘‘colorectal cancer’’, ‘‘liver metastases’’, ‘‘liver resection’’, ‘‘primary tumor’’, ‘‘chemotherapy resistance’’, ‘‘carcinogenesis’’, ‘‘metastases spread’’ and ‘‘tumoral markers’’. The search was limited to English-language publications. All titles and abstracts were reviewed, and appropriate papers were further assessed. Case reports, editorials, abstracts and reviews were excluded. 11 genes related to carcinogenesis and metastases spread and 7 genes related to mCRC chemotherapy resistance were selected. The selected genes are shown in Table 1.
123
Clin Transl Oncol Table 1 Genes determined by qPCR Genes related to carcinogenesis and metastases spread
Genes related to chemotherapy resistance
Description
Geneunigene
BAXHs.624291
BCL2-associated X protein
DPYDHs.335034
Dihydropyrimidine dehydrogenase
CCR 6Hs.34526
Chemokine (C-X-C motif) receptor 6
CDH1Hs.461086
Cadherin 1, type 1, E-cadherin (epithelial)
ERCC1Hs.435981
Excision repair cross-complementing rodent repair deficiency, complementation group 1 Glutathione S-transferase Pi
Geneunigene
Number
Number
Description
CXCR4Hs.593413
Chemokine (C-X.C motif) receptor 4
GST PiHs.523836
FAS/CD95Hs.244139
TNF factor superfamily, member 6
KRASHs.505033
V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog Topoisomerase I
h TERTHs.492203
Human Telomerase Reverse Transcriptase
TOP1Hs.
MMP 9Hs.297413
Matrix Metalloproteinase
TYMSHs.369762
Thymidylate synthase
PLAUHs.77274
Plasminogen activator, urokinase
TYMPHs.180903
Thymidine phosphorylase
SMAD4Hs.75862
SMAD family member 4
THBS1Hs.164226
Thrombospondin-1
VEGF AHs.73793
Vascular Endothelial Growth Factor A
RNA extraction and qPCR FFPE tumor sections from the initial biopsies of the 50 patients were reviewed by an expert pathologist. Thirtytwo were selected after excluding cases with insufficient tumor cells, incomplete clinical data, or unsuccessful amplification results. More than 80 % tumor cells enrichment was ensured, when necessary, by subsequent macrodissection with the use of a safety blade. Tissue sections were deparaffinated with xylol and rehydrated with downgraded alcohols. Tissues were digested with Proteinase K. RNA was extracted from 10 to 15 five-micrometer sections using the Masterpure RNA Purification Kit (Epicentre Biotechnologies) according to the manufacturer’s instructions. One microgram of total RNA was used for cDNA synthesis according to the protocol provided with the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA). Reverse transcription reactions were primed with random hexamers. qPCR was done using the ABI Prims 7900 HT Sequence Detection System (Applied Biosystems), according to manufacturer’s instructions. TLDAs were designed to include genes from published literature (18 genes). Three control genes were included in the TLDAs (GAPDH, B2M and PMSB4). Taqman assays with small amplicon sizes were used for optimal qPCR on FFPE material. The expression of each gene was measured in duplicate. Statistical analyzes Testing for differences between the gene expression profiles of primary tumors and their matched liver metastases
123
472737
was performed using Student’s t test for paired samples. Correlations between the mRNA levels of each gene in primary tumors and liver metastases were assessed using Pearson’s correlation coefficient. Statistical significance was set at a 95 % confidence level. qPCR data were normalized against the three most stable genes (GAPDH, B2M and PMSB4) prior to analysis. The stability of the expression of the candidate control genes was determined with geNorm [19].
Results We identified 50 patients with available tumor specimens and complete clinical data. In this group, qPCR analysis of both primary tumors and matched liver metastases was performed for 32 patients. This reduction in the sample size was secondary to problems in qPCR quantification and in the mRNA extraction method, which was less successful than expected. Patient’s clinical characteristics are shown in Table 2. Twelve patients (37.5 %) received 5-FU-based chemotherapy prior to liver surgery (‘‘neoadjuvant’’ treatment). 75 % patients received chemotherapy after resection of primary tumors. Chemotherapy adjuvant treatment was either a fluoropirimidin-based infusion or a combination of fluoropirimidin and oxaliplatin. The expression of 6 of the 18 genes (CXCR4, THBS1, GST Pi, TYMP, DPYD and MMP9) was significantly different in primary tumor specimens compared with their corresponding matched liver metastases. For CXCR4, TYMP and DPYD, expression was higher in the primary tumor relative to the liver metastases. In contrast, higher levels of THBS1, GST Pi and MMP 9 were detected in the metastatic liver lesions (Table 3). Two genes involved in
Clin Transl Oncol Table 2 Clinicopathological characteristics Variable
Number of patients (percentage) (n = 32)
Table 3 Difference of gene expression of genes in liver metastasis and in primary colorectal tumors Standard deviation
95 % CI
p
-1.028
1.650
-1.623 to -0.433
0.001
0.576 0.553
1.121 1.537
0.172 to 0.980 -0.008 to 1.107
0.007 0.050
D. TYMP
-0.612
1.636
-1.202 to 0.0226
0.042
E. DPYD
-0.851
1.379
-1.348 to 0.354
0.001
0.710
1.947
0.007 to 1.412
0.048
Gene
22 (68.8 %) 10 (31.3 %)
Difference in gene expression between metastasis and primary tumora
A. CXCR4
Colon
21 (65.6 %)
B. THBS1 C. GST Pi
Rectum
11 (34.4 %)
Age, years
Median 63 (range 29–83)
Gender Male Female Tumor
Clinical stage (AJCC TNM)a I
2 (6.3 %)
F. MMP 9
II
4 (12.5 %)
a
III
6 (18.8 %)
IV
20 (62.5 %)
Difference has been calculated as mRNA gene levels in the metastasis minus mRNA gene levels in the primary tumor. Only the genes showing statistically significant differences are shown
Pathological stage (AJCC TNM) I
1 (3.1 %)
II
28 (87.5 %)
III
3 (9.4 %)
Gene
Correlation coefficient (r)
p
8 (25.0 %) 17 (53.1 %)
A. SMAD4
0.447
0.010
7 (21.9 %)
B. CXCR4
0.386
0.029
C. THBS1 D. ERCC1
0.547 0.423
0.001 0.016
Adjuvant chemotherapy No Fluoropyrimidines Oxaliplatin based
Table 4 Correlation coefficients of mRNA gene levels of primary colorectal cancer and corresponding liver metastasis
ECOG 0
85 (70.2 %)
1
35 (28.9 %)
E. VEGF A
0.453
0.009
2 (0.8 %)
F. DPYD
0.473
0.006
12 (37.5 %)
Only genes with a significant correlation are shown
2 Chemotherapy prior to liver resection (‘‘neoadjuvant’’) a
Discussion
apoptosis (BAX and FAS) showed a non-significant trend towards higher expression in liver samples. No significant differences in expression were observed for the rest of the genes. Pairwise correlations were determined for the primary tumors and matched liver metastases. Correlations can be used to measure the degree of concordance between increases or decreases in gene expression in primary tumors and metastases, regardless of expression levels. Significant correlations were found for six genes: SMAD4, CXCR4, THBS1, ERCC1, VEGF A and DPYD (Table 4), CXCR4, THBS1 and DPYD showed both a significant correlation and a significant difference in expression between the two tissues. Conversely, three other genes, SMAD4, ERCC1 and VEGF A, showed significant correlations in gene expression between the primary tumors and matched liver metastases but no significant differences in expression. Therefore, only these three genes exhibited a trend towards similar expression levels in the two tissues.
Our study compared the expression of 18 genes in primary colorectal tumors and their associated liver metastases in 32 patients. We found that only SMAD4, ERCC1 and VEGF A showed similar expression levels and were positively correlated between primary tumors and their corresponding liver metastases. Among all the other genes analyzed, only CXCR4, THBS1, and DPYD (and with a marginal significance also GST Pi, TYMP and MMP9) showed a difference of mRNA expression between both tissues; however, CXCR4, THBS1 and DPYD also showed a significant correlation between both tissues. For the rest of the genes, we did not find a significant difference of gene expression between primary tumors and metastases but we could not find a significant correlation between the expressions in both tissues either. Therefore, only SMAD4, ERCC1 and VEGF A showed a similar expression in colorectal primary tumors and related colorectal liver metastases. Our results suggest that the expression of several genes might be different in liver metastases than in primary colorectal tumors, and that using only the genomic results
Clinical stage at diagnosis of primary tumors considering patients who developed metachronous liver metastases
123
Clin Transl Oncol Table 5 Series of studies of comparative analysis of genes measured in primary tumors and liver metastases Study
Genes
n
Year
Technique
Differences among primary tumors and liver metastases
Guichard et al. [29]
Topo I
Backus et al. [30]
TS, p53, Rb, FasR, FasL, bcl-2, mcl-1, bax, bcl-x, Ki-67
8
1999
Activity assay
Reduced activity (p \ 0.001)
8
2002
IHQ
Increase of TS expression (p = 0.004) and Ki-67 (p = 0.016)
TS, DPD, TP, UP, PRT
23
Decrease of Rb expression (p = 0.024) Lack of expression of Fas Different expression patterns in Bax, mcl-1 y bcl-xl
Inokuchi et al. [31]
2004
qPCR
Increase of DPD (p = 0.005) OPRT (p = 0.016), TP (p = 0.000), UP (p = 0.003) No difference of expression of TS (p = 0.280) Lack of correlation of TS (p = 0.48), DPD (p = 0.94), OPRT (p = 0.19), TP (p = 0.81), UP (p = 0.90)
Kim et al. [32]
CXCR4
29
2006
qPCR
Increase of expression (p value not found)
Ghadjar et al. [33]
CCR6
16
2006
IHQ
Reduction of expression (p = 0.02)
Kuramochi et al. [34]
VEGF
31
2006
qPCR
No expression difference (p = 0.989)
Illemann et al. [35]
MMP9
15
2006
IHQ
Different growth pattern in primary tumors and liver metastases
Kobayashi et al. [36]
TS, ERCC1
31
2008
qPCR
Correlation of TS (rs = 0.875, p = 0.0024) and ERCC1 (rs = 0.835, p = 0.0038) in primary tumors with synchronous liver metastases
Koh et al. [37]
80 genes
12
2008
Micro-arrays
Presence of correlation (rs = 0.663, p \ 0.0001)
Similar gene expression profile Different expression in individual genes
Illemann et al. [38]
uPAR, uPA, PAI-1,
14
2009
IHQ
Kuramochi et al. [39]
EGFR
31
2010
qPCR
Chen et al. [40]
c-erbB2, VEGF
44
2010
IHQ
Different growth pattern in primary tumors and liver metastases No expression difference (p = 0.99) Presence of correlation (rs = 0.78, p \ 0.001).
from primary tumor biopsies to select a treatment regime might not be the optimal approach. Previous studies have suggested that the measurement of biomarkers in metastatic tissue in addition to primary tumors should provide more information for predicting treatment response [20], and the finding of different genetic profiles in liver metastases and their primary CRC tumors has led some authors to propose the selection of targeted therapies based on the genetic properties of the metastases [21]. The study of the differences detected in gene expression during the process of carcinogenesis and metastases could also provide a better understanding of the process of progression of colorectal cancer [22, 23]. Gene expression in metastatic tumor cells might be affected by the use of chemotherapy, and it is possible that our results could have been influenced by this as 75 % of the patients had received adjuvant chemotherapy after resection of primary tumors. We detected significant differences in the expression of six genes between primary tumors and liver metastases: TYMP and DPYD—involved in 5-FU metabolism- and GST Pi—polymorphisms in this gene have been associated with increased survival of
123
No expression difference in c-erbB2 (p = 0.503) and VEGF (p = 0.285)
patients with advanced CRC receiving 5-FU/oxaliplatin chemotherapy [24]—We observed higher expression of TYMP and DPYD in primary tumors relative to liver metastases, which might suggest that primary cells would be more resistant to 5-FU chemotherapy leading subsequently to clone selection, lower expression levels in metastatic lesions, and undifferentiated liver metastases. In contrast, we found significantly higher levels of GST Pi expression in liver metastases. Considering that GST Pi has been associated with the acquisition of cisplatin resistance [25–27], liver metastases might show an increased oxaliplatin resistance compared to primary tumors. With respect to genes involved in carcinogenesis and distant metastasis, we observed significant differences in the expression of MMP 9, THBS1 and CXCR4 which would support the hypotheses that different genetic profiles are shown among primary tumors and related metastases. Among the remaining genes, where significant differences in expression could not be demonstrated, we found significant correlations between expression levels in the primary tumor and liver metastases for only SMAD4, ERCC1 and VEGF A. Therefore, only these three genes showed
Clin Transl Oncol
similar levels of expression between primary tumors and liver metastases. Our study was not designed to establish if this was secondary to changes in expression which might occur at earlier stages of CRC pathogenesis and persist during the metastatic process. Our study has several limitations. It is a retrospective single-center study, which has to be taken into account when extrapolating the results to other settings. Also, our sample was reduced from 50 to 32 patients after performing RNA extraction and qPCR, due to technical issues with qPCR amplification. These issues were related to the low quality of the tumor cDNA and the fact that we only had access to FFPE specimens: although qPCR can be performed on FFPE samples, the use of fresh or frozen tumor material is generally preferred for molecular analyzes, as this type of processing yields higher-quality tumor DNA and formalin can fragment and induce modifications in genomic DNA during the storage and fixation process [28]. As a result of the bad quality of the tissues, we did not have enough left material to perform IHC for testing protein expression and making a validation set of specimens. Seventy-five percent of our patients received adjuvant therapy after surgery of primary tumors, which might have affected the gene expression profiles, as metastatic tumor cells can be affected by chemotherapy treatment. Moreover, almost 40 % of patients received chemotherapy prior to liver resection. Our small sample size did not allow us to perform subgroup analyzes to evaluate the influence of the presence of synchronous or metachronous liver disease, the chemotherapy treatment received (fluorouracil and leucovorin versus fluorouracil, leucovorin and oxaliplatin— FOLFOX), or neoadjuvant chemotherapy on gene expression. However, all of the reviewed studies analyzing gene expression in primary CRC and metastases have included fairly small samples (Table 5) [29–40]: only one of them [40] included more patients than ours: this study analyzed two genes in a sample of 44 patients. Our results suggest that colorectal primary tumors and liver metastases can show differences in gene expression and might have the potential to behave as biologically different entities. We could not assess if this difference was secondary to chemotherapy-induced molecular changes, tumoral heterogeneity or most likely both of them. This distinction acquires special relevance with respect to potentially predictive biomarkers, and greater effort should be made to analyze biomarkers in metastases prior to therapy selection. Similar to other studies [41], our study might suggest that clinicians cannot assume that gene expression is correlated among primary tumors and liver metastases. In fact, it has been shown that although detectable genetic modifications in metastases may be retained from primary tumors, epigenetic changes in
metastases can be acquired de novo [42]. The high degree of concordance in KRAS and BRAF mutations observed between primary tumors and metastatic lesions makes genotyping of primary tumors sufficient prior to the use of anti-EGFR therapies. Nevertheless, for new mCRC therapies directed against other genes, genotyping of metastatic lesions might also be necessary to account for the heterogeneity between primary tumors and liver metastases.
Conclusion The expression levels of the genes CXCR4, THBS1, GST Pi, TYMP, DPYD and MMP9 were significantly different between primary colorectal tumors and their matched liver metastases. Among the remaining genes, where significant differences in expression could not be demonstrated, we found significant correlations between expression levels in primary tumors and liver metastases for only SMAD4, ERCC1 and VEGF A. Given that some of our patients had received adjuvant chemotherapy after resection of primary tumors and also prior to liver resection, these differences of gene expression profiles might not only be due to tumor heterogeneity but also due to the administration of chemotherapy. Nevertheless, our results suggest that that gene expression might not be necessarily correlated among primary CRC tumors and liver metastases and that genotyping potential chemotherapy resistance markers in both metastatic lesions and primary tumors could be beneficial. The development of new therapeutic agents against specific targets might therefore require gene expression profiling in liver metastases, as many genes may exhibit different expression levels in primary tumors relative to metastases. Our study could not assess whether these differences were secondary to chemotherapy-induced changes, to tumoral heterogeneity or to epigenetic changes; this could be investigated in further prospective studies. Conflict of interest All listed authors have no conflicts of interest regarding the material presented in this article.
References 1. Amado RG, Wolf M, Peeters M, Cutsem E, Siena S, Freeman DJ, et al. Wildtype KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol. 2008;26:1626–34. 2. Lie`vre A, Bachet JB, Boige V, Cayre A, Le Corre D, Buc E, et al. KRAS mutations as an independent prognostic factor in patients with advanced colorectal cancer treated with cetuximab. J Clin Oncol. 2008;26:374–9. 3. Douillard JY, Oliner KS, Siena S, Tabernero J, Burkes R, Barugel M, et al. Panitumumab-Folfox 4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013;369(11):1023–34. 4. De Roock W, Claes B, Bernasconi D, De Schutter J, Biesmans B, Fountzilas G, et al. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol. 2010;11:753–62.
123
Clin Transl Oncol 5. Sartore-Bianchi A, Martini M, Molinari F, Veronese S, Nichelatti M, Artale S, et al. PIK3CA mutations in colorectal cancer are associated with clinical resistance to EGFR-targeted monoclonal antibodies. Cancer Res. 2009;69:1851–7. 6. Oden-Gangloff A, Di Fiore F, Bibeau F, Lamy A, Bougeard G, Charbonnier F, et al. TYMP53 mutations predict disease control in metastatic colorectal cancer treated with cetuximab-based chemotherapy. Br J Cancer. 2009;100(8):1330–5. 7. Cejas P, Lo´pez-Go´mez M, Aguayo C, Madero R, Moreno-Rubio J, de Castro Carpen˜o J, et al. Analysis of the concordance in the EGFR pathway status between primary tumors and related metastases of colorectal cancer patients: implications for cancer therapy. Curr Cancer Drug Targets. 2012;12(2):124–31. 8. Miranda E, Bianchi P, Destro A, Morenghi E, Malesci A, Santoro A, et al. Genetic and epigenetic alterations in primary colorectal cancers and related lymph node and liver metastases. Cancer. 2013;119(2):266–76. 9. Zeitoun G. Cellular and molecular deregulations driving the metastatic phenotype. Med Sci (Paris) 2009, Spec No 1, p. 29–32. 10. Gerlinger M, Rowan AJ. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med. 2012;366(10):883–92. 11. Albanese I, Scibetta AG, Migliavacca M, Russo A, Bazan V, Tomasino RM, et al. Heterogeneity within and between primary colorectal carcinomas and matched metastases as revealed by analysis of Ki-ras and p53 mutations. Biochem Biophys Res Commun. 2004;325:784–91. 12. Arapantoni-Dadioti P, Valavanis C, Gavressea T, Tzaida O, Trihia H, Lekka I. Discordant expression of hormone receptors and HER2 in breast cancer. A retrospective comparison of primary tumors with paired metachronous recurrences or metastases. J BUON. 2012;17(2):277–83. 13. Jensen NF, Smith DH, Nygard SB. Predictive biomarkers with potential of converting conventional chemotherapy to targeted therapy in patients with metastatic colorectal cancer. Scand J Gastroenterol. 2012;47(3):340–55. 14. Soong RC, Sha N, Salto-Tellez M, Han HC, Ng SS, Zeps N, et al. Prognostic and predictive significance of 5-fluorouracil metabolic enzymes in colorectal cancer. J Clin Oncol, ASCO Annual Meeting Proceedings Part I, 2006, vol 24, No. 18S (June 20 Supplement). 15. Arnould S, Hennebelle I, Canal P, Bugat R, Guichard S. Cellular determinants of oxaliplatin sensitivity in colon cancer cell lines. Eur J Cancer. 2003;39(1):112–9. 16. Mathijssen RH, Loos WJ, Verweij J, Sparreboom A. Pharmacology of TOPBP1isomerase I inhibitors irinotecan (CPT-11) and TOPBP1tecan. Curr Cancer Drug Targets. 2002;2(2):103–23. 17. Schimanski C, Schwald S, Simiantonaki N, Jayasinghe C, Go¨nner U, Wilsberg V, et al. Effect of chemokine receptors CXCR4 and CCR7 on the metastatic behavior of human colorectal cancer. Clin Cancer Res. 2005;11(5):1743–50. 18. Zhang YY, Chen B, Ding YQ. Metastasis-associated factors facilitating the progression of colorectal cancer. Asian Pac J Cancer Prev. 2012;13(6):2437–44. 19. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3(7):RESEARCH0034. 20. Vermaat JS, Nijman IJ, Koudijs MJ, Gerritse FL, Scherer SF, Mokry M, et al. Primary colorectal cancers and their subsequent hepatic metastases are genetically different: implications for selection of patients for targeted treatment. Clin Cancer Res. 2012;18(3):688–99. 21. Ganepola GA, Mazziotta RM, Weeresinghe D, Corner GA, Parish CJ, Chang DH, et al. Gene expression profiling of primary and metastatic colon cancer identifies a reduced proliferative rate in metastatic tumors. Clin Exp Metastasis. 2010;27(1):1–9. 22. Iqbal S, Lenz HJ. Determinants of prognosis and response to therapy in colorectal cancer. Curr Oncol Rep. 2001;3(2):102–8. 23. Gmeiner WH, Hellmann GM, Shen Pj. Tissue-dependent and independent gene expression changes in metastatic colon cancer. Oncol Rep. 2008;19(1):245–51. 24. Stoehlmacher J, Park DJ, Zhang W. Association between glutathione S-transferase P1, T1 and M1 genetic polymorphism and survival of patients with metastatic colorectal cancer. J Natl Cancer Inst. 2002;94(12):936–42.
123
25. Ban N, Takahashi Y, Takayama T, Kura T, Katahira T, Sakamaki S, et al. Transfection of glutathione S-transferase (GST-Pi) antisense complementary DNA increases the sensitivity of a colon cancer cell line to adriamycin, cisplatin, melphalan and eTOPBP1side. Cancer Res. 1996;56(15):3577–82. 26. Nishimura T, Newkirk K, Sessions RB, Andrews PA, Trock BJ, Rasmussen AA, et al. Immunohistochemical staining for gluthatione S-transferase predicts response to platinum-based chemotherapy in head and neck cancer. Clin Cancer Res. 1996;2(11):1859–65. 27. Goto S, Iida T, Cho S, Oka M, Kohno S, Kondo T. Overexpression of gluthatione S-transferase pi enhances the adduct formation of cisplatin with gluthatione in human cancer cells. Free Radic Res. 1996;31(6):549–58. 28. Garraway Levi A. Concordance and discordance in tumor genomic profiling. J Clin Oncol. 2012;30(24):2937–9. 29. Guichard S, Terret C, Hennebelle I, Lochon I, Chevreau P, Fre´tigny E, et al. CPT-11 converting carboxylesterase and TOPBP1isomerase activities in tumor and normal colon and liver tissues. Br J Cancer. 1996;80(3–4):364–70. 30. Backus HH, Van Groeningen CJ, Vos W, Dukers DF, Bloemena E, Wouters D, et al. Differential expression of cell cycle and apoptosis related proteins in colorectal mucosa, primary colon tumors, and liver metastases. J Clin Pathol. 2002;55(3):206–11. 31. Inokuchi M, Uetake H, Shirota Y, Yamada H, Tajima M, Sugihara K. Gene expression of 5 fluorouracil metabolic enzymes in primary colorectal cancer and corresponding liver metastases. Cancer Chemother Pharmacol. 2004;53(5):391–6. 32. Kim J, Mori T, Chen SL, Amersi FF, Martinez SR, Kuo C, et al. Chemokine receptor CXCR4 expression in patients with melanoma and colorectal cancer liver metastases and the association with disease outcome. Ann Surg. 2006;244(1):113–20. 33. Ghadjar P, Coupland SE, Na IK. Chemokine receptor CCR6 expression level and liver metastases in colorectal cancer. J Clin Oncol. 2006;24(12):1910–6. 34. Kuramochi H, Hayashi K, Uchida K, Miyakura S, Shimizu D, Vallbo¨hmer D, et al. Vascular endothelial growth factor messenger RNA expression level is preserved in liver metastases compared with corresponding primary colorectal cancer. Clin Cancer Res. 2006;12(1):29–33. 35. Illemann M, Bird N, Majeed A, Sehested M, Laerum OD, Lund LR, et al. MMP9 is differentially expressed in primary human colorectal adenocarcinomas and their metastases. Mol Cancer Res. 2009;4(5):293–302. 36. Kobayashi H, Sugihara K, Uetake H, Higuchi T, Yasuno M, Enomoto M, et al. Messenger RNA expression of TS and ERCC1 in colorectal cancer and matched liver metastasis. Int J Oncol. 2008;33(6):1257–62. 37. Koh KH, Rhee H, Kang HJ, Yang E, You KT, Lee H, et al. Differential gene expression profiles of metastases in paired primary and metastatic colorectal carcinomas. Oncology. 2008;75(1–2):92–101. 38. Illemann M, Bird N, Majeed A, Laerum OD, Lund LR, Danø K, et al. Two distinct expression patterns of urokinase, urokinase receptor and plasminogen activator inhibitor-1 in colon cancer liver metastases. Int J Cancer. 2009;124(8):1860–70. 39. Kuramochi H, Hayashi K, Nakajima G, Kamikozuru H, Yamamoto M, Danenberg KD, et al. Epidermal growth factor receptor (EGFR) mRNA levels and protein expression levels in primary colorectal cancer and corresponding liver metastases. Cancer Chemother Pharmacol. 2010;65(5):825–31. 40. Chen J, Li Q, Wang C, Wu J, Zhao G. Prognostic significance of c-erbB-2 and vascular endothelial growth factor in colorectal liver metastases. Ann Surg Oncol. 2010;17(6):1555–63. 41. Vakiani E, Janakiraman M, Shen R, Sinha R, Zeng Z, Shia J, et al. Comparative genomic analysis of primary versus metastatic colorectal carcinomas. J Clin Oncol. 2012;30(24):2956–62. 42. Miranda E, Destro A, Malesci A, Balladore E, Bianchi P, Baryshnikova E, et al. Genetic and epigenetic changes in primary metastatic and nonmetastatic colorectal cancer. Br J Cancer. 2006;95(8):1101–7.
