Abstract
This study attempts to determine whether primary tumor tissue could reliably represent metastatic colorectal cancer in therapy-guiding analysis of mitochondrial microsatellite instability. Therefore, we investigated the concordance of microsatellite instability in D310, D514, and D16184 (mitochondrial DNA displacement loop), and its association with selected clinical categories and KRAS/NRAS/BRAF/PIK3CA/TP53 mutation status between primary and metastatic colorectal cancer tissue from 119 patients. Displacement loop microsatellite instability was significantly more frequently seen in lymph node metastases (53.1%) compared to primary tumors (37.5%) and distant metastases (21.4%) (p = 0.0183 and p = 0.0005). The discordant rate was significantly higher in lymph node metastases/primary tumor pairs (74.6%) than in distant metastases/primary tumor pairs (52.4%) or lymph node metastases/distant metastases pairs (51.6%) (p = 0.0113 and p = 0.0261) with more gain (86.7%) than loss (61.1%) of microsatellite instability in the discordant lymph node metastases (p = 0.0024). Displacement loop instability occurred significantly more frequently in lymph node metastases and distant metastases of patients with early colorectal cancer onset age <60 years (p = 0.0122 and p = 0.0129), was found with a significant high rate in a small cohort of TP53-mutated distant metastases (p = 0.0418), and was associated with TP53 wild-type status of primary tumors (p = 0.0009), but did not correlate with KRAS, NRAS, BRAF, or PIK3CA mutations. In conclusion, mitochondrial microsatellite instability and its association with selected clinical and molecular markers are discordant in primary and metastatic colorectal cancer, which could have importance for surveillance and therapeutic strategies.
Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the fourth leading cause of cancer-related death worldwide. 1 In addition to surgery as primary treatment, 5-fluorouracil (FU)-based chemotherapy and blocking of epidermal growth factor receptor (EGFR) have become fundamental tools to reduce recurrence in patients with stage III CRC and increase overall survival in patients with stage IV CRC, respectively. 2 However, despite improved patient selection by use of predictive biomarkers, a proportion of individuals gain little or even no benefit from these therapies. 3 Therefore, the spectrum of prognostic and therapy response markers needs being extended. The majority of studies searching for genetic chemosensitivity markers focused mainly on alterations in the nuclear DNA (nDNA) of cancer cells, even if mitochondrial DNA (mtDNA) has been shown to be more susceptible to sequence variation than nDNA.4,5 mtDNA is a 16,569-base pair (bp), double-stranded, circular DNA composed of genes and a noncoding region, the displacement loop (D-loop), located between nucleotides 16,024 and 576, which contains essential transcription and replication elements. 6 In colon cancer, mutations and microsatellite instability (MSI) of the mtDNA have been found in 8%–70% of the investigated cohorts.6–23 Furthermore, mtDNA alterations have been proven as prognostic marker, 6 indicator of resistance to FU-based6,24 and cisplatin chemotherapy, 24 and characteristic of cancer cells, which can be targeted by next-generation metal-based therapy as the recently published organometallic “half sandwich” compound [Os(η(6)-p-cymene)(4-(2-pyridylazo)-N,N-dimethylaniline)I]PF6. 25 The future potential combined diagnostic function of mtDNA alterations to identify tumor cells, which are resistant to conventional cytotoxic drugs and/or responsive to new therapeutic approaches, requires to elucidate its association with other therapeutic markers, which, for example, are related to proliferation, apoptosis, and the EGFR pathway, and to define preanalytic conditions for reliable diagnostic routine. In the case of EGFR inhibitor therapy, methodically powerful studies26–28 determined that primary CRC sufficiently represent metastatic disease for mutation analysis in EGFR pathway regulating genes. However, the literature provides only little information about the heterogeneity of mtDNA alterations, comprising comparison of mtDNA mutations in oral cancer and derived lymph node metastases (LN), 29 mitochondrial genomic variation in prostate cancer and its distant metastases (DM) 30 and mtDNA copy number in primary and lymphatic metastatic esophagus cancer. 31 Furthermore, it is currently not known, whether associations between mtDNA alteration and clinical or nDNA markers are maintained during the metastatic process of CRC. To address this issue, we analyzed mitochondrial microsatellite instability (mtMSI) and its association to selected clinical and nDNA markers in 120 primary tumors (PT), 113 LN, and 42 DM of CRC.
Materials and methods
Tissue sampling and selection
Formalin-fixed paraffin-embedded (FFPE) CRC samples (120 PT, 42 DM, 113 LN sample mixes, comprising between one and eight LN per case) and 119 nonneoplastic FFPE samples from 119 patients were collected from the tissue archive (1999–2005) at the Department of Pathology, Southern Norwegian Hospital Trust, Kristiansand. The material was partly included in a previous study. 32 Tissue and patient data were obtained and used after approval of the Regional Ethics Committee (REK) of Southern Norway in accordance with the declaration of Helsinki and the International Conference of Harmonization—Good Clinical Practice. The anonymity of the patients investigated was preserved corresponding to rules of data protection of the National Data Protection Commission (NSD) of Norway and the institutional guidelines of our hospital. All tumor samples underwent histopathologic review (B.K.). Only material containing less than 20% necrosis and less than 20% nonneoplastic adherent tissue was included in this study. Tumor response to treatment was classified according to the Response Evaluation Criteria in Solid Tumors (RECIST). 33
Immunohistochemical analysis
Analysis of Bcl-2 and Ki-67 expression is described in detail in the supplementary text material. At least moderate cytoplasmic or perinuclear staining in >10% of tumor cells was required to define Bcl-2 positivity. High Ki-67 expression was defined at an arbitrary cut off level of >30% stained tumor cells.
Molecular genetic analysis
Mutation status of KRAS, NRAS, BRAF, PIK3CA, and TP53 of the tumor tissue has already been determined in a previous study. 32 Description of DNA isolation and molecular genetic analysis of these five genes is added as supplementary material (text and Table S1).
Alterations in the mtDNA D-loop were assessed by PCR and direct sequencing of the whole control region (nt16024–nt576) in three parts using 0.3 to 0.8 ng DNA in a 12.5-µL sample volume with 15 mM Tris/HCL, 50 mM KCL, with 200-µM dNTPs, 2.5 µM MgCl2, 0.1 nM primer, 10% dimethyl sulfoxide (DMSO), and 0.5 Units AmpliTaq Gold DNA Polymerase™ (Applied Biosystems, Darmstadt, Germany). Cycle-sequencing PCR of PCR products was carried out with the BigDye Terminator kit (Applied Biosystems) in accordance with the manufacturer’s instructions in both directions and evaluated on an ABI 310 Genetic Analyzer by two independent observers. mtDNA haplogroups were determined using HaploGrep. 34
Statistical analysis
Data were analyzed using the chi-square test and the Fisher’s exact test (GraphPad QuickCalcs). 35 A p-value of less than 0.05 was considered as statistically significant.
Results
Clinical data of the patients, histopathological characteristics, and molecular genetic data of the PT, lymph node, and distant metastatic sites are listed in Table 1. According to a definition by Mekenkamp et al., 36 18 DM were synchronous (onset within 6 months after primary diagnosis of CRC) and 24 DM were metachronous (onset beyond 6 months after primary diagnosis). Postoperative chemotherapy was applied in 78 (65.5%) patients, but no patient received preoperative chemotherapy.
Molecular genetic and clinicopathological data in primary and metastatic colorectal cancer.
mtDNA: mitochondrial DNA; MSS: microsatellite stability; MSI: microsatellite instability.
Number of patients.
Number of tumor samples.
Significant differences marked gray: cp = 0.0122; dp = 0.0129; ep = 0.0275; fp = 0.047; gp = 0.0009; hp = 0.0284; ip = 0.0418.
Several known mtDNA D-loop single nucleotide polymorphisms could be detected in the patients, but PT and metastatic tissue of individual patients displayed an identical mtDNA haplotype. Mitochondrial haplogroups (established from D-loop variants) and comparison with clinicopathological and mutational data are summarized in supplementary Table S2. D-loop instability (example displayed in Figure 1) was significantly more frequently seen in LN than in PT and DM, considering both the number of included patients (p = 0.0248 and p = 0.0082) or the number of investigated tumor samples (p = 0.0183 and p = 0.0005). The distribution of mtMSI in the three different loci is displayed in Table 1. Instability of D310 occurred together with instability of D514 in one PT and one LN and together with instability of D16184 in two PT and one LN. The discordance rate was significantly higher in LN/PT pairs (74.6%) than in DM/PT pairs (52.4%) or LN/DM pairs (51.6%) (p = 0.0113 and p = 0.0261) with significant more gain (86.7%) than loss (61.1%) of mutations in the discordant lymph nodes (p = 0.0024) (Figure 2). Thirteen of the discordant DM were metachronous, five of those were biopsied or resected after application of chemotherapy. Two of 10 patients with multiple DM showed discordant mtMSI status of the metastases. mtMSI was also discordant in two synchronous PT of one patient.
Electropherogram displaying mtMSI in D310 after direct sequencing. (a) Normal tissue, (b) primary tumor, (c) lymph node metastasis. The C(n) status changes from C7TC6 in normal tissue to C8TC6 in the primary tumor to C9TC6 in the metastasis.
Concordance and discordance rates of mtMSI in matched primary tumor (PT)/lymph node metastases (LN) pairs (A), primary tumor (PT)/distant metastases (DM) pairs (B), and lymph node metastases (LN)/distant metastases (DM) pairs (C). (+) indicates mtMSI positive tissue, (−) indicates mtMSI negative tissue.
D-loop instability showed a trend to occur more frequently in PT of patients with late CRC onset age ⩾60 years (82.2%, p = 0.09) and was significantly more frequently seen in lymph node and DM of patients with early CRC onset age <60 years (p = 0.0122 and p = 0.0129), in PT with loss of Bcl-2 and high Ki-67 expression (p = 0.0275 and p = 0.047), in TP53-mutated DM (p = 0.0418), and exclusively in TP53 wild-type PT (p = 0.0009), but did not correlate with gender, tumor location, clinical stage, pTNM stage, tumor differentiation, survival, or KRAS, NRAS, BRAF, or PIK3CA mutations, neither in primary nor in metastatic CRC.
Altogether, mitochondrial haplogroups H* (1.7%), H1 (8.4%), H2 (21.8%), H3 (0.8%), H4 (1.7%), H5 (2.5%), H6 (0.8%), H11 (0.8%), H13 (0.8%), I* (1.7%), I2 (4.2%), J1 (10.9%), J2 (2.5%), R (1,7%), T1 (2.5%), T2 (8.4%), U3 (2.5%), U4 (0.8%), U5 (11.8%), U8 (1.7%), X (3.4%), K* (0.8%), K1 (2.5%), K2 (0.8%), V (2.5%), and W (1.7%) were found in the patients of this study. Comparing different mitochondrial haplogroups (displayed in supplementary Table S2), group J showed a significant low number of cases with high Ki-67 expression (p = 0.0385), group U showed absence, and group K the highest rate of TP53 mutations (p = 0.0232 and p = 0.033). Haplogroups did not correlate with other clinicopathological parameters, mutations in the other investigated genes, or survival of the patients.
Discussion
This study analyzed mtMSI in the D-loop in primary and metastatic CRC tissue and determined the mtDNA haplogroup of CRC patients (based on sequencing data of the D-loop region). The frequencies of haplogroups showed mostly only little deviation from published data for a Norwegian population, 37 with two exceptions. Haplogroup H, the predominant haplogroup for Northern and Western Europe 38 occurred only in 40%, whereas the rather rare haplogroup I (average frequency in Europe 2% or lower) was found in 6% of patients. This can probably be explained by the low number of individuals analyzed. However, if these results could be confirmed in larger CRC patient cohorts, they point to mitochondrial haplogroup I as a risk factor for this tumor type.
The proportion of mtMSI detected in this study was within the published range of frequencies (24.1%–44.4%),8,14,16,19,20,22 considering all investigated regions of the D-loop (comprising D310, D514, and D16184) or only the region D310. In contrast, our study revealed a lower mtMSI rate for the loci D514 and D16184 than reported previously for PT 22 (7.4% and 18.6%, respectively). Less data are available for metastatic CRC. The detection rate of mitochondrial instability for the LN of this study was in accordance with a recent publication on metastatic CRC 23 considering all three investigated loci (54.4%) or D16184 only (1.6%), negligibly lower than reported by these authors for D310 only (50.5%) and little higher than reported in the case of D514 only (4.9%). In contrast, the rate of mtMSI in DM of our patient cohort was less than half of that reported by others. 23 However, these authors have summarized data for hepatic and extrahepatic metastatic sites without mentioning, whether LN were included in their study. According to the literature, the widely ranging frequency of mitochondrial genomic variation between several studies could be due to varying age compositions of the investigated populations,39,40 surgical stress induced in possibly adherent normal colon tissue, 41 different epithelium-stroma ratios of the tumors, 42 varying proportions of tumors with advanced stages,23,42,43 and by differences in the applied definitions and diagnostic panels. 23 Direct sequencing, which was applied in the current study, has a limited detection level, and interpretation of data might be complicated by heteroplasmy, 23 especially, if the portion of tumor DNA is less than 20% of analyzed DNA. We tried to avoid as much as possible “contamination” of tumor samples by normal tissue, but we were not able to standardize the epithelium-stroma ratio of the investigated CRC. This study included only stage III and IV tumors with unknown nuclear MSI status, which has probably no influence on the mtMSI rate,6,23 but on the mtDNA copy number. 44
In our CRC cohort, the rate of mtMSI was significantly higher in LN than in PT, which is in line with a recent study on oral squamous cell carcinomas (SCC), 29 but in contrast to a previous investigation of esophageal SCC. 31 Furthermore, we detected, with marginal significance, mtMSI less frequently in DM than in PT, which is in contrast to data of Arnold et al., 30 who demonstrated higher mitochondrial genomic variation in DM compared to primary prostate cancer. However, these authors also described site-specific variation of mtDNA mutation rates within their metastatic prostate cancer cohort, which is in line with our finding of significantly higher mtMSI rates in lymph node compared to distant metastase. The significantly higher mtMSI discordance rate for PT/LN pairs compared to PT/DM pairs and LN/DM pairs and the fact that LN gained mtDNA instability significantly more than losing it compared to the PT could point to an adaptive mechanism. Adaptation by mitochondrial genomic variation could support “survival of the fittest” 30 in the early metastatic process, where the selective pressure is possibly higher than later in the metastatic cascade.
In this study, mtMSI showed a trend to occur more frequently in PT of patients with CRC onset age ⩾60 years, but the mtMSI status did not differ significantly between several age groups. This is in line with mtDNA mutation data for CRC.6,15 However, we detected mtMSI significantly more frequently in lymph node and distant metastases of patients with early onset age <60 years. To the best of our knowledge, this association has not been previously reported. Age-related different potential of tumor cells to adapt to selective pressure during the metastatic process might be one reason for this link, but definitive explanation of this finding requires more detailed basic research. In accordance with other data on mtDNA alteration,6,15 this study could not reveal a link between mtMSI and gender, tumor stage, or tumor location.
The varying detection rates of mtMSI could be one reason for contrary published data about the association of mtDNA with other molecular and immunohistochemical markers. Mutations of the genes KRAS, NRAS, BRAF, and PIK3CA, respectively, did not correlate with any haplogroup or D-loop instability, neither in primary nor in metastatic CRC. Although mtDNA copy number was shown to be significantly lower in BRAF mutated and higher in KRAS mutated CRC 44 and basic research demonstrated constitutive activation of the PI3K/Akt pathway by mtDNA mutations, 45 no other studies focusing on mtMSI and mutational status of KRAS, NRAS, BRAF, and PIK3CA could be found. We were not able to demonstrate a relationship between mtMSI and TP53 mutations in LN, which is in line with the data of Tsai et al., 20 but found an association between TP53 mutations and mtMSI in DM despite small sample size, which is in accordance with data of Chang et al. 18 However, we could compare our data for metastatic CRC only with studies focusing on primary CRC, because a link between TP53 and mtMSI in metastases has not been elucidated so far. In contrast to our results for lymph node and DM and to the above-named references, we found a significant correlation between wild-type TP53 and mtDNA instability in primary CRC, which has not been previously reported. Contrary data about the link between mtDNA and TP53 alterations in the literature could be partially attributed to the large variety of reported mtDNA alteration frequencies. Furthermore, our approach only to search for known mutations in selected exons of the TP53 gene might have influenced our results. In addition, mtMSI correlated significantly with loss of Bcl-2 expression in our primary CRC. The link of altered mtDNA, to both TP53 wild-type status and probably downregulated Bcl-2 protein expression in CRC, could point to an intact intrinsic apoptotic pathway, either despite of the types of mtDNA alterations found here or caused by these via a currently unknown mechanism. Preliminary results of basic research demonstrated that mutant mtDNA promotes apoptotic resistance45,46 and that TP53 enhances base excision repair, 47 thus suggesting a link between wild-type TP53 and intact mtDNA genome. According to our data, the role of mtMSI in the apoptotic pathway could be different from the role of mtDNA mutations.
To the best of our knowledge, a link between proliferative activity of CRC and mitochondrial genomic variation has not been investigated so far. Our study revealed a trend to correlation between mtMSI and higher proliferation determined by Ki-67 protein expression, which is in line with results from a study on breast cancer. 48 Immunohistochemical analysis of Bcl-2 and Ki-67 could not be performed for the metastases of this study, because of exhausted material after initial molecularpathological analyses. Further evaluation of a possible link between mtDNA and proliferation as well as apoptosis in larger CRC cohorts including metastases could be useful, because the biological background for occasionally reported poor prognosis of mtDNA-mutated CRC 6 is still poorly understood. In contrast to this study, neither mtMSI status in primary CRC nor in metastatic CRC correlated with survival of our patient cohort.
Nevertheless, we were able to demonstrate that mitochondrial D-loop variation and its association with selected clinical and molecular markers are discordant in primary and metastatic CRC. Therefore, metastatic tissue should be considered as preferential material for mtDNA analysis in a therapeutic or prognostic context, even if the definition of a diagnostic standard requires further studies.
Footnotes
Acknowledgements
The authors thank M. Tellefsen and C. Loland for their excellent technical assistance.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Ethical approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This is a retrospective study and for this type of study formal consent is not required.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
