SEARCH

SEARCH BY CITATION

Keywords:

  • mitochondrial DNA;
  • esophageal cancer;
  • hypervariable region;
  • D-loop;
  • mutation

Abstract

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Recent studies of various cancers, such as those of the breast, head and neck, bladder and lung, reported that 46–64% of somatic mutations in the D-loop region of mitochondrial DNA (mtDNA) are observed. However, in esophageal cancer, only a low rate (5%) of somatic mutations has so far been reported in one article (Hibi, K. et al., Int J Cancer 2001;92:319–321). Thus, to confirm this we analyzed the somatic mutations for hypervariable regions (HVR-I and HVR-II) in the D-loop of mtDNA to reevaluate the possibility of mitochondrial genetic instability in this cancer. We amplified both HVRs by PCR and DNA samples obtained from 38 esophageal tumors and matched normal tissues, and then sequenced them. Comparing the sequences of tumors to those of normal tissues, we found 14 somatic mutations in 13 patients (34.2%). Eleven mutations were at the C consecutive stretch from position 303 to 309 of MITOMAP in the mitochondria databank (http://www.mitomap.org/), 1 at position 215 in HVR-II and 2 at positions 16,304 and 16,324 in HVR-I. There were 41 types of germ line variations in HVR-I including 2 not so far recorded in the mtDNA databank and 17 in HVR-II including 1 not yet recorded. We also determined nuclear genome instability of these 38 specimens by analyzing 3 independent microsatellite sequences. While 4 specimens showed a single microsatellite change, which is tumor specific, we did not find any co-relation between a somatic mtDNA mutation and microsatellite instability of nuclear genome DNA. These results suggest that mtDNA mutations might show a genetic instability in esophageal cancer independently from a nuclear genome instability. © 2003 Wiley-Liss, Inc.

Mitochondria are cytoplasmic organelles in which cells produce energy as ATP through oxidative phosphorylation. Each cell contains up to 1,000 mitochondria, and each mitochondrion harbors a few copies of mitchondrial DNA (mtDNA), which is a double-stranded circular DNA.1 Human mtDNA consists of 16,569 nucleotides in which 37 genes code 2 rRNAs, 22 tRNAs and 13 polypeptides.2

While the mutation rate of mtDNA is different for each region (e.g., the coding region, D-loop region etc.), the overall rate is at least 10 times higher than that of nuclear genomic DNA.3, 4 The following reasons for this high mutation rate have been proposed: 1) mtDNA does not have histone proteins that play a protective role against damage,5 2) mitochondria possess only limited DNA repair systems,6 and 3) reactive oxygen species are continuously generated in mitochondria during the formation of ATP through oxidative phosphorylation.7 A mutation in mtDNA expands either partially (heteroplasmy) or totally replaces all mtDNA (homoplasmy). However, it is still unclear how mutated mtDNA expands in cells.

mtDNA mutations were recently reported in several human cancers such as pancreatic,8 gastric,9, 10 colon,11 ovarian,12 head and neck and bladder and lung cancers.13 Most somatic mutations in these cancers were found in the D-loop region, which is a long noncoding region that contains the leading strand for origin of replication and the major promoters for transcription.14 Two hypervariable regions (HVRs) in the D-loop are identified from position 16,024 to 16,324 in MITOMAP (HVR-I), and from position 63 to 322 in MITOMAP (HVR-II).15, 16

Esophageal cancer is one of the commonest and most aggressive cancers in the world. Multistage progression of esophageal cancer advances from a normal epithelium to basal cell hyperplasia, dysplasia and finally to invasive carcinoma. During this process, a series of genetic changes occurs in dominant oncogenes such as cyclin D117 and tumor suppressor genes such as p5318 and p16.19, 20 It was recently reported that in esophageal cancers mutations occur at a low rate (5%) in the D-loop region of mtDNA.21 Since this is the only study in the literature that analyzes mtDNA mutations in esophageal cancer, we attempted to analyze the genetic alterations in HVR-I and HVR-II of mtDNA by using direct sequencing of a PCR product, to reevaluate the possibility of mitochondrial genetic instability in esophageal cancer.

MATERIAL AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Subjects

Thirty-eight esophageal tumors and matched normal tissues were obtained from patients who were operated on at the Aichi Cancer Center Hospital between September 2000 and August 2002. Age and sex distributions of patients were shown in Table I. Written informed consent for genetic analyses was obtained from all patients before surgery. Tissues were frozen until extraction of DNA. Genomic DNA with mtDNA was isolated from those frozen tissues using a DNeasy Tissue kit (QIAGEN, Hilden, Germany) with proteinase K. Total DNA was quantified and diluted to 50 ng/μl for PCR reaction.

Table I. Age and sex distribution among patients
 No. (%)
TotalSomatic mutationNo somatic mutation
Age (yr)   
 –491 (2.6)01
 50–5912 (31.6)57
 60–6916 (42.1)610
 70–9 (23.7)27
 Mean62.6 ± 6.961.7 ± 5.763.0 ± 7.6
Sex   
 Male29 (76.3)920
 Female9 (23.7)45
 Total 381325

Mutation and variation analysis of HVRs in D-loop of mtDNA

Both HVRs in the D-loop of mtDNA were amplified by PCR according to Wilson et al.22 with slight modifications. The upstream primer, 5′-CACCATTAGCACCCAAAGCT-3′ (position 15,978 [RIGHTWARDS ARROW] 15,997 in MITOMAP), and the downstream primer, 5′-TGATTTCACGGAGGATGGTG-3′ (16,401 [LEFTWARDS ARROW] 16,420), were used for amplification of HVR-I, while the upstream primer, 5′-CTCACGGGAGCTCTCCATGC-3′ (29 [RIGHTWARDS ARROW] 48), and the downstream primer, 5′-GACTGTTAAAAGTGCATACCGC-3′ (410 [LEFTWARDS ARROW] 431), were used for amplification of HVR-II. The PCR was performed with 35 cycles of denaturation at 94°C for 30 sec, annealing at 60°C for 30 sec and extension at 72°C for 30 sec. Amplified fragments were then purified by QIAquick (QIAGEN). Sequencing was performed using the same primers as for PCR with an ABI PRISM 310 Genetic Analyzer (Perkin-Elmer, Foster, CA).

Analysis of microsatellite alteration in nuclear genome DNA

To determine microsatellite instability of nuclear genome, we employed an ABI PRISM Linkage Mapping Set (Applied Biosystems, Foster City, CA). Primer sets of 3 microsatellite markers, D9S171, D10S195 and D13S175 which were labeled by NED, FAM and NED fluorescent dye, respectively, were obtained from Applied Biosystems. We analyzed these microsatellite sequences in our study because previous studies reported rather frequent changes of them in esophageal cancer23, 24 or head and neck cancer.25 The PCR was performed in 7.5 μl of a reaction mixture containing 3 sets of primers for amplification of the microsatellite sequences with AmpliTaq Gold DNA Polymerase (Applied Biosystems). The PCR condition consisted of an initial activation of AmpliTaq Gold DNA Polymerase at 95°C for 12 min followed by 10 cycles of denaturation at 94°C for 15 sec, annealing at 55°C for 15 sec and extension at 72°C for 30 sec, and then 30 cycles of denaturation at 89°C for 15 sec, annealing at 55°C for 15 sec and extension at 72°C for 30 sec. PCR products were mixed with 12 μl of formamide and 0.5 μl of GS-500 rox size standard. After denaturation at 95°C for 2 min followed by immediate cooling, samples were analyzed by an ABI PRISM 310 Genetic Analyzer.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

We sequenced the HVR-I and HVR-II regions of mtDNA from 38 primary esophageal tumors and matched normal tissues. The sequences were compared to those in the mtDNA databank (http://www.mitomap.org/). Forty-one types of genetic variations in the HVR-I region and 17 types in the HVR-II region were found in the 38 normal tissues (Table II). Among these variations, 2 in HVR-I and 1 in HVR-II were not yet recorded in the databank described above or previous reports.26, 27 Tumor-specific somatic mutations were also found in the HVR-I region of 2 specimens and in the HVR-II region of 12 others. While double somatic mutations were found in 1 specimen (case 74), 12 specimens had a single somatic mutation (Table III). Eleven somatic mutations were found from position 303 to 309 in the consecutive C stretch, which is reported to be highly variable.11

Table II. Germline variations in hypervariable region of mtDNA in esophageal cancer
PositionBase variationCambridge sequenceFrequencyPositionBase variationCambridge sequenceFrequencyPositionBase variationCambridge sequenceFrequency
  • 1

    Not recorded in mtDNA databank.

681AG116,075CT116,232AC1
941AG116,093CT416,245TC2
146CT316,129AG716,249CT2
150TC716,140CT116,257AC1
152CT716,145AG116,261TC2
191insA116,172CT216,274AG1
194TC316,174TC116,278TC2
195CT116,176TC116,290TC7
199CT416,182CA416,294TC2
207AG116,183CA616,295TC1
235GA516,1831GA116,298CT1
249delA116,184TC216,304CT2
2941CT116,187TC516,309GA1
303–309delC7116,189CT1016,311CT1
 insCC71716,193insC216,3181GA1
 insC, C2C7716,209CT316,319AG9
 insC2C7216,216GA116,324CT3
    16,217CT116,325CT3
    16,223TC3516,344TC1
    16,227GA116,357CT1
        16,362CT23
Table III. Somatic mutations in hypervariable region of mtDNA in esophageal cancer
Case numberPositionBase changeCambridge sequencePattern1
  1. Homo, homoplasmic; hetero, heteroplasmic.

87215A [RIGHTWARDS ARROW] GAHomo [RIGHTWARDS ARROW] homo
72303–309C8,9 [RIGHTWARDS ARROW] C7C7Hetero [RIGHTWARDS ARROW] homo
74303–309C8 [RIGHTWARDS ARROW] C7C7Homo [RIGHTWARDS ARROW] homo
77303–309C8 [RIGHTWARDS ARROW] C8,9C7Homo [RIGHTWARDS ARROW] hetero
80303–309C8,9 [RIGHTWARDS ARROW] C9C7Hetero [RIGHTWARDS ARROW] homo
83303–309C8 [RIGHTWARDS ARROW] C7C7Homo [RIGHTWARDS ARROW] homo
89303–309C8,9 [RIGHTWARDS ARROW] C9C7Hetero [RIGHTWARDS ARROW] homo
91303–309C7 [RIGHTWARDS ARROW] C7,8C7Homo [RIGHTWARDS ARROW] hetero
92303–309C8,9 [RIGHTWARDS ARROW] C7C7Hetero [RIGHTWARDS ARROW] homo
97303–309C8 [RIGHTWARDS ARROW] C7C7Homo [RIGHTWARDS ARROW] homo
99303–309C8 [RIGHTWARDS ARROW] C7,8C7Homo [RIGHTWARDS ARROW] hetero
102303–309C8 [RIGHTWARDS ARROW] C7,8C7Homo [RIGHTWARDS ARROW] hetero
9016,304C [RIGHTWARDS ARROW] TTHomo [RIGHTWARDS ARROW] homo
7416,324C [RIGHTWARDS ARROW] TTHomo [RIGHTWARDS ARROW] homo

When we analyzed microsatellite instability of nuclear genome at the D9S171, D10S197 and D13S175 loci of these 38 primary esophageal tumors, 4 specimens exhibited microsatellite changes at a single locus. Three of them did not exhibit any mtDNA somatic mutations, whereas the other exhibited a frame shift mutation in mtDNA (Table IV).

Table IV. Somatic mutations in mtDNA and microsatellite changes in nuclear DNA
Nuclear genome instability mtDNA DNA mutation
 Frame shiftBase changeNone
Microsatellite change+103
 9322
Total 10325

The mean age of patients with somatic mutations was 61.7 ± 5.7, while that of patients without them was 63.0 ± 7.6 (Table I), suggesting that somatic mutations may not be due to the effect of aging. Nine male and 4 female patients had a somatic mutation (Table I). No significant difference in the rate of patients with somatic mutations was found between male and female patients. Smoking status were obtained from 35 out of 38 patients and drinking status were from 34 out of 38. Eight of 25 (32%) smokers and 3 of 10 (30%) nonsmokers had somatic mutations, and 7 of 26 (27%) drinkers and 3 of 8 (38%) nondrinkers had somatic mutations. Since statistically significant relationship between either smoking nor drinking and mtDNA instability was found, smoking or drinking status did not affect whether patients had somatic mutations in mtDNA.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Several studies have been conducted on somatic mutations of mtDNA in tumor tissues. Two of 45 gastric tumors (4%) had mutations in the D-loop of mtDNA.10 Heerdt et al.28 did not find any mutations in the D-loop region of mtDNA, whereas Polyak et al.29 found somatic mutations in 7 of 10 (70%) human colorectal tumor cell lines. Three of 15 ovarian cancers (20%) carried single or multiple mutations in the D-loop region, and 6 of 10 tumors (60%) were found to be positive for somatic mutations by a complete sequence analysis of mtDNA.12 It was recently reported that only a small number of esophageal cancers (2 of 37, 5%) had somatic mutations.21 In contrast, in the present study we have found that 13 of 38 (34.2%) esophageal cancers had somatic mutations, including 11 at the consecutive C stretch from position 303 to 309. It is possible that some characteristics of our patients such as age and clinical stage might differ from those in the study by Hibi et al.21 Our results suggest that the consecutive C stretch from position 303 to 309 at HVR-II in the D-loop region might be mutable in cancer cells, and that these mutations might be the result of a mitochondrial genome instability (mtGI) in esophageal cancer. On the other hand, it might be possible that minor allele pre-existed in normal cells expanded in tumor cells. However, in our study we were not able to determine which was the case, since we only had tumor specimens with surrounding normal tissues.

The consecutive C stretch from position 303 to 309 in HVR-II exhibited genetic diversity with 6 to 9 nucleotides (Tables II and V), and frequent somatic mutations were also detected in this region (Table III), which is known to be a “hot spot” of germline variations.11 On the other hand, the consecutive C stretch from position 311 to 315, which is downstream of the C stretch from 303 to 309, exhibited no genetic diversity at all nor any somatic mutations. In HVR-I, there are other consecutive C stretches from position 16,184 to 16,193 interrupted by T at 16,569. While no somatic mutations were found, this region displayed 7 types of genetic variations (Table V). In addition to these regions, several homopolymeric or dipolymeric tracts exist in mtDNA.30, 31, 32, 33, 34 It would be intriguing to analyze these homo- or di-polymeric tracts to evaluate whether esophageal cancer patients also exhibit genetic variations and/or somatic mutations. If somatic mutations were found, it would strongly support our notion that esophageal cancers exhibit mtGI.

Table V. Variations in consecutive C stretch in HVR-I and HVR-II
PositionVariationNumberPositionVariationNumber
303–309C6116,184–16,193C5TC422
 C711 C103
 C817 C114
 C8, C97 C122
 C92 C3TCTC45
    TC4TC41
    TC91

While tumors exhibited a homoplasmy, matched normal tissues exhibited a heteroplasmy in 4 patients (Table III). It is so far unclear whether these changes have a selective advantage, since a number of consecutive C stretch was different from each patient in tumor. It might be a result of clonal expansion in 2 patients (cases 80 and 89 in Table III) since 8 or 9 bases consecutive C stretch in normal tissues as a heteroplasmy were changed to only 9 bases consecutive C stretch in tumors as a homoplasmy. Also, it might be a result of mtGI in 2 other patients (cases 72 and 92 in Table III) since 8 and 9 bases consecutive C stretch in normal tissues as a heteroplasmy were changed to 7 bases consecutive C stretch in tumors as a homoplasmy.

Forty-one of 58 (70.7%) germline variations (Table II) were found in HVR-I; on the other hand, only 2 of 14 (14.3%) somatic mutations (Table III) were in HVR-I. The fact that germline variations include germline mutations and polymorphisms and somatic mutations include only new mutations in esophageal tumor cells in this case may cause this difference. Furthermore, a hot spot of somatic mutations existing at position 303–309 in HVR-II may also cause the reduced rate of somatic mutations in HVR-I.

In nuclear genome DNA, it is generally accepted that microsatellite instability in patients is due to a defect in mismatch repair activity.35 On the other hand, it was demonstrated in yeast cells that MSH1 protein, which is one of the mismatch repair related proteins, was present in mitochondria and that it may play a role in the mismatch repair of mitochondria genome DNA.36 However, it is still not known whether a mismatch repair system exists in the mitochondria of higher eukaryote cells.6 Although mtGI was detected in several human cancers,37 no correlation between mtGI and nuclear genome instability (nGI) was found in either breast cancer32 or colorectal cancer.29 A correlation between nGI and mtGI has so far been found only in gastric cancer,38 suggesting that the repair system in mitochondria might be different from that in nuclei. We also could not find any correlation between somatic mutations in mtDNA and nGI in our esophageal tumor specimens (Table IV). While nGI has not been frequently found in esophageal cancer,39 further analysis is necessary to determine the mechanisms that induce mtGI in esophageal cancers.

Liu et al.12 reported that among 6 somatic mutations in mtDNA, only 2 (33%) were located in the D-loop region in ovarian cancer. Fliss et al.13 also reported that 14 of 20 (70%) mutations in bladder cancer, 3 of 9 (33%) in head and neck cancer and 3 of 10 (30%) in lung cancer were located outside of the D-loop region. Since we analyzed only the HVR-I and HVR-II regions in the D-loop of mtDNA, it is possible that other mutations may exist in the D-loop as well as in the entire mtDNA. If this is the case, we may be underestimating the rate of somatic mutations. It would seem important to analyze the whole mtDNA in esophageal cancers to accurately determine the exact rate of somatic mutations.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES