Heterogeneous microsatellite instability observed within epithelium of ulcerative colitis
Version of Record online: 23 AUG 2006
Copyright © 2006 Wiley-Liss, Inc.
International Journal of Cancer
Volume 119, Issue 11, pages 2513–2519, 1 December 2006
How to Cite
Ozaki, K., Nagasaka, T., Notohara, K., Kambara, T., Takeda, M., Sasamoto, H., Jass, J. R., Tanaka, N. and Matsubara, N. (2006), Heterogeneous microsatellite instability observed within epithelium of ulcerative colitis. Int. J. Cancer, 119: 2513–2519. doi: 10.1002/ijc.22095
- Issue online: 20 OCT 2006
- Version of Record online: 23 AUG 2006
- Manuscript Accepted: 18 APR 2006
- Manuscript Received: 1 DEC 2005
- Japanese Ministry of Education, Science, Sports and Culture. Grant Numbers: 17390368, 17591403, 17591402, 14031227
- colitic cancer;
- ulcerative colitis;
- microsatellite instability;
- mismatch repair
Microsatellite instability (MSI) has been associated with colitic cancer. However, reported frequency of MSI was varied and the association of MSI with mismatch repair (MMR) deficiency was unclear. In addition, the occurrence of genetic alterations in stromal cells within ulcerative colitis (UC) is still controversial. We therefore sampled 164 microareas in various pathological lesions of UC with or without colitic cancer and studied the MSI status in relation to the DNA repair protein expressions. A total of 129 microfoci from colorectal tissue of 5 colitic cancer patients and 35 microfoci of 7 UC patients (without neoplasm) were carefully sampled by laser-capture microdissection. MSI was analyzed in each microsamples. The protein expression of MMR genes (MLH1, MSH2, MSH6), O6-methylguanine-DNA methyltransferase and p53 were assessed by immunohistochemical analysis. Variety of di-nulcleotide microsatellite markers was altered in individual microfoci from different morphological epithelial lesions, in full range of nonneoplastic epithelium to colitic cancer. Interestingly, MSI was not observed in stromal cells at any sites, including those within colitic cancer/dysplasia lesions. Expression of the MMR proteins was not lost in any of the lesions examined. Microsatellite alterations rather seem to be related to the initiation than to the progression of colitic cancer. © 2006 Wiley-Liss, Inc.
Colorectal epithelium with a long-standing ulcerative colitis (UC) has an increased risk of developing cancer.1, 2, 3, 4, 5, 6 An invasive colitic cancer often arises from a preceding flat dysplastic epithelium and tend to develop multifocally over the background inflamed epithelium, whilst sporadic colorectal cancer (CRC) develops from a preexisting adenoma, suggesting a dissimilar carcinogenic pathway in colitic cancer.7, 8, 9, 10, 11, 12, 13, 14, 15
In lesions undergoing chronic inflammation, microsatellite instability (MSI) is a possible genetic alteration, and high- and low-level of MSI (MSI-H and MSI-L) has been described in UC associated neoplasm (UCAN).16, 17, 18, 19, 20, 21, 22, 23 However, the reported frequency of MSI-H, a consequence of a defective mismatch repair (MMR) system (notably MLH1 or MSH2), and MSI-L, a consequence in part of O6-methylguanine-DNA methyltransferase (MGMT) deficiency,24 has varied considerably from 9 to 50%, and 11 to 85%, respectively.22, 25, 26, 27 In situation of considerable genetic heterogeneity within a sample, collection of samples without microdissection could possibly obscure the findings of MSI.28, 29 In fact, analysis of mixed MSI-L components could result in a false diagnosis as MSI-H. To verify true MSI status and to clarify the significance of MSI in colitic carcinogenesis, precise analysis of individual microfoci from various pathological lesions (including stroma) is required.16, 18, 19
Accordingly, we sampled 164 microfoci, including nondysplastic epithelium, dysplastic epithelium and cancers from colorectal specimens from UC patients to assess the precise (predicted heterogeneous) genetic alterations that may occur in colitic cancer. We further analyzed the protein expression of MLH1, MSH2, MSH6, MGMT and P53. Finally we investigated genetic alterations in stroma cells within cancer/dysplastic epithelium, because possible genetic alterations within stromal components in UC carcinogenesis were reported.30
Material and methods
Sample collection and microdissection
All tissue samples were obtained from the surgical specimens from 5 UC patients with developed UCAN and from 7 UC patients without neoplasm. The patients underwent surgery at our institution or our affiliated hospitals from 1998 to 2000. Multiple microareas of nonneoplastic epithelium, dysplastic epithelium and cancer were obtained using a PixCel laser capture microscope (LM100; Olympus, Tokyo, Japan) with an infrared diode laser (Arcturus Engineering, CA). A total of 164 microepithelial samples from 12 UC patients and 129 microsamples from 5 UCAN patients were collected and separately analyzed. We extracted and analyzed DNAs from 49 microfoci of nonneoplastic epithelium, 48 microfoci of 5 dysplastic lesions, and 32 microfoci of 4 cancer lesions derived from the 5 UCAN patients. We also extracted DNA from 35 microsamples of different nonneoplastic epithelium derived from the 7 UC without cancer (patient nos. 6–12). Additionally, 28 microloci from stromal areas were collected from 5 UCAN patients. Stromal microsamples within dysplastic, and nonneoplastic areas were carefully microdissected to avoid contamination of epithelial-origin cells, however, some infiltrated lymphocytes were included in the microsamples. As for cancer, stromal cells at the peripheral part of cancer were carefully collected to avoid contamination of cancer cells. Normal control tissue was collected from colonic nonstriated muscle of each specimen. Our study received approval from the institutional review board of the faculty of medicine of Okayama University, Japan.
For each sample, serial sections were cut from formalin-fixed paraffin blocks and the first section was counter-stained with hematoxylin and eosin (H&E) for histological diagnosis. The lesions were classified according to the Inflammatory Bowel Disease-Dysplasia Morphology Study Group Criteria (IBD-DMSGC) published in 1983.11 The histological diagnosis was conducted by a gastrointestinal pathologist (K. N). To maximize the purity of the target epithelial cells, we applied laser capture microdissection (LCM) technique. Tissue was transferred in a thermal polymer disk from 5 to 8 μm deparaffinized unstained serial slides by using a laser beam. This method enabled us to enrich the cellularity of the collected epithelium, at a percentage of 75 or greater. One microsample was dissected by 50–60 shots of laser beam and incubated in 20 μm of lysis buffer [20 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.5% Tween 20, 200 μg/ml proteinase K] for 24 hr at 37°C, then for 15 min at 95°C to inactivate proteinase K.
The MSI testing for each tumor was determined on the basis of an examination of 12 microsatellite markers (BAT25, BAT26, BAT40, D2S123, D5S107, D5S346, D8S87, D17S261, D17S250, D18S35, D18S58 and MYCL1) described previously.31BAT25, BAT26 and BAT40 are mono-nucleotide repeat markers and MYCL1 is the tetra-repeat markers. The rest of the markers are di-nucleotide repeat ones. We classified tumors as MSI-L and MSI-H, if 0%, <40% and ≥40% of the markers displayed MSI, respectively. Tumors displaying no MSI with any of the microsatellite markers were classified as microsatellite stable (MSS).
Polymerase chain reaction (PCR) was performed using PCR9600 thermal cycler (Perkin-Elmer, Foster City, CA) in 50 μl reaction mixtures comprising 3 μl of the 5× DNA sample, 9 μl of Gene Releaser (BioVentures, Murfreesboro, TN), 0.25 μM of each oligonucleotide primer pair (5′-end labeled with Texas-Red), 200 mM each of dNTPs, 5 μl of 10× PCR buffer, 15 mmol/l MgCl2 and 1.25 unit of Taq DNA polymerase (Ampli-TaqGold; Perkin-Elmer, Foster City, CA). After denaturation for 10 min at 95°C, each PCR was carried out for 38 cycles consisting of denaturation for 30 sec at 94°C, annealing for 1.5 min at tm (54°C for BAT25, BAT26, BAT40, D2S123, D5S107, D5S346, D8S87, D17S250 and D17S261; 58°C for MYCL1, D18S35 and D18S58), following the extension step for 10 min at 72°C. After denaturation in formaldehyde for 5 min at 95°C, the amplified PCR products were electrophoresed on denaturing 6% LongRanger-6.1 M gel on HITACHI Autosequencer SQ-5500, and analyzed by FRAGRYS Version 2 software (Hitachi, Tokyo, Japan).
Detection of K-ras mutation
Mutation of the K-ras gene on Codons 12 and 13 was examined. The DNA was amplified and analyzed by two-step PCR restriction fragment length polymorphism (RFLP) method, which is highly sensitive and specific, as previously described.31 Negative controls and positive controls (wild-type DNA and mutated DNA) were run in each analysis.
Immunohistochemical (IHC) staining for P53 (CM1, anti-P53, polyclonal; Novocastra, Newcastle upon Tyne, UK), MLH1 (Clone G168-728, 1 mg/ml; PharMingen, San Diego, CA), MSH2 (Clone FE11, 0.5 mg/ml; Oncogene Science, Cambridge, MA), MSH6 (Clone 44, 0.5 mg/ml; Transduction Laboratory, Lexington, KY), and MGMT (Clone MT3.1; PharMingen, Fremont, CA; dilution 1:100) were performed. Briefly, staining was carried out manually with formalin-fixed paraffin-embedded tissues. Thin (5 μm) sections of representative blocks were deparaffinized and dehydrated using gradient alcohol series. Slides were then immersed in citrate buffer (pH 6.0), irradiated in a microwave oven and allowed to cool at room temperature for 1 hr. After blocking endogenous peroxidase with phosphate-buffered saline containing 3% H2O2, the slides were incubated overnight in the presence of an anti-MGMT monoclonal antibody (Clone MT3.1; PharMingen, Fremont, CA; dilution 1:100). Slides incubated with normal mouse antiserum (Vector Laboratories, Burlingame, CA) provided a necessary negative control. Following a further incubation with secondary antibody (Vector Laboratories, Burlingame, CA) plus avidin–biotin–peroxidase complex (Vector Laboratories, Burlingame, CA), the slides were incubated with biotinyl-tyramide followed by streptavidin-peroxidase (TSA Biotin System Kit, PerkinElmer Life Sciences, Boston, MA). Diaminobenzidine was used as a chromogen, and hematoxylin as a nuclear counter-stain. Sections with obvious nuclear staining were deemed positive. The only foci that were scored as negative were those for which there was definite evidence of positively staining admixed (or surrounding) nonneoplastic tissues such as normal colonic mucosal cells, lymphocytes or stromal cells. The normal staining pattern for MLH1, MSH2, MSH6 and MGMT was nuclear. Tumor cells that exhibited an absence of nuclear staining, in the presence of nonneoplastic cells with nuclear staining, were considered to have an abnormal pattern.
Microarea microsatellite instability in neoplastic and nonneoplastic tissue from UC colorectum
We extracted DNA from 129 epithelial and 28 stromal microareas from 5 colitic cancer specimens. We also extracted DNA from 35 epithelial microsamples from 7 UC specimens without neoplasm (Fig. 1). MSI status on each sample was assessed using 12 microsatellite markers. Clinical characteristics of 12 patients with UC are shown in Table I. The mean age of 5 UCAN patients was 52.8 ± 10.4 years old, and the mean duration suffering from inflammation was 12.0 ± 5.43 years. All 7 UC patients without neoplasm (mean age 31.1 ± 16.7) have undergone emergency operation due to hemorrhage or perforation and were classified into 2 groups by the duration suffering from inflammation. Mean age of the patients with longer duration (Case nos. 6–8) was 48.3 ± 7.09 years old and that of the patients with shorter duration (Case nos. 9–12) was 18.3 ± 2.63 years old.
|Case no.||Age/sex||Duration (years)||Extent of disease||Tumor||No. of microsamples collected|
|UCAN developed on UC|
|UC with long duration|
|UC with short duration|
The MSI status of individual microsamples from 5 UCAN patients was mapped on Figure 2. Microareas exhibiting MSI-L were heterogeneously detected within cancer, dysplasia and nonneoplastic epithelium. Forty four out of 129 (34.1%) epithelial samples showed MSI-L and 34 out of 44 (77.3%) MSI-L samples exhibited instability in only 1 marker among the examined 12 markers. The Case 1 patient developed a rectal lesion with severe dysplasia. MSI-L was not only detected heterogeneously within the dysplastic area, but also detected in the nondysplastic epithelium of the same colon (Tables II and III). The incidence of MSI-L microsamples among the total microsamples in dysplastic lesion was the same as that in nondysplasic epithelium. For example in the Case 1, 4 out of 15 (26.7%) samples in the dysplastic lesion and 4 out of 15 (26.7%) samples in nondysplastic epithelium showed MSI-L (Table IIa). Cases 2–5 developed invasive cancers in rectum or ascending colon but also developed dysplastic lesions close to or distant from primary cancers (Fig. 2). In the case of 2, 3 and 5, MSI-L microareas were heterogeneously distributed in all 3 differing pathological areas. In Case 4, however, MSI-L was identified in 8/8 (100%) of cancer, 6/8 (75.0%) of dysplasia and 3/6 (50.0%) of nonneoplastic epithelium. Since the frequency of MSI-L samples increased along with the levels of histological grade, an underlying mechanisms causing MSI-L may have given an advantage for the cancer development in this case. However, incidence of MSI between dysplastic and nonneoplastic lesion showed no difference in other 4 cases.
|Case 1||4/15 (26.7)1||4/15 (26.7)|
|Case 2||1/6 (16.7)||6/8 (75.0)||3/6 (50.0)|
|Case 3||1/8 (12.5)||0/5 (0.0)||1/8 (12.5)|
|Case 4||8/8 (100)||6/8 (75.0)||3/6 (50.0)|
|Case 5||1/10 (0.0)||4/12 (33.3)||4/14 (28.6)|
|Case 1||0/2 (0.0)||0/2 (0.0)|
|Case 2||0/2 (0.0)||0/2 (0.0)||0/2 (0.0)|
|Case 3||0/2 (0.0)||0/2 (0.0)||0/2 (0.0)|
|Case 4||0/2 (0.0)||0/2 (0.0)||0/2 (0.0)|
|Case 5||0/2 (0.0)||0/2 (0.0)||0/2 (0.0)|
|Epithelium||Left-side colon and rectum||Right-side colon||Total|
|Case 6||2/3 (66.7)1||0/2 (0.0)||2/5 (40.0)|
|Case 7||2/3 (66.7)||1/2 (50.0)||3/5 (60.0)|
|Case 8||3/3 (100)||1/2 (50.0)||4/5 (80.0)|
|Case 9||2/3 (66.7)||0/2 (0.0)||2/5 (40.0)|
|Case 10||1/3 (33.3)||2/2 (100)||3/5 (60.0)|
|Case 11||1/3 (33.3)||1/2 (50.0)||2/5 (40.0)|
|Case 12||2/3 (66.7)||1/2 (50.0)||3/5 (60.0)|
The MSI status of 35 microepithelial samples collected from 7 UC patients, who did not develop neoplasm, was also examined (Table I). One sample from the rectum, 2 samples from the left-side colon and 2 samples from the right-side colon were collected and each sample was taken from the specimen at least in the distance of more than 15 cm of each other. Again, MSI-L was heterogeneously detected among the microdissected epithelium. MSI positive samples were identified from any of the 3 differing anatomical areas: 4/7 (57.1%) in the rectum, 9/14 (64.3%) in the left side colon and 6/14 (42.9%) in the right-side colon (Table III). None of the samples showed MSI-H. There was no difference in the frequency of MSI-L between the UC samples with longer duration and those with shorter duration.
Alteration of 12 microsatellite markers in each microfoci is shown in Figure 3. Three mono-nucleotide markers (BAT25, BAT26 and BAT40) did not show any instability in any of the samples examined both in UCAN and UC cases. D2S123 also showed no instability except in 1 case from nonneoplastic epithelium (Case no. 4). There was no specific marker revealing instability in UC. However, D5S107 showed the most frequent instability among the samples examined.
Then, we studied MSI status in 28 stromal samples from UCAN cases. The microsamples of stroma within cancer, dysplasia and nonneoplastic epithelial regions were carefully collected by LCM method. Interestingly, none of the stromal samples even next to the MSI positive epithelial lesion showed MSI (0/28) (Table II).
K-ras mutation and immunohistochemcal analysis of DNA repair proteins and P53
In addition to the MSI analysis, we examined the protein expressions of MLH1, MSH2, MSH6, MGMT and P53 in 5 UCAN cases by IHC (Table IV). All MMR proteins including MLH1 MSH2 and MSH6 were expressed normally in both neoplastic and nonneoplastic epithelial tissue. MGMT expression was kept in all of the 5 UCAN cases. Expression of P53 was observed in 3 neoplastic lesions and in 1 of the nonneoplastic epithelium. K-ras mutation at Codon 12 was observed in 1 UCAN case (Case no. 3).
|Case no.||K-ras mutation||Immunohistochemistry|
|codon 12||codon 13||P53||MLH1||MSH2||MSH6||MGMT|
This study demonstrates for the first time that the heterogeneous di- or tetra-nucleotide slippage is the primary microsatellite alterations occurred in the range of epithelial lesions of UC. Marked heterogeneity, as shown in this study, could only be achieved by the analysis of multiple (in total of 164) microsamples within the same pathological area by LCM. The microsatellite alteration already presents in inflamed epithelium suffered from short duration. The frequency of MSI, however, did not elevate along with tumor development in the majority of UC cases. Strikingly, all levels of inflamed microepithelia from nondysplastic lesion to colitic cancer showed complete heterogeneous instability in different microsatellite markers. On the other hand, the stromal cells, within any pathological area including colitic cancer, showed complete stability of microsatellite markers (MSS).
There were several studies that examined the incidence of MSI-H in colitic neoplasm. However, the reported frequency of MSI-H has varied from 9 to 50%.17, 20, 21 MSI has also been reported in colorectal epithelium of UC patients suffering from chronic inflammation.16, 18, 19 Importantly, most of the reported microsatellite analysis has been undertaken on bulk samples of either the tumor or nonmalignant epithelium and, thus, the status of MSI-H could be due to the pooling effect of individual microfoci which showed MSI(-L) in different markers.
In our study, there were no samples that showed loss of expression of the MMR proteins (MLH1, MSH2, MSH6). We also could not identify microsatellite alterations in mono-nucleotide repeat markers (BAT25, BAT26 and BAT40), alterations of which are highly specific for MMR deficiency. Recently, novel types of MSI have been reported termed type A and type B alterations.32 Type A alterations are defined as length changes of ≤6 bp and type B changes are more drastic modification of ≥8 bp. In our study on UC specimens, all types of MSI we observed belonged to the type A alterations. Type A MSI may reflect the uncorrected DNA slippage events that may be related to the MSI-L phenotype.32 Loss of MGMT is a candidate mechanism for MSI-L CRC. However, loss of MGMT expression was not observed in our study. Accordingly mechanisms other than dysfunction of MMR proteins or MGMT could contribute to microsatellite alterations that we observed. Oxidative stress is one of probable mechanisms that may affect MMR system.33, 34, 35 In both UC and Crohn's disease, the reactive oxygen species, that are produced within chronically inflamed colorectal epithelium, may lead to DNA damage. Overproduction of free radicals saturates the ability of cells to repair DNA damage prior to replication. The resulting imbalance in base excision-repair enzymes may cause MSI-L in chronic inflammation.36 Lack of folic acids in colorectal mucosa in UC patients was also reported to accelerate the deficiency in the MMR system due to the oxidative stress.18
Field carcinogenesis is one of the well-known hypotheses to explain tumor development from chronic inflammation. In juvenile polyposis syndrome (JPS) and UC, abnormal stromal cells can affect the development of adjacent epithelial cells.30 A population of cells defective in proliferation appeared to be derived from the stroma and these cells lead to the evolution of CRC through a hamartomatous change.37 Regenerations that occur to replace the damaged epithelium may increase the frequency of genetic alterations in stromal cells. This primary oncogenic effect that causes mutation in stromal cells is called “landscaper defects.”30 However, in UC specimens, stromal cells never showed microsatellite alterations in our study. Our result is different from previous report by Matsumoto et al., which showed MSI in microdissected stroma from epithelial lesions of UC with relatively high frequency.38 In our study, loss of heterozygosity (LOH) in several loci was also examined by microsatellite markers, but no LOH was observed in stromal cells (data not shown), possibly due to our strict avoidance of boundary regions between epithelium and stroma in our sample collection.
In summary, we have shown marked heterogeneity in MSI in various mucosal lesions of UC epithelium that were not restricted to colitic cancer. Such microsatellite alterations rather seem to be related to the initiation of the neoplastic transformation of UC epithelium than to the progression of the colitis–dysplasia–cancer sequence in a majority of the UC cases.
We thank Ms. Naoko Hoshijima for technical assistance.
- 18Microsatellite instability in non-neoplastic mucosa of patients with ulcerative colitis: effect of folate supplementation. Am J Gastroenterol 1998; 93: 2060–4., , , , , , , , .Direct Link: