Use of a Microsatellite Marker in Predicting Dysplasia in Ulcerative Colitis

Background  Patients with ulcerative colitis (UC) have an increased risk of developing colorectal cancer. The current screening protocol involves an annual colonoscopy and biopsy after the patient has had the disease for 8 years. This, however, does not prevent the development of colorectal cancer.

Hypothesis  A microsatellite marker for IBD1 may identify individuals who are at greater risk of developing dysplasia and therefore colorectal cancer.

Design  Case-control study.

Setting  Single surgical practice.

Patients and Methods  DNA was extracted from peripheral leukocytes of 152 patients: 22 with UC and dysplasia; 48 with UC and no dysplasia; 24 with colorectal cancer; and 58 with noninflammatory bowel disease, nonmalignant gastrointestinal tract disease who were used as control patients. A microsatellite marker for IBD1 (D16S541) was amplified by polymerase chain reaction. Genotypes were identified using autoradiography.

Results  Six alleles and 15 genotypes were identified for marker D16S541. Genotype CC was found in 33% (8/24) of cancer patients but only 12% (7/58) of controls (χ²=5.5; P=.02). Thirty-two percent (7/22) of patients with dysplastic UC also had this genotype, whereas only 8% (4/48) of patients with nondysplastic UC had the genotype (χ²=4.6; P=.03; vs controls: χ²=3.1; P=.08).

Conclusions  This microsatellite marker for IBD1, when combined with other markers, has the potential to be used as a screening tool for colorectal cancer and dysplasia in patients with UC. Such a marker would be of particular use in improving the sensitivity and specificity of the current screening protocol for dysplasia and colorectal cancer for patients with UC.

ULCERATIVE colitis (UC) and Crohn's disease, collectively known as inflammatory bowel disease, affect more than 1 million Americans. Although both disorders are associated with an increased risk of developing colon cancer, this risk is much higher for UC than for Crohn's disease.¹

This increased risk of developing colorectal cancer begins after the disease has endured for 8 years,² which is one of the primary concerns in the treatment of patients with chronic UC. This risk is related to disease duration and extent of disease and is greatest in patients with pancolitis.

Guidelines issued by the American Cancer Society³ recommend surveillance colonoscopy every 1 to 2 years beginning 8 years after the diagnosis of pancolitis. Random mucosal biopsies are performed at 10-cm intervals. The biopsy specimens are then examined for the presence of dysplasia and are used as an early predictor of malignancy. The finding of dysplasia is an indication for colectomy to prevent the development of colon and rectal cancer.⁴

This current approach has its limitations. Patients with UC may still present with an advanced colorectal cancer not previously identified at colonoscopy, despite participating in such a screening protocol.⁵ This is due, in part, to sampling error, the presence of inflammation that might obscure the histological diagnosis of dysplasia,⁶ and documented differences in the pathological interpretation of dysplasia.^7-9 In addition, a significant proportion of patients develop colorectal cancer in the absence of dysplasia.⁵

Ulcerative colitis is thought to have a multifactorial etiology, encompassing genetic and environmental factors.^10,11 It is a complex genetic trait that does not follow a simple mendelian model, and more than one gene is likely to be involved. Evidence of this genetic predisposition has been shown in family-based and twin studies.^12,13

IBD1, a susceptibility locus on chromosome16,¹⁴ has been identified as a potential candidate gene. This locus, mapped to the pericentromeric region of chromosome 16, has been associated with Crohn's disease in several genomewide linkage studies^15,16 and is also associated with UC in another study.¹⁷

As part of our ongoing research into the genetics of inflammatory bowel disease, we hypothesized that it might be possible to predict which patients are at risk of developing dysplasia and therefore colorectal cancer by using a microsatellite marker for this IBD1 locus.

This study was approved by the institutional review board of the University of Louisville, Louisville, Ky. Written informed consent was obtained from all patients.

We undertook a case-control study using 152 patients from a single surgical practice: 48 had UC with no evidence of dysplasia; 22 had UC with dysplasia; 24 had colorectal cancer; and 58 had noninflammatory bowel disease, nonmalignant gastrointestinal tract disease and were used as a control group. All patients in the control group underwent a complete colonoscopy.

The diagnosis of UC was verified by endoscopy with biopsy. The histological evidence from all patients with UC was reviewed by a single pathologist (R.P.) with a special interest in inflammatory bowel disease. The diagnosis of dysplasia by the pathologist was then used to eliminate interobserver variability in the diagnosis of dysplasia.

Ten milliliters of peripheral blood was obtained by venipuncture. Leukocyte DNA was extracted using a commercial DNA extraction kit (Puregene; Gentra, Minneapolis, Minn). A previously identified microsatellite marker for IBD1 (D16S541) was amplified by polymerase chain reaction. The primer sequence for D16S541 was:

CCACACCAGCGGTTTTTCTAA CACACTTTACACACACCTATACCC

Polymerase chain reaction was performed in a total volume of 10 µL containing 2 µL (100 ng) of DNA, 3.75 µL of water, 1.0 µL of buffer, 0.6 µL of magnesium chloride, 0.35 µL of deoxynucleotide triphosphate, 1.0 µL of each primer, 0.2 µL of Taq polymerase, and 0.1 µL of phosphorus 32 labeled–deoxycytidine triphosphate. The conditions for polymerase chain reaction included denaturation at 94°C for 2 minutes and then 25 cycles (each cycle consisting of 94°C for 20 seconds, 55°C for 30 seconds, and 72°C for 60 seconds). The final step consisted of a 72°C extension for 5 minutes. The polymerase chain reaction products were electrophoresed on 6% polyacrylamide gels and the genotypes were determined by autoradiography. Genotypes were scored by 2 independent investigators masked to disease diagnosis.

Clinical data were collected prospectively by patient interview and chart review and were then integrated with the DNA results on a customized database designed for this study.

χ² Tests were performed using statistical software (Primer for Biostatistics, version 4.02; Microsoft, Redmond, Wash). A χ² value of greater than 3 was considered to be statistically significant. Analysis of variance tests were used to determine statistical significance between the ages of the different groups.

The alleles were labeled according to size, with the smallest being A and the largest being F (Figure 1).

Six alleles and 15 genotypes were identified for D16S541. There was no statistical difference in allele frequency among any of the groups. The frequency of genotype CC was, however, different among groups. The CC genotype was found in 32% (7/22) of patients with dysplastic UC and 33% (8/24) of patients with colorectal cancer, whereas only 8% (4/48) of patients with nondysplastic UC and 12% (7/58) of controls had the genotype (Figure 2 and Table 1).

These differences were statistically significant for patients with dysplastic UC and those with nondysplastic UC (χ²=4.6; P=.03) and approached significance for patients with dysplastic UC and controls (χ²=3.1; P=.08). Differences between patients with cancer and genotype CC compared with controls were also statistically significant (χ²=5.5; P=.02).

Patient demographic data are shown in Table 2. There was no statistical difference between groups when comparing mean age or sex. Mean duration of disease and age at diagnosis were similar for both groups of patients with UC. There was a preponderance of women in most patient groups. This might, in part, be because the principal investigator is a woman surgeon.

Only 50% of patients with dysplastic UC had documented pancolitis. This diagnosis is thought to increase the risk of developing dysplasia. In addition, 9 patients had dysplasia with a disease duration of less than 8 years, and only 4 of these had a diagnosis of pancolitis.

Colorectal cancer is thought to evolve in a multistep process that may be influenced by genetic and environmental factors.¹⁸ Identification of a specific genetic alteration using a microsatellite marker could potentially be used for patient surveillance.

The risk of malignant transformation in pancolitis may reach greater than 10% after 25 years of disease duration.^19,20 This can result in colorectal cancers affecting patients younger than we would usually see with sporadic colorectal cancer.

Although dysplasia is widely accepted as a marker for malignant potential by both surgeons and gastroenterologists,²¹ inflammatory bowel disease–associated cancers are still a relatively common occurrence.²² These cancers, even if diagnosed by an endoscopic surveillance program, might not be diagnosed at an early and potentially curable stage. This calls into question the reliability and usefulness of colonoscopic screening in diagnosing dysplasia and therefore in detecting a predilection for developing colorectal cancer.

We believe that a panel of molecular markers, of which D16S541 would be only one, could potentially be used as an adjunct to current screening protocols to identify patients at increased risk of developing colorectal cancer. Such markers would assess the risk of developing colorectal cancer independent of disease duration, since they would be screened from germline DNA. Another advantage of this type of screening is that it would allow the risk of developing colorectal cancer to be assessed at any stage of the disease. This would allow for more effective screening and possibly permit diagnosis of high-risk patients before cancer develops or at least at an earlier stage. If the sensitivity and specificity of such markers were sufficient, it would even allow for less intensive screening of patients who are not deemed to be at high risk of developing colorectal cancer.

Presented at the 107th Scientific Session of the Western Surgical Association, Santa Fe, NM, November 15, 1999.

Reprints: Susan Galandiuk, MD, Department of Surgery, University of Louisville, Louisville, KY 40292 (e-mail: s0gala01@gwise.louisville.edu).

Fabrizio Michelassi, MD, Chicago, Ill: Cancer is an ever-present concern in patients with UC. The association between UC and cancer was first described by Crohn and Rosenberg in 1925. Estimates of the relative risk of developing colon cancer in UC vary between 0.8 and 23 times the relative risk for the normal population. The cancer risk increases with both the duration and the extent of the disease, as stated by Dr Hunt. The original suggestion from Rausohoff that the risk is in the range of 0.5% to 1.0% per year after 10 years of duration is still a reasonable working estimate for patients with UC. Dysplasia, as stated in the presentation, is the earliest histologically detectable neoplastic precursor of cancer. With this in mind, the work of Dr Hunt and colleagues to define a microsatellite marker in predicting dysplasia in UC has the potential of identifying patients at high risk of neoplastic transformation with genetic tests, even before the actual transformation has occurred. By using a microsatellite marker for a locus for inflammatory bowel disease, IBD1, the authors have identified one genotype frequently expressed in patients with UC complicated by dysplasia or cancer.

I have 2 questions for the authors: the IBD1 locus is located in the pericentromeric region of chromosome 16 along a very broad region encompassing hundreds of markers. Why have the authors decided on this specific marker? Which was the prestudy hypothesis, or the specific genetic reason, that led to the use of this marker?

My second question focuses on the relationship between inflammatory bowel disease and IBD. This relationship has been established for Crohn's disease and not so much for UC. The University of Chicago has recently participated in an international genetic study where the pedigree and the genetic characteristics of more than 500 IBD families have been studied. While there appears to be a significant and strong correlation between IBD1 and Crohn's disease, there appears to be no such relationship between IBD1 and UC. Why do the authors believe the IBD1 may be important in the genetic pathogenesis of dysplasia and cancer in UC?

This work parallels the work from other investigators with proto-oncogenes, such as K-ras, with tumor suppressor genes such as P53, APC, DCC, and recently with the DNA repair genes such as MSH2 and MLH1. I would like to congratulate the authors for their efforts to elucidate specific genetic pathways leading to cancer formation in UC as this knowledge will eventually help us with clinical decisions which at present are difficult both for the patient and for the physician.

Theodore X. O'Connell, MD, Los Angeles, Calif: I have a question regarding the clinical implications that the authors refer to as far as screening is concerned. Although these people with CC genes seem to be at higher risk and perhaps should be screened more often, two thirds of the patients who had cancer or dysplasia did not have this marker, CC. Therefore, do we screen without CC less? Should we really be doing that since it is clear that the marker is not the only reason these people develop cancer?

Steven J. Stryker, MD, Chicago: I would like to ask your help with a difficult clinical situation that we all encounter from time to time, ie, the patient with long-standing quiescent colitis who is found during screening colonoscopy to have an otherwise typical adenomatous polyp. The decision then is whether to treat the polyp as an isolated independent lesion or to assume that it is a dysplasia-associated lesion and therefore recommend proctocolectomy. Have you used the microsatellite marker to guide you in this situation?

Dr Galandiuk: This is, of course, preliminary data that is exciting since it could potentially help in patient management. Regarding Dr Michelassi's question, Dr Hunt also attended the international consortium meeting in Australia. As Dr Michelassi mentioned, IBD1 occupies a large area. The marker that we used was obtained from the Marshfield database. It is within the area that the international consortium is studying; therefore, we feel it is a valid marker since it is within this 4-cm area. We chose to study IBD1 since in our collective study of slightly over 600 patients, we did find an association between IBD1 with both Crohn's disease and UC. This is in agreement with the findings of Mirza et al, who also reported an association with UC. We are excited about our data and believe that our database is unique because all of the patients are surgical patients and because of this have a more severe disease phenotype. This may actually facilitate identification of genetic associations. In addition, the majority of genetic studies that have looked at IBD1 or other candidate genes utilize family studies, as Dr Michelassi described. One of the problems with such studies is that there is often no pathology review. We have found that if a single pathologist reviews all of the cases, the diagnosis will frequently change, particularly in cases of colonic IBD. Due to our finding that UC was also associated with IBD1, we began examining subgroups and found a significant association with dysplasia as presented today.

Regarding Dr O'Connell's question, again, this is early data. One cannot yet apply these findings clinically because, as we have shown, the genotype in question was only present in one third of patients with dysplasia. However, if you take other markers such as that we have reported with one of the DNA mismatch repair genes and others and combined them with this marker, one could perhaps have a 90% chance of identifying patients with dysplasia. This could be much more likely to have a clinical application.

Regarding Dr Stryker's question, that is the ideal situation where a testing method like this could be of help. I have gone over the pathology slides of all patients together with Dr Petras. It is his bias that there is no true adenoma in patients with inflammatory bowel disease, specifically UC. All villous adenomatous change in these patients is a dysplasia-associated lesion or mass unless proved otherwise. Techniques such as that presented here might help manage the patients.