TUMORS IN BLADDER REMNANT AFTER ENTEROCYSTOPLASTY 485 ever, the condition continued to deteriorate and a bone scan in October revealed widespread metastases, from which the patient died in December.

Macroscopic examination demonstrated apparently nor- mal ileal mucosa and intact anastomoses. The bladder mu- cosa was congested and the bladder wall was extensively infiltrated by tumor, predominately infiltrating the base of the left lateral wall. The kidneys were hydronephrotic, and the ureters were dilated and congested.

Microscopy confirmed the tumor to be poorly differentiated adenocarcinoma forming irregular glands and diffise sheets that extensively infiltrated the bladder base and left lateral wall. Tumor extended through the bladder wall to involve the perivesical fat. The remaining urothelium showed cystitis with glandular metaplasia and dysplasia. The ileal mucosa had nonspecific inflammatory changes. The ureters were un- remarkable but the kidneys were hydronephrotic with marked loss of parenchyma due to interstitial nephritis, con- sistent with chronic pyelonephritis. Focal intestinal metapla- sia with atypia and adenocarcinoma in situ were noted in the right renal pelvis (fig. 5). Prostate specific antigen staining of the tumor was negative.

DISCUSSION

The etiology of tumors that develop in bladders after aug- mentation is probably multifactorial but these histological findings suggest that transitional epithelium rather than bowel mucosa undergoes transformation to adenocarcinoma. However, adenocarcinoma of the bladder is uncommon and it usually arises from remnants of the urachus, or areas of cystitis glandularis or glandular metaplasia.1°

It is known that nitrosamines are elevated in the urine of patients who underwent enterocystoplasty.11 Metabolic acti- vation of nitrosamines results in the formation of alkylation products that can combine with deoxyribonucleic acid (DNA) to form abnormal adducts.12 Badawi et al showed that pro- mutagenic change in the DNA of patients with schistosomi- asis is associated with elevated urinary nitrosamines and an increased incidence of bladder cancer.13 This promutagenic adduct (06 methyl guanine) is specifically repaired by the enzyme 06-alkylguanine-DNA alkyltransferase.14 Endo- scopic biopsies of ileocystoplasties a t our hospital have shown that there are large amounts of this enzyme in the bowel segment but only small quantities in the transitional epithe- 1 i ~ m . l ~ This finding suggests that bowel mucosa is protected against nitrosamine induced alkylation damage and, hence, initiation of carcinogenesis. Nurse and Mundy thought that ileal segments used in the urinary tract did not carry the risk of malignant change" and these findings may also explain why few tumors have been reported after ileal conduit diver- sion. Therefore, DNA adduct formation may occur as a field change wherever urothelium is exposed to nitrosamines. Re flux of these carcinogenic compounds in case 4 may have caused the changes in the renal pelvis (fig. 5).

A possible explanation for the apparent preference of these tumors to be located in the anastomotic region is that in- creased cellular proliferation associated with healing tissues may result in cell division before the cell can repair the DNA adduct. Therefore, all subsequent daughter cells are abnor- mal and tumor development may be initiated. Tumor promo- tion can then occur, possibly due to growth factors such as basic fibroblast growth factor, which has been shown to be elevated in the urine of patients with enterocystoplasty and is thought to originate from the bowel mucosa.16

Most patients described underwent ileocystoplasty because of bladder contracture due to tuberculosis. It is not certain whether tuberculosis of the bladder is significant for the development of these tumors. However, nearly all early en-

dication and only time will tell whether these patients with bladder tuberculosis are at greater risk.

CONCLUSIONS

In light of this new evidence, all patients who underwent enterocystoplasty require regular followup with cystoscopy and biopsy to ensure that malignant transformation does not occur in the bladder remnant. It also appears pertinent to monitor similarly all patients with a history of chronic in- flammation and bladder contracture for possible late tumor formation.

REFERENCES

1. Mundy, A. R.: Cystoplasty. In: Urodynamics-F'rinciples, Prac- tice and Application, 2nd ed. Edited by A. R. Mundy, T. P. Stephenson and A. J . Wein. London: Churchill Livingstone,

2. Treiger, B. F. G. and Marshall, F. F.: Carcinogenesis and the use of intestinal segments in the urinary tract. Urol. Clin. N. Amer., 18: 737, 1991.

3. Stewart, M.: Urinary diversion and bowel cancer. Ann. Roy. Coll. Surg., 68: 98, 1986.

4. Seidman, H., Mushinski, M. H., Gelb, S. K and Silverberg, E.: Probabilities of eventually developing or dying of cancer- United States. CA, 35: 36, 1985.

5. Watt, P. C. H., Spence, R. A. J. and Jackson, B.: The intestines. In: Pathology for Surgeons. Bristol: Wright, pp. 81-131.1986,

6. Shousha, S., Scott, J. and Polak, J.: Ileal loop carcinoma after cystectomy for bladder exstrophy. Brit. Med. J., 2 397, 1978.

7. Sakano, S., Yoshihiro, S., Joko, K, Kawano, H. and Katsuke, N.: Adenocarcinoma developing in an ileal conduit. J. Urol., 153: 146, 1995.

8. Filmer, R. B. and Spencer, J. R.: Malignancies in bladder aug- mentations and intestinal conduits. J. Urol., 143: 671, 1990.

9. Gregoire, M., Kantoff, P. and DeWolf, W. C.: Synchronous ade- nocarcinoma and transitional cell carcinoma of the bladder associated with augmentation: case report and review of the literature. J. Urol., 149: 115, 1993.

10. Watt, P. C. H. and Spence, R. A. J.: The kidney, ureter and bladder. In: Pathology for Surgeons. Bristol: Wright, pp. 218- 259, 1986.

11. Nurse, D. E. and Mundy, A. R.: Assessment of the malignant potential of cystoplasty. Brit. J. Urol., 64: 489, 1989.

12. Shank, R. C.: Toxicology of N-nitroso compounds. Toxicol. Ap- plied Pharmacol., 31: 361, 1975.

13. Badawi, A. F., Mostafa, M. H., Robert, A. and OConnor, P. J.: Role of schistosomiasis in human bladder cancer: evidence of association, aetiological factors, and basic mechanisms of car- cinogenesis. Eur. J. Cancer Prevention, 4 45, 1995.

14. Badawi, A. F., Cooper, D. P., Mostafa, M. H., Aboul-Azm, T., Barnard, R., Margison, G. P. and OConnor, P. J.: 06- alkylguanine-DNA alkyltransferase activity in schistosomia- sis-associated human bladder cancer. Eur. J. Cancer, 30 1314, 1994.

15. Barrington, J . W., O'Connor, P. J., Cooper, D. P. and Stephenson, T. P.: DNA alkyltransferase activity following augmentation cystoplasty. Brit. J. Urol., in press.

16. Bamngton, J . W., Fraylin, L., Fish, R., Shelley, M. and Stephenson, T. P.: Elevated levels of basic fibroblast growth factor in the urine of clam enterocystoplasty patients. J. Urol., 156: 468, 1996.

pp. 457-466, 1994.

EDITORIAL COMMENT

Augmentation tumors are of special interest to the cancer biologist because they present the opportunity to study the relationship among inflammation, wound healing and malignancy. This relation- ship has been pondered for many years and it is perhaps best char- acterized by Haddow who in 1972 proposed that "the wound may & regarded as a tumor which heals itself."' This actually makes more sense when turned around, for example a tumor is a wound that never heals.

This article adds 4 more cases to the literature on augmentation tumors, unofficially totaling 19. I believe that these cases are impor- tant because, when added to the 15 others, 2 interestingobservationa arise. 1) Of the 19 Datients 13 had a history of genitourinary tuber- terocystoplasties at OUT hospital were performed for this in- . -

486 TUMORS IN BLADDER REMNANT AFTER ENTEROCYSTOPLASTY

culosis. 2 ) Of 17 tumors 13 (the pathological condition in 2 previously reported cases was inadequately described) were associated with vesicointestinal anastomosis.

The fact that these tumors commonly arise near the anastomosis may not be just an accident. This concept is familiar to urologists, who are aware of the increased risk of adenocarcinoma near the anastomosis following ureterosigmoidostomy.' I t is thought that a biological change must occur in this zone, because tumors have been reported in the ureteral stump of the ureterocolonic anastomotic site of ureterosigmoidostomies up to 15 years a h r urine was again diverted from the Furthermore, in an animal model of uret- emsigmoidostomies tumors were not observed outside the suture line.' Molecular mechanisms involving the generation of nitroso compounds in the presence of bacteria are becoming better under- stood. as indicated in the Discussion. The basic idea is that nitrates, which are normally excreted by the urine, are degraded to nitrites by bacteria. These compounds undergo further degradation to nitroso compounds, such as N-methyl N-nitroso-urea, which have the capa- bility of carcinogenic activity and, therefore, are eligible as tumor initiator^.^ The classic example relates to the demonstration that N-methyl N-nitroso-urea can induce a guanine to adenine mutation in the H ras oncogene, which is an important step in the initiation of colon and transitional cell carcinoma.6

The authors correctly describe the possibility that the products of inflammation and healing may act as tumor promoters after tumor initiation, increasing the chance of malignancy near the anastomo- sis. Certain growth factors, such as epidermal growth factor, are known to augment tumorigenesis that was induced by separate carcinogens, such as N-nitroso compounds. These substances, called promoters, need not be mutagens, as classically believed, but they may act in an epigenetic manner, for example by increasing the risk of cumulative mutations by sustaining cell proliferation.7

Keeping these lessons in mind, the apparent association of these tumors with a history of genitourinary tuberculosis becomes intrigu- ing. Our experience with bacillus Cdmette-Guerin has taught us that chronic inflammation, in this case iatrogenic, has the potential to be powerful. If it has the ability to cure cancer in 1 type of circumstance, there is no reason to believe that it cannot cause cancer in another. In this case the unique inflammation caused by tuberculous oraanisms may in fact be the uromoter that drives the

the mechanisms of action of bacillus Calmette-Guerin to understand fully this process. However, the type of tumor would simply depend on what initiation events occurred in the beginning. The promoters then simply amplify and drive the malignant signal past the body's fail-safe systems. This hypothesis is suggested by the fact that blad- der augmentation performed for tuberculous bladders in Africa, where colon cancer is much less prevalent than in the United States, is not associated with malignancy.8 The idea is that in Africa initi- ation events probably did not occur and without proper initiation, for example genetic changes, promotion has little effect on malignant transformation.

It is only through careful observation and reporting of our clinical experience that scientists will be led in the right direction. The authors are to be congratulated.

William C. DeWolf Department of Urology Beth Israel Hospital Haruard Medical School Boston, Massachusetts

1. Haddow, A,: Molecular repair, wound healing, and carcinogene- sis: tumor production a possible overhealing? Adv. Cancer Res., 16: 181, 1972.

2. Stewart, M.: Urinary diversion and bowel cancer. Ann. Roy. Coll. Surg., 68: 98, 1986.

3. Husmann, D. A. and Spence, H. M.: Current status of tumor of the bowel following ureterosigmoidostomy: a review. J. Urol., 144: 607, 1990.

4. Gittes, R. F.: Carcinogenesis in ureterosigmoidostomy. Urol. Clin. N. h e r . , 13: 201, 1986.

5. Crissy, M. M., Gittes, R. F. and Steele, G. D.: Rat model of carcinogenesis in ureterosigmoidostomy. Science, 201: 1079, 1980.

6. Zarbl, H., Sukumar, S., Arthur, A. V., Martin-Zanca, D. and Barbacid, M.: Direct mutagenesis of Ha-ras-1 oncogenes by N-nitroso-N-methylurea during initiation of mammary carci- nogenesis in rats. Nature, 315: 382, 1985.

7. Ames, B. N. and Gold, L. S.: Too many rodent carcinogens: mitogenesis increases mutagenesis. Sciknce, 249 970, 1990.

tumor. Unfortunately we db not yet have sukcient understanding of 8. Steele, G.: Personal communication.