DETECTION OF BACILLUS CALMETTE-GUERIN IN THE BLOOD BY THE POLYMERASE CHAIN REACTION METHOD OF TREATED

BLADDER CANCER PATIENTS SERDAR TUNCER, MEHMET ILTERIfj TEKIN, HALUK OZEN, CENK BILEN, SERHAT mAL AND

DOGAN REMZI From the Lkparhnent of Urology and Infectious Disease, Hacettepe Universi@, Ankam, Turkey

ABSTRACT

Purpose: Following intravesical bacillus Calmette-Guerin (BCG) instillation, we attempted to detect BCG in the blood using the polymerase chain reaction (PCR) method and correlate these findings with the occurrence of major complications due to this treatment.

Materials and Methods: Intravesical BCG immunotherapy was given to 22 consecutive pa- tients with superficial bladder tumors. In 2 patients the BCG instillation had to be discontinued due to serious side effects of therapy. Blood samples (252 aliquots) were obtained from 126 BCG courses in 22 cases, and 2 additional samples (4 aliquots) were obtained from 1 patient 1 and 3 months after cessation of therapy. All blood samples were analyzed by the PCR technique for detection of deoxyribonucleic acid tuberculosis Mycobacterium tuberculosis.

Results: Of the 126 blood samples 9 (7.196) were PCR positive for M. tuberculosis. These 9 positive samples belonged to 3 patients, all of whom were among those 4 patients who had major clinical side effects.

Conclusions: We demonstrated that rapid and sensitive detection of mycobacteremia by PCR correlated with the clinical course of these patients. We also demonstrated that PCR can be used to monitor BCG in the blood after antituberculous therapy. The early, fast and accurate diagnosis of BCG in the blood by PCR may alter the serious clinical c o m e of these patients by initiation of specific treatment early. However, further extensive studies are needed to validate the8e results.

KEY Worn: bladder neoplasms, BCG vaceine, polymerase chain reaction

Intravesical administration of bacillus Calmette-Guerin (BCG) has been shown to be effective in the treatment of prophylaxis of superficial bladder tumors and carcinoma in 8itu.l Although side effects of BCG immunotherapy are re-

Severe toxicity has also been observed afbr only a few insid- lations.2.3 Serious complications have mostly been reported

was recently reviewed by Lamm et al in 2,602 ~a8es.l While it has been shown that 95% of patients had no serious side effects, 0.7% had granulomatous pneumonitis andlor hepati- U 10

Patient chameteristics No. pta. 22 No. men 18 No. women 4

Age *.): Medim 61 Ranee 46-72

Tumor sbge.: Ta 7 T1 15

Tumorgrade: I 4

rn 8

ported to be substantially greater with intensive regimens,

as case reports. The incidence of BCG induced side effects

tis resulting from systemic BCG infection.' In the same study it was reported that 0.4% of patients had experienced life -threatening BCG sepsis.' Early diagnosis of these serious and life-threatening complications is crucial since prompt administration of specific drugs has been shown to alter the for a of 6 Each therapy of 81 mg. unfavorable clinical course. However, the conventional l a b connaught (2.2 to 6.4 x 108 co lony~fom~ per ratory investigations are timecons- and the treatment 27 =.) intravesid BCG, suspended in 50 d. hotonic so- is usually started based on clinical judgment. we attempt to &- CNonde, and &till& into the bladder for 2 hours. In 2 detect BCG in the blood using the polymerase chain readon patienb the ~0-s had to be disC0nt;inued due to (PCR) method following intravesical BCG instillation, and side effects oftherapy. *O ~ q u o b of5 d. P r i p h e d blood correlate these findings with OCcurrenCe Of major comPLca- s-ples were obtained fi.om all patien& 15 afbr

each intravesical instillation. We obtained 126 blood samples tions. (252 &quota) during BCG coulse~ and 2 additional samples (4 ahquote) ware obtained from 1 patisnt 1 and 3 months after cessation of therapy. A Merent code number was given to each aliquot and they were stored in tubes eon- ethylenediaminektraacetic acid at -20c until &- Fh.

Sample preparation. Ten ml. l y ~ b buffer containing 100 mM. sodium chloride, 10 mM. tris-hydrochloric acid (pH 8.0), 25 mM. EDTA and 0.5% sodium dodecyl sulfate were added

redon ofthe bar, and was

MATERIALS AND METHODS

Patients and samples. Intravesical BCG immunotherapy was given to 22 consecutive patients with superficial bladder tumors b e h e n May 1994 and November 1995. The tumor stages and grades of the patients are shown in the table. The treatment was started 2 weeks h r complete transurethrd

Accepted for publication May 29,1997. 2109

21 10 DETECTION OF BACILLUS CALMET'TE-GUERIN IN BLOOD BY POLYMERASE CHAIN REACTION

to each 5 ml. blood sample. The suspension was incubated on rotary shaker for 30 minutes and centrifuged at 3,000 rpm for 15 minutes. After discarding the supernatant the same procedure was repeated once more. The pellet was suspended in 2 ml. TE buffer (10 nM. tris-hydrochloric acid, 1 mM. EDTA) and filtered through 5 pm. disposable filters. The suspension was centrifuged a t 12,000 rpm for 2 minutes, and the pellet was washed with distilled water twice and sus- pended with 20 pl. distilled water. The final suspension was incubated for 10 minutes in boiled water and supernatant was used for PCR.

PCR. Primers specific to the 123-base pairs (bp) fragment of the repetitive insertion element IS6110 in the chromosome of M. tuberculosis complex, including M. bovis-BCG strain, were used in amplification reactions.4 Amplifications were camed out in 50 pl. volumes of reaction mixture containing 50 mM. potassium chloride 10 mM. tris hydrochloric acid (pH 9.0), 1% Triton X-100,2.0 mM. magnesium chloride, 50 mM. of each deoxynucleoside triphosphate, 20 pmol. of each primer, 1 unit of taq polymerase and 5 pl. deoxyribonucleic acid (DNA) sample. Forty cycles of denaturation at 94C 30 seconds and annealing, extension at 68C, 90 seconds were performed in an automated thermal cycler. Reaction mix- tures without DNA were used as negative controls. The same sample preparation and amplification protocols were done with 10-fold serial dilutions of BCG Connaught strain sus- pension in sheep and human blood for the detection of sen- sitivity level of the procedure.

"he presence of the 123-bp amplification product was an- alyzed by electrophoresis of 10 pl. amplified mixture on 2% agarose gel. Gels were stained by ethidium bromide and photographed on an ultraviolet transilluminator (see figure). If the amplification bands were observed in both aliquots of the same patient the result was considered positive. "he result was interpreted as negative if either 1 of the 2 aliquots remained negative. The detection sensitivity of PCR was found approximately 10 to 50 BCG Connaught strain per ml. of blood by this method. All of the side effects of treatment were recorded carefully.

RESULTS

Of22 patients 20 completed the aforementioned BCG ther- apy schedule. The treatment had to be stopped in 2 patients and macroscopic hematuria occurred in 2 others. The rest of the patients completed therapy without any major complica- tions and side effects. Of the 126 blood samples 9 (7.1%) were PCR positive for M. tuberculosis. These 9 positive samples

=longed to 3 of the 22 patients and were among 4 patients who had major side effects of BCG. Blood samples from the remaining 19 patients were negative.

CASE HISTORIES

M. tuberculosis PCR products from peripheral blood on agarose gel with IS6110 specific primers (123-base pairs (bp) . Lanes 1 to 7, ~sitivesarnples.Lanes9to13,5x1O4,5~1O3,5~1O2, 50and5crude M. tuberculosis lysate. Lane 8, negative control (reaction mixture without DNA). Lane M, molecular weight marker (Hae 111 digested ax174 phage DNA).

Case 1 . A 72-year-old man received intracavitary BCG immunotherapy after endoscopic resection of a stage pT1, grade I1 bladder carcinoma. Soon after the second instillation he complained of high fever (>39C) and malaise that lasted for 2 days and then disappeared. After course 4 of therapy he was hospitalized because of disseminated erythematous skin rash, high fever and malaise. Complete blood counts and blood biochemistry results were within normal limits except for elevated liver transaminase levels. Alanine aminotrans- ferase was 69 IUD. (normal 0 to 40) and aspartate amino- transferase was 77 IUD. (normal 0 to 50). Serological markers for viral hepatitis were negative. The routine and specific cultures of blood and urine for tuberculosis were also nega- tive. Intravesical BCG treatment was stopped and 300 mg. isoniazid daily were started. He was discharged from the hospital after the symptoms subsided. Blood samples re- vealed positive PCR for M. tuberculosis in every specimen except for the samples obtained after course 2.

Case 2. A 57-year-old woman received 6 courses of intra- cavitary BCG because of a recurrent, multifocal stage pTa, grade I1 bladder carcinoma. After courses 2 and 5 of therapy she had flu-like symptoms, malaise and gross hematuria. Routine bacteriological and specific cultures of urine and blood for tuberculosis remained negative. Blood samples af- ter courses 2, 4 and 5 were be PCR positive for M. tubercu- losis.

Case 3. A 54-year-old man male received 2 weekly intra- cavitary BCG instillations after endoscopic resection of a stage pTa, grade I11 multifocal bladder carcinoma. The day after course 2 he was hospitalized because of severe malaise, high fever (greater than 39.5C) and jaundice. The leukocyte count was 2,600/1nm.~ The alanine aminotransferase, aspar- tate aminotransferase, alkaline phosphatase and bilirubin levels were 262 IUfi., 304 IUfl., 594 IUD. (normal 40 to 100) and 3.2 mg./dl. (normal 0 to 1.21, respectively. Serological markers for viral hepatitis were negative. Blood and urine cultures were also negative. BCG therapy was stopped and a combination therapy of 300 mg. isoniazed, 1,500 mg. etham- butol and 1,500 mg. pyrazinamide daily were started due to the toxicity of rifampicin based combinations, although it is known that BCG is relatively resistant to pyrazinamide. After 8 days of antituberculous therapy, the fever subsided and the liver enzymes decreased. Blood samples obtained after each of the 2 courses and 1 month after the beginning of antituberculous therapy were PCR positive. Blood samples obtained after 3 months of antituberculous therapy were PCR negative.

Case 4. A 62-year-old man received 6 courses of intracavi- tary BCG immunotherapy following transurethral resection of a stage pT1, grade I1 bladder carcinoma. He complained of low grade fever (less than 38.5C) for 3 and 12 hours after courses 2 and 3 of BCG, respectively. Macroscopic hematuria was observed following instillation 4. The patient also expe- rienced severe dysuria after BCG course 5. Serological mark- ers for viral hepatitis, and routine urine and nonspecific blood cultures were negative. All blood samples obtained after each of the 6 courses were PCR negative.

DISCUSSION

BCG is the attenuated strain of M. bovis which is a mem- ber of the Mycobacterium tuberculosis complex and consists of living bacilli. The efficacy of intravesical BCG instillations in the prophylaxis and treatment of superficial bladder tu- mors and carcinoma in situ has led to extensive use of this therapy. BCG itself is not toxic to the tumor cells.5.6 The

DETECTION OF BACILLUS CALMETTE-GUERIN IN BLOOD BY POLYMERASE CHAIN REACTION 2111

bacilli attach to the damaged bladder wall and provoke an intensive local inflammatory reaction.5 It is believed that th is reaction is important for the response to BCG treatment. m e reaction is not simple and involves a series of modulators in the immunological cascade including cytokine~.~-$ Local side effects, such as hematuria, cystitis, bladder contractions and granulomatous prostatitis, have been described.' Gran- domatous hepatitis, pneumonitis, azotemia, miliary tuber- alosis and pancytopenia have been defined as rare systemic complications. Additionally, death directly related to BCG toxicity has been reported in 10 patients.' The systemic side effects are believed to result from a hypersensitivity reaction

BCG antigens,' or clinical illness directly caused by BCG mycobacteria.I0 The documentation of the presence of myco- bacteria at distant organs and in blood11 may be regarded as the Indirect evidence to support active BCG infection as the muse of the clinical illness.

The long incubation period of conventional culture meth- ods and relatively low incidence of the serious reactions hinder comprehensive clinical studies. Review of the limited clinical experience with BCG toxicity does not provide suffi- cient information to draw firm recommendations for initia- tion of the treatment. Laboratory tests that enable prompt confirmation of BCG related systemic complications are re- quired.

Advances in molecular biological techniques have provided a new approach to the rapid diagnosis of tuberculosis by nucleic acid probes and recently by PCR. PCR has already been shown to be applicable for the amplification of M. tu- berculosis DNA in various clinical samples including blood with EDTA.12-14 Accurate diagnosis of tuberculosis by PCR has reduced the number of patients who were treated solely on the basis of clinical suspicion. To our knowledge we report the first attempt for the detection of the presence of BCG mycobacteria in the blood by PCR. Three cases were PCR positive for mycobacteria, including 2 in which therapy had to be stopped because of serious systemic complications.

We used a different strategy for the sample preparation from blood for mycobacterial DNA amplification. Recently, Folgueira12 and Schlugerl3 et a1 demonstrated M. tuberculo- sis bacteremia, and Iralu et all4 have shown M. avium bac- kremia by PCR. Total lymphocyte fraction of the patients has been used in these studies for the detection of mycobac- kria in the blood. We have concentrated BCG efficiently from the whole blood by aggressive digestion followed by large pored filtration and centrifugation. With this method, it is possible to collect bacilli that are free in serum and/or at- tached to the blood cells in addition to intralymphocytic ba- cilli. Although the detection limit of procedures in previous mycobacteremia studies was not clear, we were able to detect approximately 10 to 50 Calmette-Guerin bacilli per ml. bloOd.12.13 Many copies ofthe IS6110 repetitive sequence are Present in most of the M. tuberculosis strains but only 1 to 4 Copies are found in M. bovis strains, which could limit the Sensitivity for detection of BCG. In separate experiments we have shown that this method also detects amplifications of 10

20 bacilli from different M. tuberculosis clinical isolates diluted in blood (unpublished data).

Flu-like symptoms with low grade fever were observed in cases 2 and 4, both of which had hematuria as well. Severe systemic complications of therapy resulted in sepsis in Case 3, and possible hepatitis and/or BCG inflammation occurred in Case 1. The rate of serious complications ranges from 0.1 to 2.95% in other studies.' The relatively high frequency of these serious complications could be the result of the limited number of patients included in our series. Traumatic cathe- terization and impaired immunological host defense increase

risk of systemic reactions. There was no difficulty during Mtheterization ofthe bladder in our patients. Similar to our results, Lamm et a1 have reported 20% pneumonitis andor

1 i

I

t 5

(

1 1 1 1 1 1

I

iepatitis in a small series, including 10 patients treated with L different strain.15

There are no definite data on the incidence and/or risk -atio of mycobacteremia with BCG instillations in bladder ;umor patients. In our study 9 of 126 intravesical BCG in- itillations (7.1%) were associated with mycobacteremia. In 2 :ases the clinical appearance resembled systemic BCG infec- ;ion which responded to antituberculosis therapy. Sepsis was present in Case 3, and PCR results were in concordance with the clinical picture. Disappearance of mycobacteria detected by PCR after antituberculous therapy confirms that myco- bacteremia can be detected and monitored by PCR. These results show that there is a high correlation between PCR detected mycobacterial DNA in blood and systemic complica- tions. This observation supports the importance of circulat- ing BCG organisms in the etiology of these reactions. How- ever, the presence of BCG in blood per se cannot definitely rule out the possibility of a hypersensitivity reaction to these microorganisms. In either case the detection of these micro- organisms in blood helps the clinician to initiate specific treatment a t once.

Although low grade fever was observed in Case 2, other systemic manifestations of BCG infection were absent. Flu- like symptoms and low grade fever have been accepted as the result of immune stimulation which can also be seen in patients receiving systemic interferons and cytokines. How- ever, 3 of the 6 samples showed mycobacteremia by PCR in our study. The presence of hematuria could be the reason of high load hematogenous delivery of bacilli in this patient, which resulted in positive PCR results in 3 samples. Hema- turia and low grade fever were also observed in Case 4 without serious systemic findings. BCG in the blood was not observed in this patient. However, it is impossible to rule out mycobacteremia unequivocally since the number of bacteria in the blood samples may have remained below the detection limits of our method. Another possibility is that the timing of our blood sample was not ideal to detect the highest level of mycobacteremia in these patients.

CONCLUSIONS

This pilot study was designed to detect BCG in blood by PCR and to analyze the reliability of the method in early detection of the side effects of BCG immunotherapy. Rapid and sensitive definition of mycobacteremia in these patients by PCR was possible. We have shown that in this limited group of patients detection of mycobacterial DNA in the blood stream by PCR correlates with systemic infection or septic reactions due to BCG. We were able to monitor BCG in the blood after antituberculous therapy with PCR. Further stud- ies of larger groups will not only help to validate these results but also determine the optimum time of blood sampling.

REFERENCES

1. Lamm, A. L., van der Meijden, A. P. M., Morales, A,, Brosrnan, S. A,, Catalona, W. J., Herr, H. W., Soloway, M. S., Steg, A. and Debruyne, F. M. J.: Incidence and treatment of complications of bacillus Calmette-Guerin intravesical therapy in superficial bladder cancer. J. Urol., 147: 596, 1992.

2. Brosman, S. A.: Experience with bacillus Calmette-Guerin in patients with superficial bladder carcinoma. J. Urol., 128: 27, 1982.

3. Steg, A., b l u e , C . , Debre, B., Boccon-Gibod, L. and Sicard, D.: Systemic bacillus Calmette-Guerin infection in patients treated by intravesical BCG therapy for Superficial bladder cancer. In: EORTC Genitourinary Group Monograph 6: BCG in Superficial Bladder Cancer. Edited by F. M. J. Debruyne, L. Denis and A. P. M. van der Meijden. New York Alan R. Liss, Inc., pp. 325334. 1989.

4. Eisenach, K. D., Cave, M. D., Bates, J. H. and Crawford, J. T.: Polymerase chain reaction amplification of a repetitive DNA

2112 DETECTION OF BACILLUS CALMETTE-GUERIN IN BLOOD BY POLYMERASE CHAIN REACTION

sequence specific for Mycobacterium tuberculosis. J. Infect. Dis., 161: 977, 1990.

5. Bar, B., Bernstein, I. D. and Rapp, H. J.: Suppression of tumor growth at the side of infection with living bacillus Calmette- Guerin. J. Natl. Cancer Inst., 46: 831, 1971.

6. Flippin, T., Mdherji, B. and Dayal, Y.: Granulomatous hepatitis as a late complication of BCG immunotherapy. Cancer, 46: 1759, 1980.

7. DeJong, W. H., De Boer, E. C., Van der Meijden, A. P. M., Vegt, P., Steerenberg, P. A,, Debruyne, F. M. J. and Ruitenberg, E. J.: Presence of interleukin 2 in urine of superficial bladder cancer patients a h r intravesical treatment with Bacillus Calmette-Guerin. Cancer Immunol. Immunother., 31: 182, 1990.

8. Ratliff, T. L.: Mechanisms of action of intravesical BCG for bladder cancer. In: BCG in Superficial Bladder Cancer. Edited by F. M. J. Debruyne. L. Denis and A. P. M. van der Meijden. New York Alan R. Kiss, Inc., pp. 325-334, 1989.

9. Balbay, D., &en, H., Ozkardeg, H., Barut, A., Bakkaloglu, M., Tasar, C. and Remzi, D.: Detection of urinary interleukin-2, interleukin-2 receptor and tumor necrosis factor levels in pa- tients with superficial bladder tumors after intravesical BCG immunotherapy. Urology, 43: 187,1994.

10. Pinsky, C. M., Hirhault, Y. and Oettgen, H.: Treatment of ma- lignant melanoma by intra-tumoral injection of BCG. Natl. Cancer Inst. Monogr., 39: 225, 1973.

11. Civen, R., Berlin, G. and Panosian, C.: Vertabral osteomyelitis a h r intravesical administration of Bacille Calmette-Guerin. Clin. Infect. Dis., 1 8 1013, 1994.

12. Folgueira, L., Delgado, R., Palenque, E., Aguado, J. M. and Noriega, A. R.: Rapid diagnosis of Mycobacterium tuberculosis bacteremia by PCR. J. Clin. Microbiol., 34 512, 1996.

13. Schluger, N. W., Condos, R., Lewis, S. and Rom, W. N.: Ampli- fication of DNA of Mycobacterium tuberculosis from periph- eral blood of patients with pulmonary tuberculosis. Lancet, 344: 232, 1994.

14. Iralu, J. V., Sritharan, V. K, F'ieciak, W. S., Wirth, D. F., Maguire, J. H. and Barker, R. H.: Diagnosis of Mycobacterium avium Bacterernia by polymerase chain reaction. J. Clin. Mi- crobiol., 31: 1811, 1993.

15. Lamm, D. L., Stogdill, V. D., Stogdill, B. J. and Crispen, R. G.: Complications of bacillus Calmette-Guerin immumotherapy in 1278 patients with bladder cancer. J. Urol., 135: 272, 1986.

EDITORIAL COMMENT

The authors present the first series of patients treated with intra- vesical BCG in whom routine peripheral blood sampling was accom- plished for examination of BCG bacteriemia using PCR. The obser- vation that 7% of blood samples were positive and that these positive blood samples correlated with major clinical side effects support the current practice of treating these patients with antitubercular anti- biotics. The demonstrated efficacy of PCR in detecting BCG back"- emia could result in more prompt diagnosis and treatment, and therefore reduce the systemic toxicity of BCG.

The demonstration of BCG in the blood stream of patients with systemic reaction supports the theory that these reactions are the result of circulating organisms. However, it is important to note that these circulating organisms may also induce the hypersensitivity response associated with BCG sepsis. Our clinical experience1 as well as our controlled laboratory studies suggest that treatment with isonazid, rifampin and prednisone results in optimal survival.2.3 Therefore, we would recommend double or triple antitubercular an- tibiotics plus 40 mg. prednisone daily in these patients. Experience has shown that relapse can occur if steroids are tapered too rapidly. Hypersensitivity is clearly an important aspect of this reaction but the data presented in this paper as well as our laboratory studies suggest that prednisone alone should be avoided but used in combi- nation with a 3 to 6-month course of antitubercular antibiotics.

Donald L. Lamm Department of Urology West Virginia University Morgantown, West Virginia

1. Rawls, W. H., Lamm, D. L., Lowe, D. A,, Crawford, E. D., Sarosdy, M. F., Montie, J. E., Grossman, H. B. and Scardino, P. T.: Fatal sepsis following intravesical bacillus Calmette- Guerin administration for bladder cancer. J. Urol., 144: 1328, 1990.

2. DeHaven, J. I., Traynelis, C., Riggs, D. R., Ting, E. and Lamm, D. L.: Antibiotic and steroid therapy and massive systemic bacillus Calmette-Guerin. J. Urol., 147: 738, 1992.

3. Koukol, S. C., DeHaven, J. I., Riggs, D. R. and Lamm, D. L.: Drug therapy of bacillus Calmette-Guerin sepsis. Urol. Res., 2 2 373. 1995.