Cathepsin B (CB) is involved in degradation of extracellular matrix proteins during tumor progression in human solid organ tumors (such as colorectal, bladder, and breast cancers), including human prostate cancer. Its activities are regulated by endogenous inhibitors (such as stefins or cystatins). Increased expression of cathepsin B message, protein, and membrane association have been linked to malignancy, but there are very few studies of their mRNA expression in prostate cancer using in situ hybridization techniques. Our objective was to determine the relationship of CB and stefin A (cystatin A) mRNA localization to the Gleason grading system for histologic scores in the hope of distinguishing aggressive and less aggressive variants of prostate cancer. We used a 25-base biotinylated oligonucleotide CB cDNA antisense probe to localize CB message and a 27-base biotinylated oligonucleotide stefin A cDNA antisense probe to localize stefin A message. Prostate samples from 41 prostatectomy patients were collected along with their pre-surgery serum PSA levels and clinical stage of the disease. Sections prepared from frozen prostate tissue samples were hybridized with the CB and stefin A and control pBR 322 probes using techniques reported by Sinha et al. [1] and their distribution quantitated by an image analysis system. Prostate sections treated with RNAse before hybridization or incubated with the pBR 322 control probe showed little or no reaction products, confirming that localization of CB and stefin A probes was specific. In prostate cancer, the reaction products were found in neoplastic and invasive cells and occasionally in stromal cells. The ratios of CB to stefin A were similar in normal prostate and benign prostatic hyperplasia (BPH) whereas they varied consistently within and between Gleason histologic scores for prostate cancer. These variations showed three localization patterns; namely, prostate cancers with higher levels of CB than stefin A, lower levels of CB than stefin A, and similar levels of CB and stefin A. All three patterns and ratios for CB and stefin A were found in prostate samples (22/41) represented by the Gleason histologic score 6 tumors. In these tumors, serum PSA levels ranged from 1 to 78 ng/ml and prostate cancers showed B, C, and D clinical stages. There was no correlation of CB/stefin A ratio and serum PSA values or clinical stage in a limited number of prostate cancer cases. Our data showed that there were prostate cancer cases within Gleason histologic scores which expressed high, similar, and low levels of CB when compared to stefin A. We postulate that prostate cancer cases showing higher levels of CB compared to stefin A probably represent an aggressive variant of this cancer within any one Gleason histologic score. If this is the case, aggressive variants of prostate cancer would occur within Gleason scores 3 to 10 even though higher scores are usually considered more aggressive forms of prostate cancers. Since our study is based upon a very limited number of frozen prostate samples, we emphasize that a larger series of archival prostate cancer samples along with their survival data should be analyzed to establish any relationship of CB/stefin A ratio and aggressive variants of this cancer. Therefore, our conclusion is tentative. Our study provides a partial explanation for differences in the clinical course of prostate cancer in patients. This is the first study to show that determination of CB and stefin A mRNA ratios may lead to identification of aggressive and less aggressive variants of prostate cancer within a Gleason histologic score.