Arylamine N-acetyltransferases (NATs) were partially purified from human breast tumor tissues with complete separation of the isoforms in DEAE-Cellulose ion-exchange step. NAT with activity towards p-aminobenzoic acid (PABA) was isolated and purified from human breast tumor with 77 % yield and a purification factor of 5-fold. NAT with activity towards sulfamethazine (SMZ) was isolated and purified from human breast tumor with 21 % yield and a purification factor of 3-fold. Further purification attempts by Blue Sepharose affinity column chromatography resulted in the complete loss of both enzyme activities. The NAT1 purified from human breast tumor tissues had a molecular weight (Mr) value of about 27600 and an isoelectric point (pI) around 4.8, as confirmed by SDS-PAGE, IEF and Western blotting analysis. With immunohistochemical analysis, level of intensity of NAT1 immunostaining was observed to be going from weak in reduction mammoplasty samples to strongest in malignant breast tissue. The interindividual variation in the conjugation of p-aminobenzoic acid (PABA) and of sulfamethazine (SMZ) by cytosolic arylamine N-acetyltransferases (NATs) were investigated in 30 human breast tumor and matched samples. The average specific activity against PABA was calculated as 13&amp / #61617 / 2 pmole/min/mg protein for breast control NATs, and 20&amp / #61617 / 3 pmole/min/mg protein for breast tumor NATs. The average specific activity against SMZ was calculated as 12&amp / #61617 / 2 pmole/min/mg protein for breast control NATs, and 34&amp / #61617 / 6 pmole/min/mg protein for breast tumor NATs. Wilcoxon test revealed that the difference between the control and tumor groups is statistically significant with respect to the NAT1 activities as well as NAT2 activities. In three (3/30, 10%) patients tumor and tumor-free breast tissue NAT1 activity was not detectable. Among control tissues, the percentage of measurable NAT2 activity was 77% (23/30), while in tumor tissues it increased to 91%. Chemotherapy treatment was observed to have a slight inhibitory effect on mean NAT1 and NAT2 activities. There was an indication of a possible negative association with mean NAT1 activity and estrogen receptor status, while mean tumor NAT2 activity was observed to increase among estrogen receptor positive patients. Grade of malignancy seems to be positively associated with NAT1, but no such association could be suggested for NAT2 enzyme. Menopausal state of the patient was suggested to have a significant effect on NAT2 activity. Genotype determination of NATs revealed that NAT1*4 and NAT2*5A allele being most common among 10 breast cancer patients. NAT1*11 allele was prevalent among postmenopausal women. The putative rapid NAT1 genotypes was found to display lower control and tumor mean NAT1 activities compared to normal NAT1 genotypes. Among slow NAT2 acetylators, mean tumor NAT2 activities was found to be significantly higher than respective controls.