Urinary soluble HLA class I antigen in patients with minimal change disease: a predictor of steroid response

In primary minimal change disease (MCD) and focal segmental glomerulosclerosis (FSGS), increased lymphocyte reactivity to renal antigens has been defined. Soluble HLA class I antigen (sHLA-I) is actively secreted by T and B lymphocytes when they are stimulated by mitogens, antigens and lymphokines. To determine if serum and urine sHLA-I levels could predict steroid response in patients with MCD and differentiate those from FSGS, we have investigated 45 healthy controls, biopsy-proven 17 patients with MCD (edema and 24-hour urine protein > 3.5 g/day), 8 patients with FSGS (24-hour urine protein > 1 g/day) and 10 patients with membranous nephropathy (MGN) (24-hour urine protein > 1 g/day). Before and after prednisone therapy (1 mg/kg/day or 2 mg/kg/EOD for 8 weeks), the levels of serum and urinary sHLA-I were measured by ELISA (sHLA-STAT; Sangstat Co., Calif., USA). After 8 weeks of treatment, 10 patients with MCD were responders (MCD-CR) while the other 7 patients with MCD were nonresponders (MCD-NR). Three of 7 patients with MCD-NR were re-biopsied and finally diagnosed as FSGS. They were included in the data of patients with FSGS. In healthy controls, serum sHLA-I was detected (415 +/- 256 ng/ml), but urinary sHLA-I was not. At entry, there were no differences in age, sex, serum Cr and 24-hour urine protein among the patients with MCD-CR, MCD-NR and FSGS, but serum albumin was significantly elevated in patients with FSGS and MGN (p < 0.05). Serum sHLA-I levels were notably elevated in MCD-CR (1,040 +/- 1,066 ng/ ml), in MCD-NR (668 +/- 315 ng/ml) and in FSGS (713 +/- 790 ng/ml), but not in patients with MGN (444 +/- 86 ng/ml) when compared with controls (p < 0.05). On the other hand, urinary sHLA-I levels in MCD-NR (541 +/- 239 ng/mg Cr) and in FSGS (457 +/- 239 ng/mg Cr) were significantly higher than those in MGN (125 +/- 28 ng/mg Cr) and in MCD-CR(100 +/- 42 ng/mg Cr, p < 0.05) and these substantial differences were maintained for 8 weeks. In all patients, serum and urinary sHLA-I levels were not reduced during 8 weeks of steroid therapy. We conclude that elevated serum and urinary sHLA-I levels reflect increased cellular immune response and disease activity in patients with MCD and FSGS. In patients with MCD, urinary sHLA-I may be an easily measurable indicator of predicting steroid response, while MCD-NR with high urinary sHLA-I levels might be re-evaluated for the possibility of FSGS.