The utility of urinary CD80 as a diagnostic marker in patients with renal diseases

CD80, which regulates T cell activation, may provide a differential diagnostic marker between minimal change disease (MCD) and other renal diseases, including focal segmental glomerular sclerosis (FSGS). However, recent reports show contrasting results. Therefore, we evaluated the utility of urinary CD80 as a diagnostic biomarker. We collected 65 urine samples from 55 patients with MCD (n = 31), FSGS (n = 4), inherited nephrotic syndrome (n = 4), Alport syndrome (n = 5) and other glomerular diseases (n = 11), and control samples (n = 30). We measured urinary CD80 levels by ELISA. Urinary CD80 (ng/gCr) (median, interquartile range) levels were significantly higher in patients with MCD in relapse (91.5, 31.1–356.0), FSGS (376.2, 62.7–1916.0), and inherited nephrotic syndrome (220.1, 62.9–865.3), than in patients with MCD in remission (29.5, 21.7–52.8) (p < 0.05). Elevation of urinary CD80 was observed, even in patients with inherited nephrotic syndrome unrelated to T cell activation. Additionally, urinary CD80 was positively correlated with urinary protein levels. Our results suggest that urinary CD80 is unreliable as a differential diagnostic marker between MCD in relapse and FSGS or inherited kidney diseases. Increased urinary CD80 excretion was present in all patients with active kidney disease.

However, more recent reports have shown different results for the expression of CD80 on podocytes. Reiser et al. reported that CD80 was expressed on podocytes in patients with various diseases: MCD in relapse, FSGS, SLE, and even inherited diseases 14 . In contrast to this report, Larsen et al. and Novelli et al. reported that CD80 was not found on podocytes in either MCD in relapse or FSGS 15,16 . Thus, the specificity of expression of CD80 on podocytes in MCD remains controversial 15,17 . Moreover, the usefulness of urinary CD80 levels as a differential diagnostic marker for patients with MCD has not been fully examined.
The present study was performed to determine the utility of urinary CD80 measurement for the differential diagnosis of patients with various renal diseases. We recruited patients with MCD, FSGS, and other glomerular diseases, including inherited renal diseases that are not associated with T-cell activation.

Results
Patient characteristics. The clinical characteristics of each group are shown in Table 1. There were statistically significant differences in urinary protein/creatinine between MCD in remission and MCD in relapse, FSGS, or inherited NS.
Precision of our ELISA. We produced a standard curve for CD80 with commercially quantified CD80 protein (Cat. No. 10698-H08H, Sino Biological Inc.) (Supplemental Fig. 1). In addition, we also produced a standard curve with an ELISA kit (Bender MedSystems, Burlingame, CA, USA) used in previous reports 13,18 , as a comparison with our ELISA method. Both ELISAs displayed a linear correlation between CD80 concentration and OD450; the correlation coefficient of our ELISA was higher than that of the established ELISA kit.
Urinary CD80 excretion. The CD80 (ng/gCr) levels in urine were significantly higher in patients with MCD in relapse (91.5, 31.1-356.0), FSGS (376.2, 62.7-1916.0) or inherited NS (220.1, 62.9-865.3) than in patients with MCD in remission (29.5, 21.7-52.8) (p < 0.05) (Fig. 1). Two of four patients with FSGS and one of four patients with inherited NS showed much higher urinary CD80 than 75th percentile of CD80 in patients with MCD in relapse (2337.6, 651.5, and 1038.1 ng/gCr, respectively). There was no significant difference in urinary CD80 level between patients with MCD in remission and controls (59.9, 48.6-72.5). There were no significant differences in urinary CD80 levels among patients with MCD in relapse, FSGS, or inherited NS. To investigate whether elevated urinary CD80 in MCD in relapse has specificity for other renal diseases, we compared urinary CD80 values between MCD in relapse and all renal diseases other than MCD. Notably, no differences in urinary CD80 were found between MCD in relapse and other renal diseases (Supplemental Fig. 2).
Correlation between CD80 values in serum and urine. The relationship between serum CD80 and urinary CD80 was examined in 15 patients, but no correlation was found (r = 0.07, p = 0.80) (Fig. 2). Serum CD80 remained at a normal value (20-320 pg/ml) 19 , even in patients who exhibited high values, such as urinary CD80 exceeding 1000 ng/gCr.

Correlation between urinary CD80 and urinary protein.
A positive correlation was found between urinary CD80 levels and the urine protein/creatinine ratio in total subjects (r = 0.57, p < 0.0001) (Fig. 3).
In each group, there were wide variations in urinary CD80 concentration, reflected by wide error bars (Fig. 1). Therefore, we studied the correlation between urinary CD80 and urinary protein in each group. First, we calculated the 97.5th percentile of the control group and determined that the normal range of urinary CD80 lies below 105.1 ng/gCr. Values greater than the 97.5th percentile of control were defined as high. The remaining subjects in each group were defined to exhibit normal CD80 levels. Then, we compared the amounts of urinary protein in these two groups, according to each renal disease. There were significant differences between the high and low groups in MCD in relapse, and for the group of all renal diseases (Supplemental Fig. 2). Although the same tendency was observed between inherited NS and FSGS, a significant difference was not detected because of the small sample size.

Discussion
In this study, we evaluated the concentration of urinary CD80 of 65 samples from 55 patients with various renal diseases, including inherited diseases, to determine the utility of urinary CD80 as a diagnostic biomarker. Among these 65 samples, 20 were collected from 10 patients at different two points of NS (remission and relapse). The remaining 45 samples were collected from different patients. Since Garin et al. reported the utility of urinary CD80 as a differential biomarker between the relapse phase of MCD and other renal diseases in 2009 13 , five other reports have been published concerning the use of this biomarker 12,18,[20][21][22] . All these reports showed elevation of urinary CD80 only in patients with MCD in relapse, with levels remaining normal in controls, patients with MCD in remission, FSGS or other renal diseases. Specific elevation of urinary CD80 in MCD in relapse was supported by the following reports: (1) The expression of CD80 on podocytes was strongly enhanced in MCD patients in relapse, measured by immunostaining, compared with patients with MCD in remission or those with FSGS 12 . (2) The source of urinary CD80 is the podocytes because the molecular weight of urinary CD80 (53 kDa) is close to that of CD80 derived from cell membrane, rather than that of serum CD80 (23 kDa) 12 . (3) The serum level of CD80 is normal in MCD in relapse although it is elevated in urine 13 . However, some other studies have recently reported that the immunostaining assay of CD80 on podocytes is not a reliable diagnostic marker for MCD in relapse 15,16 .  Although previous reports maintain that urinary CD80 levels clearly differentiate between MCD in relapse and FSGS 12,13,18,23 , our study did not ( Fig. 1) (Supplemental Fig. 2). There was some uncertainty, in that previous studies involved FSGS cases derived from genetic defects, whereas our cases with FSGS were derived from immunological mechanisms. To remove this concern, we measured urinary CD80 levels in patients with inherited NS; these patients also exhibited elevated CD80. This result supports our conclusion that urinary CD80 is not a reliable marker for MCD in relapse.
In our study, the concentration of CD80 in urine was higher in patients with FSGS and inherited NS in addition to MCD in relapse than MCD in remission. This might mean that excretion of urinary CD80 may increase in patients with active renal disease, rather than upon activation of T cells. CD80 has been previously measured in patients with inherited renal diseases 14,23 . Cara-Fuentes et al. reported a patient with an NPHS2 defect who showed elevation of urinary CD80. Reiser et al. reported that nephrin or α3 integrin knockout mice showed expression of CD80 on podocytes; they hypothesized that various stresses, including genetic defects, might lead to podocyte CD80 induction.
Our study showed that urinary CD80 is not affected by serum CD80 (Fig. 3). Importantly, we examined the correlation between serum and urine CD80 levels; we found that serum CD80 was within the reference range, even for serum samples from patients that simultaneously exhibited very high urinary CD80 values, suggesting that urinary CD80 is not derived from serum CD80.
We focused on the relationship between urinary protein and urinary CD80 and found a positive correlation (Fig. 2), which was observed in MCD in relapse, FSGS, and inherited NS (Supplemental Fig. 3). These results indicated that the elevation of urinary CD80 was not specific to patients with MCD in relapse, and it was correlated with urinary protein levels that might rather reflect the severity of glomerular damage, i.e., podocyte injuries. In contrast, in some samples, the correlation between urinary protein and CD80 was poor; this suggests that other factors may influence urinary CD80 excretion.
The current study has limitations. We did not perform Western blotting analysis of urinary CD80 or measure CD80 expression in kidney tissue. When examining urinary CD80, it is important to determine whether that urinary CD80 is blood-derived or kidney-derived; it is an important limitation that we did not perform these analyses. However, the method for measuring CD80 in renal tissues has been reported as low quality 15,16 ; moreover, we showed that serum CD80 levels in our patients were not correlated with urinary CD80. The second limitation is that the sample sizes remain small in our study. The number of patients with FSGS or inherited NS was only four; however, 50% or 25% of these patients showed much higher levels of urinary CD80. Therefore, we judged that urinary CD80 might not be a useful diagnostic marker for renal diseases, and we decided not to collect more samples.
In conclusion, our results showed that urinary CD80 was not a reliable differential diagnostic marker between MCD in relapse and FSGS or other kidney diseases, including inherited kidney diseases, in contrast to previous reports. The level of CD80 was elevated in all patients with active kidney disease. We suggest that measuring urinary CD80 levels is not a reliable marker for the differential diagnosis of MCD in relapse and other kidney diseases.   (Table 1). Renal biopsies were performed on all patients with MCD, FSGS, Alport syndrome, and other glomerular diseases except for the case with HUS. These patients were diagnosed using established criteria [24][25][26][27][28] ) and Alport syndrome by genetic tests. Thirty-one patients with MCD provided 41 urine samples, including 10 patients who provided samples at both relapse and remission. They were divided into two groups, including a group in remission (n = 17), whose urinary protein creatinine ratio was <2.0 g/gCre, and a group in relapse (n = 24), whose urinary creatinine ratio was >2.0 g/gCre. Immunosuppressive therapy was introduced for all patients with MCD in relapse, or FSGS, and for some patients with IgAN and HSPN. This led to a reduction in proteinuria for all treated patients. Statistical analysis. All results were described as median with interquartile range. Kruskal-Wallis and Dunn's multiple comparison tests were used for analyses of differences among groups. The Mann-Whitney U test was used for comparisons between high and low CD80 groups in each renal disease, and for comparison of urinary CD80 between MCD in relapse and renal diseases other than MCD. Spearman rank correlation coefficient was used to calculate the correlation between urinary protein and urinary CD80. The level of statistical significance was set as p < 0.05. Data graphics and statistical analysis were described with GraphPad Prism7 software.

Data Availability Statement
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.