Urinary Podocyte Count as a Potential Routine Laboratory Test for Glomerular Disease: A Novel Method Using Liquid-Based Cytology and Immunoenzyme Staining.

INTRODUCTION: This study investigated whether our urinary podocyte detection method using podocalyxin (PDX) and Wilms tumor 1 (WT1) immunoenzyme staining combined with liquid-based cytology can serve as a noninvasive routine laboratory test for glomerular disease.
METHODS: The presence of PDX- and WT1-positive cells was investigated in 79 patients with glomerular disease and 51 patients with nonglomerular disease.
RESULTS: The frequencies and numbers of PDX- and WT1-positive cells were significantly higher in the glomerular disease group than in the nonglomerular disease group. The best cutoffs for PDX- and WT1-positive cell counts for identifying patients with glomerular disease were 3.5 (sensitivity = 67.1% and specificity = 100%) and 1.2 cells/10 mL (sensitivity = 43.0% and specificity = 100%), respectively.
CONCLUSION: Because our urinary podocyte detection method using PDX immunoenzyme staining can be standardized and it detected glomerular disease with high accuracy, it can likely serve as a noninvasive routine laboratory test for various glomerular diseases.

Introduction

Podocytes are highly specialized epithelial cells that play an important role as filters that prevent the leakage of high-molecular-weight proteins, glomerular basement membrane (GBM) components, and endothelial cells [1, 2]. Podocytes are injured by various insults (genetic, mechanical, immunologic, and toxic). Upon podocyte injury, localized changes occur in the slit membrane, followed by foot process effacement [3]. This phenomenon causes a decrease in glomerular filtration function and clinically appears as proteinuria. Other podocyte injury mechanisms include hypertrophy, detachment, apoptosis, and epithelial-to-mesenchymal transition [4, 5, 6, 7]. In these cases, podocytes detach from the GBM frequently. Podocytes are terminally differentiated cells that are generally unable to replicate in response to injury [8]. Thus, to correct local podocyte loss, parietal epithelial cells (PECs) are activated and attached to the naked GBM. When podocyte loss is excessive, PECs are further activated and induced to proliferate, eventually causing glomerulosclerosis. Additionally, severe glomerular injury associated with defects of the capillary wall and GBM leads to the proliferation of PECs and crescent formation. In this situation, podocytes are severely damaged, leading to their detachment from the GBM [2, 9].

Because highly damaged podocytes detach from the GBM and accumulate in urine, several studies revealed that the detection of podocytes in the urine is a valuable noninvasive biomarker for disease activity and therapeutic efficacy in various glomerular diseases [2, 10, 11]. These previous studies relied on the combination of conventional sample preparation methods (direct smears and cytospins) and podocalyxin (PDX) immunofluorescence staining to detect urinary podocytes. However, conventional sample preparation methods have problems and difficulty regarding standardization, including air-drying artifacts and significant cell loss during the fixation process [12]. Moreover, because immunofluorescence staining requires a fluorescent microscope, darkroom, and human resources, it is difficult to perform this method as a routine laboratory test in small- to medium-sized hospitals, and this staining method cannot create permanent specimens. Therefore, we have developed a new urinary podocyte detection method as a routine laboratory test [13]. For urine cytologic specimen preparation, we adopted the SurePath method, a major liquid-based cytology (LBC) method. The advantages of the SurePath method include its higher cell recovery rate, better cell preservation, and standardized technique [13]. Additionally, this method can be used both automatically and manually. In particular, the manual procedure can be used even in small- to medium-sized hospitals and developing countries, provided a slide rack and centrifuge are available. Meanwhile, we leveraged immunoenzyme method for immunostaining. Consequently, we developed a new method for detecting urinary podocytes that combines the SurePath method and immunoenzyme staining using the Wilms tumor 1 (WT1) antibody (podocyte marker) [13]. This method proved effective for detecting glomerular disease, in line with previous reports using PDX immunofluorescence staining [10, 11], and it was also possible to detect crescent formation [14]. Although PDX is considered the best marker for detecting podocytes, in our experiments, the PDX antibody used in previous studies did not work with immunoenzyme staining in urine cytology [10, 11]. Therefore, we chose WT1, which is commonly employed for routine pathologic examination, to detect urinary podocytes in immunoenzyme staining in our previous studies. However, because WT1 is positive for cytoplasm in immunoenzyme staining using alcohol-fixed cytologic specimens [13, 14, 15], researchers accustomed to immunofluorescence staining for nucleus expression may feel uncomfortable with this technique. Thus, we have searched for a PDX antibody that works in immunoenzyme staining in urine cytology. After identifying a commercial PDX antibody, we investigated its effectiveness.

This study assessed whether our urinary podocyte detection method using PDX immunoenzyme staining can represent a noninvasive routine laboratory test for glomerular disease. Therefore, we investigated the frequencies and numbers of PDX- and WT1-positive cells in patients with and without glomerular disease. Meanwhile, we examined the correlation between the number of positive cells and renal function markers. This is the first report investigating glomerular disease using a urinary podocyte detection method combining SurePath and PDX immunoenzyme staining.

Materials and Methods

Patients, Urine Samples, and LBC (SurePath) Slides

This study used data of 79 patients (46 males and 33 females, mean age = 10.6 ± 4.1 years) with glomerular disease who underwent renal biopsy and 51 patients (32 males and 19 females, mean age = 6.0 ± 4.8 years) with nonglomerular disease (underwent orthopedic or plastic surgery) in Shizuoka Children's Hospital. The distribution of each glomerular disease is summarized in Table 1.

Table 1

The distribution of glomerular disease

Histologic diagnosis	n
IgA nephropathy	25
HSPN (IgA vasculitis)	19
MGAs	9
Membranoproliferative glomerulonephritis	6
Focal segmental glomerulosclerosis	4
Lupus nephritis	3
Membranous nephropathy	3
Alport syndrome	2
Non-IgA mesangial proliferative glomerulonephritis	1
Mitochondrial nephropathy	1
Acute tubular injury	1
Autosomal dominant tubulointerstitial kidney disease	1
Poststreptococcal acute glomerulonephritis	1
Tubulointerstitial nephritis	1
Glomerulonephritis	1
Mixed connective tissue disease	1
Total	79

All samples were residual voided urine after a routine test (glomerular disease: median = 30 mL [5–68 mL], nonglomerular disease: median = 23 mL [5–84 mL]), and catheterized urine and bladder washing urine were not included. In the glomerular disease group, voided urine was collected immediately before renal biopsy. The same urine sample was divided into two aliquots and used to prepare two LBC slides. Then, each slide was stained with PDX and WT1 antibodies. LBC slides were prepared using the SurePath (Becton, Dickinson, Franklin Lakes, NJ, USA) modified manual protocols [13].

Immunoenzyme Staining

LBC slides were fixed in 95% ethanol for 30 min and labeled with primary antibodies of PDX (dilution 1:16,000, clone EPR9518; Abcam, Cambridge, UK) and WT1 (dilution 1:1,000, clone 6F-H2; DakoCytomation, Glostrup, Denmark) using the BOND-MAX automatic immunostaining system (Leica Microsystems, Wetzlar, Germany).

Counting of PDX- and WT1-Positive Urinary Cells

The total numbers of PDX- and WT1-positive cells on each LBC slide were counted on the entire smear area under a light microscope (×10 and ×40 magnification) by 2 authors (J.S. and T.I.) blinded to the patients' diagnoses. Some patients were too young to provide constant amounts of samples. Therefore, to eliminate the effects of urine volume, the number of podocytes in 10 mL of urine was calculated for each patient.

First, we compared the frequencies and numbers of urinary PDX- and WT1-positive cells between the glomerular disease and nonglomerular disease groups. Second, we compared the numbers of PDX- and WT1-positive cells among 3 groups (i.e., IgA nephropathy, Henoch-Schönlein purpura nephritis [HSPN], and minor glomerular abnormalities [MGAs] groups), which are the top three glomerular diseases in this study. Third, we examined the correlations of the numbers of positive cells with renal function markers such as serum creatinine, urine protein-to-creatinine ratio, and occult blood in the urine in patients with glomerular disease.

Laboratory Test Data

Serum and urine creatinine contents were measured by the enzymatic method (Cygnus Auto CRE; Shino-Test Corp., Tokyo, Japan) using a BioMajesty JCA-BM9130 automatic biochemical analyzer (JEOL Ltd., Tokyo, Japan). Urine protein levels were measured by the pyrogallol red method (Micro TP-AR 2; Wako Pure Chemical Industries, Osaka, Japan) using a BioMajesty JCA-BM9130 automatic biochemical analyzer (JEOL Ltd.). Occult blood in urine scores of 1+ or more on dipstick testing (Uropaper α||| Eiken; Eiken Chemical, Tokyo, Japan) as read by an automated dipstick reader (US-3500; Eiken Chemical, Tokyo, Japan) was defined as positive.

Statistical Analysis

The Mann-Whitney U test and Kruskal-Wallis test were used where appropriate. p < 0.05 indicated statistical significance. All analyses were performed using StatFlex software (version 7.0; Artec Inc., Osaka, Japan).

Results

Comparisons of PDX- and WT1-Positive Cell Counts between the Groups

PDX- and WT1-positive cells were strongly positive in the cytoplasm. These positive cells appeared singly; they were 20–40 μm in maximum length and round or oval in shape (shown in Fig. 1). In addition, these positive cells were occasionally encased by a cast.

Fig. 1

Images of immunoenzyme staining with PDX and WT1 antibodies showing urinary podocytes. PDX- and WT1-positive cells are scattered between leukocytes (×1,000 magnification).

The numbers of PDX- and WT1-positive cells were significantly higher in the glomerular disease group than in the nonglomerular disease group (shown in Fig. 2). In addition, there were no significant differences in PDX- and WT1-positive cell numbers within the nonglomerular disease group. However, in the glomerular disease group, the number of PDX-positive cells was significantly higher than that of WT1-positive cells.

Fig. 2

Number of PDX- and WT1-positive cells in the glomerular disease and nonglomerular disease groups.

The best cutoff for urinary PDX-positive cells to differentiate glomerular disease from nonglomerular disease was 3.5 cells/10 mL (sensitivity = 67.1%, specificity = 100%, positive predictive value = 100%, negative predictive value = 66.2%), and the cutoff for WT1-positive cells was 1.2 cells/10 mL (sensitivity = 43.0%, specificity = 100%, positive predictive value = 100%, negative predictive value = 53.1%). These cutoffs produced area under the curves of 0.835 and 0.705 (shown in Fig. 3), respectively, and revealed that glomerular disease could be detected with moderate accuracy (Table 2).

Fig. 3

Receiver operating characteristic curve for the cutoff value of urinary PDX- and WT1-positive cells.

Table 2

Sensitivity and specificity of PDX- and WT1-positive cells using the best cutoff value

	Nonglomerular disease (n = 51), n (%)	Glomerular disease (n = 79), n (%)	p vaiue
PDX-positive (≥3.5 cells/10 mL)	0 (0)	53 (67.1)	<0.001
PDX-negative (<3.5 cells/10 mL)	51 (100)	26 (32.9)
WT1-positive (≥1.2 cells/10 mL)	G (0)	34 (43.0)	<0.001
WT1-negative (<1.2 cells/10 mL)	51 (100)	45 (57.0)

Comparison of PDX- and WT1-Positive Cell Counts among the Three Glomerular Disease Groups

The numbers of PDX- and WT1-positive cells were significantly higher in the IgA nephropathy and HSPN groups than those in the MGAs group (shown in Fig. 4). On the other hand, there were no significant differences in PDX- and WT1-positive cell numbers between the IgA nephropathy and HSPN groups.

Fig. 4

Number of PDX- and WT1-positive cells in the three glomerular disease groups. IgA, IgA nephropathy; HSPN, Henoch–Schönlein purpura nephritis; MGA, Minor glomerular abnormalities.

Correlations of PDX- and WT1-Positive Cell Counts with Renal Function Markers in Glomerular Disease

In the glomerular disease group, the numbers of PDX- and WT1-positive cells were not correlated with patient sex, height, weight, age, serum creatinine levels, urine protein-to-creatinine ratio, and occult blood in urine.

Discussion

PDX- and WT1-positive cells exhibited cytoplasmic expression in this study. Researchers accustomed to immunofluorescence staining may raise concerns regarding the WT1 immunoenzyme staining results (cytoplasmic positivity of WT1). Previous studies have reported that WT1 is involved in transcriptional regulation within the nucleus and RNA metabolism and translational regulation in the cytoplasm [16, 17]. Likewise, Western blotting of nuclear and cytoplasmic fractions using WT1 antibody revealed that WT1 predominantly inhabited the nucleus, although cytoplasmic fractions were also positive for WT1 [18]. For these reasons, in this study, WT1-positive cells displayed cytoplasmic expression.

The PDX- and WT1-positive cells of various sizes and shapes were observed in this study. These morphologic features are consistent with previous reports on PDX immunofluorescence staining [10, 11, 19]. One hypothesis explaining the various shapes and sizes of podocytes is miniaturization attributable to apoptosis [8]. Conversely, hypertrophy may be induced to allow the remaining podocytes to cover the GBM in denuded areas following the detachment of other podocytes [20]. Although podocytes are considered terminally differentiated cells, they have been reported to reenter the cell cycle in several glomerular diseases [20, 21, 22]. Podocytes proceed until the mitosis phase under stress or injury; however, as they cannot assemble mitotic spindles, cytokinesis is impossible. Consequently, these podocytes show hypertrophy and become multinucleated [20]. Additionally, it has been demonstrated that both podocytes and active PECs are positive for PDX and WT1 in several glomerular diseases [11, 14, 23, 24]. Therefore, the result of this study that urinary PDX- and WT1-positive cells had diverse morphologic features may be explained by the fact that these cells are derived from podocytes and active PECs. On the other hand, PDX- and WT1-positive cells' morphology was mainly round shape and about 20 μm in maximum length, and very few cells were present on the slides. Hence, it is difficult to detect these cells only by cytomorphology, and thus, immunocytochemistry is necessary. In this study, PDX- and WT1-positive cells were occasionally encased by a cast. Because casts are formed as the molds for the lumina of renal tubules and collecting ducts, these positive cells inside the cast were suggested to have arisen from nephrons [14].

Wang et al. [25] reported that in a pediatric study, PDX-positive cells were detected in 53.8% (35/65) of patients in the glomerular disease group, and the number of PDX-positive cells was higher in this group than in the healthy control group. Their differences were smaller than our findings, although the data cannot be easily compared because the urine volume, cytologic smear preparation method, and immunocytochemistry method were different. One possible explanation for the increased frequency of PDX-positive cells in our study is the improvement of cell recovery using the SurePath method [26, 27, 28]. Meanwhile, another study found that PDX immunofluorescence staining of urine sediments was positive in 92.5% (62/65) of children with glomerular disease [10]. However, this discrepancy is likely caused by the inclusion of positive casts and granules in addition to PDX-positive cells in their study. The sensitivity and specificity of WT1-positive cells in the present study (sensitivity = 43.0% and specificity = 100%) were similar to those in a previous study on adult patients with glomerular disease (sensitivity = 50.0% and specificity = 100%) [13]. The results of the present study indicated that immunoenzyme staining using the PDX antibody can detect glomerular disease with a higher accuracy than that using WT1 antibody. Some researchers may prefer PDX immunofluorescence staining, but because immunoenzyme staining is routinely performed as a pathologic examination, we consider it will be better suited for routine tests [14].

This study targeted 79 cases of glomerular diseases, but the number of cases was tiny in most of the glomerular disease types. Therefore, we compared the numbers of PDX- and WT1-positive cells among the 3 groups (IgA nephropathy, HSPN, and MGAs groups), which had a relatively large number of cases. As a result, the numbers of PDX- and WT1-positive cells were significantly higher in the IgA nephropathy and HSPN groups than that in the MGAs group. On the other hand, there were no significant differences in these positive cell numbers between the IgA nephropathy and HSPN groups. Our results were similar to those of previous studies on the detection of PDX-positive cells using immunofluorescence staining [7, 29, 30, 31].

The lack of correlations of PDX- and WT1-positive cell counts with renal function markers likely has several explanations. The number of urinary PDX-positive cells was significantly higher in patients with focal segmental glomerulosclerosis than in those with membranous nephropathy or minimal change nephrotic syndrome in prior research [29, 30]. The number of urinary PDX-positive cells was significantly higher in patients with active IgA nephropathy or HSPN than in those with inactive disease cases [7, 31]. Furthermore, in minimal change nephrotic syndrome, it is known that a large amount of protein appears in the urine even though PDX- and WT1-positive cells are rarely observed [13, 29, 30]. In IgA nephropathy, hematuria is usually present, though proteinuria is rare in the early stages. As mentioned previously, urinary PDX-positive cells and some renal function markers reflect both glomerular disease type and activity, but the current study population included patients with various disease types and activities. These findings potentially explain the lack of correlations of PDX- and WT1-positive cell counts with serum creatinine, urine protein-to-creatinine ratio, and occult blood in the urine in the present study. From this result, we would like to note that our method is a noninvasive biomarker independent of other tests, and glomerular disease can be detected with high accuracy using our cutoffs for PDX-positive cells.

In conclusion, because our urinary podocytes detection method can be standardized and it detected glomerular disease with high accuracy, we believe this method could emerge as a noninvasive routine laboratory test for various glomerular diseases. Future work should focus on the relationship between the activity of individual glomerular diseases and PDX-positive cell counts.

Statement of Ethics

This study was approved by the Ethics Committee of Shizuoka Children's Hospital (no. 2020-28) and Kobe University Graduate School of Health Sciences (no. 977) and conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from each patient's parent or legal guardian.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work was financially supported by the Shizuoka General Hospital Encouragement Research Grant and the Japanese Society of Clinical Cytology Grant.

Author Contributions

Methodology: Ohsaki H. Formal analysis: Sakane J., Inoue T., Nakamura A., and Ohsaki H; resources: Kitayama H., Yamada M., Miyama Y., Kawamura H., Iwafuchi H., and Ohsaki H; writing − original draft: Ohsaki H; writing − review and editing: Kitayama H., Nakamura A., Kamoshida S., and Ohsaki H; and supervision: Ohsaki H.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.