Podocytes are lost from glomeruli before completing apoptosis

Abstract

Although apoptosis of podocytes has been widely reported in in vitro studies, it has been less frequently and less definitively documented in in vivo situations. To investigate this discrepancy, we analyzed the dying process of podocytes in vitro and in vivo using LMB2, a human (h)CD25-directed immunotoxin. LMB2 induced cell death within 2 days in 56.8 ± 13.6% of cultured podocytes expressing hCD25 in a caspase-3, Bak1, and Bax-dependent manner. LMB2 induced typical apoptotic features, including TUNEL staining and fragmented nuclei without lactate dehydrogenase leakage. In vivo, LMB2 effectively eliminated hCD25-expressing podocytes in NEP25 mice. Podocytes injured by LMB2 were occasionally stained for cleaved caspase-3 and cleaved lamin A but never for TUNEL. Urinary sediment contained TUNEL-positive podocytes. To examine the effect of glomerular filtration, we performed unilateral ureteral obstruction in NEP25 mice treated with LMB2 1 day before euthanasia. In the obstructed kidney, glomeruli contained significantly more cleaved lamin A-positive podocytes than those in the contralateral kidney (50.1 ± 5.4% vs. 29.3 ± 4.1%, P < 0.001). To further examine the dying process without glomerular filtration, we treated kidney organoids generated from nephron progenitor cells of NEP25 mice with LMB2. Podocytes showed TUNEL staining and nuclear fragmentation. These results indicate that on activation of apoptotic caspases, podocytes are detached and lost in the urine before nuclear fragmentation and that the physical force of glomerular filtration facilitates detachment. This phenomenon may be the reason why definitive apoptosis is not observed in podocytes in vivo.

NEW & NOTEWORTHY This report clarifies why morphologically definitive apoptosis is not observed in podocytes in vivo. When caspase-3 is activated in podocytes, these cells are immediately detached from the glomerulus and lost in the urine before DNA fragmentation occurs. Detachment is facilitated by glomerular filtration. This phenomenon explains why podocytes in vivo rarely show TUNEL staining and never apoptotic bodies.

INTRODUCTION

Podocytes are highly differentiated cells and play an essential role in the glomerular filtration barrier (1, 2). However, podocytes have limited proliferation and regeneration abilities; therefore, their loss is irreversible, which eventually causes glomerulosclerosis and renal failure (2–4).

There are three ways of podocyte loss, namely, detachment, necrosis, and apoptosis. Detached podocytes can be observed in the urine. Many studies have shown that the urine of patients with various glomerular diseases contains podocytes labeled by specific markers, such as podocalyxin, and that the number of urinary podocytes reflects disease activity (5–7). Detachment can be facilitated by increased mechanical stress and/or decreases in adhesion molecules that link podocytes and glomerular basement membrane. The former is caused by hyperfiltration, and the latter is caused by dysregulation of integrins in inflammatory glomerular diseases (8–12). Necrosis has been much less frequently examined than detachment and apoptosis due to technical limitations and difficulties in detecting necrotic and prenecrotic podocytes in vivo.

However, apoptosis of podocytes has been documented in numerous in vitro studies. Various factors, including puromycin aminonucleoside, adriamycin, high glucose, oxidized low-density lipoprotein, transforming growth factor-β1, Smad7, or mutations in Wilms’ tumor 1 (WT1), have been shown to induce apoptosis in cultured podocytes at high rates via various pathways, including p38, NF-κB, and Notch (13–21). In these studies, apoptosis was often detected by multiple methods, including TUNEL, detection of activated apoptotic caspases, or translocation of phosphatidylserine on the surface of the plasma membrane. Morphologically, nuclear fragmentation and apoptotic bodies were often observed. In contrast, apoptosis in podocytes in in vivo studies was less frequent, less definitive, and sometimes inadequately documented. In many studies, apoptosis was detected by TUNEL assays. TUNEL assays are believed to be specific for apoptosis and are now known to show other types of cell death or conditions with active DNA repair (22). Most of the reported TUNEL-positive podocytes did not show nuclear fragmentation. In addition, no reports have shown transmission electron microscopy (TEM) images of apoptotic bodies, the morphological hallmark of apoptosis (22–24). The reason for this discrepancy between in vitro and in vivo podocytes has not been clarified.

Previously, we established a transgenic mouse line, NEP25, which selectively expresses human (h)CD25 in podocytes (25). By injection of an hCD25-targeted recombinant immunotoxin, anti-Tac(Fv)-PE38 (LMB2), selective podocyte injury can be induced in a dose-dependent manner (25, 26). LMB2 is composed of the antigen-binding portion of the anti-hCD25 antibody and the truncated form of Pseudomonas exotoxin. LMB2 is incorporated selectively in hCD25-expressing cells, and the toxin portion is released to the cytosol, which inhibits protein synthesis by ADP-ribosylation of elongation factor 2 (26–28). One possible mechanism of the lethal effect of LMB2 is depletion of Mcl-1, the major antiapoptotic protein in podocytes, which has a very high turnover (29). After injection of LMB2 into NEP25 mice, podocytes were lost to various degrees depending on dosage, and mice developed proteinuria, nephrotic syndrome, and glomerulosclerosis. However, TUNEL-positive podocytes and morphologically definitive apoptosis of podocytes have never been observed in any experimental conditions. This is contrasting to the fact that proximal tubular cells taking up large amount of an immunotoxin showed typical apoptotic figures (30).

In the present study, using the LMB2-induced podocyte injury model, we studied the dying process in podocytes in vitro and in vivo and clarified the reason why apoptosis is not observed in podocytes in vivo.

MATERIALS AND METHODS

In Vitro Experiments

By perfusion with Dynabeads (Invitrogen, Waltham, MA), glomeruli were isolated from wild-type female mice (C57BL/6J strain) at 5–6 wk of age (31). These cells were cultured in DMEM-F-12 (Sigma-Aldrich, St. Louis, MO) containing 5% FCS (Nippon Bio-Test Laboratories, Saitama, Japan) with 0.5% insulin-transferrin-selenium A agent (ThermoFisher Scientific, Waltham, MA) and antibiotic-antimycotic agent (ThermoFisher Scientific) on a collagen type I-coated dish at 37°C (in 5% CO₂). After 7 days of incubation, sprouted podocytes were detached by 0.25% trypsin (Wako, Osaka, Japan) and passaged. Cells after the first or second passages were used for experiments.

Plasmid DNA was introduced into cells by electroporation using the Neon Transfection System (Invitrogen) according to the manufacturer’s instructions. Briefly, 1 × 10⁵ podocytes suspended in 10 µL of Buffer R (Invitrogen) were mixed with 0.9 µg of hCD25 expression plasmid (32) or mock plasmid DNA together with 0.1 µg of enhanced green fluorescent protein (EGFP) expression plasmid DNA to monitor transfected cells. In some experiments, membrane-bound EGFP or tdTomato was used instead of EGFP. The electroporation conditions were 1,500 V, 20 ms, and one pulse. A total of 5–10 × 10⁴ cells/well were seeded in a collagen type I-coated 24-well plate.

After 48–72 h, EGFP was imaged using an Olympus IX70 Fluorescence Microscope with marking on the bottom of the culture plates to fix the visual fields. LMB2 (1 nM) was administered and cultured in medium containing 0.5% FCS for 24 h. After being stained with propidium iodide (1 µg/mL) and Hoechst 33342 (5 µg/mL), cells in the same visual field were imaged. Some cells were stained for cleaved caspase-3 or TUNEL and similarly imaged.

Supernatants were collected from wells 24 h after LMB2 administration, and lactate dehydrogenase (LDH) activity was determined using an LDH cytotoxicity assay kit (Dojindo Molecular Technologies, Mashiki, Kumamoto, Japan) according to the manufacturer’s instructions. The activity relative to that of 100% necrotic cells was calculated.

For analysis of the cell death pathway, the effect of the following reagents was tested: 10 µM Ac-DEVD-CHO (caspase-3 inhibitor, Cat. No. 10404, Biotium, Fremont, CA), 100 µM Ac-YVAD-cmk (caspase-1 inhibitor, inh-yvad, InvivoGen), 100 µM necrostatin-1 (necroptosis inhibitor, Cat. No. sc-200142, Santa Cruz Biotechnology, Dallas, TX), 10 µM ferrostatin-1 (ferroptosis inhibitor, Cat. No. SML0583, Sigma-Aldrich), and 10 nM rapamycin (autophagy inhibitor, Cat. No. R-5000, LC Laboratories, Woburn, MA).

For analysis of the role of Bak1 and Bax, siRNAs for Bak1 and Bax (NM_007523 and NM_007527, Sigma) were introduced into cells using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions 24 h after electroporation. After incubation for 96 h, LMB2 (1 nM) was administered and similarly analyzed.

Animal Experiments

The protocols for animal experiments were approved by the Animal Experimentation Committee of Tokai University School of Medicine in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

For the induction of podocyte injury, NEP25 mice (n = 6, 20–50 wk of age, female, on a C57BL/6 genetic background) were intravenously injected with 1.25 ng/g body wt LMB2. Urine samples were collected before and 7 days after the injection of LMB2, and the kidneys were then harvested. Control NEP25 mice (n = 3, 20–50 wk of age, female) were intravenously injected with saline and similarly analyzed.

In addition, eight NEP25 mice (12–20 wk of age, 3 male and 5 female) were subjected to unilateral ureteral obstruction (UUO) surgery 6 days after the injection of 0.625 ng/g body wt LMB2. Mice were euthanized 24 h later, and both kidneys were harvested and histologically analyzed.

Morphological Analyses of NEP25 Mice

Kidneys were fixed in 4% paraformaldehyde (PFA) for 2 h at 4°C, incubated in 30% sucrose-PBS overnight, embedded in OCT compound (Sakura Finetek, Osaka, Japan), and frozen. For immunofluorescence staining, 6-µm-thick frozen sections were cut, air-dried, rinsed in PBS, and used for immunostaining and TUNEL staining. Slides were observed with an Olympus IX70 Fluorescence Microscope (Olympus) or confocal microscope (LSM880, Zeiss, Jena, Germany).

For TEM, some kidneys were perfused with 4% PFA, fixed in 2% glutaraldehyde, and postfixed with 1% osmium tetroxide. After dehydration, samples were embedded in an epoxy resin, cut into 90-nm-thick sections and observed and imaged with a JEM-1400 transmission electron microscope (JEOL, Tokyo, Japan).

For scanning electron microscopy, some kidneys were perfused with 4% PFA, fixed in 2% glutaraldehyde, and postfixed with 1% osmium tetroxide. After dehydration, samples were transferred into t-butyl alcohol and frozen. After being coated with osmium, freeze-dried samples were observed and imaged with a JSM-6510 LV scanning electron microscope (JEOL). Twenty glomeruli were randomly selected in each kidney and observed.

Cleaved lamin A-positive podocytes were counted in paraffin-embedded sections subjected to double staining for cleaved lamin A and synaptopodin. In addition, podocytes in paraffin-embedded sections subjected to double staining for WT1 and synaptopodin were counted. As WT1 is often ectopically positive in cells other than podocytes and disappears earlier than synaptopodin in injured podocytes of NEP25 mice, we counted synaptopodin-positive nuclei (WT1 positive or hematoxylin positive) and regarded them as podocyte numbers. In both analyses, all glomeruli (54–96 glomeruli per kidney), except tangential glomerular sections smaller than 40 µm in diameter, were analyzed.

Analyses of Urinary Cells

Spot urine samples were collected in tubes containing 50 µL of 4% PFA. Urinary cells were attached to APS-coated microscope slides using Cytospin 3 (ThermoFisher Scientific), and slides were further fixed in 4% PFA at room temperature for 15 min. After the slides had been rinsed in PBS, they were stained for podocalyxin, cleaved caspase-3, or TUNEL.

Organoid Experiments

Nephron progenitor cells were obtained from NEP25 mouse embryonic kidneys by the culture purification method as previously reported with a few modifications (33). Concentrations of fibroblast growth factor-2 and 2-mercaptoethanol in the NPSR medium were increased to 300 ng/mL and 0.1 mM, respectively. Nephron progenitor cell aggregates were passaged every 4 days. For the induction of differentiation, aggregates with a diameter of 1 mm were transferred onto Transwell membranes and cultured on KR5-CF medium containing 4.5 µM CHIR99021 and 200 ng/mL fibroblast growth factor-2 for 2 days and then on KR5 medium. Six days later, LMB2 (20 nM) was added to the medium and cultured for additional 3 days. Organoids were fixed in 4% PFA, embedded in paraffin, and histologically analyzed.

Immunostaining and TUNEL Assay

Cultured cells and frozen samples were fixed in 4% PFA and permeabilized in 0.1% Triton X-100-PBS. The following primary antibodies were used at the indicated dilutions: cleaved caspase-3 (Cat. No. 9664, Cell Signaling Technology, 1/500), cleaved lamin A (Cat. No. 2035, Cell Signaling Technology, 1/400), podocalyxin (Cat. No. MAB1556, R&D, 1/8,000), nephrin (Cat. No. AF3159, R&D, 1/160, for frozen), nephrin (Cat. No. GP-N2, Progen, 1/200, for paraffin), Mcl-1 (Cat. No. ab32087, Abcam, 1/4,000), desmin (Cat. No. M0760, Dako, 1/50), synaptopodin (Cat. No. 65194, Progen, 1/100), WT1 (Cat. No. ab89901, Abcam, 1/200), EGFP (Cat. No. MBL598, MBL, 1/4,000), and hCD25 (Clone 33B3.1, Cat. No. ab19683, Abcam, 1/50). For podocalyxin staining, paraffin sections were pretreated with 0.1% trypsin; for other types of immunostaining, paraffin sections were heated in citrate buffer (pH 6.0). For cleaved caspase-3, cleaved lamin A, podocalyxin, nephrin (R&D), Mcl-1, EGFP, WT1, and hCD25, the CanGet signal (Toyobo) was used to amplify the staining signals. Methods for antibody validation are shown in the Supplemental Material.

TUNEL staining was performed with an in situ Apoptosis Detection Kit (TaKaRa Bio) according to the manufacturer’s instructions.

Paraffin sections from Pod-hCD25T/G mosaic mice (34), the α-actinin4 model (35), and the human immunodeficiency virus (HIV)-1-associated nephropathy model (36) were stained for cleaved caspase-3 and desmin. Paraffin sections were prepared from a part of the renal tissue obtained by renal biopsy at Tokai University School of Medicine and stained for cleaved caspase-3 and nephrin. In Pod-hCD25T/G mosaic mice, Cre recombinase reciprocally generates either an hCD25-expressing allele or an EGFP-expressing allele in podocytes. Consequently, nearly half of the podocytes express hCD25 but not EGFP, whereas the rest express EGFP but not hCD25. LMB2 directly acts only on hCD25-expressing podocytes. After LMB2 administration, EGFP-expressing podocytes were also secondarily injured, although they lacked hCD25 (34).

Images of sections double immunostained for nephrin/cleaved caspase-3, nephrin/cleaved lamin A, and podocalyxin/TUNEL were obtained with a confocal laser scanning microscope (LSM880, Zeiss) and processed with Zen 2.3 lite. Other immunofluorescent images were obtained using an Olympus IX70 microscope and processed using Photoshop 2022 (Adobe). Images of sections subjected to periodic acid-Schiff and immunohistochemical staining were obtained using Axioplan 2 (Zeiss) and processed using Photoshop 2022.

Statistical Analysis

Results are expressed as means and 95% confidence intervals. One-way ANOVA was applied for comparisons among three or more groups. P values were corrected using Bonferroni’s (Fig. 1, C–E and J) or Dunnett’s (Fig. 1K) method to minimize the inflation of type I error due to multiple comparisons. The urinary protein-to-creatinine ratio was log-transformed and compared using a paired t test. A paired t test was used for comparisons between obstructed and nonobstructed kidneys in the UUO experiment. For other comparisons between two groups, an unpaired t test was applied. All statistical analyses were performed using the free statistical software EZR version 4.0.3 (37). Statistical significance was set at a P value of <0.05.

Figure 1.

LMB2 induces apoptosis in cultured podocytes expressing human (h)CD25. A: podocytes were cotransfected with plasmids expressing enhanced green fluorescent protein (EGFP) and hCD25 at a 9:1 ratio (left). All EGFP⁺ podocytes were stained with hCD25. After LMB2 administration, EGFP disappeared, and some shrunken cells with condensed DAPI or cell debris were stained for hCD25. Control podocytes were cotransfected with mock- and EGFP-expressing plasmids (right). hCD25 staining was negative and EGFP remained unchanged after LMB2 administration. B: podocytes were transfected with plasmids expressing hCD25 and EGFP and cultured with or without LMB2 for 24 h. LMB2 markedly diminished EGFP fluorescence and induced propidium iodide (PI) uptake. C: rates of EGFP loss, which were obtained by comparing two images before and after the treatment, with or without LMB2 and hCD25. D: rates of PI incorporation with or without LMB2 and hCD25. E: lactate dehydrogenase (LDH) activity in the culture media relative to that of the well where 100% of the cells were dead by necrosis. F: podocytes were transfected with plasmids expressing hCD25 and membrane-bound EGFP and treated with LMB2. Twenty-four hours later, cleaved caspase-3 (cCasp-3) was observed in the cytosol of some EGFP⁺ cells, which had condensed nuclear staining. G: rates of cCasp-3⁺ cells with or without LMB2 treatment. H: podocytes were transfected with plasmids expressing hCD25 and tdTomato and treated with LMB2. Twenty-four hours later, TUNEL was often positive in condensed nuclei. Most TUNEL⁺ cells lost tdTomato but rarely retained tdTomato, as shown. I: rates of TUNEL⁺ cells with or without LMB2 treatment. J: effect of a caspase-3 (Casp-3) inhibitor on EGFP loss induced by LMB2. K: effect of Bak1 and Bax inhibition on EGFP loss induced by LMB2. Values are presented as means ± 95% confidence interval. Data in G and I were analyzed using an unpaired t test. Data were analyzed using one-way ANOVA followed by a Bonferroni’s test in C, D, and J and by Dunnett’s method in K. Scale bars = 200 μm in A and B as well as the low-power fields in F and H and 50 μm in the high-power fields in F and H.

RESULTS

LMB2 Induced Apoptosis in hCD25-Expressing Cultured Podocytes

Primary cultured podocytes were transfected with an hCD25-expressing plasmid. The EGFP-expressing plasmid was cotransfected to identify transfected cells and monitor cell death because loss of EGFP is a very sensitive and early marker of cell death (38). The efficiency of transfection estimated by EGFP expression was ∼60%. Immunostaining for hCD25 revealed that EGFP faithfully represented hCD25 in cotransfected podocytes without LMB2 (Fig. 1A). Twenty-four hours after LMB2 administration, 56.8 ± 13.0% (95% confidence interval) of the initially EGFP-expressing cells lost EGFP emission (Fig. 1, A–C); however, in phase-contrast images, most cells were maintained at their original positions, indicating cell death rather than detachment. Some shrunken podocytes with condensed DAPI or cell debris were stained with hCD25 (Fig. 1A). In addition, 13.6 ± 1.2% of the initial EGFP-expressing cells incorporated propidium iodide (Fig. 1D). Positive cleaved caspase-3 staining was observed in 19.6 ± 2.2% of the initially EGFP-positive cells, and positive TUNEL staining was observed in 4.5 ± 1.6% of the initially EGFP-positive cells (Fig. 1, F–I). The cleaved caspase-3-positive cells often shrank with condensed nuclei (Fig. 1F). These phenomena were not observed in hCD25-transfected cells without LMB2 and mock-transfected cells with or without LMB2. Although the majority of the hCD25-transfected cells lost EGFP within 24 h after LMB2 treatment, LDH activity in the culture medium did not significantly increase at the same time point (Fig. 1E).

The supernatant of the LMB2-treated wells contained only a small number of cells, totaling less than 2 × 10³ cells (2–4% of the seeded cells). Three percent of these cells were positive for EGFP (Supplemental Fig. S1A).

We next tested the effects of various inhibitors on the EGFP disappearance induced by LMB2 to identify the responsible pathway. Ac-DEVD-CHO, a caspase-3 inhibitor, attenuated the disappearance of EGFP by 45.5% (Fig. 1J). In contrast, Ac-YVAD-cmk (pyroptosis inhibitor), necrostatin-1 (necroptosis inhibitor), ferrostatin-1 (ferroptosis inhibitor), and rapamycin (autophagy inhibitor) did not show significant effects (Supplemental Fig. S1B).

We next tested the roles of key triggering molecules of the intrinsic apoptotic signaling, Bak1 and Bax. Transfection of Bak1 and Bax siRNAs in cultured podocytes reduced mRNA by 95.8% and 84.3%, respectively. Bak1 siRNA attenuated the EGFP disappearance induced by LMB2 by 77.6%, and Bax siRNA attenuated that by 28.4%. Moreover, a mixture of Bak1 and Bax siRNAs attenuated the EGFP disappearance induced by LMB2 by 88.8% (Fig. 1K and Supplemental Fig. S1, C and D). LMB2 inhibits protein synthesis in hCD25-expressing cells. Our previous study indicated that Mcl-1 is the most abundantly expressed antiapoptotic Bcl-2 family member (35). Mcl-1 protein has a very short half-life, estimated to be <1 h. Consistent with this, Mcl-1 staining disappeared in hCD25-transfected podocytes 24 h after LMB2 treatment (Supplemental Fig. S2A), indicating that the depletion of Mcl-1 protein may facilitate the initiation of apoptosis.

These results collectively indicated that LMB2 induces typical intrinsic apoptosis dependent on caspase-3 in hCD25-expressing cultured podocytes.

Caspase-3 Activation but Not Full Apoptosis in Podocytes of NEP25 Mice

We next investigated the dying process of podocytes damaged by LMB2 in vivo using NEP25 mice. NEP25 mice express hCD25 selectively in podocytes. LMB2 dose dependently injured podocytes in NEP25 mice. As observed in hCD25-expressing cultured podocytes, Mcl-1 staining in podocytes of NEP25 mice disappeared after LMB2 administration (Supplemental Fig. S2B), indicating that LMB2 can cause proapoptotic conditions in podocytes of NEP25 mice.

Under any experimental condition (LMB2: 0.625–25 ng/g body wt, duration: 1 to >28 days), we never observed TUNEL-positive podocytes in NEP25 mice. We therefore investigated whether an earlier event of apoptosis, such as activation of caspase-3, could be observed in the podocytes of NEP25 mice. Immunostaining revealed that cleaved caspase-3 was occasionally positive in the podocytes of NEP25 mice. Seven days after 1.25 ng/g body wt LMB2 treatment, cleaved caspase-3 was frequently observed; 14.4 ± 3.8% of the glomeruli in the frozen sections contained cleaved caspase-3-positive cells, most of which were identified as podocytes by nephrin staining (Fig. 2, B and C, and Supplemental Fig. S3A). With this dosage of LMB2, NEP25 mice showed severe glomerular damage and complete loss of the podocyte marker proteins WT1, nephrin, and podocin ∼14 days after the injection (25). On the seventh day, NEP25 mice showed mild morphological changes, including focal proteinaceous cast formation, occasional nuclear deformation (Fig. 2A, arrow), vacuolar degeneration, and podocyte detachment. However, they showed severe proteinuria, with a urinary protein-to-creatinine ratio of 69.2 ± 2.86 g/g compared with the baseline value of 0.06 ± 0.01 g/g (P < 0.001).

Figure 2.

Caspase-3 activation in podocytes of NEP25 mice after LMB2 injection. NEP25 mice (n = 11, female) were injected with 1.25 ng/g body wt LMB2 and analyzed 7 days later. The control group included NEP25 mice (n = 3, female) injected with vehicle. A: periodic acid-Schiff staining. NEP25 mice with LMB2 mice showed mild morphological changes, such as focal proteinaceous cast formation. Some glomeruli contained podocytes with distorted nuclei (black arrow). B: confocal microscopic images of nephrin and cleaved caspase-3 (cCasp-3) in NEP25 mice treated with LMB2. cCasp-3 was localized to the cytoplasm of podocytes labeled with nephrin (arrow). C: frequency of cCasp-3⁺ podocytes, as evaluated with fluorescence microscopy. NEP25 mice treated with LMB2 (N = 6) contained more cCsap3⁺nephrin⁺ cells than control mice (n = 3). D: confocal microscopic images of nephrin and cleaved lamin A (cLamin A) in NEP25 mice treated with LMB2. cLamin A was localized to the perinuclear cytoplasm of podocytes labeled with nephrin (arrow). Note that lamin A is a component of the nuclear membrane, but it is rapidly translocated to the cytoplasm on cleavage (39, 40). E: frequency of cLamin A⁺ podocytes, as evaluated with fluorescence microscopy. NEP25 mice treated with LMB2 (n = 6) contained more cLamin A⁺nephrin⁺ cells than control mice (n = 3). F: transmission electron micrographs of podocytes in NEP25 mice treated with LMB2 showing swollen mitochondria and ruptured plasma membrane (arrow). Transmission electron micrographs of NEP25 mice without LMB2 treatment are provided in Supplemental Fig. S3C. G: phase-contrast, podocalyxin (Podxl1), DAPI, and merged images of the urine of control (top) and NEP25 mice treated with LMB2 (bottom). H: number of urinary podocytes. More Podxl1⁺ cells were observed in NEP25 mice treated with LMB2 (n = 11) than in control mice (n = 3). I: cCasp-3⁺ podocytes in the urinary sediment of NEP25 mice treated with LMB2. J: rate of cCasp-3⁺ podocytes in the urinary sediment. The urine of NEP25 mice treated with LMB2 (n = 5) contained more cCasp-3⁺Podxl1⁺ cells. K: cLamin A⁺ podocytes in the urinary sediment of NEP25 mice treated with LMB2. L: rate of cLamin A⁺ podocytes in the urinary sediment. The urine of NEP25 mice treated with LMB2 (n = 3) contained more cLamin A⁺Podxl1⁺ cells. M: TUNEL⁺ podocytes in the urinary sediment of NEP25 mice treated with LMB2. N: rate of TUNEL⁺ podocytes in the urinary sediment. The urine of NEP25 mice treated with LMB2 (n = 3) contained more TUNEL⁺Podxl1⁺ cells. Values are presented as means ± 95% confidence intervals. The data were analyzed using an unpaired t test. Scale bars = 100 μm in the low-power fields of A and G, 2 μm in F, and 50 μm in the other images.

We then found that cleaved lamin A, a product of activated caspase-6, could be more clearly detected (Fig. 2, D and E, and Supplemental Fig. S3B). In line with our previous observations, we confirmed that TUNEL staining was not observed. TEM revealed that podocytes were severely damaged with vacuolar degeneration, foot process effacement, scattered microvilli, and mitochondrial swelling accompanied by loss of cristae. Notably, a few podocytes showed rupture of the plasma membrane (Fig. 2F and Supplemental Fig. S3C). However, no apoptotic bodies, the hallmarks of apoptosis, were observed.

Urinary Podocytes Were Observed in LMB2-Injured NEP25 Mice

The above findings suggest that podocytes are lost in the urine before nuclear fragmentation just after the activation of caspase-3. We then analyzed the urinary sediments of NEP25 mice 7 days after injection of 1.25 ng/g body wt LMB2. The urine contained, on average, 2.54 ± 0.45 podocytes/µL identified by podocalyxin staining (Fig. 2, G and H). Among these, 39.2 ± 7.3% and 17.8 ± 8.7% were positively stained for cleaved caspase-3 and cleaved lamin A, respectively (Fig. 2, I–L). Moreover, 20.5 ± 4.0% of urinary podocytes were positively stained for TUNEL (Fig. 2, M and N).

Glomerular Filtration Facilitates Detachment of Podocytes With Activated Caspase-3

We then examined the effect of glomerular filtration on the detachment of podocytes with activated apoptotic caspases. For this purpose, we performed UUO in NEP25 mice 6 days after injection of LMB2 and euthanized them the next day. Double immunostaining for synaptopodin and cleaved lamin A revealed that 29.3 ± 4.1% of the glomeruli contained cleaved lamin A-positive podocytes in the nonobstructed kidney. In contrast, significantly more (50.1 ± 5.4%) glomeruli contained cleaved lamin A-positive podocytes in the obstructed kidney (P = 0.0070; Fig. 3, A and B). The average number of podocytes per glomerulus was significantly greater in the obstructed than nonobstructed kidney (6.70 ± 0.30 vs. 6.07 ± 0.32, P = 0.0064; Fig. 3C). In addition, although rare, TUNEL-positive podocytes were observed in the obstructed kidney but not in the nonobstructed kidney (Fig. 3, D and E, and Supplemental Figs. S3D and S4A).

Figure 3.

Glomerular filtration facilitates detachment of podocytes with activated apoptotic caspases. NEP25 mice were injected with 0.625 ng/g body wt LMB2, underwent unilateral ureteral obstruction on day 6, and were euthanized on the next day. A: representative pictures of synaptopodin and cleaved lamin A (cLamin A) double staining of glomeruli in the obstructed (Obst.) and nonobstructed (Non-obst.) kidneys. Scale bar = 50 μm. B: number of cLamin A⁺ podocytes in the obstructed and nonobstructed kidneys. C: number of synaptopodin⁺ podocytes in the obstructed and nonobstructed kidneys. In B and C, each pair of squares connected with a line represents the average value for obstructed and nonobstructed kidneys from the same mouse. Data between the obstructed and nonobstructed kidneys of both male (n = 3) and female (n = 5) mice were compared using a paired t test. D: rare TUNEL⁺ podocytes in the obstructed kidney. Scale bar = 50 μm. E: rate of glomeruli containing TUNEL⁺podocalyxin⁺ cells, as evaluated with fluorescent microscopy. TUNEL⁺ podocytes were observed only in the obstructed kidney, but not in the nonobstructed kidney, of NEP25 mice (n = 4, female). F: scanning electron microscopy images showing detaching podocytes in the nonobstructed kidney. Scale bars = 10 μm in the left and middle images 2 μm in high-magnification image (right). A scanning electron microscopy image for NEP25 mice without LMB2 is provided in Supplemental Fig. S3E. G: rate of glomeruli containing detaching podocytes. Detaching podocytes were detected only in the obstructed kidney, but not in the nonobstructed kidney, of NEP25 mice (n = 3). Values are presented as means ± 95% confidence intervals. Data were analyzed using a paired t test. Podxl1, podocalyxin.

By scanning electron microscopy, we observed detaching podocyte cell bodies in 22.2 ± 11.5% of the glomeruli of the nonobstructed kidney but not of the obstructed kidney (Fig. 3, F and G, and Supplemental Fig. S3E).

These results indicated that the physical force of glomerular filtration facilitates the detachment of podocytes with activated apoptotic caspases.

LMB2 Induces Typical Apoptosis in Podocytes of Kidney Organoids

To further investigate the dying process of podocytes without glomerular filtration, we isolated nephron progenitor cells from embryonic day 12.5 embryonic kidneys of NEP25 mice and generated kidney organoids. The organoids contained differentiated podocytes expressing nephrin, WT1, and podocin as well as proximal tubular-like cells marked by Lotus tetragonolobus lectin and megalin staining (Fig. 4A and Supplemental Fig. S5A). Organoids selectively expressed hCD25 in podocytes, similar to the kidneys of NEP25 mice. LMB2 treatment selectively injured podocytes of organoids, as evidenced by diminished nephrin staining (Supplemental Fig. S5A) and disappearance of most podocyte clusters. In a few residual clusters of WT1-positive cells, cleaved lamin A and TUNEL staining was frequently observed (Fig. 4A). In addition, nuclei of TUNEL-positive cells were often fragmented, which was never observed in in vivo podocytes with glomerular filtration (Fig. 4A, red box and arrows).

Figure 4.

Activation of caspase-3 in podocytes in kidney organoids, other models, and human biopsy samples. A: cleaved caspase-3 (cCasp-3) and TUNEL staining in podocytes of kidney organoids. The organoids were generated from nephron progenitor cells of NEP25 mice and contained Wilms’ tumor-1 (WT1)⁺ podocytes and tubular cells showing affinity to Lotus tetragonolobus lectin (LTL). After LMB2 treatment, most podocyte clusters disappeared. In a few residual clusters of WT1⁺ cells, cCasp-3 staining and fragmented nuclei stained with TUNEL were frequently observed (black arrows). B: cCasp-3 in enhanced green fluorescent protein (EGFP)⁺ (human CD25⁻) podocytes were indirectly injured in mosaic mice injected with LMB2. C: cCasp-3 in desmin⁺ injured podocytes of transgenic mice expressing mutant α-actin-4 (arrows). D: cCasp-3 in desmin⁺ injured podocytes of human immunodeficiency virus (HIV)-1 transgenic mice (arrows). E: cCasp-3 in podocytes in human biopsy samples with antineutrophil cytoplasmic antibody-associated nephropathy (arrows). F: cCasp-3 in podocytes in human biopsy samples from patients with diabetic kidney disease (arrows). Scale bars = 50 μm. Controls for α-actin-4, HIV-1 transgenic mice, and normal human samples are provided in Supplemental Fig. S5B, C, and D, respectively. cLamin A, cleaved lamin A.

Activation of Caspase-3 in Other Podocyte Injury Models and Human Kidney Diseases

Because translational inhibition is caused not only by LMB2 but also by various cellular stresses, we investigated whether caspase-3 is activated in podocytes of other injury models or human kidney diseases. We first examined a mosaic mouse model, in which a fraction of podocytes expressed hCD25. After injection of LMB2, not only hCD25-positive podocytes but also hCD25-negative podocytes were injured, and mice developed global sclerosis. In severely injured kidneys of mosaic mice, cleaved caspase-3 signals were occasionally observed in hCD25-negative podocytes, which were labeled by EGFP (Fig. 4B). Thus, caspase-3-dependent cell death occurs in some indirectly injured podocytes. Cleaved caspase-3 signals were also observed in desmin-positive injured podocytes of transgenic mice expressing mutant α-actinin-4 or HIV-1 genes (Fig. 4, C and D, and Supplemental Fig. S5, B and C).

Unlike desmin in mice, there is no marker for injured podocytes in the human kidney. Therefore, the identification of injured podocytes is difficult in human kidneys. Nevertheless, cleaved caspase-3 signals were occasionally observed in podocytes with diminished nephrin staining in biopsy samples of individuals with antineutrophil cytoplasmic antibody-associated nephropathy or diabetic kidney diseases (Fig. 4, E and F, and Supplemental Fig. S5D).

DISCUSSION

The present study clearly showed that LMB2 very efficiently induces caspase-3-dependent typical intrinsic apoptosis in hCD25-expressing cultured podocytes. LMB2 also induced TUNEL-positive staining in podocytes of kidney organoids. In contrast, LMB2 never induced TUNEL-positive podocytes in NEP25 mice under any experimental conditions. Further analysis revealed that cleaved caspase-3 and cleaved lamin A were detected in injured podocytes of NEP25 mice. These events occur early in apoptosis and are widely used as indicators of the irreversible apoptotic process (24, 41, 42). The present study indicated that podocytes with activated apoptotic caspases are quickly detached and lost in the urine before DNA fragmentation. We speculate that this is the reason why definitive apoptosis is rarely observed in podocytes in vivo.

The TUNEL method, which detects DNA fragmentation, used to be regarded as specific for apoptosis but is now known to be observed in other types of cell death (22, 43). In addition, inadequate fixation or overexposure of the TUNEL reaction causes artifacts in TUNEL staining. Many previous studies have described apoptosis induced by various factors in cultured podocytes. Podocyte apoptosis has also been reported in many in vivo studies (13, 20, 44). In some studies, cells with TUNEL-positive staining were not definitively identified as podocytes. In most studies, TUNEL-positive cells did not show morphological characteristics of apoptosis, including chromatin condensation and fragmentation of nuclei. It cannot be ruled out that podocytes with DNA fragmentation but without apoptosis-specific morphological changes are detected. It is also possible that apoptotic podocytes are retained in glomeruli with crescent formation or under special conditions in which cell-matrix adhesion is intensified. However, our study suggested that TUNEL-positive apoptotic podocytes are less frequent than previously reported.

In the present and our previous studies, podocytes were injured in an LMB2 dose-dependent manner (25). However, the frequency of cleaved caspase-3 or cleaved lamin A did not increase even when the dose of LMB2 was increased to more than 1.25 ng/g body wt and UUO was performed 1 day before euthanasia (data not shown). Electron microscopy analysis found plasma membrane rupture in some podocytes, indicating that this may be another mechanism of podocyte loss before completion of the full apoptotic process. Exposure to exceptionally high filtration pressure (up to 40 mmHg) may cause plasma membrane rupture in podocytes with activated apoptotic caspases.

Generally, apoptotic cells maintain plasma membrane integrity and are immediately eliminated by phagocytosis (22). These features prevent the scattering of inflammatory cellular substances (damage-associated molecular patterns and alarmins). However, phagocytes have limited access to podocytes under normal conditions. On signaling of the irreversible apoptotic process by the activation of caspase-3, podocytes are quickly detached from the glomerular basement membrane and lost in the urine. Moreover, even if the plasma membrane is ruptured, toxic cellular materials will be quickly washed away by glomerular filtration. Under such circumstances, conducting the entire apoptotic process is not advantageous and unnecessary for podocytes.

Our analyses revealed more cleaved lamin A-positive podocytes in obstructed kidneys and more detaching podocytes in nonobstructed kidneys (Fig. 3, A, B, F, and G). Intuitively, one would expect that the more injured podocytes, the more detaching cells. This would be true if the filtration pressure is constant. In our experiment, podocytes on both sides of the kidney were exposed to the same LMB2 concentration and to a similar systemic milieu by the time of UUO surgery (day 6). Therefore, podocytes on both sides of the kidney were similarly injured on day 6. However, on the day after UUO surgery, the obstructed kidney showed significantly more cleaved lamin A-positive podocytes than the nonobstructed kidney. In addition, synaptopodin-positive nuclei were observed more frequently in obstructed kidneys. These results indicate that the nonobstructed kidney lost injured podocytes in the urine because their detachment was facilitated by glomerular filtration pressure. In contrast, the obstructed kidney retained injured podocytes because their detachment was prevented by increased pressure in Bowman’s space.

Many molecules related to cell adhesion, including β-catenin and focal adhesion kinase, are cleaved by activated caspases (45). This phenomenon may cause detachment of apoptotic cells from the surrounding cells and matrix. Our study indicated that the physical force of glomerular filtration facilitates the detachment of podocytes with activated apoptotic caspases. Another possibility is that damaged podocytes are first detached, and detachment then induces apoptosis, which is known as anoikis. Anoikis is triggered by unligated integrins and prevents adherence-independent cell growth and reattachment to an inappropriate matrix (46, 47). In our study, however, inhibition of detachment by ureteral obstruction did not decrease, but rather increased, podocytes with activated apoptotic caspases. This finding clearly indicates that caspase activation occurred before detachment.

LMB2 inhibits protein synthesis by inactivating elongation factor-2. The mechanism of caspase activation by LMB2 in hCD25-expressing podocytes is likely depletion of Mcl-1, an antiapoptotic protein (Supplemental Fig. S2). Depletion of the short-lived antiapoptotic protein Mcl-1 likely facilitates the activation of Bak1 and Bax via direct activators, Bid and Puma, or via themselves, thereby initiating the apoptotic pathway. Indeed, Mcl-1 staining in podocytes of NEP25 mice was attenuated after injection with LMB2 (Supplemental Fig. S2B). Because protein synthesis is blocked by various factors, such as endoplasmic reticulum stress, we examined cleaved caspase-3 in other podocyte injury models and human biopsy samples. We previously generated transgenic mice in which ∼50% of podocytes express hCD25 (32). After injection of LMB2, hCD25-negative podocytes were also secondarily injured. The present study showed that caspase-3 was activated in hCD25-negative, i.e., indirectly injured, podocytes. In addition, cleaved caspase-3 was detected in podocytes of mutant α-actinin-4 or HIV-1 transgenic mice and human biopsy samples of antineutrophil cytoplasmic antibody-associated nephropathy and diabetic kidney disease. Although the frequency of cleaved caspase-3 staining in podocytes was not high, caspase-3-dependent cell death may be involved in these conditions. Of note, the frequency of cleaved caspase-3-positive podocytes of NEP25 mice was not high, only 14.4% at the highest, even though most podocytes were lost rapidly within short periods. The present study indicated that podocytes with activated caspase-3 are rapidly lost in the urine. We found that a further increase in LMB2 dosage did not increase the frequency of cleaved caspase-3-positive podocytes. We speculate that podocytes are lost before cleaved caspase-3 becomes detectable. The present study indicated that urinary cells may be more informative than kidney tissues.

Perspectives and Significance

The present study demonstrated that podocytes dying dependently on apoptotic caspases are quickly lost before completion of the entire apoptotic process. This explains the rarity of definitive apoptosis in in vivo podocytes. Moreover, urinary sediments may be more informative than kidney tissue for the process of podocyte death.

SUPPLEMENTAL DATA

Supplemental material and Supplemental Figs. S1–S5: https://doi.org/10.6084/m9.figshare.19395467.

GRANTS

This work was supported by JSPS KAKENHI Grants JP18H02827, JP21H02936, JP18K15986, and JP21K16172.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.Y., M.O., K.T., and T.M. conceived and designed research; K.Y., M.O., I.P., T.A., T.U., M.K., and T.M. performed experiments; K.Y., M.O., K.T., T.Y., I.P., T.A., K.O., T.U., M.K., and T.M. analyzed data; K.Y., M.O., K.T., T.Y., I.P., T.A., K.O., T.U., M.K., and T.M. interpreted results of experiments; K.Y., M.O., and T.M. prepared figures; K.Y. drafted manuscript; K.Y., M.O., T.A., K.O., T.U., M.K., and T.M. edited and revised manuscript; T.M. and M.O. approved final version of manuscript.