Abstract
Diabetic kidney disease (DKD) is the main cause of chronic kidney
disease worldwide. While injury to the podocytes, visceral epithelial
cells that comprise the glomerular filtration barrier, drives
albuminuria, proximal tubule (PT) dysfunction is the critical mediator
of DKD progression. Here, we report that the podocyte-specific
induction of human KLF6, a zinc-finger binding transcription factor,
attenuates podocyte loss, PT dysfunction, and eventual interstitial
fibrosis in a male murine model of DKD. Utilizing combination of
snRNA-seq, snATAC-seq, and tandem mass spectrometry, we demonstrate
that podocyte-specific KLF6 triggers the release of secretory ApoJ to
activate calcium/calmodulin dependent protein kinase 1D (CaMK1D)
signaling in neighboring PT cells. CaMK1D is enriched in the first
segment of the PT, proximal to the podocytes, and is critical to
attenuating mitochondrial fission and restoring mitochondrial function
under diabetic conditions. Targeting podocyte-PT signaling by enhancing
ApoJ-CaMK1D might be a key therapeutic strategy in attenuating the
progression of DKD.
Subject terms: Diabetic nephropathy, Chronic kidney disease, Podocytes
__________________________________________________________________
While diabetic kidney disease is characterized by initial glomerular
injury, proximal tubule dysfunction determines the progression to
kidney fibrosis. Here, the authors show that podocyte KLF6 attenuates
proximal tubule injury via ApoJ-CaMK1D signaling in a murine model of
diabetic kidney disease.
Introduction
Diabetic kidney disease (DKD) is the leading cause of chronic kidney
disease (CKD) and end-stage kidney disease (ESKD)
worldwide^[40]1,[41]2. Endothelial injury, glomerular hypertrophy,
podocyte foot process effacement, eventual podocyte loss, and the
accompanying albuminuria are key features of early DKD^[42]3, with
tubular injury and interstitial fibrosis occurring in later stages of
DKD^[43]4. While a significant proportion of individuals with diabetes
develop CKD, a majority of these individuals do not progress to
ESKD^[44]4,[45]5. Recent studies demonstrate that while glomerular
injury is the initial driver of early injury to the kidney, factors
mediating proximal tubule (PT) dysfunction is the key determinant of
DKD progression^[46]6,[47]7.
Krüppel-like factors (KLFs) are a family of zinc-finger transcription
factors that play a critical role in fundamental cellular processes in
multiple tissues, including the kidney, to maintain homeostasis as well
as in development and in disease^[48]8. Among the 17 members of the KLF
family, Krüppel-like factor 6 (KLF6) is an early-inducible responsive
gene expressed with cell-specific diverse roles. Specifically in the
podocyte, KLF6 is a key regulator of mitochondrial function and the
conditional loss of Klf6 in podocytes reduces mitochondrial complex IV
assembly, which exacerbates glomerular injury in murine models of Focal
Segmental Glomerulosclerosis (FSGS) and DKD^[49]9,[50]10. Based on
these previous studies, we sought to investigate whether
podocyte-specific induction of human KLF6 will attenuate podocyte
injury in a murine model of DKD. Here, we demonstrate that this
induction of KLF6 specifically in podocytes attenuates podocyte injury
and glomerulosclerosis, and improves PT injury and interstitial
fibrosis under diabetic conditions in mice. While we initially
suspected this improvement in PT injury was a result of reduction in
podocyte injury, unbiased single nuclei RNA-sequencing (snRNA-seq) and
Assay for Transposase-Accessible Chromatin with sequencing (ATAC-seq)
demonstrate that the induction of podocyte KLF6 under basal conditions,
primes transcriptional changes in the first segment of the PT, proximal
to the podocytes, through the induction of calcium/calmodulin dependent
protein kinase ID (CaMK1D) signaling. We report that CaMK1D is highly
enriched in the first segment of the PT and preconditions the proximal
tubule against mitochondrial fission under diabetic conditions. In
addition, unbiased tandem mass spectrometry showed that the induction
of KLF6 in podocytes triggers the release of secretory Apolipoprotein J
(ApoJ), which subsequently undergoes cellular uptake by low-density
lipoprotein-related protein 2 (Lrp2)/megalin to activate CaMK1D
signaling in the PT to attenuate mitochondrial fission. To date, this
is the first study to demonstrate a potential mechanism by which the
podocyte preconditions the PT against injury through ApoJ-CaMK1D
signaling under diabetic conditions.
Results
Generation and characterization of KLF6^PODTA mice under basal conditions
Since the podocyte-specific loss of Klf6 increases susceptibility to
DKD^[51]10, we sought to ascertain whether the podocyte-specific
induction of KLF6 conversely attenuates the progression of DKD. Mice
with podocyte-specific expression of human KLF6 (KLF6) were initially
generated using the “tet-on” system, where the binding of chimeric
tetracycline transactivator protein (rtTA) to tet-operator and gene
activation only occurs in the presence of doxycycline (DOX).
Specifically, TRE-KLF6 mice were bred with the Podocin-rtTA (PODTA)
mice to generate mice with podocyte-specific expression of KLF6
(KLF6^PODTA) in the setting of DOX administration. To assess the
specificity of KLF6 expression, relative KLF6 mRNA abundance was
initially measured in isolated kidney cortex, glomeruli, podocyte, and
non-podocyte glomeruli fractions from KLF6^PODTA and NPHS2-rtTA mice
(Supplementary Fig. [52]1a). While KLF6 expression remained completely
undetected in NPHS2-rtTA mice across glomeruli, cortex, podocyte and
non-podocyte glomeruli fractions, expression in the KLF6^PODTA mice was
primarily enriched in the podocyte and glomeruli fractions, with a
lower level of expression detectable in the kidney cortex and
non-podocyte glomerular fractions (Supplementary Fig. [53]1a). Although
mouse KLF6 is expressed in both glomerular and non-glomerular cells,
immunostaining for KLF6 confirmed the higher expression of
podocyte-specific KLF6 in KLF6^PODTA mice as compared to NPHS2-rtTA
mice (Supplementary Fig. [54]1b, c). In addition, mKlf6 expression
remained similar in both KLF6^PODTA and NPHS2-rtTA mice (Supplementary
Fig. [55]1d). Furthermore, KLF6^PODTA mice were viable and fertile with
no significant difference in albuminuria or kidney injury as compared
to NPHS2-rtTA mice (Supplementary Fig. [56]1e).
Podocyte-specific induction of KLF6 attenuates kidney injury under diabetic
conditions
To test whether the podocyte-specific induction of KLF6 attenuates DKD,
NPHS2-rtTA and KLF6^PODTA mice underwent uninephrectomy (UNx) with
subsequent low-dose streptozotocin (STZ) treatment (Fig. [57]1a).
SHAM + vehicle-treated (SHAM-VEH) NPHS2-rtTA and KLF6^PODTA mice were
used as non-diabetic controls. The combination of UNx + STZ has been
shown to accelerate the development of glomerular lesions
representative of DKD^[58]11, thereby making it a suitable model to
test the potential renoprotective effects of podocyte KLF6 induction.
All diabetic mice exhibited high blood glucose levels (>600 mg/dl)
post-treatment, which was maintained throughout the experimental period
(Supplementary Table [59]1). Blood glucose levels of non-diabetic
groups were within the normal range throughout the experimental period
(Supplementary Table [60]1). Kidney weights and kidney to body weight
ratios were significantly increased for all diabetic groups compared to
the non-diabetic groups, with no significant differences between
KLF6^PODTA and NPHS2-rtTA mice under either condition (Supplementary
Table [61]1). While the diabetic NPHS2-rtTA mice exhibited an increase
in albuminuria, serum urea nitrogen, and serum creatinine as compared
to the non-diabetic NPHS2-rtTA mice and the diabetic KLF6^PODTA mice,
no significant differences were observed between the diabetic and
non-diabetic KLF6^PODTA mice (Fig. [62]1b–d). In addition, we observed
an improvement in survival in the diabetic KLF6^PODTA mice as compared
to the diabetic NPHS2-rtTA mice (Fig. [63]1e). Subsequent staining with
Periodic-acid Schiff (PAS) showed a significant increase in glomerular
volume, mesangial expansion, % sclerotic glomeruli in all diabetic mice
as compared to their respective controls (Fig. [64]2a–c, Supplementary
Table [65]2). However, the diabetic KLF6^PODTA mice exhibited less
glomerular hypertrophy, mesangial expansion, % sclerotic glomeruli with
proteinaceous casts as compared to the diabetic NPHS2-rtTA mice
(Fig. [66]2a–c). Immunostaining for Wilms tumor 1 (WT1), a
podocyte-specific marker, showed that the diabetic NPHS2-rtTA mice had
fewer podocytes per glomerular area as compared to non-diabetic mice.
In comparison, the podocyte number was preserved in the diabetic
KLF6^PODTA mice as compared to the diabetic NPHS2-rtTA mice
(Fig. [67]2d). Furthermore, synaptopodin, critical for podocyte actin
cytoskeleton, expression was reduced in the diabetic NPHS2-rtTA mice as
compared to the diabetic KLF6^PODTA mice and non-diabetic mice
(Fig. [68]2e). We also investigated the ultrastructural changes using
transmission electron microscopy (TEM) and found that the diabetic
NPHS2-rtTA mice had an increase in foot process effacement and
glomerular basement membrane (GBM) thickness as compared to the
non-diabetic NPHS2-rtTA mice, which was attenuated in the diabetic
KLF6^PODTA mice (Fig. [69]2a, f, g).
Fig. 1. KLF6^PODTA mice demonstrate a reduction in albuminuria and improved
kidney function and overall survival compared to NPHS2-rtTA mice under
diabetic conditions.
[70]Fig. 1
[71]Open in a new tab
a Experimental timeline for DOX treatment, UNx or SHAM procedures, STZ
or VEH treatment and experimental endpoint. b Twenty-four hour urine
albumin concentration (mg/24hrs); p = 0.0025 for NPHS2-rtTA (SHAM + VEH
vs UNx + STZ) group, p = 0.0006 for UNx + STZ group. c Serum urea
nitrogen (mg/dl); p = 0.0025 for NPHS2-rtTA (SHAM + VEH vs UNx+STZ),
p = 0.0006 for UNx + STZ group. d Serum creatinine concentration
(mg/dl) at 20 weeks of age; p = 0.016 for NPHS2-rtTA (SHAM + VEH vs
UNx + STZ), p = 0.054 for UNx+STZ group. For SHAM + VEH groups, n = 5
(NPHS2-rtTA), n = 3 (KLF6^PODTA) mice; for UNx + STZ groups, n = 7
(NPHS2-rtTA), n = 6 (KLF6^PODTA) mice; *p < 0.05, **p < 0.01,
***p < 0.001; Mann–Whitney test, two-sided; data presented as
mean ± SEM. e Survival curve for all 4 groups until 20 weeks of age
(for SHAM + VEH group, n = 5 (NPHS2-rtTA); n = 3 (KLF6^PODTA) mice,
n = 14 for UNx+STZ groups; p = 0.0094 vs all other groups; Mantel-Cox
test). Source data are provided as a Source Data file.
Fig. 2. KLF6^PODTA mice demonstrate a reduction in glomerular and tubular
injury compared to NPHS2-rtTA mice under diabetic conditions.
[72]Fig. 2
[73]Open in a new tab
a Representative images of PAS staining at full scan (10x), low (20x)
and high (40x) magnifications, synaptopodin staining, and transmission
electron microscopy (TEM) of glomerular ultrastructure. Black
arrowheads indicate sclerotic glomeruli, $ = protein casts, and
# = tubular dilation. Top panel: red scale bar = 1000 µm. TEM: red
arrows indicate effaced foot processes; red arrowheads point to healthy
foot processes; red calipers show thickened glomerular basement
membrane (GBM). Quantification of (b) glomerular volume; p = 0.0001 for
SHAM + VEH (NPHS2-rtTA vs KLF6^PODTA), p = < 0.0001 for NPHS2-rtTA
(SHAM + VEH vs UNx+STZ) and KLF6^PODTA (SHAM + VEH vs UNx+STZ),
p = 0.0228 for UNx+STZ (NPHS2-rtTA vs KLF6^PODTA), (c) mesangial
expansion; p = 0.0001 for SHAM + VEH (NPHS2-rtTA vs KLF6^PODTA),
p = <0.0001 for NPHS2-rtTA (SHAM + VEH vs UNx + STZ) and KLF6^PODTA
(SHAM + VEH vs UNx+STZ KLF6^PODTA), p = 0.0228 for UNx+STZ (NPHS2-rtTA
vs KLF6^PODTA), and (d) podocyte number (WT1^+ve Hoechst^+ve cells) per
glomerular cross-section; p = 0.0005 for NPHS2-rtTA (SHAM + VEH vs
NPHS2-rtTA), p = 0.0428 for UNx+STZ (NPHS2-rtTA vs KLF6^PODTA). (n = 20
glomeruli/mouse, n = 3–5 mice/group, *p < 0.05, **p < 0.01 ***p < 0.001
****p < 0.0001 Kruskal–Wallis test with Dunn’s post-test). e
Quantification of glomerular synaptopodin staining (n = 20
glomeruli/mouse, n = 3–5 mice/group; p = 0.0374 for NPHS2-rtTA
(SHAM + VEH vs UNx + STZ), p = < 0.0001 for UNx+STZ (NPHS2-rtTA vs
KLF6^PODTA), *p < 0.05, ****p < 0.0001; Mann–Whitney test, two-sided).
f Quantification of podocyte foot process effacement measured as foot
process width (nm); p = 0.0001 for NPHS2-rtTA (SHAM + VEH vs
UNx + STZ), p = 0.0262 for UNx + STZ (NPHS2-rtTA vs KLF6^PODTA) and (g)
GBM thickness (nm) (n = 3 mice/group; p = <0.0001 for NPHS2-rtTA
(SHAM + VEH vs UNx + STZ), p = <0.0001 for (UNx + STZ NPHS2-rtTA vs
KLF6^PODTA), *p < 0.05, ***p < 0.001, ****p < 0.0001; Kruskal–Wallis
test with Dunn’s post-test). Source data are provided as a Source Data
file.
Hematoxylin & eosin (H&E) with histopathological scoring by M.P.R.,
renal pathologist, in a blinded fashion showed an increase in
interstitial inflammation in both diabetic groups as compared to their
respective non-diabetic controls, with an increasing trend
(non-statistically significant) in the diabetic NPHS2-rtTA mice as
compared to the diabetic KLF6^PODTA mice (Supplementary Fig. [74]2a,
Supplementary Table [75]2). While both diabetic groups also had some
loss of lotus lectin staining, suggesting PT brush border loss, the
diabetic NPHS2-rtTA mice had a significant reduction as compared to the
diabetic KLF6^PODTA mice (Supplementary Fig. [76]2a, [77]b). In
addition, picrosirius red, and α-SMA staining showed an increase in
staining in both diabetic groups, which was improved in the diabetic
KLF6^PODTA mice (Supplementary Fig. [78]2a, [79]c). Interestingly,
histopathological scoring noted significant tubular injury and
interstitial fibrosis only in the diabetic NPHS2-rtTA mice
(Supplementary Fig. [80]2a, Supplementary Table [81]2). Collectively,
these data suggest that the podocyte-specific induction of KLF6
attenuated glomerular and tubulointerstitial injury under diabetic
conditions.
Podocyte-specific induction of KLF6 restores podocyte differentiation markers
under diabetic conditions
To assess the potential mechanism that mediates the renoprotective
effects of KLF6^PODTA under diabetic conditions at the single cell
level, we initially performed snRNA-seq on the kidney cortex. Rationale
for choosing snRNA-seq versus scRNA-seq was based on recent
studies^[82]12. After clearing all quality control checks, we initially
generated single nuclear transcriptomes for 78,979 nuclei
(Supplementary Fig. [83]3a, [84]b). Unsupervised clustering analysis
subsequently identified 23 cell clusters (Supplementary Fig. [85]3c)
and the top significant marker genes were compared to previously
reported cell type markers^[86]13 to assign cell type specific identity
to these clusters. Clusters with similar cell type identity were
combined to generate 18 unique clusters (Fig. [87]3a, b). The nuclei
count for each cluster, with respect to each group was reported
(Supplementary Data [88]1). To identify differentially expressed genes
(DEGs) in the setting of podocyte-specific KLF6 induction with and
without diabetes, we initially performed differential expression
analysis on all the clusters (Supplementary Data [89]2). Interestingly,
in the podocyte cluster, there were very few DEGs (0 upregulated and 2
downregulated) in the KLF6^PODTA group vs NPHS2-rtTA under non-diabetic
conditions (Supplementary Data [90]2). However, under diabetic
conditions, we observed a significant number of DEGs (59 upregulated
and 42 downregulated) in the KLF6^PODTA as compared to the NPHS2-rtTA
mice (Fig. [91]3c, d). Subsequent pathway enrichment analysis on these
DEGs was conducted on the podocyte cluster from the diabetic KLF6^PODTA
mice using Enrichr^[92]14,[93]15. Enrichment analysis using
Reactome^[94]16, WikiPathways^[95]17, and KEGG pathways^[96]18 for the
upregulated DEGs demonstrated key pathways involving N-linked
glycosylation, axon guidance, nephrin and semaphorin interactions, as
well as vascular endothelial growth factor (VEGF) signaling pathways
(Fig. [97]3c, e). In addition, the reduced expression of podocyte
structural and differentiation markers such as Nephrin (Nphs1),
Synaptopodin (Synpo), Podocalyxin (Podxl) and Dachshund family
transcription factor 1 (Dach1) in the diabetic NPHS2-rtTA mice were all
significantly restored in the diabetic KLF6^PODTA mice (Fig. [98]3c,
e). Conversely, the downregulated DEGs were enriched for pathways
involving focal adhesion, integrin-mediated cell adhesion, complement
activation and inflammation signaling pathways (Fig. [99]3d, f). To
determine the cell specificity of the DEGs in the podocyte cluster, we
plotted the expression of these genes on a dot plot showing their
expression across all the cell types (Supplementary Fig. [100]4a,
[101]b). We report that while the downregulated DEGs were not
necessarily specific to one cell type, the upregulated genes were
predominantly specific to the podocyte cluster (Supplementary
Fig. [102]4a, [103]b). This suggests that KLF6 might play a key role in
maintenance of the mature podocyte differentiation markers in the
setting of cell stress under diabetic conditions.
Fig. 3. snRNA-seq analysis showing the clustering signature and composition.
[104]Fig. 3
[105]Open in a new tab
a Uniform Manifold Approximation and Projection (UMAP) plot shows the
78,979 nuclei, mapping to 18 clusters. b Dot plot showing cluster
identities aligned to canonical cell types in the adult mouse kidney
based on a variety of cell type-specific marker genes. PT proximal
tubules, LH(DL) Loop of Henle (Descending loop), LH(AL) Loop of Henle
(Ascending loop), DCT distal convoluted tubule, CNT connecting tubule,
CD-PC collecting duct principal cell, IC intercalated cell, Mø
macrophage, PEC parietal epithelial cell, Prolif.PT Proliferating
proximal tubules, Mes mesangial, Endo endothelial, Pod podocyte. Dot
plot of all significantly (c) upregulated and (d) downregulated genes
in the podocyte cluster between KLF6^PODTA vs NPHS2-rtTA mice under
diabetic conditions. Expression level of these genes is shown by the
heat map and the percentage of cells expressing the gene in the cluster
is shown by the size. The default statistical test (Wilcoxon Rank Sum
test, adjusted p-value based on Bonferroni correction) in Seurat
package was used. Significantly expressed genes are determined by
adjusted p < 0.05. Reactome, WikiPathways and KEGG enrichment analysis
of these (e) upregulated and (f) downregulated genes in the podocyte
cluster between KLF6^PODTA vs NPHS2-rtTA mice under diabetic
conditions. The default statistical test (fisher’s exact test) in
Enrichr was used.
To explore the potential direct and indirect mechanism by which KLF6
might regulate gene expression, we initially conducted in-silico
analysis by matching the DEGs in the podocyte cluster with previously
reported KLF6 ChIP-seq expression arrays^[106]19 obtained from the
Encyclopedia of DNA Elements (ENCODE) project (Supplementary
Fig. [107]5a). Class 0 DEGs were defined as having at least 1 KLF6
binding site within ±1 kb of the transcription start site (TSS), class
1 DEGs have at least 1 KLF6 binding site between ±1 and 10 kb of the
TSS and class 2 DEGs as having no KLF6 binding site within ±10 kb of
the TSS. Of the 59 upregulated DEGs, there were 11 class 0, 5 class 1,
and 43 class 2 DEGs (Supplementary Fig. [108]5b), while the
downregulated DEGs were 9 class 0, 2 class 1, and 31 class 2 DEGs
(Supplementary Fig. [109]5c). To examine the contribution of the class
0 DEGs to enriched pathways, we compared the statistical significance
of pathway enrichment between class 0 and class 2 DEGs (Supplementary
Fig. [110]5d, [111]e). Despite having fewer numbers of DEGs,
upregulated class 0 DEGs were more highly enriched for N-linked
glycosylation pathways, axon guidance, and nephrin interaction
pathways, suggesting a potentially more direct transcriptional
regulation by KLF6 (Supplementary Fig. [112]5d). In comparison, class 0
downregulated DEGs were enriched for pathways involving integrin and
non-integrin dependent ECM interactions as compared to class 2
downregulated DEGs (Supplementary Fig. [113]5e).
To further characterize the potential mechanism by which
podocyte-specific KLF6 might mediate its salutary effects, we performed
snATAC-seq on kidney cortex samples from NPHS2-rtTA and KLF6^PODTA mice
to ascertain the degree of chromatin accessibility in the podocyte
cluster. We successfully generated chromatin accessibility libraries
for 9,894 nuclei that passed all quality control checks (Supplementary
Fig. [114]6a–c). Dimensionality reduction was performed using
R-packages Signac 1.8.0^[115]20 and Seurat 4.3.0^[116]21, and clusters
were identified and assigned using anchors transferred from the
snRNA-seq data. Label transfer was very successful in mapping the
snATAC-seq data as indicated by the high prediction scores
(Supplementary Fig. [117]6d). Nuclei from all 18 clusters in our
snRNA-seq data were identified in the snATAC-seq data (Fig. [118]4a).
To compare conformity between the snRNA and snATAC data, we plotted
gene activity metrics for the clusters using the top significant marker
genes from the snRNA-seq data (Fig. [119]4b). Gene activity score
predicts the level of gene expression based on the accessibility of
regulatory elements in the vicinity of the gene^[120]22. Majority of
the clusters showed exemplary correlations between gene expression of
the selected marker gene and its corresponding gene activity score
(Fig. [121]4b). We also observed that all the clusters exhibited unique
cell type-specific chromatin accessibility (Fig. [122]4c). To explore
the potential chromatin accessibility changes induced by KLF6 induction
in the podocyte cluster, we performed differential accessibility
analysis in the podocyte cluster between both groups (Fig. [123]4d,
Supplementary Data [124]3). We also identified several chromatin
organization genes such as Nuclear Receptor Coactivator 1 (Ncoa1),
Nuclear Receptor Coactivator 2 (Ncoa2), Bromodomain Containing 1
(Brd1), Lysine Demethylase 5 C (Kdm5c), Lysine Demethylase 2 A (Kdm2a),
PHD Finger Protein 20(Phf20), MSL Complex Subunit 2 (Msl2), GATA Zinc
Finger Domain Containing 2B (Gatad2b) and, Polybromo 1 (Pbrm1) are
differentially expressed in KLF6^PODTA podocytes (Supplementary
Fig. [125]5f), suggesting that KLF6 might regulate chromatin
reorganization. In addition, the KLF6^PODTA podocytes possessed
chromatin regions that were more accessible compared to the NPHS2-rtTA
podocytes (Fig. [126]4d). Motif enrichment analysis of the
differentially more accessible chromatin regions in the KLF6^PODTA
podocytes showed a high enrichment for several classes of transcription
factors involved in podocyte differentiation such as WT1, KLF15,
Transcription Factor 21 (TCF21), including KLF6, among others
(Fig. [127]4e, Supplementary Data [128]4). Subsequent gene ontology
(GO) enrichment analysis for the differential accessible chromatin
regions showed an enrichment for biological processes involving the
actin cytoskeleton such as cell-cell communication, Ca^2+ signaling,
VEGF signaling, receptor for tyrosine kinase signaling, and syndecan
interactions, with higher statistical significance in the KLF6^PODTA
accessible chromatin regions (Fig. [129]4f). Collectively, these data
suggest that podocyte-specific induction of KLF6 preserves podocyte
health by increasing the accessibility to podocyte pro-differentiation
transcription factors.
Fig. 4. snATAC-seq analysis showing the gene activity, chromatin
accessibility and motif enrichment analysis at baseline.
[130]Fig. 4
[131]Open in a new tab
a UMAP showing predicted annotation of the snATAC-seq data following
integration and label transfer from the snRNA-seq data. b Dot plot
showing gene activity aligned to canonical cell types in the adult
mouse kidney based on a variety of cell type-specific marker genes. PT
proximal tubules, LH(DL) Loop of Henle (Descending loop), LH(AL) Loop
of Henle (Ascending loop), DCT distal convoluted tubule, CNT connecting
tubule, CD-PC collecting duct principal cell, IC intercalated cell, Mø
macrophage, PEC parietal epithelial cell, Prolif.PT Proliferating
proximal tubules, Mes mesangial, Endo endothelial, Pod podocyte. c
Heatmap showing of average number of Tn5 cut sites within a
differentially accessible region (DAR) for all cell-types. d Heatmap
showing differential chromatin accessibility in the podocyte cluster of
KLF6^PODTA vs NPHS2-rtTA. e TF motif enrichment analysis was carried
out on genes with differential accessible chromatin in the podocyte
cluster (KLF6^PODTA vs NPHS2-rtTA); genes with C2H2 zinc finger (C2H2
ZF) motifs are shown in red, basic helix–loop–helix (bHLH) in blue,
Homeodomain in green and others are shown in gray. DNA motifs that are
overrepresented in a set of peaks that are differentially accessible
including KLF6, KLF15, WT1, and TCF21 are shown below. The default
statistical test (fisher’s exact test) in Signac package was used. f
Reactome pathways associated with DARs in the podocytes are shown.
Accessible terms in KLF6^PODTA are shown in red. The default
statistical test (fisher’s exact test) in clusteProfiler package was
used.
Podocyte-specific induction of KLF6 induces CaMK1D signaling in the 1st
segment of the proximal tubule
Initial clustering of snRNA-seq data from all 4 groups demonstrated a
transcriptionally distinct cluster resembling some, but not all
components of the known segments of the PT, which we labeled
“preconditioned-PT” (Fig. [132]3a). Interestingly, this cluster was
predominant in both the nondiabetic and diabetic KLF6^PODTA groups,
suggesting these transcriptional changes in this preconditioned-PT
cluster are driven primarily by the induction of KLF6 in the podocytes
(Fig. [133]5a). In addition, the UMAP plots for all 4 groups showed a
relative shift in cell population between the preconditioned-PT and the
other PT groups (PTS1-S2, PTS1-S3, PTS3). To characterize this
preconditioned-PT cluster and explore its functional significance, we
initially determined the top DEGs for this cluster (Supplementary
Data [134]2). Key DEGs in this cluster included calcium/calmodulin
dependent protein kinase ID (Camk1d), Cms1 ribosomal small subunit
homolog (Cmss1), aminoacylase 3 (Acy3), and glutathione peroxidase 3
(Gpx3) (Fig. [135]5b). We subsequently interrogated our
preconditioned-PT cluster using previous reported markers for PT-S1-S3,
repairing/Injured PT and failed repair PT cell markers^[136]23–[137]27
and found none of the repairing/injured or “failed repair” cell markers
are expressed by the preconditioned-PT cell cluster, suggesting that
this cluster is distinct from previously reported post-injury PT
clusters (Fig. [138]5c). Subsequent pathway enrichment analysis using
WikiPathways^[139]17 for these upregulated DEGs demonstrated an
enrichment in metabolic pathways, such as electron transport chain,
glycolysis & gluconeogenesis, amino acid metabolism, tricarboxylic acid
cycle (TCA) cycle, and peroxisome proliferator-activated receptors
(PPAR) signaling, tryptophan metabolism, and fatty acid metabolism,
which are collectively pathways known to be enriched in the PT segments
under basal conditions (Fig. [140]5d). Furthermore, among the
upregulated DEGs in the preconditioned-PT cluster, Camk1d and Cmss1
were significantly upregulated in the diabetic KLF6^PODTA as compared
to the other groups (Fig. [141]5e). To explore the potential
transcriptional paths between the tubular segments and the
preconditioned-PT, we performed trajectory analysis on the snRNA
sequencing data using monocle (Supplementary Fig. [142]7a). While the
other PT segments, namely PT(S1-S2), PT(S1-S3), PT(S3) and
PT(S3)/LH(DL), showed similar ordering across pseudotime constrained
mainly on one side of branch point 1, the preconditioned-PT spanned
across both sides of branch point 1, showing both its similarity as
well as uniqueness from the other PT segments (Supplementary
Fig. [143]7b). Expression of the key DEGs of the preconditioned-PT
(Camk1d), PT(S1-S2) (Slc5a12), PT(S1-S3) (Erc2), PT(S3) (Keg1),
PT(S3)/LH(DL) (Cyp7b1) and PEC/Proliferating PT (Cd44), localized
similarly to their pseudotime ordering patterns (Supplementary
Fig. [144]7d). We further interrogated the DEGs responsible for the
branch-dependent expression across branch point 1 and the pathways they
are involved in using clusterprofiler (Supplementary Fig. [145]7c,
[146]e). We found fatty acid beta-oxidation, glycolysis and
gluconeogenesis, oxidative phosphorylation, and electron transport
chain among the significant pathways (Supplementary Fig. [147]7e).
These data suggest that a shift towards a highly metabolically active
PT characterizes the transcriptome of the preconditioned-PT cluster.
Fig. 5. snRNA-seq analysis reveals a unique proximal tubule
(preconditioned-PT) subpopulation predominant in KLF6^PODTA mice.
[148]Fig. 5
[149]Open in a new tab
a UMAP highlighting the relative shift in the nuclei between the
preconditioned-PT and other PT clusters across the groups, red outline
indicates higher nuclei count and blue outline indicates lower nuclei
count. b Dot plot showing highly upregulated genes in the
preconditioned PT cluster. c Dot plot showing injury markers from
previously reported datasets (PT-S1-S3, repairing injured PT, failed
repair PT) in the preconditioned-PT cluster as compared to other PT
clusters. d Heatmap showing pathway enrichment analysis (WikiPathway)
for differentially upregulated genes in the preconditioned PT cluster
as compared to all other clusters. The default statistical test
(fisher’s exact test) in Seurat package was used. e Dot plot showing
differential expression of upregulated genes across the groups. f
Representative images of CaMK1D and lotus lectin immunostaining. g
Quantification of CaMK1D staining per high power field (hpf - 20x)
images (n = 10 hpf/mouse, n = 3 mice/group; p = < 0.0001,
****p < 0.0001; Mann–Whitney test). h Camk1d expression (relative to
mouse actin) in isolated primary proximal tubule cells (PTEC), parietal
epithelial cells (PEC), and podocytes (Pod) (n = 3 mice/group,
p = 0.0381 for PTEC vs PEC and p = 0.0316 for PTEC vs Pod, *p < 0.05;
ordinary one-way ANOVA; data presented as mean ± SEM). i Representative
images of CaMK1D and PHA-E staining in human biopsy tissue in healthy
donor specimens, DKD (<30% and > 30% fibrosis) (n = 3 mice/group). j
Quantification of CaMK1D staining in PHA-E^+ tubules (n = 33–35
tubules; p = 0.0036 for healthy donors vs < 30% fibrosis, p = <0.0001
for healthy control vs > 30% fibrosis, p = 0.0284 for < 30% fibrosis
vs > 30% fibrosis, *p < 0.05, **p < 0.01, ****p < 0.0001;
Kruskal–Wallis test with Dunn’s post-test). Source data are provided as
a Source Data file.
To determine the spatial location of these unique PT cell populations,
we costained for CaMK1D (high enrichment in the preconditioned-PT
cluster) and lotus lectin in NPHS2-rtTA and KLF6^PODTA mice, which
showed granular and apical staining specifically in the 1st segment of
the PT, proximal to the podocytes (Fig. [150]5f, Supplementary
Fig. [151]7f). Furthermore, this PT-specific CaMK1D expression was
increased in the KLF6^PODTA mice as compared to NPHS2-rtTA mice
(Fig. [152]5f, g). We also validated that Camk1d mRNA expression was
enriched in PT cells in isolated primary PT cells as compared to
neighboring parietal epithelial cells (PECs), and podocytes
(Fig. [153]5h). In addition, CaMK1D was enriched in the PT segments
with coexpression of CaMK1D in PHA-E^+ve cells in healthy donor
nephrectomies (Fig. [154]5i). Interestingly, this PT-specific CaMK1D
expression was reduced in kidney biopsies with early (<30% fibrosis),
prior to significant loss of PT segments, as well as late (>30%
fibrosis) DKD as compared to control specimens (Fig. [155]5i, j). Based
on these data, the sole induction of podocyte-specific KLF6 triggers
CaMK1D signaling and pro-metabolic pathways in the 1st segment of PT,
which might precondition the PT cells against diabetic injury.
The requisite role of CaMK1D in the proximal tubule
In the setting of podocyte-specific KLF6 induction, we report an
increase in CaMK1D expression and an enrichment in DEGs involving
oxidative phosphorylation and electron transport chain in the 1st
segment of PT, proximal to the podocytes. To ascertain the role of
CaMK1D in the kidney, we initially knocked down CAMK1D in HK2 cells
(CAMK1D-shRNA). CAMK1D knockdown was confirmed using qPCR, western blot
and immunofluorescence staining (Fig. [156]6a–c). CAMK1D-shRNA cells
exhibited a reduction in cell viability (cell count over time) as
compared to EV-shRNA cells (Fig. [157]6d). These findings were
validated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) assay (Fig. [158]6e). In addition, CAMK1D-shRNA cells had
reduced mitochondrial membrane potential as compared to EV-shRNA cells
(Fig. [159]6f). To assess for changes in mitochondrial structure, we
initially stained for translocase of the outer mitochondrial membrane
(TOM20), which demonstrated an increase in mitochondrial fragmentation
in CAMK1D-shRNA cells as compared to EV-shRNA cells, as demonstrated by
an increase in the % of cells with fragmented mitochondria as compared
to tubular mitochondria (Fig. [160]6g, h). Since mitochondrial fission
is inhibited with the phosphorylation of Dynamin related protein 1
(pDRP1) at Ser637^[161]28, we observed that CAMK1D-shRNA cells
exhibited a reduction in pDRP1(ser637) expression as compared to
EV-shRNA cells (Fig. [162]6i).
Fig. 6. Genetic and pharmacological inhibition of CaMK1D induces
mitochondrial injury and reduces cell survival.
[163]Fig. 6
[164]Open in a new tab
a Fold change in CAMK1D expression (n(biological replicates) = 4 for
EV-ShRNA and n(biological replicates) = 5 for CAMK1D-shRNA, p = 0.0159,
*p < 0.05; Mann–Whitney test, two-sided; data presented as mean ± SEM).
b Western blots for CaMK1D and β-actin. c Representative images for
CaMK1D and Hoechst immunostaining. d Cell count at days 0, 2, 3 and 6
(n(biological replicates) = 3, p = 0.0025, **p < 0.01; two-way ANOVA;
data presented as mean ± SEM). e Cell viability (n(biological
replicates) = 4, p = 0.0286, *p < 0.05; Mann–Whitney test, two-sided;
data presented as mean ± SEM). f Mitochondrial membrane potential
(n(biological replicates) = 3, p = 0.0071, **p < 0.01; Welch’s t-test,
two-sided; data presented as mean ± SEM). g Representative images for
TOM20 and Hoechst staining. Arrow-tubular mitochondria;
Arrowheads-indicates fragmented mitochondria. h % of cells with
tubular, intermediate, and fragmented mitochondria (n(biological
replicates) = 30, for three independent experiments, p = <0.0001 for
tubular and fragmented mitochondria ****p < 0.0001, two-way ANOVA; data
presented as mean ± SEM). i Western blots for pDRP1(637), DRP1, and
β-actin. j Cell viability in 1° PT cells (n(biological replicates) = 8,
p = 0.0039 for DMSO vs STO609 (20 µg/ml) and p = 0.0007 for DMSO vs
STO609 (50 µg/ml), **p < 0.01, ***p < 0.001; Kruskal–Wallis test with
Dunn’s posttest; data presented as mean ± SEM). k–l OCR in 1° PT cells
(n(biological replicates) = 6, p = < 0.0001 for DMSO vs STO609
(20 µg/ml and 50 µg/ml), ****p < 0.0001; one-way ANOVA; data presented
as mean ± SEM). m, n OCR with quantifications in 1° PT cells in normal
(NG) and high glucose (HG) conditions (n(biological replicates) = 5,
p = 0.0086 for basal respiration (NG + STO609 vs HG + STO609),
p = 0.0056 for ATP production (NG + STO609 vs HG + STO609), p = <0.0001
for NG + DMSO vs NG + STO609 and HG + DMSO vs HG + STO609 for all other
significant comparisons, **p < 0.01, ****p < 0.0001; one-way ANOVA;
data presented as mean ± SEM). o, p Representative images of TOM20 and
Hoechst staining and quantification of % of cells with tubular,
intermediate, and fragmented mitochondria in 1° PT cells Arrow-tubular
mitochondria; Arrowheads indicates fragmented mitochondria
(n(biological replicates) = 30, for three independent experiments,
p = 0.0389 for tubular NG vs HG, p = 0.0014 for fragmented NG vs
HG + STO609, p = 0.001 for tubular NG vs HG + STO609, *p < 0.05,
**p < 0.01, ***p < 0.001; two-way ANOVA; data presented as mean ± SEM).
q Western blots for pDRP(637), DRP1, and β-actin in 1° PT. Source data
are provided as a Source Data file.
To validate this detrimental effect of the loss of CAMK1D on
mitochondria specifically in primary PT cells, we pharmacologically
inhibited CaMK1D with STO-609, which inhibits CAMKK, leading to a loss
of phosphorylation and subsequent inactivation of
CaMK1D^[165]29,[166]30. Primary PT cells treated with STO-609 at
20 µg/ml and 50 µg/ml resulted in reduced cell viability (Fig. [167]6j)
as well as a reduction in basal respiration, ATP production, maximal
respiration, and spare respiratory capacity as compared to DMSO-treated
cells (Fig. [168]6k, l). Furthermore, STO-609-treated cells also
exhibited a reduction in basal respiration, ATP production, maximal
respiration, and spare respiratory capacity as compared to DMSO-treated
cells under normal glucose (NG, 5 mM) conditions, which was further
exacerbated under high glucose (HG, 30 mM) conditions (Fig. [169]6m,
n). TOM20 staining also showed an increase in mitochondrial
fragmentation in STO-609-treated cells as compared to DMSO-treated
cells under HG conditions (Fig. [170]6o, p). In addition, STO-609
reduced pDRP1(ser637) expression in primary PT cells as compared to
DMSO under HG conditions, suggesting an increase in mitochondrial
fission in the presence of pharmacological CaMK1D inhibition
(Fig. [171]6q). Under diabetic conditions, we also observed that STO609
increased PT injury (loss of brush border staining, tubular dilatation,
and reduced lotus lectin expression) as compared to DMSO treatment in
the KLF6^PODTA mice (Supplementary Fig. [172]8a, b). STO609-treated
mice also had an increase in interstitial fibrosis (increase in
picrosirius red and α-SMA expression) as compared to DMSO-treated mice,
suggesting that CaMK1D inhibition exacerbated kidney injury under
diabetic conditions (Supplementary Fig. [173]8a–c).
Increased release of ApoJ from the KLF6^+ podocytes triggers PT CaMK1D
signaling, leading to a reduction in PT injury
To ascertain the mechanism by which podocyte-specific KLF6 triggers
CaMK1D signaling in PT cells, primary PT cells were initially treated
with conditioned media from primary mouse podocytes isolated from
KLF6^PODTA mice as compared to NPHS2-rtTA mice under NG and HG
conditions. While the NPHS2-rtTA podocyte secretome reduced cell
respiration in PT cells, this was restored with exposure to the
KLF6^PODTA podocyte secretome under HG conditions (Fig. [174]7a, b).
Tandem mass spectrometry was subsequently performed on the conditioned
media, after removal of cell debris, to determine the differentially
secreted proteins between KLF6^PODTA and NPHS2-rtTA podocytes. A total
of 838 proteins were identified, with 321 proteins significantly
increased [fold change (FC > 1.2)] and 132 proteins significantly
decreased (FC < 0.77) in the KLF6^PODTA compared with the NPHS2-rtTA
conditioned medium (Fig. [175]7c, Supplementary Data [176]5). Key
proteins such as ApoJ, calreticulin (CALR), clathrin light chain A
(CLTA), clathrin light chain B (CLTB), collagen type IV alpha 1
(COL4A1) were differentially expressed in the conditioned medium from
KLF6^PODTA as compared to NPHS2-rtTA podocytes. In comparison, a total
of 142 proteins were identified, with 30 proteins significantly
increased (FC > 1.3) in the urine proteome of KLF6^PODTA as compared to
NPHS2-rtTA mice (Fig. [177]7d, Supplementary Data [178]5).
Interestingly, ApoJ was uniquely enriched in both the podocyte
secretome and the urine proteome of KLF6^PODTA as compared to the
NPHS2-rtTA mice, suggesting an increase in the release of ApoJ from the
podocytes in presence of KLF6 induction. ApoJ, also known as clusterin,
is a glycoprotein that has been previously reported to have a salutary
effect in the kidney^[179]31–[180]34. Specifically, pre-treating
podocytes with recombinant ApoJ has shown to be protective against
apoptosis by reducing oxidative-stress induced under diabetic
conditions^[181]33. To determine the expression of ApoJ in the
glomeruli under diabetic conditions, we interrogated previous reported
microarrays using Nephroseq to compare human biopsy specimens with DKD
as compared to healthy donor nephrectomies^[182]35,[183]36. ApoJ
expression was significantly increased in microdissected glomeruli in
DKD patients compared to healthy donors and it inversely corelates with
estimated GFR (Supplementary Fig. [184]9a, [185]b). To ascertain
potential ligand-receptor interactions between podocytes and PT cells,
a previously validated database of known ligand-receptor
interactions^[186]37 was interrogated to predict potential ligands from
the podocyte secretome and urine proteome as well as corresponding
receptors from the DEGs in the preconditioned-PT cell cluster
(Fig. [187]7e). ApoJ, which was identified in the podocyte secretome
and the urine proteome of KLF6^PODTA mice, demonstrated a potential
ligand-receptor interaction with Lrp2 in the preconditioned-PT cluster
(Fig. [188]7e). Interestingly, depending on the degree of
glycosylation, the secretory form of ApoJ has been previously reported
to serve as a molecular chaperone by binding to specific cell surface
receptors to mediate its biological effects, such as
endocytosis^[189]38. Immunohistochemistry confirmed that ApoJ was
enriched in the podocytes and the apical portion of the 1st PT segment
in KLF6^PODTA as compared to NPHS2-rtTA mice (Fig. [190]7f).
Additionally, we costained ApoJ with podocyte marker, synaptopodin, and
found increased colocalization of ApoJ in podocytes (Supplementary
Fig. [191]9c). While ApoJ is present at high levels in the plasma and
prevents complement deposition^[192]39, we did not observe an increase
in complement deposition (C3 and C5b-9 expression) in KLF6^PODTA as
compared to NPHS2-rtTA mice (Supplementary Fig. [193]9d).
Fig. 7. Secretory ApoJ from KLF6^PODTA 1° mouse podocytes restores
mitochondrial respiration under high glucose conditions.
[194]Fig. 7
[195]Open in a new tab
a, b OCR and quantifications in 1° PT treated with conditioned media
(CM) from 1° KLF6^PODTA (OE) versus NPHS2-rtTA (WT) podocytes in normal
glucose (NG) and high glucose (HG). n (biological replicates) = 7;
basal respiration: p = 0.0013:OE CM (NG) vs WT CM (HG), p = 0.0008:OE
CM (NG vs HG), p = 0.0002:WT CM vs OE CM (HG); ATP production:
p = 0.0006:OE CM (NG) vs WT CM (HG), p = 0.0004:WT CM (NG vs HG),
p = < 0.0001:WT CM vs OE CM (HG); maximal respiration: p = 0.0343:WT CM
(NG vs HG), p = 0.0389:OE CM (NG) vs WT CM (HG); spare respiratory
capacity: p = 0.0143:OE CM (NG) vs WT CM (HG), p = 0.0117:WT CM (NG vs
HG). c, d Volcano plot showing podocyte secretome and urine proteome
from mice (n = 3). e Circos plot highlighting ligand-receptor
interactions for upregulated proteins in podocyte secretome and DEGs in
preconditioned-PT from snRNA-seq. f Representative images of ApoJ
immunohistochemistry. Arrows-ApoJ staining (n = 3 mice/group, scale
bar = 25 µm). g, h OCR and quantifications in 1° PT cells treated with
CM + anti-ApoJ blocking antibody or CM + anti-IgG in HG (n (biological
replicates) = 7); WT CM vs OE CM: p = 0.0259(basal respiration),
p = 0.0026(maximal respiration), p = 0.0004(spare respiratory
capacity). i, j OCR and quantifications in 1° PT treated with CM from
differentiated human podocytes in HG (n (biological replicates) = 6,
p = 0.0001(basal respiration), p = 0.0009(ATP production),
p = 0.005(maximal respiration), p = 0.0156(spare respiratory
capacity)). k Western blot for immunoprecipitation of ApoJ in 1° PT
treated with CM in NG/HG. l, m Representative images from
co-immunostaining for ApoJ, CaMK1D, Lrp2, and Hoechst. Arrows-indicate
ApoJ/CamK1D and ApoJ/Lrp2 colocalization. (scale bar = 50 µm, 5 µm for
inset). n, o OCR and quantification in 1° PT treated with recombinant
ApoJ ± cilastatin/VEH (n (biological replicates) = 7 for control), (n
(biological replicates)=5) for ApoJ+VEH, (n(biological replicates) = 5)
for ± cilastatin; basal respiration: p = 0.0003(control vs ApoJ + VEH),
p = 0.0157(control vs cilastatin), p = 0.0329 (ApoJ + VEH vs
ApoJ + cilastatin); ATP production:p = 0.0001 (control vs ApoJ + VEH),
p = 0.0086 (ApoJ + VEH vs ApoJ + cilastatin); maximal
respiration:p = 0.0151 (control vs ApoJ + VEH). p Proposed schematic of
potential KLF6-ApoJ-CaMK1D signaling between podocytes and PT cells.
For all data: *p < 0.05, **p < 0.01 ***p < 0.001, Kruskal–Wallis test
with Dunn’s posttest; data presented as mean ± SEM. Source data are
provided as a Source Data file. p Created with BioRender.com released
under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0
International license
[196]https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en.
To determine whether the PT-protective effects of the KLF6^PODTA
podocyte secretome are driven largely by ApoJ, the conditioned media
from KLF6^PODTA and NPHS2-rtTA podocytes were incubated with goat
anti-ApoJ antibody (1:100) as compared to goat IgG antibody for 24 h
prior to being administered to primary PT cells. While the KLF6^PODTA
podocyte secretome improved PT basal respiration, maximal respiration,
and spare respiratory capacity, blocking of ApoJ attenuated these
protective effects on the mitochondria (Fig. [197]7g, h). To further
validate these findings, we initially generated human podocytes with
stable APOJ overexpression (Lenti-ORF–APOJ) as compared to control
podocytes (Lenti-ORF-control) (Supplementary Fig. [198]10a, [199]b). PT
cells treated with the conditioned media from Lenti-ORF-APOJ podocytes
had higher basal respiration, ATP production, maximal respiration, and
spare respiratory capacity as compared to Lenti-ORF-control podocytes
under HG conditions (Fig. [200]7i, j). Interestingly, the sole
induction of APOJ also increased canonical podocyte markers of
differentiated podocytes such as nephrin and synaptopodin, suggesting
that ApoJ might also have salutary effects on podocyte health
(Supplementary Fig. [201]10c).
Since calmodulin is critical component of CaMK1D activation and
signaling, we tested the interaction between ApoJ and calmodulin by
incubating the primary PT cell fractions that have been exposed to
KLF6^PODTA and the NPHS2-rtTA podocyte secretome under NG and HG
conditions using calmodulin-sepharose beads. While the ApoJ-calmodulin
interaction was reduced in the PT cells when exposed to NPHS2-rtTA
podocyte secretome, this was restored with the KLF6^PODTA podocyte
secretome under HG conditions (Fig. [202]7k). In addition,
immunofluorescence staining demonstrated ApoJ colocalized with CaMK1D
(Fig. [203]7l). Lrp2 has been reported previously as the endocytic
receptor for ApoJ^[204]40. Since the molecular weight of Lrp2 limits
co-immunoprecipitation studies, we conducted immunofluorescence
staining with quantification using CellProfiler to demonstrate that
ApoJ colocalizes with Lrp2 in the proximal tubules (Fig. [205]7m,
Supplementary Fig. [206]11). We also treated the PT cells with
recombinant ApoJ which led to an increase in PT basal respiration, ATP
respiration, and maximal respiration. Finally, concurrent treatment
with cilastatin, inhibitor of Lrp2 receptor activity^[207]41–[208]43,
attenuated the salutary effects of ApoJ in PT cells as demonstrated by
a decrease in basal respiration, ATP-linked, and maximal respiration
(Fig. [209]7n, o). Collectively, these data suggest that the
kidney-protective effects of podocyte-specific KLF6 might, in part, be
mediated through ApoJ-CaMK1D signaling in PT cells.
Discussion
A large body of literature has focused on factors that drive kidney
injury in diabetes, but little is known about mechanisms that confer
resistance to progression of DKD. These unexplored mechanisms that
delay progression of DKD might serve as potential targets for therapy
in those individuals that progress rapidly. While podocyte injury
contributing to glomerular dysfunction is an early indicator of kidney
injury in diabetes, PT injury correlates with the decline in kidney
function^[210]3,[211]44. To date, mechanisms mediating podocyte to PT
injury remain elusive in DKD. In this study, we demonstrate that the
podocyte-specific induction of KLF6 attenuates kidney injury in a
murine model of diabetes. In addition to the salutary effects in the
podocyte, the induction of podocyte-specific KLF6 preconditions the PT
from injury under diabetic conditions. By utilizing a combination of
snRNA-seq, snATAC-seq, and LC-MS/MS, we report that podocyte-specific
KLF6 triggers the release of ApoJ, which subsequently activates
CaMK1D-mediated preservation of mitochondrial dynamics and function in
the first segment of PT, proximal to podocytes (Fig. [212]7p). To date,
this is the first study to demonstrate a novel mechanism by which the
podocyte directly regulates PT mitochondrial health to attenuate the
progression of DKD.
Our previous studies demonstrated the detrimental effects of
podocyte-specific knockdown of Klf6 in murine models of glomerular
disease (i.e., Focal Segmental Glomerulosclerosis and
DKD)^[213]9,[214]10. Here, conversely, the induction of KLF6
specifically in podocytes ameliorated albuminuria and improved kidney
function as well as histological features consistent with DKD. We also
previously showed that induction of KLF6 in cultured podocytes
attenuated detrimental effects of adriamycin in cultured
podocytes^[215]9. However, similar to other tissues, this
reno-protective effect of KLF6 appears to be cell-context dependent in
the kidney. For instance, we recently reported the detrimental role of
PT-specific KLF6 in a murine model of post-DNA damage in PT
cells^[216]45. Furthermore, this contrasting cell-specific role of KLF6
is not restricted to the kidney, with opposing roles in the liver
(hepatocytes versus hepatic stellate cells)^[217]46 as well as in the
heart (cardiac myocytes versus cardiac fibroblasts)^[218]47. One
potential explanation in the kidney is that Klf6 is expressed at a
higher level in podocytes as compared to PT cells under basal
conditions^[219]48, thereby suggesting a requisite physiological role
in the podocytes. Nonetheless, studies investigating the mechanisms
regulating cell-specific contrasting roles of KLF6 will be important to
understand its biology in kidney health and disease.
CaMKs (CaMK1, CaMK2, CaMK4, CaMKK) are multifunctional
calcium/calmodulin-dependent protein kinases that have wide specificity
but regulate several critical cellular functions in multiple cell
types^[220]49–[221]52. While CaMKs are well studied in the brain, their
biology in the kidney is not well understood. A few studies have
reported on the detrimental role of CaMK2 and CaMK4 activation in the
kidney^[222]53–[223]56, but CaMK1D remains unexplored in the kidney. In
genome-wide association studies (GWAS), single nucleotide polymorphisms
in the CAMK1D loci are associated with increased risk of Type 2
Diabetes Mellitus^[224]57–[225]62. Validation of our snRNA-seq with
immunostaining demonstrates an enrichment of CaMK1D expression in the
first segment of PT, proximal to podocytes. Interestingly, genetic and
pharmacological inhibition of CaMK1D reduced pDRP1(S637) expression,
leading to an increase in mitochondrial fragmentation with a reduction
in mitochondrial membrane potential and respiration. DRP1 is a critical
regulator of mitochondrial dynamics and posttranslational modifications
are critical to its function^[226]63. Previous studies report that
CaMKI might be involved in the regulation of DRP1 activity^[227]64, but
the mechanism(s) mediating this process remains to be investigated.
Furthermore, future studies are needed to examine the interplay between
CaMK1D and other isoforms of CaMKI signaling on pDRP1 activity and
mitochondrial health^[228]64–[229]66. Nonetheless, this is the first
study, to date, to demonstrate the protective role of CaMK1D in the
kidney.
Another key finding in this study is the role of ApoJ in mediating the
salutary effects of podocyte-specific KLF6 on CaMK1D activation in the
PT. ApoJ undergoes posttranslational modification, largely
glycosylation, in the endoplasmic reticulum and golgi apparatus prior
to being released extracellular space. Secretory ApoJ serves as a
molecular chaperone to regulate various cellular processes such as
lipid transport, cell differentiation, membrane cycling, apoptosis, and
cell-cell interactions^[230]67. Several studies have implicated ApoJ in
a host of human diseases^[231]68–[232]72, ranging from myocardial
injury^[233]73 to post-ischemic brain injury^[234]74 to Alzheimer’s
disease^[235]75. In the kidney, ApoJ expression has been previously
shown to increase post-injury^[236]76–[237]79. Specifically, both
protein and mRNA expression levels of ApoJ are increased in podocytes
in diabetic mice and in kidney biopsies with early DKD^[238]33.
Similarly, we validated this in expression arrays deposited in
Nephroseq. In addition, ApoJ levels in the urine have been associated
with increased tubular damage in patients with diabetes^[239]80. We
also observed an enrichment in ApoJ levels in both the urine and
podocyte secretome from KLF6^PODTA as compared to NPHS2-rtTA mice by
LC-MS/MS. Interestingly, this increase in levels of secretory ApoJ
might be in response to podocyte injury, since we observed that
blocking ApoJ in the KLF6^PODTA podocyte secretome attenuated the
salutary effects in the PT cells under HG conditions. Furthermore, ApoJ
knockout mice develop glomerulopathy with aging^[240]76 and individuals
with glomerular disease demonstrate an overall depletion in the pool of
ApoJ with progressive disease^[241]32. In addition, salutary effects of
ApoJ have been shown in other models of kidney injury^[242]81,[243]82.
However, to date, the mechanism by which secretory ApoJ attenuates
kidney injury remains poorly understood^[244]83. Here, we demonstrate
that induction of KLF6 triggers the release of secretory ApoJ, which
undergoes cellular uptake by Lrp2, triggering the activation of CaMK1D
by binding to calmodulin in PT cells. snRNA-seq analysis demonstrated
an enrichment in pathways associated with N-glycosylation in the DEGs
with KLF6 binding sites. Glycosylated intracellular ApoJ forms may act
as a redox sensor under oxidative stress conditions and are essential
for its chaperone activity as well as release of ApoJ into the
extracellular space^[245]84,[246]85. Therefore, the induction of KLF6
could potentially increase glycosylation of ApoJ, thereby triggering
the release of secretory ApoJ from the podocytes. However, additional
studies are required to test the mechanism by which KLF6 leads to
glycosylation of ApoJ. Interestingly, ApoJ has been previously reported
to serve as a ligand for Lrp2-facilitated endocytosis in the
brain^[247]40,[248]86,[249]87 and has a putative calmodulin binding
domain containing three motifs (two 1–12, one 1–14)^[250]88. A recent
study using Ingenuity Pathway Analysis also showed KLF6 motifs in the
regulatory regions of ApoJ as well as potential interactions between
calmodulin and Lrp2^[251]89. Calcium and calmodulin also play a role in
supporting the endocytosis of ApoJ as calmodulin has been previously
reported to serve as a calcium sensor for endocytosis in
synapses^[252]90. Furthermore, the proximity of podocyte secretory ApoJ
to Lrp2 and calmodulin-CaMK1D in the first segment of the PT, in
combination with the direction of glomerular filtrate, could enhance
the feasibility of this interaction. ApoJ has also been reported to
protect endothelial cells by suppressing mitochondrial fission under
diabetic conditions^[253]91. In addition, ApoJ has been reported to
facilitate mitochondrial respiration in the healthy brain^[254]92 and
overexpression of ApoJ attenuated Drp1 activation, thereby inhibiting
mitochondrial fission^[255]93. Based on our findings, we postulate that
ApoJ-mediated activation of CaMK1D-pDRP1 signaling protects the PT
against injury by inhibiting mitochondrial fission under diabetic
conditions.
Collectively, these studies uncover a previously unreported mechanism
by which the podocyte preconditions the proximal tubule to attenuate
mitochondrial injury and subsequent deterioration of kidney function
under diabetic conditions. In addition to this requisite role of
podocyte-specific KLF6 in podocytes, we also provide evidence for a new
downstream signaling pathway involving secretory ApoJ-Lrp2-CaMK1D-pDrp1
between podocytes and the first segment of PT that might be critical
for kidney health. Finally, maintaining secretory ApoJ from podocytes
or enhancing CaMK1D signaling in the PT might be a key therapeutic
strategy in attenuating the progression of DKD.
Methods
Generation and validation of KLF6^PODTA mice
All animal studies were approved by the Stony Brook Animal Care and Use
Committee and carried out in accordance with the National Institutes of
Health standards. To generate KLF6^PODTA mice, NPHS2-rtTA mice
(FVB/N-Tg(NPHS2-rtTA2*M2)1Jbk/J, The Jackson Laboratory) were bred with
the TRE-KLF6 mice to generate mice with both transgenes on the FVB/N
background. The TRE-KLF6 transgene contained the (TetO)7/CMV regulatory
element driving the full-length human KLF6 coding sequence (ORF021674)
followed by the polyadenylation signal^[256]45. Transgene was purified
from plasmid vector sequences and microinjected into the pronucleus of
FVB/N single-celled embryos to generate TRE-KLF6 mice. DNA extraction
by Extracta DNA prep (Quanta Biosciences) from tails at 2 weeks of age
and PCR to confirm the genotype. Primers for genotyping are provided in
Supplementary Table [257]3.
To induce transgene expression, KLF6^PODTA mice were fed TestDiet
Modified Rodent Diet 5001 containing 0.15% doxycycline (DOX; El-Mel).
KLF6^PODTA mice and littermate controls remained on DOX continuously
starting at 8 weeks of age. To validate induction of podocyte-specific
KLF6 in KLF6^PODTA mice, we generated fluorescent labeled TRE-GFP mice
and bred them with KLF6^PODTA mice or NPHS2-rtTA mice. All mice were
treated with DOX. Glomeruli were subsequently isolated from these mice
using iron oxide and the glomerular and tubular fraction was separated
using a magnet after digestion with collagenase for 45 min at 37 °C as
previously described^[258]94. The purity of glomeruli was verified
under microscopy. Both fractions were digested using collagenase, DNase
and trypsin to make a single cell suspension. The single cell
suspension for the glomerular fractions was sorted using
fluorescence-activated cell sorting (FACS) to isolate GFP(+ve) cells.
GFP(-ve) glomerular and tubular cells were also collected for further
analysis. Podocyte-specific KLF6 expression was determined in GFP(+ve)
glomerular cells as compared to GFP(-ve) glomerular and tubular cells.
Animal experiments
The animals in this study were housed in our animal facility with free
access to chow (Lab diet 5053/Dox diet) and water and 12 h day/light
cycle, in ambient temperature and humidity conditions.
Baseline urine was collected from the NPHS2-rtTA and KLF6^PODTA mice.
Mice were anesthetized and UNx was performed as previously
described^[259]95. In brief, the blood vessels of the left kidney were
ligated, and the kidney removed surgically, and the mice were monitored
for a week after surgery. To induce diabetes, mice were administered
STZ (50 mg/kg) in 50 mmol/L sodium citrate buffer (pH 4.5) by
intraperitoneal injection over the course of 5 days^[260]10,[261]96.
On day 14 after the STZ injections, blood glucose was measured from the
tail vein using a OneTouch glucometer (LifeScan)^[262]94. Diabetes was
defined as sustained fasting blood glucose level > 250 mg/dL. All mice
were euthanized using intraperitoneal injection of Ketamine/Xylazine
(150/20 mg/kg) at 20 weeks of age.
For the STO-609 experiments, UNx-STZ KLF6^PODTA mice were treated with
DMSO or STO-609 (30 µm/kg) daily intraperitoneal injection for 2 weeks.
The mice were euthanized using intraperitoneal injection of
Ketamine/Xylazine (150/20 mg/kg) at 16 weeks of age.
Measurement of albuminuria, serum urea nitrogen and creatinine
The animals were housed in a single mouse metabolic cage (Tecniplast)
with free access to food and water, and urine was collected after 24 h.
Albumin concentration was measured using ELISA assay kit (Bethyl
Laboratory). Twenty-hour urine albumin concentration was calculated by
multiplying total volume of urine collected and albumin concentration.
Serum urea nitrogen levels were measured by a colorimetric detection
method (Arbor Assay) according to the manufacturer’s protocol. Serum
creatinine levels were measured using the isotope dilution liquid
chromatography-tandem mass spectrometer at the University of Alabama at
the Birmingham O’Brien Core Center.
Real-time PCR
Total RNA was extracted using TRIzol (Gibco) for cells and RNA easy kit
(Qiagen) for kidney tissue. First-strand cDNA was prepared from total
RNA using the SuperScript III First-Strand Synthesis Kit (Life
Technologies), and cDNA was amplified using SYBR GreenER qPCR Supermix
on ABI QuantStudio 3 (Applied Biosystems). Primers were designed using
National Center for Biotechnology Information Primer-BLAST and
validated for efficiency before application. Primer sequences are
listed in supplemental information (Supplementary Table [263]4). Light
cycler analysis software was used to determine crossing points using
the second derivative method. Data was normalized to the housekeeping
gene (Actb or ACTB) and presented as a fold increase compared to the
control group using the 2^-ΔΔCT method.
Histopathology and morphometric studies by bright-field light microscopy
Mice were perfused with phosphate buffer saline (PBS) and the kidneys
were fixed in 10% phosphate buffered formalin overnight and switched to
70% ethanol prior to processing for histology. Kidney tissue was
embedded in paraffin by histology core facility at Stony Brook
University and 4 μm thick sections were stained with periodic
acid-Schiff (PAS) (Sigma-Aldrich), hematoxylin and eosin (H&E), and
picrosirius red, according to previously reported protocols, and
mounted in permanent mounting medium^[264]45,[265]94.
Mesangial expansion, and glomerular volume were quantified as
previously described^[266]10,[267]94. In brief, images were scanned,
and glomerular areas were traced using ImageJ. Mean glomerular tuft
volume (GV) was determined from mean glomerular cross-sectional area
(GA) by light microscopy. GA was calculated based on average area of 20
glomeruli in each group and GV was calculated based on the following
equation: GV
[MATH: =βκ<
/mrow>×GA3/2 :MATH]
[β = 1.38, the shape coefficient of spheres (the idealized shape of
glomeruli), and κ = 1.1, the size distribution coefficient]. Mesangial
expansion was defined as the PAS-positive and nuclei-free area in the
mesangium. Quantification of mesangial expansion was based on 20
glomeruli cut at the vascular pole in each group.
Histological scoring for sclerotic glomeruli, tubular injury,
interstitial fibrosis, and inflammation was performed in a blinded
fashion using a semiquantitative scale from 0 to 3 (0 indicates none);
1 = mild (≤25%), 2 = moderate (>25%–50%) and 3 = severe (>50%) by the
kidney histopathologist (M.P.R.).
Immunofluorescence staining and immunohistochemistry
All kidney sections from mice were prepared for immunofluorescence
staining in identical fashion as previously described^[268]94.
Immunofluorescence staining was performed using mouse anti-WT1 (Santa
Cruz, sc-7385, 1:50 dilution), goat anti-synaptopodin (Santa Cruz,
sc21537, 1:200 dilution), rabbit anti-KLF6 (Santa Cruz, AP6588B, 1:150
dilution), mouse anti-α-SMA (Sigma-Aldrich, A1978, 1:10,000 dilution),
rabbit anti-CaMK1D (Invitrogen, PA5-21957, 1:100 dilution), goat
anti-ApoJ (Novus Biologicals, NBP1-06027, 1:100 dilution), mouse
anti-Lrp2 (Novus Biologicals, NB110-96417, 1:100 dilution) and rabbit
anti-TOM20 (Abcam, ab78547, 1:100 dilution), Fluorescein-conjugated
goat IgG to mouse complement C3 (MP Biomedicals, 085500, 1:100
dilution) and mouse anti-C5b-9 (Santa Cruz, sc-66190, 1:100 dilution)
antibodies. After washing, sections were incubated with a
fluorophore-linked secondary antibody (Alexa Fluor 647 Donkey
anti-mouse, Fluor 488 Goat anti-rabbit, or Fluor 568 Donkey anti-rabbit
from Life Technologies, 1:300 dilution). After counterstaining with
fluorescein-labeled lotus lectin (Vector Labs, 1:100 dilution) and/or
Hoechst (Invitrogen, 1:1000 dilution), slides were mounted in ProLong
gold antifade mounting media (Invitrogen) and photographed under a
Nikon Eclipse i90 microscope and DS-Qi1Mc camera.
Immunohistochemistry was conducted for ApoJ as previously
described^[269]9 using goat anti-ApoJ antibody (R&D, AF2747, 1:100
dilution). Briefly, slides were dewaxed, followed by rehydration with
decreasing concentrations of ethanol (100%, 90%, 70%). After antigen
retrieval using sodium citrate buffer at 120 °C, the endogenous
peroxidases were blocked using hydrogen peroxide (3%) in methanol for
20 min, followed by blocking with 2% non-fat milk. The sections were
incubated overnight with primary antibodies at 4 °C followed by
anti-goat horseradish peroxidase secondary (Sigma Aldrich, AP200P,
1:300 dilution). The slides were then incubated with diaminobenzidine
solution (Betazoid DAB Chromogen Kit, Biocare Medical) for 5 min,
slides were then dehydrated in ethanol and xylene followed by mounting
with permanent mounting media.
De-identified human biopsy specimens from University of Utah were
categorized into early-stage (<30%) and late-stage (>30%) chronic
tubulointerstitial fibrosis by a renal pathologist (M.P.R.). Control
kidney biopsy specimens were acquired from the unaffected pole of
kidneys that were removed because of renal cell carcinoma. The study
was approved by the Stony Brook University Institutional Review Board
(#798611).
Podocyte number was determined by dividing the number of WT1(+ve)
Hoechst(+ve) podocytes per glomerular cross-sectional area. Glomerular
synaptopodin expression was determined by measuring the percent area
stained for synaptopodin in the glomerular cross-sectional area.
Quantifications of lotus lectin and α-SMA in the cortex were determined
by measuring the percentage area stained per cross-sectional area of a
high-power field. All quantification was conducted using 20X high-power
digitized images in ImageJ.
Immunofluorescence staining and quantification of mitochondrial
fragmentation were done using TOM20 staining in cultured cells as
previously described^[270]9,[271]97. Briefly, mitochondrial morphology
was categorized in each cell, by an investigator blinded to the
experimental conditions, as tubular (>75% of mitochondria with tubular
length > 5 mm), intermediate (25%–75% of mitochondria with tubular
length > 5 mm), or fragmented (<25% of mitochondria with tubular
length > 5 mm).
Quantification of colocalization of ApoJ with Lrp2 was done using
CellProfiler^[272]98. The colocalization pipeline with minor
modifications was utilized to measure the colocalization between
fluorescently labeled ApoJ and Lrp2 to measure the degree of overlap
between them. Briefly, the images were split into two channels (ApoJ
and Lrp2), followed by illumination correction for both channels. The
images were aligned to ensure accurate positioning of the features in
both images (Supplementary Fig. [273]11b), and primary objects were
identified for each channel along with the outliers as shown in
Supplementary Fig. [274]11c, d. The primary objects from ApoJ channel
that colocalize with the Lrp2 channel were measured (Supplementary
Fig. [275]11e). A mask image indicating the colocalized areas was
generated and the area occupied by the colocalized objects and Lrp2 was
used for quantifications.
Histopathology by transmission electron microscopy
Mice were perfused with PBS and then immediately fixed in 2.5%
glutaraldehyde for transmission electron microscopy (TEM). Sections
were mounted on a copper grid and photographed under a FEI BioTwinG2
transmission electron microscope. Briefly, negatives were digitized,
and images with a final magnitude of ~ 10,000X were obtained^[276]9.
The quantification of podocyte effacement and glomerular basement
membrane (GBM) thickness was performed as previously
described^[277]96,[278]99. ImageJ was used to measure the length of the
peripheral GBM and the number of slit pores overlying this GBM length
was counted. The arithmetic mean of the foot process width (WFP) was
calculated using the following: WFP =
[MATH: π4
:MATH]
[MATH: ×∑GB
MLength∑slit
s :MATH]
; where ∑GBM length indicates the total GBM length measured in one
glomerulus, ∑ slits indicates the total number of slits counted, and =
[MATH: π4
:MATH]
is the correction factor for the random orientation by which the foot
processes were sectioned^[279]99. Quantification of GBM thickness was
performed as described^[280]100,[281]101. The thicknesses of multiple
capillaries were measured in 3–4 glomeruli per mouse. A mean of 120
measurements was taken per mouse (from podocyte to endothelial cell
membrane) at random sites where GBM was displayed in the best cross
section.
Cell culture
1° mouse podocytes were isolated from GFP-labeled KLF6^PODTA and
NPHS2-rtTA male mice using FACS and were maintained on DOX (1 μg/ml)
throughout the experimental conditions. Serum free media was collected
from these podocytes to carry out proteomic analysis.
Mouse 1° PT cells were isolated from the kidney cortex of male mice
after cardiovascular perfusion with PBS as previously
described^[282]102. In brief, the cortex is digested using collagenase
A at 37 ^οC for 1 h. The digested tissue is filtered, followed by
multiple washing and resuspension in complete PT cell media (Dulbecco’s
modified Eagle’s medium:F12 with 10 ng/L epidermal growth factor, 5 pM
T3, 3.5 mg/L ascorbic acid, 25 µg/L prostaglandin E1, 25 µg/L
hydrocortisone, 1 × insulin transferrin selenium supplement, 100
units/ml penicillin, and 100 µg/ml streptomycin).
Human kidney (HK2, ATCC, CRL-2190) cells with CAMK1D knockdown were
generated using the Genecopoeia lentiviral shRNA system with the
following construct, MSH079040-LVRU6GP-c, GGTGCTGTATATAAGAATCTT. In
brief, lentiviral particles were produced by transfecting HEK 293 T
cells with a combination of lentiviral expression plasmid DNA,
pCD/NL-BH ΔΔΔ packaging plasmid, and VSV-G–encoding pLTR-G plasmid. For
infection, viral supernatants were supplemented with 8 μg/ml polybrene
and incubated with cells for 24 h. Cells expressing shRNA were selected
with puromycin for 2–3 weeks before use in all studies. Real-time PCR,
western blot, and immunofluorescence staining were performed to confirm
CAMK1D knockdown.
Mouse 1° PT cells were treated with DMSO (vehicle) or CaMK1D inhibitor,
STO-609 (5, 10, 20 and 50 μg/ml), for 24 h under normal glucose
conditions. After dose optimization, the 1° PT cells were treated with
20 µg/ml of STO-609 for 24 h under normal (5 mM) and high glucose
(30 mM) conditions. For measuring the effect of conditioned media on
the 1° PT cells, the cells were treated with 5% conditioned media from
1° podocytes from KLF6^PODTA and NPHS2-rtTA mice under normal glucose
(NG) and high glucose (HG) conditions for 24 h. For the blocking
experiments, the conditioned media (CM) was incubated with goat
anti-clusterin antibody (R&D, AF2747) or goat IgG-antibody (1:100) for
24 h at 4 °C. The CM containing the blocking antibody was used to treat
the 1° PT cells under NG and HG conditions.
The LentiORF-APOJ clone was purchased from Genecopoeia, and stable APOJ
overexpression was achieved by lentiviral delivery. Cells expressing
LentiORF-APOJ were selected with puromycin for 2–3 weeks prior to use
in all studies. LentiORF-control serves as the GFP control vector. qPCR
and western was performed to confirm APOJ overexpression.
Mouse 1° PT cells were treated with recombinant mouse ApoJ-His-tag
(His-ApoJ) (20 µg/ml) (SinoBiological, 50485-M08H) and Cilastatin
(0.1 mg/ml) (MedChemExpress, HY-A0166A), for 24 h under NG conditions.
Oxygen consumption rate (OCR) and extracellular acidification rate
(ECAR) were measured using a Seahorse XFe96 Analyzer (Agilent) in the
presence or absence of mitochondrial function inhibitors as previously
reported for all the different experimental conditions^[283]45. In
brief, cells were seeded into 96 well plates at 2 × 10^6 cells per
well, and 24 h later, the growth media was removed, cells were washed
with PBS. DMSO, STO-609 or conditioned media was added with or without
NG or HG, and 24 h later the media was removed and replaced with
serum-free Seahorse DMEM, pH 7.4, supplemented with 1 mM glutamine and
2 mM pyruvate, NG or HG. After incubation in a CO[2] free incubator for
45–60 min, OCR and ECAR were measured at baseline or after acute
injections of 1.5 μM oligomycin (Agilent), 3 μM carbonyl
cyanide-p-(trifluoromethoxy) phenylhydrazone (FCCP) (Agilent), and
0.5 μM rotenone/antimycin A (Agilent) using the Seahorse XF Cell Mito
Stress Test Kit (Agilent).
Mitochondrial membrane potential was measured using MitoProbe DiIC1(5)
Assay Kit (Invitrogen). In brief, cells were trypsinized, washed with
PBS, and incubated with 1,1′,3,3,3′,3′-hexamethylindodicarbo-cyanine
iodide (DilC1) alone or with DilC1 and carbonyl cyanide 3-
chlorophenylhydrazone (CCCP) and the difference in fluorescence
intensity was measured between the groups using FACS.
Cell numbers were determined using the Countess 3 cell counter at days
0, 2, 3 and 6. Cell proliferation was also measured using the 3-(4,
5-dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide (MTT) assay.
GFP-labeled primary mouse podocytes were obtained using FACS sorting.
Other cell lines were validated in previous publications or by
manufacturer’s website
Western blot assay
Protein lysates were collected from HK2 cells and 1° PT cells with a
buffer containing 4% SDS, and protease phosphatase inhibitor. Protein
lysates were subjected to immunoblot analysis using rabbit anti-CaMK1D
(Proteintech, 13613-1-AP, 1:1000 dilution), rabbit anti-DRP1
(Invitrogen, MA5-26255, 1:1000 dilution), rabbit
anti-phospho-DRP1(Ser637) (Invitrogen, PA5-101038, 1:1000 dilution),
goat anti-ApoJ (R&D, AF2747, 1:1000 dilution), rabbit anti-ApoJ (Cell
Signaling technology, 42143, 1:1000 dilution), and mouse anti-β-actin
(Sigma-Aldrich, A1978, 1:5000 dilution) antibodies. Uncropped and
unprocessed scans are included in the Source data file.
Immunoprecipitation
Protein lysates were collected from 1° PT cells treated with 5%
conditioned media with Pierce IP Lysis Buffer (Thermo Fisher
Scientific, 87788) and Halt protease and phosphatase inhibitor cocktail
(Thermo Fisher Scientific, 78438). Pull-down was performed using
calmodulin (CaM)-sepharose beads (Abcam, ab286869) as previously
described^[284]103. Briefly, protein lysates from 1° PT cells were
incubated with CaM-sepharose beads at 4 °C for 3 h on a shaker,
centrifuged, unbound protein was removed, and the beads were incubated
with elution buffer at room temperature for 30 min on a shaker. The
beads were then centrifuged, and the bound protein was subjected to
western blot analysis to determine the amount of ApoJ pulled down with
CaM beads.
Single-nuclei isolation, sequencing, data processing, and analysis
Nuclei were isolated from the mouse kidney cortex based on previous
protocol^[285]94,[286]104. In brief, mice were perfused with PBS and
kidney cortex was stored in RNAlater for snRNA-seq, whereas for
snATAC-seq, kidney cortex without any RNAlater at -80 °C. A 2 mm^3
section of tissue was rinsed briefly with PBS and minced before adding
1 mL of lysis buffer containing 20 mM Tris-HCl (pH 8), 320 mM sucrose,
5 mM CaCl[2], 3 mM MgAc[2], 0.1 mM EDTA, 0.1% TritonX-100, 0.1% RNase
Inhibitor and 0.1% DAPI. The tissue was initially dissociated by
pipetting up and down 10X with a p-1000 tip and then being passed
through a 25 G syringe 10X. The tissue was incubated on ice for 10 min
and then passed through a 30 µm CellTrics filter. Nuclei were pelleted
by centrifugation (5 min, 500 g, 4 °C) and washed with PBS after
removal of supernatant. Nuclei were pelleted again and resuspended in
1 mL of PBS containing 0.04% BSA and RNAse inhibitor (0.2 u/µL) and
then passed through a 40 µm FlowMI filter, followed by a 5 μm
pluriStrainer Mini filter before generating counts with hemocytometer.
Nuclei were then diluted and prepared for snRNA-seq and snATAC-seq with
the 10X Chromium System according to the manufacturer’s instructions
(10X Genomics). Sequencing was performed using a NovaSeqS4 platform.
Raw sequencing data was demultiplexed and aligned to a mouse pre-mRNA
reference genome using Cell Ranger 3.1.0-v2 on SeaWulf, the HPC Cluster
at Stony Brook University. For snRNA-seq, quality control,
dimensionality reduction and clustering were performed using the
R-package Seurat 4.3.0^[287]21. Genes expressed in a minimum of 3 cells
were retained. Cells expressing < 200 or > 7500 genes were excluded.
Cells expressing > 5% mitochondrial genes were also excluded. For the
snATAC-seq, Cell Ranger atac-2.0.0, Signac 1.8.0^[288]20 and Archr
1.0.2^[289]22 were used for subsequent analysis. Quality control for
each single cell were conducted based on peak_region_fragments>100,
peak_region_fragments <60000, pct_reads_in_peaks>5,
blacklist_ratio<0.05, nuclesosome_signal<4, and TSS.enrichment>1.
For chromatin immunoprecipitation (ChIP)-enrichment analysis, KLF6
binding site data were obtained from ChIP-sequencing data deposited in
the Gene Expression Omnibus database (accession no. [290]GSE96355), and
locations of binding sites were determined using the Genomic Regions
Enrichment of Annotations Tool (GREAT)^[291]105 “basal plus extension”
approach, with a maximum extension of 10 kb from any transcription
start site (TSS). Heat maps were generated using Morpheus software
([292]https://software.broadinstitute.org/morpheus) and genes were
clustered using the one minus Pearson correlation method. Pathway
enrichment analyses of the significantly differentially expressed genes
were undertaken using Enrichr^[293]14,[294]15 libraries of KEGG 2019
(mouse) pathways^[295]18, WikiPathways 2019 (mouse)^[296]17, and
Reactome pathways^[297]16. Trajectory analysis was performed using the
R package, Monocle-v2^[298]106,[299]107. Ligand receptor interactions
were shown in the circle plot using R package, edgebundler
0.1.4^[300]108. Pathway enrichment analyses of the genes in the
trajectory analysis and for differentially accessible regions (DARs)
was carried out using clusterProfiler 4.6.2^[301]109,[302]110.
Proteomics
Tandem mass spectrometry was performed on the supernatant from mouse
primary podocytes isolated from KLF6^PODTA and NPHS2-rtTA mice (with
and without doxycycline) using a label-free proteomic approach. After
the cells reached confluency, cells were washed five times with 1 × PBS
and medium was replaced with phenol-red, insulin-transferrin-selenium
(ITS), and serum-free RPMI. After 48 additional hours, the CM harvested
and centrifuged for 5 min at 500 g and then for 10 min at 1500 g.
Samples were reduced (dithiothreitol), alkylated (iodoacetamide),
digested with trypsin, and cleaned using hydrophobic-lipophilic-balance
(HLB) pack C-18. For the urine proteomics, the urine samples from
NPHS2-rtTA and KLF6^PODTA mice were spun at 2000 g for 10 min at 4 °C,
the supernatant was mixed with a laemmli buffer and heated at 100 °C
for 5 min. The protein amount was normalized using creatinine. The
samples were run in a 12% criterion TGX gel, the gel was fixed with 40%
EtOH/10% acetic acid at room temperature for 15 min. Following a wash
with water, the gel was stained with coomassie overnight with gentle
rocking. The gel bands were excised based on their molecular weight and
cut into small pieces, before undergoing in-gel reduction
(dithiothreitol), alkylation (iodoacetamide) and destaining (ammonium
bicarbonate/acetonitrile). The dried gel pieces were digested with
trypsin and cleaned through a C-18 column. For both, CM and urine
proteomics, samples (4 μl) were injected onto a 20 cm-long ReproSil
C-18 (3 μM particle) column and run on the Q Exactive HF Hybrid
Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific) or
TripleTOF 5600+ (Sciex). Analysis was carried out using Thermo Fisher
Proteome Discoverer v2.4 and Scaffold.
Statistical analysis
All statistical analysis was performed using GraphPad Prism 9.0. Based
on the distribution of the data, the appropriate parametric or
nonparametric statistical tests were utilized. The exact test used for
each experiment is denoted in the figure legends and data presented as
the mean ± SEM.
Study approval
All animal studies conducted were approved by the Stony Brook
University Animal Institute Committee (#564062). Consent was waived due
to the de-identified human biopsy specimens and the study was approved
by Stony Brook University Institutional Review Board (#798611). The
National Institutes of Health Guide for the Care and Use of Laboratory
Animals was followed strictly.
Reporting summary
Further information on research design is available in the [303]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[304]Supplementary Information^ (2.5MB, pdf)
[305]Peer Review File^ (445.2KB, pdf)
[306]41467_2024_52306_MOESM3_ESM.docx^ (27.8KB, docx)
Description of additional supplementary files
[307]Supplementary data 1^ (10.8KB, xlsx)
[308]Supplementary data 2^ (17.9MB, xlsx)
[309]Supplementary data 3^ (8.5MB, xlsx)
[310]Supplementary data 4^ (169.6KB, xlsx)
[311]Supplementary data 5^ (24.7MB, xlsx)
[312]Reporting Summary^ (1.4MB, pdf)
Source data
[313]Source Data^ (409KB, xlsx)
Acknowledgements