Abstract

Objective

   In the fibrotic kidneys, the extent of a formed deleterious
   microenvironment is determined by cellular mechanical forces. This
   process requires metabolism for energy. However, how cellular mechanics
   and metabolism are connected remains unclear.

Methods

   A multi-disciplinary approach was employed: the fibrotic kidney disease
   models were induced by renal ischemia-reperfusion injury and unilateral
   ureteral obstruction in Calponin 2 (CNN2) knockdown mice. Proteomics,
   bioinformatics, and in vivo and in vitro molecular experimental
   pathology studies were performed.

Result

   Our proteomics revealed that actin filament binding and cell metabolism
   are the two most dysregulated events in the fibrotic kidneys. As a
   prominent actin stabilizer, CNN2 was predominantly expressed in
   fibroblasts and pericytes. In CKD patients, CNN2 levels was markedly
   induced in blood. In mice, CNN2 knockdown preserves kidney function and
   alleviates fibrosis. Global proteomics profiled that CNN2 knockdown
   enhanced the activities of the key rate-limiting enzymes and regulators
   of fatty acid oxidation (FAO) in the diseased kidneys. Inhibiting
   carnitine palmitoyltransferase 1α in the FAO pathway resulted in lipid
   accumulation and extracellular matrix deposition in the fibrotic
   kidneys, which were restored after CNN2 knockdown. Bioinformatics and
   chromatin immunoprecipitation showed that CNN2 interactor, estrogen
   receptor 2 (ESR2), binds peroxisome proliferator-activated receptor-α
   (PPARα) to transcriptionally regulate FAO downstream target genes
   expression amid kidney fibrosis. In vitro, ESR2 knockdown repressed the
   mRNA levels of PPARα and the key genes in the FAO pathway. Conversely,
   activation of PPARα reduced CNN2-induced matrix inductions.

Conclusions

   Our results suggest that balancing cell mechanics and metabolism is
   crucial to develop therapeutic strategies to halt kidney fibrosis.

   Keywords: Calponin 2, ESR2, Fatty acid oxidation, Proteomics, Chronic
   kidney disease

Highlights

     * •
       Serum CNN2 was increased in CKD patients.
     * •
       Knockdown of CNN2 alleviated kidney fibrosis in CKD models.
     * •
       Knockdown of CNN2 enhanced FAO and reduced lipid accumulation after
       CKD.
     * •
       Knockdown of CNN2 preserved kidney functions once CPT-1α was
       inhibited.
     * •
       CNN2 interacts with ESR2 to transcriptionally modulate the activity
       of PPARα.

1. Introduction

   Chronic kidney disease (CKD) is estimated to become the 5th leading
   cause of death worldwide by 2040 [[52]1,[53]2]. As the common
   consequence of CKD, kidney fibrosis refers to a heterogeneous group of
   disorders that scar kidneys, most often inevitably and
   irreversibly [[54]3]. In fibrosis, cell mechanical forces are essential
   in controlling fibroblast activation and extracellular matrix (ECM)
   stiffness, and regulating tubular cell metabolic reprogramming for
   energy [[55]4,[56]5]. Nevertheless, mechanotransduction is often
   ignored in studying CKD.

   Mechanical signals are equally important to chemical signals in
   exerting biological effects. Our proteomics data revealed that actin
   filament binding is a key event in kidney fibrosis. Serving as an actin
   stabilizer, Calponin (CNN) plays a central role in numerous fundamental
   processes, including cell proliferation, motility, and
   adhesion [[57]6,[58]7]. CNN mostly consists of α-helices with hydrogen
   bond turns. As a binding protein, CNN structure consists of three
   domains. These domains appear in order of CNN homology, regulatory
   domain, and the domain containing the CNN repeats. In the human genome,
   CNN has three isoforms: CNN1, CNN2, and CNN3 [[59]6]. In comparison to
   CNN1 and CNN3, CNN2 is expressed in more tissues and cell
   types [[60]6,[61]8]. Under pathophysiological conditions, CNN2 deletion
   attenuates inflammatory arthritis [[62]9], calcific aortic valve
   disease [[63]10], postoperative peritoneal adhesions [[64]11], and
   tumor metastasis [[65]12]. However, little is known about how CNN2
   influences kidney fibrosis.

   Cell mechanics and metabolism are intrinsically intertwined [[66]13].
   After CKD, tubular epithelial cells (TECs), the largest resident cell
   population in the kidney, can adapt to the fibrotic microenvironment by
   reprogramming their metabolism to meet the energy consumption
   requirements and improve CKD [[67]14,[68]15]. Importantly, this
   metabolic reprogramming provides energy for actin cytoskeletal
   rearrangement in response to external forces [[69]16]. TECs demand high
   energy which primarily relies on fatty acid oxidation (FAO) and
   mitochondrial oxidative phosphorylation to generate adenosine
   triphosphate (ATP) as their energy source [[70]17]. Under CKD,
   defective FAO induces intracellular lipid accumulation and ATP
   depletion. Consequently, cellular lipotoxicity and energy deprivation
   contribute to TEC dedifferentiation, leading to irreversible kidney
   fibrosis [[71][18], [72][19], [73][20]]. Fatty acid (FA) is transported
   to the kidney mainly by the cluster of differentiation 36 (CD36) or FA
   transporter 2 [[74]21,[75]22]. It is then catalyzed to fatty acyl-CoA
   by acyl-CoA synthetase (Acsm) in cells [[76]23]. Activated FA enters
   the mitochondrial matrix for β-oxidation mediated by carnitine
   palmitoyltransferase 1 (CPT1) and 2 (CPT2) [[77]24]. In mitochondria,
   key rate-limiting enzymes such as CPT1, peroxisomal acyl-CoA oxidase
   (ACOX), acyl-CoA dehydrogenase medium chain (Acadm) or long chain
   (Acadl) are required to degrade fatty acyl-CoA into the end-product,
   acetyl-CoA, by FA β-oxidation [[78]25]. Lastly, acetyl-CoA enters the
   tricarboxylic acid (TCA) cycle to generate ATP [[79]26]. Peroxisome
   proliferator-activated receptor α (PPARα) transcriptionally controls
   expression of FAO genes to balance FA uptake and oxidation and
   determine CKD outcomes [[80]15,[81]20,[82]27]. Because
   epithelial-mesenchymal communication is key for controlling CKD
   progression [[83]28], we hypothesized that CNN2 affects the FAO pathway
   and causes kidney fibrosis. Understanding cellular mechano-metabolic
   changes will create therapeutic opportunities to treat CKD.

2. Materials and methods

   Detailed methods are in the Supplementary Materials.

2.1. Mouse models of CKD

   Male Balb/c mice were obtained from the Jackson Laboratory. Mouse CKD
   models were induced by ischemia-reperfusion injury (IRI), unilateral
   ureteral obstruction (UUO), and adriamycin (ADR), as described
   previously [[84]29,[85]30]. Customized short hairpin RNA (ShRNA)
   specific for the mouse CNN2 gene was ordered from Qianlong Biotech. In
   a separate experiment, etomoxir was administrated in mice 2 days before
   IRI and nephrectomy, respectively. All animal experiments were approved
   by the Institutional Animal Care and Use Committee at the University of
   Connecticut, School of Medicine.

2.2. Human kidney biopsy specimens

   Human kidney biopsy specimens and non-tumor kidney tissue were obtained
   from the pathology archive at the University of Pittsburgh Medical
   Center. All procedures followed ethical standards and were approved by
   the Institutional Review Board at the University of Pittsburgh, School
   of Medicine. Demographic data are in [86]Supplementary Table S1.

2.3. Proteomic analysis

   Kidney tissues were processed following a shotgun proteomics workflow
   published previously [[87]31]. The LC-MS/MS analysis was performed
   using an Ultimate 3000 nanoLC and Q Exactive mass spectrometer system.

2.4. Determination of blood urea nitrogen (BUN), serum creatinine (Scr), and
alanine transaminase (ALT)

   BUN, Scr, and ALT levels were determined respectively using the
   QuantiChrom™ Urea, Creatinine, and ALT assay kits, according to the
   protocols specified by the manufacturer.

2.5. Quantitative real-time reverse transcription PCR (qRT-PCR)

   Total RNA isolation and qRT-PCR were conducted by procedures described
   previously [[88]29]. The mRNA levels of various genes were calculated
   after normalizing with β-actin or GAPDH. The primer sequences used are
   in [89]Supplementary Table S2.

2.6. Western blot analysis

   Protein expression was analyzed by western blot as described
   previously [[90]29]. Antibodies used are in [91]Supplementary Table S3.

2.7. Triglyceride and ATP measurement

   The levels of serum, kidney, and liver triglyceride or kidney ATP
   content were measured by using the Triglyceride Quantification Kit or
   the ATP Colorimetric/Fluorometric Assay Kit, according to
   manufacturers’ instructions.

2.8. Enzyme-linked immunosorbent assay (ELISA)

   The human and rat calponin 2 ELISA kits were purchased from
   MyBioSource, Inc. This assay employs the quantitative sandwich enzyme
   immunoassay technique.

2.9. Histology, and Oil Red O, immunohistochemical, and immunofluorescence
staining

   Paraffin-embedded kidney sections and cryosections were prepared by a
   routine procedure. The sections were stained with Masson Trichrome
   Staining reagents or Oil Red O staining reagents. Immunostaining was
   performed according to the established protocol as described
   previously [[92]29]. Antibodies used are in [93]Supplementary Table S3.

2.10. Cell culture, small interfering RNA (siRNA) inhibition, and treatment

   Normal rat kidney fibroblasts (NRK-49F) and human proximal tubular
   cells (HK-2) were obtained from the American Type Culture Collection.
   For conditioned media (CM) collection, NRK-49F were transfected with
   siCNN2 for 24 h and then cultured with serum-free media for 24 h.
   Serum-starved HK-2 cells were then transfected with ESR2-Dicer siRNA or
   treated with CM, CNN2 recombinant protein, etomoxir, or fenofibrate.

2.11. Chromatin immunoprecipitation (ChIP) assay

   To analyze interactions between ESR2 and the binding sites in the PPARα
   gene promoter, a ChIP assay was performed. This assay was conducted
   according to the manufacturer-specified protocols. Primer sequences
   used for the ChIP assay are in [94]Supplemental Table S2.

2.12. Statistics

   All data were expressed as mean ± SEM if not specified otherwise in the
   legends. Statistical analysis of the data was performed using GraphPad
   Prism 9. Results are presented in dot plots. P < 0.05 was considered
   statistically significant.

3. Results

3.1. Global proteome characterization after CKD

   At the genome-wide transcript level, metabolism is a key pathway in
   metabolic disease-associated CKD patients [[95]19]. To investigate
   whether metabolism is also involved in non-metabolic disease-associated
   CKD, we used quantitative proteomics to profile the kidney proteome
   landscape of controls and IRI-induced CKD mice. The principal component
   analysis (PCA) showed distinct segregation of proteins in control and
   IRI mice kidneys ([96]Figure 1A). Correlations between biological
   replicates within the same group and the distribution of protein
   intensity are presented ([97]Figure 1, B–C). In total, we identified
   6,146 proteins and 1,340 of these had significantly different
   expression (Permutation FDR 0.05) between control and ischemic kidneys
   ([98]Figure 1D). Compared to controls, 1,127 and 213 proteins were up-
   and down-regulated in the ischemic kidneys. Kyoto Encyclopedia of Genes
   and Genomes (KEGG) enrichment analysis highlighted that metabolic
   pathways are the most significantly disturbed events ([99]Figure 1E),
   consistent with the transcript-level results [[100]19].

Figure 1.

   [101]Figure 1
   [102]Open in a new tab

   The landscape of proteomes in CKD. (A) Principal component analysis of
   proteomes in control and IRI-induced CKD kidneys. (B) Correlation of
   kidney proteome profiles between control and IRI-induced CKD kidneys.
   The color scale represents R^2 values. (C) Violin plot of ANOVA
   significant proteins (Permutation FDR 0.05) among control and
   IRI-induced CKD kidneys. Label-free quantitation (LFQ) intensities of
   represented proteins were z-scored and plotted according to the color
   bar. (D) Hierarchical clustering of the intensity (plotted as z-score)
   of the proteins identified in the control and ischemic kidneys by
   quantitative proteomic analysis. Two clusters of proteins with
   different patterns of abundance profiles were highlighted in the
   heatmap. (E) KEGG pathway enrichment analyses revealed the top 20
   activated pathways in the ischemic kidneys after CKD.

3.2. Calponin 2 is a key actin filament-associated regulatory protein that
facilitates kidney fibrosis

   Besides the disturbance of metabolic pathways, Gene Ontology (GO)
   analysis of molecular function revealed that protein binding, RNA
   binding, and actin filament binding are the top 3 events that play a
   key role in facilitating kidney fibrosis ([103]Figure 2A). Considering
   kidney fibrosis is directly determined by cellular mechanical forces,
   we decided to focus our study on dissecting the role of actin filament
   binding in CKD. Earlier studies showed that cell mechanics and
   metabolism are reciprocally regulated [[104]13], we therefore screened
   the significant actin filament regulatory proteins in our proteomics
   database. To identify the most impacted cell mechanics-related proteins
   for investigation, we excluded unknown proteins and those that have
   been well studied ([105]Figure 2B). Among the remaining proteins, we
   identified CNN2 as a protein of interest ([106]Figure 2C) because it is
   predominantly expressed in the cells that are physiologically under
   high mechanical tension, have high rates of proliferation, and are
   actively migrating. To enhance the reproducibility of our proteomic
   analyses, we constructed three well-characterized CKD models induced by
   IRI, UUO, and ADR, respectively. CNN2 mRNA was significantly increased
   in all three models’ fibrotic kidneys compared to controls
   ([107]Figure 2D). Western blots demonstrated CNN2 protein was
   upregulated in the fibrotic kidneys ([108]Figure 2E and
   [109]Supplementary Figure S1, A-C). A separate single-nucleus RNA
   sequencing revealed that CNN2 was mainly expressed by kidney
   fibroblasts and pericytes after 14 days of IRI
   ([110]Figure 2F) [[111]32]. Consistently, immunohistochemical staining
   confirmed that CNN2 was predominantly localized in the fibrotic kidney
   interstitium ([112]Figure 2G). Using double immunostaining for CNN2
   (red) and fibroblast/pericyte marker platelet-derived growth factor
   receptor β (PDGFRβ, green), we confirmed that fibroblasts/pericytes are
   the major sources of CNN2 in the fibrotic kidneys after IRI
   ([113]Figure 2H).

Figure 2.

   [114]Figure 2
   [115]Open in a new tab

   Calponin 2 is a key actin filament-associated regulatory protein
   identified in the fibrotic kidneys. (A) Gene Ontology (GO) enrichment
   analysis revealed that actin filament binding is listed as one of the
   top 3 significantly upregulated events under molecular function terms.
   (B) Heatmap of the differentially expressed actin filament-associated
   regulatory proteins between controls and CKD induced by IRI. (C)
   Volcano plot showed calponin 2 (CNN2) is upregulated in the ischemic
   kidneys after CKD. (D) Quantitative RT-PCR analysis revealed that CNN2
   mRNA levels were upregulated in the fibrotic kidneys induced by IRI,
   UUO, and ADR nephropathy, respectively. Graphs are presented as
   means ± SEM. ∗P < 0.05 (n = 5–6). (E) Western blot assays demonstrated
   the expression of CNN2 protein in the fibrotic kidneys induced by IRI,
   UUO, and ADR nephropathy. Numbers indicate individual animals within
   each group. (F) Single nucleus RNA sequencing showed CNN2 mainly
   expressed by fibroblasts (Fib) and pericytes (Per) after IRI. (Data
   were extracted from the online database provided by Dr. Benjamin
   Humphrey's laboratory at the Washington University in St. Louis,
   [116]http://humphreyslab.com/SingleCell/displaycharts.php). (G)
   Immunohistochemical staining showed CNN2 distributions in the fibrotic
   kidneys after IRI, UUO, and ADR nephropathy, respectively. (H)
   Immunofluorescence staining demonstrated the co-staining of CNN2 (red)
   and platelet-derived growth factor receptor β (green) in the ischemic
   kidneys. (I) Representative immunohistochemical staining images showed
   CNN2 expression in the non-tumor normal human kidneys and the kidney
   biopsy specimens from CKD patients diagnosed with IgA nephropathy
   (IgAN), focal segmental glomerulosclerosis (FSGS), and membrane
   nephropathy (MN). Boxed areas are zoomed. (J) Serum CNN2 levels in
   healthy adults (n = 75) and non-diabetic CKD patients (n = 84). Graphs
   are presented as means ± SEM. Arrows indicate positive staining. Scale
   bar, 50 μm. IRI, ischemia reperfusion injury; UUO, unilateral ureteral
   obstruction; ADR, Adriamycin.

   To establish the clinical relevance of CNN2 expression in human CKD, we
   immunohistochemically stained kidney biopsy specimens from CKD patients
   with different etiologies for CNN2. CNN2 was induced in the
   interstitium of all CKD patients, compared with non-tumor HA controls
   ([117]Figure 2I and [118]Supplementary Figure S1D). To determine
   whether CNN2 is detectable in circulation, we further analyzed serum
   CNN2 level from 75 healthy adults and 84 non-diabetic CKD patients and
   found it was increased in CKD patients ([119]Figure 2J).

3.3. CNN2 knockdown alleviates kidney fibrosis

   To explore the role of CNN2 in kidney fibrosis, we employed a
   hydrodynamic gene delivery technique to knock down CNN2 in IRI and UUO
   models ([120]Figure 3A and [121]Supplementary Figure S2A). Compared to
   vehicles, qPCR and western blot analyses confirmed that CNN2 induction
   was reduced in the fibrotic kidneys of ShCNN2 mice ([122]Figure 3, B–C;
   [123]Supplementary Figure S2, B–C and S3A). Immunohistochemical
   staining and double staining for CNN2 (red) and PDGFRβ (green)
   consistently showed decreased CNN2 in fibroblasts/pericytes from the
   fibrotic kidneys ([124]Figure 3D and [125]Supplementary Figure S2D).
   Then, we examined whether knockdown of CNN2 could preserve kidney
   function and alleviate kidney fibrosis. After IRI, BUN and Scr levels
   were reduced ([126]Figure 3, E–F), and expression of fibrosis-related
   genes, including fibronectin (FN), α1 type I collagen (Col1α1), and α1
   type III collagen (Col3α1), were reduced in ShCNN2 kidneys compared to
   vehicles ([127]Figure 3G). Western blots assay demonstrated reductions
   of Col1α1, FN, α-smooth muscle actin (α-SMA), and PDGFRβ in ShCNN2
   kidneys ([128]Figure 3H and [129]Supplementary Figure S3B).
   Immunohistochemical staining illustrated that CNN2 knockdown repressed
   α-SMA + myofibroblast activation, Col1α1 deposition, and FN inductions
   ([130]Figure 3I and [131]Supplementary Figure S3C). Consistently,
   Masson trichrome staining (MTS) showed reduced ECM deposition in ShCNN2
   kidneys. Furthermore, qPCR analysis revealed decreased mRNA levels of
   secreted cyto-chemokines including monocyte chemoattractant protein-1
   (MCP-1), interleukin 6 (IL-6), and tumor necrosis factor-α (TNF-α) in
   ShCNN2 kidneys ([132]Figure 3J). Immunostaining showed fewer CD45+
   monocytes and CD3+ T cells in ShCNN2 kidneys ([133]Figure 3I and
   [134]Supplementary Figure S3D). Consistent with the IRI model,
   knockdown of CNN2 exhibited similar effects on reducing ECM deposition
   and inflammation in a separated UUO model ([135]Supplementary
   Figure S2, E-L).

Figure 3.

   [136]Figure 3
   [137]Open in a new tab

   Knockdown of CNN2 alleviates kidney fibrosis. (A) Experiment design.
   The first dose of ShCNN2 plasmid was administrated in mice 1 day (d)
   before unilateral IRI (UIRI) surgery. At d10 after IRI (1d before
   unilateral nephrectomy, UNX), the second dose of ShCNN2 plasmid was
   injected. The mice were sacrificed on d18. (B) Quantitative RT-PCR
   (qPCR) analysis showed the changes of CNN2 mRNA levels in the kidneys
   of ShNC and ShCNN2 mice after IRI. Graphs are presented as means ± SEM.
   ∗P < 0.05 (n = 6). (C) Western blot assay demonstrated CNN2 protein
   expression in the kidneys of ShNC and ShCNN2 mice after IRI. Numbers
   indicate individual animals within each group. (D) Immunohistochemical
   staining and immunofluorescence co-staining of CNN2 (red) and
   platelet-derived growth factor receptor β (green) showed CNN2
   expression and distribution in the kidneys of ShNC and ShCNN2 mice
   after IRI. Scale bar, 50 μm. (E, F) Blood urea nitrogen (BUN) and serum
   creatinine (Scr) levels in ShNC and ShCNN2 mice after IRI. Graphs are
   presented as means ± SEM. ∗P < 0.05 (n = 6–8). (G) qPCR analyses
   revealed the mRNA abundance of FN, α1 type I collagen (Col1α1), and α1
   type III collagen (Col3α1) in the kidneys of ShNC and ShCNN2 mice after
   IRI. Graphs are presented as means ± SEM. ∗P < 0.05 (n = 6). (H)
   Western blot assay demonstrated Col1α1, FN, α-smooth muscle actin
   (α-SMA), and platelet-derived growth factor receptor β (PDGFRβ) protein
   expression in the kidneys of ShNC and ShCNN2 mice after IRI. Numbers
   indicate individual animals within each group. Graphs are presented as
   means ± SEM. ∗P < 0.05 (n = 6). (I) Representative micrographs for
   α-SMA, Col1α1, FN staining, Masson Trichrome Staining, CD45, and CD3 in
   the fibrotic kidneys of ShNC and ShCNN2 mice after IRI-induced CKD.
   Arrows indicate positive staining. Scale bar, 50 μm. (J) qPCR analyses
   revealed the mRNA abundance of monocyte chemoattractant protein-1
   (MCP1), interleukin 6 (IL-6), and tumor necrosis factor α (TNF-α) in
   the kidneys of ShNC and ShCNN2 mice after IRI. Graphs are presented as
   means ± SEM. ∗P < 0.05 (n = 6).

3.4. Global proteomics reveals FAO is a key pathway after CNN2 knockdown

   To better understand the molecular mechanisms of how knockdown of CNN2
   alleviates CKD, we employed a label-free quantitative approach to
   profile the proteome landscape of ShNC and ShCNN2 kidneys
   ([138]Figure 4A). PCA classified ShNC and ShCNN2 kidneys according to
   their genotype ([139]Figure 4B). A t-test identified 2419
   differentially expressed proteins (Permutation FDR 0.05) between ShNC
   and ShCNN2 mice. Compared to ShNC kidneys, 906 and 1513 proteins were
   up- and down-regulated in ShCNN2 kidneys ([140]Figure 4, C-D and
   [141]Supplementary Table S4). The proteomic measurements were
   reproducible as the protein intensity distribution was similar and the
   Pearson correlation of the six biological replicates of each group was
   0.95 or higher ([142]Supplementary Figure S4, A-B). GO and KEGG
   enrichment analysis revealed that the metabolic pathway was
   significantly upregulated in ShCNN2 fibrotic kidneys and FA metabolic
   process was the most changed event ([143]Figure 4, E–G). KEGG also
   identified the downregulated pathways and biological processes
   ([144]Supplementary Figure S4, C-D). We further investigated the
   dysregulated metabolic pathways. Among the FA, carbohydrate, and amino
   acid metabolisms, peroxisome, propanoate, and branched-chain amino acid
   metabolisms were significantly enriched ([145]Figure 4F). Impressively,
   about 40 components in the FAO pathway, including Acsm, Acox, and CPT,
   were simultaneously upregulated in ShCNN2 kidneys compared to ShNC
   ([146]Figure 4H).

Figure 4.

   [147]Figure 4
   [148]Open in a new tab

   Global proteomics reveals fatty acid oxidation is a key pathway in
   mediating kidney fibrosis after knockdown of CNN2. (A) Experimental
   workflow of the global proteomic analysis. In each group, kidney
   samples from six mice were used for mass spectrometry. (B) Principal
   component analysis of global proteomes from ShNC and ShCNN2 fibrotic
   kidneys after IRI. (C) Heatmap of t-test significant proteins
   (Permutation FDR 0.05). Two clusters of proteins with different
   patterns of abundance profiles are highlighted in the heatmap. (D)
   Volcano plot showed the differential proteins (Permutation FDR 0.05)
   between ShNC and ShCNN2 fibrotic kidneys. Up- and down-regulated
   proteins (fold-change, FC) are colored in red and blue, respectively.
   (E, F) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway
   enrichment analysis highlighted upregulated metabolic pathways in
   fibrotic kidneys from ShNC and ShCNN2 mice after IRI. (G) Gene Ontology
   (GO) biological process terms in each cluster of proteins are plotted
   with their names and significance. The fatty acid metabolic process is
   highlighted to indicate the group with the largest difference in
   upregulated proteins. (H) Heatmap of the differentially expressed
   proteins in the fatty acid oxidation pathway in ShNC and ShCNN2 mice
   fibrotic kidneys after IRI.

3.5. CNN2 modifies lipid accumulation and FAO pathway activity in fibrotic
kidneys

   We then validated FAO regulation in ShNC and ShCNN2 kidneys. First, we
   examined the extent of lipid accumulation in fibrotic kidneys.
   Triglyceride content in sera or kidneys was decreased in ShCNN2 mice
   compared to ShNC after IRI ([149]Figure 5, A–B). Oil red O staining
   consistently showed less triglyceride in ShCNN2 kidney tubules
   ([150]Figure 5C). To enhance the evidence of lipid accumulation in the
   fibrotic kidneys, we employed an additional marker for lipid droplets,
   perilipin 2 (Plin2) [[151]33,[152]34]. Western blots demonstrated that
   Plin2 is dramatically decreased in ShCNN2 kidneys compared to ShNC
   ([153]Figure 5D and [154]Supplementary Figure S5A). Immunohistochemical
   staining revealed that Plin2 was predominantly distributed in ShNC
   kidney tubules ([155]Figure 5E). Interestingly, although the liver is
   the central organ that controls lipid homeostasis, we did not observe
   significant differences of liver function changes and lipid
   accumulation in liver between ShNC and ShCNN2 mice after systemically
   knockdown CNN2 in IRI-induced CKD ([156]Supplementary Figure S6, A-B).

Figure 5.

   [157]Figure 5
   [158]Open in a new tab

   CNN2 modifies lipid accumulation and fatty acid oxidation pathway amid
   kidney fibrosis. (A, B) The enzyme-linked immunosorbent assay (ELISA)
   showed triglyceride contents in serum (A) and kidney (B) collected from
   ShNC and ShCNN2 mice after IRI. Graphs are presented as means ± SEM.
   ∗P < 0.05 (n = 6). (C) Representative micrographs of Oil Red-O staining
   in the fibrotic kidneys from ShNC and ShCNN2 mice after IRI. Arrows
   indicate lipid accumulation in tubules. Scale bar, 50 μm. (D) Western
   blot assay demonstrated perilipin 2 (Plin2) protein expression in ShNC
   and ShCNN2 mice fibrotic kidneys after IRI. Numbers indicate individual
   animals within each group. (E) Immunohistochemical staining showed
   Plin2 distributions in the fibrotic kidneys of ShNC and ShCNN2 mice
   after IRI. Arrows indicate positive staining. Scale bar, 50 μm. (F)
   ELISA showed ATP levels in the total kidneys collected from ShNC and
   ShCNN2 mice after IRI. Graphs are presented as means ± SEM. ∗P < 0.05
   (n = 6). (G) Quantitative RT-PCR analyses revealed the mRNA abundance
   of PPARα, CPT1α, ACOX1-3, Acsm1, Acsm2, Acsm5, Acadm, and Acadl in
   fibrotic kidneys from ShNC and ShCNN2 mice after IRI. Graphs are
   presented as means ± SEM. ∗P < 0.05 (n = 6). (H) Western blot assays
   demonstrated PPARα, PGC1α, CPT1α, ACOX1, Acsm5, and Acadm protein
   expression in ShNC and ShCNN2 mice fibrotic kidneys after IRI. Numbers
   indicate individual animals within each group. (I) Representative
   micrographs of PPARα, CPT1α, ACOX1, and Acadm staining in the fibrotic
   kidneys from ShNC and ShCNN2 mice after IRI. Boxed areas are enlarged.
   Arrows indicate positive staining. Scale bar, 50 μm.

   Because FAO is the key contributor to intracellular ATP levels in
   tubular cells, we detected ATP consumption in the fibrotic kidneys.
   ELISA showed the ATP levels were higher in ShCNN2 kidneys than ShNC
   ([159]Figure 5F). Then, we further examined expression of key elements
   in the FAO pathway in ShNC and ShCNN2 kidneys. qPCR analyses showed
   that a key transcriptional regulator of FAO, PPARα, was upregulated in
   ShCNN2 kidneys compared to ShNC ([160]Figure 5G). Meanwhile, the key
   rate-limiting enzymes of FAO, including CPT1α, Acox1-3, Acsm1, Acsm2,
   Acsm5, Acadm, and Acadl were also increased in ShCNN2 kidneys. Western
   blots confirmed upregulation of PPARα and PPARγ coactivator-1α (PGC-1α)
   in ShCNN2 kidneys. The downstream targets, including CPT1α, Acox1,
   Acsm5, and Acadm were accordingly increased ([161]Figure 5H and
   [162]Supplementary Figure S5B). Immunohistochemical staining
   consistently showed upregulation of PPARα, CPT1α, Acox1, and Acadm in
   ShCNN2 kidney tubules ([163]Figure 5I). However, after CKD, the FAO
   pathway activity in liver has minor changes between ShNC and ShCNN2
   mice ([164]Supplementary Figure S6, C-D). Additionally, knockdown of
   CNN2 also reduced lipid accumulation and FAO defect in the UUO model
   ([165]Supplementary Figure S7).

3.6. CNN2 enhances CPT1α to preserve kidney function after CKD

   To further elucidate how FAO impacts CKD, we applied etomoxir to
   specifically inhibit CPT1α, the key rate-limiting enzyme in the FAO
   pathway ([166]Figure 6A). At mRNA and protein levels, etomoxir
   repressed CPT-1α expression, but knockdown of CNN2 restored its
   induction ([167]Figure 6, B–C and [168]Supplementary Figure S8A). As
   predicted, etomoxir increased BUN and Scr levels after IRI. However,
   after etomoxir treatment, knockdown of CNN2 preserved kidney function
   ([169]Figure 6D–E). Oil Red O and immunohistochemical staining for
   Plin2 showed reduced lipid droplet accumulation in ShCNN2 kidney
   tubules compared to ShNC after etomoxir treatment ([170]Figure 6F).
   Then, we further assessed the extent of kidney fibrosis. qPCR analysis
   revealed etomoxir increased the mRNA levels of FN, Col1α1, and Col3α1
   after IRI, but knockdown of CNN2 repressed their abundances
   ([171]Figure 6G). Western blots demonstrated that etoxomir induced FN
   and α-SMA but knockdown of CNN2 suppressed their expression
   ([172]Figure 6H and [173]Supplementary Figure S8B). Immunostaining and
   MTS showed consistent results on myofibroblast activation and ECM
   deposition ([174]Figure 6I and [175]Supplementary Figure 8C).
   Additionally, knockdown of CNN2 reduced cyto-chemokine secretion,
   including MCP-1, IL-6, and TNF-α, and infiltration of CD45+ monocytes
   in IRI-induced CKD after etomoxir treatment ([176]Supplementary
   Figure S8, D-F).

Figure 6.

   [177]Figure 6
   [178]Open in a new tab

   Knockdown of CNN2 enhances CPT1α activity to alleviate kidney fibrosis
   (A) Experiment design. CPT1α inhibitor, Etomoxir (Eto), was
   administrated in mice 2 days before unilateral IRI (UIRI) and
   nephrectomy (UNX), respectively. ShNC and ShCNN2 plasmids were injected
   at day (d) −1 and 10 through the tail vein. The mice were sacrificed on
   d18. (B) Quantitative RT-PCR (qPCR) analysis showed the changes of
   CPT1α mRNA levels in the fibrotic kidneys of ShNC mice, ShNC mice that
   received Etomoxir, and ShCNN2 mice that received Etomoxir after IRI.
   Graphs are presented as means ± SEM. ∗P < 0.05 (n = 5). (C) Western
   blot assay demonstrated CPT1α protein expression in the fibrotic
   kidneys of ShNC mice, ShNC mice that received Etomoxir, and ShCNN2 mice
   that received Etomoxir after IRI. Numbers indicate individual animals
   within each group. (D, E) Blood urea nitrogen (BUN) and serum
   creatinine (Scr) levels in ShNC mice, ShNC mice that received Etomoxir,
   and ShCNN2 that received Etomoxir after IRI. Graphs are presented as
   means ± SEM. ∗P < 0.05 (n = 5). (F) Representative micrographs for Oil
   Red O Staining and perilipin 2 (Plin2) immunohistochemical staining
   showed lipid accumulations in fibrotic kidneys of ShNC mice, ShNC mice
   that received Etomoxir, and ShCNN2 mice that received Etomoxir after
   IRI. Arrows indicate positive staining. Scale bar, 50 μm. (G) qPCR
   analyses revealed the mRNA abundance of FN, α1 type I collagen, and α1
   type III collagen in the fibrotic kidneys of ShNC mice, ShNC mice that
   received Etomoxir, and ShCNN2 mice that received Etomoxir after IRI.
   Graphs are presented as means ± SEM. ∗P < 0.05 (n = 5). (H) Western
   blot assay demonstrated FN and α-SMA proteins expression in the
   fibrotic kidneys of ShNC mice, ShNC mice that received Etomoxir, and
   ShCNN2 mice that received Etomoxir after IRI. Numbers indicate
   individual animals within each group. (I) Representative micrographs
   for immunostaining of α-SMA, Col1α1, fibronectin (FN), and Masson
   Trichrome Staining (MTS) showed myofibroblast activation and collagen
   deposition in the fibrotic kidneys of ShNC mice, ShNC mice that
   received Etomoxir, and ShCNN2 mice that received Etomoxir after IRI.
   Arrows indicate positive staining. Scale bar, 50 μm.

3.7. CNN2 knockdown promotes ESR2 binding to PPARα to enhance transcription
of FAO genes

   Given that CNN2 knockdown affected multiple FAO genes and its regulator
   PPARα expression ([179]Figure 5), we hypothesized that there is a
   common mechanism that controls their activation, so we performed
   bioinformatics analyses. Based on GeneMANIA database, the human
   protein–protein network analysis indicated that estrogen receptor 2
   (ESR2) is a strong interactor of CNN2 ([180]Figure 7A). qPCR analyses
   showed that ESR2 mRNA were increased in ShCNN2 kidneys compared to ShNC
   after IRI and UUO ([181]Figure 7B and [182]Supplementary Figure S9A).
   Western blots demonstrated increased ESR2 induction in ShCNN2 kidneys
   ([183]Figure 7C and [184]Supplementary Figure S9B). Interestingly,
   under TGF-β stress in vitro, ELISA showed that CNN2 levels in
   fibroblast (NRK-49F) conditioned medium (CM) were reduced after CNN2
   knockdown ([185]Figure 7D and [186]Supplementary Figure S9C). We then
   employed CNN2-deprived CM and CNN2 recombinant protein to treat human
   kidney proximal tubular cells (HK-2) ([187]Figure 7E). qPCR analysis
   revealed that ESR2 mRNA was induced by CNN2-deprived CM but was
   repressed by CNN2 recombinant protein after TGF-β stimulation
   ([188]Figure 7F). Western blot assays confirmed a similar trend of ESR2
   expression at the protein level ([189]Figure 7G). Because ESR2 is a
   member of the superfamily of nuclear receptor transcription factors, to
   examine whether ESR2 directly regulates PPARα, we treated HK-2 cells
   with an ESR2 agonist ([190]LY500307) and then performed a chromatin
   immunoprecipitation (ChIP) assay using an anti-ESR2 antibody. Through
   the JASPAR database, we predicted there are at least 12 putative
   binding sites for ESR2 on the PPARα gene promoter region
   ([191]Supplementary Figure S9D). The co-precipitated chromatin DNA
   fragments extracted from the cells were amplified by qPCR using two
   pairs of primers for the PPARα promoter region. We found that the ESR2
   agonist promoted ESR2 binding to the putative sites on the PPARα gene
   promoter ([192]Figure 7H). In vitro, under TGF-β stress, knockdown of
   ESR2 in HK-2 cells repressed PPARα at both mRNA and protein levels
   ([193]Figure 7, I–J and [194]Supplementary Figure S9E), but PPARα
   levels were increased in HK-2 cells after incubation with CNN2-deprived
   CM ([195]Figure 7, K and M). Conversely, CNN2 recombinant protein
   repressed PPARα expression ([196]Figure 7, L and N). However, knockdown
   of ESR2 abolished PPARα induction upon incubation with CNN2-deprived CM
   in HK2 cells ([197]Figure 7O).

Figure 7.

   [198]Figure 7
   [199]Open in a new tab

   Knockdown of CNN2 promotes ESR2 binding PPARα to enhance fatty acid
   oxidation (A) Based on the GeneMANIA database, the human
   protein–protein interaction analysis showed CNN2 interacts with
   estrogen receptor 2 (ESR2). (B) Quantitative RT-PCR (qPCR) analysis
   showed the abundance of ESR2 mRNA levels in ShNC and ShCNN2 mice
   fibrotic kidneys after IRI. Graphs are presented as means ± SEM.
   ∗P < 0.05 (n = 6). (C) Western blot assay demonstrated ESR2 protein
   expression in the fibrotic kidneys of ShNC and ShCNN2 mice after IRI.
   Numbers indicate individual animals within each group. (D) ELISA showed
   CNN2 levels in the conditioned medium (CM) collected from the cultured
   fibroblasts after knockdown of CNN2 under TGF-β stress. Graphs are
   presented as means ± SEM. ∗P < 0.05 (n = 6). (E) Schematic diagram
   showed CNN2-deprived CM from the cultured fibroblasts or recombinant
   CNN2 protein were employed to treat human kidney proximal tubular cells
   (HK-2). (F) qPCR showed ESR2 expression in HK-2 cells after incubation
   with CNN2-deprived CM or human recombinant CNN2 protein. Graphs are
   presented as means ± SEM. ∗P < 0.05 (n = 3). (G) Western blot assay
   demonstrated ESR2 expression in HK-2 cells after incubation with
   CNN2-deprived CM or human recombinant CNN2 protein. (H) ChIP-qPCR assay
   showed ESR2 binds to putative sequences in the promoter of PPARα genes.
   HK-2 cells were treated with ESR2 agonist ([200]LY500307) for 3 h.
   Chromatin preparations from the cells were immunoprecipitated using an
   anti-ESR2 antibody, and co-precipitated DNA fragments were amplified
   using primers specific for the PPARα promoter. Graphs are presented as
   means ± SEM. ∗P < 0.05 (n = 3). The upper panel showed representative
   DNA sequence logo representing the binding motif of the ESR2 gene. (I)
   qPCR analyses revealed the mRNA abundances of PPARα were reduced in
   HK-2 cells after knockdown of ESR2 under TGFβ1 stress, compared with
   scramble controls. Graphs are presented as means ± SEM. ∗P < 0.05
   (n = 3). (J) Western blot assay demonstrated that knockdown of ESR2
   repressed PPARα expression in HK-2 cells under TGFβ1 stress, compared
   with scramble controls. (K, L) qPCR analyses revealed the mRNA
   abundances of PPARα were increased in HK-2 cells by CNN2-deprived CM
   (K) but were repressed by CNN2 recombinant protein (L) under TGFβ1
   stress, compared with controls. Graphs are presented as means ± SEM.
   ∗P < 0.05 (n = 3). (M, N) Western blot assays demonstrated that
   CNN2-deprived CM enhanced PPARα expression (M) but CNN2 recombinant
   protein repressed PPARα inductions (N) in HK-2 cells under TGFβ1
   stress, compared with controls. (O) qPCR analyses revealed the mRNA
   abundances of PPARα were induced by CNN2-deprived CM but were repressed
   after knockdown of ESR2 under TGFβ1 stress. Graphs are presented as
   means ± SEM. ∗P < 0.05 (n = 3).

3.8. CNN2 knockdown in fibroblasts promotes FAO in tubular cells

   To study how CNN2 regulates FAO in tubular cells, we incubated HK-2
   with CNN2-deprived CM. qPCR analysis revealed that the key
   rate-limiting enzymes including CPT1α, Acox1, Acsm1, and Acsm5 mRNA
   levels were induced by CNN2-deprived CM under TGF-β stress
   ([201]Figure 8A). Consequently, CNN2-deprived CM reduced FN induction
   and inhibited α-SMA expression ([202]Figure 8B). However, knockdown of
   ESR2 largely abolished induction of CPT1α, ACOX1, and Acsm1 and
   increased FN and α-SMA ([203]Figure 8, C–D). To confirm that CNN2
   mediates kidney fibrosis by affecting the FAO pathway in tubules, HK-2
   cells were incubated with CNN2-deprived CM followed by etomoxir under
   TGF-β stress. Western blot assays demonstrated that etomoxir abolished
   the beneficial effects of CNN2-deprived CM on reducing induction of FN
   and α-SMA ([204]Figure 8E). In comparison, HK-2 cells treated with
   human recombinant CNN2 protein repressed mRNA levels of PGC1α, CPT1α,
   Acox1, and Acsm5 after TGF-β stimulation ([205]Figure 8F). Similarly,
   CNN2 recombinant protein increased FN and α-SMA induction
   ([206]Figure 8G). In a separate experiment, HK-2 cells were treated
   with CNN2 recombinant protein followed by the PPARα agonist,
   fenofibrate. After stimulation with TGF-β, western blots demonstrated
   that fenofibrate ameliorated CNN2-induced FN and α-SMA in vitro
   ([207]Figure 8H).

Figure 8.

   [208]Figure 8
   [209]Open in a new tab

   Knockdown of CNN2 alleviates defective fatty acid oxidation and ECM
   deposition in tubular cells. (A) qPCR analyses revealed the mRNA
   abundances of CPT1α, ACOX1, Acsm1, and Acsm5 were increased in HK-2
   cells incubated with CNN2-deprived CM under TGFβ1 stress, compared with
   controls. Graphs are presented as means ± SEM. ∗P < 0.05 (n = 3). (B)
   Western blot assay demonstrated that CNN2-deprived CM reduced the
   inductions of FN and α-SMA in HK-2 cells under TGFβ1 stress, compared
   with controls. (C) qPCR analyses revealed the mRNA abundances of CPT1α,
   ACOX1, and Acsm1 were increased by CNN2-deprived CM in HK-2 cells but
   were repressed after knockdown of ESR2 under TGFβ1 stress, compared
   with controls. Graphs are presented as means ± SEM. ∗P < 0.05 (n = 3).
   (D) Western blot assays demonstrated CNN2-deprived CM alleviated FN and
   α-SMA inductions in HK-2 cells under TGFβ1 stress but they were
   increased after knockdown of ESR2, compared with controls. (E) Under
   TGFβ1 stress, HK-2 cells were incubated with CNN2-deprived CM and
   followed by CPT1α inhibitor Etomoxir (100 μM). Western blot assays
   showed Etomoxir induced FN and α-SMA expression after incubation with
   CNN2-deprived CM. (F) qPCR analyses revealed the mRNA abundances of
   PGC1α, CPT1α, ACOX1, and Acsm5 were reduced in HK-2 cells incubated
   with CNN2 recombinant protein under TGFβ1 stress. Graphs are presented
   as means ± SEM. ∗P < 0.05 (n = 3). (G) Western blot assay demonstrated
   that CNN2 recombinant induced FN and α-SMA expression in HK-2 cells
   after TGFβ1 stimulations. (H) Under TGFβ1 stress, HK-2 cells were
   treated with CNN2 recombinant proteins followed by PPARα agonist
   Fenofibrate (Feno, 80 μM). Western blot assays showed Fenofibrate
   decreased FN and α-SMA expression after treated with CNN2 recombinant
   protein. (I) Schematic diagram depicts knockdown of CNN2 promotes ESR2
   binding to PPARα to transcriptionally regulate the FAO genes to halt
   kidney fibrosis.

4. Discussion

   This study elucidated CKD pathogenesis from the perspective of cellular
   mechanics and metabolism interactions ([210]Figure 8I). We demonstrated
   that CNN2, a prominent actin stabilizer, determines kidney fibrosis
   through modulating FA metabolism as evident by: 1) knockdown of CNN2
   alleviated kidney fibrosis in CKD models; 2) if the FAO pathway was
   activated, then less lipid accumulated in ShCNN2 fibrotic kidneys; 3)
   knockdown of CNN2 preserved kidney functions after inhibiting the key
   rate-limiting enzyme, CPT-1α; 4) CNN2's interactor, ESR2,
   transcriptionally mediates the regulator of the FAO pathway, PPARα; and
   5) serum CNN2 is increased in CKD patients. These results provide
   rationale for expanding the paradigm of mechano-metabolic programming
   after CKD.

   In mice and human fibrotic kidneys, CNN2 is mainly expressed in
   interstitial fibroblasts or pericytes ([211]Figure 2). CNN2 possesses
   the capacity of generating significant forces when
   fibroblasts/pericytes communicate with the neighboring cells due to its
   nature as an actin filament regulatory protein. It directly determines
   the extent of kidney fibrosis by forming a deleterious kidney local
   microenvironment (KLM). Due to mechanical changes in a fibrotic kidney,
   injured kidney cells receive information from the microenvironment to
   coordinate their behavior at local and systemic levels. Our results
   indicate that CNN2 affects both levels in CKD. At the local level, it
   serves as an actin stabilizer. Under physiological conditions, CNN2
   controls cell shape, polarity, migration, and establishes intercellular
   contacts to support organ architecture [[212]5]. After CKD, fibrotic
   niche-like KLM formation is a critical structural change in
   kidneys [[213]29]. The densities, sizes, and cellular/non-cellular
   components of niches determine the severity of kidney fibrosis. Amid
   these processes, all kidney cells experience pathological external
   forces. These highly varied forces may mediate individual cell
   phenotypes and kidney architecture changes [[214]4]. Our data show that
   knockdown of CNN2 alleviates kidney fibrosis after ischemic and
   obstructive CKD ([215]Figure 3 and [216]Supplementary Figure S2). From
   the aspect of mechanical changes, dysregulated CNN2 reduces the injured
   cells' ability to quickly reinforce their mechanosensitive structures
   and cytoskeleton so they do not respond to forces appropriately. After
   CKD, actin cytoskeleton disarrangement causes aberrant activation of
   fibroblasts, excessive ECM deposition, and tubular atrophy which lead
   to the irreversible change of KLM. At the systemic level, CNN2 is able
   to mediate mechanical and chemical signal interactions. It is reported
   that rapid CNN2 degradation is required for cytokines to inhibit
   malignant tumor growth [[217]35]. A recent study also indicated that
   CNN2 could be ubiquitylated and removed for efficient lysophagy during
   lysosomal membrane permeabilization [[218]36]. As well, CNN2 was
   increased in CKD patients’ serum ([219]Figure 2). Therefore, CNN2 may
   also enter the KLM via the bloodstream, regulating cell behavior to
   maintain the forces in the kidney fibrotic regions. In addition, CNN2
   expression is associated with LDL-C levels in stage 2 CKD patients
   (Data not shown). This phenomenon leaves us questioning, does CNN2
   affect lipid or FA metabolism reprogramming in fibrotic kidneys?

   Metabolic reprogramming is a hallmark of malignant
   transformation [[220]37]. This process includes activating signal
   transduction cascades, key metabolic enzymes, and transcriptional
   regulation of genes encoding metabolic enzymes [[221]38]. Although the
   interaction between the actin cytoskeleton and metabolic enzymes was
   observed decades ago [[222]39], most studies investigating cellular
   mechanics remain focused on supporting organ architecture rather than
   metabolic reprogramming. As a highly metabolically active
   organ [[223]40], kidney metabolism is markedly disturbed after CKD
   ([224]Figure 1). Our proteomics data suggest that all metabolic
   pathways, including FA, carbohydrates, and amino acids, and about 40
   FAO-related enzymes were upregulated in ShCNN2 kidneys ([225]Figure 4).
   Specifically, gene expression of the key regulators of FAO (PPARα) and
   rate-limiting enzymes (Acsm1, 2, 5, CPT1α, Acox1-3, Acadm, Acadl, and
   Acad10) were upregulated in ShCNN2 kidney TECs. Although CNN2 is
   predominantly expressed in fibroblasts/pericytes, communication between
   TECs and fibroblasts in the KLM is critical in determining the severity
   of CKD [[226]41]. As the largest cell population of kidneys, TECs are
   not bystanders in kidney fibrosis [[227]29]. TECs are the main energy
   consuming cells, and FAO is their chief ATP source in kidneys. Lower
   FAO in TECs, followed by intracellular lipid accumulation, profoundly
   switches TECs into a profibrotic phenotype and leads to kidney
   fibrosis [[228]42,[229]43]. Indeed, intracellular triglyceride levels
   were reduced in ShCNN2 kidneys after CKD ([230]Figure 6 and
   [231]Supplementary Figure S7). Most importantly, after systemic
   knockdown of CNN2 in vivo, liver triglyceride and FAO genes have minor
   changes between ShNC and ShCNN2 mice after IRI-induced CKD although the
   liver is the central organ in maintaining lipid homeostasis
   ([232]Supplementary Figure S6). Therefore, we speculated that CNN2
   independently regulates intra-renal lipid accumulation and the FAO
   pathway activation by mediating cellular mechanical and chemical
   signals. CNN2 knockdown rescued CPT1α inhibitor-induced kidney fibrosis
   by stimulating FA import into the mitochondria and initiating
   mitochondrial fission. Enhanced FAO provides more ATP to maintain the
   structural integrity of actin filaments.

   The most novel finding of this study is that ESR2 is a bridge that
   connects CNN2 and PPARα to modulate FAO after CKD ([233]Figure 7),
   which we determined via several lines of evidence. First, ESR2
   interacts with CNN2. Second, as a nuclear transcription factor, ESR2
   harbors similar conserved sequences with PPARα in its promoter regions,
   and ESR2 agonist promotes their binding. Third, CNN2 knockdown
   alleviated kidney fibrosis, whereas knockdown of both CNN2 and ESR2
   abolished these protective effects and repressed FAO. Consistent with
   the protective role of ESR2 in cancer, neuropathies, cardiovascular
   disease, osteoporosis, menopause, and metabolic disorders [[234]44],
   the finding that CNN2 controls ESR2 to influence FAO and kidney
   fibrosis provides new insight into how mechanical signals contribute to
   CKD development; although, FAO is positioned at the intersection of
   multiple chemical signals such as TGF-β, Krüppel-like factor 15, and
   CD36 [[235][45], [236][46], [237][47]]. Of note, in our in vitro
   system, whether the employed conditioned medium contains other soluble
   or transferrable factors such as a PPARα activating factor to mediate
   the FAO pathway needs to be determined.

   Despite the robust evidence that demonstrates how the actin stabilizer
   CNN2 regulates metabolism, several unanswered questions remain. First,
   could CNN2 in T-cells be equally important to its role in
   fibroblasts/pericytes during kidney disease progression after IRI? Our
   deconvolution analysis in [238]Figure 2F indicates it may be important;
   however, the exact role of CNN2 expression in immune cells and how this
   mediates kidney fibrosis remains unclear. Second, does altered glucose
   metabolism contribute to kidney fibrosis in ShCNN2 kidneys, given that
   kidneys are important for glucose reabsorption, production, and
   utilization? Third, as we did not generate CNN2 genetic mouse models,
   would genetic knockout of CNN2 have the same effects in vivo?

   In summary, our investigation of cellular mechanics and metabolism
   interactions identified that the knockdown of actin stabilizer, CNN2,
   alleviated kidney fibrosis by activating FA metabolism. The new
   insights on modulating cellular mechanical properties to tune energy
   production in response to formation of the kidney fibrotic
   microenvironment supplement our understanding of CKD pathogenesis. The
   possibility of employing approaches to target mechano-metabolic
   mechanisms may shed light on treating CKD.

Author contributions

   Conceptualization, Y.G. and D.Z.; methodology, Y.G., J.T., S.B., H.F.,
   and D.Z.; investigation, Y.G., Y.W., Z.P., J.W., P.G., Y.Q., and Y.Y.;
   writing – original draft, Y.G. and D.Z.; writing – review& editing,
   Y.G., D.K., S.L., Y.W., and D.Z.; funding acquisition, D.Z.; resources
   and supervision, D.Z.

Acknowledgments