Abstract

   AKI, due to the fact of altered oxygen supply after kidney
   transplantation, is characterized by renal ischemia–reperfusion injury
   (IRI). Recent data suggest that AKI to CKD progression may be driven by
   cellular senescence evolving from prolonged DNA damage response (DDR)
   following oxidative stress. Cellular communication factor 2 (CCN2,
   formerly called CTGF) is a major contributor to CKD development and was
   found to aggravate DNA damage and the subsequent DDR–cellular
   senescence–fibrosis sequence following renal IRI. We therefore
   investigated the impact of CCN2 inhibition on oxidative stress and DDR
   in vivo and in vitro. Four hours after reperfusion, full transcriptome
   RNA sequencing of mouse IRI kidneys revealed CCN2-dependent enrichment
   of several signaling pathways, reflecting a different immediate stress
   response to IRI. Furthermore, decreased staining for γH2AX and p-p53
   indicated reduced DNA damage and DDR in tubular epithelial cells of
   CCN2 knockout (KO) mice. Three days after IRI, DNA damage and DDR were
   still reduced in CCN2 KO, and this was associated with reduced
   oxidative stress, marked by lower lipid peroxidation, protein
   nitrosylation, and kidney expression levels of Nrf2 target genes (i.e.,
   HMOX1 and NQO1). Finally, silencing of CCN2 alleviated DDR and lipid
   peroxidation induced by anoxia-reoxygenation injury in cultured PTECs.
   Together, our observations suggest that CCN2 inhibition might mitigate
   AKI by reducing oxidative stress-induced DNA damage and the subsequent
   DDR. Thus, targeting CCN2 might help to limit post-IRI AKI.

   Keywords: acute kidney injury, ischemia–reperfusion injury, oxidative
   stress response, DNA damage response, cell cycle arrest, cellular
   communication network factor 2

1. Introduction

   Acute kidney injury (AKI) that requires renal replacement therapy is
   associated with over 50% mortality, and the severity of AKI in
   hospitalized patients correlates to mortality, length of hospital stay,
   and healthcare costs [[48]1,[49]2]. In addition, AKI predisposes to
   chronic kidney disease (CKD) development later in life, in part due to
   the excessive fibrosis associated with tubular injury [[50]3].
   Ischemia–reperfusion injury (IRI) is a leading cause of AKI, which
   occurs in 30–50% of patients receiving deceased donor kidneys and, in
   this context, often leads to delayed graft function [[51]4,[52]5].
   Understanding the immediate pathophysiological changes that underlie
   IRI in the acute phase is paramount for identifying early intervention
   options to limit subsequent AKI to CKD progression.

   During renal IRI, excessive amounts of reactive oxygen species (ROS)
   are produced [[53]6,[54]7]. ROS induce DNA damage and, consequently,
   lead to a DNA damage response (DDR) to facilitate cell repair and
   regeneration of the injured tubules through DNA repair or apoptosis
   [[55]8,[56]9]. However, a subset of cells fails to go into apoptosis,
   despite unresolved DNA damage, resulting in persistent DDR and
   senescence [[57]10]. Senescent cells alter their microenvironment
   through a proinflammatory and profibrotic secretome dubbed the
   senescence-associated secretory phenotype (SASP), thereby affecting
   organ morphology and function [[58]11]. Preventing cellular senescence
   by interfering with senescent-associated intracellular pathways has
   been proposed as a novel strategy to prevent progression to CKD
   [[59]12].

   Cellular communication network factor 2 (CCN2), previously known as
   connective tissue growth factor (CTGF), is a well-known modulator in
   CKD development [[60]13,[61]14]. CCN2 is a matricellular protein
   involved in various processes including matrix biology, angiogenesis,
   inflammation, cell proliferation, hypoxia sensing, senescence, and
   apoptosis [[62]15,[63]16,[64]17,[65]18,[66]19]. Furthermore, CCN2 is
   associated with DDR and subsequent cellular senescence in delayed graft
   function and chronic allograft dysfunction following kidney
   transplantation and in experimental IRI [[67]20], and it is also a
   potent inducer of vascular oxidative stress [[68]21].

   Differential regulation of gene expression in multiple signaling
   pathways associated with CCN2 in the immediate response, 4 h after
   experimental IRI led us to hypothesize that CCN2 can contribute to
   oxidative stress-induced DDR in the immediate phase during IRI. To
   elucidate this hypothesis, we assessed kidney function, damage, DDR,
   and oxidative stress markers in wild-type (WT) and CCN2 knockout (KO)
   mice 4 h and 3 days after IRI. In addition, we also evaluated the
   effect of CCN2 treatment in mice and of CCN2 silencing in an IRI-like
   model in cultured renal epithelial cells.

2. Materials and Methods

2.1. Animals

   Experiments were performed according to the European community
   guidelines for animal experiments and the ARRIVE guidelines and with
   consent of the Experimental Animal Ethics Committee of the University
   of Utrecht, The Netherlands [[69]22]. Generation of tamoxifen-inducible
   CCN2 full-knockout mice is extensively described elsewhere [[70]23]. In
   brief, CCN2^Flox/Flox mice were crossbred with ROSA26CreERT2 mice
   (Gt(ROSA)26Sor^tm(cre/ERT2)Tyj/J, The Jackson Laboratory, Maine, USA),
   both on a C57Bl6/J background. Male 12–14 week old mice received four
   intraperitoneal injections on alternate days over a course of 7 days
   with either 100 μL corn oil (Sigma–Aldrich, St. Louis, MO, USA) or 100
   μL of 10 mg/mL tamoxifen–corn oil solution (Sigma–Aldrich, St. Louis,
   MO, USA). After the last injection, a 14 day washout period was
   followed by the IRI operation. Sample sizes were based on published
   studies and pilot experiments. [[71]24]

2.2. Ischemia–Reperfusion Injury Model

   The surgical procedures were executed as previously described [[72]24].
   In short, renal pedicles were located through an abdominal midline
   incision and bilaterally clamped for 25 min with atraumatic
   neurovascular clamps. Clamping and subsequent reperfusion associated
   color changes were visually confirmed. IRI mice without color changes
   were excluded from further analyses. Sham operated animals underwent
   the same procedure with the exception of the pedicle clamping. The
   operations were executed under 2% isoflurane anesthesia, and body
   temperature was maintained at 37 °C. After 4 h, mice were euthanized by
   a lethal dose of a ketamine–xylazine–acepromazine injection, and the
   plasma and organs were collected and stored at −80 °C.

2.3. Mouse Recombinant CCN2 Model

   CCN2 administration studies were performed in 3 month old male C57Bl/6
   mice as previously described [[73]25]. Mice received 2.5 ng/g body
   weight recombinant human CCN2 (corresponding to the C-terminal module
   IV of CCN2; MBL International Corporation, Woburn, MA, USA) or saline
   by intraperitoneal administration.

2.4. Cell Culture

   Mouse primary renal PTECs (RPTECs, cat. no. C57-6015, Cell Biologics,
   Chicago, IL, USA) were cultured as described previously using a GasPak
   EZ Anaerobe Container System with Indicator (cat. no. 26001, BD
   Biosciences, S. Plainfield, NJ, USA) [[74]26]. Gene silencing was
   performed using a predesigned siRNA corresponding to CCN2 (siRNA ID
   MSS274358, Thermo-Fisher, Waltham, MA, USA) according to the
   manufacturer’s protocol. In short, sub-confluent cells were transfected
   in Opti-MEM reduced serum medium (Invitrogen, Waltham, MA, USA) for 24
   h with 5 ng/mL siRNA using 50 nM lipofectamine RNAi-MAX (Invitrogen,
   Waltham, MA, USA), followed by a 24 h incubation in 20% FBS medium and
   a 24 h incubation in serum-free medium. Controls were non-transfected
   cells treated with lipofectamine vehicle. Cells were subjected to 24 h
   of anoxia and 2 h of reoxygenation.

2.5. Plasma Urea

   Plasma urea was measured using the colorimetric assay conform
   manufacturer’s protocol (DiaSys, Holzheim, Germany).

2.6. Histology and Immunohistochemistry

   Renal tissue was fixed in a buffered 4% formalin solution for 24 h and,
   subsequently, embedded in paraffin blocks. Sections were mounted on
   adhesive slides (Leica Xtra) and rehydrated through a series of xylene
   and alcohol washes after which slides were rinsed in distilled water.

   For periodic acid-Schiff (PAS) staining, standard procedures were used
   (Dako, Glostrup, Denmark). Immunohistochemistry (IHC) for gamma H2AX
   (γH2AX), 4-hydroxynonenal (4-HNE), and Bcl-xL were performed as
   described elsewhere [[75]27,[76]28,[77]29]. IHC for serine 15
   phosphorylated p53 (p-p53) and nitrotyrosine was set-up based on the
   manufacturer’s protocol. First, endogenous peroxidase was blocked using
   H[2]O[2], followed by antigen retrieval by boiling in pH6 citrate
   buffer and primary antibody incubation for two hours at room
   temperature (anti-γH2AX (p Ser139), Novus Biologicals NB100-2280,
   1:500; anti-p-p53 (p Ser15), R&D Systems AF1043, 1:800; anti-Bcl-xL,
   Abcam ab178844, 1:16,000; anti-nitrotyrosine, Bioss bs-8551R, 1:500) or
   overnight at 4 °C (anti-4-HNE, Abcam ab46545, 1:150), diluted in 1% BSA
   blocking solution. For 4-HNE IHC, endogenous peroxidase was blocked
   after the primary antibody to prevent extra lipid oxidation. For p-p53,
   Bcl-xL, 4-HNE, and nitrotyrosine, secondary HRP-conjugated antibodies
   were applied and visualized using a Nova Red substrate (Vector
   Laboratories, Burlingame, CA, USA). For γH2AX,
   alkaline-phosphatase-conjugated antibody and liquid permanent red
   substrate (Dako) were used. Slides were counterstained with Mayer’s
   hematoxylin.

   Slides were scanned (NanoZoomer, Hamamatsu, Japan) and images were
   acquired by taking photographs in ImageScope. For the assessment of ATN
   severity, a blinded pathologist (TQN)-graded cortical ATN severity
   using an arbitrary score ranging from 0 (i.e., none) to 3 (i.e.,
   severe) on PAS-stained slides. ATN was scored based on loss of TECs,
   luminal debris, and mitotic activity. The score was displayed as the
   mean of the left and right kidney. Nuclear expression of γH2AX and
   p-p53 was quantitated in QuPath 0.2.3 (Edinburgh, Scotland) by counting
   the number of IHC-positive cells, relative to the number of hematoxylin
   stained nuclei in whole slides [[78]30]. Total expression of p-p53,
   4-HNE, nitrotyrosine, and Bcl-xL was quantitated in QuPath by setting a
   pixel classifier based on selected annotations corresponding to
   positive and negative areas, resulting in the percentage of positive
   area relative to total tissue area in whole slides.

2.7. RNA-Seq

   Full transcriptome RNA-sequencing was performed on a cDNA library
   constructed from RNA extracted from 18 selected samples of kidney
   cortex (4 WT sham; 4 WT IRI; 4 KO sham; 6 KO IRI). mRNA was isolated
   using NEXTflex Poly(A) Beads (Bioo Scientific, Austin, TX, USA).
   Libraries were prepared using the Rapid Directional RNA-Seq Kit
   (NEXTflex), looped in equimolarly and sequenced on Illumina NextSeq500
   to produce single-end 75 base long reads, and read-count analysis was
   performed by the Utrecht DNA Sequencing Facility (Utrecht, The
   Netherlands). RNA-seq reads were aligned to the reference genome GRCm38
   using STAR version2.4.2a (Cold Spring Harbor, NY, USA) [[79]31]. Read
   groups were added to the BAM files with Picard’s AddOrReplaceReadGroups
   v1.98 (Broad Institute, Cambridge, MA, USA). The BAM files were sorted
   with Sambamba v0.4.5, and transcript abundances were quantified with
   HTSeq-count version0.6.1p1 using the union mode [[80]32]. Subsequently,
   reads per kilobase of transcript per million mapped reads (RPKM’s) were
   calculated with edgeR’s RPKM function [[81]33]. The full data set with
   RPKM values was deposited in the GEO database (accession number:
   [82]GSE186316).

   Differentially expressed genes were identified using the DESeq2 package
   with standard settings [[83]34]. False discovery rate (FDR) adjusted
   p-values were used to determine significance [[84]35]. Pathway
   enrichment analysis was performed using Gene Set Enrichment Analysis
   (GSEA; MSigDB, UC San Diego, CA, USA) [[85]36]. GSEA analysis was
   performed on the expression data set file containing quantile
   normalized log2 transformed RPKM values, the Hallmarks and KEGG gene
   sets databases and the “Mouse_Symbol_with_Remapping_to_Human_Orthologs”
   chip platform. Gene sets with an FDR < 25% were identified as
   significantly enriched and ranked based on the normalized enrichment
   score (NES). A full list of differentially expressed genes (DEGs) and
   enriched pathways between all groups is included in [86]Supplementary
   Table.

2.8. Quantitative Real-Time PCR

   Full RNA was extracted from kidney cortical poles using TRIzol Reagent
   (Thermo-Fisher, Waltham, MA, USA). RNA purity and quantity were
   determined using the NanoDrop 2000 (Thermo-Fisher). Using 3 μg RNA per
   kidney, a cDNA library was synthesized with SuperScript III reverse
   transcriptase (Thermo-Fisher). Relative RNA expression levels were
   determined on a ViiA 7 real-time PCR system (Applied Biosystems,
   Pleasanton, CA, USA). The SYBR green primer sequences and TaQman probes
   used for quantitative real-time polymerase chain reaction (qPCR) are
   shown in [87]Table S1. TATA-box binding protein (TBP) was used as
   internal reference gene. Samples were run in duplicate. Samples free of
   mRNA and reverse transcriptase were used to control for potential
   contamination of the reaction. The ΔΔCT method was used to calculate
   relative expression levels.

2.9. Western Blot

   Cultured RPTECs were lysed and Western blot analysis was routinely
   performed as described previously [[88]26]. Experiments were repeated
   three times. Membranes were incubated with antibodies specific for the
   following proteins: 4-HNE (1:500; Abcam ab46545), nitrotyrosin (1:500;
   Bioss bs-8551R; Bcl-xL (1:500; Abcam ab178844), HMGB1 (1:500;
   Proteintech 10829-AP-1), p-p53 (1:1000, Cell Signaling Technology
   #9284), and β-actin (1:2500; Cell Signaling Technology #4967).

2.10. MDA Assay

   Lipid peroxidation was assessed in RPTECs cultured in 6-well plates. At
   the end of the reoxygenation period, malondialdehyde (MDA), the end
   product of lipid peroxidation, was measured fluorometrically in cell
   extracts with the Lipid Peroxidation (MDA) Assay Kit (cat. no.
   ab118970, Abcam, Cambridge, UK). Before MDA measurement, a Bradford
   assay was performed to adjust lysate volumes to equal protein
   concentration. These experiments were performed six times.

2.11. Statistical Analysis

   Two-way ANOVA with post hoc Tukey correction was used to compare means
   of the various parameters in all 4 groups at both time points. Discrete
   dependent variables were tested non-parametrically with the
   Kruskal–Wallis test (ATN grade). Correlation of two independent
   variables was assessed using Pearson. Values exceeding >1.5
   interquartile ranges from the mean were labeled as outliers and
   excluded. Data that showed abnormal distribution (i.e., right skewness)
   were log-transformed. Otherwise, normal distribution was assumed. The
   homogeneity of variances was tested with Levene’s test because of
   unequal sample sizes. All statistical analyses were executed using the
   statistical program SPSS (IBM SPSS Statistics 25). Error bars represent
   SEM, and p-values smaller than 0.05 were considered statistically
   significant.

3. Results

3.1. CCN2 Deletion Did Not Reduce IRI-Induced Functional Decline and Acute
Tubular Necrosis 4 h after IRI

   To evaluate CCN2-mediated changes in immediate IRI, a 4 h IRI model was
   conducted using tamoxifen-inducible CCN2 full KO mice ([89]Figure 1A).
   Tamoxifen-induced recombination resulted in 77% reduction in CCN2 mRNA
   in both sham and IRI kidneys (p = 0.005 and p = 0.008, respectively;
   [90]Figure S1A). Plasma urea increase upon IRI was similar in WT and in
   CCN2 KO mice ([91]Figure S1B). Histological examination revealed
   IRI-induced acute tubular necrosis, but no significant difference
   between WT- and CCN2 KO IRI mice ([92]Figure S1C). Concordantly,
   tubular injury markers KIM-1, NGAL, and SOX9 were equally upregulated
   in CCN2 KO IRI and WT IRI kidneys ([93]Figure S1D–F).

Figure 1.

   [94]Figure 1
   [95]Open in a new tab

   CCN2 regulated gene transcription 4 h after renal IRI. (A) Time course
   of the IRI mouse model; (B) principal component analysis plot of sham
   and IRI RNA sequence data in both WT and CCN2 KO genotypes showed
   clustering of WT IRI, CCN2 KO IRI, and sham groups but no clustering
   within the sham group; (C) pairwise comparisons of RNA sequence data of
   sham and IRI data in both WT and CCN2 KO genotypes showing the amount
   of differentially expressed genes.

3.2. CCN2 Regulated Gene Transcription of Several Signaling Pathways 4 h
after IRI

   To find CCN2-related changes in primary major regulatory events
   associated with the immediate response to IRI in the renal cortex, full
   transcriptome RNA-sequencing was performed. According to the
   literature, distinct differential gene expression patterns can be
   recognized in the kidney, as early as three hours after IRI, but the
   increase of a pro inflammatory transcriptome in those reports was
   limited [[96]37,[97]38]. In order to allow sufficient time for
   alteration of transcriptome profile with limited risk of analyzing
   secondary regulatory mechanisms, we harvested kidneys 4 h after
   reperfusion.

   Principal component analysis (PCA), as a method to identify
   similarities and differences in overall gene expression, showed that
   the data clearly segregated WT IRI, CCN2 KO IRI, and SHAM groups, but
   no evident sub-clusters formed when comparing WT SHAM with CCN2 KO SHAM
   mice ([98]Figure 1B). Pairwise comparison of transcriptome reads
   between WT IRI and CCN2 KO IRI mice yielded a total of 4167
   differentially expressed genes (DEGs) ([99]Table S2, of which 2169
   genes were significantly downregulated and 1998 upregulated in CCN2 KO
   IRI mice ([100]Figure 1C). DEGs with the largest fold down- or
   upregulation are presented in [101]Table S3. In sham mice, a total of
   21 DEGs were identified of which 15 downregulated and six upregulated
   in CCN2 KO SHAM compared to WT SHAM ([102]Figure 1C).

   Gene set enrichment analysis (GSEA) on DEGs in CCN2 KO IRI compared to
   WT IRI mice using the Hallmarks and the Kyoto Encyclopedia of Genes and
   Genomes (KEGG) databases identified significant enrichment of 16
   pathways with the WT IRI phenotype ([103]Figure S2A) and 18 pathways
   with the KO IRI phenotype ([104]Figure S2B) [[105]39]. WT IRI
   phenotype-enriched pathways included cytokine (TNFα and
   IL-6/JAK/STAT3), growth factor (TGFβ), and downstream transcription
   factor signaling (p53, PI3K/AKT/mTOR and mTORC1) as well as
   inflammatory response, hypoxia, and oxidative phosphorylation,
   suggesting overall lower activation of these pathways in CCN2 KO
   compared to WT IRI-mice.

3.3. CCN2 Deletion Reduced DNA Damage and p53 Activation 4 h after IRI

   Oxidative stress in IRI involves excessive ROS production, driving DNA
   damage and a DDR involving p53 pathway’s activation
   [[106]8,[107]40,[108]41]. Markers related to DNA damage, p53 signaling,
   and oxidative stress were analyzed by immunohistochemistry (IHC) and
   qPCR. Lower cell numbers positive for the double-strand break marker
   phosphorylated H2AX (γH2AX) were observed in CCN2 KO IRI kidneys
   compared to WT IRI kidneys, indicating reduced DNA damage (p < 0.005;
   [109]Figure 2A). Additionally, p-p53 expression, indicative for DNA
   damage-induced p53 activation [[110]42], was reduced by CCN2 KO (p =
   0.01; [111]Figure 2B). Furthermore, in CCN2 KO IRI kidneys, CCN2 mRNA
   correlated with p-p53 (r = 0.77; p = 0.04; [112]Figure 2B). Of note,
   γH2AX and p-p53 were mainly expressed in tubular epithelial cells
   (TECs). However, although we observed upregulation of p53
   transcriptional target genes GADD45A, an established immediate IRI
   response gene [[113]43,[114]44] and p21Cip1 (p21), CCN2 KO did not
   reduce the expression level of these genes on this time point
   ([115]Figure 2C,D).

Figure 2.

   [116]Figure 2
   [117]Open in a new tab

   Near total deletion of CCN2 resulted in reduced DNA damage and p53
   activation 4 h after IRI. (A,B) Representative micrographs of mouse
   renal cortex stained with gamma H2AX (γH2AX; (A) and phosphorylated p53
   (p-p53; B). Quantification of γH2AX (A) and p-p53 (B) staining showed
   that increased DNA damage and p53 activation in IRI kidneys decreased
   in CCN2 KO mice compared with WT mice. Additionally, p-p53 expression
   correlated with CCN2 mRNA in CCN2 KO kidneys (B). (C,D) Quantitative
   real-time polymerase chain reaction (qPCR) analysis showed that mRNA
   expression levels of p21CIP1 (p21) and GADD45A increased by IRI in WT
   and in CCN2 KO kidneys. Data are expressed as the mean ± SEM (n= 6 for
   WT sham; n = 6 for WT IRI; n = 4–5 for KO sham; n = 6–7 for KO IRI).
   TBP was used as an internal control. * p < 0.05, ** p < 0.01, and, ***
   p < 0.005. Bar = 100 µm.

3.4. CCN2 Deletion Reduced DDR and Apoptosis Resistance 3 Days after IRI

   To further evaluate the protective effects of CCN2 deletion, we
   performed a 3 day IRI model. In this model, we previously observed less
   DNA damage and p21 upregulation in CCN2 KO kidneys compared to WT
   kidneys 3 days after IRI [[118]20]. Furthermore, CCN2 KO IRI kidneys
   had lower numbers of p-p53 positive cells (p = 0.02; [119]Figure 3A)
   and reduced expression of GADD45A (p < 0.005; [120]Figure 3B) 3 days
   after IRI. Expression of p-p53 was predominantly nuclear and mainly
   observed in TECs ([121]Figure 3A,B). Additionally, gene and protein
   expression levels of anti-apoptotic mediator Bcl-xL (p = 0.04 and p =
   0.02; respectively; [122]Figure 4A,B) and gene expression level of
   HMGB1 (p < 0.005; [123]Figure 4C) were reduced in CCN2 KO injured mice.

Figure 3.

   [124]Figure 3
   [125]Open in a new tab

   Near total deletion of CCN2 resulted in reduced p53 activation 3 days
   after IRI. (A) Representative micrographs of mouse renal cortex stained
   with phosphorylated p53 (p-p53). Quantification of p-p53 staining
   showed that increased p53 activation in IRI kidneys, decreased in CCN2
   KO mice compared with WT mice. (B) qPCR analysis showed that increased
   mRNA expression of GADD45A in IRI kidneys reduced in CCN2 KO mice
   compared with WT mice. Data are expressed as the mean ± SEM (n = 4–5
   for WT sham; n = 7 for WT IRI; n = 4–5 for KO sham; n = 10 for KO IRI).
   TBP was used as an internal control. * p < 0.05 and *** p < 0.001. Bar
   = 100 µm.

Figure 4.

   [126]Figure 4
   [127]Open in a new tab

   Near total deletion of CCN2 expression resulted in reduced Bcl-xL and
   HMGB1 anti-apoptosis expression 3 days after IRI. (A) Representative
   micrographs of mouse renal cortex stained with Bcl-xL showed that
   increased Bcl-xL expression in IRI kidneys decreased in CCN2 KO mice
   compared with WT mice. (B,C) qPCR analysis showed that increased mRNA
   expression of Bcl-xL (B) and HMGB1 (C) in IRI kidneys reduced in CCN2
   KO mice compared with WT mice. Data are expressed as the mean ± SEM (n
   = 4–5 for WT sham; n = 7 for WT IRI; n = 5 for KO sham; n = 9–10 for KO
   IRI). TBP was used as an internal control. * p < 0.05 and *** p <
   0.001. Bar = 100 µm.

3.5. CCN2 Deletion Reduced Oxidative Stress Response 3 Days after IRI

   The role of CCN2 in oxidative stress response was evaluated by assaying
   ROS production and redox signaling. The lipid peroxidation product
   4-hydroxynonenal (4-HNE) and protein nitrosylation product
   nitrotyrosine, that mark ROS tissue levels, were not expressed 4 h
   after IRI (data not shown). IRI induced upregulation of ROS scavenging
   erythroid 2-related factor 2 (Nrf2 target) genes heme oxygenase 1
   (HMOX1) and glutamate–cysteine ligase catalytic subunit (GCLC) (p =
   0.009 and p < 0.005, respectively; [128]Figure S3A,C) and tendencies
   for increased expression after IRI were observed for (NAD(P)H quinone
   dehydrogenase 1 (NQO1) and glutamate–cysteine ligase modifier subunit
   (GCLM), but no differences in expression levels were observed between
   CCN2 KO and WT IRI kidneys ([129]Figure S3A–D).

   However, in the later model, 3 days after IRI, CCN2 KO kidneys showed
   lower 4-HNE expression ([130]Figure 5A; p < 0.005) and lower
   nitrotyrosine expression ([131]Figure 5B; p < 0.005) indicating that
   CCN2 KO reduced excessive ROS production upon IRI. Expression of these
   oxidative stress markers was mainly observed in proximal tubules of the
   inner cortex. Furthermore, reduced mRNA levels of Nrf2 target genes
   HMOX1 and NQO1 were observed in CCN2 KO compared to WT IRI kidneys (p <
   0.001 and p = 0.001; [132]Figure 5C,D).

Figure 5.

   [133]Figure 5
   [134]Open in a new tab

   Near total deletion of CCN2 expression resulted in reduced oxidative
   stress response 3 days after IRI. (A,B) Representative micrographs of
   mouse renal cortex stained with 4-hydroxynonenal (4-HNE; A) and
   nitrotyrosine (B). Quantification of 4-HNE (A) and nitrotyrosine (B)
   staining showed that an IRI-induced increase of lipid peroxidation and
   protein nitrosylation, respectively, decreased in CCN2 KO mice compared
   with WT mice. (C,D) qPCR analysis showed that increased mRNA expression
   of HMOX-1 (C) and NQO1 (D) in IRI kidneys significantly reduced in CCN2
   KO mice compared with WT mice. Data are expressed as the mean ± SEM (n
   = 4–5 for WT sham; n = 6–7 for WT IRI; n = 4–5 for KO sham; n = 8–10
   for KO IRI). TBP was used as an internal control. * p < 0.05, ** p <
   0.01, and *** p < 0.001. Bar = 100 µm.

3.6. CCN2 Induced Oxidative Stress in Mouse Kidneys

   To evaluate if CCN2 itself can induce an oxidative stress response,
   mice were injected with recombinant CCN2 and kidneys were harvested 24
   h later. Kidneys from CCN2 injected mice showed an increase in 4HNE
   expression compared to control mice. (p < 0.05; [135]Figure 6).
   Expression of 4-HNE was mainly observed in tubules ([136]Figure 6).

Figure 6.

   [137]Figure 6
   [138]Open in a new tab

   Administration of CCN2 induced oxidative stress in murine kidneys. Mice
   received intraperitoneal injections with 2.5 ng/g of CCN2 and were
   sacrificed 24 h later. Representative micrographs and quantification of
   mouse renal cortex stained with 4-hydroxynonenal (4-HNE), showing 4-HNE
   positive tubules. Data are expressed as the mean ± SEM of 4-HNE
   positive area per field (n = 5 for control and n = 7 for CCN2).

3.7. CCN2 Silencing Reduces Oxidative Stress Induced by Anoxia–Reoxygenation
Injury in PTECs

   The direct impact of CCN2 on oxidative stress in renal tubular cells
   was evaluated by silencing CCN2 in primary murine RPTECs subjected to
   anoxia–reoxygenation (AR) injury, mimicking in vivo IRI [[139]45], and
   showing a similar immediate upregulation of CCN2 [[140]20]. AR-injured
   RPTEC showed more expression of p-p53, 4-HNE, and nitrotyrosine
   compared to control cells that were cultured under normoxic conditions,
   and p-p53 and 4-HNE expression were alleviated when CCN2 was silenced
   ([141]Figure 7A). No marked differences in Bcl-xL or HMGB1 expression
   were observed (data not shown). Finally, AR-injured RPTEC showed
   increased levels of lipid peroxidation marker MDA compared to controls
   cells that were cultured under normoxic conditions, and this was
   alleviated when CCN2 was silenced (p < 0.005; [142]Figure 7B). Taken
   together, silencing of CCN2 alleviated DDR and lipid peroxidation
   induced by anoxia–reoxygenation in-jury in cultured PTECs.

Figure 7.

   [143]Figure 7
   [144]Open in a new tab

   Silencing of CCN2 alleviated oxidative stress response induced by
   AR-injury in cultured PTECs. (A) Representative Western blotting of
   protein extracts from cultured cells for 4-HNE, nitrotyrosine, and
   p-p53 showed that increased oxidative stress response and DDR in
   AR-injured PTECs was alleviated by silencing of CCN2. Data represent
   three independent experiments. (B) Malondialdehyde (MDA) levels
   indicated that increased oxidative stress response in AR-injured PTECs
   was alleviated by silencing of CCN2. Data represent six independent
   experiments. Data are expressed as the mean ± SEM. *** p < 0.001.

4. Discussion

   Our observations in the immediate early response phase following
   experimental IRI identify CCN2-dependent enrichment of several
   signaling pathways involved in the stress response to IRI. This was
   evidenced by differential gene expression at 4 h after reperfusion, and
   follow-up analyses revealed that CCN2 deletion reduces DNA damage and
   DDR already at that time point, despite similar decline in kidney
   function and acute tubular damage. Subsequently, reduced DDR,
   anti-apoptosis and oxidative stress response were observed 3 days after
   IRI. Moreover, CCN2 silencing also reduced DDR and oxidative stress in
   cultured PTECs and intraperitoneally injected recombinant CCN2 induced
   oxidative stress in murine kidneys. CCN2 has previously been shown to
   contribute to vascular ROS generation [[145]21], and to DDR and
   subsequent cellular senescence in delayed graft function and chronic
   allograft dysfunction following kidney transplantation, and in
   experimental IRI [[146]20]. These novel findings suggest that CCN2 not
   only acts as a profibrotic factor in later stages of adverse tissue
   remodeling, but that it also negatively contributes to the immediate
   early response to IRI, in particular to oxidative stress-induced DDR
   ([147]Figure 8). This could explain our previous findings that CCN2
   deletion preserves kidney function 3 days after IRI, and contribute to
   reduced cellular senescence and fibrosis at later stages [[148]20].
   This implicates that anti-CCN2 therapy, currently in phase 2 and phase
   3 trials for other indications, may be developed to reduce post-IRI
   acute and chronic kidney dysfunction by limiting IRI-induced oxidative
   stress induced DNA damage, senescent cell accumulation, and subsequent
   fibrosis.

Figure 8.

   [149]Figure 8
   [150]Open in a new tab

   Mechanisms involved in IRI linked with CCN2. Inhibition of CCN2 reduced
   the expression of γH2AX and p-p53 marking reduced DNA damage and DDR in
   the immediate early response following IRI. In the early phase upon
   IRI, this was followed by sustained reduction in DNA damage and DDR,
   along with reduced anti-apoptosis. This was associated with reduced
   oxidative stress response. Dotted arrows are used to acknowledge that
   the exact mechanisms by which CCN2 contributes to DNA damage and DDR
   remain to be established.

   This study provides novel insights into the effect of CCN2 on the
   oxidative stress response following renal IRI. Few studies have
   previously observed that CCN2 is a potent inducer of oxidative stress
   in other organs and cell types. CCN2 induces oxidative stress in murine
   aorta and in cultured vascular smooth muscle cells and endothelial
   cells [[151]21]. Furthermore, CCN2 induces accumulation of ROS in
   cultured fibroblasts [[152]46]. However, to our knowledge, this is the
   first report regarding CCN2-induced oxidative stress in the kidney, and
   the first demonstration of a protective effect of CCN2 inhibition on
   oxidative stress induced by other factors. Remarkably, CCN2
   overexpression has also been described to limit oxidative stress. For
   instance, CCN2 protected cultured cardiac myocytes from
   doxorubicin-induced oxidative stress and cell death [[153]47].
   Possibly, the complex and diverse mechanisms of action of CCN2, might
   lead to different outcomes depending on the cell type and context,
   including matrix, cytokine environment, and genotype [[154]48].

   Main mechanisms by which CCN2 might contribute to oxidative stress
   following renal IRI include increased mitochondrial oxidative
   phosphorylation and ROS generation, and decreased ROS scavenging.
   First, CCN2 may interfere with mitochondrial oxidative phosphorylation,
   which is implicated in renal IRI [[155]49]. GSEA revealed enrichment of
   the oxidative phosphorylation pathway 4 h after IRI in the direction of
   lower activation in CCN2 KO IRI- compared to WT IRI-mice. Consistently,
   in murine chondrocytes, CCN2 deletion impaired metabolism with reduced
   intracellular ATP levels [[156]50]. Upon reperfusion of the kidney,
   restored levels of oxygen stimulate mitochondrial oxidative
   phosphorylation to produce ATP, the main source of cellular energy,
   with the concurrence of harmful ROS [[157]6,[158]51]. Thus, altered ATP
   production depending on CCN2 expression level could affect the level of
   ROS and subsequent DNA damage upon renal IRI. However, a contrary
   effect of CCN2 on ROS generation has also been described in Oral
   Squamous Cell Carcinoma cells, where CCN2 overexpression impaired
   mitochondrial oxidative phosphorylation [[159]52]. A possible effect of
   CCN2 inhibition on oxidative phosphorylation and underlying mechanisms
   in the setting of renal IRI remains to be elucidated.

   Underlying pathways linking CCN2 and the oxidative stress response may
   include epidermal growth factor receptor (EGFR) activation and NAD(P)H
   oxidases (Nox)1 activity, which have previously been linked to
   CCN2-induced ROS generation [[160]21]. Crosstalk between CCN2 and ROS
   producing metabolic pathways may also involve integrin interactions.
   CCN1 induces ROS through integrin signaling [[161]53], and engagement
   of integrin α6β1 and cell surface heparan sulfate proteoglycans (HSPGs)
   [[162]46]. The mechanisms of action of CCN2 and CCN1 may well be
   similar in several respects, since they are highly homologous and bind
   the same receptors [[163]53,[164]54].

   Alternatively, reduced CCN2 might improve the handling of ROS by
   antioxidant mechanisms. However, regarding Nrf2 target genes, no
   differences in gene expression levels of GCLC and GCLM were observed,
   and reduced rather than increased expression levels of HMOX1 and NQO1
   were found in the CCN2-depleted mice.

5. Conclusions

   In aggregate, our observations indicate that DNA damage and DDR in the
   kidney following IRI are mitigated by CCN2 inhibition through reduction
   of the oxidative stress response in tubular epithelial cells. Anti-CCN2
   therapy might therefore be explored for its potential to help limiting
   post-IRI AKI and AKI to CKD progression.

Acknowledgments