Loss of activator of G-protein signaling 3 impairs renal tubular regeneration following acute kidney injury in rodents

Abstract

The intracellular mechanisms underlying renal tubular epithelial cell proliferation and tubular repair following ischemia-reperfusion injury (IRI) remain poorly understood. In this report, we demonstrate that activator of G-protein signaling 3 (AGS3), an unconventional receptor-independent regulator of heterotrimeric G-protein function, influences renal tubular regeneration following IRI. In rat kidneys exposed to IRI, there was a temporal induction in renal AGS3 protein expression that peaked 72 h after reperfusion and corresponded to the repair and recovery phase following ischemic injury. Renal AGS3 expression was localized predominantly to the recovering outer medullary proximal tubular cells and was highly coexpressed with Ki-67, a marker of cell proliferation. Kidneys from mice deficient in the expression of AGS3 exhibited impaired renal tubular recovery 7 d following IRI compared to wild-type AGS3-expressing mice. Mechanistically, genetic knockdown of endogenous AGS3 mRNA and protein in renal tubular epithelial cells reduced cell proliferation in vitro. Similar reductions in renal tubular epithelial cell proliferation were observed following incubation with gallein, a selective inhibitor of Gβγ subunit activity, and lentiviral overexpression of the carboxyl-terminus of G-protein-coupled receptor kinase 2 (GRK2ct), a scavenger of Gβγ subunits. In summary, these data suggest that AGS3 acts through a novel receptor-independent mechanism to facilitate renal tubular epithelial cell proliferation and renal tubular regeneration.—Regner, K. R., Nozu, K., Lanier, S. M., Blumer, J. B., Avner, E. D., Sweeney, Jr., W. E., Park, F. Loss of activator of G-protein signaling 3 impairs renal tubular regeneration following acute kidney injury in rodents.

Renal ischemia–reperfusion injury (IRI) is a common cause of acute kidney injury (AKI), which is associated with increased patient morbidity and mortality (1). Acute ischemia leads to ATP depletion and initiates renal tubular cell injury, which is exacerbated on reperfusion (2, 3). During the recovery phase, renal tubular regeneration occurs in a coordinated fashion, primarily by dedifferentiation of sublethally injured tubular epithelial cells (4–6). Subsequently, these cells proliferate and migrate to reestablish cell polarity and cell-cell adhesion. These processes ultimately culminate in the restoration of normal tubular morphology and function (4–7). Although this sequence of events is well characterized, the molecular mechanisms that influence renal tubular epithelial cell proliferation during tubular repair remain poorly understood.

G-protein-coupled receptors (GPCRs) activate heterotrimeric G proteins to initiate and integrate a variety of critical intracellular signals to coordinate biological responses, including those that influence cell proliferation (8). Emerging evidence suggests that heterotrimeric G-protein signaling can be regulated independently of cell surface receptors through the action of accessory proteins (9). Accessory proteins modulate heterotrimeric G-protein function by controlling signaling intensity, receptor trafficking, and accessibility to specific signaling molecules (10). These regulatory proteins are classified by the mechanism in which they control the actions of the Gα or Gβγ subunits (9, 10). In the present study, we focused on activator of G protein signaling 3 (AGS3) also referred to as G-protein signaling modulator 1 (Gpsm1) (11, 12), which was originally identified using a yeast-based screening system (11, 12). AGS3 is one member of a larger family of AGS proteins (9, 10). AGS3 acts as a guanine nucleotide dissociation inhibitor (GDI) and regulates heterotrimeric G-protein activity through the selective binding of Gα subunits to multiple G-protein regulatory (GPR) motifs (12–14). AGS3 is a highly conserved protein in which homologs can be identified in species from Drosophila to humans (15). For this reason, AGS3 has a diverse number of biological functions, including roles in the regulation of cell polarity, mitotic spindle orientation, adenylyl cyclase activity, and metabolic activity (9, 15–19). To date, the function of AGS3 in the kidney has been largely unexplored.

Recently, our group demonstrated an unexpectedly high expression of AGS3 in rodent and human kidneys with genetic defects that promote abnormal renal epithelial cell proliferation and cystogenesis (20). In this report, we demonstrate that AGS3 protein expression is induced in normal rat and mouse kidneys during the regenerative phase following renal IRI. In vitro and in vivo experiments indicated that AGS3 influences renal tubular epithelial cell proliferation and modulates renal tubular repair and regeneration following IRI.

MATERIALS AND METHODS

Animals

Male Sprague-Dawley (SD) rats (200–225 g) were obtained from Taconic Farms (Germantown, NY, USA). Gpsm1^+/+, Gpsm1^+/−, and Gpsm1^−/− were generated and genotyped, as described previously (16). Wild-type Gpsm1^+/+ mice were either bred from heterozygous parental Gpsm1^+/− mice backcrossed to C57BL/6J mice for 12 generations or obtained from Jackson Laboratories (C57BL6/J; 20–22g; Bar Harbor, ME, USA). All mice and rats were allowed ad libitum access to food and water during the course of the experimental period. All protocols used in this study were approved by the Medical College of Wisconsin Institutional Animal Care and Use Committee.

Rat model of renal IRI

SD rats underwent 30 min bilateral renal ischemia by occlusion of the renal pedicles with microvascular clamps using surgical procedures, as described previously by our group (21, 22). Rats were sacrificed at various time points (24, 48, 72, 96, and 168 h) following ischemia. In a separate set of SD rats, unilateral IRI was performed in which only the left renal pedicle was occluded for 30 min and allowed to reperfuse until sacrifice at 24 and 72 h. Sham-operated rats underwent the same surgical procedure without clamping of the renal pedicles. All of the kidneys were harvested for histology and protein extraction. Prior to sacrifice of the rats, blood was harvested by cardiac puncture and collected in heparin-coated tubes. The tubes were spun, and plasma was isolated to measure creatinine (Bioassay Systems, Hayward, CA, USA), which is an index of renal function.

Mouse model of renal IRI

C57BL/6J (or Gpsm1^+/+), Gpsm1^+/−, and Gpsm1^−/− mice were anesthetized with nembutal (40–60 mg/g body weight). Subsequently, the abdominal cavity was exposed, and a microvascular clamp was placed onto the left renal pedicle to occlude blood flow for 40 min. At the end of the ischemic period, the clamp was removed, and the left kidney was allowed to reperfuse. The left kidney was observed until the recovery of normal renal color prior to closure of the wound. As a control, sham-operated Gpsm1^+/+ mice underwent the same surgical procedure without the clamping of the left renal pedicle. At different time points, blood was harvested to isolate plasma to measure creatinine levels (Bioassay Systems). The mice were sacrificed at 168 h following the ischemia, and the left kidneys were harvested for histology and protein extraction.

Histopathologic analysis of renal injury

Kidneys from sham-operated and IRI-treated mice were formalin fixed, paraffin embedded, sectioned (4 μm), and stained with hematoxylin and eosin (H&E). Ten randomly chosen cortical and corticomedullary fields were photographed from each mouse kidney using a digital color camera (×40 view) attached to a Nikon 55i light microscope (Nikon Instruments, Melville, NY, USA). Acute renal damage and recovery by the renal tubular epithelial cells following injury were determined using two methods: ratio of damaged to total tubular area, as described by Westhoff et al. (23); and epithelial cell density, as described by Kim et al. (24). All morphometric analyses of the kidney samples were performed in a genotype-masked manner using Nis-Elements 3.03 image analysis software (Nikon Instruments). In the first analysis, damaged area and total tubular area were calculated by marking areas of necrosis, cast formation, and dilated tubules relative to the total tubular area in the high-power field. A quantitative ratio of damaged to total tubular area was calculated to provide an index of tubular injury. In the second analysis, tubular epithelial cell density was calculated by counting individual tubular epithelial cell nuclei in the corticomedullary region for each mouse kidney group. The tubular epithelial cell density was expressed as the number of tubular epithelial cell nuclei per unit area (0.1 mm²).

AGS3 expression in sham-operated and IRI-treated kidneys

Mouse and rat renal protein lysates were isolated using 1× RIPA buffer containing protease (Roche, Mannheim, Germany) and phosphatase inhibitors (Pierce, Rockford, IL, USA) from sham-operated and IRI-treated kidneys at 24, 48, 72, and 168 h following ischemia. AGS3 expression was determined using standard Western blot techniques. The primary rabbit antibody for AGS3 was generated by immunization of rabbits (Life Technologies, Carlsbad, CA, USA) with a specific antigenic peptide designed using the amino acid sequence previously described (25). As a control, AGS5/LGN/Gpsm2 was measured using a polyclonal antibody generated and validated by Blumer et al. (25). Mouse anti-GAPDH (cat. no. G879; 1:4000 dilution) and mouse anti-β-actin (A5441; 1:4000) were obtained from Sigma (St. Louis, MO, USA). Densitometry was performed to quantify band intensity using ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA).

Immunohistochemical localization of AGS3 with Ki-67 and tubule segment-specific markers in normal and IRI kidneys

Zinc formalin-fixed, paraffin-embedded rat kidney tissue was sectioned (4 μm), deparaffinized, and either individually stained for AGS3 or Ki-67, or double stained for AGS3 and Ki-67. In some kidneys, serial sections were made to localize the tubule segments associated with AGS3 expression. Phaseolus vulgaris erythroagglutinin (PVA-E) and Dolichos biflorus agglutinin (DBA) are lectins that specifically stain proximal convoluted tubules (26) and collecting ducts (27), respectively. Calbindin-D 28K was used as a marker for distal tubules (28). All staining for AGS3 and Ki-67 was performed using the Dako Autostainer Plus system (Dako, Carpenteria, CA, USA). The lectin and calbindin-D 28K staining was performed using a protocol previously described by Sweeney et al. (29). For AGS3, Ki-67, tubule-segment marker sections, rabbit polyclonal AGS3 antibody (1:600–1:750 dilution), rat Ki-67 antibody (1:150), biotinylated lectins (1:400), or calbindin-D 28K (1:1500; clone CB-955; Sigma) were used, and DAB was applied as the detection reagent following the secondary antibody (AGS3 and Ki-67) or extravidin-peroxidase (PVA-E, DBA, and calbindin-D 28K) step. Biotinylated DBA lectin and calbindin-D 28K were detected using Sigma Fast Red staining following the extravidin-alkaline phosphastase step. For the AGS3/Ki-67 double-stained sections, Ki-67 was detected using DAB, and AGS3 was detected using Vulcan Fast Red. All slides were subsequently counterstained in Mayer's hematoxylin solution and coverslipped. The number of Ki-67⁺ and AGS3⁺ cells in the renal cortex and outer medulla were counted in ≥5 randomly selected cortex and corticomedullary fields from each kidney. Representative images were acquired using a Nikon 55i light microscope and Nis-Elements image analysis software. For each kidney, >750 total tubular epithelial cells were counted.

Cell number quantification using genetically modified renal epithelial cells

Lentiviral vectors expressing control (scrambled) or AGS3-specific shRNA, or the carboxyl terminus of the G-protein-coupled receptor kinase (GRK2ct) cDNA, as described previously (20), were produced by triple plasmid transfection (30) and titered by PCR using specific WPRE primers (31). Normal rat kidney-52E (NRK) cells were serially transduced with the lentiviral vectors at MOI ∼ 10–20, and subsequently expanded for analysis to determine the effects of AGS3 knockdown on cell numbers and cAMP production. NRK cells expressing control shRNA (NRK-Csh), AGS3 shRNA (NRK-AGS3sh), and GRK2ct (NRK-GRK2) were denoted as indicated in parentheses. Cell numbers were determined using CyQuant Direct Cell Proliferation Assay (Life Technologies) at distinct time points over a 3-d period following portioning of the genetically modified cells into 96-well plates. Low-serum medium (1% FBS) was replaced each day during the course of the experiment. A standard curve generated on each plate was used to provide a quantitative value for the cell numbers between the different groups.

Measurement of cAMP levels

To measure the cAMP levels, the genetically modified cells were expanded in 60-mm plates for 48–72 h and harvested in 0.1 N HCl, and cAMP levels were determined by ELISA (Assay Designs, Ann Arbor, MI, USA), as previously performed in our laboratory (32).

Statistical analysis

AGS3 protein expression differences between groups were assessed by a nonparametric 1-way ANOVA. Dunn's post hoc analysis was used to compare individual groups with the appropriate control groups. For all other statistical analyses, we performed a t test or 1-way ANOVA followed by a Tukey post hoc test. All statistical analyses were performed using Prism 4.0 software (GraphPad, San Diego, CA, USA). In all cases, values of P < 0.05 were considered statistically significant.

RESULTS

Induction of AGS3 protein expression during the recovery phase following renal IRI

The role of AGS3 in the kidney has been largely ignored, in large part because of the lack of detectable AGS3 protein in normal kidneys, as determined in our laboratory (ref. 20 and Fig. 1Q) and others using immunoblot analysis (25, 33). In the current study, we investigated whether renal AGS3 expression could be induced following IRI, a biological stimulus that promotes renal tubular epithelial cell proliferation.

Figure 1.

AGS3 expression and localization in rat kidneys following IRI. SD rats (n=4–8/group and time point) underwent 30-min bilateral renal artery clamping and were allowed to recover between 24 and 168 h following reperfusion. A–E) H&E-stained kidney sections were examined in sham-operated (A) and IRI-treated kidneys at 24 (B), 48 (C), 72 (D) and 168 h (or 7 d; E) after ischemia. At 24 h, there was extensive necrosis (asterisks; B) of proximal tubular cells in the outer stripe of the outer medulla. At 48 h, slight improvement in the extent of tubular necrosis (asterisks; C) was noted. By 72 h, some regenerating tubules were lined by flattened epithelium (arrows; D). At 168 h, most of the renal tubules had regained their normal appearance (E). On rare occasions, tubules lined by thinned epithelium could be observed (arrowheads; E). F–P) Localization of AGS3 and tubule segment-specific lectins was determined in serial sections by immunohistochemistry in sham-operated kidneys (F–I) and IRI-treated kidneys at 72 h (J–M) and 168 h (N–P) after ischemia. Lectins for DBA (G, O) and PVA-E (H, L, P) were used as markers for collecting duct and proximal tubular epithelial cells, respectively. Calbindin-D 28K was used as a marker for the distal tubules (K). All sections were counterstained with hematoxylin. Brown DAB staining denotes AGS3⁺ cells (F, J, N). In kidneys of the sham-operated rat, AGS3 (asterisks; F) was predominantly localized to the collecting duct epithelial cells, which were DBA⁺ (red stain; G). As negative control for AGS3 staining, kidney sections were incubated with AGS3 antibody preabsorbed with the competing peptide (I, M), preimmune rabbit serum or vehicle solution (data not shown). Arrowheads (N, P) indicate corresponding regenerating proximal tubules (PVA-E⁺; P) with high expression of AGS3 (N) compared to the minimal AGS3 expression in the collecting duct (DBA⁺; O). Original magnification ×20 (A--E) and ×40 (F--P). Q) Representative immunoblot analysis of renal AGS3 in kidneys from sham-operated and IRI-treated rats at 24, 48, 72, and 168 h postinjury. Expression of LGN, a close homologue of AGS3, did not change over the same time period. GAPDH was used as a loading control. R) Renal function as measured by plasma creatinine levels, and fold-change in AGS3 levels from sham-operated and IRI-treated rats at 24, 48, 72, 96, and 168 h postinjury; n = 3–8 samples/time point and group. *P < 0.05 vs. all other time points. **P < 0.001 vs. all other groups at every time point.

In the current study, we examined renal AGS3 protein expression and localization by immunoblotting and immunohistochemistry, respectively, using a polyclonal rabbit AGS3 antibody that was generated in our laboratory. The specificity of the antibody to detect the full-length AGS3 protein (∼74 kDa) is shown in Supplemental Figs. S1 and S3. The AGS3 band was not detectable using the antibody preabsorbed with the competing peptide (Supplemental Fig. S3). The inability to readily detect AGS3 protein under normal conditions using whole-kidney lysates is likely due to its localization to the cortical and outer medullary collecting duct epithelial cells (DBA⁺ cells), which only comprise a small percentage of the total renal cell population (Fig. 1F, G). No AGS3-specific staining was detected in other cortical or medullary cell types, including glomeruli, vessels, or proximal tubules (Fig. 1F, H) in normal (or sham-operated) kidney sections. As a negative control, AGS3 immunostaining was not detectable in kidney sections incubated with AGS3 antibody preabsorbed with the competing peptide (Fig. 1I, M), preimmune sera (not shown), or vehicle solution (not shown).

Renal function was markedly decreased by 24 h following 30 min bilateral renal ischemia, and some recovery of renal function was noted by 72 h, as determined by plasma creatinine measurement (Fig. 1R). Between 24 and 48 h after ischemia, marked necrosis and sloughing of proximal tubular cells were noted, particularly in the outer stripe of the outer medulla (Fig. 1B). Tubular necrosis was less prominent by 72 h, and some tubules demonstrated tubular dilatation and thinning of the epithelial cell layer (Fig. 1C). AGS3 protein expression was not significantly increased by 24 h following ischemia (Fig. 1Q, R). In contrast, AGS3 protein expression peaked with a 57.3 ± 23.2-fold increase at 72 h (P<0.05 vs. time-control, sham-operated kidneys), which corresponded to the improvement in renal function and tubular injury (Fig. 1Q, R). In contrast to normal or sham-operated kidneys, AGS3 was localized to the thinned epithelial cells remaining on the basement membrane of the injured proximal (PVA-E⁺ cells; Fig. 1L) and not distal (calbindin-D 28K⁻; Fig. 1K) tubules at all time points following ischemia, including 72 h (Fig. 1J, L) and 168 h (Fig. 1N, P). No AGS3 staining was noted in glomeruli or vessels. By 168 h following ischemia, renal function recovered to normal levels, as determined by serum creatinine (Fig. 1R), and histological analysis demonstrated near-complete recovery of normal renal tubular architecture (Fig. 1E). However, rare tubules lined by thinned epithelium could be observed (Fig. 1E, N; arrowheads). In those tubules, the AGS3 expression remained fairly robust compared to the collecting duct cells, which expressed relatively little to no AGS3 in comparison (Fig. 1N, O). At the 168-h time point, AGS3 protein expression was returning toward normal (or sham-operated) levels (8.0±2.8-fold higher; Fig. 1Q, R). The expression of LGN, a close homologue of AGS3, did not change following ischemia (Fig. 1Q). Taken together, these findings demonstrate that AGS3 expression can be induced in rat kidneys following IRI.

Expression of AGS3 in renal tubular epithelial cells actively undergoing mitosis

Proliferation of sublethally injured tubular epithelial cells is a critical component of renal tubular regeneration (4–6). To determine whether AGS3 was expressed in regenerating renal tubules, we costained kidney sections with AGS3 and Ki-67, a marker of proliferation. Ki-67 is a nuclear antigenic marker of cell proliferation, which is only expressed in actively cycling cells (34). The percentage of total tubular cells immunoreactive for Ki-67 at 48 h was significantly elevated (P<0.05) in the cortex (25.1±4.5%) and outer medulla (48.8±4.5%) compared to the corresponding cortical (1.9±0.7%) and medullary regions (2.1±0.2%) in sham-operated rats (Fig. 2M) and was consistent with prior studies demonstrating a peak proliferative response ∼48 h after injury (6, 35). In sham-operated rat kidneys, little staining for AGS3 or Ki-67 was noted in the cortex or the outer medulla (Fig. 2A, E, I). The relatively weak staining of AGS3 in the renal cortex and medulla is consistent with previous studies in our laboratory (20). In contrast, AGS3 staining was detected in Ki-67⁺ renal tubular epithelial cells 48 h following IRI in both the cortex and outer medulla (Fig. 2B–D, F–H, J–L). Approximately 80% (80.0±3.0%; n=4 kidneys) of the tubular epithelial cells in the outer stripe of the outer medulla were AGS3⁺ 48 h following IRI (Fig. 2N), whereas only 0.52 ± 0.3% (n=3) of the total outer medullary tubular cells were AGS3⁺ in the sham-treated kidneys. About half of the tubular epithelial cells (51.3±2.7%; n=4 kidneys) were colabeled with both AGS3 and Ki-67 (Fig. 2O) following IRI. No double AGS3- and Ki-67-labeled cells were detected in the sham-treated kidneys (n=3). Taken together, these findings demonstrate that AGS3 is expressed in proliferating renal epithelial cells in regenerating renal tubules following IRI.

Figure 2.

AGS3 colocalizes with proliferation marker Ki-67. A–L) Representative images of Ki-67 in the cortex (A–D) and outer medulla (E–L) from sham-operated (A, E, I) and 48-h bilateral IRI-treated rats (B–D, F–H, J–L). Images are from the renal outer medulla in two different rats. Arrows in panels C, G, and K indicate areas magnified in panels D, H, and L, respectively. Ki-67 (brown) and AGS3 (red) were detected by DAB and Vulcan fast red staining, respectively. Arrowhead in panel C shows an AGS3-deficient tubule in the absence of Ki-67 unlike the adjacent Ki-67⁺ tubule (arrow) with detectable levels of AGS3 protein expression. All sections were counterstained with hematoxylin. M) Quantitation of Ki-67⁺ epithelial cells in the renal cortex (Ctx) and outer medulla (OM) as a percentage of total tubular epithelial cells. N) Quantitation of AGS3⁺ epithelial cells in OM as a percentage of total tubular epithelial cells. O) Quantitation of Ki-67 and AGS3 double-positive epithelial cells in OM as a percentage of the total tubular epithelial cells. n = 3–5 samples/group, as indicated on bars. *P < 0.05, **P < 0.001 vs. sham-operation group.

Loss of AGS3 protein impairs tubular regeneration following IRI

The effect of a partial or complete loss of renal AGS3 expression on the recovery of tubular epithelial cells following IRI was investigated using either hypomorph (Gpsm1^+/−; n=6) or null (Gpsm1^−/−; n=2) AGS3 mice. Renal AGS3 protein expression was similarly induced in the acutely injured kidneys following either bilateral (Fig. 1) or unilateral IRI (Supplemental Fig. S3). Unlike the bilateral IRI model (Fig. 1R), no significant difference was detected in the plasma creatinine after 24 and 168 h following the ischemic injury (Supplemental Fig. S4A). Since the unilateral IRI model exhibits similar changes in AGS3 protein induction, while minimizing the likelihood of animal mortality, we chose to perform our next set of experiments using the wild-type and AGS3-deficient mice with the unilateral IRI procedure.

The kidneys from the wild-type Gpsm1^+/+ and AGS3-deficient Gpsm1^+/− and Gpsm1^−/− mice were harvested 7 d following unilateral IRI to examine AGS3 expression and renal tubular morphology. As shown in Supplemental Fig. S4B, induction of AGS3 expression was markedly higher in the Gpsm1^+/+ mice following unilateral IRI at d 7 (n=5) compared to the sham-operated Gpsm1^+/+ mice (n=3). Interestingly, the induction of AGS3 in the heterozygous Gpsm1^+/− mouse kidneys was minimal compared to the wild-type Gpsm1^+/+ mice. As expected, no AGS3 protein expression was detected in the null Gpsm1^−/− mouse kidney.

Renal tubular architecture was similar in the uninjured wild-type Gpsm1^+/+ (Fig. 3A–E) and uninjured AGS3-deficient Gpsm1^+/− (not shown) and Gpsm1^−/− (Fig. 3K–O) mouse kidneys. In contrast, there was an increase in the number of dilated tubules, thinned tubular epithelial cells, and cast formation in both the heterozygous Gpsm1^+/− (n=6) and null Gpsm1^−/− (n=2) mouse kidneys following IRI (Fig. 3P–T) compared to injured wild-type Gpsm1^+/+ mouse kidneys (Fig. 3F–J). Renal tubular injury in the renal corticomedullary and outer medullary region was quantified by morphometric analysis of the kidney sections. At 7 d after injury, the surface area of renal tubular injury in the outer medulla (Fig. 4A) and total kidney (i.e., combination of the corticomedullary and outer medullary surface area; Fig. 4B) was significantly higher (P<0.001) by ∼2-fold in the AGS3-deficient Gpsm1^+/− and Gpsm1^−/− mouse kidneys compared to the wild-type Gpsm1^+/+ mouse kidneys. No tubular injury in the outer medulla (Fig. 4A) or total kidney (Fig. 4B) was detected in either the uninjured wild-type or the AGS3-deficient mouse kidneys. In addition, tubular epithelial cell density calculated 7 d following the initiation of reperfusion (Fig. 4C) in the IRI-treated Gpsm1^+/− and Gpsm1^−/− mouse kidneys was significantly lower (P<0.001) by ∼25% (258.5±5.0 nuclei/0.1 mm²; n=8) compared to the IRI-treated Gpsm1^+/+ mouse kidneys (343.2±20.4 nuclei/0.1 mm²; n=6). Epithelial cell density in the uninjured wild-type (226.5±6.3 nuclei/0.1 mm²; n=6) and AGS3-deficient (246.1±9.8 nuclei/0.1 mm²; n=8) mouse kidneys (226.5±6.3 nuclei/0.1 mm²; n=6) did not differ significantly from each other or from the IRI-treated AGS3-deficient mouse kidneys.

Figure 3.

Partial or complete loss of AGS3 impairs renal tubular recovery following unilateral IRI in mice. Representative H&E stained sections of the renal cortex and outer medulla from Gpsm1^+/+ (A–J) and Gpsm1^−/− mice (K–T) 7 d following sham operation (A–E, K–O) or unilateral IRI (F–J, P–T). Diluted tubules were found in the cortex and outer medulla in the IRI mouse kidneys with the genotypes Gpsm1^−/− (asterisks; P–T) and Gpsm1^+/− (not shown), but not in either the uninjured Gpsm1^−/− kidneys (K–O) or the normal Gpsm1^+/+ kidneys with IRI (F–J) or without IRI (A–E). Hypernucleated renal tubules (pound sign; H, J) in the cortex and outer medullary region were found in the Gpsm1^+/+ mice 7 d following IRI. In contrast, tubules with casts (Ca; T) or thinned epithelium (arrowheads; R, T) were frequently noted in Gpsm1^−/− (P–T) and Gpsm1^+/− (not shown) kidneys 7 d postinjury. Original view: ×10 (A, F, K, P); ×20 (B, D, G, I, L, N, Q, S); ×40 (C, E, H, J, M, O, R, and T). Scale bars = 50 μm.

Figure 4.

AGS3 expression and morphometric analysis of kidney injury following unilateral IRI in mice. Morphometric analysis was performed on histological sections obtained from the uninjured (Uninj) and IRI kidneys at 7 d following ischemia. A, B) Changes in outer medullary (A) and total kidney tubular injury (B) were quantified by histological analysis of H&E kidney sections. *P < 0.001 vs. all other groups; **P < 0.001 vs. both uninjured groups. C) Tubular epithelial cell density was histologically quantified using H&E sections from uninjured and IRI-treated kidneys. *P < 0.001 vs. IRI-treated AGS3-deficient group. n = 6–8 mice/group, as indicated on bars.

Effect of AGS3 knockdown on epithelial cell numbers

Presently, no pharmacological inhibitors of AGS3 function are available. Therefore, to evaluate whether AGS3 plays a biological role in promoting epithelial cell numbers, we pursued a genetic approach in which the expression of AGS3 was modified using lentiviral vectors expressing either control or AGS3-specific shRNA in the NRK tubular epithelial cell line. The sequences of the AGS3-specific shRNA are shown in Supplemental Fig. S2A. Because of the intrinsic ability of lentiviral vectors to integrate into the genome, control and AGS3 shRNA-expressing NRK cells were able to be serially passaged for use in the experiments described in Figs. 5–7.

Figure 5.

Effect of AGS3 on renal epithelial cell proliferation. A) Cell proliferation was determined by DNA fluorescence assay using vehicle-treated NRK-Veh epithelial cells (●; n=7), control NRK-Csh cells (○; n=8) or NRK-AGS3sh cells (▾; n=8). Cells were portioned into 96-well plates and counted by measuring DNA fluorescence intensity at distinct point points over a 3-d period. *P < 0.001 vs. corresponding control cells. B) NRK-Csh (solid bars) or NRK-AGS3sh (shaded bars) cells were portioned into 6-well dishes and counted at 24 and 48 h later by hemocytometry. Number of samples counted by hemocytometry is indicated on bars. *P < 0.001 vs. NRK-AGS3sh cells.

Figure 6.

Role of Gβγ subunit on AGS3-mediated renal epithelial cell proliferation. A) NRK-Csh epithelial cells were incubated with vehicle (●) or gallein (1.3 μM, ○; 6.5 μM, ▾) at 24 h after portioning into 96-well plates. Cell numbers were determined at distinct times over a 3-d period. n = 7–8/group/time point. *P < 0.001 vs. vehicle-treated cells. B) NRK cells were transduced with lentiviral vectors expressing either control shRNA (■) or GRK2ct (□). GRK2ct inhibits Gβγ activity by scavenging free Gβγ subunits. Cell numbers were determined 24 and 48 h after initial plating. Two separate experiments were performed using 6–8 samples/group at each time point. *P < 0.001 vs. corresponding control cells.

Figure 7.

Effect of AGS3 on cAMP production and epithelial cell numbers. A) cAMP levels were measured by ELISA using cell lysates collected from vehicle-treated NRK-Veh cells or genetically modified control NRK-Csh or NRK-AGS3sh cells. cAMP levels are graphed as percentage change normalized to values obtained from NRK-Csh group; n = 8 samples/group. B) NRK-Veh, NRK-Csh, and NRK-AGS3sh cells were portioned into 96-well plates and treated with 10 μM H-89, a selective protein kinase A inhibitor, for 48 h. Cell numbers were measured by DNA fluorescence assay; n = 4–7 samples/group as indicated on bars.

Knockdown of endogenously expressed AGS3 was achieved in NRK-AGS3sh cells (see Materials and Methods and Supplemental Fig. S2B). No effect on AGS3 expression was observed using the control NRK-Csh cells as compared to the untreated NRK (NRK-Veh) cells. The changes in epithelial cell numbers for NRK-Veh, NRK-Csh, and NRK-AGS3sh cells were evaluated at distinct time points using two different methods: DNA fluorescence assay (Fig. 5A) and hemocytometry (Fig. 5B). Epithelial cell proliferation of NRK-AGS3sh cells was significantly lower (P<0.001) compared to NRK-Veh and NRK-Csh cells (Fig. 5A). These findings were confirmed by hemocytometry in which NRK-AGS3sh cell numbers (330,408±17,365 cells) were significantly lower (P<0.001) than NRK-Csh (838,319±33,763 cells) 48 h after plating (Fig. 5B).

Regulation of epithelial cell numbers by AGS3 is dependent on Gβγ function, and not cAMP production

We postulated that sequestration of Gα_i subunits by AGS3 could alter renal epithelial cell numbers through two distinct intracellular mechanisms: preventing the reassociation of Gα_i with free Gβγ subunits, or abrogating the inhibitory effect of Gα_i on the production of cAMP.

The effect of Gβγ-subunit function on epithelial cell numbers was examined using genetically modified NRK-Csh cells. NRK-Csh cells were incubated with gallein, a small-molecule inhibitor of Gβγ subunits, at 2 concentrations (1.3 and 6.5 μM) over a 2-d period. As shown in Fig. 6A, there was a significant dose-dependent reduction (P<0.001) in epithelial cell numbers following incubation with gallein compared to vehicle. In another set of experiments, NRK cells were genetically modified to express the carboxyl terminus of GRK2ct, which acts as a scavenger of unbound Gβγ subunits (36). Epithelial cell numbers were decreased by 27 ± 4% in the GRK2ct-expressing NRK cells compared to control shRNA-expressing cells (P<0.001; Fig. 6B).

Gα_i subunits directly inhibit adenylyl cyclase activity (37). Therefore, we performed additional experiments to determine whether the AGS3-Gα_i-GDP complex abrogates the inhibitory effect of Gα_i on adenylyl cyclase. Following genetic manipulation of AGS3 expression in NRK cells, we measured changes in cAMP levels (Fig. 7A) and cell numbers (Fig. 7B). No significant differences in the intracellular levels of cAMP were noted between NRK-Csh and NRK-AGS3sh cell lines in the presence (Fig. 7A) or absence (data not shown) of a phosphodiesterase inhibitor (IBMX). Consistent with these findings, chronic blockade of protein kinase A (PKA) did not significantly change epithelial cell numbers regardless of the level of AGS3 expression in the genetically modified NRK cells (Fig. 7B). In the presence of H-89 (10 μM), a selective inhibitor of PKA, epithelial cell numbers decreased by ∼40–60% in each of the NRK cell groups. These findings demonstrate that the cAMP/PKA pathway promotes renal epithelial cell proliferation under basal conditions, but AGS3 is not a major regulator of cAMP production in renal epithelial cells. Taken together, this set of experiments suggests that AGS3 mediates changes in renal tubular epithelial cell numbers through Gβγ and not through a cAMP/PKA-dependent pathway.

DISCUSSION

Epithelial cell proliferation plays a critical role in the repair and regeneration of renal tubules following IRI, but the intracellular mechanisms that mediate this proliferative response remain poorly characterized. In the present study, we provide evidence that AGS3, a novel receptor-independent regulator of G-protein function, is induced in the kidney during the recovery phase following renal IRI. Mouse models deficient in the expression of AGS3 following IRI demonstrated impaired renal tubular recovery compared to wild-type mice. Genetic knockdown of endogenous AGS3 expression in renal tubular epithelial cells led to a significant reduction in cell numbers in vitro. These findings strongly suggested that AGS3 contributes to renal tubular repair by influencing epithelial cell proliferation.

Classically, signals from the external milieu are transmitted through the plasma membrane by cell surface GPCRs that activate heterotrimeric G-protein-mediated signal transduction cascades. Heterotrimeric G proteins have been implicated in nearly every cellular function required to maintain normal homeostasis. Moreover, alterations in the function of G proteins contribute to the cellular adaptation to a variety of stimuli. Recent studies have demonstrated that heterotrimeric G proteins can be regulated independently of cell surface receptor activation through a number of accessory proteins, including AGS3 (9, 15). These accessory proteins provide spatial and temporal responses to pathological stimuli by controlling signaling pathway intensity and integration, receptor trafficking, and accessibility to specific signaling molecules (10).

AGS3 is one member in the AGS protein family (9, 15) and functions as a guanine nucleotide dissociation inhibitor. The activation of heterotrimeric G proteins is dependent on the exchange of GDP with GTP on the Gα subunits. The intrinsic GTPase activity of Gα subunits hydrolyzes GTP to GDP to terminate the activation state. AGS3 contains multiple GPR (12) or GoLoco (12, 38, 39) motifs in the C-terminal part of the protein, and these sites preferentially bind to Gα_i subunits complexed with GDP. A single molecule of AGS3 can stoichiometrically bind up to 4 Gα_i-GDP (13, 40, 41), and it has the potential to dramatically shift the balance between activation and inactivation of heterotrimeric G proteins. AGS3 and other GPR proteins can also serve as alternate binding partners for Gα_i independent of Gβγ to generate a GPR-Gα_i signaling module that is regulated by guanine exchange factors and cell surface receptors (41–44).

Previous studies of AGS3 function have focused primarily on biological processes integral to the brain and heart, which were the mammalian organs originally found to express high levels of endogenous AGS3 (25, 33, 45). AGS3 was shown to regulate asymmetric cell division, membrane protein transport, neuronal development, and synaptic plasticity (9). A recent study by Blumer et al. (16) found that AGS3-knockout mice exhibited unexpected reductions in body weight, body fat mass, and blood pressure, suggesting that AGS3 may have additional functions in other organs. Early studies did not observe any detectable levels of AGS3 protein in the kidney (25, 33, 45). In contrast, our laboratory observed that AGS3 is highly expressed in rodent and human kidneys affected with genetic mutations that exhibit phenotypes consistent with polycystic kidney disease (20). In the absence of these genetic mutations, the renal expression of AGS3 remained largely undetectable (20). Since the mitotic index of the renal epithelia is extremely low under normal conditions, we speculated that the inability to detect AGS3 in the normal kidney may be related to the cell cycle status of the renal tubular epithelial cells. In the present study, we demonstrated that renal tubular epithelial cells in the outer stripe of the outer medulla can be induced to transiently express AGS3 following renal IRI, which is a potent stimulus for epithelial cell proliferation. The delayed postischemic induction of AGS3 suggests that the ischemia itself is not directly responsible for the changes in AGS3 expression. The similarity in magnitude of AGS3 expression in the unilateral and bilateral IRI rat models suggests that the accumulation of creatinine, urea, endogenous guanidino compounds (46, 47), and other constituents of the uremic milieu are also unlikely to be major factors in the induction of AGS3.

Although the mechanisms involved in the induction of AGS3 expression are not well defined, it is clear from the present study that renal tubular epithelial cell repair and recovery is impaired in the genetically defective Gpsm1 mice compared to the wild-type Gpsm1^+/+ mice. In the Gpsm1-defective mice, tubular epithelial cell density was significantly reduced at d 7 following unilateral IRI compared to the wild-type Gpsm1^+/+ mice. We presumed that this was attributable to a decrease in the proliferative capacity of AGS3-deficient kidneys. Indeed, this is supported by the strong expression of AGS3 in proliferating renal tubular epithelial cells in vivo and the reduction in tubular epithelial cell numbers following AGS3 knockdown in vitro.

To date, the role of AGS3 in the regulation of epithelial cell proliferation and growth is poorly understood. We demonstrated that selective knockdown of AGS3 can reduce the number of cultured rat renal tubular epithelial cells in vitro. We postulate that AGS3 alters the intracellular signaling cascades that regulate proliferation by either actively promoting dissociation of Gαβγ to release Gβγ, or scavenging Gα_i/o subunits and preventing their reassociation with Gβγ subunits. This would inhibit the function of Gα_i/o proteins and leaves Gβγ subunits free to activate downstream effectors. In the present study, the reduction in epithelial cell proliferation following AGS3 knockdown was faithfully recapitulated by pharmacologic and genetic inhibition of Gβγ function. The involvement of Gβγ subunits as an intermediary signaling molecule has been well established in the mammalian brain and neural derived cells (17, 18, 25, 48). We recently demonstrated increased expression and function of AGS3 in animal models of polycystic kidney disease and that proliferation of isolated renal collecting duct epithelial cells were inhibited by either knockdown of AGS3 or scavenging of Gβγ subunits (20). In renal epithelial cells, the mechanism of action of free Gβγ subunits is not well characterized but could involve the activation of MAPK pathways (49), regulation of mitotic spindle orientation and cell polarity (48), or direct interaction with adenylyl cyclase to promote cAMP production (19).

As an alternate mechanism of action, we investigated whether AGS3 would prevent the reformation of the Gβγ-Gα_i heterotrimer and alter the production of intracellular cAMP. It is well established that Gα_i plays a role in the inhibition of adenylyl cyclase (50). Therefore, AGS3 scavenging of Gα_i would be expected to abrogate the inhibitory effect of Gα_i on adenylyl cyclase, leading to an increase in cAMP and cell proliferation. In the present study, we found no evidence that AGS3 changed the production of cAMP in the renal tubular epithelial cells in vitro, even in the presence of a phosphodiesterase inhibitor. These findings are consistent with previous studies by Sato et al. (51), using transfected AGS3-overexpressing CHO cells, and in our laboratory, using cystic renal epithelial cells (20). However, there is evidence that AGS3 promotes the cAMP/PKA system in neurally derived cells following morphine withdrawal (19). These differences in the effect of AGS3 on cAMP production suggest that the intracellular action of AGS3 depends on both the biological stimulus and cell type. In the context of renal tubular regeneration, the lack of an effect of AGS3 on cAMP in renal epithelial cells may be beneficial, since recent studies demonstrated that increased cAMP/PKA activity in renal epithelial cells leads to reduced cell numbers (52). Taken together, our present findings in conjunction with previously published studies in extrarenal organs, suggest that there is a unique role for AGS3 in the regulation of proliferation through a noncanonical Gβγ-dependent mechanism.

In summary, this is the first study to highlight an important role for accessory proteins in the regulation of heterotrimeric G proteins during renal tubular regeneration. Our study demonstrated that AGS3 protein expression was robustly induced following ischemic renal injury and that the increased protein expression was predominantly localized to the regenerating proximal tubules in the outer medulla. Gpsm1-defective mice were unable to induce AGS3 protein expression following IRI and have a significant impairment in renal tubular repair compared to wild-type Gpsm1^+/+ mice. In all, these findings provide new insight into the biological role of AGS3 in renal epithelial cells and suggest that modulation of AGS3 function or Gβγ activity may be a novel strategy in the development of regenerative therapies for acute kidney injury.