Mesangial cell α_vβ₈-integrin regulates glomerular capillary integrity and repair

Abstract

α_vβ₈-Integrin is most abundantly expressed in the kidney, brain, and female reproductive organs, and its cognate ligand is latent transforming growth factor (LTGF)-β. Kidney α_vβ₈-integrin localizes to mesangial cells, and global β₈-integrin gene (Itgb8) deletion results in embryonic lethality due to impaired placentation and cerebral hemorrhage. To circumvent the lethality and better define kidney α_vβ₈-integrin function, Cre-lox technology was used to generate mesangial-specific Itgb8-null mice. Platelet-derived growth factor-β receptor (PDGFBR)-Cre mice crossed with a reporter strain revealed functional Cre recombinase activity in a predicted mesangial pattern. However, mating between two different PDGFBR-Cre or Ren1^d-Cre strains with Itgb8 ^flox/− mice consistently resulted in incomplete recombination, with no renal phenotype in mosaic offspring. Induction of a renal phenotype with Habu snake venom, a reversible mesangiolytic agent, caused exaggerated glomerular capillary microaneurysms and delayed recovery in Cre^+/− PDGFRB ^flox/− mice compared with Cre^+/− PDGFRB ^flox/+ control mice. To establish the mechanism, in vitro experiments were conducted in Itgb8-null versus Itgb8-expressing mesangial cells and fibroblasts, which revealed β₈-integrin-regulated adhesion to Arg-Gly-Asp (RGD) peptides within a mesangial-conditioned matrix as well as β₈-integrin-dependent migration on RGD-containing LTGF-β or vitronectin matrices. We speculate that kidney α_vβ₈-integrin indirectly controls glomerular capillary integrity through mechanical tension generated by binding RGD peptides in the mesangial matrix, and healing after glomerular injury may be facilitated by mesangial cell migration, which is guided by transient β₈-integrin interactions with RGD ligands.

integrins are extracellular matrix receptors that regulate vital functions, including cell adhesion and migration, and are composed of α- and β-glycoprotein subunits, which heterodimerize through disulfide bridges between extracellular domains. Eighteen α-subunits and eight β-subunits combine to form twenty-four heterodimers. The β₈-subunit partners with the α_v-subunit, and expression is restricted primarily to the kidney, brain, and female reproductive organs (30). Global β₈-integrin gene (Itgb8) or alpha-v integrin gene (Itgav) deletion is embryonic lethal due to placenta and brain angiogenesis defects (28, 29, 32, 50). We previously found that kidney β₈-integrin is localized to mesangial cells (22, 25). Itgb8^−/− mice bred to a genetic background that permitted postnatal survival developed a renal phenotype that included modest albuminuria and azotemia, and the mechanism appeared to be due to indirect effects of mesangial cell β₈-integrin on neighboring glomerular endothelial cells (22). However, few Itgb8^−/− mice lived beyond 6 wk, and the kidney phenotype therefore remained incompletely characterized.

The mesangium is the central region of the glomerulus, which is composed of mesangial cells and the extracellular mesangial matrix. Mesangial cells are modified pericytes, which contact the glomerular basement membrane at the base of the glomerular capillary loops to counteract hemodynamic forces associated with capillary loop distention (24). The most dramatic evidence of this function is derived from platelet-derived growth factor (PDGF)-B or PDGF-β receptor (PDGFBR) knockout mice, which develop glomerular capillary microaneurysms due to the absence of mesangial cells (27, 41). In response to developmental cues or glomerular injury, mesangial cells participate in glomerular basement membrane and extracellular matrix remodeling, phagocytosis, cytokine secretion, migration, and regulation of glomerular circulation and ultrafiltration (11, 18, 37); many of these functions are regulated by integrins (11).

To circumvent the lethality associated with global Itgb8 deletion and better define kidney α_vβ₈-integrin function, Cre-lox technology was used to generate mesangial-specific Itgb8-null mice. Complete recombination was not achieved, using three different Cre mouse strains. However, induction of mesangial injury in β₈-integrin mosaic mice unmasked a glomerular capillary phenotype. When combined with clarifying in vitro experiments, we conclude that mesangial cell α_vβ₈-integrin binds to mesangial matrix ligands and may indirectly regulate glomerular capillary structure and integrity through the generation of mechanotransduction signals and migration to sites of glomerular capillary repair (33).

MATERIALS AND METHODS

Mice.

Itgb8 ^flox/− mice were a gift from Dr. Louis Reichardt (University of California, San Francisco, CA) (32). PDGFRB-Cre mice [PDGFRB-Cre(RA) mice] were a gift from Dr. Ralf Adams (Cancer Research UK London Research Institute) (10). A second strain of PDGFRB-Cre mice, which contain a longer promoter with the transgene [PDGFRB-Cre(VL) mice], was a gift from Dr. Volkhard Lindner (Maine Medical Center Research Institute) (8). Ren1^d-Cre mice were a gift from Dr. Ariel Gomez (University of Virginia) (40). All transgenic mice were backcrossed with C57BL/6 wild-type mice for at least 12 generations to create congenic strains. Cre/+ mice were crossed with Itgb8 ^flox/− mice to generate Cre/+ Itgb8 ^flox/− and Cre/+ Itgb8 ^+/− progeny, which were then intercrossed to produce Cre/+ Itgb8 ^flox/+ and Cre/+ Itgb8 ^flox/− mice for phenotyping. Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J reporter mice were purchased from Jackson Laboratories (Bar Harbor, ME) (9). All protocols were approved by the Institutional Animal Care and Use Committees of Case Western Reserve University.

Genotyping.

All mice were genotyped by PCR analysis using tail DNA according to previously published methods (32). Primers used to identify each allele were as follows: Itgb8 wild-type (250 bp) and floxed allele (370 bp), 5′-GAGATGCAAGAGTGTTTACC-3′ and 5′-CACTTTAGTATGCTAATGATGG-3′; Itgb8-null allele (450 bp), 5′-AGAGGCCACTTGTGTAGCGCCAAG-3′ and 5′-GAGGCATACAGTCTAAATTGT-3′; and Cre (400 bp), 5′-CTGGCAATTTCGGCTATACGTAACAGGGTG-3′ and 5′-GCCTGCATTACCGGTCGATGCAAC-3′.

Immunoblot analysis.

Methods for the immunoblot analysis have been previously described (1). Briefly, cell monolayers were lysed and denatured in boiling buffer [125 mM Tris (pH 6.8), 2% SDS, 5% glycerol, 1% β-mercaptoethanol, and 0.003% bromphenol blue] for 5 min. Samples (20 μg protein/lane) were resolved by SDS-PAGE and transferred to polyvinylidine difluoride membranes. Blots were blocked in 5% dried milk and probed with anti-β₈-integrin IgG (Santa Cruz Biotechnology, Santa Cruz, CA; 1:1,000 in 5% BSA, overnight, 4°C) and then horseradish peroxidase-conjugated IgG (1:10,000, 1 h, room temperature). Band intensity was detected by enhanced chemiluminescence. Blots were exposed to stripping buffer (Thermo Scientific, Waltham, MA; 10 min, room temperature) and then reprobed with anti-α-tubulin IgG (Santa Cruz Biotechnology; 1:3,000, 1 h, room temperature) followed by horseradish peroxidase-conjugated IgG (1:10,000, 1 h, room temperature).

Quantitative RT-PCR.

Functional Itgb8 deletion was verified by measuring β₈-integrin mRNA content by quantitative RT-PCR from microdissected mouse glomeruli, which was normalized to GAPDH expression, according to previously described methods (22, 25). Oligonucleotide primer sequences were as follows: β₈-integrin (66 bp), 5′-ACTTCTCCTGTCCCTATCTCCA-3′ and 5′-ATCTGCCACCTTCACACTCC-3′; and GAPDH (625 bp), 5′-GGAGCCAAACGGGTCATC-3′ and 5′-TGTTGCTGTAGCCGTATTCAT-3′.

Cre reporter mouse.

Cre recombinase activity was screened in double-fluorescent Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J mice (9), which contain a membrane-targeted tdTomato transgene and a stop codon flanked by loxP sites (Fig. 1A). Mice constitutively fluoresce red in all tissues, but in cells expressing active Cre, the dtTomato transgene is deleted, and removal of the stop codon allows expression of downstream, membrane-targeted enhanced green fluorescence protein (EGFP).

Fig. 1.

Cre-lox recombination efficiency assays. A: reporter gene map. B: multiplex PCR strategy to identify one lox, two lox, and wild-type (WT) β₈-integrin alleles. EGFP, enhanced green fluorescent protein.

Recombination efficiency assay.

A multiplex PCR strategy was used to identify the extent of recombination after crosses between the Cre strains and Itgb8-floxed mice using DNA extracted from tails and whole kidneys, according to published methods (48). Upstream primer sequences were 5′-GTGGTTAAGAGCACCGATTG-3′ (F1) and 5′-GAGATGCAAGAGTGTTTACC-3′ (F2); the downstream primer sequence was 5′-CACTTTAGTATGCTAATGATG-3′ (R1). The cycling conditions were as follows: preincubation at 95°C for 10 min and then 45 cycles at 95°C for 10 s (denaturation), 58°C for 45 s (annealing), and 72°C for 1 s (extension). The amplicons from wild-type DNA were predicted to be 1,847 and 247 bp in size; recombined DNA with only one loxP site was predicted to yield a single 340-bp band; unrecombined DNA with both loxP sites intact yielded 2,093- and 360-bp bands (Fig. 1B).

Renal phenotype induction.

Tails of anesthetized mice were warmed with cotton balls immersed in hot water to stimulate vasodilation before an intravenous injection with Habu toxin (Sigma, St. Louis, MO, 2–10 μg/g) or saline vehicle control (3, 4). Mice were monitored twice daily for the ensuing 72 h for signs of distress and then daily thereafter. Mice were euthanized at 2, 7, and 14 days for phenotype evaluation.

Kidney histology.

Kidneys were fixed in formaldehyde, embedded in paraffin, sectioned, deparaffinized, and then stained with periodic acid-Schiff reagent or Masson's trichrome stain. Ki-67 and TUNEL staining experiments were conducted as previously described (22, 34). Glomerular capillary dilation was defined as the percentage of glomeruli (from a total of 50 glomeruli/kidney, 5 kidneys/genotype) containing a capillary that exceeded 25 μm in diameter. Mesangial hypercellularity was defined as clusters of greater than four mesangial cells (16).

Renal function assays.

Albuminuria was measured using the Mouse Albumin ELISA Quantitation Kit (ab108792, Abcam, Cambridge, MA) as previously described (23). Serum and urine creatinine were measured by an enzymatic assay (ab65340, Abcam), which is superior to colorimetric and comparable with HPLC methods in mice (21).

Cell culture.

Mouse mesangial cells and glomerular endothelial cells [gift from Dr. Michael Madaio (Medical College of Georgia) (2)] were harvested and maintained in cell cultures according to previously described methods (22). Temperature-sensitive, conditionally immortalized mouse podocytes were maintained in the undifferentiated, proliferating state at 33°C and induced to differentiate as previously described (42). Mesangial cells derived from Itgb8^−/− mice have previously been characterized (25). CHO-B2/v7 fibroblasts, which are devoid of β-integrins and stably express α_v-integrin, were transiently transfected with β₈-integrin cDNA or empty vector and maintained in cell cultures as previously described (25).

Cell adhesion assays.

The mesangial matrix was isolated using the methods of Langhofer et al. (26). Briefly, mouse mesangial cells in primary culture were grown to confluence, washed with PBS, and then lysed in NH₄OH (20 mM, 15 min, room temperature). The residual extracellular matrix, which remains coated on wells, was washed with sterile water followed by PBS. Adhesion was then assayed in wild-type and Itgb8^−/− mesangial cells and in β₈-integrin-expressing (CHO-B2/v7-β8) and β₈-integrin-null (CHO-B2/v7) fibroblasts. All cell lines were maintained in serum-free media overnight, washed in PBS, and then seeded in 96-well plates (2 × 10⁴ cells/well, 30 min) with or without GRGDSP [Arg-Gly-Asp (RDG)] or GRGESP [Arg-Gly-Glu (RGE)] peptides (100 μg/ml) on the mesangial matrix or poly-l-lysine (PLL; 5 μg/ml). Background was established in wells coated with 0.2% BSA and subtracted from total binding. Adherent cells were fixed in paraformaldehyde, and nuclei were stained with crystal violet, which was eluted with sodium acetate and ethanol and quantitated by spectrophotometry (absorbance wavelength: 550 nm), as previously described (19).

Cell migration assays.

Cell migration in wells coated with 5 μg/ml LTGF-β, latency-associated peptide-β₁ (LAP), vitronectin, or PLL was determined by scratch wound assays, whereby a sterile 20-μl pipette tip was used to create a wound of ∼300 μm width in confluent cell culture monolayers. Images were captured at 0 and 14 h with a Nikon epifluorescence microscope (Tokyo, Japan) in phase-contrast mode at ×10 magnification, and the extent of migration was calculated by counting the number of cells migrating into the wound gap, which was normalized to scratch length. For experiments with mouse mesangial cells, which had a tendency to migrate in sheets, a second measure of migration was also performed. The baseline gap created by the scratch was measured at 10–15 locations over a 1.5- to 2.0-mm distance using Metamorph software (Molecular Devices, Sunnyvale, CA). The distance across the wound was then remeasured at the same locations after 15 h of migration.

Immunocytochemistry.

Cells were evaluated by previously described methods (25). Briefly, CHO-B2/v7 and CHO-B2/v7-β8 fibroblasts were plated on glass coverslips with or without LAP (5 μg/ml). After 2 h, cells were fixed in paraformaldehyde (4%, 10 min, room temperature), blocked, and then permeabilized with 5% BSA in 0.2% Triton X-100. Actin was labeled with Alexa 568-conjugated phalloidin (1:40, room temperature). Coverslips were mounted in 4′,6-diamidino-2-phenylindole-containing Vectashield medium (Vector Laboratories, Burlingame, CA) and viewed with a Nikon epifluorescence microscope. Images were generated with a Spot Digital System camera (Diagnostic Instruments, Sterling Heights, MI) and Image Pro software (Media Cybernetics, Silver Spring, MD).

Statistics.

All experiments were conducted a minimum of three times. Results are expressed as means ± SE. Two-tailed paired t-tests were used for statistical analysis between two groups. Two-way ANOVA was used for comparisons between more than two groups. Statistical significance was defined as P < 0.05.

RESULTS

Efficiency of PDGFRB-Cre activity.

We (22, 25) have previously shown that α_vβ₈-integrin is expressed in a glomerular mesangial pattern by in situ hybridization and immunohistochemistry. We verified by immunoblot analysis that β₈-integrin is not expressed in cell lines derived from the other major resident glomerular cells, podocytes and glomerular endothelial cells (Fig. 2A). To understand the role of α_vβ₈-integrin in kidney mesangial cell function, we used transgenic mice harboring floxed alleles of Itgb8. Floxed mice were then crossed with the PDGFRB-Cre strain because the PDGF-β receptor is abundantly expressed in mesangial cells and pericytes but not in cells of placental and neuroepithelial origin, which are responsible for global Itgb8^−/− embryonic lethality (50). Cre recombinase activity was screened in double-fluorescent Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J mice (9), as described in materials and methods. Figure 2, B–D, shows a glomerulus from the Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J reporter mouse expressing the PDGFRB-Cre(RA) transgene, which showed tdTomato fluorescence in all nonmesangial glomerular cells, whereas EGFP fluorescence was only observed in a mesangial distribution.

Fig. 2.

Platelet-derived growth factor-β receptor (PDGFBR)-Cre activity in a mesangial pattern. A: β8-integrin immunoblot of cell culture lysates from mouse mesangial cells (MMC), undifferentiated and differentiated mouse podocytes (Podo), and mouse glomerular endothelial cells (MGEC). PDGFBR-Cre mice were crossed with floxed reporter mice. B: the red tag marks nonrecombined cells. C: the green tag marks cells where recombination has occurred. D: merged image.

Itgb8 deletion in only a fraction of mesangial cells.

Because previous studies indicated that mesangial Cre activity was robust, it was surprising to consistently detect β₈-integrin mRNA in isolated glomeruli from PDGFRB-Cre(RA)Itgb8 ^flox/− (Fig. 3A) as well as PDGFRB-Cre(RA)Itgb8 ^flox/flox mice (not shown). To determine whether these results were specific to crosses involving PDGFRB-Cre(RA) mice, we tested two different Cre strains, each driven by a mesangial-selective promoter [PDGFRB-Cre(VL) and Ren1^d-Cre mice], and noted similar residual β₈-integrin mRNA levels (Fig. 3A). Since mice with the Itgb8 ^flox/− genotype have only one functional Itgb8 gene, we hypothesized that the source of β₈-integrin mRNA in Cre-expressing animals is unrecombined, floxed, Itgb8. To investigate recombination efficiency, we used a PCR-based method, which determines the proportion of functional (two loxP sites preserved) versus truncated (collapsed to one loxP site) alleles in Cre/+ mice (48). Figure 3B, top, shows amplicons corresponding to both one lox and two lox alleles in tails from PDGFRB-Cre(RA)Itgb8^flox/− mice. Figure 3B, bottom, shows the persistence of the unrecombined floxed β₈-integrin allele in isolated glomeruli from PDGFRB-Cre(RA)Itgb8^flox/− mice. Since each diploid Itgb8 ^flox/− cell contains a null allele paired with either a floxed or recombined β₈-integrin allele, tissue showing three PCR bands would be composed of two cell types: β₈-integrin floxed/null or β₈-integrin recombined/null, thus generating mosaic tissue in these animals.

Fig. 3.

Floxed β₈-integrin is incompletely deleted by Cre recombinase. A: β₈-integrin mRNA expression was determined by quantitative RT-PCR in isolated glomeruli derived from crosses between floxed β₈-integrin and three different Cre mouse strains. B: to investigate incomplete recombination as a possible mechanism, a PCR-based strategy was used, whereby complete recombination collapses the target gene to one loxP site (1 lox), whereas the unrecombined target gene retains both loxP sites (2 lox). Results from two mice per genotype revealed incomplete recombination in PDGFRB(VL)-Cre/+ flox/− mice and variable, but sometimes complete, recombination in PDGFRB(VL)-Cre/+ flox/+ mice.

Embryonic lethality in Cre-lox mice.

Of the crosses between floxed mice and the three Cre strains, PDGFRB-Cre(RA)Itgb8^flox/− and Ren1^d-Cre Itgb8 ^flox/− progeny were observed in normal Mendelian ratios, whereas PDGFRB-Cre(VL)Itgb8 ^flox/− progeny were born at less than the expected frequency (Table 1). Cre(VL)/+ flox/− survivors demonstrated incomplete recombination of the floxed allele, and, interestingly, complete recombination was achieved in some Cre(VL)/+ mice, but only if the floxed allele was accompanied by a wild-type copy of β₈-integrin providing functional mRNA, independent of the Cre-lox system. The data shown in Table 1 and Fig. 3 suggest that absence of β₈-integrin function in a PDGFBR-Cre-expressing cell lineage leads to embryonic lethality. We proceeded by determining if there was a phenotype difference in PDGFRB-Cre(VL) flox/− versus PDGFRB-Cre(VL) flox/+ mice, hereafter designated as Cre/+ flox/− and Cre/+ flox/+, respectively.

Table 1.

Genotypes of progeny from crosses between floxed β₈-integrin gene and three different Cre strains

	PDGFRB(RA)-Cre Mice		Ren1d-Cre Mice		PDGFRB(VL)-Cre Mice
	Cre/+ flox/+	Cre/+ flox/−	Cre/+ flox/+	Cre/+ flox/−	Cre/+ flox/+	Cre/+ flox/−
Anticipated	84	84	44	44	65	65
Observed	74	94	50	38	98	31*
Complete recombination	0	0	0	0	67	0*

PDGFBR, platelet-derived growth factor-β receptor.

^*P < 0.001 by χ²-test.

Glomerular phenotype in Cre/+ flox/− mice.

Significant differences in baseline renal function were not observed between Cre/+ flox/− and Cre/+ flox/+ mice, as determined by serum creatinine (Fig. 5B), albuminuria (Fig. 5C), or histology (not shown). Because kidney α_vβ₈-integrin expression is restricted to mesangial cells, Cre/+ flox/− and Cre/+ flox/+ mice were treated with Habu snake venom to induce mesangiolysis and a renal phenotype (3, 4). Dose finding experiments in wild-type mice revealed that 10 μg/g was lethal within 24 h, whereas concentrations of 2–6 μg/g yielded modest but discernible glomerular lesions (Fig. 4). Mice were therefore treated with 2 μg/g Habu toxin to allow phenotype accentuation in conditional mutants. All mice lost weight after they received Habu toxin, but weight loss was more pronounced in Cre/+ flox/− mice compared with Cre/+ flox/+ mice (Fig. 4A). Interstitial fibrosis was rarely observed in Cre/+ flox/− kidneys (Fig. 4B), and the major lesion was glomerular capillary dilation (Fig. 4C), which was observed significantly more frequently in Cre/+ flox/− mice and persisted for up to 2 wk after Habu administration (Fig. 4D). During the recovery phase, mesangial hypercellularity was less in Cre/+ flox/− kidneys compared with Cre/+ flox/+ kidneys (Fig. 4, E and F).

Fig. 5.

Renal function in Habu-treated mice. Renal function was determined by serum creatinine (A and B) or by the urine albumin-to-creatinine ratio (C). Data in A were generated 2 days after Habu venom or saline control injection in Cre/+ flox/+ and Cre/+ flox/− mice. Data in B and C were generated at the indicated time points after Habu toxin or saline injection.

Fig. 4.

Phenotypes in Habu venom-injected mice. A: body weights in Cre/+ flox/+ and Cre/+ flox/− mice after Habu injection. B: a Masson's trichrome-stained Cre/+ flox/− kidney revealed interstitial fibrosis 14 days after Habu injection. C: glomerular capillary dilation in a Cre/+ flox/− kidney 7 days after Habu injection. D and E: glomerular capillary dilation was quantitated in Cre/+ flox/+ and Cre/+ flox/− mice at the indicated time points. E: mesangial hypercellularity in a Cre/+ flox/− kidney 7 days after Habu injection. F: mesangial hypercellularity was quantitated in Cre/+ flox/+ and Cre/+ flox/− mice at the indicated time points. Results are expressed as means ± SE. *P < 0.05 compared with Cre/+ flox/+ kidneys at the same time point.

Renal function in Habu toxin-treated mice.

Habu toxin administration caused a modest rise in serum creatinine compared with saline control (Fig. 5A), but the difference between Habu-treated Cre/+ flox/− and Cre/+ flox/+ mice was not significant at any time point (Fig. 5B). A surrogate marker of glomerular function, albuminuria, also did not differ between Habu-treated Cre/+ flox/− and Cre/+ flox/+ mice (Fig. 5C).

β8-Integrin effects on apoptosis and proliferation.

The mechanism of healing after Habu-induced glomerular injury involves initial mesangial cell proliferation, which is followed by mesangial cell apoptosis, to prevent persistent hypercellularity (15). To investigate the mechanism for impaired recovery from Habu toxin-induced glomerular capillary dilation in Cre/+ flox/− mice, we determined whether diminished mesangial hypercellularity is attributed to decreased hyperplasia or increased apoptosis by Ki-67 and TUNEL staining, respectively. Figure 6 shows that there were no differences in proliferation or apoptosis between Cre/+ flox/+ and Cre/+ flox/− glomeruli at 2 or 7 days after Habu toxin injection.

Fig. 6.

β₈-Integrin effects on apoptosis and proliferation. Mice were evaluated for mesangial cell proliferation by Ki-67 labeling (A) or apoptosis by TUNEL (B) 2 and 7 days after Habu injection.

β₈-Integrin regulates adhesion to the mesangial matrix.

To test whether cell migration, a previously described β₈-integrin function (19) and a process by which mesangial cells populate glomeruli during development and repair (18, 27), could be impaired and therefore responsible for delayed glomerular healing in Habu-treated Cre/+ flox/− mice, we first established that α_vβ₈-integrin-expressing cells bind canonical RGD peptide ligands within the mesangial matrix. Mesangial cells derived from Itgb8^+/+ or Itgb8^−/− mice were cultured on plastic, grown to confluence, and then removed, leaving the deposited extracellular matrix behind (see materials and methods). Itgb8^+/+ or Itgb8^−/− mesangial cells were then replated, and adhesion to the residual matrix was assayed in the presence or absence of RGD or control RGE peptides. Figure 7, A and B, shows the increased adherence of wild-type compared with Itgb8^−/− mesangial cells. Adhesion was significantly inhibited by RGD but not by RGE peptide preincubation. Similar results were observed for adhesion to the matrix deposited by Itgb8^−/− mesangial cells (Fig. 7, C and D), suggesting that β₈-integrin is not required for synthesis or deposition of the mesangial matrix. Of note, Itgb8^+/+ mesangial cells assumed a more spread morphology compared with Itgb8^−/− mesangial cells (Fig. 7, E and F).

Fig. 7.

β₈-Integrin regulates adhesion to the mesangial matrix. A–F: mesangial cells derived from Itgb8^+/+ and Itgb8^−/− kidneys were seeded in 24-well plates coated with an extracellular matrix deposited by Itgb8^+/+ (A, B, E, and F) or Itgb8^−/− (C and D) mesangial cells with or without GRGDSP [Arg-Gly-Asp (RDG)] or GRGESP [Arg-Gly-Glu (RGE)] peptides. Adherent cells were fixed in paraformaldehyde, and nuclei were stained with crystal violet, which was eluted and quantitated, as described in materials and methods. A and C: representative plates with samples in duplicate. B and D: quantification of adhesion. E and F: representative phase-contrast microscopy fields for Itgb8^+/+ (+/+; E) and Itgb8^−/− (−/−; F) mouse mesangial cells at ×40 magnification. G: CHO-B2/v7 and CHO-B2/v7-β8 cells were maintained in serum-free media overnight, washed in PBS, and then seeded in 96-well plates with or without RGD or RGE peptides on a matrix deposited by wild-type mesangial cells or poly-l-lysine (PLL). Adhesion was assayed as described in A–D. * P < 0.05 compared with cells on a mesangial matrix without RGD peptide.

Mesangial cells express other α_vβ₈-integrin partners, which recognize RGD sequences, consistent with incomplete inhibition of adhesion with RGD peptide preincubation. To confirm that α_vβ₈-integrin mediates adhesion to the mesangial matrix, adhesion assays were repeated using CHO-B2v7 fibroblasts, which are devoid of β-integrin subunits. Figure 7G shows that CHO-B2v7 cells transfected to express β₈-integrin (CHO-B2/v7-β8) adhered better than β₈-integrin-null CHO-B2/v7 cells to the wild-type, mesangial-conditioned matrix, implying the presence of a β₈-integrin ligand within the mesangium. Figure 7G also shows that CHO-B2/v7-β8 binding was inhibited by RGD peptide preincubation to levels observed with PLL, which supports adhesion through integrin-independent pathways. CHO-B2/v7 adhesion was not inhibited by RGD, and neither cell line was affected by RGE peptide competition.

β8-Integrin regulates migration on RGD-containing matrix.

To simulate the impact of α_vβ₈-integrin on recovery from Habu injury, cell migration was measured in Itgb8^+/+ and Itgb8^−/− mesangial cells as well as CHO-B2/v7 and CHO-B2/v7-β8 fibroblasts plated in wells coated with the α_vβ₈-integrin ligand LTGF-β. Figure 8, A and B, shows the enhanced migration in wild-type compared with Itgb8^−/− cells. Similar disparities were observed between Itgb8^+/+ and Itgb8^−/− mesangial cell migration on vitronectin (Fig. 8, C and D), an alternative RGD ligand for α_vβ₈-integrin that is expressed exclusively in diseased glomeruli (36, 39). Cells migrated less well on vitronectin compared with LAP, perhaps reflecting the relatively weaker affinity of vitronectin for β₈-integrin (31). Figure 8, E and F, shows the negligible migration on PLL. Mesangial cell migration was quantitated by two different methods. Figure 8G shows the number of cell migrating into the scratch over 15 h. Because mesangial cells tended to migrate in sheets, data are also expressed as mean distances of migration into the wound (Fig. 8H).

Fig. 8.

Mesangial cell β₈-integrin regulates migration on a RGD-containing matrix. A–F: scratch migration assays for wild-type (Itgb8^+/+; A, C, and E) and Itgb^−/− (B, D, and F) mesangial cells plated on latent transforming growth factor (LTGF)-β (A and B), vitronectin (VN; C and D), or PLL (E and F), all at 5 μg/ml, for 15 h. G and H: quantification of the scratch migration data. Results are representative of 5 scratches/well. *P < 0.05 compared with the similarly treated Itgb8^−/− group. Bar = 300 μm.

Mesangial cell data were confirmed in CHO-B2/v7-β8 and CHO-B2/v7 fibroblasts (Fig. 9, A–F) and quantitated (Fig. 9G). Of note, CHO cells migrated much more robustly compared with mesangial cells and more commonly as individual cells rather than in sheets. LAP-stimulated CHO-B2/v7-β8 cells exhibited prominent lamellipodia (Fig. 9H), which was lacking in cells maintained on glass (Fig. 9I). These morphological changes are consistent with β₈-integrin-regulated Rac1 signaling (25), which is associated with lamellipodia formation in motile cells (14).

Fig. 9.

β₈-Integrin regulates migration on a RGD-containing matrix. A–F: scratch migration assays for fibroblasts expressing β₈-integrin (CHO-B2/v7-β8; A, C, and E) and devoid of β₈-integrin (CHO-B2/v7; B, D, and F) plated on latency-associated peptide-β₁ (LAP; A and B), VN (C and D), or PLL (E and F), all at 5 μg/ml, for 14 h. Lines represent the width of the scratch at time 0. G: quantification of the scratch migration data. Results are representative of 3 scratches/well. *P < 0.05 compared with CHO-B2/v7 cells. H and I: morphology of CHO-B2/v7-β8 cells on LAP (5 μg/ml, 2 h; H) or glass coverslips (I). Actin was stained in red with phalloidin. Arrows mark lamellipodia.

DISCUSSION

In this report, we evaluated conditional knockout mice to determine the role of β₈-integrin in mesangial cell function. The major finding is that β₈-integrin influences glomerular capillary morphology and repair after mesangial injury. Using in vitro approaches with β₈-integrin-expressing and null cells, we confirmed the preference of β₈-integrin for RGD peptide ligands, including within the mesangial matrix. To reconcile the in vivo and in vitro results, we demonstrated that mesangial cell β₈-integrin binds ligands within mesangial matrix and that this process is likely to regulate glomerular capillary architecture in a homeostatic fashion. This paradigm extends the interpretation of our previous study (22) implicating mesangial β₈-integrin as a governor of TGF-β release to juxtaposed glomerular endothelial cells. In addition, our data support the notion that during recovery from glomerular injury, β₈-integrin-expressing mesangial cells migrate to areas of capillary repair, using paths provided by mesangial matrix ligands (Fig. 10).

Fig. 10.

Schematic diagram of the β₈-integrin influence on glomerular capillary injury and repair. The reduced Cre-lox recombination in Cre/+ flox/− mice, resulting in mosaicism, is shown by the greater number of mesangial cells, which do not express β₈-integrin (shown in red). The reduced number of β₈-integrin-expressing mesangial cells (shown in green), causes enhanced capillary dilation with injury and delayed mesangial cell migration during the recovery phase in Cre/+ flox/− glomeruli. GEC, glomerular endothelial cell.

Cre/+ flox/− mice expressed 40–60% as much glomerular β₈-integrin mRNA as Cre/+ flox/+ mice, which was sufficient to prevent a renal phenotype. However, spontaneous and induced phenotypes are influenced by genetic background, and congenic C57BL/6 background mice, which were used for all experiments, tend to be disease resistant (12, 20). Rather than breeding mice to a more permissive background, we opted to induce a phenotype after mesangial injury. Habu toxin was chosen to stimulate mesangiolysis, which reliably causes widespread glomerular capillary microaneurysms in mice with prior unilateral nephrectomy (4, 15). In our experiments, the nephrectomy step was omitted to cause a milder baseline lesion, which permitted evaluation of disease accentuation. Using this model, β₈-integrin-deficient mice developed more severe glomerular capillary microaneurysms and delayed repair of these lesions. This β₈-integrin function in recovery may be vital, since the aftermath of mesangiolysis and glomerular capillary injury has been implicated in glomerulosclerosis (43) and diabetic nodular sclerosis (47). A similar role has been proposed for α₈β₁-integrin (13, 15), suggesting that there could be cooperation and/or redundancy among mesangial integrins.

α_vβ₈-Integrin has been assigned a variety of biological functions, which could be relevant to glomerular capillary repair, including cell migration (19), proliferation (7), angiogenesis (6, 50), adhesion to the extracellular matrix (45), and regulation of cytokines (22, 31). Our data demonstrated that adherence to the mesangial matrix was dependent on ligation with RGD peptide-containing proteins, and, although RGD ligation is a feature of all α_v-integrins, our results are specific to α_vβ₈-integrin, since CHO-B2v7 cells do not express β-integrin subunits that partner with α_v-integrin subunits (49). β₈-Integrin regulated adhesion to the mesangial matrix, in agreement with mesangial cell migration as the predominant mechanism in glomerular capillary remodeling after Habu-induced injury (4). Specific ligands tested for β₈-integrin-dependent migration were LAP and vitronectin, both of which localize to injured glomeruli (22, 36). The in vitro migration on LTGF-β or LAP had not previously been described and suggests that LTGF-β imbedded within the extracellular matrix may provide guidance cues for migrating mesangial cells during recovery from glomerular injury.

Cre-lox technology represents a valuable technique to delineate specific gene functions, but there are limitations that must be considered. A frequent issue is incomplete recombination, resulting in mosaic mice. This was first suspected in our experiments when β₈-integrin transcript persistence was observed in PDGFRB-Cre(RA)Itgb8 ^flox/− mice and then confirmed using PCR-based strategies to assess recombination efficiency. The most common explanation for deficient recombination is weak Cre activity (35). However, in experiments with the Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J reporter mouse, Cre activity, driven by a PDGFRB promoter, corresponded to a mesangial pattern. That mosaic mice resulted from crosses between floxed Itgb8 and three different Cre strains suggests that the target allele is resistant to recombination. Indeed, significantly different recombination frequencies have been described between the same Cre and different floxed alleles as well as with the same Cre-lox combination in different tissues (46). Because recombination was routinely better when the accompanying allele was wild type rather than null (Fig. 1E), we propose that the mechanism could be due to an altered local chromatin structure (46).

The observation that Itgb8 ^flox/− progeny resulting from breeding with PDGFRB-Cre(VL) mice were born in reduced numbers implies that a minimum Itgb8 dose in PDGFBR-expressing cells is required to sustain survival. We speculate that cells capable of bypassing PDGFBR-Cre-mediated recombination of the floxed β₈-integrin allele gain a competitive advantage over cells in which floxed β₈-integrin recombination is achieved, a mechanism for which there is precedence (5). Since we did not observe a kidney development phenotype in global β₈-integrin knockout mice (22) and renal dysfunction does not generally cause embryonic lethality, we reason that the relevant, vital PDGFBR-expressing cells are located in nonrenal tissues.

The alternative possibility that persistent glomerular β₈-integrin expression stems from nonmesangial β₈-integrin expression seems unlikely, since β₈-integrin mRNA and protein expression correlated with a mesangial pattern in vivo (22, 25), and neither β₈-integrin mRNA (unpublished observations) nor protein (Fig. 1A) expression was detected in podocytes or glomerular endothelial cells, the other major resident cells of the glomerulus. However, α_vβ₈-integrin is expressed in dendritic cells (44), which may be rarely present in the kidney, though much more commonly in the tubulointerstitium than glomerulus (17, 22, 38).

In this study, which combines in vivo and in vitro techniques to define kidney α_vβ₈-integrin function, we conclude that the β₈-integrin indirectly controls glomerular capillary integrity, which we speculate occurs through mechanical tension generated by binding RGD peptides in the mesangial matrix. Healing after glomerular injury may be facilitated by mesangial cell migration, which is guided by transient β₈-integrin interactions with RGD ligands.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01-DK-067528 and R01-DK-072348.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: S.L.-R., S.K., and J.R.S. conception and design of research; S.L.-R., V.L., T.D.A., and S.K. performed experiments; S.L.-R., T.D.A., and J.R.S. analyzed data; S.L.-R., T.D.A., and J.R.S. interpreted results of experiments; S.L.-R. and J.R.S. prepared figures; S.L.-R. and J.R.S. drafted manuscript; S.L.-R., S.K., and J.R.S. edited and revised manuscript; S.L.-R., V.L., T.D.A., S.K., and J.R.S. approved final version of manuscript.