Targeting αvβ8 Integrin with mAb MEDI8367 Prevents Fibrosis in Preclinical Models of CKD

Visual Abstract

Abstract

Key Points

• αvβ8 integrin was upregulated in kidney tubules and associated with fibrosis and reduced kidney function in human CKD.

• MEDI8367 blocked αvβ8-mediated TGFβ activation and reduced fibrosis in mouse unilateral ureteral obstruction model.

• In diabetic kidney disease mouse models, αvβ8 inhibition improved kidney injury markers and preserved kidney function.

Background

CKD is a global health issue exacerbated by the rising prevalence of diabetes and obesity. Renal fibrosis, characterized by the accumulation of extracellular matrix proteins, is a critical factor in CKD progression. TGF-β, a key profibrotic cytokine, plays a pivotal role in this process. However, systemic inhibition of TGF-β has been limited by associated toxicities.

Methods

This study explores the role of αvβ8 integrin in renal fibrosis and its potential as a therapeutic target in CKD. We used various preclinical models of CKD, including humanized αvβ8 mice and the db/db uninephrectomy model, to investigate the role of αvβ8 integrin in renal fibrosis. Gene set variation analysis was used to assess fibrotic gene signatures in human kidney biopsies. The therapeutic potential of MEDI8367, a monoclonal antibody targeting αvβ8 integrin, was evaluated in vitro and in vivo.

Results

Our findings demonstrate that αvβ8 integrin is upregulated in the tubulointerstitium of CKD kidneys, particularly in diabetic kidney disease, and correlates with TGF-β activation and renal function decline. MEDI8367 effectively inhibited αvβ8-mediated TGF-β activation in vitro and attenuated murine unilateral ureteral obstruction-induced renal fibrosis. Notably, inhibition of αvβ8 reduced kidney damage and improved kidney function in models of diabetic kidney disease and hypertensive nephropathy.

Conclusions

Our study highlights the potential of MEDI8367 to mitigate renal fibrosis and improve kidney function, offering a novel approach to CKD treatment that complements existing therapies.

Introduction

CKD is becoming a worldwide epidemic, driven by the increase in the prevalence of diabetes and obesity; hence, novel targets and treatments for CKD are urgently needed. Renal fibrosis is a major component of disease progression in CKD and is characterized by the accumulation of extracellular matrix (ECM) proteins in the glomerulus and tubulointerstitium (TI), leading to the destruction of the organ architecture and impairment of kidney function.¹

TGF-β is a key profibrotic cytokine that drives matrix synthesis, inhibits matrix degradation, and promotes myofibroblast activation. It is one of the main drivers of tubulointerstitial fibrosis.^1,2 Targeting TGF-β is associated with adverse events owing to the pleiotropic function of TGF-β signaling. Pan-TGF-β inhibitors have been in development for decades; however, associated toxicities were identified in the preclinical and clinical settings, limiting their efficacy at doses that were safe, or a lack of efficacy impeded their development.^3–5 TGF-β is secreted as a latent complex bound to latency-associated peptide (LAP) and latent TGF binding proteins, and extracellular activation is required for TGF-β activity.⁶ The αv integrins have a unique ability to activate latent TGF-β1 and TGF-β3 isoforms through the binding to the RGD domain of the LAP, leading to the release of active TGF-β and therefore, modulating fibrotic processes.⁶ Consequently, integrins have emerged as promising therapeutic targets in many fibrotic diseases, including kidney.⁷

Tissue-specific expression of integrins determines the activation of TGF-β in specific organs. Mice lacking the function of both αvβ6-and αvβ8-integrin display abnormalities that recapitulate the major abnormalities in Tgfb1-null mice and Tgfb3-null mice, highlighting their importance during in vivo TGF-β activation.⁸ It has been reported that αvβ8 has a higher affinity to the RGD binding domain in the LAP compared with any other integrin,⁹ therefore, making it more specific for TGF-β activation than other cell adhesion functions.^10,11 Furthermore, unlike other integrins, αvβ8 has been described to be constitutively active because of its extended-closed conformation, enabling extracellular ligand interactions.^12,13

Although the role of TGF-β in driving renal fibrosis in CKD is well established, the role that αvβ8 integrin plays has not been thoroughly explored. In our study, data show αvβ8 is a kidney-enriched integrin and is upregulated in the tubules in human diabetic kidney disease (DKD), where it colocalizes with fibrotic areas. Correlation of αvβ8 integrin expression with TGF-β activation and renal function decline was also observed in CKD kidney biopsies. We describe the use of MEDI8367 our lead anti-αvβ8 molecule, derived from previously described 37E1B5.¹⁴ This molecule demonstrated in vivo efficacy in the unilateral ureteral obstruction (UUO) preclinical model of renal fibrosis in humanized αvβ8 mice. Moreover, modulation of αvβ8 integrin reduced kidney damage in the db/db uninephrectomy model of DKD and the 5/6 nephrectomy model of hypertensive nephropathy, demonstrating the role of αvβ8 in the development of kidney fibrosis. Overall, our results suggest targeting αvβ8-integrin-mediated local TGF-β-activation may represent a promising approach for the treatment of DKD.

Methods

Targeted Analysis of Gene Expression in Human Kidney Biopsies

Human renal biopsies from the European Renal cDNA Bank (ERCB) cohort were microdissected for glomeruli (GM) and tubulointerstitial compartment separation and profiled by microarray analysis as described previously.^15,16 Gene set variation analysis (GSVA) was performed in this dataset, and a signature score (scaled as –1≤×≤1) was calculated on the basis of expression data from a predefined gene set.¹⁷ ITGB8, epithelial-mesenchymal transition (EMT)-related genes, and collagen expression in biopsies from living kidney donors were compared with expression in biopsies from patients with various CKD etiologies.

Data from the Karolinska Institute/Sahlgrenska University Hospital (KI/SU) human kidney cohort were used to analyze inflammatory marker expression in patients with DKD.¹⁸ Briefly, RNA sequencing was performed on paired, microdissected glomerular and tubulointerstitial tissue from patients diagnosed with DKD (n=19; 15 male patients; median age, 61 [range: 30–85] years; CKD stages 1–4) and living kidney donors (n=20; 12 males; median age: 56 [30–70] years), as previously described.¹⁸ To identify EMT markers and proteins that were significantly differentially expressed in GM and tubulointerstitium, we adjusted the P values for multiple hypothesis testing using the Benjamini–Hochberg method. The threshold for significance was set at an adjusted P value of <0.05.

Immunohistochemistry of Human Kidney Biopsies

Formalin-fixed paraffin-embedded 3–4 µm human kidney sections (n=3) were stained using an anti-αvβ8 antibody (Calico labs) on an automated immunohistochemistry robot (Flatbed Autostainer Plus, Dako; Santa Clara, CA).

Active TGF-β Estimation in Human Urine Samples

Free active TGF-β levels in urine samples of healthy individuals and patients with CKD were measured using the LEGEND MAX Free active TGF-β1 ELISA kit (Biolegend) and Simoa assay kit (Quanterix) following the manufacturer's instructions.

In Vitro Experiments

αvβ8 Binding ELISA

Recombinant human αvβ-1, -3, -5, -6, -8 were diluted in PBS (2 μg/ml) and coated on Nunc 96-well plates overnight at 4°C and blocked with 3% BSA for 1 hour at room temperature (RT). After washing, MEDI8367 was incubated for 1 hour. After three washes in PBS, peroxidase-conjugated anti-human Fc (1:20,000) was used as the detection antibody. The wells were washed three times, and TMB substrate solution (KPL) was added. The reaction was stopped with 2 N sulfuric acid, and the absorbance at 450 nm was measured. Mouse binding was detected with peroxidase-conjugated anti-mouse Fc (1:20,000).

LAP and αvβ8 Binding Inhibition

Recombinant human LAP diluted in PBS (4 μg/ml) was coated on a Nunc 96-well plate overnight at 4°C and blocked with 3% casein (Thermo Scientific) for 2 hours at RT. After washing, 50 μg/ml αvβ8 was mixed with serially diluted MEDI8367 and added to the LAP-coated wells. After 2 hours RT incubation, the plate was washed three times in PBS. The biotinylated-anti-αv antibody (1:1000) was used to detect αvβ8 binding and incubated at RT for 1 hour. After washing, peroxidase-conjugated streptavidin (1:3000) was added for 30 minutes. The wells were then washed three times, and TMB substrate solution (KPL) was added. The reaction was stopped with 2 N sulfuric acid, and the absorbance at 450 nm was measured.

TGF-β Bioassay

TGF-β activation was measured using transformed mink lung epithelial cells stably transfected with a portion of the plasminogen-activated inhibitor 1 promoter linked to a luciferase reporter (cells provided by Daniel Rifkin, New York University) and cultured as described.¹⁹ HeLa-B8 cells were cocultured with TMLCs in a 96-well plate overnight in DMEM high glucose (Life Technologies/Thermo Fisher) supplemented with 10% FBS and 10 U/ml Penicillin G, 10 μg/ml streptomycin G sulfate, with or without test antibody. After 16 h, supernatants were removed, and cells were lysed in 100 µl of cell lysis buffer (Promega) and luciferase activity determined using the luciferase assay system (Promega). Samples were read immediately on a luminometer and shown as either relative luciferase units or percent maximal response, determined by using TMLCs alone as the baseline or 0% control and TMLCs cocultured with HeLa-B8 cells as maximal or 100% response in the assay.

Animal Studies

5/6 Nephrectomy CKD Model

Eighteen male C57Bl6 and 12 humanized avB8 transgenic mice of 6–8 weeks of age were purchased from MRC Harwell. Animals were allowed ad libitum access to food (Harlan 8640 rodent chow) and water, group-housed under standard conditions, and allowed to acclimate for at least 1 week before study inception. All studies were approved by the Institutional Animal Care and Use Committee. Before the surgery, mice were stratified by body weight and then randomly assigned to different treatment groups. Anesthesia was induced with isoflurane 3%–4% and maintained at 2.5% with an oxygen flow of 1.5  L/min during the whole procedure. Animals underwent 5/6 nephrectomy or sham surgeries in a 2-stage procedure, as previously described at Plato BioPharma.²⁰ At stage 1, the right kidney poles were removed by excision; at stage 2, after 2 weeks of recovery, left nephrectomy was performed. Subtotal nephrectomy mimics the progressive renal failure after loss of renal mass in humans. After surgery, mice were allowed to recover in a clean, heated cage before being returned to normal vivarium conditions. All mice received one injection (subcutaneous [s.c.]) of 0.05 mg/kg buprenorphine immediately after completion of surgery. Mice were sacrificed 12 weeks after left nephrectomy, and samples were collected for analysis.

UUO Fibrosis Model

Thirty-two male humanized αvβ8 transgenic mice aged 7–8 weeks were purchased from MRC Harwell. Animals were allowed ad libitum access to food (Harlan 8640 rodent chow) and water, group-housed under standard conditions, and allowed to acclimate for at least 1 week before study inception. Mice were randomized on the basis of body weight 2 days before UUO surgery. The experiment was designed with two arms: Mice underwent UUO surgery on day 0, and the experiment was terminated after 5 days. On day 0, all mice were anesthetized with isoflurane on a nosecone, and then a laparotomy (sham operation) was performed for the control group, while all other disease induction groups underwent permanent right unilateral UUO surgery using heat sterilized instruments and aseptic surgical technique. Mice were allowed to recover in a clean, heated cage before being returned to normal vivarium conditions. All mice received one injection (s.c.) of 0.05 mg/kg buprenorphine immediately after completion of surgery. Mice received IP administration of NIP228 (10 mg/kg; isotype control), TGF-β neutralizing antibody (1D11; 3 mg/kg), or MEDI8367 (10 mg/kg) every other day starting the day before UUO surgery.

Db/db Uninephrectomy DKD Model

Thirty-six males at age of 5–6 weeks were acquired from Charles River Italy. Animals were allowed ad libitum access to food (5008 diet from Labdiet) and water, group-housed under standard conditions, and allowed to acclimate for at least 1 week before study inception. Before the surgery, mice were randomized by body weight into different treatment groups. On day 0, all mice were anesthetized with isoflurane on a nosecone. Then, the mice underwent uninephrectomy of the left kidney by making a 1–2 cm incision site. The adrenal gland and kidney capsule were separated carefully from the kidney before the kidney was removed completely using sterilized instruments and aseptic surgical technique. Mice were then allowed to recover in a clean, heated cage before being returned to normal vivarium conditions. All mice received one injection (s.c.) of 0.05 mg/kg buprenorphine immediately after completion of surgery. 3G9 (αVβ6 antibody), RAD264 (αVβ6/8 antibody), and control antibody were dosed at 20 mg/kg (s.c.) twice weekly. All mice were terminated at 10 weeks postnephrectomy.

Urine Albumin/Creatinine and TGF-β, Phospho-Smad2/3

To collect urine samples, mice were housed in metabolic cages with free access to food and water for an 18-hour period. To reduce evaporation, the urine collectors were placed in a prechilled plastic thermoblock (Tecniplast; Buguggiate VA, Italy). Urinary creatinine and albumin concentrations were measured using the Cobas C111 analyzer (Roche; Basel, Switzerland). Total and phospho-Smad2/3 were measured in either urine or kidney lysates using the PathScan Phopsho-Smad2/3 sandwich ELISA kit (Cell Signaling Technology).

Immunohistochemistry of Murine Kidneys

Kidneys were collected at study termination, fixed in 10% neutral buffered formalin for 48 hours, and paraffin-embedded. Fixed 3- to 4-μm sections were immunostained with anti-αvβ8 antibody (Calico Labs) for estimating the expression of αvβ8 in various kidney regions. Analysis was performed by using the Aperio ePathology digital image analysis software. Images (n=10/depth per kidney) of PSR stained renal cortex were obtained in a blinded manner using a Zeiss AxioImager.A2 at 200× (sufficient to obtain a representation of approximately 60%–70% of the renal cortical area) and subjected to quantitative image analysis. Collagen volume fraction (percent total area imaged) was expressed as the average positive stain across sampled images from right (sham or obstructed) kidneys.

Statistics

Data are presented as mean±SD, mean±SEM, or as median±1.5 quartile when data are visualized as a boxplot. Statistical analysis was performed using GraphPad Prism 9.0 software or R v3.3.2 and ggpubr package. Correlation analysis was performed using Spearman correlation. One-way (Kruskal–Wallis) or a two-way ANOVA was performed to compare experimental groups. Statistical significance was depicted as *P < 0.05, **P < 0.01, ***P < 0.001, or ****P< 0.0001. Signature scores were calculated by the GSVA.¹⁷ For gene and signature contrasts to a living donor (LD) baseline, a Wilcoxon test was applied and adjusted by Bonferroni correction for multiple tests. This procedure was also applied to analyze gene expression across CKD etiologies contrasted with LD.

Results

Associations between Urinary Active TGF-β Levels, Fibrotic Gene Signatures in Kidneys, and Decline in eGFR among Patients With CKD

TGF-β is known to drive fibrotic responses through EMT and collagen deposition. Accordingly, we found that urine active TGF-β levels were elevated in patients with CKD. Active TGF-β/creatinine ratio was increased by 4.4-fold (P = 0.0003) in participants with CKD (median: 7.5 µg/g, N=20) compared with non-CKD controls (1.4 µg/g, N=20; Figure 1A). Furthermore, using human kidney biopsies from the ERCB database with different CKD etiologies, we used GSVA to show increased expression of EMT-related genes as well as collagen in both GM and TI of patients with CKD compared with LD controls. This elevation in fibrosis gene profiles was particularly notable from CKD stage 3 (Figure 1B). To reinforce the involvement of TGF-β signaling in these responses, we derived a TGF-β activation signature from 30 downstream genes (Supplemental Table 1) across the ERCB GM and tubulointerstitium and the Clinical Phenotyping Resource and Organ Biobank ECore I datasets, using z-scoring. In brief, 30 genes were Z-scored across the mentioned datasets, the average Z-score of the genes was generated, and the TGF-β activation score was assessed across diseases relative to LDs and correlated with GFR. Our analysis revealed a negative correlation of the TGF-β score with eGFR decline in patients with CKD (N=163, R=0.60, P < 0.01; Figure 1C). Together, our data demonstrate that increased levels of active TGF-β in the urine of patients with CKD are associated with increased expression of fibrotic signature genes in the kidneys and the decline in eGFR.

Figure 1.

Associations between urinary active TGF-β levels, fibrotic gene signatures in kidneys, and decline in eGFR among patients with CKD. (A) Active TGF-β/creatinine ratio was increased by 4.4-fold (P = 0.0003) in participants with CKD compared with non-CKD controls (N=20). Student t test was used for analysis. (B) GSEA showed increased expression of EMT related genes and collagen in both GM and TI of patients with CKD (N=150). Data presented as Box and Whisker plots and analyzed using one-way ANOVA. The P values were adjusted for multiple hypothesis testing using the Benjamini–Hochberg method. EMT markers and collagen were identified as significantly differentially expressed in GM and TI. The threshold for significance was set at an adjusted P value of <0.05. (C) TGF-β score negatively correlates with eGFR decline in different stages of CKD (N=163). Data were analyzed using Pearson correlation analysis. EMT, epithelial-mesenchymal transition; ERCB, European Renal cDNA Bank; GM, glomeruli; GSEA, gene set enrichment analysis; GSVA, gene set variation analysis; LD, living donor; TI, tubulointerstitium.

Kidney αvβ8 Expression Increases during CKD Progression in Patients

To explore mechanisms of fibrosis in CKD, we investigated αvβ8 in human CKD cohorts because previous reports have shown that αvβ8 is an essential mechanism of profibrotic TGF-β activation.⁶

Through an immunohistochemistry screening of αvβ8 protein expression in 33 healthy human tissues, we observed preferential expression of the integrin in kidneys, specific brain regions, and nerve bundles (Figure 2A and Supplemental Figure 1A). In addition, increased expression of ITGB8, the gene encoding integrin subunit β8 of αvβ8, was observed in commercially resourced kidney biopsies from patients with CKD (n=4; Figure 2B). Subsequent analysis of kidney biopsies from patients with CKD with different etiologies within the ERCB cohort (n=129) revealed elevated kidney ITGB8 gene expression specifically in diabetic nephropathy and rapidly progressing GN (RPGN; Figure 2C). In an independent DKD KI/SU patient cohort, we found that the increase in ITGB8 gene expression was predominantly observed in the tubulointerstitial compartment (Figure 2D). Finally, we performed post hoc analysis on published single cell RNA sequencing from KPMP cohort²¹ (Supplemental Figure 1B). These data highlight enriched expression in proximal tubule regions and collectively, these findings support the conclusion that αvβ8 expression is enriched in healthy kidneys and increased in the tubulointerstitial compartment in patients with CKD.

Figure 2.

Kidney αvβ8 expression increases during CKD progression in patients. (A) Representative images of immunohistochemistry staining of αvβ8 in healthy human kidneys. Isotype antibody was used as a control. (n=3). (B) The mRNA expression of αvβ-1, -3, -5, -6, -8 revealed an increased expression of αvβ8 compared with other integrins in kidneys of patients with CKD (n=4). P < 0.05 compared with ITGB8 (C) Elevated kidney ITGB8 gene expression in DN and RPGN in the ERCB cohort (N=129). (D) Elevated ITGB8 gene expression in GM and tubulointerstitial compartment in DKD KI/SU patient cohort. Data are presented as Box and Whisker plots and analyzed using one-way ANOVA. The adjusted P values threshold for significance was set at <0.05. Scale bar=25 µm. DKD, diabetic kidney disease; DN, diabetic nephropathy; HT, hypertension; IgA, IgA nephropathy; KI/SU, Karolinska Institute/Sahlgrenska University Hospital; RPGN, rapidly progressing GN.

Associations between αvβ8 Upregulation, Fibrotic Signatures in Kidneys, Disease Severity, and Decline in eGFR among Patients With CKD

To further delineate the localization of αvβ8 expression within kidneys, we conducted immunohistochemical staining of kidney biopsies from patients with DKD. Consistent with ITGB8 gene expression, we observed an augmented αvβ8 expression primarily confined to the tubulointerstitial compartment, colocalizing with regions exhibiting increased collagen 1 expression (Figure 3A), thus suggesting an association between αvβ8 expression and ECM deposition in the kidneys of patients with DKD. Furthermore, in patients across multiple CKD etiologies in the ERCB cohort, we noted that the elevation in tubulointerstitial ITGB8 gene expression was particularly pronounced from CKD stage 3 (Figure 3B), which corresponds to the increase in fibrotic gene profiles previously described (Figure 1, B and C). In addition, we observed a positive correlation between the TGF-β activation score and ITGB8 gene expression in ERCB GM and TI in the Clinical Phenotyping Resource and Organ Biobank ECore TI dataset (Figure 3C). Notably, we also identified an inverse correlation between the ITGB8 gene expression and eGFR in patients with CKD relative to LDs in the ERCB CKD patient cohort (Figure 3D). Together, our findings illustrate that increased αvβ8 expression in the tubulointerstitial compartment is linked to kidney fibrosis and the decline in eGFR among patients with CKD.

Figure 3.

Associations between αvβ8 expression, fibrotic gene signatures in kidneys, and decline in eGFR among patients with CKD. (A) Representative images of immunohistochemistry staining of αvβ8 and collagen 1 in kidney biopsies of patients with DKD. (n=3) (B) The tubulointerstitial ITGB8 gene expression in kidneys of patients with CKD in ERCB cohort at different stages of CKD (N=129). Data are presented as Box and Whisker plots and analyzed using one-way ANOVA. The adjusted P values threshold for significance was set at <0.05. (C) TGF-β score correlated with kidney ITG8 expression (N=163). (D) An inverse correlation between the ITGB8 gene expression and eGFR in patients with CKD relative to LDs in ERCB CKD patient cohort (N=129). Data were analyzed using Pearson correlation analysis. Scale bar=20 µm.

Monoclonal Anti-αvβ8 Integrin Antibody MEDI8367 Prevents TGF-β Activation In Vitro

MEDI8367 (Figure 4A) specifically binds to human integrin β8 and works allosterically, reducing its affinity for the LAP domain, hence preventing β8-mediated TGF-β activation, but not its cell adhesion function, as previously described.¹⁴ We confirmed using in vitro assays MEDI8367 was specific to humans and not mouse cross reactive (Figure 4C). Showing that humanized transgenic αvβ8 mice are necessary for target validation and in vivo disease models. We also checked that human αvβ8 protein binds with comparable affinity with the recombinant human and mouse LAP proteins suggesting that the αvβ8-dependent TGF-β activation system should be functional in humanized mice, to test the efficacy of MEDI8367 (Figure 4D). Finally, through a HeLa-TMLC reporter assay, we showed that MEDI8367 effectively inhibited αvβ8-dependent TGF-β activation in vitro, although to a lesser extent than the pan αv antibody (Figure 4E).

Figure 4.

Monoclonal anti-αvβ8 integrin antibody MEDI8367 prevents TGF-β activation in vitro. (A) Schematic of MEDI8367 binding residues (B) and MEDI8367 specifically bound to human integrin β8. (C) Testing of MEDI8367 for human but not mouse integrin β8 cross reactivity and (D) testing of human β8 affinity for both human and mouse LAP. (E) Neutralizing activity using a HeLa TMLC reporter assay showing MEDI8367 fully and specifically blocked αvβ8-dependent TGF-β activation in vitro. The adjusted P values threshold for significance was set at a <0.05. LAP, latency-associated peptide; TMLC, transformed mink lung epithelial cells.

Preclinical Models of CKD Show Inhibition of αvβ8 Integrin Benefits Kidney Fibrosis

We used three preclinical models of CKD to explore the role of αvβ8 integrin in kidney fibrosis. Using humanized αvβ8 mice as hypomorphs, which had very low Itgb8 expression compared with wild-type mice (Supplemental Figure 2). Because humanized Itgb8 hypomorphs would have reduced αvβ8 functionality, we first tested in the 5/6 nephrectomy CKD model. We showed no activation of TGF-β in 5/6 nephrectomy humanized hypomorphs, measured by renal active TGF-β and urinary phospho-SMAD2/3 (Figure 5, A and B). Moreover, despite no change in renal αvβ8 expression, we observed a rescue of kidney functional decline in αvβ8 hypomorphs, as evidenced by reduced urine albumin-to-creatinine ratio (UACR; Figure 5C and Supplemental Figure 3).

Figure 5.

Reduced ITGB8 improves kidney function in 5/6 nephrectomy preclinical model. (A and B) Using hypomorphs, we showed no activation of renal TGF-β in 5/6 nephrectomy humanized mice, measured by active TGF (A) and phospho-SMAD2/3 (B) (n=6–9). (C) Rescue of kidney function decline in humanized αvβ8 hypomorphs after 5/6 nephrectomy with reduced UACR (C; n=10). *** and **** represented P < 0.01 and P < 0.001, respectively. NX, nephrectomy; UACR, urine albumin-to-creatinine ratio; WT, wildtype.

Next, we wanted to evaluate MEDI8367 in a severe kidney fibrosis model. We induced kidney fibrosis through UUO in humanized αvβ8 mice. Interestingly, despite very low expression of Itgb8 in hypomorphs, we still noted elevated αvβ8 protein expression in kidney tubules with increased collagen and fibronectin RNA (Figure 6, A and B). Blocking αvβ8 with MEDI8367 significantly prevented the activation of TGF-β signaling as measured by reduced renal phospho-SMAD2/3 (Figure 6C). Moreover, treatment with MEDI8367 led to decreased kidney fibrosis as indicated by collagen content in obstructed kidneys (Figure 6D). Anti-TGF-β antibody was used as a positive control to inhibit TGF-β activation.

Figure 6.

MEDI8367 reduces fibrosis in UUO model. (A and B) UUO model in humanized ITGB8 mice showed elevated αvβ8 protein expression in kidney tubules (A) with increased collagen and fibronectin RNA (B; n=9). (C) Blocking αvβ8 with MEDI8367 (10 mg/kg) significantly prevented the activation of TGF-β signaling as measured by reduced renal phospho-SMAD2/3 (n=5–8). (D). MEDI8367 treatment (10 mg/kg) decreased collagen content (D). (n=9–10). *** and **** represented P < 0.01 and P < 0.001, respectively. Scale bar=1 mm. UUO, unilateral ureteral obstruction.

Finally, to explore in a DKD-relevant setting, we used the db/db uninephrectomy DKD model. Because MEDI8367 was not mouse-cross reactive (Figure 4C), we used surrogate murine antibodies 3G9 and RAD2640 in db/db uninephrectomy. Both antibodies target murine αvβ6 integrin, but RAD2640 also exhibits cross-reactivity with murine αvβ8 integrin. We showed that RAD2640, but not 3G9, declined the progression of UACR (Figure 7, A–C). These data indicated that blocking of αvβ8 is required for the functional benefit of the kidney in this DKD model.

Figure 7.

RAD2640 improves kidney function in db/db UNX mice. (A–C) MEDI8367 is not mouse cross-reactive, we used surrogate murine αvβ6 and αvβ6/8 blocking antibodies (3G9 and RAD2640). RAD2640, but not 3G9 improved UACR in db/db UNX mice over 12–17 weeks (A), individual progression (B) and progression rate (C) over 17 weeks. (n=9). *** represents P < 0.01. UACR, urine albumin-to-creatinine ratio; UNX, uninephrectomy.

Together, we conclude that blocking αvβ8 integrin with MEDI8367 attenuated fibrosis in humanized αvβ8 mice and blocking αvβ8 integrin may have the potential to improve kidney function in DKD.

Discussion

We hypothesized that αvβ8 integrin plays a pivotal role in driving kidney fibrosis through the activation of TGF-β signaling. Moreover, we proposed that blocking αvβ8 integrin could protect kidney function during CKD. Our findings shed light on the involvement of TGF-β and αvβ8 integrin in CKD, along with their association with fibrotic gene signatures and declining kidney function.

Initially, we established a TGF-β activation signature using publicly available human CKD datasets (ERCB, C-PROBE) and subsequently observed a clear association with the decline in renal function in patients with CKD. This correlation indicated the potential involvement of TGF-β in driving fibrotic responses, such as EMT and collagen deposition, particularly prominent in patients at stage 3 CKD and those with DKD. These results underscore the critical need to identify drugs capable of inhibiting kidney fibrosis development. Furthermore, our data demonstrated increased levels of active TGF-β in urine, suggesting its potential utility as a biomarker for predicting kidney fibrosis in patients with CKD.

In addition to its well-established profibrotic role, TGF-β also contributes significantly to tissue homeostasis and the modulation of immune responses.^1,22 Notably, systemic blockade of TGF-β has been associated with severe adverse effects, including an inflammatory phenotype, progression of premalignant lesions, and limited efficacy at safe doses for patients.³ This underscores the necessity for targeted inhibition of TGF-β in the kidney. Our study revealed that αvβ8 integrin, essential for profibrotic TGF-β activation, is preferentially expressed in areas of ECM deposition within the kidneys, along with a negative correlation with eGFR during CKD across multiple independent human CKD cohorts. This suggests that targeting αvβ8 integrin could prevent kidney fibrosis.

We developed a human-specific monoclonal anti-αvβ8 integrin antibody, MEDI8367. MEDI8367 exerted incomplete inhibition of TGF-β activation, which might be due to other TGF-β activating pathways, such as other αv integrins, being present on the reporter cells.¹⁴ To further investigate the role of αvβ8 integrin, we used humanized αvβ8 mice with reduced αvβ8 expression. These hypomorphs exhibited no activation of renal TGF-β, resulting in decreased kidney fibrosis in a 5/6 nephrectomy-induced CKD model compared with sham-operated mice, highlighting the significance of αvβ8 integrin in kidney fibrosis. In addition, in a more robust kidney fibrosis model, UUO-induced CKD,²³ MEDI8367 inhibited αvβ8 integrin and delayed the progression of kidney fibrosis in humanized αvβ8 hypomorph mice. Furthermore, in vivo experiments in a DKD setting indicated that blocking αvβ8 integrin delayed kidney function decline, emphasizing its role as a driver of kidney damage. Interestingly, although previous studies have demonstrated a role for αvβ6 integrin in organ fibrosis, we did not observe a reduction in UACR progression in our DKD model after specific αvβ6 inhibition, which may be attributable to differences in the mode of inhibition, for example, we used an anti-αvβ6 antibody, whereas prior studies used genetic αvβ6 deficiency or small molecule inhibitors.^24–27 Moreover, in an independent DKD KI/SU patient cohort, ITGB8 gene expression was elevated only in the tubulointerstitial compartment, whereas ITGB6 gene expression was increased in both the GM and tubulointerstitial compartments. While not conclusive, this differential expression pattern may partially contribute to the observed lack of UACR reduction after αvβ6 inhibition in our model and highlights the complexity of integrin signaling in DKD.

We have highlighted variability of αvβ8 expression across the kidney, but with some enrichment in tubulointerstitial compartment across multiple CKD cohorts. We also describe changes in αvβ3 between healthy and diabetic GM. More work will be needed if this integrin has a role in diabetic nephropathy. Interestingly, because ITGAV is highly expressed in both GM and tubulointerstitial and ITGB8 is lower, this may be interesting to study if enriched expression of ITGAV is driven by coupling with other β integrins.²⁸

While analyzing the ITG8 gene expression in the ERCB cohort, we observed elevated levels in both RPGN and DKD. Although αvβ8 upregulation in RPGN may reflect an acute-phase inflammatory response and requires further investigation, we propose that it may also contribute to profibrotic signaling in RPGN related to inflammation and tissue damage and pathological fibrosis in DKD. This hypothesis is supported by our findings that αvβ8 expression is increased in the UUO model and is associated with TGF-β activation, leading to fibrosis in humanized αvβ8 mice, therefore, emphasizing a mechanistic link to fibrosis.

Our findings underscore the involvement of αvβ8 integrin in driving fibrotic responses in CKD/DKD and highlight its potential as a therapeutic target to mitigate kidney fibrosis and improve kidney function in patients with CKD/DKD. However, we cannot fully exclude the possibility that αvβ8 integrin may exert effects independent of TGF-β activation, and warrants further investigation. Furthermore, we did not explore sex-specific differences in this study, and this important aspect should be addressed in future research.

In summary, the introduction of MEDI8367 as a promising mAb for inhibiting αvβ8-mediated TGF-β activation may offer new prospects for therapeutic intervention in CKD. These results provide new insights into the molecular mechanisms underlying kidney fibrosis and may help the development of novel treatments for CKD.

Supplementary Material

kidney360-7-0942-s001.pdf ^{(3.4MB, pdf)}

kidney360-7-0942-s002.pdf ^{(589.1KB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/KN9/B434.

Author Contributions

Conceptualization: David Baker, Barbara Hebeis, Richard Stebbings, Kevin J. Woollard.

Data curation: Lihuan Liang, Elena Liarte Marin, Asha Seth, Richard Stebbings.

Formal analysis: James Conway, Sonja Hess, Jean-Martin Lapointe, Lihuan Liang, Elena Liarte Marin, Shrikant R. Mulay.

Investigation: Elena Liarte Marin.

Methodology: Elena Liarte Marin.

Project administration: Barbara Hebeis.

Writing – original draft: Lihuan Liang, Elena Liarte Marin, Shrikant R. Mulay.

Writing – review & editing: David Baker, James Conway, Pernille B.L. Hansen, Jean-Martin Lapointe, Michael McCarthy, Shrikant R. Mulay, Asha Seth, Kevin J. Woollard.

Funding

AstraZeneca.

Declarative Statements

All animal experiments were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals or an equivalent standard that meets or exceeds the ethical and welfare requirements outlined in the NIH Guide. All protocols were approved by the appropriate institutional animal care and use committee.

Data Availability Statements

All original data, including deidentified patient-level data or individual laboratory data measurements, are included in the manuscript and/or supplemental material.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/KN9/B435.

Supplemental Table 1. Thirty gene signatures of TGF-β activation used for target validation.

Supplemental Figure 1. (A) ITGB8 (αvβ8) protein expression in 33 healthy human tissues, we observed preferential expression of the integrin in kidneys, specific brain regions, and nerve bundles. Representative examples for each tissue are shown. Scale bar=0.1 mm. (B) Post hoc analysis of ITGB8 in published single-cell RNA sequencing from KPMP cohort in CKD and control donors (BB Lake et al. Nature, 2023, 585–594). Normalized UMI. Error bars represent 95% confidence interval.

Supplemental Figure 2. Humanized αvβ8 mice as hypermorphs. Humanized mice showed very low ITGB8 expression compared with wild-type mice (n=7–10).

Supplemental Figure 3. ITGB8 (αvβ8) protein is upregulated by UUO in humanized αvβ8 mice. Humanized mice showed increased αvβ8 expression after UUO compared with wild-type mice (n=5), but no change in expression in 5/6 nephrectomy (n=6).