Indicators of Kidney Fibrosis in Patients with Type 2 Diabetes and Chronic Kidney Disease Treated with Dulaglutide

Chronic kidney disease (CKD) is estimated to affect almost half of patients with type 2 diabetes (T2D), and diabetes is the leading cause of kidney failure worldwide [1, 2]. Progressive decline in the glomerular filtration rate (GFR) is associated with structural abnormalities including glomerular and interstitial fibrosis that results from increased collagen deposition due to imbalance in the formation and degradation of the extracellular matrix proteins [3]. PRO-C6 is a biomarker for type VI collagen formation [4], while C3M is a biomarker for type III collagen degradation. Increases in serum PRO-C6 and lower levels of urine C3M have been reported to correlate with CKD progression, lower estimated GFR (eGFR), and higher levels of tubulointerstitial fibrosis after kidney transplant and in IgA nephropathy and type 1 and 2 diabetes [5‒11].

The glucagon-like peptide-1 receptor (GLP-1R) is present on both intrinsic and invading immune cells within the diabetic kidney [12, 13]. When bound by a ligand, GLP-1R activation increases intracellular cyclic adenosine monophosphate which, in turn, inhibits a series of signals for transcriptional upregulation of proinflammatory and profibrotic mediators. Notably, GLP-1R agonists (GLP-1 RAs) have been observed to reduce albuminuria and slow eGFR decline in cardiovascular outcome trials [14‒16]. Dulaglutide is a once-weekly GLP-1 RA that is not cleared by kidney [17] but is presumed to be catabolized by proteolytic degradation. In the Assessment of Weekly AdministRation of LY2189265 (dulaglutide) in Diabetes-7 (AWARD-7) trial of glycemic control in participants with T2D and moderate-to-severe CKD, dulaglutide slowed the decline in eGFR and decreased the urine albumin-to-creatinine ratio (UACR) compared with insulin glargine [18]. The aim of this exploratory analysis of the AWARD-7 study was to evaluate changes in serum PRO-C6 and urine C3M as biomarkers of kidney fibrosis in the overall safety population as well as the macroalbuminuric subgroup.

The AWARD-7 clinical trial (NCT01621178) was a 52-week, randomized, multicenter, open-label (dulaglutide dose masked), parallel-arm study. Participants were randomized (1:1:1) to receive weekly dulaglutide 1.5 mg, dulaglutide 0.75 mg, or daily insulin glargine with premeal insulin lispro, as needed. Details of the study design and eligibility criteria have been previously published [18]. The main inclusion criteria included adults (age ≥18 years) with T2D and moderate-to-severe CKD (eGFR ≥15 and <60 mL/min/1.73 m²), glycated hemoglobin A1c (HbA1c) of 7.5–10.5% (58.5–91.3 mmol/mol), treatment with insulin alone or in combination with an oral antihyperglycemic drug, and treatment with an angiotensin-converting enzyme inhibitor and/or an angiotensin receptor blocker at the maximal tolerated dose. Exclusion criteria were type 1 diabetes, stage 5 CKD (eGFR <15 mL/min/1.73 m²), dialysis, acute kidney injury, or expectation of dialysis or kidney transplantation in the next 6 months.

The study was conducted in accordance with the Declaration of Helsinki. The protocol was approved by the Local Institutional Review Boards. All participants provided written informed consent.

Frozen stored samples of serum and urine from participants in the dulaglutide 1.5 mg and insulin glargine arms were used for these analyses. Samples were collected at baseline, week 26, and week 52 or at early termination. Serum PRO-C6 and urine C3M were measured using competitive enzyme-linked immunosorbent assays developed and produced at Nordic Bioscience (Herlev, Denmark) according to manufacturer’s instructions. The PRO-C6 assay used a monoclonal antibody detecting the sequence KPGVISVMGT, corresponding to the last 10 amino acids of the type VI collagen α3 chain. The C3M assay used a monoclonal antibody detecting the sequence KNGETGPQGP, corresponding to the cleavage site of MMP-9 in position 610 of type III collagen. C3M concentrations were divided by urine creatinine to adjust for urine concentration. Urine creatinine was measured by Jaffe reaction on an ADVIA 1800 chemistry system (Siemens, Germany).

The eGFR was measured at baseline, week 26, and week 52 with the 2009 CKD-Epidemiology Collaboration (CKD-EPI)-creatinine equation and the CKD-EPI-cystatin C equation [19]. Changes from baseline in fibrosis biomarker levels at 26 and 52 weeks were analyzed in the safety population (all patients who received at least one dose of study treatment) that had any post-dose biomarker data. The patients in the overall safety population with baseline UACR >300 mg/g were categorized as the macroalbuminuria subgroup.

Data were statistically analyzed using a mixed-effect model for repeated measures of log-transformed values with the following covariates: baseline log-transformed values for biomarkers, log of CKD-EPI-creatinine eGFR, HbA1c, systolic blood pressure, macroalbuminuria region, CKD severity categories at baseline, treatment, time, and treatment by time. Tests of potential treatment effects were conducted at a two-sided alpha level of 0.05 and two-sided 95% confidence interval unless otherwise indicated, with no adjustments for multiplicity. Pearson correlation analyses were conducted to assess associations between changes in biomarkers and changes in CKD-EPI-creatinine or CKD-EPI-cystatin C eGFR measures in the safety population with post-dose biomarker data, and by the treatment group at 52 weeks. All statistical analyses were implemented using R software and SAS version 9.1 or higher.

Baseline characteristics were similar between treatment groups in the overall safety study population that had any post-dose biomarker data (Table 1). Women comprised half of this population (N = 330), and the mean ± SD age was 65 ± 8 years old. The racial and ethnic distribution included persons who identified as American Indian/Alaskan Native (8%), Asian (4%), Black or African American (13%), Hispanic or Latino/a (41%), and White (70%). Their duration of diabetes was 18 ± 9 years, and their known duration of CKD was 4 ± 5 years. Mean baseline eGFR was 37.3 ± 12.7 mL/min/1.73 m² and 38.3 ± 14.7 mL/min/1.73 m² by CKD-EPI-creatinine and CKD-EPI-cystatin C, respectively. The median (interquartile range) for UACR was 196 (37–862) mg/g. At baseline, the overall serum PRO-C6 and urine C3M geometric mean values (%coefficient of variation) were 16.6 (2.4) ng/mL and 21.8 (5.3) µg/mmol, respectively.

From baseline to week 26 and week 52, the serum PRO-C6 levels increased in both treatment groups (Table 2; Fig. 1a). However, the increase in serum PRO-C6 level was significantly less in the dulaglutide group than in the insulin glargine group at week 26 and week 52. The proportional change from baseline (CFB) to 52 weeks in serum PRO-C6 biomarker was significantly lower in the dulaglutide group as compared with the insulin glargine group (least squares mean ± standard error; −0.2% ± 2.2 vs. 8.0% ± 2.3, p = 0.005; Fig. 1c). The urine C3M levels significantly increased overtime in the dulaglutide treatment group, while they decreased in the insulin glargine treatment group (Table 2; Fig. 1b). The proportional CFB to 52 weeks in urine C3M biomarker was also significantly higher in the dulaglutide group as compared with the insulin glargine group (20.7% ± 8.8 vs. −16.9% ± 6.4, p < 0.001; Fig. 1d).

At baseline, the macroalbuminuria subgroup had numerically higher serum PRO-C6 levels and lower urine C3M levels than the overall study population (Table 2; Fig. 1a, b). At weeks 26 and 52, serum PRO-C6 levels were significantly lower and urine C3M levels were significantly higher in the dulaglutide treatment group compared with the insulin glargine treatment group (Fig. 1).

The CFB to week 52 in serum PRO-C6 negatively correlated with eGFR in both treatment groups. In the dulaglutide group, R = −0.78 and p < 0.001 for CKD-EPI-cystatin C eGFR (Fig. 2a); R = −0.77 and p < 0.001 for CKD-EPI-creatinine eGFR (Fig. 2c). In the insulin glargine group, R = −0.33 and p < 0.001 for CKD-EPI-cystatin C eGFR (Fig. 2a); R = −0.45 and p < 0.001 for CKD-EPI-creatinine eGFR (Fig. 2c).

The CFB to week 52 in urine C3M positively correlated with eGFR in both treatment groups. In the dulaglutide group, R = 0.22 and p = 0.032 for CKD-EPI-cystatin C eGFR (Fig. 2b); R = 0.32 and p = 0.002 for CKD-EPI-creatinine eGFR (Fig. 2d). In the insulin glargine group, R = 0.18 and p = 0.092 for CKD-EPI-cystatin C eGFR (Fig. 2b); R = 0.28 and p = 0.007 for CKD-EPI-creatinine eGFR (Fig. 2d).

Patients with T2D and moderate-to-severe CKD treated with dulaglutide 1.5 mg weekly had lower levels of a biomarker for type VI collagen formation, serum PRO-C6, and higher levels of a biomarker for type III collagen degradation, urine C3M, compared to insulin glargine. Similar patterns were observed in the macroalbuminuric subgroup, representing 43% of the overall population, although baseline levels of serum PRO-C6 appeared higher and those of urine C3M trended lower, along with larger magnitudes of change associated with dulaglutide. Taken together, these findings are consistent with a reduction in collagen deposition associated with fibrosis within the kidney and perhaps elsewhere. Reduction of type VI collagen formation could be potentially beneficial. In a prospective, observational study, higher collagen VI formation and serum PRO-C6 predicted cardiovascular events, all-cause mortality, and CKD progression in patients with T2D and microalbuminuria [11]. Serum PRO-C6 is also a measure of endotrophin, a bioactive molecule derived from the type VI collagen pro-peptide, which has been associated with proinflammatory and pro-fibrotic effects in different organs [20‒24]. Hence, the reduction in serum PRO-C6 levels may also reflect the systemic reduction of a potential pathological protein.

Additionally, either creatinine- or cystatin C-based eGFR negatively correlated with serum PRO-C6 and positively correlated with urine C3M, which may be interpreted to reflect an association of lower collagen deposition with lesser eGFR decline. Both serum PRO-C6 and urine C3M have previously shown good correlation with eGFR after kidney transplant and in IgA nephropathy and type 1 diabetes [5‒7]. The AWARD-7 study demonstrated slower eGFR decline with dulaglutide compared to insulin glargine over 1 year of treatment, an effect that was particularly evident in participants with more severe CKD as indicated by macroalbuminuria [18]. This study cohort was selected for people diagnosed with moderate-to-severe CKD (eGFR ≥15 and <60 mL/min/1.73 m²) at study entry. Notably, in cardiovascular outcome trials that enrolled participants with a much broader range of kidney function, those with eGFR <60 mL/min/1.73 m² appeared to benefit more than those with higher eGFR >60 mL/min/1.73 m² from GLP-1 RAs for reducing the rate of eGFR decline [14, 16]. Although the mechanistic basis for these observations remains to be determined, it is plausible later stage CKD in diabetes is characterized by more severe glomerular and tubulointerstitial damage that may be amenable to anti-inflammatory and anti-fibrotic effects of GLP-1 RAs [13, 25, 26].

Besides the potential for direct effects on the kidney, the GLP-1 RAs also have favorable effects on CKD risk factors that could indirectly contribute to reducing fibrotic injury to the diabetic kidney. However, levels of HbA1c as well as systolic and diastolic blood pressure were comparable between the dulaglutide and insulin glargine treatment groups in AWARD-7, but body weight was substantially lower with an approximately 4 kg difference on average between dulaglutide 1.5 mg- and insulin glargine-treated participants [18]. However, in a post hoc analysis, no association between weight change and either creatinine- or cystatin C-based eGFR was observed [27]. A mediation analysis performed for the Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial and Trial to Evaluate Cardiovascular and Other Long-Term Outcomes With Semaglutide in Subjects With Type 2 Diabetes (SUSTAIN-6) also found no convincing evidence that changes in body weight predicted the favorable effects of liraglutide or semaglutide to preserve eGFR [28].

This study has valuable strengths including the availability of stored samples from a rigorously conducted clinical trial for biomarker analyses, which can help uncover potential pathways for the preservation of kidney function associated with some GLP-1 RAs. The present data showing associations between biomarkers that reflect a lessened fibrosis process and preservation of eGFR can serve to guide subsequent mechanistic investigations. Nevertheless, this study is limited by being a post hoc exploratory analysis that must be viewed as primarily hypothesis-generating. Reverse causality is also possible. For example, those with higher eGFR could have greater clearance of serum PRO-C6 or perhaps these individuals have an alternate basis for a larger amount of urine C3M. In addition, the measured biomarkers are not organ specific and, therefore, are not unique for kidney tissue. However, both biomarkers have previously been associated with extent of fibrosis in the kidneys [8, 9, 29]. Although Hispanic or Latino/a ethnicity was well represented in the study (approximately 40%), the majority of participants identified as White race (approximately 70%), which may limit generalizability of the findings to other racial groups.

In conclusion, dulaglutide treatment was associated with biomarkers of lower type VI collagen formation and higher type III collagen degradation compared to treatment with insulin glargine. These findings were more evident in the macroalbuminuric subgroup, an observation consistent with the concept that GLP-1 RAs may have particular benefit in patients with T2D and more advanced CKD. Future research is needed to evaluate whether there are anti-fibrotic effects associated with dulaglutide treatment that are mechanistically linked to slower decline in kidney function.

We thank Kathleen Kelly-Boruff from Eli Lilly and Company for her assistance in handling the kidney fibrosis biomarker samples. We thank Shirin Ghodke, PhD, and Amelia Torcello Gomez, PhD, from Eli Lilly and Company for their writing and editorial support.

The appropriate National/Local Institutional Review Boards and Ethics Committees approved the protocol. The Copernicus Group IRB, Durham, NC, USA, was the master Ethics Committee and approved the protocol on April 3, 2012. All participants provided written informed consent before participating in the study.

K.R.T. is supported by NIH research grants R01MD014712, U2CDK114886, UL1TR002319, U54DK083912, U01DK100846, OT2HL161847, UM1AI109568, and CDC contract 75D301-21-P-12254; and reports other support from Eli Lilly and Company; personal fees and other support from Boehringer Ingelheim; personal fees and other support from AstraZeneca; grants, personal fees, and other support from Bayer AG; grants, personal fees, and other support from Novo Nordisk; grants and other support from Goldfinch Bio; other support from Gilead; and grants from Travere outside the submitted work. F.G. and M.A.K. are employees and stockholders of Nordic Bioscience and have no other conflicts of interest. J.M.W., Y.L., H.-R.Q., K.L.D., and F.T.B. are employees and stockholders of Eli Lilly and Company and have no other conflicts of interest.

Assessment of Weekly AdministRation of LY2189265 (dulaglutide) in Diabetes-7 (AWARD-7) study was sponsored by Eli Lilly and Company (Indianapolis, Indiana). All the authors have nothing to disclose with regard to the manuscript.

All authors critically reviewed/edited the manuscript and took responsibility for the integrity and accuracy of the data presented. K.R.T. and F.T.B. contributed to the AWARD-7 study design and drafted the manuscript. J.M.W., F.G., M.A.K., K.L.D., and F.T.B. contributed to this post hoc analysis design. F.G. and M.A.K. conducted the biomarker assays. Y.L. and H.-R.Q. conducted and interpreted the statistical analyses.

Eli Lilly and Company provides access to all individual participant data collected during the trial, after anonymization, with the exception of pharmacokinetic or genetic data. Data are available to request 6 months after the indication studied has been approved in the USA and EU and after primary publication acceptance, whichever is later. No expiration date of data requests is currently set once data are made available. Access is provided after a proposal has been approved by an independent review committee identified for this purpose and after a receipt of a signed data sharing agreement. Data and documents, including the study protocol, statistical analysis plan, blank or annotated case report forms, will be provided in a secure data sharing environment. For details on submitting a request, see the instructions provided at www.vivli.org. Further inquiries can be directed to the corresponding author.

Previous presentations: Part of these data were previously presented as virtual posters at the 80th Annual Scientific Sessions of American Diabetes Association held on June 12–16, 2020 and at Kidney Week 2020 of the American Society of Nephrology on October 22–25, 2020.