Diabetic Kidney Disease: Drug Discovery and Clinical Development Challenges

Journal Articles

Agarwal A, Duffin KL, Laska DA, et al. A prospective study of multiple protein biomarkers to predict progression in diabetic chronic kidney disease. Nephrol Dial Transplant. 2014;29(12):2293-302.

Beckerman P, Ko YA, Susztak K. Epigenetics: a new way to look at kidney diseases. Nephrol Dial Transplant. 2014;29(10):1821-7.

Breyer MD. Drug discovery for diabetic nephropathy: trying the leap from mouse to man. Semin Nephrol. 2012;32(5):445-51.

Breyer MD. Translating experimental diabetic nephropathy studies from mice to men. Contrib Nephrol. 2011;170:156-64.

Breyer MD, Coffman TM, Flessner MF, et al. Diabetic nephropathy: a national dialogue. Clin J Am Soc Nephrol. 2013;8(9):1603-5.

Brosius FC 3rd. Susceptible mice: identifying a diabetic nephropathy disease locus using a murine model. Kidney Int. 2010;78(5):431-2.

Brosius FC 3rd, Alpers CE. New targets for treatment of diabetic nephropathy: what we have learned from animal models. Curr Opin Nephrol Hypertens. 2013;22(1):17-25.

Brosius FC 3rd, He JC. JAK inhibition and progressive kidney disease. Curr Opin Nephrol Hypertens. 2015;24(1):88-95.

Coresh J, Turin TC, Matsushita K, et al. Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality. JAMA. 2014;311(24):2518-31.

Fabian SL, Penchev RR, St-Jaques B, et al. Hedgehog-Gli pathway activation during kidney fibrosis. Am J Pathol. 2012;180(4):1441-53.

Formentini I, Bobadilla M, Haefliger C, et al. Current drug development challenges in chronic kidney disease (CKD) — identification of individualized determinants of renal progression and premature cardiovascular disease (CVD). Nephrol Dial Transplant. 2012;27 Suppl 3:iii81-8.

Gerich JE. Role of the kidney in normal glucose homeostasis and in the hyperglycaemia of diabetes mellitus: therapeutic implications. Diabet Med. 2010;27(2):136-42.

Greene T, Teng CC, Inker LA, et al. Utility and validity of estimated GFR-based surrogate time-to-event end points in CKD: a simulation study. Am J Kidney Dis. 2014;64(6):867-79.

Humphreys BD. Targeting pericyte differentiation as a strategy to modulate kidney fibrosis in diabetic nephropathy. Semin Nephrol. 2012;32(5):463-70.

Jiang R, Wang S, Takahashi K, et al. Generation of a conditional allele for the mouse endothelial nitric oxide synthase gene. Genesis. 2012;50(9):685-92.

Jourdan T, Demizieux L, Gresti J, et al. Antagonism of peripheral hepatic cannabinoid receptor-1 improves liver lipid metabolism in mice: evidence from cultured explants. Hepatology. 2012;55(3):790-9.

Jourdan T, Godlewski G, Cinar R, et al. Activation of the Nlrp3 inflammasome in infiltrating macrophages by endocannabinoids mediates beta cell loss in type 2 diabetes. Nat Med. 2013;19(9):1132-40.

Jourdan T, Szanda G, Rosenberg AZ, et al. Overactive cannabinoid 1 receptor in podocytes drives type 2 diabetic nephropathy. Proc Natl Acad Sci U S A. 2014;111(50):E5420-8.

Ju W, Brosius FC 3rd. Understanding kidney disease: toward the integration of regulatory networks across species. Semin Nephrol. 2010;30(5):512-9.

Kang HM, Ahn SH, Choi P, et al. Defective fatty acid oxidation in renal tubular epithelial cells has a key role in kidney fibrosis development. Nat Med. 2015;21(1):37-46.

Kramann R, Humphreys BD. Kidney pericytes: roles in regeneration and fibrosis. Semin Nephrol. 2014;34(4):374-83.

Kramann R, Schneider RK, DiRocco DP, et al. Perivascular Gli1(+) progenitors are key contributors to injury-induced organ fibrosis. Cell Stem Cell. 2014. [Epub ahead of print]

Ledo N, Ko YA, Park AS, et al. Functional genomic annotation of genetic risk loci highlights inflammation and epithelial biology networks in CKD. J Am Soc Nephrol. 2014. [Epub ahead of print]

Levey AS, Inker LA, Matsushita K, et al. GFR decline as an end point for clinical trials in CKD: a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration. Am J Kidney Dis. 2014;64(6):821-35.

Little MH, Brown D, Humphreys BD, et al. Defining kidney biology to understand renal disease. Clin J Am Soc Nephrol. 2014;9(4):809-11.

Loghman-Adham M, Kie Weber CI, Ciorciaro C, et al. Detection and management of nephrotoxicity during drug development. Expert Opin Drug Saf. 2012;11(4):581-96.

Mol PG, Maciulaitis R, Vetter R. GFR decline as an end point for clinical trials in CKD: a view from Europe. Am J Kidney Dis. 2014;64(6):838-40.

Moll S, Chaykovska L, Meier M, et al. Targeting the epithelial cells in fibrosis: a new concept for an old disease. Drug Discov Today. 2013;18(11-12):582-91.

Moll S, Meier M, Formentini I, et al. New renal drug development to face chronic renal disease. Expert Opin Drug Discov. 2014;9(12):1471-85.

Prunotto M, Budd DC, Meier M, et al. From acute injury to chronic disease: pathophysiological hypothesis of an epithelial/mesenchymal crosstalk alteration in CKD. Nephrol Dial Transplant. 2012;27 Suppl 3:iii43-50.

Shahinfar S, Dickson TZ, Ahmed T, et al. Losartan in patients with type 2 diabetes and proteinuria: observations from the RENAAL Study. Kidney Int Suppl. 2002;(82):S64-7.

Shahinfar S, Lyle PA, Zhang Z, et al. Losartan: lessons learned from the RENAAL study. Expert Opin Pharmacother. 2006;7(5):623-30.

Susztak K. Understanding the epigenetic syntax for the genetic alphabet in the kidney. J Am Soc Nephrol. 2014;25(1):10-7.

Tam J, Cinar R, Liu J, et al. Peripheral cannabinoid-1 receptor inverse agonism reduces obesity by reversing leptin resistance. Cell Metab. 2012;16(2):167-79.

Tam J, Godlewski G, Earley BJ, et al. Role of adiponectin in the metabolic effects of cannabinoid type 1 receptor blockade in mice with diet-induced obesity. Am J Physiol Endocrinol Metab. 2014;306(4):E457-68.

Thompson A, Lawrence J, Stockbridge N. GFR decline as an end point in trials of CKD: a viewpoint from the FDA. Am J Kidney Dis. 2014;64(6):836-7.

Woroniecka KI, Park AS, Mohtat D, et al. Transcriptome analysis of human diabetic kidney disease. Diabetes. 2011;60(9):2354-69.

Zhang Z, Shahinfar S, Keane WF, et al. Importance of baseline distribution of proteinuria in renal outcomes trials: lessons from the reduction of endpoints in NIDDM with the angiotensin II antagonist losartan (RENAAL) study. J Am Soc Nephrol. 2005;16(6):1775-80.

Sponsors

The Biochemical Pharmacology Discussion Group is proudly supported by

Mission Partner support for the Frontiers of Science program provided by

Grant Support

This program is supported in part by a grant from Merck and Co., Inc.

Funding for this conference was made possible (in part) by 1 R13 DK103523-01 from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). The views expressed in written conference materials or publications and by speakers and moderators do not necessarily reflect the official policies of the Department of Health and Human Services; nor does mention by trade names, commercial practices, or organizations imply endorsement by the U.S. Government.

Diabetic nephropathy is thought to be a microvascular complication of diabetes caused by angiopathy of the capillaries in the glomerulus, the kidney's main filtration unit. The weakened capillary walls become leaky, resulting in higher levels of protein passing from the blood into urine—a condition known as proteinuria. The leakiness also decreases blood flow and the consequent drop in oxygen levels leads to cellular death, causing glomerular scarring or sclerosis. The interstitial space between the glomerular tubules expands and the tubules are pushed further apart by deposits of collagen and other matrix proteins, leading to tubulointerstitial fibrosis as the disease progresses.

Not all patients with type 1 or type 2 diabetes develop diabetic nephropathy, pointing to environmental and genetic factors in its initiation and progression. But a major impediment to understanding the genetic and molecular mechanisms of DN, and identifying and validating potential drug targets, is the lack of animal models that accurately and completely mimic human disease.

Mouse models of DN do not accurately mimic human diseases because the mice do not develop progressive disease or fibrosis. But the models do share some molecular pathways with human disease. (Image courtesy of Frank C. Brosius)

Frank C. Brosius uses a systems biology approach to overcome the shortcomings of the animal model. By comparing human disease and mouse models at the molecular level, he is creating humanized mice with the dysregulated genetic and signaling pathway seen in diabetic kidney disease in humans but not seen in diabetic mice. He explained how this approach has validated the JAK/STAT pathway as a potential target for therapeutic intervention in DN.

For regulatory approval, drug candidates in human trials must either reduce the likelihood of dialysis or reduce levels of serum creatinine, a marker of renal function. Mouse models of DN do not develop progressive or late-stage disease and therefore cannot be used to test these endpoints. Matthew D. Breyer discussed three new mouse models that develop progressive disease and can be used to test therapeutics with human endpoints.

DN in humans is a progressive disease characterized by increasing levels of protein in the urine, decreasing kidney filtration (as measured by decreasing glomerular filtration rate), and hypertension. (Image courtesy of Matthew D. Breyer)

Although DN is considered a microvascular complication of diabetes, there is growing evidence that it also involves dysfunction in kidney epithelial cells and loss of podocytes, which line the capillaries in the glomerulus. Katalin Susztak is looking for causal pathways in DN progression by analyzing genetics and epigenetics. Her work in mouse models has uncovered a novel way to protect against fibrosis, a hallmark of disease progression in DN.

Fibrosis is triggered in DN by the deposition of collagen and other matrix proteins in the interstitial spaces of the glomerular tubules by myofibroblasts. Benjamin D. Humphreys identified progenitors of these cells, whose origins had been a mystery, as a population of stem cell–like cells that express a marker protein called Gli1. Selective ablation of these cells, Humphreys has shown, can protect against fibrotic progression.

Pericytes are multifunctional cells in the kidney vasculature that provide stability and architectural support to blood vessels and act as sensors that can also repair damage. (Image courtesy of Benjamin D. Humphreys)

Current therapies for DN aim to slow the progression of disease by controlling hypertension using drugs that block the RAAS pathway. But new compounds in the drug development pipeline employ new therapeutic strategies, and Shahnaz Shahinfar introduced the most promising candidates. Matthias Meier, while lauding these developments, explained challenges in drug discovery, translational research, and preclinical testing. Possible solutions include modifying clinical trial design and considering new endpoints.

Aliza Thompson from the U.S. Food and Drug Administration also discussed new endpoints for chronic kidney disease (CKD) drug trials, focusing on decline in glomerular filtration rate (GFR) as a leading choice. GFR is the volume of fluid filtered per unit time from the kidney's glomerular capillaries into a structure called the Bowman's capsule. There are several ways to calculate or estimate GFR, which is typically recorded as milliliters per minute. The normal adult GFR, adjusted for body surface area, is 100–130 ml/min/1.73m². This rate naturally decreases with age. The severity of CKD (stages 1–5) is defined based on GFR levels. In stage 5 disease, in patients who have kidney failure and need dialysis, GFR is less than 15 ml/min/1.73². Thompson described a recent FDA-sponsored workshop that examined how slowed decline in GFR could be used as an endpoint in CKD trials.

The symposium also included two short presentations by early career investigators. Tony Jourdan is investigating the cannabinoid signaling system in podocyte cells and whether blocking this system could be a therapeutic strategy in DN. Kathleen Lincoln is looking for ways to improve treatments by combining current drugs with an agent that blocks an inflammatory pathway in DN.

Speakers:
Frank C. Brosius, University of Michigan
Katalin Susztak, University of Pennsylvania
Benjamin D. Humphreys, Harvard Medical School; Brigham and Women's Hospital

Highlights

• The arrest and reversal of DN in a humanized mouse model upon injection of a JAK1/JAK2 inhibitor offers proof of concept for human trials. Several such inhibitors, including baracitinib, are now in clinical trials.
• Mouse studies suggest that fibrosis progression in DN could be caused by loss of fatty acid oxidation, which is required for optimal kidney function.
• Inhibitors of the Hedgehog–Gli1 signaling pathway could protect against the progression of fibrosis in DN by preventing the proliferation of Gli1-expressing myofibroblasts in the kidney's microvasculature.

A new therapeutic intervention for diabetic nephropathy

The lack of an animal model that adequately replicates human diabetic nephropathy has stalled efforts to understand its genetic and molecular mechanisms. Diabetic kidney disease (DKD) is pathologically dissimilar in humans and mice. Humans develop progressive glomerular sclerosis and tubular interstitial fibrosis with a gradual decline in kidney function; mice show mild or no glomerular or interstitial disease and no loss of kidney function. Despite some common molecular networks implicated in kidney disease, the imperfect mouse models have at best provided limited or indirect information about disease initiation and progression in diabetic patients. In addition, these models have not been useful in preclinical validation of drug targets or drug candidates.

Frank C. Brosius of the University of Michigan used transcriptome (gene expression) analysis of biopsies from patients with CKD and DN to implicate the JAK/STAT signaling pathway in DN progression and kidney function decline. By correlating biopsy results with clinical data, he showed that the pathway is highly activated in different tissue compartments in diseased kidneys.

Several other groups have confirmed elevated JAK/STAT activity, particularly JAK2 activity, in human CKD. But gene expression profiles in diabetic mice did not show these changes. Brosius and his colleagues have therefore created a humanized mouse model of DN by coaxing diabetic mice to express JAK2 protein only in podocytes, cells in the kidney's glomerulus that are critical for its filtration function. When injected with Ang II, a protein that triggers glomerular injury, these mice developed DN within weeks of birth, as shown by histological changes (such as decreased podocyte density and increased glomerular sclerosis) and signs of functional disruptions (such as increased albumin and urea).

The direct effect of JAK2 activity on DN progression was shown when these changes were arrested and reversed after the mice were injected with a JAK1/JAK2 inhibitor. JAK inhibitors are currently used to treat some cancers and autoimmune diseases. Brosius is leading a phase II clinical trial to evaluate baracitinib, an oral JAK1/JAK2 inhibitor, in DKD.

In diabetic mice engineered to express JAK2 exclusively in kidney podocyte cells, progression of DN is halted following injection of a JAK1/JAK2 inhibitor. (Image courtesy of Frank C. Brosius)

Staving off fibrosis in CKD

Renal fibrosis correlates with decline in kidney function in CKD. Katalin Susztak of the University of Pennsylvania is looking for genetic and epigenetic biomarkers of renal disease progression, integrating histological, phenotypic, transcriptional, and epigenetic analyses of patient samples and using mouse models to identify causal pathways.

A transcriptional profile analysis comparing healthy adult and progressive DKD kidney biopsies identified two gene clusters, one involved in inflammation and the other in fatty acid metabolism. A functional genomics approach, investigating how polymorphisms associated with CKD vary the expression levels of nearby genes, then highlighted the dysregulation of the same inflammatory gene cluster. This study found an inverse correlation between kidney function and inflammatory gene expression, pointing to the possible causal role of inflammation in diabetic kidney disease.

Susztak next investigated the dysregulation of fatty acid metabolism in DKD. Epithelial cells in the tubulointerstitium mainly depend on fatty acid oxidation, not glucose metabolism, for energy. Without that energy supply, the cells undergo apoptosis (programmed cell death), leading to tissue atrophy and fibrosis. In CKD tissue samples, key regulators and enzymes required for fatty acid oxidation had decreased activity, and a consequent accumulation of lipids correlated with decreased kidney function.

In mouse models, researchers could prevent accumulation of lipids to protect against fibrosis by increasing the activity of enzymes such as PPARα to increase fatty acid oxidation. (Image courtesy of Katalin Susztak)

Further analysis in CKD mice linked fibrosis not to the mere accumulation of fatty acids (lipotoxicity) but to the loss of fatty acid oxidation. Transgenic CKD mice engineered to express a critical enzyme called pgcA1—which promotes fatty acid oxidation—in tubulointerstitium cells did not develop fibrosis. Susztak observed the same protective effect against fibrosis when CKD mice were injected with fenofibrate, a molecule that triggers the fatty acid oxidation pathway by activating its key enzyme, PPARα.

There are no studies investigating the protective effects of fenofibrate in humans, but data from large trials, including ACCORD (Action to Control Cardiovascular Risk in Diabetes), show that PPARα activation decreases albuminuria, a major diagnostic marker of CKD.

A new therapeutic target in diabetic nephropathy

A hallmark of progressive kidney disease is fibrosis in the tubulointerstitium caused by the deposition of matrix proteins secreted by myofibroblasts. Benjamin D. Humphreys of Harvard Medical School and Brigham and Women's Hospital is studying myofibroblasts' pathogenic role in DN and whether the cells can be therapeutically targeted in kidney disease.

Lineage-tracing experiments in mice showed that the cellular origin of myofibroblasts is a heterogeneous population of perivascular fibroblasts and pericytes. Pericytes are extensively branched cells located in the microvasculature that maintain blood vessel stability, architecture, and function, and can sense and repair damage through functions such as phagocytosis.

The hunt for a molecular marker that would identify myofibroblast progenitors from among these heterogeneous cells in the perivascular space led Humphreys to the Hedgehog signaling pathway, which regulates embryonic development and establishes body plans and organ formation. Extracellular Hedgehog ligands bind to Patch cell receptors to trigger cell signaling that activates Gli1, a transcription activator that turns on various gene networks.

In the kidney, and in the perivascular niches in other organs such as the heart, Humphreys found Gli-expressing cells in the interstitial space and around vascular structures—exactly in the areas where myofibroblasts are found. He showed that these Gli1⁺ perivascular cells have mesenchymal stem cell–like properties, such as self-renewal and the ability to form colonies, and differentiate into multiple cell types. Injecting mice with tamoxifen to simulate kidney injury caused Gli1⁺ cells in the perivascular kidney tissue, and not elsewhere in circulation, to differentiate into myofibroblasts.

To show that these cells are responsible for fibrosis and to understand whether destroying them could protect against fibrosis, Humphreys developed transgenic mice in which only Gli1⁺ cells are engineered to express a receptor for diphtheria toxin, and hence are destroyed upon injection of the toxin. In these mice, erasing the Gli1⁺ population after simulating the development of fibrosis stopped the progression of fibrotic disease.

Arsenic trioxide is an inhibitor of Hedgehog signaling. Humphreys found that an arsenic derivative compound called darinaparsin significantly dampens kidney fibrosis in mice. By selectively knocking down levels of individual proteins in the Hedgehog–Gli pathway using RNAi (RNA interference), he identified the mechanism involved: when Hedgehog signaling is blocked, Gli1 activity is reduced, which in turn downregulates the activity of retinoblastoma protein (pRb) and upregulates a cell cycle inhibitor called p21, thereby causing cell cycle arrest only in myofibroblasts.

Targeting this pathway could be a new therapeutic strategy against fibrotic progression in DN. Several Hedgehog signaling inhibitors are already in clinical trials for various types of cancer. Whether any of these drugs will also work in fibrotic renal disease remains to be seen.

Speakers:
Matthew D. Breyer, Eli Lilly and Company
Matthias Meier, F. Hoffmann-La Roche Ltd.
Aliza Thompson, U.S. Food and Drug Administration
Shahnaz Shahinfar, S. Shahinfar Consulting Inc.; Children's Hospital of Philadelphia

Highlights

• Preclinical evaluation of drug targets and candidates will require new mouse models that mimic features of human DN such as high serum creatinine and decreasing GFR.
• Researchers need a better understanding of disease pathophysiology and of the pharmacokinetic properties of drug candidates, as well as prognostic biomarkers and acceptable surrogate endpoints.
• The FDA supports the use of GFR as a surrogate endpoint in DN clinical trials and considers a 40% decline in GFR to be adequate.
• In addition to drugs blocking the RAAS pathway, drugs targeting growth factors or components of the metabolic, inflammatory, or oxidative stress pathways are under investigation in DN clinical trials.

Clinically translatable mouse models and drug candidates for human trials

Diabetic nephropathy is characterized clinically by an increase in proteinuria, a decrease in kidney function (as defined by the glomerular filtration rate, or GFR), and hypertension. Later in disease progression patients exhibit an increase in serum creatinine, a byproduct of muscle metabolism that kidneys routinely remove from circulating blood. In patients with progressive kidney disease, a decline in GFR correlates with rising serum creatinine levels.

A doubling of creatinine levels, usually seen in patients who have had renal disease for more than 15 years, is used as an endpoint in clinical trials because it predicts the development of end-stage renal disease. Another endpoint is need for dialysis. Only drugs that delay or reduce the likelihood of such primary endpoints that are composites of mortality receive regulatory approval.

This requirement has posed a problem for preclinical validation of drug candidates in mouse models because mice do not develop hypertension or progressive kidney failure. Several years ago, Matthew D. Breyer of Eli Lilly and Company developed a mouse model with hypertension and low GFR, but these symptoms did not worsen in all animals. The model is a type 2 diabetic mouse that is also deficient in endothelial nitric oxide synthase, a vasodilator enzyme that regulates vascular permeability. He has further characterized the model to identify mechanisms of DN.

Creatinine doubling is a clinical trial endpoint that is only seen in patients who have had DKD for more than 15 years. Use of this endpoint has limited trials to patients with advanced disease and limited the utility of preclinical studies in mice, which do not develop progressive kidney disease. (Image courtesy of Matthew D. Breyer)

In a large population of these inbred mice, including animals with progressive kidney disease as well as non-progressors, Breyer discovered the progressors had higher serum creatinine levels as well as increased levels of certain inflammatory proteins, as found in human DN patients. A comparison of renal gene expression profiles in the two mouse populations and in human DN samples identified similar patterns for genes involved in cellular apoptosis and different patterns in JAK/STAT signaling and mitochondrial function among the three groups.

Similar studies in new mouse models of type 1 and type 2 diabetes—both developing high serum creatinine levels and progressive disease—identified genetic signatures of renal failure that are common to both models of DN. The next step is to define the molecular mechanisms of renal failure in these mice. The rationale is that data corroborated across three clinically relevant models will help identify therapeutic approaches that are highly likely to succeed in the clinical setting.

Some inbred eNOS^−/−;LepR^db/db mice develop disease signatures reminiscent of human DN, showing increasing serum creatinine within 6 months of birth. (Image courtesy of Matthew D. Breyer)

Drug discovery and development challenges

Moving on from talks that focused on challenges in drug target discovery and validation, Matthias Meier of F. Hoffmann-La Roche looked at challenges in drug development. Randomized controlled trials have largely failed in DN. Reasons for this failure include inadequate understanding of disease pathophysiology, sparse data on anti-hypertension drug activity in patients, lack of understanding of the pharmacokinetic properties needed in new therapeutic agents, and unavailability of prognostic biomarkers for testing drug candidates.

A multidisciplinary approach that explores changes in the kidney during progression to DKD from mechanistic, molecular, and structural perspectives could improve results. (Image courtesy of Matthias Meier)

Although systems biology approaches and improved tools and assays have identified potential therapeutic targets, the field still lacks information on mechanisms and biomarkers of progressive disease. Despite the availability of new animal models that better replicate human CKD, these animal models are not evaluated using human trial endpoints—hampering the translation of preclinical data to clinical testing.

Primary endpoints in CKD trials are often events that occur late in the disease, driving up study time and costs. Therefore, novel biomarkers are needed to identify patients progressing quickly for recruitment into trials that would reach endpoints sooner.

A surrogate endpoint is a biomarker that can substitute for a clinically meaningful endpoint that occurs too infrequently or too late in the course of a disease. The effect of a therapy on the surrogate is expected to predict the effect the therapy would have on the clinical outcome of interest. Attractive to pharmaceutical companies, surrogate endpoints can lead to faster and cheaper drug development. Sometimes use of surrogate endpoints is the only feasible option for studying a therapy in a disease that progresses very slowly. Acceptable surrogate endpoints for regulatory approval in DN are high on Meier's wish list.

Citing the success of new cancer therapies approved based on surrogate endpoints such as progression-free survival (in place of primary endpoints such as survival benefit), Meier hopes regulatory authorities will approve drugs for CKD based on surrogate endpoints, including decline in estimated GFR over 2 to 3 years.

A clinical trial design that uses biomarkers to selectively focus on patients who are likely to have faster disease progression could speed trials. (Image courtesy of Matthias Meier)

Regulatory approval: GFR as a surrogate endpoint

The discussion of regulatory approval was continued by the FDA's Aliza Thompson, who first explained the agency's position on surrogate endpoints and then reported on a 2012 scientific workshop organized to solicit the CKD community's input on GFR as an endpoint in clinical trials.

All therapies currently used to treat DN were approved based on trials in which the drug slowed the doubling of creatinine levels, which occurs late in the disease course. An analysis published in 2001 showed that the median time between creatinine doubling and end-stage renal disease is 9.3 months. Therefore, using creatinine doubling as an endpoint requires long trials with large sample sizes to detect treatment effects—because not all patients progress to this stage of disease—especially in studies that enroll patients in early stages. A decline in GFR (which reflects smaller changes in creatinine) could be an alternative endpoint.

A 2012 FDA-sponsored workshop published its findings on using GFR as a surrogate endpoint in DN clinical trials in place of endpoints such as creatinine doubling. (Image courtesy of Aliza Thompson)

Findings from the FDA workshop exploring this idea were recently published in a series of peer-reviewed articles. The workshop evaluated data from observational studies, clinical trials, and simulated studies to understand the relationship between established and alternative surrogates with clinical endpoints. Based on this composite analysis, the committee has proposed that "under some circumstances, a GFR decline of 30% could be a valid and useful surrogate end point for progression to kidney failure in clinical trials of CKD." However, the "evidence is stronger for a GFR decline of 40% (which is equivalent to 50% increase in serum creatinine) as the end point, which represents a more cautious approach and is likely to be more widely applicable."

Using the 40% decline in GFR as an endpoint, Thompson cautioned, would probably shorten the size and duration of DN trials, with fewer study subjects progressing to end-stage renal disease. Compared to trials using creatinine doubling as the endpoint, the new trials would therefore provide less direct evidence of how a therapy affects progression to end-stage disease and might require more regulatory oversight on safety issues. She speculated, however, that phase III trials in DN will include either or both of these endpoints as well as composite cardio-renal endpoints.

Moving beyond RAAS blockade: targeting new pathways

The standard of care in DN is to slow disease progression by controlling risk factors such as blood pressure, glucose levels, and lipid levels. The only approved drugs for diabetic kidney disease target the renin–angiotensin–aldosterone system (RAAS), a hormone network that regulates blood pressure and fluid balance. Renin, an enzyme, helps convert angiotensinogen into angiotensin I, which is then converted by angiotensin converting enzyme (ACE) into angiotensin II, a potent vasoconstrictor, thereby increasing systemic and intraglomerular pressure. Angiotensin II also has non-hemodynamic effects in the kidney that contribute to fibrosis and renal scarring.

ACE inhibitors and angiotensin II receptor antagonists can slow kidney disease progression but do not protect patients from end-stage renal disease. Hence, there is a need to look beyond the RAAS blockade to develop therapeutic approaches targeting other pathways and mechanisms.

Like other speakers, Shahnaz Shahinfar of S. Shahinfar Consulting and the Children's Hospital of Philadelphia stressed this point. She guided the audience through the most promising candidates in phase II and phase III trials. While molecules that target the RAAS pathway (such as mineralocorticoid receptor antagonists) continue to be studied, most compounds she highlighted work against growth factors or target components of metabolic, inflammatory, or oxidative stress pathways.

One approach targets the metabolic pathway using inhibitors of sodium-glucose transporter 2 (SGLT-2) rececptors, which are found in the proximal tubule of the nephron and control 90% of glucose reabsorption in the kidney. SGLT-2 inhibitors such as canagliflozin increase elimination of blood glucose through urine, thereby exerting glycemic control, and decrease reabsorption of sodium in proximal tubules. The inhibitors also increase delivery of sodium to distal tubules and the macula densa, which may affect tubular glomerular feedback and the vasoconstriction of preglomerular vessels that decreases glomerular pressure. Canagliflozin is currently in trials evaluating its ability to prevent DN.

Although metabolic and hemodynamic factors are the main causes of DN, recent studies suggest that DKD is an inflammatory process. Metabolic and hemodynamic conditions such as hyperglycemia, dyslipidemia, increased intraglomerular pressure, aldosterone secretion, and oxidative stress can stimulate immunity. Monocyte chemoattractant protein-1 is released from endothelial cells, mesangial cells, podocytes, and tubular epithelial cells, which in turn attract monocytes, macrophages, and T cells, resulting in inflammation and recruiting other pro-inflammatory and profibrotic cytokines (such as TNFα, TGFβ, and CTGF). This immune response promotes fibroblast migration and activation and mesangial cell transformation, leading to fibrosis. Molecules that target CCR2, a chemokine receptor that plays a role in this signaling, are undergoing testing in DN.

Another pathway under investigation is oxidative stress; hyperglycemia activates NADPH oxidases (NOX family enzymes). A phase II study in DN is testing a compound that blocks NOX1/NOX4.

Speakers:
Tony Jourdan, National Institute on Alcohol Abuse and Alcoholism, NIH
Kathleen Lincoln, Boehringer Ingelheim Pharmaceuticals

Highlights

• Inhibiting cannabinoid receptor signaling, which is increased by high glucose levels and leads to lipid production and insulin resistance, is a new therapeutic strategy against DN.
• Nitric oxide–driven endothelial dysfunction and inflammation is another pathway contributing to DN; activating an enzyme that counters the inflammatory effects of NO reduces the progression of DN in a rat model.

Suppressing cannabinoid signaling to reverse the development of DN

The final talks were short presentations on new interventions against pathogenic pathways in DN. The first focused on the role of the endocannabinoid/cannabinoid receptor (CB1R) system in driving kidney complications in diabetes. Endocannabinoids are naturally occurring lipid molecules whose engagement of CB1R promotes food intake, increases lipid production, and causes insulin resistance. The link between an overactive endocannabinoid/cannabinoid receptor system (ECS) and obesity led to the development of CB1R antagonists that reduced body weight and improved metabolic abnormalities.

One such compound, rimonabant, was initially approved in Europe but later withdrawn because it triggered psychotic side effects. Preclinical data later revealed that these adverse effects were related to the compound's interaction with CB1R receptors in the brain; the drug's benefits were due to its activity on CB1R receptors outside the brain. CB1R antagonists that act only in the periphery (do not penetrate the blood–brain barrier) have since been developed for further studies.

To investigate the cannabinoid system in DN, Tony Jourdan of the National Institute on Alcohol Abuse and Alcoholism studied a rat model that develops DN with many features seen in human disease, such as decline in GFR and RAAS activation. This animal model also shows increased CB1R expression in the glomerulus and a loss of podocytes—cells with finger-like projections that wrap around the capillaries in the glomerulus and regulate blood filtering.

Jourdan found that treating these animals with a peripheral CB1R blocker—either before diabetes and accompanying DN development or in later-stage DN—reversed DN, reduced CB1R expression in the glomerulus, and rescued podocyte numbers. Other experiments suggest that both RAAS signaling and high glucose levels increase CB1R signaling in podocytes, driving DN onset. Thus peripheral CB1R blockade is a possible therapeutic strategy.

Improving renal and cardiovascular outcomes in DN

Approved treatments for DN, which include anti-hypertension drugs such as ACE inhibitors, are only moderately effective. Up to 40% of patients with type 2 diabetes who are on these therapies develop the condition. Kathleen Lincoln of Boehringer Ingelheim Pharmaceuticals is seeking to improve the standard of care (SOC) by evaluating the effect of combining enalapril, which is currently in use, with a candidate compound that blocks a nitric oxide–driven endothelial inflammatory pathway in DN.

Normally, the enzyme eNOS (endothelial nitric oxide synthase 3) generates nitric oxide (NO) in blood vessels when activated. Once released, the NO binds to an enzyme called soluble gluanylate cyclase (sGC) to increase the formation of cGMP, a signaling coupler in many cellular processes, resulting in anti-inflammatory effects. In the diseased state, with endothelial damage in the kidneys, sGC is no longer able to bind NO, leading to a decrease in cGMP production and subsequent endothelial dysfunction and inflammation, both of which contribute to renal fibrosis in DN.

Lincoln therefore tested the hypothesis that an sGC activator could restore cGMP levels and slow or arrest the progression of DN. She tested a combination of a candidate compound and enalapril, currently used to treat DN, in a rat model that develops obesity, hyperglycemia, hypertension, and altered renal characteristics reminiscent of DN in humans (proteinuria, glomerulosclerosis, and interstitial lesions).

To mimic the human clinical setting in which patients are on a stable dose of SOC before a new therapy is introduced, Lincoln first treated the rats with enalapril alone and showed that a clinically relevant dose elicited outcomes such as lower proteinuria, lower blood pressure, and fewer interstitial lesions, comparable to results seen in humans. Addition of the candidate sGC activator to the treatment regimen resulted in further significant improvements to all these outcomes. The possibility that the sGC activator in combination with enalapril could reduce the progression of DN beyond that achieved by SOC in humans has yet to be explored.