Vascular and Neural Complications in Type 2 Diabetic Rats: Improvement by Sacubitril/Valsartan Greater Than Valsartan Alone

Peripheral neuropathy is a devastating complication affecting ∼50% of the population with diabetes (1). Good glycemic control and improved lifestyle are cited as the only options for treatment; however, in subjects with type 2 diabetes, this has provided little benefit (1). Preclinical studies of diabetic animal models have identified multiple factors that can lead to damage of neurons, Schwann cells, and the vasculature (2,3). Given its complex etiology, a successful treatment for diabetic peripheral neuropathy will likely require a combination of early detection, lifestyle changes, and pharmaceutical interventions targeting the mechanisms deemed most responsible for the pathogenesis.

Previously, we have reported that the vasopeptidase inhibitor ilepatril was an effective treatment for vascular dysfunction and peripheral neuropathy in rodent models of types 1 and 2 diabetes (4–9). Vasopeptidase inhibitors are a combination drug consisting of an inhibitor of neprilysin and ACE (10). However, undesirable side effects of this class of drugs in clinical trials prevented further development (11). This led to the development of angiotensin receptor neprilysin inhibitors, which combine the beneficial effects of angiotensin II receptor blockers with neprilysin inhibition while reducing the risk of angioedema (11). LCZ696 was the first representative of this class of drug and combines the dual action of sacubitril and valsartan (11). Sacubitril is a neprilysin inhibitor, and valsartan is angiotensin II receptor antagonist. In a recent clinical trial, sacubitril/valsartan was shown to be an effective treatment for heart failure, with reduced ejection fraction and side effects comparable to ACE inhibitor enalapril (12). Sacubitril/valsartan has subsequently obtained U.S. Food and Drug Administration approval. Because of the targeting similarities of vasopeptidase inhibitors and sacubitril/valsartan, we were interested in examining the efficacy of sacubitril/valsartan on diabetic vascular and neural impairment. We had previously demonstrated that treating diabetic rats with an ACE inhibitor or angiotensin receptor antagonist provided similar benefits on vascular and neural complications (13). We had also previously demonstrated that pharmaceutically or genetically blocking neprilysin activity improved vascular and neural deficits in diabetic rodents (4–9,14). Neprilysin is a protease that degrades a number of peptides that have neuroprotective and vasoactive properties for which expression is increased by hyperglycemia and diabetes (14). Increased expression of neprilysin could compromise vascular and neural function, and blocking activity of both angiotensin II and neprilysin would be expected to reduce oxidative stress and protect biologically active peptides necessary for normal vascular and neural function (14). Thus, we hypothesized that treating a rat model of type 2 diabetes with sacubitril/valsartan would prevent or reverse vascular and neural dysfunction and possibly provide a new source of treatment for diabetic neuropathy.

Sigma-Aldrich (St. Louis, MO) was the primary source of all chemicals used in these studies. Valsartan and sacubitril/valsartan were provided by Novartis (East Hanover, NJ).

Sprague-Dawley (Harlan Sprague Dawley, Indianapolis, IN) male rats 11 weeks of age were housed in an Association for Assessment and Accreditation of Laboratory Animal Care–certified animal care facility, and food (#7001; Harlan Teklad, Madison, WI) and water were provided ad libitum. All institutional (ACORP #5081487) guidelines for use of animals were followed. At 12 weeks of age, rats were separated into six groups. Three groups remained on the control diet (#7001; 4.25% kcal as fat and 3.0 kcal/g; Envigo Teklad, Madison, WI), and the other three groups were fed a high-fat diet (D12451 [45% kcal as fat and 4.7 kcal/g]; Research Diets, New Brunswick, NJ). Rats were maintained on these diets for 8 weeks. Afterward, the rats fed the high-fat diet were treated with streptozotocin (30 mg/kg in 0.1 mol/L citric acid buffer, pH 4.5, i.p.; EMD/Millipore, Billerica, MA) to induce hyperglycemia (7). The rats on the control diet were treated with vehicle. Blood glucose was evaluated 96 h later using glucose-oxidase reagent strips (Aviva Accu-Chek; Roche, Mannheim, Germany), and rats having a blood glucose level of ≥250 mg/dL (13.8 mmol/L) were considered to be diabetic. These rats are a model for late-stage type 2 diabetes and not hyperinsulinemic (6). At 4 (early intervention) and 16 (late intervention) weeks posthyperglycemia, a group of the control and diabetic rats were treated with valsartan (31 mg/kg body wt daily by oral gavage) or sacubitril/valsartan (68 mg/kg body wt daily by oral gavage). The dose of sacubitril/valsartan used was based upon a previous study examining effect of the drug on cardiac remodeling and dysfunction after myocardial infarction (15). Other pharmacokinetic and pharmacodynamic studies have shown sacubitril/valsartan to be safe and well tolerated (16). Valsartan was first dissolved at 12-fold higher than dosing concentration in 1 N sodium hydroxide. This stock solution was diluted with distilled water, and its pH was adjusted to 8.5 with 1 N hydrochloric acid. The solution volume was further increased with distilled water to generate final concentration of 15.5 mg/mL, and the dosing volume was 2 mL/kg body wt. Sacubitril/valsartan was dissolved in distilled water at 34 mg/mL, and the dosing volume was 2 mL/kg body wt. Untreated control and diabetic rats received vehicle. The treatment phase for each of the early and late intervention groups was 12 weeks. A group of control and diabetic rats was examined prior to beginning treatment of the early intervention group to establish a baseline for the pathology present after 12 weeks of high-fat diet and 4 weeks of untreated hyperglycemia.

Glucose utilization was determined by injecting rats with a saline solution containing 2 g/kg glucose, i.p., after an overnight fast (5).

Thermal nociceptive response in the hind paw was measured using the Hargreaves method (7). Data were reported in seconds. Corneal sensation was determined using two separate methods. First, a Cochet-Bonnet filament esthesiometer was used in unanesthetized rats (Luneau Ophtalmologie, Prunay le Gillon, France) (4). Rats were gently restrained by hand and the 6-cm filament advanced to touch the eye. If the rat blinked, the length of the filament was recorded. If the rat did not blink, the filament was shortened by 0.5 cm and the procedure repeated. This process was continued until the rat blinked. Each eye was evaluated. The data were reported in centimeters.

Corneal sensation was also measured by applying buffered isotonic (290 mOsm/L) and hypertonic eye drops (5% sodium solution; Muro 128, Bausch and Lomb; Valeant Pharmaceuticals, Bridgewater, NJ) to the eye of unanesthetized rats (17). For this method, rats were placed in a custom-made restraining apparatus, and 10 min was allowed for the animal to acclimate to the restraint device and lighting. Six complementary metal-oxide semiconductor cameras (Imaging Development Systems, Obersulm, Germany), three per each eye, were positioned in order to observe both eyes simultaneously. Custom software was used to synchronize video streams and obtain images (MATLAB R2012a; The MathWorks Inc., Natick, MA). Video recording was started 30 s before the first epoch. An image from this recording period was used to establish a baseline area of eyelid opening. The first recording epoch began with the addition of 20 µL isotonic solution to the right eye. The recording period was 3 min followed by a 5-min washout and recovery period. The next epoch began with the addition of the hypertonic solution and recorded for 3 min. An image collector was used offline to retrieve video frames from each epoch. Fiji image analysis software was used by a masked technician to measure the visible surface area of both eyes between the upper and lower eyelids. These areas were expressed as a percentage of the baseline area of eyelid opening. These data were plotted versus time, and the area under the curve (AUC) used to quantify corneal sensitivity to topical isotonic and hypertonic saline, based on narrowing of the palpebral fissure in response to corneal nerve reactivity.

On the day of the terminal studies, rats were weighed and anesthetized with Nembutal (50 mg/kg, i.p.; Abbott Laboratories, North Chicago, IL). Subepithelial corneal nerves were imaged using the Rostock cornea module of the Heidelberg Retina Tomograph confocal microscope (4). The investigator acquiring these images was masked with respect to identity of the animal condition. Corneal nerve fiber length was defined as the total length of all nerve fibers and branches (in millimeters) present in the acquired images standardized for area of the image (in square millimeters). A minimum of six images was acquired for each animal. The corneal fiber length for each animal was the mean value obtained from the acquired images and expressed as millimeters per square millimeter.

Motor and sensory nerve conduction velocity was determined as previously described using a noninvasive procedure in the sciatic-posterior tibial conducting system and digital nerve, respectively (7). Motor and sensory nerve conduction velocities were reported in meters per second.

Videomicroscopy was used to investigate in vitro vasodilatory responsiveness of epineurial arterioles vascularizing the region of the sciatic nerve (5). The arterioles used in this study should be regarded as epineurial rather than perineurial vessels. Following isolation and suspension of the vessels, cumulative concentration-response relationships were evaluated for acetylcholine (10⁻⁸–10⁻⁴ mol/L) and calcitonin gene-related peptide (10⁻¹¹–10⁻⁸ mol/L). At the end of each dose-response curve, papaverine (10⁻⁵ mol/L) was added to determine maximal vasodilation.

Immunoreactive nerve fiber profiles innervating the skin from the hind paw were visualized using standard confocal microscopy combined with immunohistochemistry (7). Profiles were counted by two individual investigators that were masked to the sample identity. All immunoreactive profiles were counted and normalized to length, and the data are presented as profiles per millimeter.

Nonfasting blood glucose was determined with Aviva Accu-Chek strips (Roche). Serum was collected for determining levels of free fatty acid, triglyceride, free cholesterol, and activity of ACE using commercial kits from Roche Diagnostics (Mannheim, Germany), Sigma-Aldrich, BioVision (Mountain View, CA), and Bühlmann Laboratories (Schönenbuch, Switzerland), respectively. Serum was also used to determine thiobarbituric acid–reactive substances, nitrotyrosine, and soluble intercellular adhesion molecule-1 (ICAM-1; using the Quantikine rat sICAM-1/CD54 ELISA kit [R&D Systems, Inc., Minneapolis, MN], according to the manufacturer’s instructions) (18). Liver was used to determine activity of neprilysin (19).

Results are presented as mean ± SEM. Comparisons between groups were conducted using a one-way ANOVA and Bonferroni test for multiple comparisons (Prism; GraphPad Software, San Diego, CA). A P value of <0.05 was considered significant.

The experimental design consisted of an early and late intervention protocol, and data for weight change, blood glucose, and serum lipids and oxidative/inflammatory stress markers are provided in Tables 1 and 2, respectively. For all data to be presented, the results shown for the control and untreated diabetic rats for the early intervention are the condition of the control and diabetic rats at start of treatment for the late intervention.

All animals weighed the same at the beginning of the study. At the end of the early intervention protocol, diabetic rats weighed less than control rats, and this was not altered by treatment. Weight was not statistically different between untreated or treated control and diabetic rats in the late intervention protocol. All diabetic rats were hyperglycemic at the end of the studies, and this was not changed with treatment. Data in Supplementary Fig. 1 demonstrate that glucose clearance, a marker of insulin resistance, was impaired in type 2 diabetic rats at the time of treatment for the early intervention protocol. Early or late intervention treatment of control or diabetic rats with valsartan or sacubitril/valsartan did not improve glucose utilization (data not shown).

Serum triglycerides, free fatty acids, and cholesterol were significantly increased in diabetic rats, and this was not changed with early intervention with valsartan or sacubitril/valsartan treatment (Table 1). With late intervention, treatment with valsartan or sacubitril/valsartan reduced serum triglyceride and cholesterol levels compared with untreated diabetic rats (Table 2). Serum triglyceride and cholesterol levels were increased with duration of hyperglycemia (compare untreated diabetic rats between Tables 1 and 2). Treatment of diabetic rats in the late intervention protocol prevented the further increase in serum triglyceride or cholesterol levels.

Serum biomarkers of oxidative and inflammatory stress were significantly increased in diabetic rats after 36 weeks of a high-fat diet and 28 weeks of hyperglycemia (end of late intervention) (Table 2). Treatment of diabetic rats with valsartan or sacubitril/valsartan slowed the increase in thiobarbituric acid–reactive substances and soluble ICAM-1 levels. Serum nitrotyrosine levels remained elevated in treated with valsartan or sacubitril/valsartan diabetic rats. In the early intervention study, only serum nitrotyrosine levels were significantly increased in untreated diabetic rats compared with control rats. Treating diabetic rats early with valsartan or sacubitril/valsartan prevented the increase.

For these studies, we also measured activity of ACE in serum and neprilysin activity in liver in rats prior to early intervention. Serum ACE activity was increased ∼65% in untreated diabetic rats (70 ± 6 and 114 ± 6 units/mL in control and diabetic rats, respectively; P < 0.01 compared with control rats; n = 6). Liver neprilysin activity in these same rats was 2.2 ± 0.1 and 3.5 ± 0.3 mmol 7-amido-3-methylcoumarin/min/mg protein for control and diabetic rats, respectively (P < 0.01 compared with control rats; n = 6).

Determination of motor and sensory nerve conduction velocities has been a standard neurological end point for preclinical and clinical studies. Data in Fig. 1 demonstrate that motor and sensory nerve conduction velocities are decreased in untreated diabetic rats in both early and late intervention protocols with duration of diabetes having little impact on the final measurement. At the time of treatment for early intervention, motor and sensory nerve conduction velocities were 55.4 ± 1.7 and 34.4 ± 0.6 vs. 37.7 ± 1.6 and 28.9 ± 0.8 m/s for control and diabetic rats, respectively (P < 0.05 compared with control rats; n = 12). Early intervention with valsartan did not significantly improve motor or sensory nerve conduction velocity deficit (Fig. 1A). In contrast, early intervention with sacubitril/valsartan interrupted the progression of slowing of both motor and sensory nerve conduction velocity with significant improvement versus untreated diabetic rats observed for motor nerve conduction velocity. In the late intervention study, valsartan treatment and to a greater extent treatment with sacubitril/valsartan significantly improved both motor and sensory nerve conduction velocity (Fig. 1B).

Recently, measurement of density of sensory nerves in the skin and cornea as well as associated biosensitivity assays of these nerves have been promoted as possible surrogate markers for peripheral neuropathy (20,21). At the time of early intervention, thermal sensitivity and intraepidermal nerve fiber density of control and diabetic rats were 11.7 ± 0.6 s and 24.5 ± 1.0 profiles/mm vs. 20.0 ± 0.7 s and 20.2 ± 0.5 profiles/mm, respectively (P < 0.05 compared with control rats; n = 12). After the early intervention phase, thermal sensitivity and intraepidermal nerve fiber density were significantly impaired in untreated diabetic rats, with a greater loss of intraepidermal nerve fibers occurring with an additional 12 weeks of no treatment (Fig. 2A). Early intervention with valsartan partially improved/delayed the impairment in both thermal sensitivity and intraepidermal nerve fiber density, whereas early intervention with sacubitril/valsartan completely reversed the early decrease in thermal sensation and loss of sensory nerve fibers in the skin. With late intervention, valsartan significantly improved thermal sensitivity, whereas intraepidermal nerve fiber density was not improved (Fig. 2B). Late intervention with sacubitril/valsartan also significantly improved thermal sensitivity and partially improved intraepidermal nerve fiber density.

Data in Fig. 3, Supplementary Fig. 2, and Tables 1 and 2 demonstrate the effect of diabetes and treatment with valsartan or sacubitril/valsartan on subepithelial cornea nerve fiber length and cornea sensitivity as determined by a response to a hyperosmotic solution and Cochet-Bonnet filament esthesiometer. Subepithelial nerve fiber length of the cornea at time of treatment for the early intervention was 9.0 ± 0.3 and 5.7 ± 0.5 mm/mm² for control and diabetic rats, respectively (P < 0.05 compared with control rats; n = 12). Corneal sensitivity as measured by responsiveness to a hyperosmotic solution at time of early intervention was 35.5 ± 5.5 and 68.9 ± 6.2 AUC for control and diabetic rats, respectively (P < 0.05 compared with control rats; n = 12). Applying an isotonic solution to the eye of control or diabetic rats elicits no response (data not shown) (17). Over the 12 weeks for the early intervention phase, both cornea nerve fiber length and sensitivity were further impaired (Fig. 3A). Treatment with valsartan provided no benefit, whereas early treatment with sacubitril/valsartan significantly improved/reversed diabetes-induced decrease in cornea nerve fiber length and sensitivity. Following the late intervention phase, cornea nerve fiber length remained significantly decreased compared with control rats but was not further reduced when compared with untreated rats following the early intervention (Fig. 3B). In contrast, cornea sensitivity was further impaired. Treating diabetic rats with valsartan trended to improve cornea nerve fiber length and sensitivity, but these outcome measures remained significantly impaired compared with control rats. Treating diabetic rats with sacubitril/valsartan significantly improved/reversed diabetes-induced impairment in both cornea nerve fiber length and sensitivity. Representative images of subepithelial cornea nerve fibers following the late intervention phase are provided in Supplementary Fig. 2. Cornea sensitivity was also measured by Cochet-Bonnet filament esthesiometer and demonstrated that both valsartan and sacubitril/valsartan treatment following early and late intervention improved cornea sensitivity (Tables 1 and 2). Considering that the rats are being physically handled during this evaluation, it may not be as sensitive as the examination using the hyperosmotic solution.

We have previously demonstrated that a decrease in vascular reactivity of epineurial arterioles, blood vessels that provide circulation to the sciatic nerve, is manifest before detection of impaired motor nerve conduction velocity (22). Therefore, we believe that if treatments are to be successful for diabetic peripheral neuropathy, they need also to provide protection to the neural vasculature. Data in Fig. 4A demonstrate that vascular relaxation to acetylcholine by diabetic rats was further impaired after an additional 12 weeks of no treatment (compare Fig. 4A to Supplementary Fig. 3 [top panel]). Early intervention in diabetic rats with valsartan prevented the progressive impairment, whereas early intervention in diabetic rats with sacubitril/valsartan reversed impaired vascular reactivity to acetylcholine. Late intervention with valsartan was unable to reverse vascular impairment of relaxation to acetylcholine. In contrast, late intervention with sacubitril/valsartan reversed the vascular dysfunction demonstrated by vasodilation to acetylcholine.

Calcitonin gene-related peptide is the most potent vasodilator of mammalian vessels (23). We have also shown that sensory nerves innervating epineurial arterioles of the sciatic nerve contain calcitonin gene-related peptide (24). In untreated diabetic rats, vascular relaxation to midrange doses of calcitonin gene-related peptide is decreased early in epineurial arterioles (Supplementary Fig. 3 [bottom panel]). Data in Fig. 5A and B demonstrate that early or late intervention in diabetic rats with valsartan did not significantly improve calcitonin gene-related peptide-mediated vascular relaxation. However, early or late intervention in diabetic rats with sacubitril/valsartan did significantly improve vascular relaxation to calcitonin gene-related peptide compared with untreated diabetic rats.

Sacubitril/valsartan, a dual-action drug that blocks the angiotensin II receptor and neprilysin activity, has been approved by the U.S. Food and Drug Administration for the treatment of heart failure, and a recent meta-analysis of six randomized trials with ∼12,000 subjects showed that sacubitril/valsartan was associated with less drug risk than a placebo (25–28).

In addition to the widely publicized effect of sacubitril/valsartan on heart failure, other organ systems and disease states may potentially benefit from inhibition of the renin-angiotensin system and neprilysin. In spontaneously hypertensive rats, treatment with sacubitril/valsartan improved endothelial dysfunction (29). Compared with irbesartan alone, dual inhibition of the renin-angiotensin system and neprilysin with irbesartan and thiorphan provided better protection against diabetic retinopathy in rats (30). In humans, neprilysin activity correlated with BMI and measures of insulin resistance with increasing levels in subjects with multiple cardiovascular risk factors (31). In a recent study, Jordan et al. (32) demonstrated that 8 weeks of treatment of obese subjects with hypertension with sacubitril/valsartan was associated with a significant increase in insulin sensitivity. It was postulated that sacubitril/valsartan treatment restored natriuretic peptide deficiency and reduced renin-angiotensin system activity, which are associated with impaired oxidative metabolism and type 2 diabetes. Plasma neprilysin levels have been shown to increase in high-fat–fed mice (31). In this study, we have demonstrated that neprilysin activity is increased in the liver of type 2 diabetic rats. High levels of fatty acids and glucose have been shown to increase neprilysin activity in human microvascular endothelial cells (33). We have demonstrated that neprilysin immunoreactivity and superoxide are increased in epineurial arterioles of diabetic rats, and this increase is associated with vascular dysfunction, which can be prevented with vasopeptidase inhibitor treatment (6,9). We have also demonstrated that endothelial cells of epineurial arterioles express C-type natriuretic peptide that has vasodilatory properties that are decreased by diabetes but protected with vasopeptidase inhibitor treatment (9).

Patients with chronic kidney disease are at increased risk of progression to end-stage renal disease and cardiovascular events and ultimately renal transplant (34). Vasopeptidase inhibitors had shown promise as a treatment for chronic kidney disease in animal models, but as previously discussed, clinical development of this class of drugs was discontinued (35). Sacubitril/valsartan is now being studied as a potential treatment in a population of patients with chronic kidney disease as well as in animal models (34,36). This literature demonstrates that widespread studies are ongoing examining the potential benefits of angiotensin receptor-neprilysin inhibition in renal and cardiovascular diseases. However, the potential benefit of this class of drug on neural complications associated with diabetes has not been reported.

The preclinical studies presented in this study clearly demonstrate that early and late intervention with sacubitril/valsartan improved multiple end points associated with peripheral neuropathy as well as improved vascular reactivity of epineurial arterioles, blood vessels that provide circulation to the sciatic nerve. The effect of sacubitril/valsartan was superior to valsartan alone, indicating that the combination of inhibition of the renin-angiotensin system and neprilysin is more efficacious. We have previously demonstrated that suppressing the renin-angiotensin system in diabetic rats through inhibiting ACE or blocking the angiotensin II receptor improved both vascular and neural function, in part by reducing oxidative stress in vascular tissue (5,7,13,37). We have also demonstrated that blocking neprilysin activity improves neural function (4–9). Malik et al. (38), in a randomized double-blind controlled trial, demonstrated that the ACE inhibitor trandolapril improved diabetic neuropathy in normotensive patients. In cardiovascular autonomic neuropathy, neuropeptides and their receptors play a key regulatory role (39). These neuropeptides include natriuretic peptides and calcitonin gene-related peptide, which are degraded by neprilysin (39,40). Dysregulation of the expression of these neuropeptides can negatively affect cardiac homeostasis (39). Calcitonin gene-related peptide also plays a significant role in peripheral nerve regeneration and Schwann cell proliferation (41,42). Because neprilysin expression is increased in diabetes, protecting calcitonin gene-related and natriuretic peptides, including C-type natriuretic peptide, from abnormal degradation by blocking neprilysin activity is perhaps one mechanism by which sacubitril/valsartan treatment is improving or reversing diabetic peripheral and autonomic neuropathy.

There are a number of other mechanisms that have been shown to cause damage of the nervous system in diabetic peripheral neuropathy including the polyol pathway, accumulation of advanced glycation end products, and activations of poly(ADP-ribose) polymerase, hexosamine pathway, and protein kinase C (43). Many of these pathways and mechanisms lead to dysregulation of the mitochondria and an increase in oxidative and nitrosative stress (14,43). Blocking the renin-angiotensin system has been shown to reduce oxidative stress in obesity and diabetes and improve insulin sensitivity (44).

In summary, the primary finding from these preclinical studies was that sacubitril/valsartan was able to slow progression of diabetes-induced vascular and neural complications as evidence of the results from the early intervention protocol and also stimulates restoration as demonstrated by outcome measures from the late intervention protocol in a rat model of type 2 diabetes. The mechanisms contributing to the beneficial effects observed with sacubitril/valsartan treatment likely include reducing oxidative stress and protecting neuro- and vasoactive peptides. These results provide rationale for the continued study of sacubitril/valsartan as a potential new treatment for diabetic peripheral neuropathy.

Duality of Interest. This material is based upon work supported by Novartis (LCZ696BUS10T). No other potential conflicts of interest relevant to this article were reported.

Author Contributions. E.P.D. performed studies, evaluated data, prepared figures, and reviewed the manuscript. L.J.C., H.S., and A.O. performed studies, evaluated data, and reviewed the manuscript. M.A.Y. performed studies, evaluated data, and wrote the manuscript. M.A.Y. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 27th Annual Meeting of the Diabetic Neuropathy Study Group of the European Association for the Study of Diabetes (NEURODIAB), Coimbra, Portugal, 9–11 September 2017.