A Potent and Specific CD38 Inhibitor Ameliorates Age-Related Metabolic Dysfunction by Reversing Tissue NAD+ Decline

Abstract

Aging is characterized by the development of metabolic dysfunction and frailty. Recent studies show that a reduction in nicotinamide adenine dinucleotide (NAD⁺) is a key factor for the development of age-associated metabolic decline. We recently demonstrated that the NADase CD38 has a central role in age-related NAD⁺ decline. Here we show that a highly potent and specific thiazoloquin(az)olin(on)e CD38 inhibitor, 78c, reverses age-related NAD⁺ decline and improves several physiological and metabolic parameters of aging, including glucose tolerance, muscle function, exercise capacity, and cardiac function in mouse models of natural and accelerated aging. The physiological effects of 78c depend on tissue NAD⁺ levels and were reversed by inhibition of NAD⁺ synthesis. 78c increased NAD⁺ levels, resulting in activation of pro-longevity and health span-related factors, including sirtuins, AMPK, and PARPs. Furthermore, in animals treated with 78c we observed inhibition of pathways that negatively affect health span, such as mTOR-S6K and ERK, and attenuation of telomere-associated DNA damage, a marker of cellular aging. Together, our results detail a novel pharmacological strategy for prevention and/or reversal of age-related NAD⁺ decline and subsequent metabolic dysfunction.

Keywords: CD38; NAD(+); SIRTUINS; acetylation; aging; exercise capacity; glucose; progeroid; skeletal muscle.

Figure 1. Characterization of 78c as a specific CD38i

(A) Hydrolase activity of recombinant human CD38 (rhCD38) in the presence of 78c, using 1,N6-ethenoadenine dinucleotide (ε-NAD⁺) as substrate (n=3 experiments, IC50 17.7 nM). Inset shows the structure of 78c. R₁=H; R₂=Me; R₃=trans-4- OCH₂CH₂OMe-cyclohexyl. (B) Effect of different substrate concentration on 78c inhibition of rhCD38 hydrolase activity. Experimental data were fitted by steady state equations derived from the kinetic model in panel (C) Kinetic model of CD38 inhibition by 78c. Model depicts 2 interconnected cycles, one for the hydrolytic activity of rhCD38 (left side), and the other for the cyclase activity of the enzyme as well as the inhibition by its products nicotinamide and ADP- ribose (right side). In the reaction scheme, E represents the enzyme CD38; k_(1-7), the substrate binding constants; K_i, the inhibitor binding constant. (D) NADase activity in tissue homogenates from 1-year-old WT and CD38 KO mice, assayed in the presence of 78c (n=3 experiments). IC50: liver (3.8 NM), skeletal muscle (4.4 nM), spleen (0.7 nM), brain (14.8 nM), ileum (0.4 nM), subcutaneous fat (5.2 nM). AFU= arbitrary fluorescence units. (E) NAD⁺ levels in WT and CD38 KO MEFs treated for 24 hours with 0.2 μM 78c (n=4–6 experiments). NAD⁺ levels were calculated relative to control WT MEF. (F) NAD⁺ levels in tissues of 1-year-old WT and CD38 catalytically inactive (CI) mice treated with 78c or vehicle (Control) for 8 days (n=4 mice per group). Spleen protein lysates of WT and CI mice were immunoblotted for CD38 and Tubulin (right panel). (G) Activity of human recombinant PARP1 in the presence of 78c or the PARP inhibitor olaparib (n=4 experiments). ((H) NAD⁺ levels in A549 cells treated for 24 hours with 0.5 μM 78c and/or 5 μM olaparib (n=3–6 experiments). (I) Activity of human recombinant SIRT1 in the presence of 100 nM 78c or 100 μM SIRT1 inhibitor suramin (n=3 experiments). (J) Activity of human recombinant NAMPT in the presence of 100 nM 78c or 20 μM of NAMPT inhibitor FK866 (n=3 experiments). All values are mean ± SEM. *P < 0.05. NS=not significant. IC50=half maximal inhibitory concentration. See also Figure S1.

Figure 2. 78c ameliorates metabolic dysfunction-associated features in chronologically aged mice

Aged (2-year-old) mice were treated with vehicle (Control) or 78c for up to 14 weeks. (A) NAD⁺ levels in multiple tissues (n=4–13 mice per group) at the end of the treatment. (B) Intraperitoneal glucose tolerance test (ipGTT) at baseline and after 7 weeks of treatment, and corresponding area under the curve (AUC) graph. Analysis by two-way repeated measures ANOVA with Bonferroni's post-tests shows significant interaction between the glucose curve for control and 78c-treated mice. Results in panels B and C show average of two independent experiments (n=18–24 mice per group). (C) Weekly body weight measurements. (D) Serum insulin measurement (ELISA) during ipGTT and corresponding AUC after 5 weeks of treatment (n=5 mice per group). (E) Homeostatic model assessment index for insulin resistance (HOMA IR) (n=5 mice per group). (F) Intraperitoneal insulin sensitivity test (ipIST) and corresponding AUC after 4 weeks of treatment (n=5 mice per group). (G) Liver mRNA levels of glucose metabolism-related genes, determined by quantitative RT-PCR (n=6–19 mice per group) at the end of the treatment. (H) Intraperitoneal pyruvate tolerance test (ipPTT) and corresponding AUC after 9 weeks of treatment (n=5 mice per group). (I) ipGTT of 2-year-old mice after 4 weeks of treatment with vehicle (Control), 78c, FK866 (NAMPT inhibitor), or 78c+FK866 and corresponding AUC (n=5 mice per group). Statistical differences were determined using Two-way repeated measures ANOVA followed by multiple-comparison testing using Bonferroni’s post hoc analysis; *P < 0.05 Control compared with the 78c group. All values are mean ± SEM. *P < 0.05. NS=not significant. See also Figure S2.

Figure 3. 78c promotes muscle functional improvement and protection from damage and fibrosis in chronologically aged mice

1-year-old mice and 2-year-old (aged) mice were treated with vehicle (Control) or 78c for up to 14 weeks. (A) Physical performance assessment on a motorized treadmill. Measurements of distance, maximal running speed, running time, and work after 8 weeks of treatment (n=6–10 mice per group). (B) Measurements of locomotor activity by Comprehensive Laboratory Animal Monitoring System (CLAMS) after 4 weeks of treatment: total activity, rearing, ambulation, and active energy expenditure (AEE) (n=6–9 mice per group). (C) Representative images (upper panel) of laminin-immunostained skeletal muscle tissue sections from vehicle (Control) and 78c-treated 2-year-old mice. The white arrows point to areas with centrally located nuclei (DAPI nuclear staining /blue) within the myofibers marked by laminin (green). Graph shows number of centrally located nuclei (CLN) per unit area of muscle (n=6–9 mice per group). (D) Representative images (upper panel) of skeletal muscle tissue sections of vehicle (Control) and 78c-treated 2-year-old mice immunostained for CD45. CD45-positive cells (red) are indicated by white arrows. Graph shows quantification of CD45-positive cells per unit area of muscle (n=6–10 mice per group). (E) Representative images (upper panel) of skeletal muscle tissue sections of vehicle (Control) and 78c-treated 2-year-old mice immunostained for intracellular immunoglobulin G (IgG). IgG-positive fibers represent necrotic fibers (yellow arrows). Graph shows number of necrotic fibers per unit area of muscle (n=6–9 mice per group). (F) Distribution of myofiber minimum Feret diameter in tibialis anterior (TA) muscle of vehicle (Control) and 78c-treated 1-year-old and 2-year-old mice (n = 6–9 mice per group; >300 fibers counted per mouse). (G) Skeletal muscle mRNA levels of fibrosis and inflammation-related genes in aged mice, determined by quantitative RT-PCR (n = 5–6 mice per group, * P < 0.05, versus control mice). All values are mean ± SEM. Feret diameter linear regression curves were compared using a sum-of-squares F test, *P <0.05. Images show their corresponding scale bars. See also Figure S3.

Figure 4. CD38⁺ cells regulate NAD⁺ levels in CD38^- cells by influencing the availability of NAD⁺ precursors

(A) Immunofluorescent localization of CD38 (red) expression in mouse liver. Sections were co-stained for the pan-leukocyte marker CD45 (green). Hoechst-stained nuclei are shown in blue. Arrow heads (white) indicate lack of CD38 in hepatocytes. Arrows (yellow) show sinusoidal distribution of CD38. Right image depicts a lobular area with a centrally located immune cells cluster. CV= central vein. Scale bar represents 50 μm. (B) Protein lysates of primary mouse hepatocytes, Kupffer cells and AML12 cells were immunoblotted with CD38 and Tubulin antibodies. (C) NAD⁺, nicotinamide mononucleotide (NMN), and nicotinamide riboside (NR) levels in liver and skeletal muscle of aged mice treated with vehicle (Control) or 78c for 10 weeks. Metabolites were measured by high-performance liquid chromatography (HPLC)-mass spec (n=5–6 mice per group). (D) Schematic drawing of the co-culture model. In co-culture experiments, CD38 null cells were plated in the lower chamber and CD38 positive cells were platted in the upper chamber. The lower and upper chambers were separated by a microporous membrane, as demonstrated by the drawing. (E) AML12 cells were plated in the lower chamber of co-culture plates. Upper chamber had no cells (n=8). 0.5 μM 78c was added to the upper chamber, and both chambers were incubated together for 24 hours. Graph shows NAD⁺ levels in the AML12 cells. (F) AML12 cells were plated in the lower chamber of co-culture plates. Upper chamber had Jurkat T cells. 0.5 μM 78c was added to the cells in the upper chamber 4 hours before addition of 100 μM NMN. 4 hours later, both chambers were incubated together for an additional 20 hours. Graphs shows NAD⁺ levels in the AML12 cells (n=3). (G) AML12 cells were plated in the lower chamber of co-culture plates. 293T cells were plated in the upper chamber and then transfected with vector, CD38, or CD38.CI. 24 hours later, 100 μM NMN was added to the upper chamber. 4 hours later, both chambers were incubated together for an additional 20 hours. Graphs shows NAD⁺ levels in the AML12 cells (n=4–6). (H) AML12 cells were plated in the lower chamber of co-culture plates. 293T cells were plated in the upper chamber and then transfected with vector or CD38 plasmid. 1 μM 78c was added to the upper chamber during transfection. 16 hours later, 100 μM Nicotinamide mononucleotide (NMN) was added to the upper chamber and 4 hours later both chambers were incubated for an additional 20 hours. Graphs shows NAD⁺ levels in the AML12 cells (n=8). (D). All values are mean ± SEM. * P < 0.05. (I) AML12 cells were incubated in the presence or absence of 100 ng/ml rhCD38, 100 μM NMN, and 1 μM 78c for 18 hours. Graphs shows NAD⁺ levels in the AML12 cells (n=4). (J) NAD⁺, NMN, and NR levels in liver of 2-year-old mice treated with vehicle (Control), 78c, NR, or 78c+NR. Mice received two doses of vehicle or 78c 22 hours and 6 hours prior to euthanasia. One dose of NR was given by gavage 6 hours before euthanasia. Metabolites were measured by high-performance liquid chromatography (HPLC)-mass spec (n=7–10 mice per group). All values are mean ± SEM. *P < 0.05. NS=not significant. See also Figure S4.

Figure 5. 78c increases sirtuins and PARP activities, activates longevity signaling pathways, and decreases accumulation of DNA damage in vivo

(A-C) 2-year-old (aged) mice were treated with vehicle (control) or 78c for 14 weeks. Protein lysates from skeletal muscle were immunoblotted with specific antibodies. Graph shows quantification of immunoblots. (A) lysine acetylation (ack) and Tubulin (n=5 mice per group). (B) succinyl lysine (suK) and GAPDH (n=6 mice per group). (C) malonyl lysine (maK), Sirt5, and Tubulin (n=5–6 mice per group). (D–J) Telomere Immuno-FISH analysis from liver (n=6–9 mice per group) and skeletal muscle (n=5 mice per group) of 2-year-old mice treated with vehicle (control) or 78c for 14 weeks. (D) Mean number of telomere-associated DNA damage foci (TAF) per nuclei of hepatocytes. (E) Percentage of TAF-positive cells (50–100 hepatocytes were counted). (F) Mean number of γ-H2AX foci per nuclei of hepatocytes. (G) Representative images of γH2A.X immuno-FISH in hepatocytes. Images are Huygens (SVI) deconvolved Z projections of stacks taken with a ×100 objective. White arrows indicate colocalization, and colocalizing foci are amplified in the right panel (amplified images are from single Z planes where colocalization was found). Scale bar: 10 μm. (H) Mean number of TAF per nucleus of skeletal muscle cell. (I) Percentage of TAF-positive cells (40–50 skeletal muscle cells were counted). (J) Mean number of γ-H2AX foci per nucleus of skeletal muscle cell. (K) Representative images ofγ-H2A.X immuno-FISH in myofibers. Images are Huygens (SVI) deconvolved Z projections of stacks taken with a ×100 objective. Yellow arrows indicate colocalization, and colocalizing foci are amplified in the right panel (amplified images are from single Z planes where colocalization was found). Scale bar: 10 μm. All values are mean ± SEM. * P < 0.05. See also Figure S5.