Exploratory study reveals far reaching systemic and cellular effects of verapamil treatment in subjects with type 1 diabetes

Abstract

Currently, no oral medications are available for type 1 diabetes (T1D). While our recent randomized placebo-controlled T1D trial revealed that oral verapamil had short-term beneficial effects, their duration and underlying mechanisms remained elusive. Now, our global T1D serum proteomics analysis identified chromogranin A (CHGA), a T1D-autoantigen, as the top protein altered by verapamil and as a potential therapeutic marker and revealed that verapamil normalizes serum CHGA levels and reverses T1D-induced elevations in circulating proinflammatory T-follicular-helper cell markers. RNA-sequencing further confirmed that verapamil regulates the thioredoxin system and promotes an anti-oxidative, anti-apoptotic and immunomodulatory gene expression profile in human islets. Moreover, continuous use of oral verapamil delayed T1D progression, promoted endogenous beta-cell function and lowered insulin requirements and serum CHGA levels for at least 2 years and these benefits were lost upon discontinuation. Thus, the current studies provide crucial mechanistic and clinical insight into the beneficial effects of verapamil in T1D.

Fig. 1. Serum CHGA in response to verapamil treatment of subjects with T1D.

CHGA as assessed by LC-MS/MS (y-axis represents relative abundance levels in zero-centered log2 form) in serum at baseline (BL) or at 1 year (Y1) of individual control subjects with T1D receiving placebo (black) (n = 5) (NS) (a) or verapamil (red) (n = 5) (t₄ = 5.966, *P = 0.0040) (b). Comparison of the changes in CHGA (BL to Y1) as assessed by LC-MS/MS in the verapamil and the placebo group (t₈ = 3.674, *P = 0.0063) (c). Serum CHGA levels at BL or Y1 as assessed by ELISA in individual control subjects with T1D receiving placebo (n = 6) (NS) (d) or verapamil (n = 9) (t₈ = 5.44, *P = 0.0006) (e). Comparison of the changes in CHGA (BL to Y1) as assessed by ELISA in the verapamil and the placebo group (t₁₃ = 4.497, *P = 0.0006) (f). Serum CHGA levels in healthy, non-diabetic volunteers (blue) (n = 9) as assessed by ELISA and compared to subjects with T1D at baseline (T1D BL) (white) (n = 15), subjects with T1D getting placebo for 1 year (Control T1D Y1) (n = 6) or receiving verapamil for 1 year (Verapamil T1D Y1) (n = 9) (F_3,35 = 4.392, *P = 0.01) (g). Correlation of C-pep AUC and serum CHGA (R = −0.62, P = 0.0026) (h). Bars represent means ± SEM. For a, b, d, e, two-tailed, paired Student’s t-test. For c, f, two-tailed Student’s t-test. For g, one way ANOVA and for h, repeated measures correlation coefficient by mixed model. Subject characteristics are listed in Supplementary Table 1. Source data are provided as a Source Data file.

Fig. 2. Insulin requirements, beta cell function, and CHGA over 2 years of T1D treatment with verapamil.

Changes over time in serum CHGA as assessed by ELISA (F_4,19 = 8.723, P = 0.0003) (a), C-peptide AUC (F_4,19 = 4.346, P = 0.012) (b), daily insulin dose (F_4,23 = 3.094, P = 0.036) (c) and blood glucose control as assessed by HbA1C (d) in subjects with T1D receiving verapamil for 2 years (Verapamil; n = 5), discontinuing verapamil after the first year (Disc V; n = 4), or not taking any verapamil (Control; n = 6). Means ± SEM are shown, two-way repeated-measures ANOVA. Subject characteristics of these subgroups are shown in Supplementary Table 4. Source data are provided as a Source Data file.

Fig. 3. Effects of T1D and verapamil treatment on T-cells.

Expression of the T-cell markers CD4 (NS) (a), CXCR3 (NS) (b), STAT4 (NS) (c), CXCR5 (H₃ = 14.011, *P = 0.003) (d) and IL21 (H₃ = 11.516, *P = 0.009) (e) as assessed by qPCR in PBMCs from healthy, non-diabetic volunteers (blue) (n = 7), subjects with T1D at baseline (T1D BL) (white) (n = 10), subjects with T1D getting placebo for 1 year (Control T1D Y1) (grey) (n = 5) or receiving verapamil for 1 year (Verapamil T1D Y1) (red) (n = 5). Serum levels of the proinflammatory cytokine IL-21 as assessed by ELISA (H₃ = 11.847, *P = 0.008) (f). Bars represent means ± SEM; Kruskal-Wallis (nonparametric ANOVA) and Dunn’s multiple comparisons. Source data are provided as a Source Data file.

Fig. 4. Gene expression profile changes in human islets in response to verapamil treatment.

RNA sequencing was performed on isolated human islets from three different individuals (A–C) treated for 24 h with or without verapamil (100 µM) with each donor serving as its own control. Volcano plot contains all genes with a baseMean expression of >500. Those genes with an adjusted DESeq2 P-value < 0.05 (calculated using a Wald test and the Benjamini–Hochberg method) are shown in blue (downregulated) and red (upregulated) (a). Key pathways modulated by differentially expressed genes (b). Heatmap showing key genes changed after treatment with verapamil (color scale represents log2 fold change) (c). Display of the normalized read counts for key genes before and after verapamil treatment of islets from each of the individual islet donors A–C (d).