Bacillus Calmette-Guerin 's beneficial impact on glucose metabolism: evidence for broad based applications

Abstract

Bacillus Calmette-Guerin (BCG) vaccinations improve glycemic control in juvenile-onset Type I diabetes (T1D), an effect driven by restored sugar transport through aerobic glycolysis. In a pilot clinical trial, T1D, but not latent autoimmune diabetes of adults (LADA), exhibited lower blood sugars after multidose BCG. Using a glucose transport assay, monocytes from T1D subjects showed a large stimulation index with BCG exposures; LADA subjects showed minimal BCG-induced sugar responsiveness. Monocytes from T1D, type 2 diabetes (T2D), and non-diabetic controls (NDC) were all responsive in vitro to BCG by augmented sugar utilization. Adults with prior neonatal BCG vaccination show accelerated glucose transport decades later. Finally, in vivo experiments with the NOD mouse (a T1D model) and obese db/db mice (a T2D model) confirm BCG's blood-sugar-lowering and accelerated glucose metabolism with sufficient dosing. Our results suggest that BCG's benefits for glucose metabolism may be broadly applicable to T1D and T2D, but less to LADA.

Keywords: Human metabolism; Immunology.

Graphical abstract

Figure 1

In vivo BCG vaccinations decrease HbA1c in T1D but not LADA patients (A) Juvenile-onset T1D diabetic subjects were compared with LADA diabetic subjects for responsiveness to BCG vaccines using the Tokyo strain over a 2-year period. Juvenile-onset diabetes was defined as age of onset <21 years. LADA subjects were defined as having diabetes onset >21 years. Percent change from baseline in HbA1c of open-label T1D patients receiving BCG treatment with an age of onset (AOO) ≤ 21 years old (n = 6) (mean age of onset of 11 ± 3 years) compared with a T1D reference population not receiving BCG treatment. The fall in HbA1c is a significant trend (repeated-measures ANOVA p = 0.01; left). The trend in the LADA patients was not significant (n = 10; repeated-measures ANOVA p = 0.58; right). Their mean age of onset was 31 ± 2 years. The current chronological ages of the two adults groups were for 28 ± 3 years for the T1D and 45 ± 4 years for LADA subjects. The duration of diabetes in the T1D group was 18 ± 3 years, and the duration of diabetes in the LADA group was 19 ± 2 years. (B) Percent change from baseline in HbA1c of the current open-label T1D patients with AOO ≤21 years (n = 6; 11 ± 3.0 years) receiving Tokyo BCG treatment is shown in red (n = 6) as compared with the previously published Phase 1 Sanofi BCG clinical trial data with similarly early onset (11 ± 5.8 years) (black, open triangle). Also shown are the Phase 1 placebo group (open squares, and a reference population (closed squares).

Figure 2

Glucose transport by T1D, LADA, and NDC human monocytes (A) Example of flow cytometry showing gating and glucose uptake measured with 2-NBDG MFI analysis of nondiabetic control (NDC) monocytes treated in vitro with BCG for 24 hours and then allowed to transport 2-NBDG (sugar) for 1 hour. (B) LADA (red open dots) and T1D (red closed dots) monocytes differ both at baseline (left) and after the in vitro BCG exposure (right) with respect to glucose transport. At baseline, T1D monocytes' basal glucose transport was 7,641 ± 325 and LADA monocytes' was 6,804 ± 170. After BCG, the MFI was 9,299 ± 421 for T1D and 8,331 ± 240 for LADA (p < 0.05 for both comparisons). At baseline LADA have insufficient glucose transport compared with augmented baseline transport in T1D. The comparison of T1D +/− BCG and of LADA +/− BCG is both significant at p < 0.0001. The number of subjects: T1D n = 14; LADA n = 37. (C) We next measured the stimulated glucose index of monocytes from subjects treated with BCG in vivo. As presented above, T1D subjects had improved blood sugar control with BCG but LADA subjects at year 2 had no improvement in their blood sugar control measured in the HbA1c assay. Comparison of the BCG stimulation index (stimulated glucose uptake - baseline) from isolated monocytes from the clinical trial subjects shows increased glucose uptake and greater accelerated uptake of glucose in T1D as compared with LADA. There was no significant difference between the BCG stimulation index of T1D and NDC. Red closed dots represent T1D monocyte samples; black triangles represent nondiabetic control monocytes, open red dots represent LADA monocyte samples. Student's t-testing (unpaired, 1-tailed) or a student's t-testing (paired, 1-tailed) was represented as: p < 0.05 ∗; p < 0.01 ∗∗; p < 0.001 ∗∗∗, p < 0.0001∗∗∗∗. An unpaired t test was used comparing T1D with either LADA or controls. A paired test was used comparing internal to self as untreated monocytes to BCG treated monocytes. The number of subjects: T1D n = 24; LADA n = 12; NDC n = 16. (D) Summary depiction of how, after in vitro BCG exposure, the 2-NBDG sugar uptake assay detects measurable changes in monocytes' sugar transport.

Figure 3

Neonatal BCG vaccinations show sustained and lifelong changes in glucose metabolism (A) Demographics of control subjects vaccinated with BCG at birth is presented by the country of birth, the BCG strain used in that county, and the current age of subjects (a reflection of years since the vaccination). (B) Glucose uptake in monocytes from nonvaccinated control subjects (n = 21, black dots) compared with neonatal BCG vaccinated control subjects (n = 13, green dots). Data present baseline uptake of sugar by monocytes (left) compared with sugar uptake after 24 hour in vitro exposure to BCG (right). (C) Baseline (top) and BCG-stimulated (bottom) monocyte glucose uptake of NDC patients (n = 21) compared on the same scale as NDC subjects given neonatal BCG vaccinations (n = 13) (2-tailed, unpaired t test revealed a significant difference at baseline and with stimulation. Both baseline sugar uptake (top, green) and BCG-stimulated sugar uptake (bottom, green) were monocyte characteristics from normal controls vaccinated at birth with BCG. (D) Baseline (left) glucose uptake in monocytes of unvaccinated T1D patients (n = 42) following a one day incubation with or without in vitro BCG treatment (right). The sample marked in green represents a T1D individual with remarkable diabetes management and a mass found in their lungs suspicious for latent Mycobacterium tuberculosis infection. p < 0.05 ∗; p < 0.01 ∗∗; p < 0.001 ∗∗∗

Figure 4

T2D monocytes transport more sugar with exposures to BCG in culture, in many ways their glucose uptake profile is similar to NDC (A) Untreated monocytes at baseline showing glucose uptake for NDC (n = 21), T1D (n = 42), and T2D (n = 13) subjects. (B) Glucose uptake for monocytes from NDC, T1D, and T2D patients after in vitro treatment with BCG for 24 hrs. (C) Comparison of glucose uptake by monocytes from NDC (n = 17) and T2D (n = 13). Overnight culture in BCG increased glucose uptake in both NDC and T2D monocytes. p < 0.05 ∗; p < 0.01 ∗∗; p < 0.001 ∗∗∗. Red dots represent T1D monocyte samples; black dots represent nondiabetic control (NDC) monocytes, and blue dots represent T2D monocyte samples.

Figure 5

Metformin inhibits glucose uptake in cultured NDC, T1D, and T2D human primary monocytes (A) Metformin inhibited glucose uptake (2-NBDG) in untreated NDC monocytes (n = 9, p = 0.00075). (B) Metformin inhibited glucose uptake (2-NBDG) in NDC monocytes cultured in the presence of BCG (n = 9, p = 0.0047). (C) Metformin inhibited glucose uptake (2-NBDG) in untreated T1D monocytes (n = 18, p = 0.0014). (D) Metformin inhibited glucose uptake (2-NBDG) in T1D monocytes cultured in the presence of BCG. n = 18, p = 3.11 × 10⁻⁶). (E) Metformin inhibited glucose uptake (2-NBDG) in untreated T2D monocytes (n = 6, p = 0.014). (F) Metformin inhibited glucose uptake (2-NBDG) in T2D monocytes cultured in the presence of BCG (n = 6, p = 0.0012). All p values are from paired, 2-tailed student's t test. p < 0.05 ∗; p < 0.01 ∗∗; p < 0.001 ∗∗∗. Red dots represent T1D monocyte samples, black dots represent nondiabetic control monocytes, and blue dots represent T2D monocyte samples.

Figure 6

Mouse models of T1D and T2D support BCG therapeutic effects; in vitro glucose transport assays are consistent (A) Kaplan Meyer survival plot for NOD mice that were untreated (CTRL n = 16), treated with BCG only (+BCG n = 11), treated with metformin only (+Met n = 15), or treated with metformin and BCG (+Met and +BCG n = 15). Metformin treatment started at 6 weeks of age in vivo; BCG injections were administered at week 7 and week 13. Survival was improved after treatment with BCG, but treatment with metformin alone or with a combination of metformin plus BCG resulted in reduced animal survival due to severe hyperglycemia. (B) Glucose uptake (2-NBDG) studies in isolated bone marrow cells from untreated, BCG-treated, or metformin-treated diabetic NOD mice show significant differences between untreated and BCG-treated mice (p = 0.015), as well as between BCG-treated and metformin treated mice (p = 0.011). In all cases, metformin was started at 6 weeks of age, and this was then followed by BCG treatment at 10 weeks. The number of mice: NOD n = 16; NOD + BCG n = 11; NOD + metformin n = 15. (C) Change in blood sugar and BCG treatment regimen for obese db/db mice, a model of T2D. Arrows indicate BCG treatment times. The BCG-treated mice have substantially lower blood sugars. Open symbols depict time points where the difference between CTRL and BCG groups was significant in Student's t testing. Number of mice: db/db + BCG n = 16; db/db untreated n = 11. (D) Change in body weight in the severely obese db/db mice suggests that BCG-treated mice are healthier since they have lower body weight. Open symbols depict time points where the difference between CTRL and BCG groups was significant in student's t testing. The number of mice: db/db + BCG n = 16; db/db untreated n = 11. (E) In vivo BCG treatment of normal BALB/c mice and of diabetic db/db mice both result in increased glucose uptake although the Balb/c mouse effect did not reach statistical significance as measured by 2-NBDG assays in bone marrow cells compared to untreated control mice. BALB/c: n = 4 + BCG and n = 4 untreated; p = 0.31; db/db: n = 10 BCG and n = 10 untreated; p = 0.04.