Short Communication: Metformin Reduces CD4 T Cell Exhaustion in HIV-Infected Adults on Suppressive Antiretroviral Therapy

Abstract

Increased negative immune checkpoint receptors (NCR) on T cells are linked to T cell exhaustion, dysfunctional effector responses, and HIV viral persistence. Metformin, an oral hypoglycemic agent used for diabetes, may have previously unrecognized beneficial immunologic effects. Using cryopreserved blood from a 24-week pilot study involving 12 virally suppressed HIV-infected individuals randomized 1:1 to metformin versus observation (OBS), we assessed change in the frequencies of T cell activation (CD38⁺HLA-DR⁺) and NCR [programmed cell death protein 1 (PD1), T cell immunoreceptor with Ig and ITIM domains (TIGIT), and T cell mucin-domain containing-3 (TIM3)]. No differences in 24-week change were seen between arms in CD4 or CD8 T cells, in the CD4/CD8 ratio, or in activated (CD38⁺HLA-DR⁺) CD4 or CD8 T cells. However, metformin over 24 weeks led to decreases compared with OBS in single PD1⁺ (percent decrease: −9.6% vs. 7.5%, p = .015), in dual PD1⁺TIGIT⁺ (−15.0% vs. 10.4%, p = .002), and in triple PD1⁺TIGIT⁺TIM3⁺ (−24.0% vs. 8.1%, p = .041) CD4 T cells. Metformin led to no changes in CD8 T cell NCR frequencies. Metformin decreases the frequency of PD1⁺, PD1⁺TIGIT⁺, and PD1⁺TIGIT⁺TIM3⁺ expressing CD4 T cells. This may have relevance to HIV cure strategies and to efforts to mitigate the risk of chronic complications of HIV.

Introduction

Negative immune checkpoint receptors (NCR) are co-inhibitory receptors that actively inhibit T cell activation and are essential in maintaining immune homeostasis. However, continuous overexpression is associated with T cell exhaustion and dysfunctional T cell effector responses. NCR expression is increased during chronic HIV on both CD4 and CD8 T cells. CD4 T cells coexpressing various NCR such as programmed cell death protein 1 (PD1), T cell immunoreceptor with Ig and ITIM domains (TIGIT), and lymphocyte-activation gene-3 (LAG-3) are specifically associated with viral persistence and enrichment of HIV-infected cells.¹ Metformin is a Food and Drug Administration (FDA)-approved drug used as first-line therapy for type 2 diabetes with long-term (>10 years) record of safety. Of interest, this drug is also known to have cardioprotective, anti-inflammatory, and anti-neoplastic benefits that appear to be independent of its ability to lower glucose.² We utilized cryopreserved peripheral blood mononuclear cells (PBMCs) and assessed the effect of metformin on the frequencies of HLA-DR, CD38, PD1, TIGIT, and T cell immunoglobulin and mucin-domain containing-3 (TIM3) expressing T cells.

Materials and Methods

Cryopreserved PBMC were retrieved from a previously conducted, open-label, 24-week clinical trial of metformin versus observation (OBS) conducted in 12 HIV-infected individuals, age >45 years, stable >1 year on antiretroviral therapy with plasma HIV RNA <50 copies/mL (ClinicalTrials.gov NCT02383563). The study randomized participants 1:1 to receive either metformin 500 mg extended release (ER) daily increasing to 1,000 mg ER daily at week 4 or to OBS. The study excluded individuals with history of diabetes, hepatitis B or C co-infection, uncontrolled medical condition or cancer, calculated creatinine clearance <60 mL/min, and recent history of illicit substance or alcohol use likely to interfere with the patient's ability to comply with protocol requirements. The study was approved by the Committee on Human Subjects of the University of Hawaii and written informed consents were obtained from all participants. All participants also signed a separate informed consent document agreeing to the use of their banked specimens for other research related to HIV.

Cryopreserved PBMCs were rapidly thawed, stained for viability with a live/dead aqua amine reactive dye, washed again with phosphate-buffered saline supplemented with 2% fetal bovine serum followed by fluorochrome-conjugated monoclonal antibodies (mAbs) as previously described.³ The following mAbs were used to identify exhausted and activated T cell populations: ECD-conjugated anti-CD3 (Clone: UCHT1) (Beckman Coulter, Fullerton, CA), Alexa Flour 700-conjugated anti-CD4 (RPA-T4), PE-Cy5-conjugated anti-CD38 (HIT2), FITC-conjugated anti-HLA-DR (G46-6) (BD Biosciences, San Jose, CA), Qdot 605-conjugated anti-CD8 (3B5) obtained from Invitrogen (Carlsbad, CA), PerCP-eFluor 710-conjugated anti-TIGIT (MBSA43) (eBiosciences, San Diego, CA), APC-conjugated anti-PD1 (EH12.2H7) (Biolegend, San Diego, CA), and anti-TIM3 (344823) (R&D Systems, Minneapolis, MN). Cells were washed twice and fixed with 1% paraformaldehyde (PFA; Electron Microscopy Sciences, Hatfield, PA) before acquiring (within 12 h) on a custom BD Bioscience four laser LSR Fortessa flow cytometer. Data were analyzed using FlowJo Software version 9.5 (Treestar, Ashland, Oregon). Median values are presented with the first and third quartiles. NCR data are presented as both absolute changes (week 24 – entry value), and percent changes (absolute change divided by entry value × 100). Wilcoxon rank sum test was used to compare 24-week changes in various parameters.

Results

Study participants consisted of 12 individuals, 6 randomized to the metformin arm and 6 to the OBS arm. Study participants had a median age of 57 years, 92% were men and 50% were Caucasian, with median CD4 count of 598 cells/μL. The two arms did not differ significantly by demographic characteristics or antiretroviral regimens. The metformin arm compared with OBS showed nonsignificant but somewhat longer duration since HIV diagnosis (22 vs. 13.5 years), slightly lower current (489 vs. 725 cells/μL), and nadir (10 vs. 52 cells/μL) CD4 count and a slightly higher frequency of TIGIT⁺ CD4 T cells (15.4% vs. 10.8%) (Supplementary Table S1).

No significant differences in change over 24 weeks were noted in Homeostatic Model Assessment for Insulin Resistance (HOMA-IR), or in CD4 or CD8 T cells, in the CD4/CD8 ratio or in activated (CD38⁺HLA-DR⁺) CD4 or CD8 T cells. Compared with OBS, metformin led to significant decreases over 24 weeks in the percentages of some single, dual, and triple combinations of PD1, TIGIT, and TIM3-expressing CD4 T cells. The metformin arm showed a median absolute decrease in PD1⁺ CD4 T cell frequency of −1.6 (−4.7, 0.2) compared with an increase of 2.0 (0.4, 3.7) in the OBS arm (p = .026), corresponding to a 9.6% decrease from baseline (−16.1, 0.9) in the metformin arm and 7.5% increase (1.7, 20.7) in the OBS arm (p = .015). Similar decreases in the metformin arm compared with the OBS arm were seen for double PD1⁺TIGIT⁺CD4 T cells [absolute change −0.9 (−2.9, −0.1) vs. 0.8 (0.2, 1.2), p = .002; percent change −15.0% (−25.1, −2.6) vs. 10.4% (7.6, 20.2), p = .002] and for triple PD1⁺TIGIT⁺TIM3⁺CD4 T cells [absolute change −0.9 (−1.3, −0.1) vs. 0.3 (0.07, 0.6), p = .041; percent change −24.0% (−28.0, −5.8) vs. 8.1% (3.6, 19.3), p = .041]. These changes are visually given in Figure 1. Entry and week 24 levels of CD38⁺HLA-DR⁺CD4 T cells correlated positively with their corresponding levels of single, double, and triple NCR-positive CD4 T cells (all p < .05).

FIG. 1.

Percent of single PD1⁺ (top row A and B), double PD1⁺TIGIT⁺ (middle row C and D), and triple PD1⁺TIGIT⁺TIM3⁺ (bottom row E and F) CD4 T cells at week 0 (entry) and at week 24 shown for each individual and as medians (bold line) separately for the OBS arm (left column) versus metformin arm (right column). Percent change [(week 24 – entry)/entry * 100] is shown under each graph. For each row, change over 24 weeks in the metformin arm was significantly different compared with change in the OBS arm (all p < .05). OBS, observation; PD1, death protein 1; TIGIT, T cell immunoreceptor with Ig and ITIM domains; TIM3, T cell mucin-domain containing-3.

In contrast, there were no significant differences in 24-week change in any single, double, or triple NCR-expressing CD8 T cell population. Supplementary Tables S2 and S3 detail changes in various CD4 and CD8 T cell NCR combinations.

Discussion

Our study found that metformin over 24 weeks leads to a 9.6% decrease in PD1⁺, a 15.0% decrease in dual PD1⁺TIGIT⁺, and a 24.0% decrease in triple PD1⁺TIGIT⁺TIM3⁺ CD4 T cells in virally suppressed HIV-infected adults. These changes were unique to CD4 T cells and not observed in the CD8 T cell compartment. As the primary mode of action of metformin involves a decrease in cellular respiration induced by a mild and specific inhibition of the respiratory chain complex 1 (NADH: ubiquinone oxidoreductase) in the mitochondria,⁴ it is possible that its impact on NCR expression stems from its ability to alter the cellular metabolism of CD4 T cells. In line with the increasing interest in targeting immuno-metabolism for HIV cure and the prevention of chronic HIV-associated inflammation,⁵ the finding that metformin leads to a decrease in NCR coexpressing CD4 T cells is of possible interest in both areas. Studies by our group and others have identified PD1 and TIGIT expressing and coexpressing CD4 T cells as highly enriched for integrated HIV DNA and strongly associated with the size of the viral reservoir even after adjusting for nadir and current CD4 T cells.^1,3 Metformin may therefore have utility as an adjunctive drug in HIV cure strategies.

In the general population, studies have reported that metformin reduces cancer, cardiovascular disease, dementia, and other age-related comorbidities.^6–8 We have recently reported that higher PD1, TIGIT, and TIM3 on CD4 T cells are associated with subclinical atherosclerosis as assessed by coronary artery calcium (CAC)⁹ and lower neuropsychological performance.¹⁰ The potential use of metformin as adjunctive therapy to decrease age-related conditions of HIV might be considered.

Our study is limited by the small sample size and the short duration of therapy. However, the concept of “repurposing” metformin as a drug to decrease HIV viral persistence and/or to decrease the risk of age-related complications in this population is attractive as the drug is inexpensive and well tolerated. Further research on metformin should be considered.

Funding Information

This study was supported by P30GM103341 Shohet, R (PI), P20GM113134 Gerschenson, M (PI) and U54MD007584 Hedges, J and Mokuau, N (PI).

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3