Letter to the editor: Mitochondrial DNA patterns among individuals living with HIV-1 on dolutegravir with hyperglycemia

Dear Editors,

While DTG is globally recommended for first-line HIV treatment due to its efficacy and safety profile[1], reports from Sub-Saharan Africa, including Uganda[2–4], have raised concerns about incident hyperglycemia, particularly in ART-experienced individuals with prior exposure to thymidine analogs. The mechanistic link between DTG and hyperglycemia remains unclear. One possibility is that DTG-related hyperglycemia reflects cumulative mitochondrial toxicity from prior ART[5], which becomes clinically manifest after DTG initiation. Deletions in mtDNA can impair oxidative phosphorylation, contributing to insulin resistance and beta-cell dysfunction[6].

We sought to assess the 4977 base pair mitochondrial DNA (mtDNA) common deletion which can disrupt several electron transport chain genes and tRNAs and has been associated with various metabolic disorders[7] in individuals living with HIV who developed hyperglycemia after initiating dolutegravir (DTG)-based antiretroviral therapy (ART) and those who did not.

We report findings from a nested case-control study conducted within a retrospective cohort of HIV-positive adults receiving antiretroviral therapy (ART) at the Infectious Diseases Institute (IDI) in Kampala, Uganda. The parent cohort, known as the “Outcomes of HIV-Infected Individuals After Ten Years on Antiretroviral Treatment” (ALT study), has been previously described in detail [8]. In brief, the ALT study was initiated in April 2004 and enrolled 1,000 adult participants who were followed every three months for ten years to evaluate long-term outcomes of ART. During each visit, comprehensive data were collected on demographics, history of previous and current opportunistic infections, non-HIV-related clinical events, WHO clinical stage, vital signs, and ART regimen. A physical examination was conducted, and laboratory tests—including CD4 cell count, HIV-1 viral load, and blood chemistry—were performed. In addition, plasma/serum and whole blood specimens were stored for future analyses.

Participants were eligible for inclusion in the nested case-control analysis if they had documented normal blood glucose test results within three months before or after switching to dolutegravir (DTG), and at least one follow-up blood glucose measurement performed after being on DTG for a minimum of three months. Participants were excluded if they had only a single available blood glucose result or if hyperglycemia was documented within three months before or after the switch to dolutegravir.

Cases were defined as individuals who developed incident hyperglycemia following the initiation of dolutegravir. Hyperglycemia was classified according to the Division of AIDS (DAIDS) criteria [9], which define fasting blood glucose levels of ≥ 6.11 mmol/L or random blood glucose levels of ≥ 6.44 mmol/L as hyperglycemic. Controls were individuals from the same cohort who did not develop hyperglycemia during follow-up. Controls were matched to cases in a 1:1 ratio based on age and sex.

We used qPCR to assess the mtDNA⁴⁹⁷⁷ deletion in whole blood, amplifying the deletion breakpoint and a reference region (HVR2) using SYBR Green chemistry. Relative deletion levels were calculated using the 2^−ΔΔCt method. To evaluate mitochondrial function, mRNA expression of mitochondrial genes (ND4, ND5, ND1, ND6, CYTB, COX1) was measured in cases with ≥ Grade 2 hyperglycemia and matched controls. Total RNA was extracted, reverse transcribed, and quantified by qPCR, using RPL32 as the endogenous control.

Primary data were obtained from the institutional database and combined with laboratory results from the nested study. The primary outcome was the difference in mtDNA⁴⁹⁷⁷ deletion levels between cases and controls. Cases were categorized into Grade 1 and ≥ Grade 2 hyperglycemia. Numerical data were summarized as medians (IQR) and categorical data as frequencies. One-way ANOVA compared baseline characteristics across groups. Differences between cases and controls were analyzed using the Wilcoxon rank sum test. A sub-analysis compared seven ≥ Grade 2 hyperglycemia cases to matched controls. Statistical significance was set at a P-value < 0.05.

The primary and nested studies received ethical approval from the Joint Clinical Research Centre Institutional Review Board and the Makerere University School of Biomedical Sciences Research Ethics Committee, respectively. Both studies were also approved by the Uganda National Council for Science and Technology under protocols HS 1586 and HS 2962ES. Written informed consent was obtained from all participants prior to the initiation of study procedures, in accordance with the Declaration of Helsinki.

Results

Dolutegravir was initiated in 711 of the 1,000 participants enrolled in the ALT cohort (supplementary material). Of these, 648 had both a baseline and a follow-up blood glucose measurement. Among the 648, 259 participants had their first blood glucose test performed within three months before or up to three months after initiating dolutegravir. Forty of these 259 individuals were excluded due to the presence of hyperglycemia within the three-month window surrounding dolutegravir initiation, leaving 219 eligible participants.

Among the 219 participants, 40 (18%) developed incident hyperglycemia following dolutegravir initiation, including 4% with grade ≥2 hyperglycemia. These 40 cases were matched 1:1 by age and sex with 40 controls who remained normoglycemic, resulting in 80 total samples selected for analysis. Three cases failed qPCR amplification, and their matched controls were also excluded, leaving 74 samples for the final mitochondrial DNA (mtDNA⁴⁹⁷⁷) deletion analysis. mRNA expression analysis was conducted on the seven cases with grade ≥2 hyperglycemia and their matched controls. Two controls failed amplification, resulting in a final sample of seven cases and five matched controls for the mRNA analysis.

The mean baseline body mass index (BMI) was higher among cases compared to controls (supplementary material), with borderline statistical significance (mean [SD] BMI: 25 [4] kg/m² vs. 23 [4] kg/m² p = 0.05). Prior exposure to stavudine, a known mitochondrial toxic agent[5], was similar between the two groups.

There was no statistically significant difference in median mtDNA⁴⁹⁷⁷ deletion levels (as measured by 2^−ΔΔCt values) between all cases and controls. Cases had a median 2^−ΔΔCt value of 1.1 (interquartile range [IQR]: 0.8–1.9), while controls had a median of 1.1 (IQR: 0.6–1.7) (p = 0.111) (Figure 1A). However, participants with grade ≥2 hyperglycemia had significantly higher median 2^−ΔΔCt values [2.0 (IQR: 1.1–2.8)] compared to their matched controls [1.1 (IQR: 0.6–1.9)], with a statistically significant difference (p = 0.048) (Figure 1B).

Figure 1.

Comparison of the relative extent of deletion in the 4977 bp region of the mitochondrial genome (as measured by the 2^−ΔΔct method) between individuals living with HIV on dolutegravir with hyperglycemia (cases) and without (controls) hyperglycemia. A: Main analysis. Cases (participants on dolutegravir with ≥ grade 1 hyperglycemia, n=37) matched with controls (participants on dolutegravir with normal blood glucose, n=37). Dolutegravir-associated hyperglycemia was defined as fasting blood glucose ≥ 6.11 mmol/L or random blood glucose ≥ 6.44 mmol/L following a normal baseline result taken ±3 months following switch to dolutegravir. There was no difference in relative deletion levels in cases compared to controls (p=0.111, Mann- Whitney U test). B: Sub-analysis. Cases (participants on dolutegravir with ≥ grade 2 hyperglycemia, n=7) matched with controls (participants on dolutegravir with normal blood glucose, n=7). Dolutegravir-associated hyperglycemia was defined as fasting blood glucose ≥ 6.95 mmol/L or random blood glucose ≥ 8.89 mmol/L following a normal baseline result taken ±3 months following switch to dolutegravir. Cases had significantly higher 2^−ΔΔct of deletion than controls (p=0.048, Mann-Whitney U test). 2^−ΔΔct is the change in the gene amplification of cases in reference to controls when the amplification of the gene of interest is normalized to a housekeeping gene. The error bars represent interquartile range.

In the sub-analysis of cases with grade ≥2 hyperglycemia versus matched controls, cases exhibited higher expression levels across all mitochondrial genes assessed (ND4, ND5, ND1, ND6, CYTB, and COX1). However, these differences were not statistically significant (supplementary material).

Discussion

Our findings suggest that mtDNA deletions may contribute to DTG-associated hyperglycemia, particularly in more severe cases. While overall mtDNA deletion levels did not differ between all cases and controls, the elevated deletions observed among individuals with grade ≥2 hyperglycemia support a potential threshold effect. This aligns with the concept of mitochondrial heteroplasmy, in which disease manifests when the proportion of mutated mtDNA exceeds a functional threshold[10].

Mitochondrial dysfunction is a known contributor to metabolic disorders such as type 2 diabetes mellitus (T2DM)[11], and prior ART—particularly stavudine—is known to induce mtDNA toxicity[5]. In this context, DTG may unmask or exacerbate latent mitochondrial injury in susceptible individuals. The uniform trend of increased mitochondrial gene expression among more severely affected cases may reflect an early compensatory mechanism preceding overt dysfunction.

This study provides preliminary mechanistic insight into DTG-associated hyperglycemia in ART-experienced individuals, particularly within African populations where integrase inhibitor-related T2DM risk appears highest. Genetic or epigenetic factors influencing mitochondrial resilience warrant further exploration.

Conclusions

While no significant differences in mtDNA deletions were observed across all hyperglycemia cases, a subset with grade ≥2 hyperglycemia exhibited elevated deletions, supporting a potential mitochondrial contribution in a retrospective nested study that had too few hyperglycemic outcomes to be conclusive. These preliminary findings underscore the need for larger mechanistic studies focused on mitochondrial integrity and its role in DTG-associated metabolic outcomes.