Does dolutegravir/lamivudine reach lymph node HIV reservoirs?

The question behind the dual regimen

Dual antiretroviral therapy with dolutegravir and lamivudine is a recommended first-line regimen, yet it was unclear whether two drugs maintain enough antiviral pressure inside lymph nodes, a major anatomical HIV reservoir. Peripheral blood measurements can look reassuring while tissue compartments behave differently, and lymph node architecture, fibrosis and uneven drug distribution could in principle leave infected cells under-exposed.

The DUALITY trial addressed this by comparing ART-naive participants starting DTG/3TC (the two-drug regimen, 2DR) with those starting dolutegravir plus emtricitabine/tenofovir alafenamide (the three-drug regimen, 3DR). Its prespecified primary endpoint, change in proviral HIV-1 DNA in peripheral blood CD4+ T cells from baseline to week 48, was similar between arms. This paper reports prespecified secondary endpoints: viral persistence and drug penetration and distribution in lymph node tissue.

One lymph node biopsy per participant

Forty participants consented to a single ultrasound-guided inguinal lymph node excision, and 39 were included in the final analysis after one participant in the 2DR group declined. Biopsy timing was randomly assigned within each arm to baseline or weeks 2, 4, 12 or 48, with four participants planned per arm per timepoint. The population was young men with a median plasma viral load of 23,828 copies/mL and a median CD4+ T cell count of 460 cells/mm³ at entry.

Because each person contributed only one biopsy, the time course is built from independent samples pooled across participants rather than repeated sampling of the same lymph node. The authors state that no formal hypothesis testing between the 2DR and 3DR arms was performed, and that the substudy was not designed to establish superiority or equivalence. Mild seroma formation occurred in eight participants (20.5%) and fell to 11.1% after a 48-hour post-biopsy compressive bandage was introduced; no severe complications were observed.

Infected cells fall fast in both arms

HIV-1 DNA-positive and RNA-positive cells were detected by DNAscope and RNAscope in situ hybridization on lymph node sections. Before treatment, both cell types were detectable in all evaluable samples. Baseline counts were numerically higher in the 3DR group: median vDNA was 2.29 versus 8.51 positive cells per 10⁵ cells, and median vRNA was 6.01 versus 15.72 per 10⁵ cells, for 2DR versus 3DR respectively.

After treatment began, both markers declined rapidly in both arms. The median fold change from baseline to week 48 for vDNA-positive cells was 0.158 (95% CI 0.036–0.416) in the 2DR group and 0.055 (0.007–0.692) in the 3DR group. For vRNA-positive cells the corresponding values were 0.049 (0.002–0.532) and 0.001 (<0.001–0.009). The authors caution that the numerically larger reductions in the 3DR arm should be read cautiously given baseline differences between arms and the very limited sample size.

Lymph node vDNA-positive cell levels correlated with total HIV-1 DNA (Spearman's ρ = 0.55, p = 0.0005) and intact HIV-1 DNA (ρ = 0.67, p = 0.0039) in peripheral blood CD4+ T cells, and lymph node vRNA-positive cells correlated with blood cell-associated HIV-1 RNA (ρ = 0.59, p = 0.0221). These exploratory correlations link the tissue measure to the easier-to-sample blood measure.

Penetration ratio is not the same as tissue amount

Drug concentrations in plasma and lymph node homogenates were quantified by LC-MS/MS, including the phosphorylated active metabolites of the nucleoside reverse transcriptase inhibitors. Paired samples were collected roughly 24 hours after dosing in 37 participants (median 26.7 h, range 19.5–32.0 h); two participants sampled earlier were excluded because their concentrations did not represent trough values.

The lymph node-to-plasma tissue penetration ratio was above 1 for tenofovir (4.34), emtricitabine (1.58) and lamivudine (1.46), but well below 1 for dolutegravir (0.18). That ratio does not mean dolutegravir was scarce in tissue: dolutegravir had the highest absolute lymph node concentrations, which the authors suggest likely reflects its high plasma protein binding. Plasma and lymph node concentrations correlated strongly for dolutegravir (ρ = 0.819, p < 0.0001) and weakly for the nucleoside reverse transcriptase inhibitors.

The tissue metabolite-to-parent ratio in lymph node was higher for lamivudine (4.19) than for emtricitabine (1.59) or tenofovir (1.58), which the authors interpret as more efficient phosphorylation to the active metabolite of lamivudine in lymph node tissue. This is a comparison of ratios within tissue, not a direct measure of antiviral effect.

Drug signal surrounds most infected cells

Spatial distribution was mapped by infrared matrix-assisted laser desorption electrospray ionization mass spectrometry imaging, where each measurement corresponds to a voxel of 100×100×10 μm³. Tenofovir concentrations fell below the method's lower limit of quantitation, so spatial analyses of drug distribution and colocalization with infected cells focused on participants in the 2DR group. Because the imaging detection limits exceeded the reported in vitro IC90 values for both dolutegravir and lamivudine, every location where either drug was detected held concentrations above those reference values.

Dolutegravir and lamivudine were consistently detectable, accumulating mainly in the extra-follicular T-cell zone or medulla rather than B-cell follicles. Distribution was heterogeneous, with median relative standard deviations of 60% for dolutegravir and 43% for lamivudine. The median proportion of lymph node tissue covered by at least one detectable drug was 65.3%, driven mainly by dolutegravir at 58.8%, while combined detectable dolutegravir and lamivudine covered 32.0% of tissue area.

Direct colocalization of dolutegravir or lamivudine with infected cells occurred in a median of 45% of vDNA-positive cells and 50% of vRNA-positive cells. When neighboring voxels were included, 60% of vDNA-positive cells and 44% of vRNA-positive cells had detectable drug in adjacent sampling locations, and 100% of vDNA-positive cells and 100% of vRNA-positive cells had at least one neighboring voxel containing dolutegravir or lamivudine. The authors note that CD4+ T cells are highly motile in lymph nodes, so infected cells may move in and out of high-exposure regions.

Fibrosis tracks with weaker drug exposure

Collagen I, measured by immunohistochemistry as a marker of lymph node fibrosis, occupied a median of 9% of lymph node tissue area and showed no consistent change over time after treatment began in either arm. It was not associated with clinical parameters or peripheral blood reservoir measures.

In the 2DR group, colocalization between antiretrovirals and collagen was lower than total antiretroviral tissue coverage, at a median of 49%, and most of that colocalization involved dolutegravir. Greater collagen deposition was associated with lower dolutegravir lymph node-to-plasma ratios, lower dolutegravir tissue concentrations, and reduced drug coverage of lymph node tissue and infected cells, but these relationships did not reach statistical significance. They are hypothesis-generating associations, not demonstrated causes of weaker penetration.

What the spatial evidence does and does not show

The paper's contribution is spatial: it extends peripheral blood observations into lymph node tissue and shows that DTG/3TC reaches regions containing HIV-infected cells, with drug detected at concentrations above reported in vitro inhibitory reference values wherever it was measurable. Combined with the rapid decline in infected cells in both arms, the authors argue this supports biologically plausible antiviral exposure within lymphoid tissue for the dual regimen.

The design limits how far that argument travels. The substudy was exploratory, with small numbers per arm and timepoint, no formal between-arm testing, and baseline imbalance in infected-cell counts. Mass spectrometry imaging could not quantify tenofovir or the phosphorylated metabolites, so the spatial analysis covers dolutegravir and lamivudine only and cannot attribute reservoir decay to individual drugs. The authors explicitly state the study was not designed to establish a causal relationship between local drug distribution and reservoir decay.

The most useful reading is a priority rather than a verdict: for judging whether a reduced-drug regimen is biologically adequate in a tissue reservoir, the relevant comparison is not the lymph node-to-plasma ratio alone but whether detectable drug reaches the microenvironments where infected cells sit. Here dolutegravir's low penetration ratio coexisted with the highest absolute tissue concentrations and the widest tissue coverage, and nearly all infected cells had detectable drug in their immediate neighborhood.

Sources used in this answer