Second-Generation Human Immunodeficiency Virus Integrase Inhibitors Induce Differentiation Dysregulation and Exert Toxic Effects in Human Embryonic Stem Cell and Mouse Models

Abstract

Background: Each year, approximately 1.1 million children are exposed in utero to human immunodeficiency virus antiretrovirals, yet their safety is often not well characterized during pregnancy. The Tsepamo study reported a neural tube defect signal in infants exposed to the integrase strand transfer inhibitor (InSTI) dolutegravir from conception, suggesting that exposure during early fetal development may be detrimental.

Methods: The effects of InSTIs on 2 human embryonic stem cell (hESC) lines were characterized with respect to markers of pluripotency, early differentiation, and cellular health. In addition, fetal resorptions after exposure to InSTIs from conception were analyzed in pregnant mice.

Results: At subtherapeutic concentrations, second-generation InSTIs bictegravir, cabotegravir, and dolutegravir decreased hESC counts and pluripotency and induced dysregulation of genes involved in early differentiation. At therapeutic concentrations, bictegravir induced substantial hESC death and fetal resorptions. It is notable that first-generation InSTI raltegravir did not induce any hESC toxicity or differentiation, at any concentration tested.

Conclusions: Exposure to some InSTIs, even at subtherapeutic concentrations, can induce adverse effects in hESCs and pregnant mice. Given the increasingly prevalent use of second-generation InSTIs, including in women of reproductive age, it is imperative to further elucidate the effect of InSTIs on embryonic development, as well as their long-term safety after in utero exposure.

Keywords: HIV; antiretroviral; hESC; integrase inhibitor; pregnancy.

Figure 1.

Bictegravir (BIC)- and dolutegravir (DTG)-containing combination antiretroviral therapy (cART) are cytotoxic to CA1S and H9 human embryonic stem cells (hESCs). Cell count (A and B), viability (C and D), and apoptosis (E and F) after exposure to no treatment (culture media without 0.1% dimethyl sulfoxide [DMSO], Ø), or 1× C_max (unless indicated) cART regimens, normalized to corresponding 0.1% DMSO vehicle controls (dashed lines) in CA1S (left, n = 5) and H9 (right, n = 6 with an additional n = 4 at 0.5× C_max) cells, respectively. Cell count collected via Trypan blue; viability and apoptosis obtained via flow cytometry. Each data symbol represents an independent replicate of the indicated base drug in combination with 1 of 4 backbones distinguished by their shape (see legend). Regimens were grouped according to base drug and compared with the 0.1% DMSO vehicle control by paired t test. *, P < .05; **, P < .01; ***, P < .001.

Figure 2.

Exposure to bictegravir (BIC)-, cabotegravir (CAB)-, and dolutegravir (DTG)-containing combination antiretroviral therapy (cART) results in loss of pluripotency markers. SSEA-3+ (A and B) and TRA-1-60+ (C and D) expression normalized to corresponding 0.1% dimethyl sulfoxide (DMSO) vehicle controls (dashed lines), after exposure to no treatment (culture media without 0.1% DMSO, Ø), or 1× C_max (unless indicated) cART regimens, in CA1S (left, n = 5) and H9 (right, n = 6, with an additional n = 4 at 0.5× C_max) cells, respectively. Data collected via flow cytometry. Each data symbol represents an independent replicate of the indicated base drug in combination with 1 of 4 backbones distinguished by their shape (see legend). Regimens were grouped according to base drug and compared to the 0.1% DMSO vehicle control by paired t test. *, P < .05; **P < .01; ***, P < .001.

Figure 3.

Bictegravir (BIC), cabotegravir (CAB), and dolutegravir (DTG) human embryonic stem cell (hESC) toxicity and differentiation occur in a dose-dependent manner. Live cell count (A and B), viability (C and D), apoptosis (E and F), SSEA-3+ (G and H), and TRA-1-60+ (I and J) normalized to 0.1% dimethyl sulfoxide (DMSO) control in CA1S (left, n = 3) and H9 (right, n = 6) hESCs treated with 5 (4 in H9) different integrase strand transfer inhibitors at various doses for 3.5 days. Representative SSEA3 + flow plot of DTG dose response in CA1S hESCs compared with the fluorescence minus 1 (FMO) and 0.1% DMSO controls; vertical line denotes SSEA-3 positivity (K). Cell count collected via Trypan blue; all other variables obtained via flow cytometry. Data are presented as mean and standard deviation (A, C, E, G, and I) or as mean and 95% confidence interval (B, D, F, H, and J). Exposed H9 hESCs were compared with the 0.1% DMSO vehicle control by paired t test. *, P < .05; **, P < .01; ***, P < .001.

Figure 4.

Bictegravir (BIC), cabotegravir (CAB), and dolutegravir (DTG) appear to induce differentiation dysregulation of human embryonic stem cells (hESC)s. Heatmap showing the mRNA expression patterns of pluripotency and early germ layer lineage gene markers in (A) CA1S hESCs and (B) H9 hESCs. Results after exposure to 4 different integrase strand transfer inhibitors at 0.5× C_max for 3.5 days (n = 3: drug treatments and untreated [∅] are depicted on each panel). Raw computed tomography values were obtained via reverse-transcription quantitative polymerase chain reaction and normalized to housekeeping genes (GAPDH and ACTB) followed by normalization to the experimental 0.1% DMSO control. Each integer value increase or decrease on the scale indicates a 2-fold change.

Figure 5.

Bictegravir (BIC) exposure reduces litter size and induces resorption in a pregnancy mouse model. Schematic representation of pregnant mice treated with 1× C_max equivalent BIC, dolutegravir (DTG) , cabotegravir (CAB), raltegravir (RAL), or H₂O as a control, starting from gestational day (GD) 0.5 until the day before sacrifice GD 15.5 (A). Table representing number of litters and number of fetuses per treatment group (B). Litter size (C), litter mean fetal weight (D), litter mean placenta weight (E), litter mean fetal to placental weight ratio (F), and resorption rate (G). Data are shown as dot plots with each dot representing 1 litter and the line indicating the mean. Treatment groups were compared by analysis of variance (C, D, E, F, G) or Kruskal-Wallis. *, P < .05; **, P < .01.