Transplacental exposure to AZT induces adverse neurochemical and behavioral effects in a mouse model: protection by L-acetylcarnitine

Abstract

Maternal-fetal HIV-1 transmission can be prevented by administration of AZT, alone or in combination with other antiretroviral drugs to pregnant HIV-1-infected women and their newborns. In spite of the benefits deriving from this life-saving prophylactic therapy, there is still considerable uncertainty on the potential long-term adverse effects of antiretroviral drugs on exposed children. Clinical and experimental studies have consistently shown the occurrence of mitochondrial dysfunction and increased oxidative stress following prenatal treatment with antiretroviral drugs, and clinical evidence suggests that the developing brain is one of the targets of the toxic action of these compounds possibly resulting in behavioral problems. We intended to verify the effects on brain and behavior of mice exposed during gestation to AZT, the backbone of antiretroviral therapy during human pregnancy. We hypothesized that glutamate, a neurotransmitter involved in excitotoxicity and behavioral plasticity, could be one of the major actors in AZT-induced neurochemical and behavioral alterations. We also assessed the antioxidant and neuroprotective effect of L-acetylcarnitine, a compound that improves mitochondrial function and is successfully used to treat antiretroviral-induced polyneuropathy in HIV-1 patients. We found that transplacental exposure to AZT given per os to pregnant mice from day 10 of pregnancy to delivery impaired in the adult offspring spatial learning and memory, enhanced corticosterone release in response to acute stress, increased brain oxidative stress also at birth and markedly reduced expression of mGluR1 and mGluR5 subtypes and GluR1 subunit of AMPA receptors in the hippocampus. Notably, administration during the entire pregnancy of L-acetylcarnitine was effective in preventing/ameliorating the neurochemical, neuroendocrine and behavioral adverse effects induced by AZT in the offspring. The present preclinical findings provide a mechanistic hypothesis for the neurobehavioral effects of AZT and strongly suggest that preventive administration of L-acetylcarnitine might be effective in reducing the neurological side-effects of antiretroviral therapy in fetus/newborn.

Figure 1. Experimental design of the study.

Figure 2. Effect of AZT and LAC treatments on spatial learning and memory assessed in the Morris Water Maze.

A) Training sessions: Latencies to reach the hidden platform throughout the four sessions of the task. The higher latencies shown by AZT-treated mice in comparison to Control mice throughout the four sessions clearly indicate a learning impairment (* p<0.05, Fisher's LSD post-hoc performed on significant three-way ANOVA interaction), counteracted by LAC treatment, as the AZT+LAC group had latencies comparable to the Control group in the 3rd and 4th session, and significantly different from those of the AZT group (# p<0.05). LAC treatment alone did not modify learning abilities. B) Probe trial: AZT-treated mice spent less time than Control mice in the quadrant where the platform was located during training (* p<0.05 AZT vs Control after t test), indicating a deficit in memory recall. AZT+LAC mice did not differ from Controls. C) No treatment-induced difference was recorded in the swimming speed assessed in a trial with a visible platform. Values are expressed as means ± S.E.M (n = 7–8 mice per group).

Figure 3. Effect of AZT and LAC treatments on plasma corticosterone secretion measured after 15-min restraint stress.

In basal conditions, no significant differences were observed in plasma corticosterone levels between groups (see text for details), while following acute stress AZT-treated mice had markedly enhanced corticosterone release in comparison to Control mice (* p<0.05 AZT vs Control, Student's t test). LAC administration protected from AZT-induced increase of corticosterone. Corticosterone levels were calculated as delta between restraint stress values and basal values. Values are expressed as means ± S.E.M. (n = 7 mice per group).

Figure 4. Effect of AZT and LAC treatments on hippocampal mGlu receptors expression.

(A) Representatives of immunoblots with mGlu1a and mGlu5 receptor antibodies showed a 142 kDa and 130 kDa bands, corresponding to receptor monomers, respectively. Blots with antibodies recognize an epitope common to mGlu2 receptors monomer(s) (100 kDa) and mGlu3 receptors dimers (206 kDa). (B) Results are expressed as the ratio of the optical density (OD) of the mGluR1a, mGluR5 or mGlu2/3 band and the β-actin band. The AZT prenatal treatment caused a significant reduction of mGlu1a and mGlu5 receptors expression in respect to Control mice (* p<0.05, Student's t-test) counteracted by LAC treatment as AZT+LAC mice differ from Control. No differences between groups were found in the mGluR2/3 (summary of OD monomers and dimers). Values are expressed as means ± S.E.M. (n = 6 mice per group).

Figure 5. Effect of AZT and LAC treatments on hippocampal AMPA and NMDA receptors expression.

(A) Representatives of immunoblots with GluR1 and GluR2 receptor antibodies showed a 110 kDa and 100 kDa bands corresponding to AMPA receptor subunits, respectively. Blot with NR1 receptor antibody showed a 130 kDa band, corresponding to NMDA receptor subunit. (B) Results are expressed as the ratio of the optical density (OD) of the GluR1, GluR2 and NR1 band and the β-actin band. AZT prenatal treatment caused a significant reduction of the GluR1 subunit of AMPA receptors expression in respect to Control mice (* p<0.05 AZT vs Control, Student's t-test) that was fully prevented by LAC treatment: in AZT+LAC mice the Glu1 subunit was increased in respect to AZT mice (# p<0.05 vs AZT, Student's t-test) and did not differ from Controls. Expression of the GluR2 subunit of AMPA and NR1 subunit of NMDA receptors was not affected by either treatment. Values are expressed as means ± S.E.M. (n = 6 mice per group).

Figure 6. Effect of AZT and LAC treatments on brain oxidative stress in newborn and adult mice.

(A) Levels of F2-Isoprostane, as marker of lipid peroxidation and oxidative stress, in brain homogenates from newborn mice. F2t-Isop levels in whole brain homogenates showed a trend to increase in samples from AZT prenatally-treated pups, but not in those from LAC or AZT+LAC groups. (B) Protein carbonyl content (Oxyblot) of hippocampal homogenates. The AZT prenatal treatment caused a significant increase of the protein carbonyl content in respect to Control mice. Both LAC and AZT+LAC groups showed a significant decrease of carbonyls in respect to Control mice indicating an antioxidant effect of the LAC treatment. (* p<0.05 vs Control, Student's t-test). Values are expressed as means ± S.E.M. (n = 6–8 mice per group).