Combinatorial assessments of brain tissue metabolomics and histopathology in rodent models of human immunodeficiency virus infection

Abstract

Metabolites are biomarkers for a broad range of central nervous system disorders serving as molecular drivers and byproducts of disease pathobiology. However, despite their importance, routine measures of brain tissue metabolomics are not readily available based on the requirements of rapid tissue preservation. They require preservation by microwave irradiation, rapid freezing or other methods designed to reduce post mortem metabolism. Our research on human immunodeficiency virus type one (HIV-1) infection has highlighted immediate needs to better link histology to neural metabolites. To this end, we investigated such needs in well-studied rodent models. First, the dynamics of brain metabolism during ex vivo tissue preparation was shown by proton magnetic resonance spectroscopy in normal mice. Second, tissue preservation methodologies were assessed using liquid chromatography tandem mass spectrometry and immunohistology to measure metabolites and neural antigens. Third, these methods were applied to two animal models. In the first, immunodeficient mice reconstituted with human peripheral blood lymphocytes then acutely infected with HIV-1. In the second, NOD scid IL2 receptor gamma chain knockout mice were humanized with CD34+ human hematopoietic stem cells and chronically infected with HIV-1. Replicate infected animals were treated with nanoformulated antiretroviral therapy (nanoART). Results from chronic infection showed that microgliosis was associated with increased myoinostitol, choline, phosphocholine concentrations and with decreased creatine concentrations. These changes were partially reversed with nanoART. Metabolite responses were contingent on the animal model. Taken together, these studies integrate brain metabolomics with histopathology towards uncovering putative biomarkers for neuroAIDS.

Fig. 1

FBMI euthanasia design and testing. (a) FBMI euthanasia was performed in a Muromachi 10 kW Microwave fixation system. (b) Mice are anesthetized and placed in an animal holder. The holder includes a water compartment filled with ionic buffer solution that aligns the mouse brain in the center of the waveguide. (c) An ASTIM phantom was manufactured in order to develop standard operating microwave irradiation protocols that enable adequate heat distribution in the brain without boiling. (d) The ASTIM phantoms were used to test the effect of irradiation time and buffer solution for the rodents

Fig. 2

¹H MRS evaluations of brain metabolites. (a) Region of interest selected for single voxel ¹H MRS of the mouse brain in vivo and post-microwave fixation. Results shown here demonstrate the effect of adequate and inadequate heat stabilization on metabolites determined retrospectively by comparing individual ex vivo to in vivo lactate levels (12 scans, 16 hs). (b) Lactate is stabilized with adequate heat (open circles, n = 4) and does not change significantly over a 16 hour scan series, but levels double by the first scan interval (60 min) with inadequate heat. (c) Residual anaerobic respiration causes a sharp decrease in NAA levels by the first scan interval when adequate heat is not applied (closed circles, n = 5). Metabolite concentrations are reported as institutional units (IU). Means at each timepoint are displayed with error bars indicating standard error of the mean

Fig. 3

LC-MS/MS measured metabolites of heat stabilized brain tissues. (a) Metabolite levels measured with LC-MS/MS demonstrate increased Glu, GABA and Cho and reduced PCho when flash frozen (solid bars) relative to FBMI fixed tissue levels (open bars). Values shown are calculated as percentage of FBMI group mean. Error bars show standard error of the mean. *statistically significant group differences with p < 0.05

Fig. 4

Morphology of FBMI prepared brain tissue. Immunohistochemical analyses of astroglial (GFAP), microglial (Iba1), neuronal (MAP-2 and NF) and synaptic antigens (VGlut1, VGAT, and Snp1). These are readily observed in brain tissues recovered after FBMI and following with or without PFA fixation and paraffin embedding. GFAP, VGlut1 and Snp1 demonstrate increase signal with FBMI alone. Antibodies against Iba1 and synaptic antigens are preserved and enhanced with FBMI and PFA compared to no FBMI. FBMI reduces NF and nuclear staining with both hematoxylin and DAPI. *significant differences (p < 0.05) compared to non-FBMI. # significant differences (p < 0.05) compared to FBMI alone

Fig. 5

HIV-hu-PBL LC-MS/MS metabolite analyses. FBMI was used in studies of HIV-1 infected human PBL reconstituted immune deficient mice. Metabolites were measured post mortem by LC-MS/MS in multiple subregions following ¹H MRS confirmation of Lac and NAA stability. Individual dissected regions did not demonstrate group differences results in (a) cortex or (b) hippocampus in mice reconstituted with human lymphocytes (open bars, n = 4) compared with HIV-1 infected hu-PBL mice (solid bars, n = 4). Error bars indicate standard error of the mean

Fig. 6

Brain pathology in HIV-1 infected hu-PBL mice. Neuropathologic alterations in the cerebellum of an HIV-1 infected mouse. Human lymphocytes invade the parenchyma adjacent to blood vessel with p24 positive HIV-infected cells. GFAP and Iba1 staining shows inflammatory process in the vicinity. At this acute stage of inflammation gross neuronal morphology visualized by NF and MAP-2 remains intact

Fig. 7

LC-MS/MS metabolomic profiling of brain tissue from HIV-1 infected humanized mice. Metabolomic profiles were generated for flash frozen hippocampal tissues of humanized mice controls (C, n = 7), HIV-1 infection (H, n = 6) and HIV-1 infection treated with nanoART (N, n = 6). Myo-Inositol (mInos), choline (Cho) and phosphorylcholine (PCho) increased and creatine (Cre) decreased during HIV-1 infection compared to humanized mice controls. NanoART reduction of HIV-1 disease reduced mI levels below the increase with HIV-1 infection. Error bars indicate standard deviation. *significant mean differences (p < 0.05) compared with uninfected humanized mouse brain controls, ^#significant mean differences (p < 0.05) between HIV-1 infected and HIV-1 infected nanoART treated

Fig. 8

Hippocampal microgliosis follows HIV-1 infection of humanized mice. Humanized NSG mice were infected with HIV-1_ADA for 18 weeks. Brains were acquired after cervical decapitation and flash freezing. Following immunohistochemical stains of hippocampal brain slices numbers of Iba1 reactive cells were counted in the hippocampus in HIV-1 infected, HIV-1 infected and nanoART-treated and uninfected (controls). Photomicrographs of Iba1-stained sections are shown at 20× with 40× magnifications in areas of the dentate gyrus. (scale bar = 0.2 mm)