Vitamin A Liver Reserves and Serum Markers of Vitamin A in US Adults at Time of Death

Minimal human data exist on actual liver vitamin A compared with blood biomarkers. One blood biomarker, the percent of total serum retinol (vitamin A) in the form of retinyl esters, has been suggested to diagnose hypervitaminosis A with cutoffs of 5% and 10%. In this study, investigators aim to compare total liver vitamin A reserves with the percent total serum retinol as retinyl esters to evaluate hypervitaminosis A using autopsy samples from US adults. Investigators also evaluate the sensitivity (the ability of the biomarker to correctly identify those with deficiency) and specificity (the ability of the biomarker to correctly identify those without deficiency) of serum retinol to determine vitamin A deficiency, variation of liver vitamin A concentration among lobes, and liver alpha retinyl ester concentrations, a cleavage product of alpha-carotene, a vitamin A precursor.

To conduct the study, matched serum and liver samples were procured from 27 US adult cadavers (from donors age 49-101 years) and their vitamin A biomarkers were analyzed. Matched human liver and serum samples were collected from 27 cadavers, 13 male and 14 female, and acquired as a consecutive sample through a request to the National Disease Research Interchange service. De-identified clinical history summaries were also obtained. The donor criteria were as follows: no restrictions on race or sex, no history of chemotherapy and/or radiation, no history of prolonged endotracheal intubation or ventilation, no evidence of sepsis or infectious disease (negative serology panel), and age ranges of 21-54 y, 55-74 y, and >75 y (target collection was 5 male and 5 female donors per age range to provide a non-homogenous sample). After 10 individuals were acquired in each age range (i.e., 55-74 y and >75 y), samples were no longer requested. All tissues and sera were procured within 16 h post-mortem from September 2013 through August 2016. The procurement protocol numbers were DWEN1 002 022, 002 023, and 002 024 for frozen liver, embedded liver, and serum, respectively. Serum samples were isolated via gel separation and then frozen; primary liver samples were snap frozen in liquid N2; a second liver sample from each donor was embedded in optimal cutting temperature compound (OCT) and snap frozen in a cryomold. Two livers were procured almost whole and evaluated for VA distribution within and among lobes. Samples were shipped on dry ice to Madison, WI, and stored in a -80°C freezer.

OCT-embedded liver samples were prepared as 5-µm frozen sections using a cryostat and stained using routine hematoxylin and eosin and Masson's trichrome protocols and reagents (Newcomer Supply, Middleton, WI). Images were captured using a Nikon E-600 bright field microscope (Melville, NY) and Olympus DP-71 digital microscopy camera (Center Valley, PA). Liver samples (1 g) for VA and carotenoid analyses were weighed and then ground with sodium sulfate (4-5 g) in a mortar and pestle using published methods with minor modifications. Purified C-23 beta-apo-carotenol was added to determine extraction efficiencies. Samples were extracted repeatedly with dichloromethane and filtered through WhatmanTM #1 (GE Healthcare Life Sciences, Pittsburgh, PA) into 50-mL volumetric flasks. An aliquot (5 mL) was dried under nitrogen, re-dissolved in 300 µL methanol:dichloroethane (75:25, volume:volume) and an aliquot of 1 µL was injected into a Waters Acquity H-Class ultra-pressure liquid chromatograph (UPLC®) system equipped with a photodiode array detector (Milford, MA). Aliquots were further saponified for quantification of total alpha-retinol as described below for the lobe analyses. For the liver lobe sample evaluations, 1 g samples were extracted in triplicate from each available lobe. A 5-mL aliquot of the extract was saponified with 400 mL potassium hydroxide:water (50:50, weight:volume) at 45°C for 1 h. The reaction was quenched with 0.5 mL water and extracted three times with 0.5 mL hexanes. The hexane layers were pooled and dried under nitrogen, redissolved in 150 µL methanol:dichloroethane (75:25, volume:volume), and an aliquot of 1 µL was injected into the UPLC® system.

Serum samples (0.5 mL) were aliquoted into test tubes for retinol and RE determination; 1.25 X volume of ethanol was added to denature proteins. The internal standard was C23-β-apo-carotenol. Samples were extracted three times with 0.75 mL hexanes; the supernatant fractions were pooled and dried under nitrogen and reconstituted in 100 µL methanol: dichloroethane (75:25, volume:volume). Two µL was injected into the UPLC® under the conditions described below. Serum C-reactive protein (CRP) and alpha1-acid glycoprotein (AGP) were analyzed to evaluate inflammation using ELISA kits (CRP: Cayman Chemical Company; AGP: Abcam). Lycopene and alpha- and beta-carotene concentrations were determined by routine HPLC analysis.

Ultra-high Performance Liquid Chromatograph (UPLC) method A Waters Acquity UPLC HSS C18 1.8 µm VanGuard pre-column was used in conjunction with a Waters Acquity UPLC® HSS C18 column (1.8 µm, 2.1 x 150 mm). The method utilized two solvent mixes programmed for a 29-min gradient. Solvent A was acetonitrile-water-isopropanol (70:25:5, volume:volume:volume) with 10 mmol/L ammonium acetate and solvent B was methanol-isopropanol (75:25, volume:volume). The column temperature was set to 32°C and flow rate was 0.4 mL/min. The gradient began by holding 100% solvent A for 7 min, followed by a 4-min linear transition to 5% A, and another transition to 1% A over 12 min. The solvent gradient was then reversed to 100% A in 2 min and equilibrated for 4 min. Chromatograms were generated at 311 nm for α-retinol and α-REs, 325 nm for retinol and REs, and 450 nm for carotenoids. alpha-REs, derived almost exclusively from beta-carotene in the diet, were quantified as a proxy for long-term vegetable exposure. The concentrations were calculated using standard curves derived from authentic HPLC-purified standards of α-retinol and retinol.