Host microbiome depletion attenuates biofluid metabolite responses following radiation exposure

Abstract

Development of novel biodosimetry assays and medical countermeasures is needed to obtain a level of radiation preparedness in the event of malicious or accidental mass exposures to ionizing radiation (IR). For biodosimetry, metabolic profiling with mass spectrometry (MS) platforms has identified several small molecules in easily accessible biofluids that are promising for dose reconstruction. As our microbiome has profound effects on biofluid metabolite composition, it is of interest how variation in the host microbiome may affect metabolomics based biodosimetry. Here, we ‘knocked out’ the microbiome of male and female C57BL/6 mice (Abx mice) using antibiotics and then irradiated (0, 3, or 8 Gy) them to determine the role of the host microbiome on biofluid radiation signatures (1 and 3 d urine, 3 d serum). Biofluid metabolite levels were compared to a sham and irradiated group of mice with a normal microbiome (Abx-con mice). To compare post-irradiation effects in urine, we calculated the Spearman’s correlation coefficients of metabolite levels with radiation dose. For selected metabolites of interest, we performed more detailed analyses using linear mixed effect models to determine the effects of radiation dose, time, and microbiome depletion. Serum metabolite levels were compared using an ANOVA. Several metabolites were affected after antibiotic administration in the tryptophan and amino acid pathways, sterol hormone, xenobiotic and bile acid pathways (urine) and lipid metabolism (serum), with a post-irradiation attenuative effect observed for Abx mice. In urine, dose×time interactions were supported for a defined radiation metabolite panel (carnitine, hexosamine-valine-isoleucine [Hex-V-I], creatine, citric acid, and Nε,Nε,Nε-trimethyllysine [TML]) and dose for N1-acetylspermidine, which also provided excellent (AUROC ≥ 0.90) to good (AUROC ≥ 0.80) sensitivity and specificity according to the area under the receiver operator characteristic curve (AUROC) analysis. In serum, a panel consisting of carnitine, citric acid, lysophosphatidylcholine (LysoPC) (14:0), LysoPC (20:3), and LysoPC (22:5) also gave excellent to good sensitivity and specificity for identifying post-irradiated individuals at 3 d. Although the microbiome affected the basal levels and/or post-irradiation levels of these metabolites, their utility in dose reconstruction irrespective of microbiome status is encouraging for the use of metabolomics as a novel biodosimetry assay.

Impact of Partial Body Shielding from Very High Dose Rates on Untargeted Metabolomics in Biodosimetry

“…Our validated serum metabolite panel for the current TBI cohorts (16 metabolite panel) was higher than what we established for our previous VHDR studies (3 metabolite panel) (Tables 3 and 4). 30,31 The number of metabolites and their corresponding metabolic pathways are in line with other studies following IR exposure in mouse 54 and NHP models, 68 where we primarily observed changes in amino acids, energy metabolites, and lipids. Of these, lipid concentrations tended to change irrespective of shielding or total body exposure (LysoPC [14:0], [16:1], [18:1], and 18:3).…”

“…TML is a methylated lysine derivative that is a precursor for carnitine synthesis, which was first proposed as a radiation marker after being identified in humans undergoing radiotherapy . Increases in mouse urinary TML levels have been identified in several experiments recapitulating realistic radiation exposures. ,, It also increases at higher doses independent of impaired inflammatory pathways and in mice with depleted microbiomes …”

“…30,31,34 It also increases at higher doses independent of impaired inflammatory pathways 53 and in mice with depleted microbiomes. 54 A previous experiment examining variable low-dose rate IR exposure replicating nuclear fallout revealed an interesting urinary spectral feature (m/z 393.2234, retention time = 1.3 min) with up to a 150-fold change in irradiated mice at early time points (<1 week). 34 Through elucidating the tandem MS spectra, a structure consisting of a hexosamine with a valineisoleucine dipeptide (Hex−V-I) was derived, confirmed to an MSI level 1 identification with synthesis of a chemical standard, and then validated across separate cohorts of mice.…”

“…34 Through elucidating the tandem MS spectra, a structure consisting of a hexosamine with a valineisoleucine dipeptide (Hex−V-I) was derived, confirmed to an MSI level 1 identification with synthesis of a chemical standard, and then validated across separate cohorts of mice. 30,31,54 Mixed model interactions showed significant interaction effects between the microbiome, dose, and time post-irradiation for urinary levels of Hex−V-I in addition to carnitine and creatine. 54 Although we had hypothesized Hex− V-I was associated with kidney function, it was not detected in kidney tissue following IR exposure (unpublished data) and shows relatively equal fold changes for both LBI and UBI cohorts (Figure 3, Tables 1 and 2).…”

“…30,31,54 Mixed model interactions showed significant interaction effects between the microbiome, dose, and time post-irradiation for urinary levels of Hex−V-I in addition to carnitine and creatine. 54 Although we had hypothesized Hex− V-I was associated with kidney function, it was not detected in kidney tissue following IR exposure (unpublished data) and shows relatively equal fold changes for both LBI and UBI cohorts (Figure 3, Tables 1 and 2). Hex−V-I remains a promising candidate for radiation biodosimetry as it shows among the highest fold changes post-irradiation along with creatine, carnitine, and taurine (Table 2).…”