Trace mineral imbalances can have significant effects on animal health, reproductive success, and survival. Monitoring their status in wildlife populations is, therefore, important for management and conservation. Typically, livers and kidneys are sampled to measure mineral status, but biopsies and lethal-sampling are not always possible, particularly for Species at Risk. We aimed to: 1) determine baseline mineral levels in Northern Mountain caribou (Rangifer tarandus caribou; Gmelin, 1788) in northwestern British Columbia, Canada, and 2) determine if hair can be used as an effective indicator of caribou mineral status by evaluating associations between hair and organ mineral concentrations. Hair, liver, and kidney samples from adult male caribou (n­­_Hair= 31; n_Liver, n_Kidney= 43) were collected by guide-outfitters in 2016-2018 hunting seasons. Trace minerals and heavy metals were quantified using inductively-coupled plasma mass spectrometry, and organ and hair concentrations of same individuals were compared. Some organ mineral concentrations differed from other caribou populations, though no clinical deficiency or toxicity symptoms were reported in our population. Significant correlations were found between liver and hair selenium (rho=0.66, p<0.05), kidney and hair cobalt (rho=0.51, p<0.05), and liver and hair molybdenum (rho=0.37, p<0.10). These findings suggest that hair trace mineral assessment may be used as a non-invasive and easily-accessible way to monitor caribou selenium, cobalt, and molybdenum status, and may be a valuable tool to help assess overall caribou health.

Study Population and Sample Collection

We studied 7 herds of the Northern Mountain ecotype of woodland caribou (R. t. caribou) occurring in the traditional territory of the Tahltan Nation (northwestern British Columbia, Canada) (Fig 1). No animals were killed for the purposes of this study. Caribou were legally hunted in their natural habitat by authorized non-resident hunters during the fall hunting season (15 August – 15 October), accompanied by guide-outfitter members of the Tahltan Guide and Outfitters Association (TGOA), per regulations outlined in the B.C. Hunting and Trapping Synopsis under the Wildlife Act (Government of British Columbia). Samples were contributed by guide-outfitters participating in a harvest-based wildlife health sampling program initiated in 2016 and use for this study was approved per the BC Wildlife Permit MRSM 18-285261 (Government of British Columbia Ministry of Forests, Lands, Natural Resource Operations, and Rural Development (FLNRORD)) and the Animal Use Protocol AC-18-0093 (University of Calgary Animal Care Committee). 

Fig 1. Study Area: the Tahltan Nation Traditional Territory with herd boundaries of 7 Northern Mountain woodland caribou herds included in this study.

The study population included adult male Northern Mountain Caribou hunted by guided hunters in the fall of 2016, 2017, and 2018. Sixty-three sample kits were collected from animals hunted between 25th August and 11th October from 2016 to 2018. Participating guide-outfitters collected a standard set of samples and data from harvested animals [18]. This included hair from the dorsal shoulder region of the harvested animal, a 10 cm² section of skin from the dorsal rump area, samples of liver tissue (~ 5 cm³ section), and whole left kidneys. In some cases, participants submitted partial left kidneys rather than the entire organ. Samples were stored in Whirl Pak™ sterile sample bags and frozen at -20°C until processing and analysis. Central incisors were submitted to Matson’s Laboratory, Manhattan, Montana for aging by cementum ring analysis [22].

Sample Analysis

Hair Analysis

Hair from the shoulder was preferentially used. In cases where insufficient sample was collected, hair from rump skin sections of the same individual was collected in the lab by shaving as close to the skin surface as possible. Visible debris such as soil and vegetation were removed from hair samples using plastic forceps. Samples were then washed twice in 96% ethanol and ultrapure Type 1 reagent-grade water to remove further external contamination then placed in clean paper envelopes and oven-dried at 50°C for at least 24 hours. 30-50 mg of dried hair was weighed and added to 2 mL of 70% HNO3 in a plastic vial (TMF Vessel, 100mL; Milestone, Shelton, CT, USA). The vials were closed with air-tight caps and digested using a high-pressure microwave reactor (ETHOS EZ Microwave Digestion System; Milestone, Shelton, CT, USA). The temperature in the reactor was gradually increased from room temperature to 220°C over one hour, and then gradually cooled to room temperature over one hour. 2 mL of each digested sample was transferred to a falcon tube and diluted with ultrapure Type 1 water to a total volume of 4 mL and stored at 5C until analysis. Each sample was further diluted with Type 1 water to a final dilution of 1:10 and hair mineral concentrations were determined using high-resolution inductively coupled plasma mass spectrometry (ICP-MS, 8800 Triple Quadrupole ICP-MS, Agilent) at the Alberta Center for Toxicology, University of Calgary. Instrument calibration verification for Quality Assurance (QA) were performed before, during, and after sample analyses using certified reference materials (Trace Elements in Natural Water (NIST1640a); Multi-Element Standard (SCP Science); and Environmental Calibration Standard (Agilent)). For each digestion (15 vials per run), 1 sample was a blank sample containing only acid to check for any contamination in laboratory procedure; 1 sample consisted of reference material for QA; and 13 vials contained samples. Of these samples, 1 sample was randomly selected to be run in duplicate for QA for each run. A maximum deviation limit of 20% between duplicates was set for the results in the run to be accepted, and all samples run in duplicate met these criteria when amount of mineral detected was greater than the method Limit of Quantitation (LOQ). For samples run in duplicate, the average of the two mineral concentration values was used for analysis. The LOQ (wet weight, digested sample) for Co, Pb, and Mo was 0.005 mg/L, for Mn and Se was 0.001 mg/L, for Cd, Cu, Zn was 0.005 mg/L, and for Fe was 0.05 mg/L. Mineral concentrations detected but falling below LOQ were included in the analysis. In cases where concentrations fell below detection limits, values of half the detection limit were assigned to assess correlations [23], and omitted for reporting baseline hair concentrations. Quality assurance was further confirmed in each batch, using certified reference materials (NIST2976 freeze-dried mussel tissue, National Institute of Standards and Technology; and NRC DORM-4 “Fish Protein Certified Reference Material for Trace Metals”, National Research Council Canada) as positive controls, and blank samples as negative controls. Blanks were negligible for all samples, and the laboratory positive controls were measured within acceptable ranges of certified reference values for all elements studied. Results are reported in mg/kg dry weight.

Kidney and Liver Tissue Analysis

The outermost portions of liver samples and partial kidney samples were removed to minimize external trace mineral contamination from handling. Sterile stainless-steel scalpel blades were used, with efforts made to minimize instrument use where possible. Extra tissues and the renal capsule were removed from kidneys. When only partial kidney samples were submitted, those samples with unequal cortex:medulla ratios were discarded. Samples (minimum 5 g tissue, wet weight) were submitted to a commercial laboratory (ALS Environmental, Vancouver, BC). Metals analysis was conducted as described in by Horvath [24], where tissues were homogenized and then subsampled prior to being hot block digested with nitric and hydrochloric acids combined with hydrogen peroxide. Analysis for tissue concentrations of various trace minerals was done by collision cell - inductively coupled plasma - mass spectrometry (CC-ICP-MS), modified from the standard US-EPA Method 6020A [25]. Moisture content of tissues (% Moisture) was determined gravimetrically by drying each sample at 105°C for a minimum of 6 hours. 

Samples were run in duplicate. A maximum deviation of 20% between duplicates was applied to qualify samples to be included in data analysis, and all samples met these criteria. Quality assurance was confirmed in each batch, using certified reference materials (NRC DORM-4 “Fish Protein Certified Reference Material for Trace Metals”, National Research Council Canada) and laboratory control samples as positive controls and method blank samples as negative controls. Blanks were negligible for all samples, and the concentrations measured were within acceptable ranges of certified reference values for all elements, meeting data quality objectives set by the commercial laboratory. In cases where concentrations fell below detection limits, values of half detection limit were assigned [23].  Results are reported in mg/kg dry weight (DW).

Statistical analysis

Liver and kidney trace mineral concentrations were compared to those published for other Canadian caribou herds/ecotypes using t-tests with Bonferroni multiple comparisons corrections. Shapiro-Wilk tests and visualization of residual plots were used to assess normality and homoscedasticity in the dataset. Where necessary and possible, data were logarithmically transformed to meet normality assumptions. Spearman rank correlations between hair, liver, and kidney mineral concentrations were assessed (Tables 2 and 3). The Cook’s distance test was used to identify influential multivariate outliers based on a standard cut-off (4 x mean), and univariate outliers were identified as values deviating from the mean by greater than 3 standard deviations. Outliers were maintained in the dataset and impacts of these outliers on relationships found between tissue mineral concentrations were assessed by removal for a sensitivity analysis. ANOVAs with Tukey’s post-hoc tests were used to evaluate differences in hair trace mineral concentrations based on body location of hair collection (rump versus shoulder). When bivariate normality was not possible, non-parametric tests were used instead (Kruskal-Wallis rank tests and Pairwise-Wilcoxon rank sum tests in place of 1-way ANOVA and Tukey post-hoc tests). In cases where significant differences were noted between the means of element concentrations in samples from different body locations, Spearman rank correlations were reassessed separately for different sample types as a further sensitivity analysis of the initial relationships found based on pooled data. Statistical tests were applied using the following packages: ‘tidyverse’ Version 1.2.1 (2019); ‘dplyr’ Version 0.8.3 (2019); and ‘ggplot2’ Version 3.2.1 (2019). All statistical analyses were done in R Version 3.6.0 [26].

Microsoft Excel