Deer capture and sampling
We captured 100 mule deer (54 females, 46 males) during November 2018–February 2019, avoiding capture and sampling of juveniles. We attempted to distribute captures throughout the ~23 km2 study area described by Miller et al.5 to minimize spatial disparities in comparing contemporary and past data, and to assure marks were widely distributed for December ground counts to estimate deer abundance5,35,36. Field and sampling methods generally followed those used elsewhere5,31,37. Field procedures were reviewed and approved by the CPW Animal Care and Use Committee (file 14–2018).
We pursued deer on foot and darted them opportunistically, delivering sedative combinations intramuscularly via projectile syringe. Premixed immobilization drug combinations included either nalbuphine (N; 0.9 mg/kg) or butorphanol (B; 0.5 mg/kg) combined with azaperone (A; 0.2 mg/kg) and medetomidine (M; 0.2 mg/kg)38, with standard total doses for respective combinations based on an estimated mass of 70 kg (average drug volume per animal was 1.3 ml NMA, 1.4 ml BAM). We collected rectal mucosa biopsies to determine CWD infection status37. We also collected whole blood and marked all deer with individually identifiable ear tags and some with telemetry (n = 51) or visual identification (n = 12) collars. Ages were estimated to the nearest year via tooth replacement and wear patterns39; observers used a pocket reference guide in the field to help assure consistency. To antagonize sedation upon completion of handling and sampling, each deer received 5 mg atipamezole/mg M administered, injected intramuscularly.
Prion diagnostics
Formalin-fixed tissue biopsies were processed and analyzed by immunohistochemistry (IHC) at the Colorado State University Veterinary Diagnostic Laboratory (Fort Collins, Colorado USA; CSUVDL) for evidence of CWD-associated prion (PrPCWD) accumulations using monoclonal antibody F99/97.6.1 (VMRD Inc., Pullman, Washington, USA)40 and standard IHC methods24,37,41, except that the CSUVDL’s IHC staining machine (Leica Microsystems Inc., Buffalo Grove, Illinois, USA) was different from that used in earlier studies (Ventana Medical Systems, Oro Valley, Arizona, USA). Biopsies were evaluated microscopically and classified as positive (infected) or not detected (negative) based on PrPCWD presence or absence; the same pathologist (T. R. Spraker) read biopsies for both the current and prior5 studies.
We included only data from deer with biopsies providing ≥3 lymphoid follicles in analyses involving infection status in order to maintain a relatively high (≥90%) probability of detecting infected individuals24. Two animals with low follicle counts that died shortly after capture were excepted by substituting postmortem IHC results. Limiting the acceptable follicle count excluded seven females (two 225SS, five 225SF) and two males (one 225SS, one 225SF) from some analyses. One male deer was 225FF and one female deer was missing a blood sample and thus not assigned to a PRNP gene group; these two individuals also were excluded from some analyses (e.g., Table 1).
PRNP genotyping
We used DNA extracted from whole blood buffy coat aliquots (n = 99) to screen for the presence of sequences at PRNP gene codon 225 that encode for serine (S) and/or phenylalanine (F) in the mature prion polypeptide, classifying individuals as 225SS, 225SF, or 225FF16,36,42. Methods generally followed those described by Jewell et al.16. Briefly, we extracted DNA using the DNeasy® blood and tissue kit (Qiagen, Valenica, California, USA). We amplified the complete open reading frame (ORF) plus 25 bp of 5′ flanking sequences and 53 bp of 3′ flanking sequences in the PRNP coding region using polymerase chain reaction (PCR). Purified DNA was combined in a 0.2 ml PCR tube containing a puReTag Ready-To-Go PCR bead (illustra™, GE Healthcare Bio-Sciences Corp, Piscataway, New Jersey, USA). Each PCR bead contained 2.5 units puReTag DNA polymerase, 10 mM Tris-HCI, 50 mM KCl, 1.5 mM MgCl2, 200 µM of each deoxynucleoside triphosphate, and stabilizers, including bovine serum albumin. For each PCR assay, 1 μL of each primer at 200 nM, 22 μL of RNase-free water and 1 μL of approximately 100 ng total genomic DNA was added for a final volume of 25 μL. Primers used for amplification were forward (MD582F, 5′-ACATGGGCATATGATGCTGACACC-3′) and reverse (MD1479RC, 5′-ACTACAGGGCTGCAGGTAGATACT-3′) described by Jewell et al.16. Reactions were thermal-cycled in a PTC 100 (MJ Research) at 94 C for 5 min and then 32 cycles of 94 C for 7 s, 62 C for 15 s, 72 C for 30 s and a final cycle of 72 C for 5 min, and kept at 4 C until inspected for successful amplification by agarose gel electrophoresis. As confirmed by LaCava et al.19, the MD582F and MD1479RC primers developed by Jewell et al.16 specifically amplify the functional PRNP gene ORF, thereby excluding confounding effects that could arise from the presence of a processed pseudogene that occurs in a majority of deer (Odocoileus spp.)42.
We used EcoRI restriction digestion of the PCR-amplified PRNP region16—a validated assay targeting the singular polymorphism at codon 225 in mule deer—to screen all 99 samples for presence of S or F codons. Aliquots (10 μl) of completed PCR reactions were incubated with 10 U EcoRI (New England Biolabs) in a total volume of 12 μl containing 50 mM NaCl, 100 mM Tris/HCl, 10 mM MgCl2, 0.025% Triton X-100 (pH 7.5) at 37 C for 2–16 h followed by the addition of 2.5 μl 6× concentrate gel loading solution (Sigma- Aldrich) per sample, and the inspection of products by agarose gel electrophoresis for the presence of one 897bp-sized band for 225SS, two bands—one 897 bp and one 719 bp—for 225SF, or one 719 bp-sized band for 225FF. As noted by Jewell et al.16, occurrence of TTC (the F codon) at position 225 creates an EcoRI recognition DNA sequence and cleavage site GAATTC from codons 224–225, whereas TCC (the S codon) creates the sequence GAATCC, which is not cut by EcoRI. When incubated with EcoRI, PCR products with a TTC codon at position 225 yielded cleavage fragments of the predictable sizes listed16. Because no other sites within the PRNP ORF DNA sequence are potentially transformable to GAATTC with one base change, this represents a specific genotyping method for assessing the S225F polymorphism in mule deer16.
To confirm findings from EcoRI screening, we examined sequences of the complete PRNP ORF from 20 samples that showed evidence of cleavage indicating 225*F and 6 samples without cleavage identified as 225SS. For DNA sequencing, we used primers 245 (5′-GGTGGTGACTGACTGTGTGTTGCTTGA-3′), 12 (5′-TGGTGGTGACTGTGTGTTGCTTGA-3′) and 3FL1 (5′-GATTAAGAAGATAATGAAAACAGGAAGG-3′; Integrated DNA Technologies). Sanger sequencing was done on purified PCR product by Eurofins Genomics (Louisville, Kentucky, USA). Sequence chromatograms were viewed and DNA sequence alignments and comparisons were made using the MAFFT multiple sequence alignment program v7.450 module, software platform v2020.2.3 of Geneious Prime. Sequencing confirmed the presence of coding for F in all samples identified as 225*F by EcoRI digestion, as well as the absence of such coding in samples identified by EcoRI digestion as 225SS. Moreover, presence of AGC at codon 138 in all sequenced samples reconfirmed that the primers we used had amplified the functional PRNP gene42.
Statistics and reproducibility
For analyses, we tabulated IHC-positive and -negative results to estimate apparent prevalence of prion infection. We also tabulated the number of individuals assigned to PRNP genotypes and to age groupings as described. Age groupings were selected based on relevance to CWD epidemiology in mule deer1,5,8,12,16,17,18,20,24,31,37. Assuming a ~2-year disease course5,8,17 and relative scarcity of end-stage disease in 225SS deer <2 years old1,5,8,20, binning deer into 1–2-, 3–4-, and ≥5-year-old groups helped establish roughly when exposure and disease-associated losses were occurring. Binning also afforded a way to isolate the 1–2-year-old age group least likely to be affected by prion infection in other comparisons to earlier data. We compared proportions related to specific hypotheses using two-tailed Fisher’s exact tests43. Descriptive statistics for survival data were derived from Kaplan–Meier life table analysis43, and we compared survival curves for infected deer killed by mountain lions or succumbing to clinical disease using Mantel’s logrank test (two-tailed, without continuity correction)43. For all comparisons, we used α = 0.05 to ascribe differences.
Our sample of 54 female and 46 male deer approximated an a priori goal to compare data from 50 individuals of each sex based on original hypotheses about expected changes in apparent prevalence. (We expected prevalence to have increased, especially among females.) Sample sizes were limited but representative because the Table Mesa study herd is relatively small. Based on mark-resight results, we sampled nearly half of the non-juvenile individuals estimated to be present in the herd during winter 2018–2019. Large proportions of both sexes and younger adult age classes were represented even though the resulting numbers were somewhat small for estimating proportions with precision, especially when data were parsed across PRNP genotypes or age groupings.
We used data from biopsy and genetic sampling and mark-resight inventories done at Table Mesa in 2005 in assessing changes in apparent prevalence, PRNP genotype distributions, and deer abundance over time. In our earlier study5, captures occurred during September–December and the infection status of deer was based on tonsil biopsy IHC. Because PrPCWD can be detected by IHC in tonsillar lymphoid follicles somewhat earlier in the disease course than in follicles associated with the rectal mucosa37, our contemporary estimates of prevalence may be slightly underestimated. Despite differences in biopsy site, median survival of infected 225SS deer after capture did not differ between the two study periods (for 2005–08: 279 [95% CL 177–381] days, range 1–661 days, n = 44; for 2018–19: 215 [95% CL 95–335] days, range 27–415 days, n = 9; Mantel’s logrank test, two-tailed, without continuity correction, P = 0.173).
Of the measurements taken, age estimation was perhaps the most vulnerable to potential bias. The vast majority of age estimates (92%) were made by one of four observers, including one (LLW) from the 2005 study. The use of pocket guides appeared helpful in assuring consistency among observers. Pairwise comparisons of the frequency distribution of age group assignments among the four main observers revealed no differences either overall (χ2 ≤ 6.632, Šidák-adjusted P ≤ 0.199) or relative to the most experienced observer (likelihood-ratio χ2 ≤ 6.185, P ≤ 0.136, with P values Bonferroni-adjusted for three comparisons).
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Source: Ecology - nature.com
