Edited by Stanley B. Prusiner, University of California, San Francisco, CA, and approved May 7, 2019 (received for review March 7, 2019)
Our gene-targeting strategy addresses several critical issues relating to chronic wasting disease (CWD), a contagious, lethal neurodegenerative prion disease affecting wild as well as captive cervids, which poses an uncertain risk to humans. First, we show that residue 226, the sole primary structural difference between deer and elk prion protein, dictates the selection and propagation of different CWD strains. Because the strain properties of prions affect their host-range potential, these findings suggest that CWD from elk and deer pose distinct risks to humans and other species. Second, we show that these gene-targeted mice offer an unprecedented means to address distinctive aspects of CWD peripheral pathogenesis and horizontal transmission that are not afforded by previously generated transgenic mouse models.
Although the unifying hallmark of prion diseases is CNS neurodegeneration caused by conformational corruption of host prion protein (PrP) to its infective counterpart, contagious transmission of chronic wasting disease (CWD) results from shedding of prions produced at high titers in the periphery of diseased cervids. While deer and elk PrP primary structures are equivalent except at residue 226, which is glutamate in elk and glutamine in deer, the effect of this difference on CWD pathogenesis is largely unknown. Using a gene-targeting approach where the mouse PrP coding sequence was replaced with elk or deer PrP, we show that the resulting GtE226 and GtQ226 mice had distinct kinetics of disease onset, prion conformations, and distributions of prions in the brains of diseased mice following intracerebral CWD challenge. These findings indicate that amino acid differences at PrP residue 226 dictate the selection and propagation of divergent strains in deer and elk with CWD. Because prion strain properties largely dictate host-range potential, our findings suggest that prion strains from elk and deer pose distinct risks to sympatric species or humans exposed to CWD. GtE226 and GtQ226 mice were also highly susceptible to CWD prions following intraperitoneal and oral exposures, a characteristic that stood in stark contrast to previously produced transgenic models. Remarkably, disease transmission was effective when infected mice were cohoused with naïve cagemates. Our findings indicate that gene-targeted mice provide unprecedented opportunities to accurately investigate CWD peripheral pathogenesis, CWD strains, and mechanisms of horizontal CWD transmission.
Prions are infectious proteins that are responsible for a group of related, fatal, and incurable mammalian neurodegenerative disorders, which continue to increase in variety in novel species. The emergence of bovine spongiform encephalopathy (BSE), and its zoonotic transmission as variant Creutzfeldt Jakob disease (vCJD) (1), highlight the substantial economic, political, and public health impacts of these disorders, and epitomize their unpredictable epidemiology and potential for interspecies transmission. Chronic wasting disease (CWD), a burgeoning epidemic of deer, elk, and other cervids is of particular concern (2). The unparalleled efficiency of its contagious transmission in wild as well as captive animals (3, 4) is related not only to extensive peripheralization and shedding of CWD prions during infection (5), but also to their seemingly unlimited environmental persistence (6). Both the geographic range of CWD and the number of naturally susceptible host species continue to increase. At the time of this writing, 26 states and three Canadian provinces report disease in farmed and free-ranging animals, with prevalence in some circumstances as high as 50% (2). Affected North American species include mule deer (Odocoileus hemionus), black-tailed deer (Odocoileus hemionus columbianus), white-tailed deer (Odocoileus virginianus), Rocky Mountain elk (Cervus elaphus nelsoni), North American elk (Cervus canadensis), and Shira’s moose (Alces alces shirasi) (2). In recent years CWD has acquired global epidemic status. Following importation of subclinically diseased North American elk, CWD was inadvertently introduced to South Korea, where it subsequently spread to native red deer (Cervus elaphus elaphus), sika deer (Cervus nippon), and offspring of crosses between red and sika deer (7⇓–9). In 2016 CWD was identified in free-ranging Norwegian reindeer (Rangifer tarandus tarandus), marking the first time the disease was detected in Europe, or in this species (10). Additional Norwegian cases were subsequently detected in European moose (Alces alces alces) (11) and red deer. CWD was detected in a moose from Finland in 2018 (http://www.dailyfinland.fi/national/4548/moose-dies-of-cwd-for-1st-time-in-finland) and in a moose from Sweden in 2019 (https://www.sva.se/en/animal-health/wildlife/map-of-chronic-wasting-disease-cwd).
An important aspect of prion diseases is their transmissibility. Exposure of individuals to prions from the same species results in fatal central nervous system (CNS) degeneration, typically with prolonged incubation periods. During this clinically silent phase, prion replication results from the conformational corruption of a host-encoded glycophoshphatidyl inositol-anchored sialoglycoprotein, referred to as the cellular prion protein (PrPC), by its abnormally conformed and biochemically distinct counterpart, PrPSc, which constitutes the infectious agent (12). While transmission of prion diseases from one species to another has been demonstrated repeatedly, it is generally a less efficient process. At the molecular level, the efficiency of interspecies prion transmission is governed by two predominant factors. First, transmission is facilitated by primary structural identity between PrPSc and the PrPC conversion template expressed in the newly infected host (13). Underscoring this concept, multiple studies demonstrate that transgenic (Tg) expression of PrP from various species, typically in the absence of endogenous mouse PrP (14), eliminates the resistance of mice to prions from those species. Second, while they lack nucleic acids, prions share with conventional infectious agents the ability to propagate heritable strain phenotypes, the properties of which also influence the ability of prions to breach species barriers (15). Strain properties are operationally defined by differences in times to disease onset after prion exposure (incubation period), clinical signs, and the relative intensity and distribution of PrPSc and neuropathological lesions in the CNS (16). At the molecular level, these distinguishable and heritable strain characteristics are enciphered by distinct PrPSc conformations (17, 18). To reconcile the interactive effects of conformationally dictated strain properties with PrP sequence variations, the conformational selection model postulated that particular PrPC primary structures are optimized to select and propagate distinct quasispecies from an ensemble of conformations produced during prion replication in an infected host (19). However, it is worth noting that this model appears to be incompatible with nonadaptive prion amplification, where certain interspecies transmissions do not entail emergence of new strains or conformational changes in PrPSc, but instead produce prions that replicate in the species of prion origination and not the new host (20).
Naturally occurring single amino acid PrP polymorphisms can have profound effects on prion disease susceptibility. For example, in humans, variation at PrP residue 129 dictates the propagation of distinct prion states, including those causing variant, sporadic, and iatrogenic forms of CJD (21). The PrP coding sequences of CWD-susceptible cervids are equivalent except at codon 226: deer express PrPC with glutamine (Q) at this position (Q226-cervid PrPC), while elk PrPC contains glutamate (E) (E226-cervid PrPC) (22), and the red deer PrP coding sequence is dimorphic (23). Whether differences at residue 226 influence the portfolio of selected strains in CWD-affected deer and elk and, by extension, whether the strain properties of prions causing CWD in deer are distinct from those causing CWD in elk is largely unknown.
We previously produced prototype CWD-susceptible Tg(CerPrP)1536+/− mice that overexpress Q226-cervid PrPC, and Tg(CerPrP-E226)5037+/− mice overexpressing E226-cervid PrPC (24, 25), referred to as Tg1536+/− and Tg5037+/−. While problematic approaches for studying CWD in the natural host or poorly susceptible experimental animals were largely superseded by the development of these, and other similar Tg mice [reviewed in ref. 16], undefined influences of random transgene insertions (position effects), and the use of different transgene expression vectors to create each line were among the drawbacks that prevented a full assessment and understanding of the influence of amino acid differences at residue 226 on CWD pathogenesis. Moreover, while studies in Tg1536+/− indicated the existence of two predominant, naturally occurring CWD strains (26), subsequent findings in other Tg mice drew attention to the potential for artifactual effects of transgene overexpression and resultant aberrant strain mutation that potentially complicate studies designed to assess strain prevalence (27). Finally, while Tg mouse models designed to abrogate species barriers to foreign prions have been primarily engineered with disease outcomes in the CNS in mind, it is unclear whether—or to what extent—they also accurately recapitulate aspects of peripheral prion pathogenesis. This consideration is particularly important in the context of CWD, given the crucial importance of peripheral prion replication and shedding in the contagious transmission of CWD in natural host species (5).
To address these shortcomings, we employed a gene-targeting (Gt) strategy in which the coding sequence of the gene encoding mouse PrP, referred to as Prnp, was precisely replaced by homologous recombination in embryonic stem (ES) cells with elk or deer PrP coding sequences, resulting in mice referred to as GtE226+/+ and GtQ226+/+. While this approach was previously employed to elegantly address the effects of mouse PrP polymorphisms on prion pathogenesis (28), other previously generated Gt mice designed to abrogate species barriers to bovine and human prions were unexpectedly found to be poorly susceptible to BSE and CJD prions compared with their Tg counterparts (14, 29⇓⇓⇓⇓⇓⇓⇓–37). In contrast, we show here that GtE226+/+ and GtQ226+/+ are fully susceptible to CWD prions from a range of diseased animals by a variety of inoculation routes. Because their genetic backgrounds are identical except at codon 226, and because they express equivalent, physiological levels of PrPC, GtE226+/+ and GtQ226+/+ mice afforded the additional prospect of precisely determining the effects of variability at PrP residue 226 on CWD pathogenesis and strain selection. Finally, we investigated the proposal (38) that gene-targeting strategies provide an accurate means to model peripheral pathogenesis in prion diseases, this being the underlying feature of CWD that contributes to its unprecedented contagious transmission.
Using ES cells from FVB mice, we developed two lines of Gt mice by direct replacement of the Prnp coding sequence with the corresponding elements from deer or elk PrP, which differ only at codon 226 (Fig. 1A). The resulting GtE226+/+ and GtQ226+/+ mice are homozygous for the targeted alleles that express elk (E226-cervid PrPC) and deer (Q226-cervid PrPC), respectively. Because deer and elk PrP expression in these circumstances is under the control of Prnp regulatory elements, GtE226+/+ and GtQ226+/+ mice produced equivalent, physiological levels of mRNA and E226-cervid PrPC and Q226-cervid PrPC (Fig. 1 B and C). Because they were engineered and maintained on an inbred FVB background, GtE226+/+ and GtQ226+/+ are syngeneic, except at codon 226 of the PrPC coding sequence.
Generation and characterization of Gt mice expressing cervid PrPC. (A) Structure of the Prnp locus, targeting vectors, the targeted alleles, and resultant targeted Prnp loci after homologous recombination and removal of puromycin resistance gene, referred to as knockin alleles. Light blue rectangles designated number 3 represent Prnp exon 3 containing the complete PrP coding sequence, shown in dark blue for mouse PrP, gray for cervid (Cer) PrP and, more specifically, magenta and orange for deer and elk PrP, respectively, and the 3′UTR. The location of amino- and carboxyl-terminal signal peptidase cleavage sites are indicated as black vertical lines in coding sequences. In the targeting vectors, the puromycin resistance (PuroR) selection cassette (red box), flanked by loxP sites (green arrowheads), was inserted 531 nucleotides upstream of the deer or elk PrP translation initiation codons within intron 2 of Prnp (gray line). Targeting vectors contained an additional 3.2 kbs (kbp) of intronic sequence upstream of the floxed PuroR cassette. A thymidine kinase cassette was engineered 6.5-kbp distal to the coding sequences. The 5′ and 3′ homology arms of the targeting vectors were composed of FVB genomic sequences. (B) Levels of mRNA expression in brains of age-matched mice FVB (blue), GtE226+/+ (orange), and GtQ226+/+ (magenta) mice (n = 3 in each case) were determined by real-time quantitative reverse-transcription PCR using β-actin as an internal control. (C) Western blot analysis of PrPC expressed in the brains of GtE226+/+ and GtQ226+/+ mice compared with FVB/N mice using mAb PRC5 (57). Prnp0/0 knockout mice maintained on an inbred FVB background (41) served as a negative control. The position of protein molecular weight standards at 36, 29, and 19 kDa are shown.
To assess whether or not physiological expression of elk and deer PrPC in mice conferred susceptibility to CWD prions, we challenged GtE226+/+ and GtQ226+/+ mice intracerebrally (ic) with prions isolated from the CNS of CWD-affected North American deer and elk. All intracerebrally inoculated Gt mice, referred to as GtE226+/+(ic) and GtQ226+/+(ic), developed disease (Table 1), and their brains accumulated cervid PrPSc (Figs. 2A and 3). The kinetics of disease onset was more rapid in GtE226+/+(ic) than in GtQ226+/+(ic) mice, resulting in isolate mean incubation times that were 20–38% shorter in GtE226+/+(ic) mice (Table 1). When we assessed the collective responses of all intracerebrally inoculated Gt mice, the 262 ± 4 (±SEM) d mean time to disease onset in GtE226+/+(ic) mice (n = 87) was 26% shorter (P ≤ 0.0001) than the 355 ± 7 d mean time to disease in GtQ226+/+(ic) mice (n = 69) (Fig. 2C). We also intracerebrally inoculated Gt mice with elk CWD prions that were previously passaged in Tg5037+/− mice (25). Like the ancestral 04-0306 CWD prions from which they originated, Tg5037+/−-derived prions produced disease with a time to onset that was 28% faster (P ≤ 0.001) in GtE226+/+(ic) than in GtQ226+/+(ic) mice (Table 1). We conclude that Gt mice expressing deer or elk PrP are highly susceptible to intracerebral challenge with CWD prions. Because their genetic backgrounds are otherwise identical, and because PrPC expression is equivalent in both lines, we ascribe differences in the kinetics of disease onset in GtQ226+/+ and GtE226+/+ mice to the effects of amino acid variation at PrPC residue 226.
Disease susceptibility of Gt mice expressing elk or deer PrP following intracerebral inoculation with cervid prions
The effects of cervid PrP residue 226 and cervid PrPC expression on the kinetics of disease onset and CNS PrPSc accumulation following intracerebral challenge with CWD. (A) Western immunoblots of PrP in the CNS of diseased Gt mice intracerebrally inoculated with CWD prions. (Upper) Gt+/+ mice; (Lower) Gt+/− mice. Samples were either untreated (25 μg) or treated (100 μg) with PK as indicated, and immunoblots were probed with mAb PRC5. The positions of protein molecular weight standards at 36, 29, and 19 kDa are shown. (B) Time to disease onset in intracerebrally challenged Gt and Tg mice expressing either E226-PrPC (orange) or Q226-PrPC (magenta) is proportional to levels of PrPC expressed in the CNS. Tg5037+/− and Tg1536+/− mice express transgene-encoded PrPC in the hemizygous state; Tg5037+/+ are homozygous for the transgene array. (C) Differences at residue 226 of deer and elk PrPC affect the tempo of disease onset in intracerebrally inoculated Gt mice. Time to disease onset of CWD inoculated GtE226+/+ (orange), and GtQ226+/+ (magenta) following intracerebral challenge. Curves represent the aggregate responses of all CWD inocula used to challenge Gt mice in Table 1. In B and C, aggregate incubation times of multiple CWD isolates are expressed as mean ± SEM, as well as the number (n) of diseased mice, with significance determined using Student’s t test. ****P ≤ 0.0001. (D) Resistance of PrPSc in the brains of diseased, intracerebrally inoculated Gt mice to PK following treatment with varying concentrations of GdnHCl. Immunoblots were probed with mAb PRC5. Brain extracts from GtE226+/+ (orange), GtQ226+/+ (magenta), and GtE/Q226 (green). Each point is the mean value derived from the analysis of eight diseased mouse brains, and error bars correspond to the SD of those mean values. Fapp, fraction of apparent PrPSc (maximum signal-individual signal)/(maximum signal-minimum signal). Sigmoidal dose–response curves were plotted using a four-parameter algorithm and nonlinear least-square fit. Statistical differences between the concentrations of GdnHCl producing half-maximal PrPSc denaturation were calculated. ****P ≤ 0.0001.
PrPSc distribution in brains of diseased Gt mice intracerebrally challenged with CWD prions. For consistency, coronal sections through the hippocampus and thalamus are shown in all cases. (A–H) PK-treated histoblots probed with mAb PRC5. (I–T) Immunohistochemical analysis with Fab D18. (Magnification: 2×.) (A–L) Primary transmissions of CWD isolates from deer or elk to either GtE226+/+ (orange), GtQ226+/+ (magenta), or heterozygous GtE/Q226 (green). (Q–T) Primary transmissions of CWD isolates from deer or elk to either hemizygous GtE226+/− or GtQ226+/−. (M–P) Secondary transmissions (P2) of CWD prions, originally passaged in GtE226+/+ (inoculum designated in orange) or GtQ226+/+ (inoculum designated in magenta).
We questioned whether the rapid and uniform induction of disease in GtE226+/+ and GtQ226+/+ mice reflected idiosyncratically virulent properties of CWD prions. Arguing against this notion, when we challenged GtE226+/+ and GtQ226+/+ mice with cervid prions from the CNS of diseased Tg1536+/− mice infected with either sheep SSBP/1 prions or mouse-adapted Rocky Mountain Laboratory (RML) scrapie prions (39, 40), all inoculated GtE226+/+ mice developed disease with ∼250–260 mean incubation times that were comparable to those produced by CWD isolates (Table 1). In contrast to CWD prion transmissions, the mean incubation time of cervid-adapted SSBP/1 prions was more rapid in GtQ226+/+ than in GtE226+/+ mice. We conclude that GtQ226+/+ and GtE226+/+ mice are susceptible to different cervid prion strains, and that amino acid variation at residue 226 has distinct effects on the propagation of CWD and cervid-adapted scrapie prions.
To assess the consequences of codon 226 heterozygosity found naturally in red deer, we bred GtE226+/+ with GtQ226+/+ mice, and intracerebrally challenged the resulting heterozygous F1 offspring, referred to as GtE/Q226(ic), with CWD prions. The mean incubation times of deer CWD isolates D10 and Db99 in GtE/Q226(ic) were equivalent to their relatively short incubation times in GtE226+/+ (P = 0.2–0.7), but distinct from their longer mean incubation times in GtQ226+/+ mice (P < 0.05). In contrast, the mean incubation times of elk CWD isolates 99W12389 and 001-44720 in GtE/Q226(ic) were intermediate between those registered in GtE226+/−(ic) and GtQ226+/−(ic) mice, producing incubation times that were distinct from those registered in GtE226+/+ mice (P < 0.0001 to <0.01) (Table 1).
The use of the inbred FVB background to produce Gt mice expressing physiological levels of deer or elk PrP, as well as previously characterized Tg1536+/− and Tg5037+/− mice overexpressing deer or elk PrP, respectively (24, 25), allowed us to assess the effects of primary structural differences at residue 226 on the kinetics of CWD onset over a range of cervid PrPC expression in a defined genetic context. We compared the responses of Tg and Gt mice challenged with matched CWD inocula, with the exception of isolate D92 which was only transmitted to Tg mice (Table 1 and SI Appendix, Table S1). The collective 264 ± 7 d mean incubation time resulting from intracerebral challenge of Tg1536+/− mice [Tg1536+/−(ic)] (n = 60), which overexpress Q226-cervid PrPC in the CNS at levels ∼fivefold higher in wild-type mice (24), was 24% shorter (P ≤ 0.0001) than the 347 ± 8 d mean time to disease in GtQ226+/+(ic) mice (n = 46). Similarly, the collective 204 ± 10 d mean incubation time in Tg5037+/−(ic) mice (n = 53), which overexpress E226-cervid PrPC in the CNS at levels ∼fivefold higher in wild-type mice (25), was 20% shorter (P ≤ 0.0001) than the 256 ± 5 d mean time to disease in GtE226+/+(ic) mice (n = 52). Consistent with our findings that disease occurred more rapidly in GtE226+/+(ic) than GtQ226+/+(ic) mice, the collective 204 ± 10 mean time to disease onset in Tg5037+/−(ic) mice (n = 53) was 23% shorter (P ≤ 0.0001) than the collective 264 ± 7 mean incubation time in Tg1536+/−(ic) mice (n = 60) (Fig. 2B and SI Appendix, Table S1). Our use of a consistent inbred background and equivalent prion inocula allow us to conclude that the ∼20–25% more rapid mean incubation times in intracerebrally inoculated Tg mice compared with intracerebrally inoculated Gt mice, matched at residue 226, results from ∼fivefold higher transgene expression in Tg mice.
The availability of Prnp0/0 knockout mice maintained on an inbred FVB background (41) allowed us to generate haploinsufficient counterparts of GtE226+/+ and GtQ226+/+ mice, referred to as GtE226+/− and GtQ226+/−, and an additional means to assess the effects of cervid PrPC expression on CWD incubation times in a controlled genetic context. All intracerebrally challenged GtE226+/− and GtQ226+/− mice, referred to as GtE226+/−(ic) and GtQ226+/−(ic), developed disease (Fig. 2B and Table 1), and accumulated PrPSc in the CNS (Fig. 2A). The collective 454 ± 7 d mean incubation time of GtQ226+/−(ic) mice (n = 26) was 32% longer (P ≤ 0.0001) than the 344 ± 11 mean time to disease in GtQ226+/+(ic) mice (n = 30) receiving the same isolates, while the collective 379 ± 8 d mean incubation time in GtE226+/−(ic) mice (n = 23) was 47% longer (P ≤ 0.0001) than the 258 ± 6 mean time to disease in similarly challenged in GtE226+/+(ic) mice (n = 42) (Fig. 2B). In support of the effects of residue 226 on the kinetics of disease onset, the mean time to disease onset was 17% shorter in GtE226+/−(ic) than GtQ226+/−(ic) mice (P ≤ 0.0001) (Fig. 2B). Collectively, our findings in Tg and Gt mice demonstrate that CWD incubation times following intracerebral inoculation are inversely related to levels of CNS cervid PrPC expression in the host. Accordingly, incubation times of CWD prions in Tg5037+/+(ic) mice, which are homozygous for the transgene array, were 23% more rapid than in counterpart hemizygous Tg5037+/− mice (Fig. 2B). Moreover, the reproducibly more rapid time to disease in mice expressing E226-cervid PrPC compared with their counterparts expressing Q226-cervid PrPC across a range of transgene expression underscores the consistent impact of residue 226 on the kinetics of CWD.
In light of the effects of residue 226 on the kinetics of disease onset, we sought to define differences in the characteristics of E226-cervid PrPSc and Q226-cervid PrPSc constituting E226 and Q226 prions produced in the CNS of intracerebrally challenged Gt mice. Although the electrophoretic migration patterns, and proportions of di-, mono-, and a-glycosyl forms of E226-cervid PrPSc and Q226-cervid PrPSc from GtE226+/+(ic) and GtQ226+/+(ic) mice were indistinguishable (Fig. 2A), they exhibited distinctive denaturation profiles following treatment with a range of guanidinium hydrochloride (GdnHCl) concentrations before digestion with proteinase K (PK) (42) (Fig. 2D). The 2.10 M concentration of GdnHCl producing half-maximal denaturation (GdnHCl1/2) of E226-cervid PrPSc from the CNS of diseased GtE226+/+ mice was distinct from the 2.52 M GdnHCl1/2 value of Q226-cervid PrPSc (P ≤ 0.0001) produced in the CNS of GtQ226+/+ mice. In addition, the unfolding characteristics of PrPSc produced in the brains of heterozygous GtE/Q226(ic) mice, and the resultant 2.30 M GdnHCl1/2 were distinct from, and intermediate between those of E226-cervid PrPSc and Q226-cervid PrPSc. These results indicate that residue 226 influences the conformational properties of CWD prions; and that the conformation of CWD-derived Q226-cervid PrPSc is more stable to denaturation than that of CWD-derived E226-cervid PrPSc.
We used histoblotting (43) and immunohistochemical (IHC) analysis to assess cervid PrPSc distribution in the brains of diseased intracerebrally inoculated Gt mice (Fig. 3). Infection of GtQ226+/+(ic) mice with either E226 (Fig. 3A) or Q226 (Fig. 3B) CWD prions resulted in a relatively disorganized distribution of Q226-cervid PrPSc characterized by intensely staining asymmetric deposits throughout the cortex, hippocampus, thalamus, and hypothalamus (Fig. 3I) that were consistently ipsilateral to the inoculation site. In contrast, the CNS of CWD-infected GtE226+/+(ic) mice exhibited two distinct patterns of E226-cervid PrPSc deposition, depending on whether disease resulted from infection with either E226 or Q226 CWD prions. In E226 CWD prion-infected GtE226+/+(ic) mice, E226-cervid PrPSc distribution was relatively diffuse, and bilaterally symmetrical in cortical, thalamic, and hypothalamic regions (Fig. 3 C, D, and J), whereas in GtE226+/+(ic) mice receiving Q226 CWD prions, E226-cervid PrPSc was asymmetrically deposited in the thalamus, again ipsilateral to the inoculation site, with relative sparing of cortical and hippocampal regions (Fig. 3 E, F, and K). Patterns of PrPSc deposition in the CNS of CWD infected heterozygous GtE/Q226(ic) mice (Fig. 3 G, H, and L) more closely resembled the asymmetrical, disorganized patterns of PrPSc in the CNS of diseased GtQ226+/+(ic) mice than the symmetrically diffuse PrPSc pattern in GtE226+/+(ic) mice. In GtE226+/−(ic) mice, E226-cervid PrPSc deposition patterns (Fig. 3 Q and R) were comparable to, but less intense than E226-cervid PrPSc found in the cortex, hippocampus, thalamus, and hypothalamus of homozygous GtE226+/+(ic) mice (Fig. 3J), while in CWD-infected haploinsufficient Gt226Q+/−(ic) mice (Fig. 3 S and T), Q226-cervid PrPSc deposition patterns resembled the disorganized amalgamations of aggregates found in the cortex, hippocampus, thalamus, and hypothalamus of their diseased homozygous GtQ226+/+(ic) counterparts (Fig. 3I). We conclude that residue 226 influences the targeted accumulation of cervid PrPSc in the CNS following intracerebral infection of Gt mice.
We performed serial transmissions of CWD prions produced in the brains of diseased Gt mice to additional Gt mice by intracerebral inoculation (Table 1). The relatively disorganized pattern of Q226-cervid PrPSc aggregates in the cortex, hippocampus, thalamus, and hypothalamus observed upon primary transmission of Q226 CWD prions to GtQ226+/+(ic) mice (Fig. 3 A, B, and I) was maintained when the resulting Q226 prions were passaged to additional GtQ226+/+(ic) mice (Fig. 3M). Similarly, the widespread, diffuse granular deposition of E226-cervid PrPSc in the cortex, hippocampus, and thalamus observed upon primary transmission of E226 CWD prions to GtE226+/+(ic) mice (Fig. 3 C, D, and J) was preserved upon serial passage of the resulting E226 prions to additional GtE226+/+(ic) mice (Fig. 3 N and O). However, the restricted, bilaterally asymmetrical thalamic pattern of E226-cervid PrPSc produced upon primary transmission of Q226 CWD prions to GtE226+/+(ic) mice (Fig. 3 E, F, and K) was not sustained when the resulting E226 prions were serially passaged to GtE226+/+(ic) mice (Fig. 3P). Instead, the new E226-cervid PrPSc pattern corresponded to the widespread, diffuse granular pattern of E226-cervid PrPSc produced upon primary transmissions of E226 CWD prions to GtE226+/+(ic) mice (Fig. 3 C, D, and J). These findings provide further support for the notion that cervid PrPSc accumulation in Gt mice is dictated by the effects of residue 226 in host-expressed cervid PrPC.
We investigated the responses of Gt mice to CWD prions introduced by the intraperitoneal (ip) and oral (po) routes, referred to as Gt(ip) and Gt(po). Gt(ip) and (po) mice developed prion disease (Fig. 4 A and B and Table 2) and accumulated PrPSc in the CNS (Figs. 4E and 5). The collective 315 ± 3 d mean incubation time of all inoculated GtE226+/+(ip) mice (n = 58) was 14% shorter (P ≤ 0.0001) than the 367 ± 13 d mean incubation time of all GtQ226+/+(ip) mice (n = 22) (Fig. 4 A and B), indicating that the residue 226 dimorphism affected the kinetics of CWD onset following intraperitoneal as well as intracerebral challenge.
Response of Gt and Tg mice expressing cervid PrPC to peripheral challenges with CWD prions. (A and B) Aggregate responses of intraperitoneally challenged (squares) GtE226+/+ (orange), GtQ226+/+ (magenta) following CWD inoculation with isolates from Table 2, compared with intracerebrally inoculated mice (circles) challenged with the same isolates (Table 1). (C) Aggregate responses of intraperitoneally challenged Tg1536+/− mice (magenta squares) compared with intracerebrally inoculated Tg1536+/− mice (circles) challenged with the same isolates. Incubation times are expressed as mean ± SEM, as well as the number (n) of diseased mice with significant differences determined using Student’s t test. ****P ≤ 0.0001; ns, not significant (P > 0.05). (D) Time to disease onset in GtE226+/+ mice intracerebrally inoculated with elk isolate 99W12389 (orange circles) and cohoused (ch) uninoculated GtE226+/+ mice (orange crossed circles). (E) Immunoblots of PrP in the CNS of diseased Gt mice intraperitoneally (ip) or orally (po) inoculated with CWD prions. (F) Immunoblots showing PrPSc accumulation in the spleens of diseased GtE226+/+ (orange, Upper), and GtQ226+/+ mice (magenta, Lower) following intracerebral and intraperitoneal challenge, compared with age matched mock-infected mice. (G) Western immunoblots showing PrPSc accumulation in the spleens of diseased GtQ226+/+ mice but not Tg1536+/− mice that were intraperitoneally inoculated (Top) or intracerebrally inoculated (Middle) with CWD prions, compared with mock infected, age matched GtQ226+/+ mice. PrPSc was detected in brain extracts of diseased Tg1536+/− mice that failed to accumulate PrPSc in spleens (Bottom). In F and G, Western blots of 100-μg spleen tissues not treated with PK (−), and 1 mg treated with PK (+) were probed with mAb PRC5. The positions of protein molecular mass standards of 36, 29, 19 kDa, from top to bottom, are shown.
Susceptibility of Gt mice expressing deer or elk PrP to CWD prions following i.p. or oral inoculation
PrPSc distribution in brains of diseased Gt mice following peripheral challenges with CWD prions. Coronal sections through the hippocampus and thalamus are shown in all cases. (A–H) PK-treated histoblots probed with mAb PRC5. (I–L) Immunohistochemical analysis with Fab D18. (Magnification: 2×.) GtE226+/+ (orange) or GtQ226+/+ mice (magenta) were inoculated either intraperitoneally (ip) or by oral gavage (po) as indicated with either elk (orange) or deer (magenta) CWD isolates. Ip-ic, secondary transmission by the intracerebral route of brain homogenates from ip challenged mice. Ch, uninoculated GtE226+/+ mice cohoused with ic-challenged GtE226+/+ mice.
When we compared the kinetics of disease onset following inoculation by the intracerebral and intraperitoneal routes, the collective 315 ± 3 d mean incubation time of GtE226+/+(ip) mice was 23% longer than the 257 ± 4 d (n = 66) time to disease in GtE226+/+(ic) mice receiving the same inocula (P ≤ 0.0001) (Fig. 4A). This outcome is in accordance with previous studies of scrapie prion transmissions in rodent models that have consistently demonstrated a protracted time to disease onset following intraperitoneal compared with intracerebral challenge (44). We were therefore surprised to detect no difference (P > 0.05) in the collective 367 ± 13 mean time to CWD onset in GtQ226+/+(ip) mice (n = 22) compared with the collective 344 ± 11 mean time to disease of GtQ226+/+(ic) mice (n = 30) inoculated with the same isolates (Fig. 4B).
To explore this finding further, we compared the responses of Tg1536+/− mice to CWD challenges by the intracerebral and intraperitoneal routes. While the collective mean incubation time of Tg1536+/−(ic) mice challenged with three CWD isolates was 251 ± 11 d (n = 21), Tg1536+/−(ip) mice receiving the same isolates failed to register disease until after 490 ± 14 d (n = 20), which was a 95% prolongation of disease onset compared with Tg1536+/−(ic) mice (P ≤ 0.0001) (Fig. 4C), and a 34% extension compared with GtQ226+/+(ip) mice (Fig. 4 B and C). While the spleens of diseased GtE226+/+ and GtQ226+/+ mice accumulated PrPSc following intraperitoneal or intracerebral inoculation (Fig. 4F), only the CNS but not the spleens of CWD inoculated Tg1536+/− mice accumulated PrPSc after either intracerebral or intraperitoneaal challenges (Fig. 4G). We conclude that the more efficient disease induction in GtQ226+/+(ip) than Tg1536+/−(ip) mice is the consequence of more accurate control of peripheral cervid PrPC expression by Prnp regulatory elements in Gt mice.
Consistent with previous studies, which established that the oral route of prion infection is less efficient than either the intraperitoneal or intracerebral routes (45), incubation times of CWD isolates in Gt(po) mice were generally longer than the corresponding Gt(ic) and Gt(ip) transmissions, and attack rates were not always uniformly complete (Table 2). However, in those mice that did develop disease, the 366 ± 26 aggregate time to disease onset in GtE226+/+(po) mice (n = 17) was 23% faster (P ≤ 0.001) than the 474 ± 19 mean time to onset in GtQ226+/+(po) mice (n = 18).
Histoblotting and IHC analyses demonstrated that cervid PrPSc in the brains of diseased Gt(ip) and Gt(po) mice accumulated as bilaterally symmetrical deposits in cortical and thalamic regions, with relative sparing of the hippocampus (Fig. 5). The general pattern of PrPSc deposition in the CNS of CWD infected Gt(ip) and Gt(po) mice was different from the patterns of PrPSc accumulation in the brains of diseased Gt(ic) mice (Fig. 3). Although patterns of PrPSc accumulation in GtQ226+/+(ic) and GtE226+/+(ic) mice receiving E226 CWD prions were recognizably different from the patterns resulting from Q226 CWD prion infection (Fig. 3), these distinctions were not obvious upon evaluation of GtE226+/+(ip) and GtQ226+/+(ip) mice, nor when comparing GtE226+/+(po) with GtQ226+/+(po) mice (Fig. 5).
When we intracerebrally challenged GtE226+/+ mice with prions from the CNS of diseased GtE226+/+(ip) mice inoculated with the 99W12389 elk CWD isolate, mean incubation times in the resulting diseased mice, referred to as GtE226+/+(ip-ic), decreased from ∼300 d in GtE226+/+(ip) to ∼200 d (Table 2), which was similar to the mean incubation time upon primary transmission of this isolate to GtE226+/+(ic) mice (Table 1). Additionally, whereas the pattern of E226-cervid PrPSc deposition in the CNS of diseased GtE226+/+(ip) (Fig. 5A) was distinct from that observed in GtE226+/+(ic) (Fig. 3C), serial transmission of prions from the CNS of GtE226+/+(ip) mice produced a pattern of E226-cervid PrPSc accumulation in the CNS of GtE226+/+(ip-ic) mice that resembled that of GtE226+/+(ic) mice (Fig. 5G).
In light of their efficient responses to peripherally administered CWD prions, we asked whether Gt mice could be used to model contagious transmission, which is the hallmark of CWD in natural hosts (46). Five days after intracerebral inoculation with CWD, GtE226+/+(ic) mice were introduced to new cages containing additional uninfected GtE226+/+ mice. Disease induction in uninoculated, cohoused GtE226+/+ mice, referred to as GtE226+/+(ch), was remarkably effective. The mean time to disease of GtE226+/+(ic) mice was 238 ± 10 d (n = 10), while that of their uninoculated, but cohoused GtE226+/+(ch) counterparts was 342 ± 28 d (n = 7) (P ≤ 0.001) (Fig. 4D). Histoblotting of whole coronal sections of diseased GtE226+/+(ch) mice (Fig. 5H) revealed that E226-cervid PrPSc was diffusely, and bilaterally symmetrically distributed in cortical, thalamic, and hypothalamic regions, in a pattern that was reminiscent of E226-cervid PrPSc in the brains of GtE226+/+(ic) mice receiving E226 CWD prions (Fig. 3 C and D).
Notwithstanding the considerable advantages of Tg mouse models for studying prions, including the capacity to decrease incubation times by transgene overexpression (13), the inability to control random integration of transgene arrays of variable copy number at undefined genomic locations complicates the interpretation of certain analyses, in particular comparative studies designed to assess the effects of PrP primary structural differences on prion pathogenesis. Here we describe the production and initial characterization of two Gt mouse models expressing the PrP coding sequences of elk and deer, which differ only at residue 226. We draw three overarching conclusions from our studies: first, replacement of the coding sequence for mouse PrP with that of deer or elk PrP renders the resulting Gt mice susceptible to CWD prions; second, amino acid variation at residue 226 of deer and elk PrP controls the tempo of disease onset, as well as the targeted accumulation and conformational properties of the resulting prions, reflecting the section and propagation of distinct cervid prion strains; and third, Gt mice offer an unprecedented means to address distinctive aspects of peripheral CWD pathogenesis and contagious transmission that were not afforded by previous Tg mouse models.
Our findings detailing uniform susceptibility, and relatively rapid responses of GtE226+/+ and GtQ226+/+ mice to elk and deer CWD prions, stand in contrast to the behaviors of previously produced gene targeted mice designed to abrogate species barriers to bovine and human prions. While expression of bovine PrPC conferred susceptibility to BSE prions in Tg mice, even when the level of bovine PrPC expression was lower than that normally found in the brains of cattle (29⇓–31), counterpart Gt mice were paradoxically less susceptible to BSE than wild-type mice (32, 33). Similarly, despite repeated examples of high susceptibility and rapid incubation times of human prions in Tg mice, even at wild-type levels of human PrPC expression (14, 34⇓–36), counterpart Gt mice exhibited longer vCJD prion incubation times compared with wild-type mice (32), as well as prolonged incubation periods and incomplete attack rates in response to sporadic CJD prions (37). The discrepant susceptibilities of these Tg mice and their Gt equivalents were puzzling because they appeared to undermine the primacy of PrP sequence homology in the control of prion transmission barriers (38). In contrast, the differences we observed in incubation times between CWD-susceptible Tg and Gt mice were surprisingly modest. When the effects of residue 226 on the kinetics of disease onset are considered, our findings in Tg and Gt mice expressing cervid PrP confirm the longstanding inverse relationship between levels of CNS PrPC expression and prion incubation times following intracerebral challenge (13).
Our findings and conclusions make it difficult to reconcile the poor responses of previously produced Gt mice to bovine and human prions on the basis of transgene expression alone, and force consideration of several alternate possibilities. First, the likelihood that the relatively facile responses of GtE226 and GtQ226 mice to prion infection are reflective of unusually virulent properties of CWD prions seems remote given that Gt mice were also equally susceptible to cervid-adapted sheep SSBP/1 and cervid-adapted mouse RML prions (Table 1). Second, previously produced Gt mice were generated on a 129 mouse background (32, 37), while the Gt mice reported here were generated using FVB. Whether the discrepant properties of Gt mice result from the effects of mouse genetic backgrounds can be resolved by studying the outcomes of inbreeding Gt mice onto a 129 background, and correspondingly by repeated backcrossing of bovine and human PrP Gt mice with FVB mice. Third, while GtE226 and GtQ226 mice express elk and deer PrP coding sequences, respectively, previously produced Gt mice were engineered to express chimeric mouse–bovine and mouse–human coding sequences, where the signal peptides of the expressed transgenes originated from mouse PrP (32, 37, 47). While such arrangements are not expected to affect the primary structures of mature bovine and human PrP after signal peptidase cleavage and processing, it is nonetheless well established that amino acid substitutions in signal peptide sequences have a considerable impact on PrP-mediated neurodegeneration by affecting the production of a trapped transmembrane isoform of PrP termed CtmPrP (48). Additional studies highlight the unpredictable effects of expressing chimeric PrP constructs on disease outcomes. While Gt mice, referred to as Ki-ChM, expressing a chimeric mouse–human PrP coding sequence containing the mouse PrP signal peptide—as well as mouse PrP sequences at the C terminus—had variable responses to human prions, certain CJD isolates produced disease with mean incubation times in the range of ∼140–180 d (49, 50). When transmitted to Gt mice expressing a different chimeric mouse–human PrP where only the signal peptide is derived from mouse PrP, referred to as Ki-Hu129M/M, mean incubation times of CJD prions that induced disease in Ki-ChM within ∼150 d were prolonged by >300 d (47). With these considerations in mind, and in light of the results reported here, it appears that expression of wild-type human or bovine PrP coding sequences affords the best prospect of conferring susceptibility to CJD and BSE prions in Gt mice.
The use of a Gt approach allowed us to unequivocally ascertain the effects of amino acid differences at codon 226 of deer and elk PrP on CWD pathogenesis. Because GtE226+/+ and GtQ226+/+ mice (and for that matter GtE226+/− and GtQ226+/− mice) produce equivalent amounts of elk and deer PrPC, and because they are otherwise syngeneic, the consistently different kinetics of CWD onset can be unambiguously ascribed to the effects of amino acid variation at residue 226. Our findings also demonstrate that Gt mice are susceptible not only to CWD prions, but also to experimentally adapted cervid prions, including cervid-adapted SSBP/1 and cervid-adapted RML, with variable kinetic effects of the residue 226 dimorphism. The production of Gt mice on an FVB background, and the availability of PrP knockout and previously produced Tg mice on the equivalent inbred mouse background, provided the added advantage of allowing us to compare the effects of residue 226 over a range of transgene expression in a controlled genetic context. The consistently more rapid time to clinical disease in mice expressing elk PrPC compared with their Gt and Tg counterparts expressing deer PrPC across a spectrum of transgene expression emphasizes the impact of residue 226 on CWD. Indeed, the comparative analyses of Tg and Gt mice reported here confirms that the variable responses of Tg5037+/− and Tg1536+/− mice to CWD prions originally reported in limited transmission studies (25), resulted from effects of amino acid differences at residue 226 rather than uncontrolled artifacts of transgene expression.
Our findings are in accordance with the notion that residue 226 in elk and deer PrPC acts to select and propagate distinct strains from an ensemble of PrPSc conformers by a mechanism that is compatible with the conformational selection model (19). In support of this view, GtE226+/+ and GtQ226+/+ mice have distinct times to disease onset and neuropathological outcomes in response to intracerebral challenge with the same CWD prion inocula, and the resulting E226 and Q226 prions produced in Gt mice have distinct conformational properties. These findings raise the possibility that the strain properties of CWD prions produced in diseased elk are distinct from those of prions causing CWD in deer. Limited analyses showing distinct pathogenic outcomes between CWD-affected deer and elk (51, 52) are consistent with this interpretation. In light of the influence of strain properties on interspecies prion transmission, our findings that elk and deer PrPC select and propagate different strains are relevant when considering the risk of interspecies CWD transmission, in particular when considering the barrier controlling human susceptibility.
Of note, the effects of amino acid differences at residue 226 on the biological and biochemical properties of CWD prions reported here are supported by structural investigations showing that the distal region of α-helix 3, which contains residue 226, participates with residues in the β2–α2 loop to form a solvent-accessible discontinuous epitope (53, 54). Our subsequent molecular dynamics analyses (55) indicated that E and Q variations at residue 226, as well as the protective polymorphism at the adjacent residue 225 in deer PrP, affected hydrogen bonding between key residues in the β2–α2 loop that modify the strength of interaction between these subdomains, supporting the notion that the plasticity of this discontinuous epitope modulates the efficiency of PrPC to PrPSc conversion.
Replication of CWD prions to high titers in the peripheral lymphoid system and their resultant shedding in excrement and bodily fluids of diseased cervids are considered to be the underlying features contributing to the unparalleled transmission efficiency of CWD (5). Because they are purported to express PrP at physiological levels in all appropriate cell types, Gt mice have been endorsed as ideal models in which to study peripheral pathogenesis of prion diseases (38). Our findings confirm that this is the case. Gt mice offer a refined improvement on previously described Tg models of CWD, and a new application to explore peripheral CWD pathogenesis and contagious spread of CWD prions. Remarkably, GtQ226+/+ mice showed no difference in time to disease onset following intraperitoneal or intracerebral challenges with CWD (Fig. 4B). In contrast, despite robust PrP over expression in the CNS of Tg1536+/− mice (24), which resulted in more rapid times to disease following intracerebral challenge than counterpart GtQ226+/+(ic) mice expressing physiological levels of PrP, the response of Tg1536+/− mice to intraperitoneal challenge with CWD prions was highly inefficient, reflected in prolonged times to disease onset compared with GtQ226+/+(ip) mice (Fig. 4C). We interpret this to mean that the cos.SHaTet cosmid used to create Tg1536+/+ mice (24) failed to reconstitute appropriate PrP expression in all compartments required for effective peripheral pathogenesis. In accordance with this interpretation, while spleens of diseased GtQ226+/+ mice accumulated PrPSc following either intracerebral or intraperitoneal inoculation, PrPSc was undetectable by Western blotting in spleens of diseased Tg1536+/+ mice that were challenged by the intracerebral or intraperitoneal route (Fig. 4G).
We also show that the route of CWD prion inoculation elicited a pronounced effect on the distribution of cervid PrPSc-containing lesions in the CNS of Gt mice. Accordingly, while intracerebral inoculation of GtE226+/+ mice with 99W12389 CWD prions resulted in widespread, diffuse, and bilaterally symmetrical PrPSc deposition in cortical, hippocampal, thalamic, and hypothalamic regions (Fig. 3), the pattern of E226-cervid PrPSc deposition in GtE226+/+ mice challenged intraperitoneally with the same CWD prions was different from, and more restricted than in GtE226+/+(ic) mice (Fig. 5). Interestingly, when prions from the brains of these mice were serially passaged to additional GtE226+/+ mice by intracerebral inoculation, the pattern of cervid PrPSc deposition in these GtE226+/+(ip-ic) mice reverted to that seen in GtE226+/+ mice intracerebrally challenged with 99W12389 (Fig. 5G). These findings are consistent with the results of previous findings in wild-type mice infected with experimentally adapted scrapie prions, which showed that route of administration affected the profile of neuropathological lesions (56), and underscore the importance of inoculation route when defining prion strain properties. In light of their highly efficient responses to peripherally administered prions, we reasoned that Gt mice might provide an effective means to model horizontal CWD transmission. Our preliminary findings show Gt mice to be a tractable experimental model in which to ascertain the underlying basis for the remarkably contagious transmission of CWD prions (46).
In conclusion, the Gt mice described here represent an important resource with which to address multiple unresolved issues relating to CWD. Our findings describing the role of amino acid differences at residue 226 of deer and elk PrP in dictating the species-specific selection and propagation of cervid prion strains has important implications for assessing the zoonotic potential of established, as well as newly emergent forms of CWD. Of immediate significance, the origins of CWD in Europe (10, 11), and for that matter North American CWD, are enigmatic, and the relationship, if any, between CWD prions from these locations is currently unclear. Understanding whether residue 226 plays a role in the selection and propagation of novel emergent strains in this context would appear to be of particular importance. While prions causing CWD in South Korea are considered to be derivative of North American CWD (7, 8), whether the properties of these CWD prions were altered as a result of passage through indigenous cervid species also remains to be determined. Comparing the properties of North American, European, and Asian CWD prions using Gt mice represents an important first step to assess the prospect of novel global CWD epidemics with unpredictable ecological and public health consequences. Finally, the refined strategies for PrP gene replacement reported here suggest opportunities for improved Gt mouse models in which to study aspects of human and animal prion biology that have until now not been possible using conventional Tg approaches. The generation of optimized Gt mice expressing human PrP rather than chimeric mouse–human PrP constructs would appear to be of particular importance in this regard.
Detailed descriptions of the methods and materials used for the production and characterization of Gt mice, including RNA extraction and analysis by qRT-PCR; prion transmission studies; analyses of PrPC and PrPSc by Western blotting, histoblotting and IHC; and analysis of the conformational stability of PK-resistant PrPSc by GdnHCl denaturation are included in SI Appendix. Production and characterization of Tg1536+/+ and Tg5037+/+ mice has been described previously (24, 25). All procedures used in this study were performed in compliance with the Colorado State University Institutional Animal Care and Use Committee.
↵1J.B. and J.R.C. contributed equally to this work.
Author contributions: G.C.T. designed research; J.B., J.R.C., J.A.M., S.J.K., V.K., J.G., and S.K. performed research; J.B. and G.C.T. analyzed data; and G.C.T. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1903947116/-/DCSupplemental.
This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).
