Introduction
The variability of the clinical presentation in Creutzfeldt-Jakob disease (CJD) has been known for many years and has led to the identification of several clinical variants [
5]. In subsequent years, we showed that the phenotypic heterogeneity of the sporadic form of CJD (sCJD) is largely driven by the pairing of two sets of major determinants of disease phenotype: the genotype ̶ MM, MV, VV ̶ at the methionine (M)/ valine (V) polymorphic codon 129 of the human prion protein (PrP) gene and the type, 1 or 2, of the pathogenic, disease-related form of the prion protein (PrP
D) [
12,
24,
25,
27,
29]. PrP
D type 1 and 2, which were initially distinguished by the electrophoretic mobilities as species with 21 kDa (type 1) and 19 kDa (type 2) proteinase-resistant cores, resPrP
D, of the unglycosylated forms, were shown to represent two fundamental prion strains associated with sCJD [
3,
29,
31]. This concept led to a new classification of sCJD that theoretically included six phenotypes or subtypes: MM, MV and VV, each of which could be associated with PrP
D type 1 or 2. Indeed, as expected, sCJD patients with the 129 MM genotype coupled with PrP
D type 1, commonly identified as sCJDMM1 (hereafter MM1), as well as sCJDVV2, VV1 and MV2 were associated with distinct disease phenotypes (Table
1). Some of these phenotypes mimicked the clinical variants originally proposed, while some were novel. However, detailed analyses revealed two apparent inconsistencies. First, MM1 and MV1 were shown to share histopathological phenotype and PrP
D electrophoretic profile, prompting their grouping into one sCJD subtype, namely MM (MV)1 [
12,
27]. Second, MM2 was associated with two quite distinct phenotypes: one with sCJD-like features including severe spongiform degeneration (SD) of the cerebral cortex, named sCJDMM2 (or MM2C for cortical), and the other, very rare, characterized by insomnia and severe thalamic atrophy that is commonly referred to as sporadic fatal insomnia (sFI or MM2T for thalamic) [
8]. Specific variations of PrP
D as well as distinct transmission characteristics indicate that MM2 and sFI are associated with distinct prion strains [
3,
9,
19,
26].
Table 1Synopsis of molecular classification and phenotypes of sCJD subtypes and variants
MV2K; ~ 8 | SD pseudo-laminar in CC (Fig. 1a-d); kuru pl. in Cbl. IHC: fine punctate in CC; Cbl (Fig. 1a-d) | Type 2 (19 kDa) + Type ~ 1 (20 kDa) |
MV2C; ~ 2 | SD: large confluent vacuoles mostly in CC. IHC: co-distributing with SD (Fig. 1e-h). | Type 2 (19 kDa) |
MV1; 5 | SD with fine vacuoles predominantly in CC; IHC: fine punctate pattern in SD regions; in Cbl “brushstroke” pattern (Fig. 1s-u) | Type 1 (21 kDa) |
VV2; 15 | SD like MV2K in CC; no kuru plaques but plaque-like aggregates in Cbl (Not shown) | Type 2 (19 kDa) |
MM2; 4 | | Similar but not identical to MV2C |
MM1; 64 | | As MV1 |
Recently, the issue of phenotypic heterogeneity within the MV2 subtype has also emerged [
23,
28] (Table
1). The MV2 subtype was shown to comprise two basic histopathological phenotypes (or histotypes) as well as two electrophoretic profiles of resPrP
D: a variant that was indistinguishable from the MM2 subtype (named sCJDMV2C) and a second phenotype that, although generally mimicking sCJDVV2, differed from the latter with regard to the prominent presence of kuru plaques (MV2K) as well as the presence of a small additional resPrP
D component in the otherwise type 2 resPrP
D profile. The coexistence in the same case of the two histotypes in various proportions (MV2K-C) was also observed.
In 2016, Moore et al. carried out a mass spectrometric study to determine the relative amount of resPrP
D-129 M and -129 V in PrP
D enriched preparations extracted from MV1, MV2C and MV2K cases as well as cases with mixed histotypes, such as MV2K-C and MV1-2C [
20]. Based on the finding that the ratios of resPrP
D-129 M and -129 V allotypes were highly variable, the authors concluded that, in MV2 and MV1 cases, normal or cellular PrP (PrP
C)-129 M and -129 V had “similar tendency to misfold” into PrP
D regardless of the MV histotype, and that “factors other than the PrP
Sc allotype abundance must influence the clinical progression and the phenotype of heterozygous cases of CJD”.
Here, we have re-examined the issue of resPrPD-129 M and -129 V relative abundance in the sCJDMV case subset using mass spectrometry. Given the variability, sometimes subtle, of the phenotype in these cases, we extensively analyzed the histotype and PrPD characteristics to select MV2C and MV2K cases associated with only one phenotype (called “pure”), along with MV2K-C mixed cases, where the two histotypes coexisted in well-defined proportions. In all MV2 cases examined, we consistently observed a positive correlation between the relative dosage of resPrPD-129 M and -129 V allotypes and the MV2C and MV2K histotypes. However, surprisingly, in the MV1 variant, resPrPD-129 M and -129 V were equally represented.
Discussion
It is increasingly clear that successful treatment of prion diseases may need to be tailored to individual prion strain [
13]. This notion underscores the need for the detailed understanding of the mechanisms of strain emergence and selection that lead to the heterogeneity of human prion diseases. This mechanistic insight will also likely have implications for development of therapeutic strategies in other neurodegenerative diseases that are characterized by prion-like propagation and structural polymorphism of misfolded protein aggregates.
Previous studies showed that both the PrP
D type 1 and 2 selection as well as phenotypic heterogeneity of sCJD are largely under the control of the 129 allotype (for review see [
12,
34]). However, this scenario does not explain the phenotypic heterogeneity of the sCJDMV2 subtype, where the heterogeneity occurs without any variation of the 129 allotype and PrP
D type (Table
1).
To bridge this gap, here we have performed detailed characterization of different subtypes of sCJDMV cases, and for each of these cases we have used a mass-spectrometry based approach to determine the relative proportion of resPrPD-129 M and -129 V. This analysis revealed a novel, versatile mechanism by which residue 129 polymorphism of resPrPD determines phenotypic heterogeneity, resulting in multiple variants within the MV2 subtype of sCJD. In particular, our data show that the occurrence of two variants, referred to as MV2C and MV2K, either with pure or mixed phenotypes, is directly related to the relative abundance of resPrPD-129 M and 129 V. The pure MV2C phenotype is characterized by large predominance of resPrPD-129 M. By contrast, in the pure MV2K phenotype, the predominant form of resPrPD contains Val at position 129, with resPrPD-129 M accounting only for ~ 23% of total resPrPD. Furthermore, consistent with the notion that the nature of amino acid at position 129 (M or V) is a major determinant of disease phenotype in heterozygous cases of sCJD, we found that, in MV2K-C mixed phenotypes, the relative abundance of resPrP-129 M and -129 V correlates with the representation of the corresponding histotypes in these mixed cases. Of note, the second MV2K-C case shows that histotype-related quite different allotypic proportions of resPrPD may occur in two separate regions of the same brain (in this case, cerebral and cerebellar cortices), suggesting that local mechanisms play a role in determining this topographic heterogeneity.
Different proportions of resPrPD corresponding to two allotypes as determined in this study likely result from the preferential templated conversion of the normal form of PrP (PrPC) that matches the allotype of the dominant resPrPD. One cannot completely rule out the possibility that the observed different proportions of resPrP-129 M and -129 V in MV2C and MV2K cases result from differences in the resistances to PK digestion. However, this possibility is highly unlikely given that the MV2C and MV2K variants share PK titration profiles and conformational stabilities or both with the VV2 and MM2 subtypes, which, being 129 homozygous, are associated with PrPD isoforms allotypically homogeneous.
Remarkably, even in the histotypically pure cases, neither resPrP
D-129 M nor the resPrP
D-129 V allotype accounted for 100% of total resPrP
D, with the minor allotype always observed. Even though one cannot absolutely rule out the possibility that this apparent coexistence of both resPrP
D allotypes in pure phenotypes is due to a systematic error of our mass spectrometry-based measurements, this is rather unlikely. Thus, the presence of a small population (~ 20%) of resPrP
D-129 V in MV2C cases appears to be an intrinsic property of this phenotype. This is very intriguing given the similarity of the MV2C and MM2 subtypes, both with regard to the histotype and resPrP
D conformational stability index. In this variant, PrP
C-129 V might be converted into PrP
D that adopts the PrP
D MM2-like conformation and contributes to the MM2-like histotype. Alternatively, this minor resPrP
D component might remain “silent” with its histotype undetectable [
4,
7].
The coexistence of resPrP
D-129 V and -129 M (77:23 ratio) in the MV2K variant raises even more challenging questions, especially given the presence in this variant of two resPrP
D electrophoretic components of 19 kDa and 20 kDa. Data from our and other laboratories allowed for detailed characterization of these two components. With regard to the allotype, our sole successful attempt at sequencing the 20 kDa resPrP
D component indicated the presence of the M residue at position 129. However, transmission of the MV2K variant to knock-in 129 MM, MV and VV transgenic mice demonstrated that the 20 kDa component can be replicated only in the presence of at least one 129 M allele [
16] . Altogether, these data identify the 20 kDa fragment as the sole component populating resPrP
D-129 M in the MV2K subtype. Furthermore, further bioassays in 129 M allotypic mouse lines have shown that the 20 kDa resPrP
D fragment can also be replicated upon inoculation of VV2 sCJD prions (which lack the 20 kDa resPrP
D component) [
16]. This observation suggests that the 20 kDa resPrP
D component in the MV2K subtype does not result from a direct templated conversion of 129 M PrP
C but is rather due to the adaptation of the VV2 19 kDa strain to the presence of the 129 M allele. Our sequencing efforts have identified residues G82 and G86 as N-termini of the 20 kDa resPrP
D component. This is consistent with the two most N-terminal residues that we observed in MV2 cases in a previous systematic sequencing study of different sCJD subtypes [
29]. These two N-termini of the 20 kDa resPrP
D component differ from those previously identified (as residues G78 and G82) in typical type 1 resPrP
D associated with the MM1 phenotype [
29]. Furthermore, our conformational stability data surprisingly indicate that, despite the immunoreactivity properties and N-terminus characteristics consistent with resPrP
D type 1, the 20 kDa resPrP
D species in the MV2K subtype appears to have conformational features of resPrP
D type 2, which significantly differ from those of all type 1 subtypes. Finally, a body of evidence indicates that the 20 kDa component of resPrP
D is the determinant of kuru plaque formation [
6,
16] and [Cali et al. unpublished data]. Altogether, these features define PrP
D associated with the MV2K subtype as a unique prion strain in sCJD.
Based on the widely accepted notion that the MV1 variant is indistinguishable from the MM1 subtype, both with regard to the histotype as well as PrPD and transmission characteristics, we expected the former variant to show higher proportion of resPrPD-129 M than resPrPD-129 V, as is the case for the MV2C variant. The finding of equal representation of the 129 M and 129 V allotypes in the MV1 variant raises questions as to the role of these two components in histotype determination. One plausible scenario would require PrPC-129 V to be converted into PrPD with structural characteristics of PrPD-129 M type 1, participating in determination of the same histotype. Should this be the case, MV1 PrPD would be the first example of human prion strain which, despite being comprised of similar amounts of two different resPrPD allotypes, is associated with a homogeneous histotype.
Finally, it must be noted that our findings are at odds with the previous study of Moore and coworkers [
20], who examined a similar case population using a similar mass spectrometry-based approach and found no systematic phenotype-specific differences between the populations of two resPrP
D isoforms. This is likely related to less stringent criteria used in the previous study for case selection, which could have resulted in the disproportionate representation of mixed (rather than pure) cases. Indeed, 10 out of the 14 sCJDMV cases examined in the previous study were mixed cases, either MV1-2C/K (4 out of 5 cases) or MV2K-C (6 out of 8 cases). Thus, it seems not surprising that the examination of a population of such mixed cases would lead to inconclusive results. However, when one considers only the “pure” cases used by Moore and colleagues, the relative proportions of the two resPrP
D isoforms appear to be comparable to those determined in the present work. Furthermore, all but one (possibly atypical or mixed) of the MV2K iatrogenic CJD cases classified in the study of Moore et al. as “pure” show the predominance of resPrP
D 129 V over 129 M. Therefore, applying more stringent case selection criteria, the findings by Moore et al. are consistent with ours.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.