We report here that the biallelic R136S PRNP mutation causes inherited human prion disease with early onset of symptoms (30s–40s). The same mutation presented in heterozygous state in other family members did not induce the same phenotype. Clinically, the patients presented with early-onset dementia developing motor disturbances at follow-up. Neuropathological examination revealed the presence of abundant multicentric PrP amyloid plaques with mild neuritic component, along with severe neuronal loss, gliosis, and microglial proliferation with only mild spongiform change, which is consistent with a GSS neuropathological phenotype. In contrast to previous description of homozygous mutations in PRNP [10,11,12,13, 27, 31] (Table 1), the disease in R136S families presented with an autosomal recessive pattern of inheritance, keeping the proven or obligated heterozygous family members protected from developing the disease, at least, at early ages. In vitro propagation and in vivo transmissibility studies suggest that the PrPSc variant resulting from the R136S mutation failed to propagate both in vivo and in vitro in the conditions analyzed, suggesting that this variant is biologically less active than other mutations causing genetic prion diseases.
Previous described pathogenic mutations cause human prion disease in the heterozygous state and show an autosomal dominant inheritance pattern. There are scarce examples in the literature of patients affected by prion diseases carrying homozygous genetic variants. Homozygous E200K carriers are associated with consanguinity in areas of high prevalence of this variant. Homozygous carriers of the E200K mutation present an earlier age of onset than heterozygous carriers but no other relevant differences in clinical features compared with heterozygous carriers . One case of prion disease linked to homozygous Q212P has been described, but also another case has been reported in a heterozygous carrier . Komatsu et al.  reported the case of a homozygous V203I patient, in addition to three previously reported heterozygous patients for this variant. However, the lack of family history of disease in all the V203I and the presence of this variant in normal controls causes that its pathogenicity is still under discussion [3, 11, 31]. Recently, Hassan et al.  also reported a case with CJD and a homozygous E200D mutation, similar to Q212P .
Minikel et al.  reported in 2016 the presence of the R136S variant in heterozygosis in two alleles from 60.706 population control exomes but not in patients, suggesting that this represents a very rare variant in the normal population even if the R136S variant is predicted to be “probably damaging” in several pathogenicity prediction software. In our center, we had not identified this genetic variant in more than 300 Spanish subjects (including both patients and controls) analyzed so far. In 2006, Pacheco et al.  described one Portuguese patient (Fig. 1B, subject II.1) with similar clinical and neuropathological features of those of our patients (early-onset dementia, motor symptoms in her early 50s and a GSS neuropathological phenotype) linked to the presence of the homozygous R136S variant. Two of her siblings presented history of similar disease in their early 50s, but not the parents who should be obligated carriers.
In both families (the present family and the Portuguese family) described heterozygous carriers were apparently protected from disease. As the genetic study in the Portuguese index case’s mother (Fig. 1B, subject I.2) and another family member neurologically preserved in their 80s and 90s revealed the presence of heterozygous R136S linked to 129MV, the authors suggested that the presence of valine in the trans-allele could prevent the expression of the disease. In contrast, in the Spanish family, one heterozygous carrier, homozygous for methionine at codon 129 is neurologically preserved in their 70s, suggesting that the expression of the disease does not depend on the codon 129 of the trans-allele but in the homozygous state or “dosage” of the mutant allele.
The expression of the disease only in homozygous subjects could be explained by a dose effect as it has been suggested for an earlier age of onset in homozygous E200K carriers . A larger number of R136S PrP molecules (or the absence of wild-type protein) in the homozygous patient’s brain may increase the chance of spontaneous conversion. Even if homozygous patients are at higher risk for the spontaneous conversion to PrPSc than heterozygous patients, on the other hand, the subsequent infective process based on PrPSc-PrPC interaction might be less effective as there is only mutant PrPSc leading to longer disease duration.
Even if we cannot rule-out a late onset of the disease, the presence of known or obligated heterozygous carriers in neurological preserved family members that were already in their 70s, 80s, or 90s make this possibility much less probable during the usual lifespan. In this sense, we believe that both family trees (Fig. 1) provide evidence for a recessive pattern of inheritance in human prion diseases.
The clinical phenotype in the patients resembled that of insertion mutations (OPRI), which are characterized by early-onset progressive dementia with late motor dysfunction , whereas the neuropathological features were typical of GSS. The observed clinical phenotype is atypical for GSS, which more commonly present with motor signs followed by dementia, although it has also been linked to other point mutations (Y218N)  and OPRI [31, 32]. Irrespectively of the clinical syndrome and the relative regional distribution of lesions, GSS is defined by the presence of multicentric PrP amyloid plaques in the cerebral and cerebellar cortices [7, 18]. In contrast to other human prion diseases, the presence of spongiform change can be variable, from absent to severe, which could explain the lack of signal intensity changes in diffusion-weighted imaging in MRI scans. In addition, Western blot studies typically show, as in the present case, PK-resistant non-glycosylated PrPSc fragments with a low molecular weight that varies between 6 and 8 kDa depending on the PRNP pathogenic variant. Occasionally, however, the immunoblot profile also shows larger truncated PrPres fragments with a molecular weight of 19–21 kDa, as observed in other sporadic or genetic prion diseases [33,34,35,36].
Animal modeling of human inherited prion diseases is not easy. Practically all attempts to generate transgenic mouse models for human inherited prion diseases using the human-PrPC sequence had been unsuccessful apart from the recent reported exception of an A117V GSS model . The rest of successful transgenic mouse models that develop spontaneously neurologic disorders when harboring mutations that have been reported in human inherited prion diseases have been made in mouse or bank vole PrP sequences or in chimeric mouse/human PrP proteins . It seems that mouse and bank vole PrP proteins are more prone to spontaneous misfolding while human PrP is apparently more resistant. Alternatively, interactions between mouse PrP and other than PrP mouse factors may be important for the spontaneous generation of prions. In general terms for GSS modelling, models overexpressing mouse, mouse/human chimeric, and cow PrP harboring the respective equivalents to human P102L mutation spontaneously developed a prion disease along with neuropathological changes . However, brain PrPres detection and transmissibility to wild-type mice was only achieved in the particular case of equivalent P113L mutation in cow-PrPC sequence producing a prion agent with features resembling those of classical bovine spongiform encephalopathy . When the A117V mutation was overexpressed in the mouse sequence (A116V) at 4–6x levels, transgenic mice showed signs of prion disease including neuropathology markers but no PrPres were detected . However, A117V GSS cases were successfully transmitted into A117V human PrP transgenic mice Tg30 and Tg31 (2x and 3x overexpression respectively) [37, 41] from which Tg30 developed a spontaneous disease that was later transmissible to both Tg30 and Tg31 mice as well as to wild-type V129 human PrP transgenic mice . It would be of interest to further investigate the effects of this homozygous mutation in an animal model, as well as the production of heterozygous animals to further investigate to which extent the presence of the wild-type allele dumpers the spontaneous misfolding of the mutated one. The recessive inheritance pattern displayed by this disease in humans points to a protective effect exerted by the wild-type variant and the in vitro studies performed in this work also supports this hypothesis. However, a slower misfolding rate of the wild-type allele that may finally contribute to disease development at elder ages cannot be ruled out.
The in vitro propagation and spontaneous misfolding proneness studies, performed by recPMCA, indicate that the R136S mutation does not confer an enhanced misfolding propensity to the human recombinant PrP. No differences were detected on misfolding proneness of this protein compared to the wild type, neither induced or spontaneously. In fact, prion ability to cause direct cellular damage (toxicity) has often been dissociated from the capacity of successfully transmit from an infected host to an uninfected recipient (transmissibility) . This, added to the difficulty of modelling inherited prion disease that is further discussed below, may explain the lack of in vivo and in vitro propagation of the prion disease associated to R136S mutation. The synthesis of peptides containing the mutation and testing their effect on cultured cells, apart from allowing determining differences in the misfolding proneness of mutated proteins as the PMCA assays carried out here, could allow evaluating potential changes in their neurotoxic properties .
This report has some limitations. First, the diagnosis of genetic prion disease is a hurdle task and some members of both extended families might have been underdiagnosed, lowering the possibilities of describing this mutation in other patients. Second, due to the lack of neuropathological evaluation in any of the asymptomatic heterozygous carriers, we could not rule-out subclinical prion disease at late ages, although this possibility seems less probable.