Activation of the unfolded protein response and granulovacuolar degeneration are not common features of human prion pathology

Human prion diseases are rare, rapidly progressive, invariably lethal neurodegenerative diseases, symptomatically characterized by severe memory impairment and a general decline in cognitive functions, which may include motor, linguistic, executive and social skills [1]. Most often, human prion diseases have a sporadic aetiology (e.g. sporadic Creutzfeldt-Jakob disease (sCJD)), but hereditary (e.g. fatal familial insomnia (FFI) and Gerstmann-Sträussler-Scheinker syndrome (GSS)), and infectiously acquired (e.g. iatrogenic CJD (iCJD), kuru and variant CJD (vCJD)) forms of the disease also exist [2]. The clinical duration varies between the subgroups of the disease, with a mean disease duration ranging from 2 to 3 months for sCJD, 12 months for kuru and 5 years for GSS [3]. Prion diseases have also extensively been described in animals, including bovine spongiform encephalopathy (BSE) in cattle and scrapie in sheep [2]. At the neuropathological level human prion diseases are characterized by the accumulation of pathological prion protein (PrP^Sc), neuronal loss, astrogliosis and spongiosis [4, 5]. PrP^Sc arises after the post-translational conformational conversion of the cellular prion protein (PrP^C). PrP^C is a constitutively expressed protein of uncertain function that contains a glycosylphosphatidylinositol (GPI) anchor, facilitating its association with the cell membrane. During human prion diseases PrP^C is converted into insoluble, β-sheet rich PrP^Sc aggregates that are usually resistant to digestion by proteinase-K. This pathological PrP^Sc conformer exhibits intriguing characteristics that, once formed, ensure conversion of native PrP^C into PrP^Sc and propagation of pathology to neighbouring cells [6–10].

Previous research has pointed out that the endoplasmic reticulum (ER), an organelle essential for protein processing, might constitute a link between prion pathology and neurodegeneration [11–18]. If the functioning of the ER is disturbed, for example when un- or misfolded proteins accumulate in the lumen of the ER or in case of an aberrant calcium concentration, an adaptive programme called the unfolded protein response (UPR) is induced [19]. The UPR aims to protect the cell by restoring protein homeostasis. To this end, three canonical signalling cascades are activated by detachment of the ER luminal chaperone BiP/Grp78 from the ER stress sensors dsRNA-activated protein kinase-like ER kinase (PERK), activating transcription factor 6 (ATF6) and inositol-requiring enzyme 1 α (IRE1α). The net effect of UPR activation is relief from the ER stress, via attenuation of translation and upregulation of genes that promote chaperone synthesis and protein degradation. However, prolonged activation of the UPR can drive its outcome from adaptation to apoptosis, as programmed cell death can then be induced via among others the pro-apoptotic proteins CCAAT/enhancer-binding-protein homologous protein (CHOP), jun NH₂-terminal kinase (JNK) and caspase-12 [19–22]. In human brain tissue activation of the UPR has been detected in several neurodegenerative diseases including Alzheimer’s disease (AD), frontotemporal lobar degeneration with tau pathology (FTLD-tau), Parkinson’s disease (PD) and amyotrophic lateral sclerosis (ALS) (for review see [23]). Prominent UPR activation is observed in AD and FTLD-tau in close association with the accumulation of phosphorylated tau [24–26]. In addition, different UPR activation markers are associated with granulovacuolar degeneration (GVD). GVD is characterized by basophilic granules surrounded by a clear zone measuring 1–5 μm in diameter, occurring predominantly in hippocampal neurons in AD and marked by the increased appearance of a variety of proteins related to cell stress [25, 27].

Studies investigating UPR activation in human prion diseases are very limited and have so far yielded inconclusive results. One study reported increased expression of BiP and several other ER chaperones in post-mortem brain samples of sCJD and vCJD patients [12]. On the other hand, immunohistochemistry on post-mortem brain tissue of human prion disease patients could not detect activation of the PERK pathway [28]. In order to investigate the possibility that other routes of the UPR are activated in human prion disease, we assessed the presence of phosphorylated IRE1α (pIRE1α) on post-mortem frontal cortex tissue of an extensive cohort of sporadic, inherited and acquired human prion disease patients. To align with the earlier report, immunohistochemistry for phosphorylated PERK (pPERK) was performed in our independent cohort. In addition, we studied the presence of a marker for GVD, as a pathological hallmark associated with the increased presence of UPR activation and cell stress-related markers [27].

Many neurodegenerative diseases are characterized by the deposition of misfolded protein aggregates. Therefore, it is suggested that these diseases share the involvement of similar cellular signalling mechanisms associated with protein folding and clearance. Previously, we have demonstrated increased presence of several UPR activation markers, including pPERK and pIRE1α, in post-mortem brain tissue of AD, FTLD-tau and PD [24, 40, 41]. In the present study, we aimed to further elucidate the involvement of the UPR in human prion pathology. To this end, we selected cases from the Dutch cohort of human prion disease patients [29] and performed immunohistochemistry to visualize UPR activation markers on post-mortem frontal cortex sections. Immunoreactivity for pIRE1α and pPERK was not increased in human prion disease cases compared to non-neurological controls. In addition, immunoreactivity for CK1δ, a marker associated with GVD and cellular stress, was not elevated in human prion disease cases.

Earlier observations indicate a close spatiotemporal relationship between UPR activation and the presence of tau pathology [24–26]. Several mechanistic links between tau pathology and UPR activation have been described [42–49]. Previously, Unterberger and colleagues reported that immunoreactivity for pPERK in human prion diseases was observed to be associated with tau pathology only in cases with concomitant AD pathology [28]. Human prion disease patients often present modest tau positive lesions that are believed to be secondary to the prion pathology itself. This type of tau amyloidosis comprises small neuritic and extracellular inclusions consisting of hyperphosphorylated tau [34, 37]. In the present study, almost all human prion disease cases presented this type of tau pathology in the absence of profound UPR activation, as assessed by the UPR activation markers pIRE1α and pPERK. The data presented in this study suggest that this modest type of tau pathology secondary to prion pathology is not functionally linked to UPR activation in human prion diseases. In contrast, other reports on human neuropathology do support the involvement of UPR activation in neurodegenerative diseases with extensive tau pathology [23]. Interestingly, Radford and colleagues showed neuroprotective effects of a selective PERK inhibitor in a mutant tau transgenic mouse model, providing proof-of-concept for addressing the UPR as therapeutic target for the treatment of primary tauopathies [50].

In this study we detected profound levels of the UPR activation marker pPERK, the GVD marker CK1δ and neurofibrillary tau pathology in the frontal cortex of a patient presenting an unusual GSS phenotype (case #14). Genetic analysis of this patient revealed a premature stop codon mutation (Q227X) that resulted in a truncated prion protein devoid of the GPI anchor that normally facilitates implementation of the prion protein in the cell membrane [39]. Previous studies in cellular and murine models of genetic prion disease have shown delayed maturation, prolonged retention and accumulation of mutant PrP in the ER, which could be a primary cause of ER stress [11, 15]. The absence of Aβ deposits in the frontal cortex of case #14 indicates that the severe tau pathology is related to the unusual PrP pathology rather than to concomitant AD pathology. Neurofibrillary tau pathology has been documented earlier in other rare hereditary prion disease cases in which mutations in PRNP cause truncation of PrP and absence of the GPI anchor [51, 52] as well as other forms of GSS caused by distinct PRNP mutations [53–55]. Notably, the presence of GVD was previously mentioned by a case report on a longstanding GSS patient (P102L) with pathological features comparable to case #14, showing neurofibrillary tau pathology secondarily induced by prion amyloidosis [53]. Molecular properties of the prion protein itself as well as the duration of the clinical course could be requisites for the emergence of neurofibrillary tau pathology and UPR activation in human prion diseases.

Data obtained from some experimental models of prion diseases suggest an involvement of the UPR in prion disease pathology. The expression levels of BiP and several other ER chaperones are (transiently) increased in N2A neuroblastoma cells after treatment with PrP^Sc derived from scrapie-infected mice and in several brain regions in a murine scrapie model [12, 14]. In addition, the UPR was implied as part of the pathological process of prion disease in a study detecting upregulation of BiP in RNA samples derived from the brainstem of BSE-infected cattle [16]. In contrast, another study did not detect increases in BiP and CHOP expression and XBP1-splicing, a downstream effect of IRE1α activation, in the brains of mice transgenic for mutant PrP and in primary neurons and HEK293 cells transfected with mutant PrP conformers [56]. Discrepancy between pathogenic mechanisms in play in experimental models of prion disease may arise from experimental methodology, e.g. by differences in strain and titer of the administered prions and the use of PrP^C overexpressing models.

One of the down-stream effects of UPR activation is a reduction of protein synthesis through phosphorylation of the eukaryotic initiation factor 2α (eIF2α), as a results of increased kinase activity of PERK. Prion infection of PrP overexpressing transgenic mice increases the levels of UPR activation markers, including pPERK and phosphorylated eIF2α [18]. In these mice, reduction of eIF2α phosphorylation rescues synaptic deficits, neuronal loss, and increases survival of prion-diseased mice [18]. In addition, oral administration of a kinase inhibitor of PERK or the small molecule ISRIB, which restores translation downstream of eIF2α, conferred neuroprotection in these prion-diseased mice [57, 58]. The discrepancy between on the one hand indications that the PERK pathway is activated in prion-diseased mice and on the other hand the absence of markers indicative of PERK activation in human prion disease (this study and [28]) could be attributed either to species differences or to the experimental methodology used for prion-diseased mice. Nevertheless, the involvement of the UPR in prion disease models and the therapeutic potential of targeting the UPR observed in these models contribute to understanding the role of the UPR in protein misfolding diseases.

A possible caveat of this study could be an effect of the post-mortem delay on the preservation of phosphorylated antigens that are needed in order to detect UPR activation by immunohistochemistry. However, this concern is undermined by i) the clear presence of pIRE1α and pPERK in the single prion disease cases and AD positive controls that do show immunoreactivity with comparable handling and post-mortem delay, ii) by the marked detection of pIRE1α and pPERK immunoreactivity in AD, but not human prion disease, hippocampus of cases with a post-mortem delay comparable to about half of the human prion disease cases included in this study, iii) by the detection of phosphorylated tau using the phospho-specific AT8 antibody in almost all prion disease cases and iv) by the use of CK1δ as non-phospho-specific surrogate marker of UPR activation.

In conclusion, our data suggest that the IRE1α branch of the UPR is not implied in human prion diseases. In addition, we confirm earlier data on the absence of increased levels of pPERK in human prion disease in an extensive independent cohort [28]. Furthermore, in this study we report no profound increase in CK1δ immunoreactive structures that resemble GVD in human prion diseases. Since UPR and other markers for cellular stress are detected alongside in neurons undergoing GVD (for review see [27]) future research should address the functional relation between these markers in order to understand the pathogenesis of GVD. In regards to prion disease, future research should focus on the involvement of alternative UPR activation markers in post-mortem brain tissue and atypical activation of the UPR in prion disease models. Results obtained in prion-infected mice indicate alongside the involvement of the PERK pathway in neurodegeneration, that pharmacological inhibition of this pathway is a potential therapeutic target [57, 58]. While available data on UPR activation in human prion diseases remain limited, this type of neuropathological data is essential for proper interpretation of recent therapeutic advances in the field of prion disease.