Atypical Creutzfeldt-Jakob disease with PrP-amyloid plaques in white matter: molecular characterization and transmission to bank voles show the M1 strain signature

Patients and tissues

We studied 5 subjects affected by CJDMM1 associated with PrP^Sc plaque-type deposits in white matter (hereafter indicated as p-CJDMM1) and 8 cases affected by typical CJDMM1 (hereafter indicated as np-CJDMM1). All cases were referred for diagnosis to the Laboratory of Neuropathology, University of Bologna, Italy between 2005 and 2016 as part of the National Surveillance program on CJD and related disorders or (one p-CJDMM1) in the context of a collaborative effort with the Dutch Surveillance Centre for Prion Diseases on the molecular characterization of autopsy confirmed prion cases [10]. The 8 selected np-CJDMM1 control cases were representative of the spectrum of clinical and histopathologic features of the sCJDMM1 subtype [21, 25] including disease duration (range 1–14 months).

Brains were obtained at autopsy, one half, or tissue blocks from representative areas, were immediately frozen at −80 °C, whereas the rest was fixed in formalin.

Clinical and diagnostic evaluation

We collected and reviewed all available medical information from hospital reports, including results of neurologic examination(s), cerebral magnetic resonance imaging (MRI) studies and electroencephalographic (EEG) recordings. We defined the date of disease onset as the time when unexplained progressive neurological or psychiatric symptoms first occurred, and as ‘onset symptom(s)’ the first neurological disturbance(s) complained by the patient. We measured total tau (t-tau) protein levels in the cerebrospinal fluid (CSF) by quantitative ELISA (INNOTEST hTAU Ag, Innogenetics) according to the manufacturer’s instructions, considering as an optimal cut-off value 1250 pg/mL on the basis of receiver operating characteristic curve analysis, as previously described [16]. Semi-quantitative detection of CSF 14–3-3 protein was performed by western blotting, as previously described [16].

Genetic analysis

Genomic DNA was extracted from blood or frozen brain tissue. Genotyping of the PRNP coding region was performed as described [10].

Neuropathology

We semi-quantitatively evaluated gray matter spongiform change and astrogliosis in 10 brain regions on hematoxylin and eosin stained sections, as reported [21].

For PrP immunohistochemistry, paraffin sections from formalin-fixed and formic acid treated blocks were processed using the monoclonal antibody (mAb) 3F4 (1:400, Signet Labs), according to published protocols [11, 22], with some modifications. Briefly, after de-waxing and re-hydration, sections were incubated for 15 min in 8% hydrogen peroxide solution in methanol to block endogenous peroxidase. Sections were then washed, immersed in 98% formic acid for 1 h, rewashed and microwaved in 1.5 mM HCl for 25 min, incubated with reagent A of Histostain-Plus IHC Kit (Thermo-Fisher Scientific) for 10 min and then probed overnight with mAb 3F4. After two sequential incubations with reagent B and C of Histostain-Plus IHC Kit interspersed with washing steps in TBS 1X, sections were treated with Romulin AEC Chromogen (Biocare Medical) for 5 min and Mayer’s hematoxylin for 15 s before being dehydrated, cleared and coverslipped.

We carried out a specific assessment of white matter changes of myelin, axons, astrocytes and microglia by means of Luxol Fast Blue (LFB) and several immunohistochemical stainings in section from the frontal, temporal and occipital cortices and the cerebellum. To this aim the following antibodies were used: 1) anti-myelin proteolipid protein (PLP) (1:3000, Biorad-Serotec MCA739G) for the assessment of demyelination (in combination with LFB staining), 2) anti-APP (1:10000, Merck-Millipore MAB348), anti-synaptophysin (1:100, Monosan Monx 10,779) and anti-neurofilament protein (1:100, Dako M0762) for axonal damage, 3) anti-GFAP (1:100 Dako M0761) for astrocytosis and 4) anti-HLA-DR (1:400, Dako M0775) for microgliosis. All antibodies but the anti-PLP required an antigen retrieval step (15 min microwave incubation after boiling) in Sodium Citrate buffer pH 6.0 (APP, GFAP, HLA-DR and neurofilament) or Tris/EDTA buffer pH 9.0 (synaptophysin).

For LFB staining, slides were immersed overnight in LFB solution (final concentration, 0.1% solvent blue and 0.5% acetic acid in 95% alcohol) at 60 °C. After immersion in 95% alcohol and washing, sections were immersed 5 s in 0.05% lithium carbonate and rewashed. The latter steps were repeated until suitable gray matter discoloration. The obtained sections were then processed for PAS staining through immersion in periodic acid for 10 min and, after a washing step in deionized water, incubation in dark condition with Schiff’s reagent for 15 min. Subsequently, slides were washed, incubated with Mayer’s hematoxylin for 1 min, immersed in warm water and rewashed.

Transmission to bank voles

Brain tissue from the p-CJDMM1 index case (case #1 described below) and from 4 control cases without plaques (three sCJDMM1 and one sCJDMV1) were homogenized at 10% (w/v) concentration in phosphate buffered saline (PBS) and stored at −80 °C. Two genetic lines of bank voles, Bv109M and Bv109I carrying methionine or isoleucine homozygosity at PRNP codon 109, were injected by the intracerebral route (20 μl) into the left cerebral hemisphere under ketamine anesthesia. Beginning one month after inoculation, voles were examined twice per week until the appearance of neurological signs, and evaluated daily thereafter. The animals were sacrificed with carbon dioxide when they reached the terminal stage of the disease. Survival time was calculated as the interval between inoculation and sacrifice, attack rate as the number of animals developing disease with respect to the total number of inoculated animals [18].

The lesion profile was based on the severity of vacuolation, with a score from 0 to 5, in nine grey-matter brain areas on hematoxylin and eosin-stained sections, as previously described [7]. Vacuolation scores derived from at least 6 individual voles per group and were reported as mean ± standard error (SEM).

Preparation of human and bank vole brain total homogenates (THs) and PK digestion

Tissues from occipital cortex grey matter (human) and vole brain (50 mg) were homogenized (10% w/v) in lysis buffer (100 mM NaCl, 10 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate, 100 mM Tris) at pH 6.9 (as reported in [19]) and digested with proteinase K (PK) (Roche Diagnostics) at a final concentration of 8 U/ml for 1 h at 37 °C. Digestion was blocked with phenylmethylsulfonyl fluoride (PMSF, final concentration 3.6 mM), then samples were boiled in sample buffer (final concentration: 3% SDS, 4% β-mercaptoethanol, 10% glycerol, 2 mM EDTA, 62.5 mM Tris) for 6 min at 100 °C.

Preparation of white matter total homogenates and PK digestion

Frontal and parietal cortical white matter was obtained from case #1 (p-CJDMM1), and one np-CJDMM1. PrP^Sc was purified from 350 mg of white matter, following a previously published protocol [29] and re-suspended in 200 μl of lysis buffer at pH 6.9. PK digestion was carried out at a final concentration of 4 U/ml for 1 h at 37 °C.

PrP deglycosylation

N-Linked glycans were removed by using a peptide-N-glycosidase F kit (New England Biolabs) according to the manufacturer’s instructions.

PK titration curves

Grey matter tissues were homogenized (10% w/v) in lysis buffer at pH 8. Total protein concentration was measured by means of a standard colorimetric method based on bicinchoninic acid (Pierce) and then adjusted to a final value of 4200 μg/ml. Samples were digested using serial dilutions of PK activity ranging from 2 to 256 U/ml, for 1 h at 37 °C. Digested samples were treated as previously described.

Thermo-solubilization assay (TSA)

TSA was performed as described [6]. Briefly, grey matter THs (10% w/v in lysis buffer at pH 6.9) were digested with 8 U/ml PK for 1 h at 37 °C with mild shaking (300 rpm). PK digestion was inactivated with PMSF (final concentration, 3.6 mM). Aliquots were mixed with an equal volume of loading buffer (final concentrations, 1.5% SDS, 2% β-mercaptoethanol, 5% glycerol, 1 mM EDTA, 31.2 mM Tris) and heated to temperatures ranging from 25 °C to 95 °C (ΔT = 10 °C) for 6 min with shaking in a thermomixer at 1000 rpm before loading.

Western blot

Samples were run in a 7 or 15 cm long separating gel and transferred to Immobilon-P membranes (Millipore). After blocking in 10% non-fat milk in Tween-Tris-buffered saline, membranes were probed overnight with the monoclonal antibody 3F4 with epitope at PrP residues 108–111 at 1:30000 working dilution (human samples). Immunoblots from bank voles samples were incubated overnight at 4 °C with the monoclonal antibody 9A2 (1:8000, PrP residues 99–101) [26] instead of 3F4. In addition, all immunoblots were probed with the C-terminal antibody SAF60 (1:2000, PrP residues 157–161) [20] in order to detect the CTF13. After four washings in Tween-Tris-buffered saline, membranes were incubated for 1 h at room temperature with an anti-mouse secondary antibody conjugated to horseradish peroxidase (GE Healthcare; working dilution, 1:4000) and washed again four times in Tween-Tris-buffered saline. The immunoreactive signal was visualized by enhanced chemiluminescence (Immobilon Western, Millipore) on an LAS 3000 camera (Fujifilm).

Quantitative analysis of protein signal

Densitometric analysis was performed using the software AIDA (Image Data Analyzer v.4.15, Raytest GmbH). For PK titration, a semi-logarithmic curve was obtained by plotting the percentage of protein remaining after digestion (with respect to the sample digested with 2 U/ml) against the corresponding PK concentration. The ED₅₀ (i.e. the PK concentration needed to digest 50% of PrP^Sc) for each sample was calculated by means of the equation of the straight line that best fitted the linear portion of the curve (r² ≥ 0.95). For TSA, the percentage of protein solubilized after heating treatment (with respect to the sample treated at 95 °C) was plotted against the corresponding heating temperature. The T₅₀ (i.e. the temperature needed to solubilize 50% of PrP^Sc) for each sample was calculated from the equation describing the sigmoidal curve that best fitted the data (r² ≥ 0.95).

Statistical analyses

All statistical analyses were performed with SigmaPlot 12.5 (Systat Software Inc.). Depending on the data distribution, Student’s t test or Mann-Whitney test were used to detect differences between two groups, while one-way analysis of variance (ANOVA), followed by Dunn’s or Holm-Sidak post hoc tests, was applied for three or more groups comparisons. P value <0.05 was considered statistically significant.

Clinical findings and diagnostic investigations

Genetic analysis and PrP^Sc typing

All 5 p-CJDMM1 cases carried MM at PRNP codon 129; in addition case #4 carried the E200K mutation while sequencing of the PRNP coding region excluded pathogenic mutations in the other four. PrP^Sc typing showed type 1 in all cases with the co-occurrence of type 2 in case #5 (Fig. 1). Notably, there were no significant differences in PrP^Sc electrophoretic mobility between p-CJDMM1 and np-CJDMM1 cases in both gray (Fig. 1) and white matter (Additional file 1: Figure S1A).

Neuropathology

The histopathological features largely overlapped with those previously reported for the np-CJDMM1 histotype. Specifically, the pathological changes mainly involved the cerebral cortex, striatum, thalamus and cerebellum, whereas the hippocampus and brainstem were relatively spared (Additional file 2: Table S1). Consistently with the relatively long disease duration, the histopathological changes were, in most cases, rather severe. Indeed, marked atrophy and status spongiosus were the main findings in the most affected areas, whereas typical spongiform change with microvacuolation was best seen in less affected structures (i.e. the hippocampus) (Fig. 2 a,b). As the only exception, the case with the shortest duration (case #5) showed typical spongiform change and, overall, moderate rather than severe histopathological lesions in most affected gray matter areas (Additional file 2: Table S1). Immunohistochemical analysis of PrP^Sc showed the typical synaptic deposition pattern in the molecular layer of cerebellum and in the cerebral cortex in all cases (Fig. 2c).

Finally, the sections stained with the anti-APP and anti-synaptophysin antibodies revealed that the deposits of these proteins in the damaged white matter of p-CJDMM1 were, occasionally, co-localizing with PrP-amyloid plaques (Fig. 2 e,f).

Biochemical comparison between p-CJDMM1 and np-CJDMM1

Besides the presence of a typical type 1 fragment, the overlap of biochemical PrP^Sc properties between p-CJDMM1 and np-CJDMM1 extended to the CTF13, as revealed by antibody SAF60. As the only exception, in case #5, as expected given the mixed types 1 + 2 molecular phenotype [20], the ratio between PrP27–30 and CTF13 differed from that of the other cases (Additional file 3: Figure S2; Additional file 4: Table S3). Analysis of PrP^Sc glycoform ratio demonstrated a slightly higher percentage of diglycosylated PrP^Sc in p-CJDMM1 than in np-CJDMM1, whilst confirming a similar glyco-pattern (Additional file 5: Table S3).

As in grey matter, the analysis of CTF13 in white matter did not show any significant difference between p-CJDMM1 and np-CJDMM1 cases (Additional file 1: Figure S1B).

PK titration curves, performed on gray matter homogenates at pH 8 in both sporadic and genetic (E200K) cases, did not exhibit consistent differences between the two groups. P-CJDMM1 and np-CJDMM1 groups showed an ED₅₀ (expressed as mean ± standard deviation) of 7.87 ± 1.09 and 9.14 ± 2.92, respectively. No statistically significant differences were observed (Fig. 3 a,c). Case #5 was excluded from this analysis, since MM2-PrP^Sc is associated with a higher PK-resistance than MM1-PrP^Sc [29].

TSA, performed on THs at pH 6.9, also revealed comparable values among the analyzed samples. The calculated T₅₀ (expressed as mean ± standard deviation) was 79.68 ± 0.72 in p-CJDMM1 and 80.38 ± 2.90 in np-CJDMM1 (Fig. 3 b,d).

Transmission to bank voles

PrP^Sc extracted from the brains of infected bank voles was subjected to Western blot analysis to detect possible differences induced by the two types of inocula. As for PrP^Sc in the CJD brains, PrP^Sc fragments were indistinguishable between bank voles inoculated with p-CJDMM1 or np-CJDMM(V)1 (Additional file 6: Figure S3A). At variance with the human brain, vole PrP^Sc was characterized by a predominance of the diglycosylated form, as previously reported [18]; however, likewise in CJD inoculated samples, PrP^Sc glycoform ratio in voles did not show any statistically significant difference related to the inoculum (p-CJDMM1 or np-CJDMM(V)1) (Additional file 6: Figure S3B). Similarly, a comparable amount of CTF13 [26] was detectable, after sample deglycosylation, in voles infected with the two inocula (Additional file 7: Figure S4).

Neuropathology in bank voles

The lesion profiles of case #1 and inoculated controls showed an identical distribution of spongiform change (Fig. 4). In Bv109M, in both first and second passage, spongiform change was more prominent in superior colliculi, thalamus, hippocampus and retrosplenial and cingulate cortices. The other areas presented none or few scattered vacuoles. The Bv109I line displayed a slightly more pronounced spongiform change in the medulla, hypothalamus and septum, which was also maintained upon the two passages. Immunohistochemistry revealed a synaptic pattern of PrP deposition, while no plaque-like deposits were seen in the affected animals of both lines.