Minimal change prion retinopathy: Morphometric comparison of retinal and brain prion deposits in Creutzfeldt-Jakob disease

Sporadic Creutzfeldt-Jakob disease (sCJD) is the most commonly diagnosed human prion disease caused by the abnormal misfolding of the ‘cellular’ prion protein (PrPC) into the transmissible ‘scrapie-type’ prion form (PrPSc). Neuropathologic evaluation of brains with sCJD reveals abnormal PrPSc deposits primarily in grey matter structures, often associated with micro-vacuolar spongiform changes in neuropil, neuronal loss, and gliosis. Abnormal PrPSc deposits have also been reported in the retina of patients with sCJD, but few studies have characterized the morphology of these retinal PrPSc deposits or evaluated for any retinal neurodegenerative changes. We performed histopathologic and morphometric analyses of retinal and brain prion deposits in 14 patients with sCJD. Interestingly, we discovered that the morphology of retinal PrPSc deposits generally differs from that of brain PrPSc deposits in terms of size and shape. We found that retinal PrPSc deposits consistently localize to the outer plexiform layer of the retina. Additionally, we observed that the retinal PrPSc deposits are not associated with the spongiform change, neuronal loss, and gliosis often seen in the brain. The stereotypic morphology and location of PrPSc deposits in sCJD retinas may help guide the use of ocular imaging devices in the detection of these deposits for a clinical diagnosis.


Introduction
Prion diseases are a heterogeneous class of rare fatal neurodegenerative diseases caused by the abnormal misfolding of the 'cellular' prion protein (PrP C ) into the transmissible 'scrapie-type' prion form (PrP Sc ) (Collinge, 2016;DeArmond and Prusiner, 1995;Halliday et al., 2014;Prusiner, 1991Prusiner, , 1998Prusiner et al., 1990;Weissmann, 2004). The PrP C peptide is ubiquitously expressed throughout the body and highly expressed within neurons via a single exon in the PRNP gene (Bendheim et al., 1992;DeArmond and Prusiner, 1995). Human prion diseases are generally classified into three groups: sporadic (85-90% idiopathic cases), genetic (10-15% inherited cases), and acquired (1-3% infectious cases) (Takada and Geschwind, 2013). Of the human prion diseases, sporadic Creutzfeldt-Jakob disease (sCJD) is the most commonly diagnosed, with a global incidence of 1-1.5 cases per 1 million population per year (Uttley et al., 2020). Although there are currently six major subtypes of sCJD with varying clinical and pathologic phenotypes (Parchi et al., 2009), patients typically present with rapidly progressive dementia associated with the accumulation of neurologic deficits, which culminate in death. Neuropathologic evaluations of brains with sCJD reveal abnormal PrP Sc deposits in grey matter structures often associated with micro-vacuolar spongiform changes in the neuropil, neuronal loss, and gliosis.
Electroretinographic b-wave abnormalities have been reported in patients with CJD (de Seze et al., 1998;Ishikawa et al., 2009). Histopathologic changes in CJD retinas have been variable. Several case reports of post-mortem CJD eyes described gliosis in the optic nerve and nerve fiber layer (NFL) and retinal ganglion cell (RGC) loss (Lesser et al., 1979;Roth et al., 1979;Sato et al., 1992;Tarkkanen and Haltia, 1980;Tsutsui et al., 1986); and two other studies also described spongiform changes in inner retinal structures (de Seze et al., 1998;Tsutsui et al., 1986). However, other CJD studies reported no structural abnormalities and no spongiform changes in the retina (Head et al., 2003;Roth et al., 1979;Tarkkanen and Haltia, 1980). With the advent of antibodies that detect PrP Sc in tissue, PrP Sc deposition has also been reported in retinas of patients with CJD (Head et al., 2003(Head et al., , 2005Orru et al., 2018;Takao et al., 2018). Head and colleagues examined the eyes of two sCJD and two variant CJD (vCJD) autopsy cases and observed PrP Sc deposition in the inner plexiform (IPL) and outer plexiform (OPL) layers of the retina (Head et al., 2003(Head et al., , 2005. Takao and colleagues reported similar findings in eyes from 16 CJD cases, which included sporadic, familial, and iatrogenic CJD forms (Takao et al., 2018). Orru and colleagues also observed PrP Sc deposition in the OPL and to a lesser extent in the IPL in the eyes of 11 patients with sCJD (Orru et al., 2018). Together, these reports independently corroborate the deposition of PrP Sc in the eyes of sCJD patients, most notably confined to the synaptic retinal laminae.
The aim of this present study is to expand upon previous research to provide a detailed characterization of the retinal pathology caused by PrP Sc deposition in the eyes of patients with sCJD. We examined post-mortem eyes obtained from 14 cases of clinically and neuropathologically-documented sCJD, including the aforementioned 11 sCJD cases examined by Orru and colleagues (Orru et al., 2018). We compared the morphology of PrP Sc deposits in the retina to those in the brain. We histologically evaluated the retina for evidence of spongiform change and neurodegeneration. Ultimately, we found intriguing differences in prion aggregate morphology, degree of spongiotic change, and neuronal loss in the eyes of decedents with sCJD compared to the brain.
From July 2015 to July 2017, eyes and brains were collected from 14 autopsies of clinically suspected sCJD at the UCSF Memory and Aging Center, 11 of which were previously described (Orru et al., 2018). Subjects consisted of four males and ten females with a mean age of 63 ± SD 8.96 (median 60, range 51-80) years old. The mean clinical disease course from symptom onset to death was 10.5 ± SD 8.38 (median 8, range 1.5-27) months, and 5 out of the 14 patients reported visual symptoms during the disease course (Table  1). Prion disease was confirmed in all cases by Western blot and immunohistochemistry (IHC) for PrP Sc (Fig. 1B). Prion-related protein gene (PRNP) sequencing analysis showed no mutations in 13 cases, and one case was not sequenced. All three PRNP gene codon 129 polymorphisms were represented in the 13 sequenced cases: MM (n = 6), MV (n = 5), VV (n = 2).
Control cases consisted of eyes and brains from autopsies of five males and one female with a mean age of 70.3 ± SD 13.84 years (median 67.5, range 51-90) with no histories of retinal diseases or ocular surgeries except cataract surgeries (n = 3). The retinal histopathology in 4 of 6 control cases showed common age-related changes such as peripheral cystoid degeneration and drusen. The retinal histopathology in the remaining 2 control cases (3 and 6) was unremarkable. No PrP Sc retinal deposits were found in any control retinas (Fig. 1C).
Two of the control patients had cognitive symptoms: control brain 4 showed "high" Alzheimer's disease neuropathology by 2018 NIA-AA framework criteria (A2-Thal, B3-Braak, C2-CERAD); control brain 6 was sent to the National Prion Disease Surveillance Center and was negative for sCJD neuropathology. The remaining 4 control brains had no neurodegenerative disease neuropathology (Table 1). No prion deposits were found in any control brains by PrP Sc IHC (Fig. 1A).
Formalin-fixed, paraffin-embedded brain and eye sections from sCJD and control cases were evaluated for further histologic and immunohistochemical PrP Sc analysis.
Ocular tissues in fourteen patients with sCJD and six controls and brain tissues across the cerebral cortex, caudate, putamen, thalamus, and cerebellum in a case of sCJD (MV2; patient n° 4 in Table 1) were used for IHC staining and PrP Sc quantification in this study. The eye tissues were formalin-fixed and immersed in 98% formic acid for 1 h. They were postfixed in formalin prior to paraffin embedding and sectioning. Sections of 4 μm thickness were placed onto positively charged silanized glass slides and deparaffinized. After being immersed in 96% formic acid for 5 min, the slides were washed with distilled water for 5 min and treated with 5 μg/ml of proteinase K for 7 min to remove natively expressed PrP C . Following another 5-min wash with distilled water, they were placed in citrate buffer and heated in a pressure cooker for 20 min for antigen retrieval. After being cooled and rinsed in distilled water the sections were incubated with anti-PrP Mab 12F10 (Cayman Chemical; 1:200 dilution) for 45 min, followed by anti-mouse IgG conjugated to biotin (Jackson Immunolabs; 1:250) and streptavidin-HRP (Jackson Immunolabs; 1:2000) for 30 min each. Labeling with diaminobenzidine (DAB) reagent (Thermo Fisher Scientific) was performed for brown chromogen visualization, and the sections were counterstained lightly with hematoxylin. PrP Sc deposits within the eyes of sCJD patients were then evaluated and compared with normal controls and coarse/perivacuolar deposits within the brain.
Following preparation of autopsy brain and retina slides with prion IHC, 202 retinal deposits across 14 sCJD cases and 195 brain PrP Sc deposits across the cerebral cortex and thalamus in a case of sCJD were photographed, and the greatest diameter of each PrP Sc deposit was measured (Olympus CellSens Imaging Software, Waltham, MA).
The difference in the diameters of the PrP Sc depositions in the brain and retina was statistically tested using the Mann-Whitney U test in SPSS. The difference in the mean diameters between brain and retina PrP Sc deposits was statistically tested using Welch's unpaired t-test. Standard deviations of the two groups were tested with the Brown-Forsythe test for their difference. P values of less than 0.05 were considered statistically significant.
In the 14 sCJD cases (Table 1) granular PrP Sc deposition was found by IHC in several synaptic laminae of the retina ( Fig. 1D and E). Similar to prior studies, PrP Sc deposition was strongest in the OPL (14/14) ( Fig. 1D and E). Fainter patchier PrP Sc staining was also occasionally observed in the IPL in some cases (Fig. 1D and E). No consistent relationship between PRNP genotype and IHC staining was seen, although this is limited by our small sample size. No PrP Sc deposition was seen in other retinal layers including retinal pigment epithelia (RPE), outer nuclear layer (ONL; rod and cone photoreceptor nuclei), inner nuclear layer (INL; retinal interneurons and Muller glia), ganglion cell layer (GCL), and nerve fiber layer (NFL). PrP Sc deposits were not detected elsewhere in the eye including sclera, cornea, lens, choroid, and optic nerve. PrP Sc deposition was completely absent in retinas of all 6 control cases (Fig. 1C).
Within the OPL, PrP Sc strongly accumulated in discrete, ovoid deposits spaced along the length of the OPL, reminiscent in appearance to "beads-on-a-string" (Fig. 1D and E). These PrP Sc deposits in the OPL also arose at fixed distances between the ONL and INL along the length of the lamina ( Fig. 1D and E). Measurement of 202 ovoid PrP Sc retinal deposits from the OPL showed an average greatest diameter of 4.94 μm with little variation (SD: ± 0.47 μm, Fig. 1H). In contrast to the stereotypic appearance of PrP Sc deposits in the OPL of the retina, PrP Sc deposits in the brain had highly irregular morphologies and widespread deposition (Fig. 1B). PrP Sc deposits in the brain also had an average greatest diameter of 32.45 μm with large variation (SD: ± 23.55 μm, Fig. 1G), which was significantly larger than that of retinal PrP Sc deposits in the OPL (p < 0.001, Mann-Whitney U test; p < 0.001, Welch's unpaired t-test, Fig. 1I). The difference of standard deviations between the two groups was statistically significant (p < 0.001, Brown-Forsythe test).
As expected, extensive spongiform change and neuronal cell loss were evident in H&Estained slides prepared from the sCJD decedent autopsy brain (Fig. 1B). In contrast, histologic findings of severe neurodegeneration were not observed in sCJD retinas. The nuclear lamina of the retina (RPE, ONL, INL, and GCL) were all intact and of similar thickness (a proxy for retinal neuronaldensity) in sCJD retinas compared to the normal control retinas (Fig. 1C and D). Similarly, the synaptic/axonal/dendritic lamina of the retina (NFL, IPL, and OPL) also appeared intact and of similar thickness in all sCJD retinas and normal control retinas, despite strong PrP Sc deposition observed in the OPL (Fig. 1D, E, 1F).
Our findings confirm prior reports of retinal PrP Sc deposition in sCJD cases (Head et al., 2003(Head et al., , 2005Orru et al., 2018;Takao et al., 2018) as well as in mouse models (Striebel et al., 2021). As with these studies, we find the strongest PrP Sc deposition in the OPL where discrete ovoid deposits of PrP Sc result in a "beads-on-a-string" appearance along the horizontal length of the lamina. This pattern of PrP Sc immunoreactivity is similar to the strong ovoid OPL PrP Sc immunohistochemical labeling observed previously (Head et al., 2003;Takao et al., 2018). We did not observe the "fine-dot" PrP Sc staining in the INL, ONL, GCL, and NFL specifically reported by Takao and colleagues in some of their CJD cases (Takao et al., 2018). Possible reasons for the differences between our study and this prior report could arise from differences in the antibodies used to detect PrP Sc , tissue preparation conditions (possible autoclaving of tissue in Takao and colleague's study vs proteinase K treatment in our study), or sCJD subtypes in the different patient cohorts. Their study included a higher proportion of MM1 sCJD cases (9 out of 11) than in our patient group (2 out of 14), and their cohorts also included 3 familial and 1 iatrogenic CJD case. Furthermore, we found that the greatest diameters of the PrP Sc deposits in the OPL of the retina were significantly smaller and with reduced variability in size compared to deposits in the brain. The greatest diameters of PrP Sc deposits were also previously measured in the brains of 11 patients with vCJD by Armstrong and colleagues (Armstrong et al., 2005). In their study, the mean greatest diameters of diffuse and florid-type PrP Sc deposits ranged from 7.5 μm to 42.3 μm across 12 different brain regions and also showed a positively skewed distribution similar to the sizes and distribution observed with the PrP Sc deposit measurements in our sCJD brain autopsy tissues. These findings further support that brain PrP Sc deposits differ markedly in size and morphology compared to retinal deposits.
Our current study focused on PrP Sc deposits in the OPL because these are consistently the strongest sites of deposition in the sCJD retina. In contrast to the highly variable morphology of PrP Sc deposits in the sCJD brain, there is a much more uniform and stereotypic pattern of PrP Sc deposition in the OPL of the retina. Although further ultrastructural investigations of the structures involved in the OPL PrP Sc deposits are warranted, we hypothesize that PrP Sc is accumulating at the synaptic ribbon where the presynaptic terminals of rod and cone photoreceptors from the ONL connect with the postsynaptic dendrites of bipolar and amacrine interneurons from the INL. Evidence of this PrP localization has recently been reported by Striebel et al. (2021), who identified PrP Sc accumulation in ribbon synapses and cilia of photoreceptors in mouse models of prion disease. The presence of PrP Sc deposits at retinal synaptic junctions could alter or disrupt this retinal circuitry, result in photoreceptor degeneration, and perhaps contribute to visual symptoms reported in sCJD.
Counterintuitively, the extensive spongiform change, neuronal loss, and gliosis seen in association with PrP Sc deposition in sCJD brain tissues were not found in association with PrP Sc deposition in the retina. The absence of retinal degeneration in sCJD could potentially explain the clinical observation that patients experience symptomatic visual disturbances rather than frank anopsia. Although the etiology is unknown and ultrastructural investigations are needed, this phenomenon of minimal change prion retinopathy observed in our sCJD cohort could be attributed to either a delayed time course of PrP Sc deposition relative to that of the brain or may be related to the genotype. It is worth noting that brain PrP Sc deposition is not always associated with vacuolation in some prion diseases, such as in fatal familial insomnia and sporadic fatal insomnia (Abu-Rumeileh et al., 2018;Halliday et al., 2014).
Our observation of minimal change prion retinopathy lends credence to a previous investigation by Head and colleagues who reported no evidence of spongiform change in the retina from two vCJD cases and one sCJD case (Head et al., 2003). Head and colleagues did, however, observe increased GFAP expression in the nerve fiber layer (NFL) in the retina of one of their CJD cases (Head et al., 2003). Additionally, NFL thinning, reduced RGC numbers and optic nerve gliosis and atrophy have been reported in other sCJD patient case reports (de Seze et al., 1998;Lesser et al., 1979;Roth et al., 1979;Sato et al., 1992;Tarkkanen and Haltia, 1980;Tsutsui et al., 1986). More recently, a small yet statistically significant thinning of the NFL was observed in the retinas of sCJD patients relative to normal controls through the use of premortem optical coherence tomography by Orru and colleagues (Orru et al., 2018). Altogether, these findings suggest that neurodegenerative sequelae observed in the inner retina structures (GCL and NFL) and the optic nerve during the course of sCJD, may be either a direct consequence of OPL PrP Sc deposition, an indirect consequence of other sCJD related ocular stressors (e.g. accumulation of redox-active iron), or a combination of such contributing factors (Chaudhary et al., 2021).
Although retinal lamination was normal, and the thickness of retinal lamina appeared similar in our sCJD cases compared to controls, we cannot exclude gliosis or microglial changes as we did not have sufficient tissue sections to stain for GFAP or other markers. Another limitation of our study is that we did not quantify retinal neuron numbers in our samples and therefore, cannot exclude neuronal loss in our sCJD cases.
Minimal change prion retinopathy may be a pathology limited to human prion disease. Several animal studies found photoreceptor degeneration, but not RGC degeneration, in scrapie-infected mice (Kercher et al., 2007;West Greenlee et al., 2016), and spongiform retinal degeneration in transmissible mink encephalopathy-affected cattle (Smith et al., 2009). The study of Kercher and colleagues also found activated microglia associated with neurodegeneration in the mice retina (Kercher et al., 2007). Future studies to quantify photoreceptor numbers in human sCJD retina may better identify signs of neurodegeneration. Also, investigating the status of microglia and Muller glia in sCJD retina may provide insight into the cellular responses to PrP Sc deposition in the eye.
One possible explanation for the resilience of human retina compared to the brain to PrP Sc neuropathology is a recent report that found natural abundance of the β-cleavage product of PrP C , N2, in human ocular tissue in contrast to the predominant α cleavage of PrP C in the brain producing N1 and C1 products (Chaudhary-Suman, Association for Research Vision Ophthalmology. 2021 annual meeting, poster). Several studies suggest that the N2 fragment has a neuroprotective, antioxidant effect (Haigh et al., 2015;Watt et al., 2005). Given the oxidative stress and inflammatory milieu previously observed in prion diseases (Brazier et al., 2006;Guentchev et al., 2000;Lee et al., 1999;Singh et al., 2020), the antioxidant effect of the N2 byproduct may play a role in inhibiting significant retinal degeneration and explain why retinal PrP Sc deposits are not associated with significant spongiform change.
Another possibility is that the retina may be affected later than the brain in sCJD and our observations might be attributed to early retinopathy -e.g. overt neurodegenerative retinopathy might only be observed in atypical sCJD cases of prolonged duration.
Overall, our study corroborates prior reports that PrP Sc deposits in sCJD retinas. In particular, PrP Sc preferentially accumulates in the OPL of the retina where they adopt a stereotypic "beads-on-a-string" pattern in the form of ovoid aggregates at periodic intervals along the horizontal axis of the lamina. Regardless of spongiform change, neuronal loss, and gliosis, this stereotypic pattern of PrP Sc deposition in the OPL could potentially help guide the development of diagnostic strategies for sCJD using retinal imaging modalities, such as adaptive optics (Qin et al., 2020). Similar to how the detection of retinal beta-amyloid in Alzheimer's disease has been proposed as a diagnostic biomarker, viewing the pattern of PrP Sc deposition in the OPL of the retina could allow for non-invasive testing and earlier diagnosis of clinically suspected sCJD cases (Koronyo et al., 2017).    Patient demographic data and clinical features of pathology-confirmed sporadic creutzfeldt jakob cases a and controls. This table reports 14 cases of patients diagnosed with brain autopsy-confirmed sCJD and 6 control patients (no clinical or brain autopsy evidence of sCJD). The first 3 sCJD patients represent the new cases studied and the remaining 11 cases were previously described in Orru et al. (2018). Patients with sCJD comprised 10 females and 4 males, with the age of onset ranging from 51 to 80 years of age (Mean: 63, Median: 60, SD: 8.96) and with the disease duration ranging from 1.5 to 27 months (Mean: 10.5, Median: 8, SD: 8.38). The PRNP genotypes of these patients were analyzed and classified into subtypes