Presence of tau pathology within foetal neural allografts in patients with Huntington’s and Parkinson’s disease

Foetal tissue grafts in patients with Huntington’s or Parkinson’s disease have been shown to develop Huntington- or Parkinson-like pathology. Cisbani et al. reveal that such grafts may also contain tau pathology, further implicating tau in these disorders, and suggesting that tau may spread via transynaptic propagation from the host.

Abstract

Cell replacement has been explored as a therapeutic strategy to repair the brain in patients with Huntington’s and Parkinson’s disease. Post-mortem evaluations of healthy grafted tissue in such cases have revealed the development of Huntington- or Parkinson-like pathology including mutant huntingtin aggregates and Lewy bodies. An outstanding question remains if tau pathology can also be seen in patients with Huntington’s and Parkinson’s disease who had received foetal neural allografts. This was addressed by immunohistochemical/immunofluorescent stainings performed on grafted tissue of two Huntington’s disease patients, who came to autopsy 9 and 12 years post-transplantation, and two patients with Parkinson’s disease who came to autopsy 18 months and 16 years post-transplantation. We show that grafts also contain tau pathology in both types of transplanted patients. In two patients with Huntington’s disease, the grafted tissue showed the presence of hyperphosphorylated tau [both AT8 (phospho-tau Ser202 and Thr205) and CP13 (pSer202) immunohistochemical stainings] pathological inclusions, neurofibrillary tangles and neuropil threads. In patients with Parkinson’s disease, the grafted tissue was characterized by hyperphosphorylated tau (AT8; immunofluorescent staining) pathological inclusions, neurofibrillary tangles and neuropil threads but only in the patient who came to autopsy 16 years post-transplantation. Abundant tau-related pathology was observed in the cortex and striatum of all cases studied. While the striatum of the grafted Huntington’s disease patient revealed an equal amount of 3-repeat and 4-repeat isoforms of tau, the grafted tissue showed elevated 4-repeat isoforms by western blot. This suggests that transplants may have acquired tau pathology from the host brain, although another possibility is that this was due to acceleration of ageing. This finding not only adds to the recent reports that tau pathology is a feature of these neurodegenerative diseases, but also that tau pathology can manifest in healthy neural tissue transplanted into the brains of patients with two distinct neurodegenerative disorders.

Introduction

The discovery of Lewy body pathology in allografted ventral mesencephalic tissue in patients with Parkinson’s disease (Kordower et al., 2008, 2017; Li et al., 2008) led to the concept that neurodegenerative diseases may develop through the spread of pathological proteins into otherwise healthy cells (Kordower et al., 2008; Li et al., 2008). In a recent post-mortem report, we also demonstrated the heterotopic presence of mutant huntingtin protein (mHtt) aggregates within foetal striatal allografts in the brains of patients with Huntington’s disease who received transplants over a decade earlier (Cicchetti et al., 2014). This was the first demonstration of mHtt ‘spread’ in humans, a finding which suggested that pathogenic proteins could seed pathology in a non-cell autonomous manner, even in a genetic disorder. These clinical pathological observations have now been further supported in several in vitro and in vivo studies for both Huntington’s disease (Yang et al., 2002; Ren et al., 2009; Costanzo et al., 2013; Pecho-Vrieseling et al., 2014; Babcock and Ganetzky, 2015; Pearce et al., 2015; Jeon et al., 2016) and Parkinson’s disease (Desplats et al., 2009; Brundin et al., 2010; Hansen et al., 2011; Angot et al., 2012; Luk et al., 2012a, b; Tyson et al., 2017), building on earlier reports of a similar phenomenon in models of Alzheimer’s disease (Clavaguera et al., 2009; de Calignon et al., 2012; Liu et al., 2012; Iba et al., 2013; Walker and Jucker, 2015; Sanders et al., 2016; Espuny-Camacho et al., 2017).

The above findings raise the question of whether other pathological proteins participate in neurodegenerative processes. It has recently been reported that tau neuronal inclusions and neuropil threads, as seen in Alzheimer’s disease, could be found in various brain structures in Huntington’s disease (Fernandez-Nogales et al., 2014; Gratuze et al., 2015, 2016; Vuono et al., 2015) as well as in Parkinson’s disease (Duka et al., 2013; Sengupta et al., 2015). We have therefore investigated whether tau pathology can also be found in foetal tissue transplants placed in the brains of patients with Huntington’s and Parkinson’s diseases.

Patients and methods

This report focuses on the histological evaluation of four brains from both Huntington’s disease and Parkinson’s disease transplanted patients (see Table 1 for details). The brains from the transplanted Huntington’s disease cases came from our initial seven Huntington’s disease patient cohort in the open label study on foetal striatal allografting conducted at the University of South Florida (Patients 1 and 7) (Hauser et al., 2002). These two brains were chosen because they had the best histological evidence of long-term graft survival from our cohort. Patient 1 had 42 CAG repeat (Hauser et al., 2002; Cicchetti et al., 2009; Cisbani et al., 2013), while Patient 7 had 53 CAG repeats (Hauser et al., 2002; Cicchetti et al., 2009; Cisbani et al., 2013). These two females were transplanted at 58 and 28 years of age, respectively, and died 9 and 12 years post-transplantation of causes unrelated to surgery. At autopsy, they were classified grade 3 and grade 4 according to the Vonsattel post-mortem rating scale (Vonsattel et al., 1985).

Table 1.

Demographic details of the cases involved in this study

Case	Sex	Clinical diagnosis	Age at autopsy	Time post-grafting	Post-mortem interval (h)
Transplanted cases
HD	F	HD (42 CAG/Grade 3)	67	9 years	5
HD	F	HD (53 CAG/Grade 4)	40	12 years	2
PD	F	Sporadic PD	74	16 years	10
PD	M	Sporadic PD	59	18 months	4
Controls
CTRL	M	Bronchopneumonia	70	NA	5
CTRL	M	Myocardial infarction	41	NA	13
CTRL	M	Pulmonary complications	12	NA	14

CTRL = control; HD = Huntington’s disease; NA = not available; PD = Parkinson’s disease.

All relevant details of the two brains from the transplanted Parkinson’s disease cases are provided in previously published work (Kordower et al., 1995, 1996; Olanow et al., 2003). In brief, the Parkinson’s disease case who came to autopsy 18 months post-transplantation was a 59-year-old male while the second case who came to autopsy 16 years post-transplantation was a 58-year-old female (Kordower et al., 1995, 1996). Both were diagnosed clinically with idiopathic Parkinson’s disease prior to their enrolment in the Mt. Sinai/Rush/Tampa Neural grafting programme and this diagnosis was pathologically confirmed by a board-certified neuropathologist (Olanow et al., 2003). Unfortunately, clinical data relating to Braak staging, Lewy body types or other genetically-related factors for dementia/tauopathies [THAL amyloid (Murray et al., 2014) or APOE (Gao et al., 2011)] were not available for these cases.

The age-matched control brain for the Huntington’s disease Patient 1 and the two Parkinson’s disease cases was a 70-year-old female previously described in Cicchetti et al. (2009). The age-matched control brain used for the Huntington’s disease Patient 7 was a 41-year-old male who died of myocardial infarction. The control brain from a non-Huntington’s disease 12-year-old boy, who died from pulmonary complications of multiple trauma (Cisbani et al., 2013), was used to match the developmental stage of the grafted foetal tissue of both Huntington’s disease and Parkinson’s disease transplanted cases with long-term survival (see Table 1 for case details). All post-mortem analyses of human brain tissue were approved by the Comité d'éthique de la recherche du CHU de Québec (#A13-02-1137).

Donor tissue preparation and transplantation

For the Huntington’s disease cases, methods for tissue preparation, transplantation and immunosuppression, as well as clinical and radiological evaluation have been described previously (Freeman et al., 1995, 2000; Hauser et al., 2002; Cicchetti et al., 2009). Briefly, all patients received foetal tissue transplants from 5 to 8 striatal primordia per site. Solid tissue transplants measuring 0.5 to 1 mm³ were derived from the far lateral portion of the lateral ganglionic ventricular eminence (FLVE) to optimize the percentage of tissue of striatal origin (Freeman et al., 1995). More specifically, Patient 1 received one striatal anlage in the left and right caudate and four in the left and right putamen. Patient 7 received one FLVE in the left and right caudate and five and six FLVEs in the left and right putamen, respectively. All graft sites could be identified macroscopically at post-mortem (Figs 1G and 2D).

Figure 1.

Presence of phosphorylated tau within neuronal foetal allografts in a 67-year-old patient with Huntington’s disease 9 years post-transplantation. (A–L) Double immunohistochemistry for distinct forms of phosphorylated tau (nickel-enhanced DAB) labelled with the antibody AT8 (phospho-tau Ser202 and Thr205) and CP13 (pSer202) as well as neuronal elements stained for NeuN (DAB; A–I’’) or MAP2 (J–L). Presence of AT8+ tau neurofibrillary tangles, neuropil threads and neuronal inclusions in the cortex (A–C) and striatum (D–F) of a patient with Huntington’s disease. (G) Macroscopic view of tissue grafted into the caudate nucleus highlighting the transplant architectural subdivisions (P-zones, dotted lines). (H and I) Higher power magnification showing two distinct P-zone and non P-zone areas of the graft where AT8+ inclusions and neuropil threads were found (insets; H’, H’’, I’, I’’). CP13+ fibrils detected in the P-zones and non P-zones of the grafted tissue (J–L). Scale bars: A = 150 µm; B and D = 60 µm; C, H’ and H’’ = 40 µm; E = 30 µm; F and H = 100 µm; G = 500 µm; I and J = 200 µm; I’, I’’, J’, J’’, K and L = 20 µm. NeuN = neuronal nuclei.

Figure 2.

Presence of phosphorylated tau within neuronal foetal allografts in a 40-year-old patient with Huntington’s disease 12 years post-transplantation. (A–K’) Double immunohistochemistry for distinct forms of phosphorylated tau (nickel-enhanced DAB) labelled with the antibody CP13 (Tau-pSer202) or AT180 (phospho-PHF-tau pThr231) as well as neuronal elements stained for MAP2 showing the presence of CP13+ tau inclusions in the cortex (A) and striatum (B and C) of the Huntington’s disease patient. (D) Macroscopic view of the grafted tissue highlighting the transplant architectural subdivisions into P-zones (dotted lines) as observed in the putamen. (E–I’’) Higher power magnification showing the distinct P-zones and non P-zones of the graft and the presence of CP13+ inclusions (insets; E’, G’, H’, I’ and I’’). AT180+ inclusions were also detected in the P-zones and non P-zones of the grafts (J and K). Scale bars: A–C = 50 µm; D = 500 µm; E, G = 50 µm; E’, F, G’, H, I’, I’’, J’, J’’and J’’’ = 20 µm; I = 200 µm; J, K = 100 µm.

For the Parkinson’s disease cases, methods for tissue preparation, transplantation and immunosuppression, as well as clinical evaluation have been described previously (Kordower et al., 1995, 1996). Briefly, both cases received a total of four foetal ventral mesencephalon implants into the right and left post-commissural putamen, respectively (Kordower et al., 1995, 1996). The implants were from donors between 6.5 and 9 weeks post-conception and of which the tissue was dissected into small solid grafts. At post-mortem evaluation, all graft sites could be identified macroscopically (Fig. 4D and P).

Figure 4.

The presence of phosphorylated tau within neuronal foetal allografts in patients with Parkinson’s disease is time-dependent. (A–I) Triple immunofluorescence for phosphorylated tau (AT8; phospho-tau Ser202 and Thr205) and neuronal elements as detected with MAP2 and TH showing the presence of AT8+ tau inclusions in the cortex of a 74-year-old patient with Parkinson’s disease 16 years post-transplantation. (D) Macroscopic view of the grafted tissue as observed in the putamen and various examples of the phosphorylated tau staining patterns observed within the grafted tissue (E–I). (J–R’) Triple immunofluorescence for phosphorylated tau (AT8; phospho-tau Ser202 and Thr205) and neuronal elements stained for MAP2 or TH showing the presence of AT8+ tau inclusions in the cortex (J–L) and striatum (M–O) of a 59-year-old patient with Parkinson’s disease 18 months post-transplantation. (P) Macroscopic view of the grafted tissue as observed in the putamen. (Q–R’) Higher power magnification showing the absence of any type of phosphorylated tau staining within the transplanted tissue. Scale bars: A, B, E-H, K, L, O, Q, R = 50 µm; C, J, M, N = 25 µm; D = 155 µm; I = 100 µm; P = 310 µm.

Tissue preparation for post-mortem histological evaluation

The brains of the transplanted patients were removed within 5 h (Huntington’s disease Patient 1), 2 h (Huntington’s disease Patient 7), 4 h (Parkinson’s disease patient who died 16 years post-transplantation) and 10 h (Parkinson’s disease patient who died 18 months post-transplantation) of death and processed as previously described (Kordower et al., 1995; Freeman et al., 2000; Olanow et al., 2003; Cicchetti et al., 2009). For the control brains, post-mortem delays varied between 5 and 14 h (Table 1). The control brains were also bisected, sliced into 2-cm thick slabs that were fixed by immersion in 4% paraformaldehyde at 4°C for 3 days. They were then stored at 4°C in a 0.1 M phosphate-buffered saline (PBS), pH 7.4 containing 15% sucrose and 0.1% sodium azide. Both the transplanted and control brains were sectioned using a freezing microtome. These slightly different methods of tissue preservation did not appear to affect the quality of the histological staining.

Immunohistochemistry

For the Huntington’s disease cases, brains were processed using methods previously published (Freeman et al., 2000; Cicchetti et al., 2014). Double immunohistochemical staining was performed to identify neuronal elements using an antibody against neuronal nuclei (NeuN, anti-mouse, MAB377, Millipore, 1:2500) or microtubule associated protein-2 (MAP2, anti-mouse, M1406, Sigma, 1:1000) in combination with AT8 (phospho-tau Ser202 and Thr205, MN1020, ThermoFisher, 1:500), AT180 (phospho-tau Thr231, MN1040, ThermoFisher, 1:500), PHF-1 (phospho-tau Ser396 and Ser404, 1:500, generous gift from Dr Peter Davies, Feinstein Institute for Medical Research, NY, USA) and CP13 (phospho-tau Ser202, 1:500, generous gift from Dr Peter Davies) to identify phosphorylated tau (Weaver et al., 2000). Additionally, some sections were double-stained for ubiquitin (ubiquitin, anti-rabbit, Dako, Z0458, 1:2000, a marker of ubiquitinated mHtt), and CP13 to assess the potential co-localization of these two pathological proteins. Antigen retrieval (25 mM Tris-HCl, 1 mM EDTA, 0.5% SDS, pH 8.5; 20 min at 95°C) was performed with the antibodies AT180, PHF-1 (Weaver et al., 2000) and CP13. Photomicrographs were taken using Picture Frame software (Microbrightfield) linked to an E800 Nikon microscope (Nikon Instruments). See Supplementary Table 1 for antibody details.

Immunofluorescence

For the Parkinson’s disease cases, we opted to perform a series of triple immunofluorescent staining to localize phosphorylated tau with AT8 (phospho-tau Ser202 and Thr205, anti-mouse, MN1020, ThermoScientific, 1:500) in combination with antibodies against microtubule associated protein-2 (MAP2, anti-rabbit, 17490-1-AO, Cedarlane, 1:500) and tyrosine hydroxylase (TH, anti-sheep, AB1452, Millipore, 1:1000) (Supplementary Table 1). We further performed co-localization studies between phosphorylated forms of α-synuclein (α-synuclein phospho129 EP1536Y, rabbit, ab51253, Abcam, 1:250) and AT8. All immunofluorescent staining experiments included multiple controls such as brain sections (paraffin-embedded sections) from healthy age-matched controls and negative controls where the primary antibody was omitted from the incubation media (data not shown). Cross-reactivity between antibodies was further prevented by sequential revelations in immunohistochemistry and the use of antibodies made is different species for the immunofluorescence. Photomicrographs were taken using Zeiss Zen Imaging software linked to a Zeiss Imager Z.2 AXI0 microscope.

Stereological quantification

Data on immunostanining patterns were carefully collected by two independent and blinded investigators and compiled for comparison. Observations included location of stainings (brain area, graft area), types of tau staining (neurofribrillary tangles, neuropil threads and inclusions) and frequency of positive staining findings. Additionally, stereology counts were performed on one sample per brain using the optical fractionator method (Glaser and Glaser, 2000). To ensure accurate distinctions between the various forms of tau pathology, observations/counts were made at 60× using the following parameters: SRS grids (2000 µm× 2000 µm), counting frame size (500 µm× 500 µm) and guard zone thickness (2 µm). Positive staining was not counted if they intersected forbidden lines. All quantifications were performed using the Stereo Investigator software (MicroBrightfield, Colchester, VT, USA) paired with an E800 Nikon microscope (Cicchetti et al., 2014).

Western blotting for 3R/4R tau isoforms

Proteins were extracted from fixed brain sections of cortex and striatum from both Huntington’s disease and control patients, as well as the grafted tissue, according to the published protocol (Cicchetti et al., 2014). The membranes were immunoblotted with the 4-repeat tau isoform (4R; anti-mouse antibody, #05-804, Millipore 1:4000) overnight at 4°C followed with horseradish peroxidase-conjugated goat anti-mouse IgG antibody (Jackson ImmunoResearch; 1:25 000) and detected by the addition of chemiluminescence reagents (Luminata Forte; EMD Millipore). The membranes were then stripped and reprobed for 3-repeat tau isoform (3R; anti-mouse antibody, #05-803, Millipore, 1:4000) and actin (anti-actin; anti-mouse antibody, #G043, ABM, 1:5000). Images of the membranes were acquired using myEcl Imager (Thermo Fisher Scientific). Band intensities were quantified and compared using ImageJ software (NIH). The graphs were generated using the Prism software (v6.01; GraphPad Software, San Diego, CA). Note that the western blot analyses were exclusively conducted on the transplanted Huntington’s disease patient who died 12 years post-transplantation and the age-matched control, as tissue samples from the other patients (the second Huntington’s disease transplanted patients and the two Parkinson’s disease transplanted cases) were not sufficient.

Image preparation

All images were prepared using Combine ZP and Adobe Photoshop CS5. When necessary, brightness and contrast adjustments were made. Panels were assembled using Adobe Illustrator CS5.

Results

All of the immunohistochemical stainings confirmed the presence of phosphorylated tau in the parenchyma of the host brain of both Huntington’s disease cases; within the cortex (Figs 1A–C and 2A) and striatum (Figs 1D–F and 2B and C), similar to what has been reported (Fernandez-Nogales et al., 2014; Gratuze et al., 2015; Vuono et al., 2015). The AT8 staining revealed neurofibrillary tangles, neuropil threads and neuronal inclusions (Fig. 1A–F, similar to previous reports) (Fernandez-Nogales et al., 2014; Vuono et al., 2015), which were also observable in the younger grafted patient (40-year-old) using the CP13 antibody (Fig. 2A–C). All forms of tau-related pathology (i.e. neurofibrillary tangles, neuropil threads and neuronal inclusions) were quantified by stereology in the main structures available on each slide (e.g. cortex, caudate, putamen, grafts) and plotted as the number of events observed per surface area (Fig. 3A and B).

Figure 3.

Stereological quantifications of tau-related pathological elements in Huntington’s disease and Parkinson’s disease transplanted cases. (A) Quantification of AT8+ (phospho-tau Ser202 and Thr205) tau staining in the Huntington’s disease patient who passed away 9 years post-transplantation. (B) Quantification of CP13+ (Tau-pSer202) tau staining in the Huntington’s disease patient who passed away 12 years post-transplantation. (C) Quantification of AT8+ tau staining in the Parkinson’s disease patient who passed away 16 years post-transplantation. (D) Quantification of AT8+ tau staining in the Parkinson’s disease patient who passed away 18 months post-transplantation. Neurofibrillary tangles, neuropil threads and inclusions were quantified using stereology and plotted as the number of events encountered per surface area expressed in mm² of tissue.

To confirm this finding, we used four distinct antibodies against different forms of the phosphorylated protein: AT8 (phospho-tau Ser202 and Thr205), CP13 (pSer202), AT180 (pThr231) as well as PHF-1 (pSer396 and pSer404; data not shown). In the two transplanted Huntington’s disease cases investigated, the grafted tissue was macroscopically identifiable in the host brain (Figs 1G and 2D) and was composed of P-zones, areas typically rich in striatal cells, and non-P-zones, areas devoid of such cells (Cicchetti et al., 2009; Freeman et al., 2011). AT8+ neuropil threads were found throughout the P-zones (Fig. 1H, H' and H'') and the non-P-zones of the grafts (Fig. 1I, I' and I''). Similar results were obtained with the CP13 antibody (Figs 1J–L, 2E–I'' and 3). The AT180 antibody revealed the presence of more fragmented and punctuated tau+ inclusions (Fig. 2J–K'). In all cases, the antibodies showed tau+ neurofibrillary tangles, neuropil threads and inclusions. Interestingly, neuropil threads were sometimes seen overlapping/crossing P-zone and non-P-zone (Fig. 1I''). Western blot analyses further revealed an increase in the ratio of 4R/3R tau isoforms within the grafted tissue of the Huntington’s disease patient who died 12 years post-transplantation (data not shown). Tissue samples were not sufficient to run the same analysis on the second transplanted Huntington’s disease patient.

Immunofluorescence staining confirmed the presence of phosphorylated tau in the cortex and striatum (both caudate and putamen) of the two Parkinson’s disease cases again in the form of neurofibrillary tangles, neuropil threads and inclusions (Figs 3C, D, 4A–C and J–O) using the AT8 antibody, in addition to the neuronal marker MAP2 and TH. In the 16-year post-transplantation case, AT8 staining was observed within the graft and at the graft–host border in the form of neurofibrillary tangles, neuronal inclusions and some neuropil threads (Figs 3C and 4D–L). In the 18-month post-transplantation case, AT8+ staining was present in both the cortex (Fig. 4J–L) and striatum (Fig. 4M–O). However, the graft was completely devoid of phosphorylated tau staining in this case (Fig. 4P–R’). In summary, tau pathology was found in the form of neurofibrillary tangles, neuropil threads and inclusions within all of the transplants in both Huntington’s disease and Parkinson’s disease cases with the exception of the Parkinson’s disease patient who came to autopsy 18 months post-transplantation.

Finally, double immunohistochemistry for ubiquitin and CP13 failed to reveal any co-localization between these pathological proteins within the grafted tissue sampled in the transplanted Huntington’s disease patients (Supplementary Fig. 1A–C). Double immunofluorescence for phosphorylated α-synuclein and AT8 in the Parkinson’s disease brains did, however, indicate co-localization of these proteins within both the striatum and the grafted tissue (Supplementary Fig. 1D–G).

To further validate these findings, we used the same immunohistochemical markers in the brain of a 70-year-old control subject. We found, as expected in a brain of this age, several examples of phosphorylated tau inclusions, similar to that seen in the age-matched Huntington’s disease patient (Vuono et al., 2015) (Fig. 5). In contrast, control brains of younger patients were largely (40-year-old), if not completely (12-year-old) devoid of any such phosphorylated tau pathology (Fig. 5).

Figure 5.

Tau-related pathology in control brains and as a function of age. (A–I) Representative images of double immunohistochemistry for neuronal elements (NeuN or MAP2; DAB) and phosphorylated tau (AT8, CP13, AT180; nickel-intensified DAB) in control brains of 70, 41 and 12 years of age. (A–C) The brain of a healthy 70-year-old control, matching the age of the Huntington’s disease patient presented in Fig. 1 and the Parkinson’s disease patient presented in Fig. 4, display age-related tau pathology. (D–F) Only a small number of phosphorylated tau neuronal threads (black arrows) can be seen in the brain of a healthy 41-year-old control, age-matched with the Huntington’s disease patient presented in Fig. 2. (G–I) In contrast to aged controls, the brain from a 12-year-old control, corresponding to the age of the grafted tissue, is completely devoid of any tau inclusions. Scale bars: A = 30 µm; B, G, H = 120 µm; C, F, I = 100 µm; D = 50 µm; E = 60 µm. NeuN = neuronal nuclei.

Discussion

We show for the first time that various forms of hyperphosphorylated tau can be found within genetically unrelated allografted tissue in patients with both advancing Huntington’s disease and Parkinson’s disease. The observation that tau pathology can manifest in two different types of foetal transplants in two distinct disease environments implies that this phenomenon may be common to several neurodegenerative diseases. Our study also confirms the presence of tau pathology in the brains of Huntington’s disease and Parkinson’s disease patients, suggesting that a number of proteins may overlap to contribute to the complex pathophysiology of these disorders.

Despite the fact that mechanisms of action cannot be explored in human post-mortem material, we can speculate on a number of pathways that may have contributed to the appearance of tau pathology within the grafted tissue. As suggested by the accumulating scientific evidence, one plausible scenario is that tau protein may potentially spread from the host to the unrelated, otherwise healthy, neuronal tissue. There is a growing body of in vivo reports that provide support for the potential of tau to propagate between cellular elements, trans-synaptically and in a prion-like fashion (Walker and Jucker, 2015; Braak and Del Tredici, 2016; Sanders et al., 2016). Intraparenchymal injections of tau aggregates into transgenic animals overexpressing human tau induce such pathology along connected brain networks (Clavaguera et al., 2009; Iba et al., 2013), while transgenic animals expressing tau within the entorhinal cortex eventually show pathology in remote brain areas synaptically linked (de Calignon et al., 2012; Liu et al., 2012; Iba et al., 2013). Two very recent independent reports further demonstrate that various tau strains as well as tau fibrils purified from Alzheimer’s disease brains have the capacity to spread within the cerebral tissue (Duka et al., 2013; Kaufman et al., 2016), with each strain characterized by a distinct spreading potential yielding multiple neuropathological patterns (Kaufman et al., 2016; Espuny-Camacho et al., 2017). The developmental regulation of exon 10 splicing of the tau protein results in foetal brains expressing only 3R isoforms of tau. Adult human brains, on the other hand, express equivalent amounts of both the 3R and 4R isoforms (Goedert et al., 1989; Medina and Avila, 2014), while Huntington’s disease brains express more 4R tau (Fernandez-Nogales et al., 2014; Vuono et al., 2015). For this reason, a greater ratio of 4R/3R in the graft would suggest pathology spreading. While it was performed only on one patient (n = 1, due to lack of available tissue), the analysis of 4R/3R tau ratio by western blot showed an increase of 4R tau in both Huntington’s disease striatum and Huntington’s disease graft when compared to control striatum, which suggests that transplants may have acquired tau pathology from the host brain. The less abundant phosphorylated tau staining within the 9, 12 and 16-year-old transplants in comparison to the host brains may reflect a time-dependent process of host pathology spreading to, or occurring within the grafts. This is supported by the lack of tau pathology in a Parkinson’s disease patient 18 months post-transplantation. The presence of tau within the grafts relative to similarly aged hosts may also be due to an accelerated ageing process of the transplanted foetal tissue as the diseased environment may indeed precipitate this phenomenon. Similar to our results, others have reported co-localization of phosphorylated tau and α-synuclein stainings, suggesting that tau may co-aggregate with α-synuclein in Lewy bodies (Ishizawa et al., 2003).

While the presence of tau pathology in the genetically unrelated allografted tissue may be a result of protein propagation and templating, it remains possible that it originates from other processes as well. For example, studies have found that the grafts may be surrounded by inflammatory markers (Brundin and Kordower, 2012), and tau pathology can be exacerbated by inflammation (Zilka et al., 2012). This theory finds support in a recent post-mortem case study where Lewy body disease found within the transplanted tissue in a Parkinson’s disease case may have been provoked by significant surrounding gliosis, as suggested by the authors (Ahn et al., 2012). Oxidative stress and excitotoxicity have been suggested to enhance α-synuclein pathology in Parkinson’s disease grafts (Brundin and Kordower, 2012), and both have been shown to induce tau phosphorylation (Ong et al., 2013; Alavi Naini and Soussi-Yanicostas, 2015). The presence of 4R tau in grafted cells may well be the result of oxidative stress causing an acceleration of ageing on the cells. This is supported by the pigmentation of grafted dopaminergic neurons previously reported in transplanted Parkinson’s disease cases, which is unlikely to be passed on from surrounding striatal tissue (Li et al., 2008). Another possibility is that mHtt or α-synuclein pathologies that have propagated to the grafts would generate hyperphosphorylated tau locally, as both pathologies can induce tau hyperphosphorylation (Blum et al., 2015; Kawakami and Ichikawa, 2015; Gratuze et al., 2016). Along these lines, it has been demonstrated that mechanisms such as suppression of metabolism or thermogenesis, as experienced during hibernation, can induce and reverse hyperphosphorylation of proteins and enzymes (Arendt and Bullmann, 2013). Finally, although remote, we could hypothesize that the foetal tissue selected for transplantation was genetically predisposed to the development of tau pathology. Indeed, it has recently been established that there is more severe cognitive decline in Huntington’s disease patients with H2 haplotype of the tau gene (MAPT) compared with Huntington’s disease patients of the H1 haplotype (Vuono et al., 2015). This is also true for Parkinson’s disease where APOE (Gao et al., 2011) or THAL amyloid (Murray et al., 2014) assays have revealed susceptibilities for co-morbidities of Parkinson’s disease and Alzheimer’s disease. Unfortunately, the methodologies needed to confirm such genetic features require a significant amount of mRNA, which was not accessible in our samples both because of tissue preservation (fixed tissue) and quantities available.

A number of recent publications have reported tau pathology in Huntington’s disease patients as well as in vivo and in vitro models of the disease (Fernandez-Nogales et al., 2014; Blum et al., 2015; Gratuze et al., 2015; Vuono et al., 2015). Collectively, these findings provide convincing evidence that Huntington’s disease pathology is accompanied by abnormalities in tau expression and function. In the initial human studies, tau pathology was detected using only the anti-phospho tau antibody AT8 (Fernandez-Nogales et al., 2014; Vuono et al., 2015). Subsequently, the AT8 and pS199 antibodies were used to identify hyperphosphorylated tau in the brains of Huntington’s disease patients by immunohistochemistry and western blotting of the sarkosyl-insoluble tau fraction (Gratuze et al., 2015). More recently, two antibodies recognizing abnormal tau conformation, MC1 and Alz50, have also shown the presence of pathological tau in the Huntington’s disease post-mortem material (Gratuze et al., 2015, 2016). Here, we provide further evidence for tau pathology in Huntington’s disease using two additional antibodies, CP13 and AT180. It is probable that most tau pathology in Huntington’s disease brains does not progress to the argyophilic state, since they do not stain with the Bielschowsky (another silver stain) method (Jellinger, 1998). Indeed, the majority of studies reporting tau pathology in Huntington’s disease patients used phospho-tau antibodies (Gratuze et al., 2015, 2016). Similarly, we failed to identify MC1 staining within the grafts, an antibody that detects advanced stages of pathological conformational changes in tau, prior to tau aggregates becoming argyophilic (Jicha et al., 1997). The tau pathology that we observed within the grafted tissue is of neuronal origin with numerous neuropil threads typical of tau lesions, and neuronal inclusions within cell bodies (Braak et al., 2006). Finally, the absence of co-localization of mHtt (observed with ubiquitin) and AT8 within the grafted tissue in our Huntington’s disease cases may relate to the fact that we were extremely limited in terms of tissue availability and therefore may not have been able to observe this protein interaction in our samples, as has been previously reported in the brain of Huntington’s disease patients (Vuono et al., 2015).

In Parkinson’s disease cases, immunohistochemical evidence of tau pathology has recently been described with PHF-1 (phospho Ser396/404) showing the protein to be present as aggregates within the cell bodies of neuronal elements of the substantia nigra (Duka et al., 2013). AT8+ hyperphosphorylated tau aggregates surrounded by granular T22+ tau oligomers were also reported in the cortex of Parkinson’s disease patients (Sengupta et al., 2015). We also observed the co-localization of α-synuclein and AT8 within the grafted tissue of the 16-year post-transplantation Parkinson’s disease case, as previously reported (Sengupta et al., 2015).

We herein provide the first evidence for the presence of tau pathology in healthy tissue transplanted to Huntington’s disease and Parkinson’s disease patients more than a decade post-grafting. Among the possible explanations for this, we suspect that the tau protein, found in other types of neurodegenerative diseases, including Alzheimer’s disease, may spread through the nervous system via similar mechanisms or result from a number of factors impaired in the diseased surrounding tissue and therefore negatively impacting on the grafted cells. These observations not only extend the growing body of evidence that protein transfer is common to many neurodegenerative disorders but that tau plays an active role in Huntington’s disease and Parkinson’s disease pathophysiology.