Neural transplants in patients with Huntington's disease undergo disease-like neuronal degeneration: Q&A with the authors.

The work reported in PNAS by Cicchetti et al (2009) adds to a growing number of reported cases of autopsies of patients with Huntington's disease (HD) who received fetal neural transplants.  The authors describe three HD patients who died 10 years post-transplantation.  Two of the patients had multiple surviving transplants, which appear to have differentiated into appropriate cell types and received innervating glutamatergic and dopaminergic projections from the host brain.  However, the grafted cells, particularly the ones with striatal-like differentiation (P-zones), exhibited degenerative changes, including cytologic swelling and vacuolation, abnormal chromatin deposits, apparent diminished calbindin, and expression of the apoptosis-associated protein caspase 3, and were associated with astrogliosis, inflammatory infiltrates, and microglial activation.  The authors characterize the degeneration as accelerated compared with host neostriatum, and suggest excitotoxicity due to host-graft glutamatergic innervation or glutamate-releasing microglia, as well as impaired host neurotrophic support, as possible etiologies.  They also show no reduction in host neostriatal degeneration, and conclude that the "combination of an unfavorable risk-benefit profile [propensity of patients to develop subdural hematoma as a complication of transplantation], short-term and mild [clinical] benefits at best, and significant disease-like graft degeneration makes future trials of fetal-cell transplantation using these techniques potentially unwarranted."

Our group has examined four HD transplant patients from a different clinical trial (Kopyov et al, 1998) who survived at least six years following fetal neural transplantation.  We found healthy grafts with patchy host-graft astrogliosis but no evidence of immune rejection, degeneration, or HD inclusions, but also no indication of connectivity with the host (Keene et al., 2007; 2009; unpublished observations).  A critical difference between the trials is the source of donor tissue.  Freeman's group (current study; see Hauser et al., 2002 for details) used lateral-most lateral ganglionic eminence, whereas we studied patients who had received total lateral ganglionic eminence (Kopyov et al., 1998).  Cicchetti et al (2009) describe preferential degeneration of P zone regions, which is important because P zone areas have cytologic and architectural similarity to striatal tissue that degenerates in HD.  Early experimental P-zone differentiation studies showed these regions were derived primarily from lateral-most lateral ganglionic eminence (Freeman et al, 1995).  The findings in this study, compared with other studies in which P zone differentiation is less apparent, suggest that mechanisms of selective striatal degeneration in HD may not require formation of ubiquitinated neuronal intranuclear inclusions (since none were identified in grafted cells), that donor-derived P-zone (striatal) neurons may be particularly susceptible to intrinsic or extrinsic degenerative stimuli compared with neurons of other striatal lineages, that significant host-to-graft innervation may be dependent on P-zone graft differentiation, and that innate and acquired immune response to grafted cells may be P-zone-dependent.  Evaluation of long-surviving patients from European trials who utilized whole ganglionic eminence (Bachoud-Levi et al., 2006) will shed more light on the relative importance of P-zone differentiation and graft neurodegeneration. 

Some of the findings in this study are encouraging.  The surviving grafts differentiated into striatal-type parenchyma by morphologic and immunohistochemical criteria.  Unlike the case we described recently in which a patient with HD suffered significant mass effect due to the formation of donor-derived cysts and transplant overgrowth (Keene et al., 2009), no such pathology was reported by Cicchetti et al (2009).  Moreover, in 2 of 3 patients, multiple grafts survived 10+ years, albeit with degenerative changes, and demonstrated some host-to-graft connectivity even after this extended period of time.  No apparent graft rejection was identified, although chronic low-level host-to-graft response could not be excluded.  Finally, ubiquitinated proteinaceous inclusions characteristic of HD were not identified in grafted neurons.

Critical indicators of neural transplant success are long-term safety, therapeutic efficacy, and graft survival and integration with host.  No study has reported long-term clinical benefit in a more than anecdotal fashion in HD patients who received neural transplants. We previously identified long-term surviving healthy grafts with no graft-host connectivity (Keene et al., 2007) and described cyst formation and graft overgrowth 10 years post-transplantation (Keene et al., 2009).  Now, Cicchetti et al (2009) reaffirm the increased risk of subdural hemorrhages (hematomas) as a serious complication of neural transplantation in HD, and show graft-host connectivity with appropriate striatal differentiation in association with graft degeneration.  Although continued negative results bode poorly for neural transplantation as a cell-replacement strategy in HD going forward, the findings reported by Cicchetti et al. and others are of critical importance, since the analysis of graft pathology and comparison between different transplant methodologies with relation to clinical and histopathologic outcomes provides valuable insight into disease pathophysiology and graft-host interactions, and will help guide development of future transplant trials in other neurological disorders. 

The paper, Cicchetti et al. (2009)with the title “Neural transplants in patients with Huntington’s disease undergo disease-like neuronal degeneration” claims based on 2 post-mortem cases transplanted a decade earlier, using simple measures of cell density and morphology, that there is evidence of some disease (presumably Huntington’s disease - HD) affecting the transplants. However, there is no evidence for this. The article presents data of cell density, measured in regions of the transplants that either contain the primordial striatum (“P-zones”) from the elected aborted dissection, or parts of the surviving pieces that contain other fetal brain regions transplanted into the brains of these advanced patients with HD.  These cell density measures showed higher neuronal density in the host HD brain than in some parts of the transplant, and the authors interpret this as indicating that there is more cell loss in the transplant than there is in the patients’ brain. This illustrates a lack of understanding, because cell density is specific to transplants and should not be equivalently compared to the host. In the HD brain, where neurons die, there is also a loss of glia over time, so the neuronal-glial ratio remains relatively constant despite up to 70-80% loss of neurons. Consequently, cell density measures as used here are invalid and distorted.

Even if the alleged loss of striatal neurons was occurring in transplants, this could simply be due to processes of ischemia or inflammation that has been described to be occurring not only in developing transplants but also in Huntington’s disease, Parkinson’s disease and other neurological diseases (Meade et al., 2000). The changes in this tissue transplanted would depend on the relative vulnerability to age-related changes specific to the regions (e.g. loss of medium-sized spiny neurons) and would occur in any transplant condition, a “healthy” recipient or with inflammation in the brain, and thus not be disease specific. Likely, the neural transplants will simply age over time and would be adversely influenced by any inflammatory processes, which are also present in HD (Silvestroni et al., 2009).

Another problem with data interpretation is the lack of reasoning and understanding around clinical impact of these transplants. The authors describe a simple ratio measure composed of striatal (P-zone) vs. nonstriatal transplant tissue that has been calculated in rodents.  The index used indicates that 50% of the transplant needs to be striatal-like to have a functional effect. However, this ratio assumes a meaningful replacement of total loss of striatal host tissue and neurons. The authors themselves describe that their transplants only replaced 3-4% of the patients’ striatum. Clearly, a few percent reconstitution of the striatum is insufficient to provide a clinical and meaningful effect. Nonetheless, the article uses the high measures of cell density and “P zone” (striatal-like tissue) ratios present as an argument against any future clinical efforts using cell therapy for HD. This paper infers like 2 related prior papers on Parkinson’s disease(Kordower et al., 2008; Li et al., 2008), that transplants have signs of disease somehow specifically caused by a disease process present in the host, instead of the obvious effects of aging in a transplant condition. Such interpretations lead to fear and scenarios of transmission of disease. With more data and interpretations like this, I am concerned about the effects on research to the entire field of cell and regenerative therapy that would delay potentially helpful – even curative methods to develop. New innovations and methodology with the advent of stem cells biology or even regenerating cells within the tissue must continue to be explored or this may delay or prevent these promising applications from reaching the patients.

Cicchetti and collaborators (Cicchetti et al, 2009) report on unique and important findings regarding the long-term fate of neural grafts in three Huntington’s disease (HD) patients. Almost a decade after surgery, significant numbers of the grafted embryonic brain cells were still alive - and expressed markers of medium spiny striatal neurons - in two of the patients.

The most important finding, however, is that the grafted cells exhibit signs of neurodegeneration. The authors speculate that the degenerative changes are due to excitotoxicity. They suggest that the excitotoxic damage might depend upon glutamatergic input from the host neocortex and they provide morphological evidence that such connections were present. The observations in the Cicchetti study are particularly interesting because they suggest that non-cell autonomous degenerative mechanisms may operate in HD. This means that mutant huntingtin might not just trigger degeneration and death in the cell where it expressed. Instead, cells might be the victims of secondary events triggered by mutant huntingtin in other brain regions, e.g. via excitotoxicity, or in closely neighboring cells. While Cicchetti and collaborators discussed the action of glutamate, released from cortical afferents, on the grafted neurons, there is also the possibility that mutant huntingtin affects host microglia directly and thereby promotes neurodegeneration in the grafts. In a series of studies, Schwarcz, Guidetti, Muchowski and colleagues (for review see Schwarcz et al, 2009) have suggested that microglia play a key role in the neurodegeneration in HD. Specifically, they have shown that mutant huntingtin perturbs the kynurenic acid pathway, both in neurons and microglia. As a result, levels of neurotoxic 3-hydroxykynurenine and quinolinic acid are elevated and may promote excitotoxic damage in neighboring neurons. This is particularly interesting because Cicchetti et al demonstrated activated host microglia around the transplanted neurons. Thus, these microglia may have been activated directly by mutant huntingtin and mediated the degeneration in the grafts.

Another intriguing observation in the Cicchetti paper is that the grafted neurons display no signs of huntingtin aggregation. Tissue sections were examined with an antibody against ubiquitin and an antibody (called EM48) that is directed against the first 256 N-terminal amino acids of huntingtin and believed to bind to only the aggregated form of huntingtin. At first glance, the lack of protein aggregation in the grafted cells may seem expected and unremarkable, since the grafted cells obviously do not express mutant, aggregation-prone huntingtin with an expanded polyglutamine stretch. However, earlier this year Ren and coworkers demonstrated that a synthetic protein containing an expanded polyglutamine protein is able to penetrate the outer membrane of cultured cells (Ren et al, 2009). Once inside the new cell the synthetic protein could act as a seed and promote nucleation of endogenous wild-type polyglutamine-containing protein, such as huntingtin. Similarly, aggregates of mutant tau (of particular relevance to Alzheimer’s disease and frontotemporal dementia) were recently found to be able to move across cell membranes and promote aggregation of normal tau (Frost et al, 2009). Furthermore, another recent study demonstrated that when mutant human tau is injected into brains of mice it causes aggregation of wild-type tau and the pathology can spread from the injection site to neighbouring brain regions (Clavaguera et al 2009). Alpha-synuclein, which is implicated in the pathogenesis of Parkinson’s disease, has also been shown to be capable of moving from one cell to another, in cultures (Lee 2008).

Why then are there no obvious signs of transfer of polyglutamine protein aggregation pathology from the host brain to the grafted neurons? Possibly the phenomenon of intercellular transfer of polyglutamine pathology is only observed in cell cultures. Maybe misfolded huntingtin is never released into the extracellular space in HD and therefore never available for uptake by neighboring cells. When neurons die in HD, they have been suggested to undergo apoptosis that would imply that cellular remnants are carefully engulfed by microglia without their cytoplasmic content accessing adjacent cells. Alternatively, aggregation may eventually occur in grafted cells but requires more time than the approximate 9 years that the two patients lived after surgery. There are reasons to believe that the time elapsed since surgery is an important factor. As discussed by Cicchetti, other studies have shown that a subpopulation of grafted neurons in Parkinson patients develops Lewy bodies, which represent the pathological hallmark of the disease. To date there are at least 7 Parkinson cases, either published or reported orally, displaying Lewy bodies in grafted neurons (Brundin et al, 2008; Kordower et al 2008a, 2008b; Li et al, 2008). The phenomenon seems to occur in the vast majority of grafted cases and develops gradually starting after about a decade. The fact that Cicchetti and coworkers do not observe protein aggregates in the grafted neurons in HD is indeed interesting. It does not, however, distract interest from the emerging interest that a prion-like transmission of proteinopathy might take place as degeneration spreads from one brain region to the next in several of common slowly progressing neurodegenerative diseases (Aguzzi, 2009).

Even though the Cicchetti’s observations in a small number of grafted patients are exciting, naturally they cannot provide all the answers regarding neurodegenerative mechanisms in HD. For example, the ultimate control – i.e. striatal tissue implants made into the striatum in healthy, non-Huntington brain - is missing for obvious reasons. It is conceivable that grafted striatal neurons will gradually undergo degenerative changes in the long-term, regardless of the genotype of the host, simply due to them not forming absolutely normal connections. As long as there are no striatal tissue grafts in normal individuals (and these should hopefully never take place) we will never know. Experiments in animal models of HD are needed to address these issues. Dunnett and collaborators have already shown that intrastriatal grafts of striatal tissue in the R6/2 transgenic mouse model survive well and do not exhibit obvious neuropathology (Dunnett et al, 1998). Possibly, longer survival times are needed, or the R6/2 model that only expresses exon 1 of the mutant HD gene is not appropriate. Undoubtedly, the insightful observations made by Cicchetti et al will trigger several exciting experimental studies that may eventually help us understand how HD and other neurological proteinopathies such as Parkinson’s disease develop.

Dear Sirs,

We are grateful for the thoughtful comments by Dr. Brundin, and agree that our recent findings in PNAS open many new scientific lines of inquiry into the mechanisms of neurodegenerative diseases in general, and HD in specific.

However, we found Dr. Isacson’s comments to be misconceived and troublesome.

First, he stated that our paper described findings from two patients with HD who received neural transplants and then had autopsies when they died. In fact, the paper reported new data from three patients, and compared results from a fourth patient reported earlier in a paper published in 2000 (Freeman et al., 2000) (also in PNAS with the same authors and methods, but with Dr. Isacson also as a co-corresponding author on the 2000 paper).

He suggested that our paper reported no findings of HD-like degeneration within the grafts, and presented a convoluted discussion about cell densities within the graft vs. the host. The evidence in our paper supporting the conclusion that the grafts had HD-like degeneration was abundant: 1) grafted projection neurons degenerated preferentially in comparison to interneurons, which recapitulates the pattern of neural degeneration in HD; 2) cells degenerated at least in part via apoptotic mechanisms; 3) cells had a markedly degenerated appearance in comparison with our earlier report at 18 months after transplantation in a patient with HD using identical techniques; and 4) mechanisms of neural degeneration via targeted microglial infiltration and glutamate neurotoxicity of projection neurons were postulated based on supporting findings in our new report (Cichetti et al., 2009). With regard to his arguments about cell densities were misdirected. In the HD autopsy study we published at 18 months after transplantation, the graft was far more cellular than the host. In the new study, 10 years after transplantation, the host was now more cellular than the graft (Cichetti et al., 2009). His discussions regarding P-zones have nothing to do with what is a very straight forward observation regarding the relative speed of degenerative changes in the graft with respect to the host at 1½ vs. 10 years after transplantation. Finally, the third autopsy in the paper (which he forgot to notice that we reported) had at most a few surviving cells. In our experience with numerous autopsies from the University of South Florida program, in both PD and HD, this has NEVER been observed before (in over 10 publications spanning 14 years). Likewise, the lack of survival of any graft in the caudate in all 4 HD patients is novel. These findings add further evidence that the grafts are undergoing neural degeneration in HD patients long-term, and that the transplant milieu in HD may be unique in comparison to PD, particularly with respect to the caudate (where the findings of HD are first evident).  

We agree that ischemia may cause similar findings in rodent brain, but his ischemia hypothesis with relation to our study, in particular, is questionable. Graft death due to ischemia at the time of transplantation would be expected to be quite rapid. As he knows from the 2000 paper in which he was a co-corresponding author, graft survival at 18 months was in fact quite robust (Freeman et al., 2000). As none of the three patients surviving 10 years had strokes, and advanced degeneration was observed bilaterally in these patients at 10 years after transplantation, this argument not persuasive.

We agree with him that inflammation was a critical part of the degeneration process, and stated that explicitly within the results and discussion section of the paper.

We cannot prove that simple aging of grafts did not play a role in our findings, but this also seems unlikely. First, the cells are only 10 years old, and advanced degeneration of cells is not routinely seen in 10 year old brains, even in those with the gene for HD. Next, the degeneration seen in our grafts of nigral tissue in patients with PD resembled disease-like degeneration rather than age related changes. Specifically, alpha-synuclein is evident in its cytoplasmic form in advanced aging as well as Parkinson’s disease. However, aggregated alpha-synuclein (seen in 5-8% of grafted cells in two reports from our group, and a total of four separate studies at this time) was only characteristic of dopamine neurons in PD, not dopamine neurons in aging patients. Dr. Isacson has made similar arguments in the PD transplant domain, and we are not surprised that he is therefore making similar arguments in the HD transplant domain as well. As a group, none of the authors of our paper have any interest in pursuing a similar drawn-out and ultimately fruitless discussion with Dr. Isacson.

Our “lack of reasoning and understanding around the clinical impact” of our studies is insulting. First, the clinical data that we reported was discussed in the most conservative possible terms, with the limitations of any open label, phase one study in general. Even with this caveat, we reported modest, short lived benefits at best in only two of the three patients. It is quite reasonable to question the relationship of possible short term benefits with time-related graft degeneration. Finally, his argument that grafts only occupied 3-4 percent of striatal volume did not seem to bother him when he was the co-corresponding author along with Dr. Freeman on the 18 month autopsy report finding nearly identical graft volumes as reported at 10 years, but this time without him as an author. What this suggests to us instead is a need for stem cells, where greater quantities of cells for transplantation may provide more benefit.

Finally, he states that our paper may “lead to fear and scenarios of transmission of disease.” He then goes on to say that he is “concerned about the effects on research to the entire field of cell and regenerative therapy…” Isn’t truth the most important requirement for scientific inquiry? We are saddened that he interprets science based on his fears rather than with a passion to use new observations as a window into opportunities for future research about potentially novel scientific and therapeutic discoveries.

Sincerely yours,

Thomas Freeman MD and Francesca Cicchetti Ph.D.

The truths about transplants to HD?
I am truly sorry to see how offended Drs Tom Freeman and Francesca Cicchetti apparently became by reading my short commentary “Problems with interpretations of patients' cell transplantation data” about their recent paper. My comments questioned the negative and alarmist reactions to neural transplantation in patients, as such reactions will potentially impede interest and efforts placed in new clinical treatment approaches, which are using different cell preparations and transplantation methodologies than those reported in their paper. I also honestly find fault with a large number of their scientific measures and interpretations. In their reaction, posted on this web-site with the title “Neural transplants in Huntington’s disease: Truth vs Fear”; they begin and end by resorting to philosophy by asking whether truth is compromised by the questions raised about their interpretations and measures, including alarmist recommendations that clinical trials of cell transplantation are not warranted. We agree about the superordinate importance of truth. However, from their comments it is now clear that these investigators believe themselves to carry “truths” rather than results and data. The truths about transplants are not yet established (for example, there are approximately 50 more HD patients currently in clinical transplantation trials in Europe), and that is why it is important to examine any current post-mortem results and conclusions in detail. Contrary to these authors, others may take the combined findings of the early clinical neural transplantation field to date to mean that neural cell transplants have survived remarkably well in patients, considering it is a decade or more after transplantation.

As is probably clear by now, members of PD Online Research post at will and without censorship.  This is precisely how the site is designed: as a platform for qualified professionals to share their expert opinions without editorial interference.  We depend on the community of scientific professionals to bring their expertise to the discussion:  If you agree or disagree or have questions to raise,  post your views and move the discussion forward. 

This discussion so far has raised several interesting points, and the one that jumps out at me is this: Leaving aside the thorny question of defining ‘disease-like’ pathology in a graft, there are remarkable differences between the patterns of survival of neural grafts in HD brains as compared to PD brains.  What does this tell us about the cellular environment in each of these diseased brains? What kinds of factors may be driving this differential pathology in ‘healthy’ graft tissue, and might these factors present likely targets for therapeutic development or for improving graft health?