|Home | About | Journals | Submit | Contact Us | Français|
Transplantation of human fetal neural tissue into adult neostriatum is an experimental therapy for Huntington's disease (HD). Here we describe a patient with HD who received ten intrastriatal human fetal neural transplants and, at one site, an autologous sural nerve co-graft. Although initially clinically stable, she developed worsening asymmetric upper motor neuron symptoms in addition to progression of Huntington's disease, and ultimately died 121 months post transplantation. Eight neural transplants, up to 2.9 cm, and three ependymal cysts, up to 2.0 cm, were identified. The autologous sural nerve co-graft was found adjacent to the largest mass lesion, which, along with the ependymal cyst, exhibited pronounced mass effect on the internal capsules bilaterally. Grafts were composed of neurons and glia embedded in disorganized neuropil; robust Y chromosome labeling was present in a subset of grafts and cysts. The graft-host border was discrete, and there was no evidence of graft rejection or HD pathologic changes within donor neurons. This report, for the first time, highlights the potential for graft overgrowth in a patient receiving fetal neural transplantation.
Huntington's disease (HD) is an autosomal dominant neurodegenerative disease characterized by progressive motor, cognitive, and psychiatric impairment mediated largely by selective loss of the medium spiny GABAergic projection neurons of the caudate nucleus and putamen. Neural transplantation as a method of cell replacement or neuroprotection has been widely investigated in multiple rodent and non-human primate models of HD that show robust interconnectivity between graft and host without negative outcomes, such as overgrowth of grafted tissue, cyst formation, or aberrant differentiation, and partial to complete recovery of lesion-induced deficits in multiple transplant paradigms (see  for review). These studies led to two case reports describing fetal neural transplantation in patients with HD [14,23] followed by four clinical trials investigating the safety and efficacy of this procedure. The first of these clinical trials was performed by Kopyov et al. [10,21], followed by trials in Florida , France [1-3] and United Kingdom .
Three patients from these trials have come to autopsy. The first was described by Freeman et al. , in which, at approximately 18 months post-transplantation, examination showed graft survival without evidence of significant rejection, as well as appropriate differentiation of the grafted tissues, but very limited graft-host connectivity and no mass or cyst formation. We recently described two autopsies of long-surviving HD patients following fetal neural transplants, and showed graft survival, striatal differentiation, and lack of immune rejection or adoption of HD pathologic changes; however, graft-host connectivity was again very limited . Here we describe autopsy findings of a patient from the original HD neural transplantation trial, in which, unlike the prior autopsies, we find evidence of graft overgrowth in association with autologous sural nerve co-graft and prominent ependymal cyst formation.
The neural transplantation and clinical protocols were approved by the Institutional Review Board at Good Samaritan Hospital in Los Angeles, CA, and patients were enrolled only after appropriate informed, written consent was obtained. Patients with pathologic CAG trinucleotide repeat expansion in exon 1 of the Huntington gene were selected based on criteria proposed in the Core Assessment Program for Intracerebral Transplantation for Huntington's disease . Inclusion and exclusion criteria are detailed elsewhere  but included chorea as the primary clinical symptom, radiographically confirmed striatal atrophy, family history of HD, PET scan-confirmed striatal hypometabolism, and no serious complicating medical or psychiatric conditions. Patients were evaluated neuropsychologically and with the original Unified Huntington's Disease Rating Scale (UHDRS) prior to transplantation and post-operatively for two years.
Donor fetuses were obtained according to recommendations established by the National Institutes of Health and as described in detail previously . In this patient, four fetuses between 20-32 mm crown-to-rump length (approximate gestational age 9-10 weeks) were used. Whole ganglionic eminence was carefully dissected, and leptomeninges and ependymal lining carefully removed. The lateral ganglionic eminence (LGE) was separated from medial ganglionic eminence (MGE) and immediately divided into 0.8 to 1.0 mm3 pieces. The dissection was performed under sterile conditions in Hank's Balanced Salt Solution supplemented with 10 ug/ml of Gentamicin and 0.25 g/ml of Fungizone (amphotericin B). The LGE pieces were washed 10 times in this solution and stored at 4°C until the time of surgery, when tissue viability of >85% was required in order to proceed.
A detailed description of methods for determination of surgical coordinates and implantation procedures has been previously described . The number of grafts and transplant coordinates were determined immediately prior to surgery using brain MRI imaging (GammaPlan software, Version 2.01, Electa, Sweden). In the patient described here, bilateral craniotomies were performed, and a single piece of fetal tissue (80-120 μL) was stereotactically injected at the appropriate coordinates with concomitant retraction of the cannula. At the most posterior right putaminal implantation site, a small length of the patient's sural nerve was deposited along with a fetal graft. Cyclosporine immunosuppression was instituted pre-operatively at a dose of 15 mg/kg and maintained 12 months after the surgery with routine renal function testing.
Consent for post-mortem examination was performed after informed written consent was obtained by the patient's legal next of kin according to the patient's wishes and according to protocols approved by the Institutional Review Board of the University of Washington. After immersion fixation in formalin for approximately four weeks, the cerebrum was coronally sectioned into approximately 0.5 cm slices while the brainstem and cerebellum were sectioned axially. Representative sections of many brain regions were evaluated while the neostriatum was divided into blocks for thorough analysis. Tissue processing and staining was performed exactly as described previously .
IHC was performed according to published protocols using diamino-benzidine tetrachloride as the chromogen substrate . Primary antibodies were purchased from Chemicon (high molecular weight neurofilament N52, 1:600; microtubule associated proteins MAP-1B, 1:1500 and MAP-2, 1:1500; synaptophysin, 1:400; ubiquitin, 1:20000; myelin basic protein (MBP), 1:1000; acetylcholinesterase, 1:250), Dako (glial fibrillary acidic protein (GFAP), 1:500; low + high molecular weight neurofilament 2F11, 1:1700; calretinin, 1:10; CD20, 1:300; CD68, 1:4000; S100, 1:9000; and Ki-67, 1:125), Sigma (calbindin-D28K, 1:1500; parvalbumin, 1:1500), and Novacastra (CD3, 1:200).
Portions of host putamen, graft, and graft-associated peripheral nerve were processed for ultrastructural analysis by fixation in 3% glutaraldehyde, and post-fixation in 1% osmium tetroxide. After polymerization, 1 micron thick sections were cut and stained with Richardson's stain (methylene blue, sodium borate, azure II) for preview. Ultra thin sections (0.1 um) were cut on selected blocks (Reichert-Jung Ultracut E) and stained with uranyl acetate and lead citrate. Sections were examined using a Philips Electron Microscope (EM410) with high resolution digital camera (AMT XR-400).
Sections were deparaffinized and then denatured in formic acid/H2O2 followed by pretreatment with citrate buffer at 80°C. Protein was digested with pepsin followed by incubation with SpectrumGreen-labeled-X/SpectrumOrange-labeled-Y chromosome probes (Abbott Molecular) in hybridization buffer and Denhyb-2 (Insitus Biotechnologies) overnight. Immunofluorescence was performed on some of the hybridized sections with primary antibodies to GFAP (1:500), S100 (1:900), and MAP2 (1:50), followed by incubation with appropriate secondary antibodies (1:400) conjugated with Cy5 (Jackson ImmunoResearch). All sections were coverslipped with Fluorguard containing DAPI nuclear counterstain (Insitus Biotechnologies).
The subject's father died of HD at age 50 and her brother had HD. She initially became symptomatic at age 27, when she noticed increased clumsiness and difficulty navigating hills and stairs, mild confusion, computational difficulties, and depression. Genetic testing revealed 48 CAG repeats in the HD gene locus. At age 29, while still relatively high-functioning, she received 10 fetal neural transplants: 4 per putamen and 1 per caudate nucleus, bilaterally. In addition, in order to assess the potential neurotrophic benefits of peripheral nerve and nerve sheath, an autologous sural nerve graft was implanted into the most posterior graft site of the putamen. Her post-operative course was uncomplicated, and clinical follow-up was performed at 3, 6, 9, 12, 18, and 24 months post transplantation.
The patient felt positive about her surgery and insisted that she benefited from the procedure, particularly with respect to memory, attention, concentration, and mood. Results of proscribed clinical assessments, including UHDRS rating scales, during the two years of structured clinical follow-up, showed stabilization and improvement in some indices, although a subset of these were beginning to worsen at her final clinical assessment (2 yrs post transplant; see Table 1). After the conclusion of the study in 1998, the patient was followed by her primary neurologist, who in 2001 first noted bilateral upper motor neuron and dystonic features, worse on the right, prompting an MRI in 2003 that showed a 3.1 cm cyst in the left putamen and a similarly sized nodular mass in the right putamen (Figure 1). No signs of increased intracranial pressure were reported. The patient's upper motor neuron symptoms seemed to progressively worsen, but became increasingly difficult to distinguish from coincident HD progression. By six years post transplant the patient exhibited choreiform movements of head, face, tongue, and feet; increased posturing and dystonia of head, face, and right leg; and worsening depression, agitation, paranoia, and two possible suicide attempts that prompted transition to skilled nursing care seven years post transplant. By nine years post transplant she exhibited coarse choreiform movements, severe dysarthria, incoordination, hyperreflexia, painful extensor spasms, repetitive and sometimes violent behaviors, progressive agitation, hallucinations, and possible seizures. She died 121 months post transplant from complications of advanced HD.
The dura had four single sutures overlying the frontal lobes bilaterally. There was no evidence for chronic subdural hematoma. The brain weighed 985 grams and showed mild frontal cortical atrophy. Seven cannula tracts (3 left, 4 right) in the middle frontal gyri were identified without evidence of hemorrhage. Coronal sections revealed multiple lesions (Figure 1; Table 2). Three mass lesions and one large cyst were present in the left caudate and putamen, with marked mass effect on the internal capsule. Five mass lesions and two cysts (one inside a mass) were identified in the right caudate nucleus and putamen, also with prominent mass effect on the posterior limb of the internal capsule. The mass lesions bilaterally were well-circumscribed and surrounded by scant amounts of firm neostriatal tissue and white matter. Although marked atrophy of the host caudate nuclei corresponded to grade IV Huntington's disease grossly , the lateral ventricles were not enlarged. There was slight right to left midline shift (Fig 1).
Histologic examination showed intense gliosis of bilateral caudate nuclei and putamen, corresponding to microscopic grade IV HD . Examination of the lesions revealed circumscribed masses composed of unorganized neuropil with scattered microcalcifications (Fig. 2) and a diffuse, biphasic population of mostly small and some large neurons interspersed with islands of reactive astrocytes and scattered oligodendrocytes (Fig. 3). The masses contained individual and bundled myelinated axons that rarely appeared to cross the graft-host boundary. The grafts were densely immunopositive for synaptophysin (data not shown) and microtubule-associated protein-2 (MAP-2), and focally positive for myelin basic protein (MBP) (Fig. 2). In addition, graft neurons were uniformly MAP-2 positive, and variably immunoreactive for MAP1B and neurofilaments N52 (high molecular weight) and 2F11 (low/high molecular weight). Calretinin neurons were abundant, and sometimes occurred in clusters, while calbindin- and parvalbumin-positive cells were rare. Neither axons nor dendrites convincingly crossed between host parenchyma and graft tissue, although glial fibrillary acidic protein (GFAP) immunostaining showed only patchy transplant border gliosis (Fig. 3). Tyrosine hydroxylase (TH) immunopositive fibers were present, but rare, within the graft tissue, and no definitive crossover of TH fibers was identified (Fig 3). Acetylcholinesterase staining was diffuse and weak within the grafts (data not shown); no clear P-zones were identified. Ubiquitinated neuronal intranuclear inclusions were present in surviving host striatal neurons, but none of these lesions were identified in neurons within graft-derived tissues (data not shown). Tissue autolysis prevented reliable assessment of cell proliferation with Ki-67 antibodies, but no mitotic figures, areas of abnormally increased cellular density, or necrosis was identified in any of the grafts.
The autologous sural nerve graft was identified in the host parenchyma adjacent to the largest mass (Fig. 2), and was characterized by dense collagen and N52- and 2F11-positive unmyelinated axons; no myelinated axons remained in this segment of peripheral nerve. Moreover, while no significant inflammatory infiltrate was identified in any mass lesion, a focally prominent mixed inflammatory infiltrate, composed of CD3, CD20, and CD68 immunoreactive T cells, B cells, and macrophages, respectively, was present in the segment of sural nerve (data not shown). Although in close proximity, no connection between graft or host parenchyma and nerve segment was identified. A small cluster of adipocytes was present within the nerve sheath.
Each cyst was surrounded by piloid gliosis except for the right caudate nucleus lesion, which contained nodular graft neuropil along portions of its wall. The cysts were lined by GFAP-immunopositive, ciliated ependymal cells (Figure 2).
Ultrastructural analysis was compromised by autolysis. However, prominent glial and neuronal intermediate filaments as well as myelinated axons were identified in graft neuropil (data not shown). The peripheral nerve contained abundant collagen fibers surrounding occasional non-myelinated axons. Host putamen revealed prominent gliosis.
All cells in the host parenchyma exhibited an XX karyotype (data not shown). Karyotypic distribution of grafts and cysts is depicted in Table 2 and in Fig 4. The largest mass lesion in the posterior putamen contained a majority of XY chromosome nuclei (172/200 cells counted). Rare XY cells (2/100 nuclei) could be identified in the host parenchyma immediately adjacent to the graft, but no XY cells were identified greater than a fraction of a mm from the graft in the host parenchyma. XX karyotype cells were present within the male grafts, but were most often endothelial cells or occasional astrocytes (data not shown). Although the largest (left putamen) cyst ependyma was XX karyotype, one of the two smaller cysts (right putamen) was lined by karyotypically male ependyma. The autologous sural nerve graft contained only XX karyotype cells. Co-immunofluorescence labeling in the male graft (Figure 4) revealed donor-derived oligodendrocytes (S100), astrocytes (S100, GFAP), and neurons (MAP2).
We show ten year survival of karyotypically confirmed transplanted tissue in a patient with HD who showed initial clinical stabilization and, by some indices, improvement. However, we also show, for the first time, the potential for fetal neural transplanted tissue to exhibit overgrowth resulting in mass effect on critical structures of the CNS. Moreover, contamination with karyotypically-confirmed fetal ependyma likely resulted in the formation of multiple cysts, contributing to mass effect on the internal capsule with progressive impairment.
Early on Peschanski et al (18, 19) discussed transplant overgrowth as a possible complication of neural transplantation in HD. In contrast to other long term survivors of this procedure who have come to autopsy , the singular feature of this case was development of graft-derived mass lesions. Indeed, about 100 microliters of tissue was transplanted at each site, but the largest mass occupied roughly 15 milliliters by the time of death, an approximate 150-fold increase in grafted tissue volume. At the time of surgery, peripheral nerve “co-grafts” for presumed neurotrophic support had shown beneficial effects in host  and transplant tissue  in a non-human primate model of Parkinson's disease, and at least one clinical trial utilizing autologous peripheral nerve and adrenal medulla “co-grafts” was underway . Interestingly, the largest graft-derived mass in our patient occurred in immediate proximity to the implanted sural nerve segment; however, transplant overgrowth also was observed in other grafts in the ipsilateral and contralateral hemispheres.
Other factors that may have influenced proliferation and graft overgrowth include the relative number of neural stem/progenitor cells initially implanted, graft neuropil formation, angiogenesis, latency from transplant surgery to death, and graft-host sex mismatch. The ganglionic eminence contains abundant germinal matrix, which is composed of dense progenitor cell networks, and utilizing this tissue may be an inherent risk of this procedure. The 10 years' latency in this patient is similar to 6-7 years' latency in the prior autopsies that did not show mass lesions , and therefore seems an unlikely explanation for graft overgrowth. The masses were well-vascularized by XX vessels, suggesting angiogenesis from the host, which was likely necessary but not clearly sufficient for transplant proliferation. Finally, since the largest graft was of male origin, the possibility of sex-mismatch overgrowth must be considered; however, this phenomenon has not been reported in other HD transplant patients or in other transplant paradigms. Thus, while many factors may have contributed to graft overgrowth, the exact mechanism may never be determined in this case.
In addition to the mass effect from graft overgrowth, significant compression of the left posterior limb of the internal capsule was mediated by an ependymal cyst in the putamen; two other ependymal cysts also were identified within the neostriatum. Ependymal cysts could be host-derived from ependymal rests or ventricular tissues carried by transplant cannulae to the neostriatum, or donor-derived. Although painstaking efforts were made at the time of dissection to remove all ependymal tissues, the adherence of the ependymal cell monolayer makes this virtually impossible in LGE dissection. One of the smaller intrastriatal ependymal cysts was XY karyotype and therefore donor-derived, but the other two cysts were lined by female cells, making their origin uncertain.
Importantly, although there was slight right-to-left midline shift, clinically the patient's right side was more severely affected, arguing for more significant mass effect from the large ependymal cyst. Compression of surrounding brain by mass lesions was similar bilaterally at autopsy. Since imaging data is only available from 2003, it is unclear when the cystic and solid mass lesions developed in this patient. Although the patient's asymmetric upper motor neuron symptoms were first reported in 2001, symptomatic onset is not known. However, since the relative size of the left putaminal cyst and right putaminal graft in 2003 was similar to the size of the same lesions at the time of autopsy, it is likely that the mass lesions were relatively stable between the time of imaging and death.
In support of this, we found no evidence of ongoing cellular proliferation (mitotic activity). Graft overgrowth is not due to the amount of implanted tissue, since a defined volume (0.8-1.0 mm3) was implanted at each site. Therefore, transient graft cell proliferation waxing from the time of surgery may have resulted in early, robust volumetric expansion, with possible ongoing neuropil production by post-mitotic neurons contributing to overall transplant size. This was likely influenced by initial graft composition, host factors, and possibly the sural nerve co-graft. Regardless, we interpret graft overgrowth in this case as an excess of “normal” graft tissue, rather than a neoplastic proliferation.
Although graft overgrowth and cyst formation are unique to this case among the three previously reported autopsies of patients with HD who received fetal tissue transplants [7,9], some features are consistent among these four patients: grafts can differentiate and survive without immunosuppression for many years, few if any neuronal projections cross the graft-host interface, and no HD-type pathologic changes are adopted by graftderived cells despite advanced changes of HD in the adjacent host tissue. This latter characteristic is particularly important since, while no neuronal intranuclear inclusions were identified in graft neurons in this case, two of three autopsy studies in long-surviving Parkinson's patients identified Lewy bodies within graft-derived neurons [11, 13, 17]. Consistent lack of neuronal projections crossing the graft-host interface in patients with HD is perplexing since connectivity of transplanted tissue occurs in rodent models of HD and autopsy studies of fetal neural transplants into neostriatum of patients with PD show robust host-graft connectivity . Impaired connectivity repeatedly observed in patients with HD may be intrinsic to transplanted fetal LGE. Another possibility is that reduced connectivity in patients with HD may be related to an unfavorable environment created by the processes of HD. It is interesting to note that graft-derived masses with male karyotype contained blood vessels lined by XX cells, suggesting that, in contrast to neuronal projections, host blood vessels readily transit the graft-host interface.
Two prior reports of pathologic outcomes in neural transplant recipients with Parkinson's disease (PD) have been reported [6, 15]; both patients died approximately two years post transplant, and both appeared to culminate in ventricular obstruction, hydrocephalus, and herniation. In the Folkerth and Durso report , inappropriate fetal elements, noted by Kordower et al.  as most likely compromised by “non-CNS structures”, were transplanted into caudate nucleus and putamen as well as the lateral ventricle. Mamelak et al.  reported a patient who received bilateral putamen transplants resulting in the formation of a large left anterior putamen cyst, which they attribute to inadvertently transplanted fetal choroid plexus; however, Ulrisch et al.  refer to additional and unrelated intracranial surgical procedures as possible contributors.
Although only a single case, and despite the lack of connectivity at the time of her death, our patient demonstrated clinical stabilization, and possibly even improvement, in multiple motor, cognitive, and behavioral assessments within two years of her surgery. Graft-mediated trophic support of diseased basal ganglia may have contributed, but whether this trophic support continued until her death and was simply overwhelmed by the other pathologic processes, or was transient, cannot be determined. It is important to note that this patient received her transplants early in the course of her disease and at a young age, which may have contributed to her apparent initial improvement.
The results of this investigation are both encouraging and cautionary. The patient's perceived initial clinical improvement made her memorable to those involved in the original trial. It is also encouraging that we can again confirm long-term fetal LGE graft survival and appropriate differentiation without evidence of immune rejection or adoption of intrinsic pathology, including neuronal intranuclear inclusions. However, consistent lack of evidence for long term connectivity with host brain raises serious challenges for this approach. Importantly, this case highlights the need to identify those factors that underlie the wide variation in graft (over) growth among transplant recipients.
The authors wish to thank Randi Small, Regina Bowman, Susan Rozell, Christiane Ullness, Lynne Greenup, Aimee Schantz, Chiyen Miller, and Mike Hobbs for expert technical support. We would also like to thank Dr. Michael Laflamme for helpful discussions. This research was supported by National Institutes of Health grants P50-AG005136 and 5T32-AG000238, Veterans Affairs research funds, the University of Washington Huntington's Disease Society of America Center of Excellence, and the Nancy and Buster Alvord Endowment.