Bilateral human fetal striatal transplantation in Huntington’s disease

Methods.

Subject selection, enrollment, and evaluation were performed according to a modified version of a standardized protocol (Core Assessment Program for Intracerebral Transplantation in Huntington’s Disease [CAPIT-HD]).26 Consecutive patients who met eligibility criteria and who were willing to provide informed consent were enrolled in the study. Subjects were required to meet inclusion and exclusion criteria at entry and before surgery. Functionally independent adults (aged >18 years) with neurologic signs consistent with HD and with genetic confirmation (trinucleotide repeat length >37)27 were eligible to participate. Exclusion criteria included dementia that would preclude providing informed consent and clinically remarkable medical or laboratory abnormalities (including serum creatinine concentration >1.8 mg/dL, or HIV-1, HIV-2, or HTLV-1 antibody positive). All subjects provided written informed consent before entering the study, before PET scanning, and before each surgery.

Subjects underwent clinical evaluations every 3 months for 1 year before surgery (−12,−9, −6,−3, and 0 months) and at 1, 3, 6, 9, and 12 months after surgery. Attempts were made to maintain medications unchanged throughout the study. Clinical evaluations included the Unified HD Rating Scale (UHDRS)28 and a battery of neuropsychological tests. Alternate forms of neuropsychological tests were used when available. All neuropsychological raw scores were converted to z scores for analysis to correct for age, sex, and education, where available.

Subjects underwent PET scanning to assess regional glucose metabolism and dopamine D1 and D2 receptor binding before and 6 to 18 months after surgery. All scans were performed on a Siemens/CTI 953B tomograph (Knoxville, TN) in three-dimensional mode, with an in-plane resolution of 5.6 mm and a slice thickness of 3.4 mm. Reconstruction with corrections for scatter and normalization was achieved as previously described.29

Dopamine D1 receptor binding was assessed following injection of 185 MBq of [¹¹C]SCH 23390 (specific activity > 37 GBq/mmol) administered by IV bolus over 60 seconds. Dopamine D2 receptor binding was assessed following the injection of 185 MBq of [¹¹C]raclopride (specific activity > 70 GBq/mmol) administered by IV bolus over 60 seconds. Sixteen PET scans were obtained over 60 minutes, starting at the midpoint of tracer injection. Regional glucose metabolism was determined using [¹⁸F]2-fluoro-2-deoxy-D-glucose (FDG). A total of 185 MBq of FDG was administered via an IV cannula over 60 seconds. Beginning 40 minutes after FDG administration, two PET scans were obtained, each of 10 minutes’ duration.

Fetal striatal tissue was transplanted stereotactically in two staged procedures. Subjects received cyclosporine 6 mg/kg/day beginning 7 days before the first operation and continuing through 14 days after the second (contralateral) operation. The cyclosporine dose was then reduced to 2 mg/kg/day for 6 months and then discontinued. Serum creatinine concentration was monitored monthly and cyclosporine dosages were reduced if creatinine concentration increased above normal.

Embryonic tissue was obtained in accordance with state, federal, and NIH guidelines and stored as previously described.30,31⇓ The developing striatum was dissected from the lateral half of the lateral ventricular eminence of donors 8 to 9 weeks postconception.12,15,16⇓⇓ Each striatum was dissected into 0.5-1 mm³ pieces and deposited along a needle tract 9 to 15 mm in height. Deposits were separated by up to 5 mm in a three-dimensional array. Two to eight donor striata were implanted into each host striatum. The amount of tissue implanted depended on donor tissue availability. The primary goal was to transplant into the “motor” striatum (postcommissural putamen).30 Transplantation into striatal areas associated with cognitive function (caudate nucleus and precommissural putamen) was also performed if sufficient tissue was available.4,20,32⇓⇓

Beginning at the time of surgery, subjects received antibiotics to provide prophylaxis against routine vaginal flora.33 These included piperacillin 3 grams IV every 6 hours, vancomycin 1 gram IV every 12 hours, and fluconazole 100 milligrams by mouth every day, for 3 days. Donor tissue samples and all reagents were cultured to determine the need for additional antibiotic therapy.

Experimental surgery and hospitalization were provided without cost to the subjects. Costs to subjects were approximately $20,000 to cover nonfunded aspects of the study, including travel expenses for subjects and spouses, PET and MRI scans, cyclosporine, and home health administration of antibiotics.

Results.

Seven subjects (six women, one man) underwent transplantation and have been followed up for 12 months. An additional subject entered the clinical evaluation phase of the study but did not meet eligibility criteria for surgery because she became nonambulatory and developed dementia that precluded providing informed consent for surgery.

For the seven subjects undergoing transplantation, mean (±SD) expanded CAG repeat length was 44.4 ± 4.2 (range, 42 to 53; table 1). Age at time of transplantation was 50.1 ± 12.4 years (range, 28 to 64 years) and time from diagnosis to first surgery was 5.9 ± 3.5 years (range, 2 to 12 years). Subjects received 4.5 ± 1.2 (range, 2 to 6) striata implanted into each putamen and 1.1 ± 0.5 (range, 0 to 2) striata implanted into each caudate nucleus. All subjects had the second procedure within 4 weeks of the first procedure, except Subject 6, for whom 14 weeks elapsed between surgical procedures owing to lack of tissue availability.

None of the subjects received dopamine antagonists or depletors during the study. Three subjects (Subjects 2, 6, and 7) received selective serotonin reuptake inhibitors (SSRIs) at stable doses (fluoxetine 40 mg/day, sertraline 200 mg/day, fluoxetine 20 mg/day). Subject 1 was initially taking fluoxetine 20 mg/day; she discontinued taking it 6 months before her first operation. Subject 5 was initially taking trazodone 50 mg/day and the dosage was increased to 150 mg/day 9 months after second surgery.

Efficacy.

Results of neurologic and neuropsychological evaluations are presented in tables 2 through 4⇓⇓. Mean UHDRS motor scores were 32.9 ± 6.2 at baseline and 29.7 ± 7.5 at 12 months (p = 0.24). The mean annual rate of UHDRS motor score worsening was 5.7 ± 5.3 units per year pretransplant and 1.8 ± 8.9 units per year post-transplant (p = 0.31; figure 1). Mean total functional capacity (TFC) scores were 6.6 ± 1.0 at baseline and 7.0 ± 2.2 at 12 months (p = 0.46). The mean annual rate of TFC score change was −1.1 ± 1.7 units per year pretransplant and +0.25 ± 1.3 units per year post-transplant (p = 0.06; figure 2).

View this table:

• View inline
• View popup

Table 3.

Neurologic scores at baseline and 12 months after transplantation

View this table:

• View inline
• View popup

Table 4.

Neuropsychological tests at baseline and 12 months after transplantation

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Individual subject Unified Huntington’s disease Rating scale (UHDRS) motor scores before (left) and after (right) transplantation. Mean (heavy) line calculated using mean slopes and intercepts from least squares line for each subject. Black diamonds = Subject 1; black squares = Subject 2; black triangles = Subject 3; open squares = Subject 4; open circles = Subject 5; black circles = Subject 6; open triangles = Subject 7.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Individual subject total functional capacity (TFC) scores before (left) and after (right) transplantation. Mean (heavy) line calculated using mean slopes and intercepts from least squares line for each subject. Black diamonds = Subject 1; black squares = Subject 2; black triangles = Subject 3; open squares = Subject 4; open circles = Subject 5; black circles = Subject 6; open triangles = Subject 7.

A post-hoc subgroup analysis was performed excluding Subject 5, who experienced cognitive and motor deterioration after the development of symptomatic bilateral SDHs. In the remaining six subjects, UHDRS motor scores were 33.8 ± 6.2 at baseline and 27.5 ± 5.2 at 12 months (p = 0.03).

Discussion.

SDHs were the most notable adverse events observed in this study. Two subjects developed a total of three SDHs that required surgical evacuation out of a total of 14 transplant procedures. Intraoperative CSF loss causes brain “shrinkage” and shift in patients with cerebral atrophy.5 This leads to targeting difficulties6 and expands the subdural space, thereby predisposing to postoperative SDHs. This safety profile differs dramatically from similar transplant procedures in subjects with advanced PD who typically lack the degree of atrophy seen in even moderate HD. In the same surgeon’s experience (T.F.), no surgically relevant SDH occurred in 66 PD transplant procedures. The incidence of SDH after transplantation in HD might be reduced in the future by the use of drains placed at the time of surgery or the intraoperative use of fibrin glue to cover the burr hole and minimize CSF loss during surgery. However, the risk of SDH may be directly related to the degree of cerebral atrophy. This poses substantial safety concerns and leads us to believe that if further trials are undertaken they should be performed in patients with less atrophy.

One patient died 18 months post-transplantation of an apparent cardiac arrhythmia in the setting of underlying severe three-vessel coronary artery disease and aspiration pneumonia. His death was most likely unrelated to transplantation, although we cannot exclude the possibility that surgery could have worsened swallowing function, leading to aspiration and resulting in an arrhythmia. An accurate and reliable assessment of the effect of transplantation on mortality would require a large, randomized, controlled, and blinded study.

We have demonstrated that fetal striatal tissue transplantation in HD is possible and that transplanted cells can survive for up to 18 months after transplantation and 12 months after discontinuation of cyclosporine.6 In the patient who died, clearly demarcated grafts of typical developing striatal morphology were seen on autopsy, with host-derived dopaminergic fibers extending into the grafts. There was no evidence of immune rejection or abnormal expression of mutated huntingtin protein within the fetal allografts. Thus, the disease process did not appear to adversely affect the grafts.

It is not clear that cyclosporine is necessary for graft survival. We have demonstrated survival of transplanted human embryonic cells for 12 months after cyclosporine discontinuation in both HD and PD.6,23-25⇓⇓⇓ At the doses used, cyclosporine appears to be well tolerated, although creatinine concentrations must be monitored and cyclosporine dosages adjusted as appropriate.

HD is associated with progressive clinical decline. One study36 found a worsening in UHDRS motor scores of 5.97 units per year and a mean rate of decline in TFC scores of 0.56 units per year in a group of 78 patients with HD with a mean age of symptom onset of 37.7 years and a mean duration of illness of 6.7 years. Another study37 found a mean annual rate of TFC score decline of 0.97 units per year in 575 patients with HD with baseline TFC scores of 7 to 13.

In our subjects, no significant changes were observed in UHDRS motor scores from baseline to 12 months. However, a lack of significant worsening might reflect clinical benefit in a progressive neurodegenerative disease. Nonetheless, we did not definitively demonstrate a lack of significant worsening owing to the small number of subjects and the open-label design of the study. We also cannot exclude the influence of random chance, bias, or placebo effects. Comparisons to historical control subjects are of limited usefulness. This problem is compounded by the fact that worsening is not linear throughout the disease. Patients in earlier stages decline faster than patients with more advanced disease.37

We also performed a post-hoc subgroup analysis excluding Subject 5. This subject and her husband reported that she initially improved after her second transplant procedure but experienced motor and cognitive deterioration after a fall that resulted in the development of symptomatic bilateral SDHs. She improved partially after evacuation of SDHs but did not return to baseline status and subsequently exhibited the worst clinical course of the seven subjects. It is our impression that her relatively poor outcome was related to her SDHs rather than to the transplantation per se. If the transplantation procedure can be modified to avoid SDHs in the future, the group results excluding this subject may be more representative of the potential clinical outcome that might be achieved with striatal transplantation. In this group of six subjects, UHDRS motor scores were 33.8 ± 6.2 at baseline and 27.5 ± 5.2 at 12 months (p = 0.03). However, this is a post-hoc analysis and we cannot be certain that the excluded subject would have done better if she had not experienced SDHs.

PET studies of patients with HD have found that striatal glucose metabolism and dopamine receptor binding decrease. One study38 found a mean annual decline of nearly 7% in striatal glucose metabolic activity and another39 found a mean annual decline of 5% in striatal D1 binding. In our study, maintenance of lentiform metabolic activity and putamenal D1 binding may suggest survival of region-appropriate transplanted neurons, but this remains to be proven. In the caudate nucleus, lack of maintenance of glucose metabolism and D1 and D2 binding may be a function of the limited quantity of tissue transplanted into this area. The reason for the lack of maintenance of putamenal D2 binding is unclear but may potentially relate to selective survival of transplanted neurons or differences in the time course or capacity for expression of these receptors.

Other investigators have reported clinical benefit after neural transplantation in HD.32,40,41⇓⇓ One study32 reported motor and cognitive improvement or maintenance and increased striatal glucose metabolism on PET in three of five transplanted patients at 2 years. However, group analyses were not performed. Another study40 reported some improvements in measures of visuoperceptual/conceptual skills, learning and memory, and executive functions in three patients 6 months after transplantation. It is difficult to compare clinical results across these studies because of their descriptive nature and the small number of subjects. Generally, in both series some patients improved and some worsened on various clinical measures and there was no consistent pattern of change across subjects. However, the association of clinical improvement with increased metabolic activity on PET as observed in one study32 suggests that the improvements may be graft related.

There were several important methodologic differences between our transplantation procedure and those of other investigators.32 We transplanted tissue from a very restricted dissection of the developing striatum and only utilized the far lateral aspect of the lateral ventricular eminence.6,16⇓ In an effort to implant sufficient tissue, two to eight developing striata were transplanted per side. Other investigators used a nonselective dissection consisting of the entire ventricular eminence, which is much larger and includes not only the developing striatum but also numerous other brain regions.12 Owing to the comparatively large size of the whole ventricular eminence, these investigators transplanted approximately one striata per side. It is not known what the impact of these differences in dissection may be on mechanism (i.e., connectivity with host vs graft-induced trophic effects) or clinical outcome.4,16⇓ Despite our use of multiple donors, only 8 to 10% of the striatal region transplanted in our autopsy case was occupied by graft material.6 We have not found it possible to obtain eight or more satisfactory33 fetal striatal grafts on a consistent basis. If a large amount of graft tissue is required for optimal results, less specific dissection techniques or striatal cell proliferation methods will be required.

We also distributed tissue differently than other investigators.32 In our trial, tissue was transplanted predominantly into the postcommissural putamen, which is primarily involved in motor circuits.20,30⇓ Additional tissue was placed in the anterior putamen and caudate nucleus if available. These areas are primarily associated with cognitive and neuropsychological function.4,19,20⇓⇓ In contrast, other investigators32 distributed approximately 80% of transplanted tissue into the precommissural putamen and caudate nucleus. It may be important to transplant sufficient tissue into both the anterior and posterior striatum to optimize clinical benefit.

We speculate that striatal transplantation may provide a better benefit to risk ratio if performed earlier in the disease. From a safety perspective, we expect this would reduce the risk of SDH and improve surgical targeting accuracy related to atrophy-associated intraoperative CSF loss.6 Early grafting may also improve efficacy because less gliosis may provide a better “transplant milieu” for graft survival. In our HD subject who died 18 months after transplantation, autopsy revealed grade II disease42 and only 6 of 10 transplant tracts survived, even though there was no evidence of immune rejection.6 In comparison, autopsies revealed 100% of grafts survived in our two subjects with PD who died.23-25⇓⇓ Grafts may also provide trophic support to the host brain to prevent or slow further striatal and second-order degeneration in remote sites.4,6,16,43-45⇓⇓⇓⇓⇓ This potential but unproven benefit would likely be most valuable before extensive neuronal loss has occurred.