The recent publication by Freed and colleagues1 of the first double-blind trial of fetal transplantation for Parkinson's disease (PD) has generated controversy in both the scientific and popular press. The results have been widely interpreted as disappointing, but we believe that many of the criticisms are misplaced. Here we present a meta-analysis of several published studies, from which we conclude that the results are in fact quite promising. We also suggest a plausible explanation for the side effects that were seen in the Freed study.

The progression of Parkinson's disease can be monitored by the Unified Parkinson's Disease Rating Scale (UPDRS), and a comparison with other studies2,3,4,5,6,7 shows that the UPDRS motor scores in the Freed study improved at a rate similar to what has been previously reported elsewhere (Fig. 1a). The double-blind protocol was concluded after only one year, although subsequent follow-up indicated that many of the patients continued to improve after that time, consistent with other studies. The Fluoro-dopa uptake measurements demonstrate that the grafted dopamine neurons survived for at least one year, and that the level of F-dopa uptake was comparable to previous studies (Fig. 1b). This does not mean, however, that the clinical improvement was at a maximum; F-dopa can be taken up by immature neurons, so the increase in uptake that occurs after grafting usually precedes the functional maturation and integration of the graft by months or even years. Consistent with this, a time lag has been demonstrated between the increase in F-dopa uptake and improvement of motor symptoms6,7.

Figure 1: Results from Freed et al. (ref. 1 and C. Freed, personal communication) compared with previous transplantation studies.
figure 1

(a) In the first year, UPDRS motor scores improve to a similar extent in all studies. No comparisons can be established beyond the first year, because Freed et al. reported only global UPDRS scores for their three-year follow-up. (b) Increase in Fluoro-dopa uptake in the grafted putamen was slightly (20%) lower in the Freed et al. study than in previous reports.

Full size image

Freed and colleagues1 reported that 15% of their patients developed dyskinesias, whereas several other exploratory studies, in a total of about 30 patients, have shown significant clinical benefits of transplantation with no such side effects5,6,7,8. This may be a consequence of the technique used by Freed and colleagues, which differs from the other studies in many respects. For example, Freed and colleagues used a transcallosal surgical approach, and administered L-DOPA throughout the period of transplant growth. Moreover, they used a different procedure to prepare their cells. Some studies3,4,5,6,7 used either small tissue fragments or suspensions of cells that had been freshly dissociated using proteolytic enzymes. In contrast, the so-called 'noodle' cell preparation technique used in the Freed study includes a cell culture step that was requested by NIH reviewers to provide a measurement of dopamine production before implantation. The culture conditions might alter the cells and/or select for particular cell types, leading to a population that differs from that obtained with freshly dissociated preparations.

In most current transplantation studies, about 80–90% of the transplanted cells fail to survive the implantation process, and further selection may occur at this stage. The fragments of fetal tissue from which grafts are derived contain about 10% dopamine neurons, and it is known that these neurons are heterogeneous. For example, the A10 dopamine neurons from the midline region (which are calbindin positive) have different functional specializations and outgrowth patterns than the neighboring A9 neurons. It is the A9 neurons that degenerate in Parkinson's disease, and these neurons are also more vulnerable to 'dopaminotoxic' drugs such as MPTP9. It is therefore plausible that they are selectively lost during transplantation; however, current methods do not distinguish what types of dopamine neurons are transplanted. Differential survival of functional subsets of dopamine neurons could therefore explain some of the observed variation of outcome reported by different groups.

In summary, cell transplantation therapy is still a very novel approach, and current studies should be considered as 'works in progress.' There is further scope for optimization at many levels, including cell preparation and transplantation, study design, and evaluation of both clinical outcomes and underlying neurobiological mechanisms. We remain optimistic that a combination of laboratory experiments and clinical trials will eventually bring this promising approach to fruition.