The present study showed that in situ continuous delivery of low doses of CNTF by encapsulated transgenic cells restores cognitive and motor functions specifically impaired in a primate model of progressive striatal degeneration. This striking result might stem from the dual properties of CNTF. We show that, first, CNTF exhibits the neuroprotective properties previously described in acute lesion models (Emerich et al., 1996, 1997) and actually reduces the progressive striatal neuronal loss induced by 3NP.

Moreover, we clearly demonstrate in nonhuman primates the neurorestorative properties of CNTF (Cayouette et al., 1998) on a dysfunctional striatal circuitry, in a chronic degenerative paradigm particularly relevant to HD gradual neuropathology.

Direct and indirect mechanisms of CNTF induced neuroprotection

The cellular mechanisms by which CNTF may display such neuroprotective and neurorestorative effects are not yet fully elucidated.

However, concerning the neuroprotective effects of this cytokine, the present results rule out any direct modulatory action of CNTF on SDH activity (Michel et al., 1996) as well as any interaction between 3NP and CNTF at the level of SDH (the target enzyme specifically inhibited by 3NP; Coles et al., 1979). As the neuroprotective effect of CNTF has also been observed in an acute lesion model using a toxin that does not interfere directly with SDH (Emerich et al., 1996, 1997), all together these data indicate that mechanisms of neuroprotection by CNTF are not mediated by an alteration of the 3NP induced SDH inhibition.

We also demonstrated a partial neuroprotection of cortical NFP immunoreactive neurons in the CNTF treated animals. CNTF intraparenchymal diffusion and direct effect on neurons around the site of delivery likely explain the mechanisms of neuroprotection observed in the striatum. However, the partial neuroprotection observed in cortical neurons is unlikely to occur through a diffusion mechanism.

It may therefore be speculated that the partial neuroprotection of cortical neurons is more likely due to the preservation of the striatal targets, as previously suggested after a similar analysis in an acute lesion model of striatal degeneration in the nonhuman primate (Emerich et al., 1997).

In the same way, CNTF has been shown to prevent apoptosis in striatal neurons expressing the mutated huntingtin in vitro (Saudou et al., 1998). The development of transgenic mice expressing mutated huntingtin (Mangiarini et al., 1996; Hodgson et al., 1999) raised the interesting possibilities of testing CNTF neuroprotective effects in a genetic animal model.

Restoration of neuronal functions by CNTF

The neurorestorative effect of CNTF observed in the present study is of major interest, but the underlying mechanisms by which this cytokine elicits a neuronal functional recovery are, at this moment, still a matter of speculation.

However, several known effects of CNTF may provide some clues (for review, see Segal and Greenberg, 1996; Ip, 1998). The main targets of CNTF seem to be the immediate-early genes, which are activated via the Jak-STAT pathway (Hirano et al., 1994). Among them is the nociceptin/orphanin FQ gene, which is upregulated by CNTF in striatal neurons (Buzas et al., 1999) and is able to reduce their sensitivity to glutamate (Shu et al., 1998) as well as their intracellular calcium concentration (Knoflach et al., 1996). Nociceptin and orphanin are structural opioid-like neuropeptides (Reinscheid et al., 1995) and, like opioids enantiomers, are able to reduce N-methyl- D-aspartate (NMDA)-induced neuronal toxicity (Choi and Viseskul, 1988).

Glutamate-mediated indirect excitotoxicity and calcium cellular overloading are both implicated in the 3NP-induced toxicity as well as in HD neurodegenerative processes (Lipton and Rosenberg, 1994; Schulz et al., 1997). Therefore, the neurorestorative effects of CNTF reported here could be due to a reduction in calcium neuronal overload and in NMDA-mediated neurotoxicity via overexpression of such opioid-like messengers.

In addition, or alternatively, CNTF may also facilitate neurotransmission in striatal neurons in vivo by increasing synaptic efficacy, as has been shown in vitro (Stoop and Poo, 1995, 1996). Through this mechanism, CNTF could induce recovery of the synaptic function in striatal projection neurons, and thereby restore basal gangliam ediated functions. Moreover, Johnson et al. (1994) have reported that the binding of CNTF to its receptor induces significant increases in cellular metabolism . Such a metabolic effect is of particular interest since the 3NP model as well as HD pathology have been associated with both chronic neuronal energy impairment and deficiency in the electron transport chain (Beal et al., 1993; Brouillet et al., 1999).

Finally, CNTF may also have an indirect effect on striatal neurons via the astrocytes, which are known to be responsive to this cytokine (Clatterbuck et al., 1996; Lisovoski et al., 1997) and to play a key role in sustaining neuronal activity and metabolism.

In conclusion, the present study demonstrates that sustained intracerebral CNTF delivery may have a strong therapeutic effect in HD patients, not only by opposing the degenerative process but also by inducing recovery of motor and frontostriatal functions in presymptomatic and early-stage HD patients.

In contrast to systemic injections of CNTF (Miller et al., 1996a,b), the continuous delivery of low doses of this cytokine by genetically engineered cells intracerebrally implanted in the immediate vicinity of target structures precludes side effects.

This delivery protocol has already been shown to be safe in patients suffering from amyotrophic lateral sclerosis (Aebischer et al., 1996a,b).

To be developed as an effective clinical treatment for HD, the macroencapsulation technique described here needs further refinements, in particular to overcome the progressive downregulation of CNTF production observed with time.

Since these experiments were concluded further work has been done leading to the establishment of longerterm (6 months), reliable, and reproducible cell viability using genetically engineered C2 C12 cells. In addition, further primate studies have also demonstrated the feasibility and safety of the frequent replacement of capsules implanted for more than 3 months into the brain parenchyma.

All together, these data demonstrate the feasibility of this approach as a means of achieving direct continuous intracerebral delivery of therapeutically relevant doses of CNTF.

Phase I clinical trials in HD patients have now been initiated using CNTF as a neurotrophic/ neuroprotective agent and the macroencapsulation technique as a gene delivery system.

This work was supported by grants from the Commissariat à l’Energie Atomique, the Centre National pour la Recherche Scientifique, the Institut National pour la Recherche Agronomique, INSERM, the Association Huntington-France, the National Institutes of Health, the Mount Sinai School of Medicine, and European Community Neuroget grant no. QL K3-CT-1999-00702.