Loss of dopaminergic neurons in Parkinson’s disease (PD) and PD animal models has been extensively documented to cause global changes in electrophysiological activity throughout the cortico-basal ganglia network. However, such loss is also associated with a range of morphological alterations of neurons forming this network, most notably the medium spiny neurons (MSNs) that are the main output neurons of the striatum. MSNs are the primary synaptic target of both nigrostriatal dopaminergic and cortico-striatal glutamatergic afferents. Both afferents converge onto dendritic spines, the critical site for synaptic integration in MSNs. In advanced PD there is a marked atrophy of dendrites and spines in these neurons, indicative of dysfunctional signal integration in the striatofugal pathway. Similar pathology, triggered by a dysregulation of intraspine Cav1.3 L-type Ca channels (Day et al., 2006), has been observed in rodent and primate models of PD (Day et al., 2006; Neely et al., 2007; Scholz et al., 2008). The significance of such dendritic atrophy and spine pruning for the symptoms and the treatment of PD has remained poorly understood. However, there is increasing awareness that these morphological alterations represent a major obstacle for therapeutic approaches to enhance striatal function (Schuster et al., 2009). Most notably, the efficacy of dopamine cell replacement strategies, designed to restore nigrostriatal connectivity, may be hampered by striatal dendritic and spine atrophy. In order for grafted dopamine neurons to re-establish functional connections, the morphological target of such reinnervation would need to be preserved or reestablished. In this issue of EJN, Soderstrom et al. (2010) report the results of a study on the impact of dendritic spine preservation in MSNs upon both antiparkinsonian and prodyskinetic effect of fetal mesencephalic cell grafts. The authors elegantly and convincingly show that administration of the L-type Ca channel blocker nimodipine prevented loss of spines in MSNs in unilaterally lesioned rats that were grafted with embryonic ventral midbrain cells. Nimodipine treatment also resulted in improved therapeutic benefit and reduced graft-induced behavioral abnormalities of these hemi-parkinsonian rats. Specifically, the results indicate that graft-mediated anti-parkinsonian efficacy was not potentiated by the prevention of spine loss; however, the impact of the graftand levodopa-induced side-effects was greatly diminished by nimodipine treatment. Interestingly, these effects were not due to increased survival of grafted cells but correlated with a greater reinnervation of the affected striatum. These results underscore the importance of prevention (or reversal) of spine loss in striatofugal neurons for effective therapy based on dopamine cell replacement. They extend a previous report of reduced levodopa-induced dyskinesia by prior treatment with L-type Ca channel antagonists (Schuster et al., 2009). The results described in Soderstrom et al. (2010) suggest that unless MSN spine loss and dendritic atrophy are reversed by appropriate pharmacological treatment, therapeutic interventions may be of limited efficacy or even cause unwarranted outcome. The findings and conclusions from the study by Soderstrom et al. (2010) need to be taken into consideration when designing dopamine cell replacement therapies in PD, using embryonic or stem cell-derived neurons in combination with appropriate drug cocktails to block MSN atrophy.