We observed a marked increase of BrdU+ cell numbers in young mutant mice relative to control mice in the GFP-global and striatal lines. We saw an increase in the number of BrdU+ cells in brain regions, predicted on the basis of the Cre activator line used, to sustain maximal levels of tox-176-mediated death of Drd1a-expressing cells. These regions are the cortex and striatum (including the periventricular region) of GFP-global line mutant mice, and the striatum (including the periventricular region) of striatal line mutant mice. As BrdU is a marker of DNA synthesis , our study suggests that the brain produces new, dividing cells in an attempt to compensate for the regionally restricted tox-176-mediated loss of Drd1a-expressing cells. A limitation of BrdU as a proliferative marker is that it not only incorporates into nascent DNA during S phase of the mitotic cycle but it may also integrate into cellular DNA during DNA repair and in the context of apoptosis. Although both DNA repair and apoptosis represent theoretical possibilities, this would not explain the uniform nuclear staining, the number of BrdU+ doublets, and more importantly the large number of BrdU positive cells (~ 720 cells/5 sections-see Figure 2A and Figure 1D-F) seen in global ablation mutant mice. Furthermore, TUNEL staining undertaken in global ablation mutant mice  aged between 2 and 9 weeks identified very few TUNEL positive cells. In addition, TUNEL-positive cells were not seen in the cortex. A small number of TUNEL-positive cells (approximately five/section) were seen in the hippocampus and thalamus of a two-week old mutant, brain regions known to express the Drd1a-receptor. The degree of striatal volume loss is no different in the striatal-restricted line and the mechanism of cell death is the same (i.e. tox-176 mediated cell death) and so there is no reason to believe that apoptosis would be any more prominent in this line. Taken together, these data suggest that it is highly improbable that the large number of BrdU-positive cells seen in our mutant lines represent anything other than active cell turnover in response to the targeted attenuated diphtheria-toxin mediated cell death of Drd1a-expressing cells.
There was also a marked increase in the number of dividing microglia/macrophages in young mutant mice belonging to the GFP-global and striatal lines. This increase was confined to areas of the brain targeted by tox-176-mediated Drd1a-expressing cell death. The robust regional increase in the number of newly synthesized microglia is likely to reflect an inflammatory response elicited by the primary death of Drd1a-expressing cells. Microglia, with their elaborate processes, respond to the presence of necrotic tissue by phagocytosing and removing debris [10–13]. Microglia have other roles including modulation of synaptogenesis and neurogenesis , processes that may be particularly relevant in our models of neurodegeneration.
There was a marked enhancement of the number of dividing astrocytes in the cortex and periventricular region of young mutant mice belonging to the GFP-global line although, no such enhancement was observed in mutant mice belonging to the striatal line. Astrocytes are known to participate in long-term brain repair and maintenance [15–18]. Their functions include the release of trophic factors that influence neurite growth, and the formation of a glial scar through the process of reactive gliosis. The survival of neurons is known to be dependent on astrocyte functions such as free radical scavenging and glutamate uptake/release.
In order to determine why a pleiotropic proliferative/repair response appears to have been activated in mutant mice from the GFP-global line but not from striatal line mutant mice, it is worthwhile comparing the time of onset of death of Drd1a-expressing cells in each line. The GFP-global line is known to undergo death of Drd1a-expressing cells at an earlier stage than the striatal line. The death of these cells commences at about 1–2 weeks of age in the GFP-global line, following Cre activation [3, 6], and 5 weeks of age in the striatal line . Our GFP-global mutant and GFP-control mice, therefore, had been subjected to several weeks of cell death prior to the commencement of the BrdU injection regime at 4 weeks of age. In contrast, the death of Drd1a-expressing cells would not have commenced in striatal line mutant mice until 5 weeks of age, which is at a point halfway through the 2-week BrdU injection regime administered to these animals commencing at 4 weeks of age. The death of these cells would, therefore, have been relatively advanced in GFP-global as compared with striatal mutant mice. It is conceivable that the enhanced number of dividing astrocytes in the mutant mice belonging to the GFP-global line reflected the comparatively early time of death of the Drd1a-expressing cells in this line, which allowed a greater time for a proliferative astrocyte response to develop.
The presence of BrdU+/NeuN+ double positive cells in the cortex, striatum and periventricular regions of young control and mutant mice belonging to the GFP-global and striatal lines confirmed that neurons were generated from dividing cells in the post-natal period in all lines. The number of BrdU+ neurons was increased relative to control mice in the striatum and periventricular region of GFP-global mutant mice but not in striatal line mutant mice. These results are consistent with published reports indicating that while constitutive neurogenesis occurs throughout the post-natal brain , albeit at low levels [20–22], enhanced neurogenesis in response to the loss of Drd1a-expressing cells in our study was restricted to the GFP-global line. This result would appear to correlate with a comparatively subtle orofacial phenotype observed in global line mutant as compared with striatal line mutant mice . A caveat for the observed difference between the GFP-global and striatal lines may be that the temporal profile of Cre-expression and therefore tox-176 mediated death of Drd1a-expressing cells in the striatal line was incompatible with a fully evolved proliferative response to become manifest within the time frame of the study design.
The generated NeuN+ cells are likely to do one of two things. First, they may differentiate into Drd1a-expressing striatal medium spiny GABAergic projection neurons or Drd1a-expressing cortical glutamatergic corticostriatal projection thereby replacing cells that are killed by tox-176 expression. Owing to their molecular signature, these cells are likely to ultimately suffer the same fate as the cells they replace. Second, the generated cells may differentiate into Drd2-expressing striatal projection neurons, non-Drd1a-expressing striatal interneurons or cortical non-Drd1a-expressing pyramidal cells. Expansion of the non-Drd1a-expressing compartments would thereby functionally compensate for the primary loss of Drd1a-expressing cells. We have evidence for up-regulated D2-compartment expression for the global-ablation mutant mice [3, 24, 25] lending support to this compensation hypothesis. An experiment quantifying BrdU+ cells co-expressing Drd2 in young BrdU injected global-mutant mice on a D2-GFP reporter background would directly address this question.
The number of dividing Olig2+ cells was increased in the cortex and striatum of young mutant mice from the GFP-global line but not young mutant mice from the striatal line. Olig2 is a transcription factor found in cells that differentiate into oligodendrocytes . The differentiated cells supply the axons of neurons with their myelin coats, an important function that correlates with neurogenesis and neural maturation/remodelling as newly generated axons require myelin . It is conceivable that our finding of enhanced numbers of dividing Olig2+ cells in the mutants from the GFP-global line reflects the necessary partnership between the generation of new oligodendrocytes and the generation of new neurons, neuronal processes and neuronal interconnections in these animals.
The expression of Olig2 is not confined to the differentiation path that produces mature oligodendrocytes. Olig2+ cells may also differentiate into neurons [28, 29], ependymal cells  and astrocytes [31, 32]. It is conceivable that dividing Olig2+ cells in the cortex and striatum of the mutants from the GFP-global line may ultimately differentiate into neurons or astrocytes in these regions of the brain, and thus contribute to the observed up-regulation of these types of cells in response to tox-176-mediated death of Drd1a-expressing cells.
In the cortex of control and mutant mice for the GFP-global and striatal lines, virtually 100% of the dividing cells were identified on the basis of marker expression, as microglia, astrocytes, neurons or oligodendrocytes. However, in the striatum and to a greater extent in the periventricular region, a large percentage of the dividing cells were negative for all markers suggesting that an increasing proportion of cells were progenitor cells that remained uncommitted to any particular lineage. Judging from the distribution profile of identifiable dividing cells in our study, it would appear that new cells that were manufactured in the sub-ventricular zone of the periventricular region divided and differentiated as they migrated through the striatum and to the cortex. By the time they reached the cortex, differentiation was more or less complete, and thus based on co-expression of lineage specific markers, all BrdU+ cells were readily assigned a phenotype. We propose that some of the phenotype negative BrdU+ cells in the striatum and periventricular region may consist of immature neurons that are yet to express NeuN. Validation of this idea would require further experiments such as; a time course analysis of the regional distribution of BrdU positive cells after BrdU-pulsing, staining for co-expression of BrdU-positive cells with markers seen in migrating and differentiating neuronal precursors such as doublecortin and PSA-NCAM , or an analysis of Ki67-positive cell distribution in a younger mutant cohort. Neurogenesis occurs in the adult mammalian brain, including that of humans [34, 35]. Neurogenesis is known to occur in the sub-ventricular zone [36, 37], the region of the brain that lines the lateral ventricular cavity. It is also possible that rather than originating from periventricular zone stem cells, NeuN/BrdU double positive cells could originate in situ from quiescent stem cells found in the striatum.
Young global and striatal ablation mutants assessed using BrdU as a proliferative marker displayed a marked up-regulation in the number of dividing cells. In contrast, older mice of the global and cortical mutant lines showed markedly reduced Ki-67+ cell numbers when compared to control mice. Experimental confirmation of this finding would involve a direct comparison of the proliferative profile in 4 week and 80 weeks old mice using established Ki-67 immunofluorescence and BrdU protocols. Our finding of reduced Ki-67+ cell numbers in older mutant lines appears to be at odds with some published studies, which underscore an increased proliferative response in humans with HD and rodent HD models [38, 39], although other studies in HD models have identified region specific impaired neurogenesis. Decreased hippocampal dentate gyrus neurogenesis was seen in the transgenic YAC128 model . Olfactory bulb neurogenesis was down-regulated in the hippocampus  and olfactory bulb  of R6/2 transgenic HD mice and striatal cell neurogenesis was impaired in a HD knock-in line . However, it is important to note the age (i.e. 80 weeks) of the mice used in this study. The comparative difference between mutant and control mice may relate to the advanced age of the mice used in this study. We postulate that the aged mutant brain is unable to mount an effective proliferative response  owing to paradigm driven increased proliferative demand in early adulthood and ultimate stem cell pool depletion. There is ample evidence for age-related stem cell depletion in the CNS [19, 45]. Although neural stem cells isolated from the subventricular zone of aged animals are able to divide and differentiate into functional neurons and other neural lineages, they do so with reduced efficiency . A large number of factors may impact on this age-related decline in functional stem cell depletion such as changes in the neurogenic niche and neural progenitor cell specific variables such as altered gene transcription and telomerase activity .
Another explanation for the reduced number of Ki-67+ cells in the mutant mice is the possibility that the proliferative pool contains Drd1a-expressing cells. In this scenario, neural precursor cells that express Drd1a in a transient fashion undergo tox-176-mediated death before they have an opportunity to enter a permissive phase of the cycle and express Ki-67. Cre expression would occur only if the CamKIIa or EMX1 transgenes are expressed in neural precursor cells of the global and cortical lines respectively. If Drd1a is transiently expressed in these cells, then Cre must be expressed before Drd1a to enable Cre-mediated recombination of LoxP sites and Drd1a promoter-driven production of tox-176. If neural progenitor cells express Drd1a, death of these cells will ensue resulting in the absence of a proliferative response in this paradigm and the uniformly low Ki-67+ cell numbers seen in the mutant mice. There are two strong counter arguments. The first is that we did not identify any GFP/Ki-67 double positive cells in GFP-control mice and the second is that the BrdU data from young mice show a clear increase in proliferative response in mutant mice that are designed to shed Drd1a-expressing cells at the time of Drd1a receptor expression.
A further possibility is that the difference in Ki-67+ cell numbers between mutant and control mice in the rostral striatum is not an indication of a decrease in proliferation in the mutant mice, but, rather reflects an increase in migration and terminal differentiation of proliferating cells in response to the molecular pathology in mutant mice. It may be the case that there is, indeed, an increase in proliferation in mutant mice but that the proliferating cells are also differentiating and migrating from the periventricular region/medial striatum towards areas of high cell death density. If this were the case, and considering that Ki-67 is only a marker for early cell division , these differentiating and migrating cells would not be labelled with Ki-67. This is an unlikely explanation given the BrdU data indicate a gradient of phenotype-negative cells maximal in the periventricular region.