**3.2 Striatal infusions of quinpirole**

While infusion of SK channel agonists (1-EBIO or riluzole) into the normal adult mouse SNc recruited ~500 more TH+ SNc cells (Aumann et al. 2008, Aumann et al. 2011), infusion of the D2 DA receptor agonist quinpirole into the normal adult mouse striatum was far more potent in this respect; recruiting ~3000 new TH+ SNc cells (Aumann et al. 2011). We therefore examined the behavioral and cellular effects of striatal quinpirole infusion in 6- OHDA-lesioned rats.

Four different behaviors were measured before and after unilateral 6-OHDA lesion of the SNc and during treatment (100nM quinpirole or vehicle). These are plotted in figure 4A-D. While there was no obvious effect of lesion or treatment on locomotor and rearing behavior (figure 4A-B), an asymmetry in left/right forelimb use (cylinder test and corridor test) and possibly also left/right side attention (corridor test) was evident following lesion (figure 4C-D). The lesion was made in the right SNc in this experiment, resulting in left forelimb

Activity-Dependent Regulation of the Dopamine

Phenotype in the Adult Substantia Nigra: Prospects for Treating Parkinson's Disease 287

Fig. 4. Effects of the D2 DA receptor agonist quinpirole on motor behavior and the number of TH+ SNc cells in 6-OHDA-lesioned adult rats. (A-D) 6-OHDA was injected into the right SNc of 8-week old male Sprague-Dawley rats to reduce the number of SNc TH+ cells to ~50% of normal (see vehicle-treated rats, black symbols in E). Over the ensuing 19 weeks locomotion (A), rearing (B), right/left forelimb use in the cylinder test (C) and right/left feeding in the corridor test (D) were examined. At week 12, 100nM quinpirole (or vehicle) infusion directly into the right dorsal striatum commenced. The mean ± SE of each of the behavioral measures is plotted over time and treatment (vehicle = black, quinpirole = red). 6-OHDA resulted in an increased ratio of right/left forelimb use in the cylinder test (C) and of right/left side feeding in the corridor test (D). There were no effects on locomotion (A) and rearing (B) beyond the expected habituation over time. Quinpirole had no effect over vehicle on any of the behaviors examined (two-way ANOVAs). (E) At the end of week 19, there was no difference in the number of TH+ cells in the 6-OHDA-lesioned SNc between vehicle- (black symbols) and 1-EBIO-infused (red symbols) rats (p=0.769, t-test). See figure 1

These studies were carried out because administration of 1-EBIO, riluzole and quinpirole to normal mice leads to a robust, rapid and reproducible increase in DA cells in the SNc

legend for details about how these data are represented.

**4. Discussion** 

Fig. 3. Effects of the SK channel agonist 1-EBIO (200µM) on rotational behavior in response to amphetamine and the number of TH+ SNc cells in 6-OHDA-lesioned adult rats. (A) 6- OHDA was injected into the right SNc of 8-week old male Sprague-Dawley rats to reduce the number of SNc TH+ cells to ~35% of normal (see vehicle-treated rats, black symbols in B). Over the ensuing 15 weeks the rotational behavior in response to amphetamine was examined. The relative mean ± SE ratio of ipsiversive-to-contraversive rotations performed over a 60 minute interval, beginning immediately following amphetamine injection, is plotted over time for vehicle-treated (black bars) and 1-EBIO-treated (red bars) rats. 6- OHDA resulted in progressive increase in the ratio of ipsiversive:contraversive rotations from weeks 1-5, although this was not evident in the vehicle-treated group and was not statistically significant in either group. At week 6, 200µM 1-EBIO (or vehicle) infusion directly into the right SNc commenced. There was a non-significant trend for 1-EBIO to improve the rotational bias at 9 & 11 weeks, but not at 13 & 15 weeks. There were no effects of treatment or time over the course of this experiment (two-way ANOVA). (B) At the end of week 15, there was no difference in the number of TH+ cells in the 6-OHDA-lesioned SNc between vehicle- (black symbols) and 1-EBIO-infused (red symbols) rats (p=0.672, t-test). See figure 1 legend for details about how these data are represented.

akinesia (and possibly left side attention deficit), evidenced by increases in right:left forelimb use in figure 4C and right:left feeding in figure 4D. There was a degree of normalization of these asymmetries following treatment onset (week 12), however there were no differences in the extent of these normalizations in quinpirole- versus vehicleinfused rats (figure 4C-D).

At the experiment end-point the number of SNc TH+ cells was not different between quinpirole- and vehicle-infused rats (figure 4E).

Fig. 4. Effects of the D2 DA receptor agonist quinpirole on motor behavior and the number of TH+ SNc cells in 6-OHDA-lesioned adult rats. (A-D) 6-OHDA was injected into the right SNc of 8-week old male Sprague-Dawley rats to reduce the number of SNc TH+ cells to ~50% of normal (see vehicle-treated rats, black symbols in E). Over the ensuing 19 weeks locomotion (A), rearing (B), right/left forelimb use in the cylinder test (C) and right/left feeding in the corridor test (D) were examined. At week 12, 100nM quinpirole (or vehicle) infusion directly into the right dorsal striatum commenced. The mean ± SE of each of the behavioral measures is plotted over time and treatment (vehicle = black, quinpirole = red). 6-OHDA resulted in an increased ratio of right/left forelimb use in the cylinder test (C) and of right/left side feeding in the corridor test (D). There were no effects on locomotion (A) and rearing (B) beyond the expected habituation over time. Quinpirole had no effect over vehicle on any of the behaviors examined (two-way ANOVAs). (E) At the end of week 19, there was no difference in the number of TH+ cells in the 6-OHDA-lesioned SNc between vehicle- (black symbols) and 1-EBIO-infused (red symbols) rats (p=0.769, t-test). See figure 1 legend for details about how these data are represented.

#### **4. Discussion**

286 Neuroscience – Dealing with Frontiers

Fig. 3. Effects of the SK channel agonist 1-EBIO (200µM) on rotational behavior in response to amphetamine and the number of TH+ SNc cells in 6-OHDA-lesioned adult rats. (A) 6- OHDA was injected into the right SNc of 8-week old male Sprague-Dawley rats to reduce the number of SNc TH+ cells to ~35% of normal (see vehicle-treated rats, black symbols in B). Over the ensuing 15 weeks the rotational behavior in response to amphetamine was examined. The relative mean ± SE ratio of ipsiversive-to-contraversive rotations performed over a 60 minute interval, beginning immediately following amphetamine injection, is plotted over time for vehicle-treated (black bars) and 1-EBIO-treated (red bars) rats. 6- OHDA resulted in progressive increase in the ratio of ipsiversive:contraversive rotations from weeks 1-5, although this was not evident in the vehicle-treated group and was not statistically significant in either group. At week 6, 200µM 1-EBIO (or vehicle) infusion directly into the right SNc commenced. There was a non-significant trend for 1-EBIO to improve the rotational bias at 9 & 11 weeks, but not at 13 & 15 weeks. There were no effects of treatment or time over the course of this experiment (two-way ANOVA). (B) At the end of week 15, there was no difference in the number of TH+ cells in the 6-OHDA-lesioned SNc between vehicle- (black symbols) and 1-EBIO-infused (red symbols) rats (p=0.672, t-test).

akinesia (and possibly left side attention deficit), evidenced by increases in right:left forelimb use in figure 4C and right:left feeding in figure 4D. There was a degree of normalization of these asymmetries following treatment onset (week 12), however there were no differences in the extent of these normalizations in quinpirole- versus vehicle-

At the experiment end-point the number of SNc TH+ cells was not different between

See figure 1 legend for details about how these data are represented.

infused rats (figure 4C-D).

quinpirole- and vehicle-infused rats (figure 4E).

These studies were carried out because administration of 1-EBIO, riluzole and quinpirole to normal mice leads to a robust, rapid and reproducible increase in DA cells in the SNc

Activity-Dependent Regulation of the Dopamine

treatment; and (4) whether there are species differences.

striatum to the SNc operates to control the number of SNc DA cells.

EBIO is infused into the midbrain or quinpirole is infused into the striatum.

Phenotype in the Adult Substantia Nigra: Prospects for Treating Parkinson's Disease 289

The caveats discussed above are tantalizing and the importance of a positive outcome justifies further experiments that attend to: (1) ensuring drug delivery remains patent; (2) whether continual drug delivery is necessary to maintain the recruited population of DA neurons; (3) whether lesion size (and rate) are confounding factors in the responsiveness to

We will now turn to discussion of what the present results might reveal about the underlying biology. There is an extensive literature showing that expression of TH in cells is regulated by changes in membrane potential (neuronal activity) and intracellular Ca2+. We recently reported that this is true also in the SNc of normal (i.e. unlesioned) adult mice (Aumann et al. 2011, Aumann et al. 2008). When drugs targeting the activity of SNc neurons are infused into the brain for 2 weeks, both the amount of TH protein/SNc cell and the number of SNc TH+ cells are altered (Aumann et al. 2011, Aumann et al. 2008). Specifically, our data show that when the number of TH+ cells increases, the TH immunoreactivity of each cell decreases, and vice versa. Also, when the number of TH+ cells increases, the number of SNc cells that are not TH immunoreactive (TH-) decreases by the same amount, and vice versa [i.e. the net number of SNc cells (TH+ & TH- combined) does not change]. From a broad perspective these data indicate that TH expression in adult SNc neurons is activity-dependent. More closely they imply a homeostatic mechanism(s) regulating DA neurotransmission in the striatum, but acting at the levels of TH expression and DA phenotype recruitment/loss in SNc. We propose that this homeostatic mechanism(s) operates at two levels. At the level of an individual SNc cell, TH expression is activity- and Ca2+-dependent and is brought about by altered Ca2+-dependent DA gene expression. At the level of nigrostriatal circuitry, striatal DA receptor signaling and feedback circuitry from the

The rapid (within 2 weeks) recruitment/loss of the DA phenotype by SNc cells in normal mice (Aumann et al. 2011, Aumann et al. 2008) implies the existence of a population of relatively mature (but not DA) neurons located in and around SNc that can be recruited into and out of the DA population. In other (unpublished) studies we have evidence that significant numbers of new neurons (NeuN+) are generated in the adult mouse midbrain [from Nestin+ (but BrdU-) neural precursor cells], many (~420 over an 8 week period) of which end up in SNc, see also (Shan *et al.* 2006). A very small number of these new midbrain neurons express TH, i.e. demonstrate a capacity to acquire the DA phenotype, but so far these have been observed around the ventral midline and in the ventral tegmental area (VTA), not in SNc. It could be that these newborn NeuN+ but TH- SNc neurons are the same cells in which the DA phenotype can be recruited should appropriate signals arrive. We believe that one of these signals is a change in their electrical activity, such as occurs when 1-

In this context, it is relevant to consider what might have happened to this putative DA phenotype recruitment when confronted with rapid depletion of SNc DA neurons following direct 6-OHDA injection. Evidence from our laboratory and others (Sauer & Oertel 1994) shows that following 6-OHDA, and in the absence of any further treatment or manipulation, the number of SNc TH+ cells spontaneously recovers toward normal, while at the same time the number of SNc TH- cells declines (Stanic et al. 2003). This suggests DA phenotype

(Aumann et al. 2011, Aumann et al. 2008). However, these agents failed to alleviate motor deficits and failed to consistently recruit "new" TH+ SNc cells in mice or rats with prior 6- OHDA-induced depletion of SNc DA neurons. The following discussion focuses on caveats of this conclusion, and on what the present experimental outcomes might reveal about the underlying biology. We conclude that these agents warrant further experimentation, perhaps using animal models that more faithfully recapitulate slow nigrostriatal degeneration in PD.

The first caveat is technical and around drug delivery. Given the credible increase in number of SNc TH+ cells brought about by 2-weeks 1-EBIO infusion in some 6-OHDAlesioned mice (figure 1), we opted for a longer infusion in the rat experiments, expecting cell recruitment would improve. However, the opposite was true. After 9-weeks 1-EBIO infusion in 6-OHDA-lesioned rats, there was no difference in behavior (week 15 in figure 3A) or in the number of SNc TH+ cells (figure 3B), compared with vehicle infusion. However, there was evidence of some behavioral improvement, and presumably nigrostriatal cell recovery, earlier in the course of the 1-EBIO infusion (weeks 9 & 11 in figure 3A). Why then did this potential improvement lapse at weeks 13 & 15? Degradation of the drug is not the explanation because a new pump with a fresh supply of 1-EBIO was introduced at week 12. However, it could have been due to inadvertent and premature termination of drug delivery around week 12. At the experimental endpoint (week 15) it was noted that the majority of pumps were disconnected from their cannula; this was not the case when the pumps were replaced (week 12). Unfortunately we cannot know precisely when this detachment, and therefore cessation of drug infusion, occurred. However, if it was around week 12, it is possible 1-EBIO was having a beneficial effect. Therefore we believe the experiment should be repeated. If future experiments reveal 1-EBIO does facilitate recruitment of new SNc DA neurons and improves motor symptoms in 6-OHDA-lesioned rodents, the present data indicate that these benefits are acutely dependent on the presence of the drug, and therefore continuous drug delivery will be necessary to maintain them.

A second caveat relates to lesion size. In the mouse 1-EBIO experiments (figure 1), lesions were similar in size, evidenced by the relatively tight cluster of data from vehicle-treated mice (black symbols, figure 1). Therefore it is likely that: (1) lesions in the animals receiving 1-EBIO infusion were also of similar size; and (2) 1-EBIO treatment is recruiting new SNc DA neurons in some 6-OHDA lesioned mice. We propose that responsiveness to 1-EBIO may be determined by lesion size, with large 6-OHDA lesions being relatively unresponsive compared to smaller lesions. Our reasoning for this is discussed later. Suffice to point out here that in the 1-EBIO rat experiment (figure 3) [and in the quinpirole rat experiment (figure 4)], lesion size was much more variable than in the 1-EBIO mouse experiments. Moreover, the size of the lesions was ≥ the upper range of mouse lesions [i.e. <40% of TH+ cells remaining, figure 3B (and 4E)] in a significant proportion of rats. Therefore it may be that many of the rats in these experiments were unable to respond to 1-EBIO (or quinpirole) because their lesions were too big (and possibly also too rapid, see discussion below).

A less likely caveat in our opinion is that rats are less able to respond to 1-EBIO or striatal quinpirole than mice. The ability of the rodent SNc to spontaneously compensate following 6-OHDA insult, and in different ways (e.g. recruitment of TH+ SNc cells or sprouting of surviving nigrostriatal axons), is no different in mice versus rats in our experience.

(Aumann et al. 2011, Aumann et al. 2008). However, these agents failed to alleviate motor deficits and failed to consistently recruit "new" TH+ SNc cells in mice or rats with prior 6- OHDA-induced depletion of SNc DA neurons. The following discussion focuses on caveats of this conclusion, and on what the present experimental outcomes might reveal about the underlying biology. We conclude that these agents warrant further experimentation, perhaps using animal models that more faithfully recapitulate slow nigrostriatal

The first caveat is technical and around drug delivery. Given the credible increase in number of SNc TH+ cells brought about by 2-weeks 1-EBIO infusion in some 6-OHDAlesioned mice (figure 1), we opted for a longer infusion in the rat experiments, expecting cell recruitment would improve. However, the opposite was true. After 9-weeks 1-EBIO infusion in 6-OHDA-lesioned rats, there was no difference in behavior (week 15 in figure 3A) or in the number of SNc TH+ cells (figure 3B), compared with vehicle infusion. However, there was evidence of some behavioral improvement, and presumably nigrostriatal cell recovery, earlier in the course of the 1-EBIO infusion (weeks 9 & 11 in figure 3A). Why then did this potential improvement lapse at weeks 13 & 15? Degradation of the drug is not the explanation because a new pump with a fresh supply of 1-EBIO was introduced at week 12. However, it could have been due to inadvertent and premature termination of drug delivery around week 12. At the experimental endpoint (week 15) it was noted that the majority of pumps were disconnected from their cannula; this was not the case when the pumps were replaced (week 12). Unfortunately we cannot know precisely when this detachment, and therefore cessation of drug infusion, occurred. However, if it was around week 12, it is possible 1-EBIO was having a beneficial effect. Therefore we believe the experiment should be repeated. If future experiments reveal 1-EBIO does facilitate recruitment of new SNc DA neurons and improves motor symptoms in 6-OHDA-lesioned rodents, the present data indicate that these benefits are acutely dependent on the presence of the drug,

and therefore continuous drug delivery will be necessary to maintain them.

A second caveat relates to lesion size. In the mouse 1-EBIO experiments (figure 1), lesions were similar in size, evidenced by the relatively tight cluster of data from vehicle-treated mice (black symbols, figure 1). Therefore it is likely that: (1) lesions in the animals receiving 1-EBIO infusion were also of similar size; and (2) 1-EBIO treatment is recruiting new SNc DA neurons in some 6-OHDA lesioned mice. We propose that responsiveness to 1-EBIO may be determined by lesion size, with large 6-OHDA lesions being relatively unresponsive compared to smaller lesions. Our reasoning for this is discussed later. Suffice to point out here that in the 1-EBIO rat experiment (figure 3) [and in the quinpirole rat experiment (figure 4)], lesion size was much more variable than in the 1-EBIO mouse experiments. Moreover, the size of the lesions was ≥ the upper range of mouse lesions [i.e. <40% of TH+ cells remaining, figure 3B (and 4E)] in a significant proportion of rats. Therefore it may be that many of the rats in these experiments were unable to respond to 1-EBIO (or quinpirole) because their lesions were too big (and possibly also too rapid, see discussion below).

A less likely caveat in our opinion is that rats are less able to respond to 1-EBIO or striatal quinpirole than mice. The ability of the rodent SNc to spontaneously compensate following 6-OHDA insult, and in different ways (e.g. recruitment of TH+ SNc cells or sprouting of

surviving nigrostriatal axons), is no different in mice versus rats in our experience.

degeneration in PD.

The caveats discussed above are tantalizing and the importance of a positive outcome justifies further experiments that attend to: (1) ensuring drug delivery remains patent; (2) whether continual drug delivery is necessary to maintain the recruited population of DA neurons; (3) whether lesion size (and rate) are confounding factors in the responsiveness to treatment; and (4) whether there are species differences.

We will now turn to discussion of what the present results might reveal about the underlying biology. There is an extensive literature showing that expression of TH in cells is regulated by changes in membrane potential (neuronal activity) and intracellular Ca2+. We recently reported that this is true also in the SNc of normal (i.e. unlesioned) adult mice (Aumann et al. 2011, Aumann et al. 2008). When drugs targeting the activity of SNc neurons are infused into the brain for 2 weeks, both the amount of TH protein/SNc cell and the number of SNc TH+ cells are altered (Aumann et al. 2011, Aumann et al. 2008). Specifically, our data show that when the number of TH+ cells increases, the TH immunoreactivity of each cell decreases, and vice versa. Also, when the number of TH+ cells increases, the number of SNc cells that are not TH immunoreactive (TH-) decreases by the same amount, and vice versa [i.e. the net number of SNc cells (TH+ & TH- combined) does not change]. From a broad perspective these data indicate that TH expression in adult SNc neurons is activity-dependent. More closely they imply a homeostatic mechanism(s) regulating DA neurotransmission in the striatum, but acting at the levels of TH expression and DA phenotype recruitment/loss in SNc. We propose that this homeostatic mechanism(s) operates at two levels. At the level of an individual SNc cell, TH expression is activity- and Ca2+-dependent and is brought about by altered Ca2+-dependent DA gene expression. At the level of nigrostriatal circuitry, striatal DA receptor signaling and feedback circuitry from the striatum to the SNc operates to control the number of SNc DA cells.

The rapid (within 2 weeks) recruitment/loss of the DA phenotype by SNc cells in normal mice (Aumann et al. 2011, Aumann et al. 2008) implies the existence of a population of relatively mature (but not DA) neurons located in and around SNc that can be recruited into and out of the DA population. In other (unpublished) studies we have evidence that significant numbers of new neurons (NeuN+) are generated in the adult mouse midbrain [from Nestin+ (but BrdU-) neural precursor cells], many (~420 over an 8 week period) of which end up in SNc, see also (Shan *et al.* 2006). A very small number of these new midbrain neurons express TH, i.e. demonstrate a capacity to acquire the DA phenotype, but so far these have been observed around the ventral midline and in the ventral tegmental area (VTA), not in SNc. It could be that these newborn NeuN+ but TH- SNc neurons are the same cells in which the DA phenotype can be recruited should appropriate signals arrive. We believe that one of these signals is a change in their electrical activity, such as occurs when 1- EBIO is infused into the midbrain or quinpirole is infused into the striatum.

In this context, it is relevant to consider what might have happened to this putative DA phenotype recruitment when confronted with rapid depletion of SNc DA neurons following direct 6-OHDA injection. Evidence from our laboratory and others (Sauer & Oertel 1994) shows that following 6-OHDA, and in the absence of any further treatment or manipulation, the number of SNc TH+ cells spontaneously recovers toward normal, while at the same time the number of SNc TH- cells declines (Stanic et al. 2003). This suggests DA phenotype

Activity-Dependent Regulation of the Dopamine

disease. *Neuroscience,* 151, 1142-1153.

disease in the rat. *Neuroscience,* 74, 971-983.

NNIPPS study. *Brain,* 132, 156-171.

*Neurobiology of disease,* 4, 186-200.

MPTP monkey model. *Eur J Pharmacol,* 356, 101-104.

*Neurochem,* 116, 646-658.

1529-1541.

**6. Acknowledgements** 

Greg Thomas.

**7. References** 

Phenotype in the Adult Substantia Nigra: Prospects for Treating Parkinson's Disease 291

of SNc neurogenesis. Progress in these areas promises to be vital for better treating the motor symptoms of PD, but will also be relevant for other disorders involving dysfunctional midbrain DA signaling (e.g. attention deficit hyperactivity disorder, schizophrenia, drug

This study was supported by grants from the National Health & Medical Research Council of Australia, the Bethlehem Griffiths Research Foundation, & the CASS Foundation. The Authors gratefully acknowledge the technical assistance of Doris Tomas, Brett Purcell &

Aponso, P. M., Faull, R. L. and Connor, B. (2008) Increased progenitor cell proliferation and

Aumann, T. D., Egan, K., Lim, J. et al. (2011) Neuronal activity regulates expression of

Aumann, T. D., Gantois, I., Egan, K., Vais, A., Tomas, D., Drago, J. and Horne, M. K. (2008)

Barneoud, P., Mazadier, M., Miquet, J. M., Parmentier, S., Dubedat, P., Doble, A. and

Bauer, M., Meyer, M., Grimm, L., Meitinger, T., Zimmer, J., Gasser, T., Ueffing, M. and

Bensimon, G., Ludolph, A., Agid, Y., Vidailhet, M., Payan, C. and Leigh, P. N. (2009)

Bezard, E., Stutzmann, J. M., Imbert, C., Boraud, T., Boireau, A. and Gross, C. E. (1998)

Bjorklund, A., Rosenblad, C., Winkler, C. and Kirik, D. (1997) Studies on neuroprotective

Bowenkamp, K. E., David, D., Lapchak, P. L., Henry, M. A., Granholm, A. C., Hoffer, B. J.

substantia nigra pars compacta. *Exp Neurol,* 213, 419-430.

astrogenesis in the partial progressive 6-hydroxydopamine model of Parkinson's

tyrosine hydroxylase in adult mouse substantia nigra pars compacta neurons. *J* 

SK channel function regulates the dopamine phenotype of neurons in the

Boireau, A. (1996) Neuroprotective effects of riluzole on a model of Parkinson's

Widmer, H. R. (2000) Nonviral glial cell-derived neurotrophic factor gene transfer enhances survival of cultured dopaminergic neurons and improves their function after transplantation in a rat model of Parkinson's disease. *Human gene therapy,* 11,

Riluzole treatment, survival and diagnostic criteria in Parkinson plus disorders: the

Riluzole delayed appearance of parkinsonian motor abnormalities in a chronic

and regenerative effects of GDNF in a partial lesion model of Parkinson's disease.

and Mahalik, T. J. (1996) 6-hydroxydopamine induces the loss of the dopaminergic phenotype in substantia nigra neurons of the rat. A possible mechanism for restoration of the nigrostriatal circuit mediated by glial cell line-derived

addiction), and for adult neurogenesis and brain plasticity more generally.

recruitment, possibly as part of a process that has directed the homeostasis described above toward repair. The failure of our attempts to facilitate this recruitment using 1-EBIO or quinpirole in the present study might be due to depletion of the population of recruitable cells by this spontaneous recovery. It also indicates the rate of replenishment of the recruitable population (i.e. neurogenesis) is not very high, since not enough were generated within the timeframe of our experiments (4-17 weeks) to significantly impact the TH+ population. This would fit with the consensus in the literature that the rate of adult midbrain neurogenesis is not very high (Shan et al. 2006, Zhao *et al.* 2003) or zero (Aponso *et al.* 2008, Chen *et al.* 2005, Cooper & Isacson 2004, Frielingsdorf *et al.* 2004, Lie *et al.* 2002, Peng *et al.* 2008, Yoshimi *et al.* 2005). This point brings us to another caveat of the present experiments, which is the very rapid depletion of SNc DA neurons by 6-OHDA injection may not be the best PD model in which to study activity-dependent DA phenotype recruitment. Perhaps a model in which the rate of degeneration is much slower, which better mimics the situation in PD, would provide enough time for the recruitable population of cells to be sustained.

#### **5. Conclusion**

In summary, our working model is: (1) The nigrostriatal DA pathway is under homeostatic control to maintain a constant level of striatal DA signaling; (2) This is achieved at the level of individual SNc cells via activity- and Ca2+-dependent alterations in DA (TH) gene expression; (3) It is also achieved at the level of D2 DA receptor-mediated striatonigral feedback circuitry (the indirect pathway) leading to changes in the number of SNc DA cells; (4) A population of relatively mature NeuN+ but TH- cells located in and immediately surrounding SNc is available to be rapidly recruited into and out of the DA population; (5) This recruitable population of cells can be rapidly depleted (i.e. recruited into the DA population) in response to perturbations of the system (e.g. altered SNc neuronal activity, nigrostriatal DA signaling, or 6-OHDA), but is only slowly replenished by neurogenesis. The implication of this homeostasis for nigrostriatal DA cell-replacement therapies is that any manipulation designed to increase the number of SNc DA cells is likely to be offset by a homeostatic response. The effects of infusing D2 agonists into the striatum suggest that striatal DA signaling may be a factor. Thus in PD or its models the effect of homeostatic control may be difficult to predict. On the other hand, pharmaceuticals that increase SNc DA cells in normal mice fail to do so in 6-OHDA-lesioned rodents, possibly because the population of recruitable cells is also extensively depleted. Thus, researchers looking to help develop cell-replacement therapies to treat the motor symptoms of PD should consider the effects their interventions might have on nigrostriatal DA homeostasis because: (1) homeostatic responses may be confounding interpretation of the effects of their interventions; and (2) homeostatic responses may need to be addressed also, as part of an overall strategy to increase cell-based nigrostriatal DA transmission.

Future work should therefore aim to better understand these homeostatic responses by: (1) identifying downstream signaling pathways mediating activity- and Ca2+-dependent changes in DA gene expression; (2) identifying mechanisms of SNc DA phenotype recruitment; (3) identifying the phenotype of recruitable cells; (4) characterizing the ontogenesis of newborn SNc neurons; and (5) investigating mechanisms regulating the rate of SNc neurogenesis. Progress in these areas promises to be vital for better treating the motor symptoms of PD, but will also be relevant for other disorders involving dysfunctional midbrain DA signaling (e.g. attention deficit hyperactivity disorder, schizophrenia, drug addiction), and for adult neurogenesis and brain plasticity more generally.
