Nigrostriatal rAAV-mediated GDNF Overexpression ... - Cell Press

original article

© The American Society of Gene Therapy

Nigrostriatal rAAV-mediated GDNF Overexpression Induces Robust Weight Loss in a Rat Model of Age-related Obesity Fredric P Manfredsson1, Nihal Tumer2,3, Benedek Erdos2,3, Tessa Landa2,3, Christopher S Broxson2,3, Layla F Sullivan1, Aaron C Rising1, Kevin D Foust1, Yi Zhang2,3, Nicholas Muzyczka4, Oleg S Gorbatyuk4, Philip J Scarpace2,3 and Ronald J Mandel1 1 Department of Neuroscience, Powell Gene Therapy Center, McKnight Brain Institute, University of Florida College of Medicine, Gainesville, Florida, USA; 2Department of Veterans Affairs Medical Center, Geriatric Research, Education and Clinical Center, Gainesville, Florida, USA; 3Department of Pharmacology and Therapeutics, College of Medicine, University of Florida, Gainesville, Florida, USA; 4Department of Molecular Genetics and Microbiology, Powell Gene Therapy Center, Genetics Institute, College of Medicine, University of Florida, Gainesville, Florida, USA

Intraventricular administration of glial cell line–derived neurotrophic factor (GDNF) in primate and humans to study Parkinson’s disease (PD) has revealed the potential for GDNF to induce weight loss. Our previous data indicate that bilateral continuous hypothalamic GDNF overexpression via recombinant adeno-associated virus (rAAV) results in significant failure to gain weight in young rats and weight loss in aged rats. Based on these previous results, we hypothesized that because the nigrostriatal tract passes through the lateral hypothalamus, motor hyperactivity mediated by nigrostriatal dopamine (DA) may have been responsible for the previously observed effect on body weight. In this study, we compared bilateral injections of rAAV2/5-GDNF in hypothalamus versus substantia nigra (SN) in aged Brown-Norway X Fisher 344 rats. Nigrostriatal GDNF overexpression resulted in significantly greater weight loss than rats treated in hypothalamus. The nigral or hypothalamic GDNF-induced weight loss was unrelated to motor activity levels of the rats, though some of the weight loss could be attributed to a transient reduction in food intake. Forebrain DA levels did not account for the observed effects on body weight, although GDNF-induced increases in nucleus accumbens DA may have partially contributed to this effect in the hypothalamic GDNF-treated group. However, only nigrostriatal GDNF overexpression induced activation of phosphorylated extracellular signal-regulated kinase (p-ERK) in a small population of corticotrophin-releasing factor [corticotrophin-releasing hormone (CRH)] neurons located specifically in the medial parvocellullar division (MPD) of the paraventricular nucleus of the hypothalamus. Activation of these hypothalamic CRH neurons likely accounted for the observed metabolic effects leading to weight loss in obese rats. Received 31 October 2008; accepted 11 February 2009; published online 10 March 2009. doi:10.1038/mt.2009.45

Introduction Glial cell line–derived neurotrophic factor (GDNF), a distant member of the transforming growth factor-β family of trophic factors, was first described in 1993 and identified as a potent survival factor for central nervous system dopamine (DA) neurons.1,2 GDNF and GDNF family ligands bind to a GFRα family of receptors. Upon GDNF family ligand binding, this ligand–receptor complex then binds to the transmembrane receptor tyrosine kinase, rearranged during transfection (RET), and induces RET dimerization and its subsequent activation.3 GDNF, as opposed to other GDNF family ligands, preferentially binds to the GFRα1 receptor,4,5 leading to intracellular signals acting through the phosphorylation and activation of the phosphatidylinositol 3-kinase/akt pathway as well as activation of the extracellular signal-regulated kinase (ERK) pathway.6,7 These pathways have been shown to promote the survival, maturation, and neurite outgrowth of several populations of neurons in the nervous system.8 Thus, although GDNF has garnered interest as a protein that may treat the progressive neurological disorder, Parkinson’s disease (PD), owing to its effects on nigrostriatal DA neurons, there is biologically significant GDNF, RET, and GFRα1 expression outside the nigrostriatal tract.9 Along these lines, an often observed side effect when delivering exogenous GDNF intracerebroventricularly to rats, rhesus macaques, or humans has been loss of body weight.10–12 To the contrary, weight loss has not been reported in PD clinical trials where GDNF was delivered directly in the putamen.13,14 We have recently shown that bilateral recombinant adeno-associated virus (rAAV)-mediated overexpression of GDNF in the hypothalamus produced significant weight loss in aged rats and reduced the trajectory of normal weight gain in young rats suggesting that circuitry outside the basal ganglia may be involved in GDNF’s ability to induce weight loss.15 The hypothalamus contains dopaminergic neurons and plays an important role in ingestive behavior and energy homeostasis. However, hypothalamic DA levels were unaffected by long-term rAAV-mediated GDNF overexpression.15 Indeed, no obvious biochemical changes in hypothalamus were observed in GDNF-treated animals.15 In addition to resident

Correspondence: Ronald J Mandel, Powell Gene Therapy Center, McKnight Brain Institute, University of Florida College of Medicine, PO Box 100244, Gainesville, Florida, USA. E-mail: [email protected]

980

www.moleculartherapy.org vol. 17 no. 6, 980–991 june 2009


hypothalamic DA neurons, the nigrostriatal DA tract passes through the lateral hypothalamus in the form of the medial forebrain bundle (MFB).16 The nigrostriatal DA tract is known to be involved in feeding and weight loss. Accordingly, complete bilateral destruction of the nigrostriatal tract in rodents leads to severe deficits in feeding and severe weight loss.17 Moreover, striatal DA’s role in maintenance of feeding behavior has been underappreciated as demonstrated by recent studies of the DA deficient mouse where specific DA replacement in areas of the striatum (not the NAcc) reinstated eating in these mice.18,19 Thus, intrahypothalamic GDNF overexpression may have affected the dopaminergic nigrostriatal tract as the MFB courses in the lateral hypothalamus in our previous study.16 Therefore, in this study, first, we wanted to compare the effect of rAAV2/5-mediated GDNF overexpression in the hypothalamus versus the substantia nigra (SN) on body weight in the senescent Fisher-344 X Brown-Norway (F344 X BN) rodent model for age-related obesity. We hypothesized that rAAV2/5-mediated overexpression in the SN might modulate the dopaminergic function in the nigrostriatal tract and possibly affect motor activity enough to result in weight loss. Second, we wanted to further characterize GDNF overexpression in the hypothalamus and its downstream effectors. We hypothesized that continuous hypothalamic GDNF overexpression would enhance the hypothalamic–pituitary–adrenal axis, thereby affecting the sympathetic nervous system, thus resulting in a weight loss, independent of that of altered motor behavior. Our findings unexpectedly indicate that rAAV2/5-mediated nigrostriatal GDNF overexpression temporarily produces significantly more anorexia as well as significantly greater loss of body weight than that of hypothalamic GDNF overexpression.

Results Transgene expression after SN injections The main goal of this study was to compare GDNF overexpression in the nigrostriatal tract versus local overexpression in the hypothalamus. We have previously reported rAAV2/5-GDNF mediated weight loss when injected in the hypothalamus along with careful documentation of the anatomical distribution of GDNF.15 To determine the precise anatomical distribution of rAAV-mediated transgene expression in the nigrostriatal tract, we performed a preliminary experiment using a single vector expressing both GDNF and green fluorescent protein (GFP) under different promoters. For this preliminary study, the rats were injected unilaterally and GDNF and GFP expression were imaged simultaneously. Nigral rAAV2/5-GDNF-GFP injection resulted in widespread GDNF immunoreactivity, which essentially filled the striatum (Figure 1a). To compare the actual transduction pattern of the vector with that of the spread of the transgene, we compared the expression of GFP with that of GDNF. The expression of GFP was typical of that seen in nigral injections of rAAV2/5 (ref. 20) with expression of the marker protein seen extending from the SN (Figure 1b,c) throughout the MFB (Figure 1d–i) to the striatum (Figure 1j,k). As expected, GDNF immunoreactivity was observed in cells coexpressing GFP, but diffused extracellular GDNF was predominant Molecular Therapy vol. 17 no. 6 june 2009

Nigrostriatal GDNF Over-expression and Weight Loss

(e.g., Figure 1b).21 In addition, GDNF transgene was also observed in nontransduced cells throughout the nigrostriatal tract (Figure 1e), which we interpret as uptake and accumulation of released GDNF.21 In addition, the distribution of the GDNF was observed in the lateral portions of the hypothalamus (Figure 1d,h). The observed pattern of GDNF and GFP staining suggests that the vast majority of GDNF is secreted when overexpressed in the nigrostriatal pathway.21

Body weight, food intake, and activity To assess the effect of GDNF overexpression in the hypothalamus as compared to the SN, F344 X BN rats were randomly divided into four rAAV2/5 vector injection groups consisting of rAAV2/5-GDNF bilaterally injected into the SN and hypothalamus in separate groups (SN-GDNF and hyp-GDNF, respectively), and rAAV2/5-GFP was identically injected to serve as controls for transduction in two additional groups (SN-GFP and hypGFP, respectively). Within the first 2 weeks following the viral injection, all groups displayed a modest decrease in body weight (Figure 2a), probably as a response to the surgical procedure. However, following this initial 2-week period, both the rAAV2/5GFP-treated groups displayed an increase in bodyweight, and at 10 weeks following the injections, both groups (SN-GFP and hyp-GFP) had returned to, and surpassed, the presurgery body weight. Conversely, both the hyp-GDNF and SN-GDNF-injected animals maintained a steady decrease in weight throughout the entire experiment, resulting in a significantly lower body weight than that of the respective control group (SN-GDNF 85% or hyp-GDNF 87%). Furthermore, the SN-GDNF group displayed a greater loss of body weight (≈ −80 g) as compared to the hyp-GDNF group (≈ −40 g). This difference in body weight can partially be explained by changes in food intake (FI). All animals displayed a reduction in FI in the weeks immediately following the rAAV2/5 injection (Figure 2b). However, whereas both control groups returned to preinjection FI levels, the SN-GDNF-injected groups displayed a significant but transient reduction in FI (75% of control) 2–5 weeks postinjection. The hypGDNF group displayed lower feeding (89% of control) 4–5 weeks postinjection. In addition, the SN-GDNF displayed 11% less FI for a shorter period of time (2–3 weeks postinjection) when compared to the hyp-GDNF group. Cumulative FI postinjection was significantly lower in the SN-GDNF (185 g) group as compared to the SN-GFP (203 g) and hyp-GDNF (201 g) groups. However, in this measure, there was no significant difference between the two hypothalamic injection groups (hyp-GFP = 211 g; Figure 2b inset). Furthermore, following the initial period of variability in FI, there was no significant time × group interaction between the two GDNF groups in terms of body weight, there was, however, a significant group × time interaction when comparing hypothalamic or nigral injection groups (data not shown). Thus, even after FI was restored, the body weight of the animals when compared with their respective controls was reduced. There was no significant difference between treatment groups in activity levels 45 days postinjection, at a time where feeding was essentially equal between all the groups (Figure 2c). Furthermore, there was no change in energy expenditure among the groups, as measured by oxygen consumption (data not shown). 981



Transgene expression GDNF protein levels were assessed by enzyme-linked immunosorbent assay in tissue samples taken from the striatum, NAcc, and the hypothalamus. Only undetectable background GDNF levels were present in rAAV2/5-GFP-injected animals (data not shown). In rAAV2/5-GDNF-injected animals, transgene levels were equal in all sites assayed regardless of the injection site, however, the hypothalamus contained significantly higher GDNF levels per mg tissue as compared to the striatum and the NAcc (Figure 2d). Histology: hypothalamic injections In addition to enzyme-linked immunosorbent assay measurements, transgene was also observed by immunohistological detection. GDNF expression from the bilateral hypothalamic injections was identical to that previously described.15 Briefly, widespread diffuse GDNF expression was observed stretching from the medial preoptic nucleus throughout the hypothalamus to the posterior lateral hypothalamic area (Figure 3c–e). In addition, GDNF expression at the level of the striatum was also observed in the MFB and

a

the bed nucleus of the stria terminalis (Figure 3a). Furthermore, no GDNF expression was seen in the SN (Figure 3b). Because GDNF signaling is known to lead to increased phosphorylation of ERK, we also evaluated the extent of ERK phosphorylation in the nigrostriatal tract as a result of rAAV2/5-GDNF injections. Hypothalamic injections of rAAV2/5-GDNF resulted in phosphorylated extracellular signal-regulated kinase (p-ERK) immunoreactivity throughout the paraventricular hypothalamic (PVH) area (Figure 4a,b) whereas rAAV2/5-GFP injections showed no phosphorylated ERK (Figure 4c,d).

Histology: nigral injections Bilateral rAAV2/5-GDNF injections resulted in GDNF expression patterns identical to those shown for unilaterally injected rats in Figure 1. When injected bilaterally into the SN, rAAV2/5-GDNF led to detectable p-ERK in the SNc specifically (Figure 4e,f), as well as in the PVH (Figure 4g,h); however, p-ERK was not observed in the terminal regions of the striatum or NAcc for either of the injection paradigms (data not shown).

c

b SNc Lv

STR

AC

SNr

d

VTA

e

f

h

i

mt ic fx

g

fx

ic

mt

3v fx

j

STR

Lv

k

l

Figure 1 GDNF distribution and viral transduction pattern. (a) Sections from an animal injected in the SNc with the CBA-GDNF/HSV-TK-GFP virus were stained for GDNF to show the level of transduction of the entire striatum using bright field microscopy, and fluorescence to show the transduction pattern with precise anatomical tracing of the tract as assessed by GFP (green) and the distribution of GDNF (red) relative to the GFP+ fibers. (b) At the level of the SN, GDNF immunoreactivity was observed throughout the SN both intra- and extracellularly, with transgene observed relatively distal to the transduced cells, arrows point to the substantia nigra pars compacta (SNc). (c) Higher magnification of area outlined in b. (d–f) GDNF immunoreactivity was observed in a wide area adjacent to the MFB throughout the nigrostriatal tract, including the posterior portions of the lateral hypothalamus. (e,f) Increased magnifications of area outlined in d. (g–i) GDNF and GFP expression at the level of the anterior hypothalamus. (i) Higher magnification of area outlined in h. Conversely, no GDNF or GFP expression was observed in the contralateral nigrostriatal tract. (j) In the terminal region of the striatum, GDNF immunoreactivity was seen throughout the striatum while a majority of GFP+ fibers were observed in the ventral portions of the striatum. (k) Higher magnification of area outlined in j. (l) Again, no GDNF or GFP expression was observed in the contralateral hemisphere. Bars in a–c = 1 mm, d, g, h, j, and l = 0.5 mm, in c, e, i, and k = 0.1 mm, and in e = 25 µm. AC, anterior commissure; fx, fornix; GDNF, glial cell line–derived neurotrophic factor; GFP, green fluorescent protein; HSV-TK, herpes simplex virus thymidine kinase; ic, internal capsule; LV, lateral ventricle; MFB, medial forebrain bundle; mt, mammillothalamic tract; STR, striatum; SNr, substantia nigra pars reticulata; VTA, ventral tegmental area; 3v, third ventricle.

982

www.moleculartherapy.org vol. 17 no. 6 june 2009


40

b 25

Change in body mass

Weight change (g)

0 #%

−20

%

−40

*

−60

#

−80 −100 −120

c

hyp-gfp hyp-gdnf sn-gfp sn-gdnf

15

* #* #

10

*

*

Cumulative Food Intake Post-Injection

% 250

%

5 0

−2 −1 0 1 2 3 4 5 6 7 8 9 10 Week

hyp-gfp hyp-gdnf sn-gfp sn-gdnf

200 150 100

GDNF (ng/ml)/mg tissue

14 12 10 8 6 4 2 Hyp-GDNF Hyp-GFP SN-GDNF

0

SN-GFP

hyp-gfp hyp-gdnf sn-gfp

sn-gdnf

−2 −1 0 1 2 3 4 5 6 7 8 9 10 Week

d 50

Activity 16

Total beam breaks (thousand/72 h)

20

50

*

18

0

Daily food intake (g)

20

Food intake

Grams

a


GDNF ELISA

40 sn-gdnf hyp-gdnf

30 20 10 0

Straitum

NAcc

Hypothalamus

Figure 2 Animal measures. (a) Weekly change in body mass and weekly FI (b) before and after rAAV2/5-administration (week 0). Open circles represent the hyp-GFP group (n = 7), open triangles SN-GFP (N = 8), closed circles hyp-GDNF (n = 6), and closed triangles SN-GDNF (n = 7). The SN-GDNF group displayed a significant decrease in body weight as compared to the GFP control from week 2 (*) and when compared to the hypGDNF (#) from week 3. The hyp-GDNF group displayed a significant weight loss as compared to the GFP control from week 6 (%) (P < 0.01). There was also a transient decrease in FI (b) in the SN-GDNF animals when compared to the GFP control (25%) 2–5 weeks following the rAAV2/5 injection (*) and the hyp-GDNF group (11%) (weeks 2,3) (#). The hyp-GDNF-injected animals consumed 11% less food 3–4 weeks following the injections as compared to the GFP control (%) (P < 0.001). Cumulative FI postinjection (b-inset) was significantly lower in the SN-GDNF group when compared to the GFP control and the hyp-GDNF group (P < 0.001). Activity measures (c) done at ~45 days postsurgery indicate no significant differences in activity among the groups. (d) ELISA measurements specific for the transgene indicate that there was no difference due to injection site, however, hypothalamic GDNF levels were >50-fold in measurements taken from the striatum and NAcc. ELISA, enzyme-linked immunosorbent assay; FI, food intake; GDNF, glial cell line–derived neurotrophic factor; GFP, green fluorescent protein.

To elucidate the specific area of the PVH showing ERK activation, double labeling of p-ERK and oxytocin was performed (Figure 5a–f). Oxytocin was observed immediately surrounding the area of ERK activation, with very little colocalization. Considering the distinct profile of oxytocin expression in the parvocellullar PVH, this almost certainly indicates that the area of ERK activation was that of the medial parvocellullar division (MPD) of the PVH almost exclusively containing corticotrophin-releasing hormone (CRH) neuroendocrine neurons.22,23 Finally, nigral rAAV2/5-GFP injections resulted in no detectable levels of p-ERK (Figure 5g,h) indicating that p-ERK activation was not due to nonspecific nigrostriatal transgene overexpression.

Catecholamine levels To ascertain whether altered catecholamine levels due to GDNF overexpression might explain the observed differences in body weight, tissue samples were collected from the striatum, NAcc, and the hypothalamus. The SN-GFP group displayed significantly lower DA levels in the striatum as well as in the NAcc (52 and 67% reduction, respectively) than the hyp-GFP group (Figure 6a,c); however, no significant difference was observed when compared to the SN-GDNF group. Animals overexpressing hypothalamic GDNF showed a significant 59% increase in levels of DA in the NAcc as compared with the SN-GDNF group (Figure 6c). DA levels in the hypothalamus were not different Molecular Therapy vol. 17 no. 6 june 2009

among the groups (Figure 6b). In addition, because elevated levels of hypothalamic norepinephrine (NE) has been shown to result in increased activation of the MPD,23 NE measurements were performed in the same tissue sites as for DA. No difference was seen between the GDNF injection groups and their respective controls in the striatum or the NAcc. In contrast, hypothalamic NE levels were 50% higher in the SN-GDNF group compared to the GFP controls; however, this difference did not reach statistical significance (P < 0.07, data not shown). To evaluate any parasympathetic effect due to GDNF overexpression, serum NE levels were evaluated and the SN-GDNF group displayed a roughly tenfold increase of serum NE over the SN-GFP group (Figure 6d). Because of the enigmatic effect on catecholamine levels observed in the SN-GFP group (Figure 6a), we also evaluated tyrosine hydroxylase (TH) immunoreactivity in the midbrain. No observable difference in TH staining intensity was seen when comparing the SN-GDNF (Figure 6e) with the TH intensity of the SN-GFP group (Figure 6f) indicating that the nonsignificant reduction of striatal DA in the SN-GFP group compared to the SN-GDNF group was not due to a lesion effect of nigrostriatal GFP overexpression. This lack of effect is in contrast to GDNFinduced alterations of striatal TH reported previously,21,24 but differences between the duration of GDNF overexpression and site of injection in this study preclude direct comparison with those previous reports. 983



a

a MS

b

Str

mt

Bed Nu

mfb fx

AC

3v

c

d

b mfb SN

mt fx 3v

mp mm

e c

d

f SN

mt

VTA

mt mfb

mfb fx

3v

fx

3v

g

h

mt

e mt fx mfb

fx

3v 3v

Figure 3 GDNF expression in the hyp-GDNF group. Sections were stained for GDNF expression using a transgene specific antibody. Images were taken throughout the entire axis of transgene expression (a) stretching from the striatum, (b) to the midbrain, (c–e) throughout the hypothalamus. (a) GDNF expression in the area of the striatum at approximately the level of bregma. (b) GDNF expression at the level of the midbrain. (Bregma −4.8 mm). (c) GDNF staining at the anterior (bregma −2.8 mm), (d) medial (bregma −3.6 mm) and posterior (bregma −4.2 mm), (e) portions of the hypothalamus. AC, anterior commissure; BedNu, bed nucleus of the stria terminalis; GDNF, glial cell line–derived neurotrophic factor; MFB, medial forebrain bundle; mm, medial mammilary nucleus; mp, mammillary peduncle; MS, medial septum; Str, striatum, SN, substantia nigra; Bar = 1 mm.

TH TH is the rate-limiting enzyme in DA synthesis, and to investigate whether the observed effect of GDNF on DA levels was a result of a change in TH expression, we also quantified TH levels from the same anatomical areas. HYP-GFP animals had significantly more hypothalamic TH than any of the other treatment groups (Figure 7c). The biological significance of this difference is unclear. In addition, to evaluate peripheral catecholaminergic function, the adrenal glands were also evaluated for TH content. Western and northern analysis of the adrenal glands showed no significant increase in TH protein or messenger RNA levels in either group (Figure 7a,b). 984

Figure 4 Phosphorylated-ERK immunoreactivity as a result of GDNF overexpression. Sections were stained using a p-ERK specific antibody. (a) Hypothalamic rAAV2/5-GDNF injections resulted in a significant increase in ERK phosphorylation in the hypothalamus. (b) Higher magnification of area indicated in a. (c,d) However, hypothalamic injections with rAAV2/5-GFP yielded no detectable p-ERK in the hypothalamus. (d) Higher magnification of area outlined in c. (e,f) Likewise, injections in the SN with rAAV2/5-GDNF resulted in p-ERK immunoreactivity in the SNc as well as (g,h) in the hypothalamus (g,h). Panels f and h are high magnification images of areas indicated in e and g, respectively. Bars in a and c = 1 mm, b, d, f, and h = 50 µm, e, g = 0.5 mm. GDNF, glial cell line–derived neurotrophic factor; p-ERK, phosphorylated extracellular signal-regulated kinase.

Neuropeptide Y and AT1 receptor Levels of neuropeptide Y message in the adrenal gland were significantly increased in the SN-GDNF group when compared to the GFP control and the hyp-GDNF groups (200%, Figure 7e). No significant difference between the two hypothalamic injection groups was observed (Figure 7e). In addition, the hyp-GDNF displayed a trend toward increased hypothalamic AT1 receptor as compared to the GFP control (P = 0.07, Figure 7d). Adiposity White adipose tissue was examined in the perineum (parametrial white adipose tissue), the retroperitoneum (retroperitoneal www.moleculartherapy.org vol. 17 no. 6 june 2009



a

b

c

d

e

f

g

h

dp pml mpd mpv

3v

Figure 5 Anatomical distribution of p-ERK immunoreactivity. (a–f) Animal injected in the SNc with rAAV-GDNF was analyzed for oxytocin and p-ERK immunoreactivity to characterize the subset of activated paraventricular neurons. The area of p-ERK immunoreactivity (red) was largely distinct from that of oxytocin (blue). (a) Mapping adapted from Swanson.49 (b,c) Individual channels of a. (d–f) Increased magnification of a–c. Yellow arrow identifies the same oxytocin expressing cell in all panels. Green arrow identifies the same p-ERK expressing cell in all panels. Abbreviations are all for subdivisions of the PVH: dp, dorsal parvocellular; pml, posterior magnicellular; mpv, medial parvocellular ventral zone; mpd, medial parvocellular dorsal zone. (g,h) Injections of rAAV2/5-GFP in the SN did not result in hypothalamic p-ERK immunoreactivity. (h) Higher magnification of area outlined in g. Bars in a–f = 0.1 mm and g, h = 0.5 mm. GDNF, glial cell line–derived neurotrophic factor; PVH, paraventricular hypothalamic; rAAV, recombinant adeno-associated virus.

white adipose tissue), and the epididymis (epididymal white adipose tissue), and brown adipose tissue was measured as well. In almost all measurements, the GDNF-injected animals displayed dramatically lower fat content; the hyp-GDNF group displayed a 41% loss in total white adipose tissue and the SN-GDNF group a 43% loss when compared to control (Table 1). This loss in fat content was distributed over the various deposits; the hypGDNF and SN-GDNF groups displayed a significant reduction in parametrial white adipose tissue, retroperitoneal white adipose tissue, and epididymal white adipose tissue as compared to the respective control. Brown adipose tissue content was also lower in both GDNF-treated groups. Furthermore, parametrial white adipose tissue content was significantly lower in the SN-GFP group when compared to the hypothalamus injection control. In addition, the fat content of the heart and the soleus was measured; the soleus and the left ventricle fat content was equal in all groups, however, the right ventricle contained significantly less fat in the hyp-GDNF and SN-GFP groups when compared to the hyp-GFP control. The hyp-GDNF group also contained 15% less fat content in the right ventricle than the SN-GDNF group. Molecular Therapy vol. 17 no. 6 june 2009

Discussion We have previously shown that rAAV2/5-mediated hypothalamic GDNF delivery results in a significant reduction in weight gain in young rats and weight loss in aged animals.15 Several groups have shown that intraventricular GDNF overexpression in both animals10,25–28 and humans11,12 has led to weight loss, presumably as a result of the effect on dopaminergic neurons of the hypothalamus or the dopaminergic nigrostriatal system. Our previous work on hypothalamic GDNF overexpression left several questions unanswered. We previously found that although GDNF clearly affected body weight in young and old rats, GDNF-mediated alterations in hypothalamic DA levels did not explain our results. However, in that experiment, we neither examined striatal DA levels nor did we monitor motor hyperactivity levels. GDNFinduced increases in striatal DA and a concomitant increase in basal activity are potential explanations for the observed hypothalamic GDNF effects because the nigrostriatal dopaminergic tract passes through the lateral hypothalamus16 and GDNF may have affected the passing DA fibers.21,29 Moreover, striatal DA signaling has been implicated as an important mediator of ingestive behavior.30 Thus, we chose to 985



a

b

Striatum

8 7

0.6 pg DA/g tissue

pg DA/g tissue

6 5 4 3 2

c

Hyp-GFP

Hyp-GDNF

SN-GFP

d

*

3.5

*

0.3 0.2

Hyp-GFP

Hyp-GDNF

SN-GFP

SN-GDNF

Serum norepinephrine

12

*

10 ng NE/ml serum

3 pg DA/g tissue

0.4

0

SN-GDNF

NAcc

4

2.5 2 1.5 1

8 6 4 2

0.5 0

0.5

0.1

1 0

Hypothalamus

0.7

*

Hyp-GFP

Hyp-GDNF

SN-GFP

0

SN-GDNF

e

Hyp-GFP

Hyp-GDNF

SN-GFP

SN-GDNF

f

Figure 6 Catecholamine analysis. (a) DA analysis of the striatum indicated a slightly lower DA content in the SN-GFP (n = 7) injected animals as compared to the hyp-GFP group (n = 6) (P < 0.05 by ANOVA) and the SN-GDNF group (n = 6) (P = 0.07). (b) No differences were observed in hypothalamic DA content in-between the various groups. (c) DA content in the NAcc was significantly elevated in the hyp-GDNF group when compared to the SN-GDNF group, in addition, the hyp-GFP group contained slightly higher DA levels when compared to the SN-GFP control (P < 0.05 by ANOVA). (d) Evaluation of serum norepinephrine content indicated a significant increase in the SN-GDNF group when compared to the GFP control (P < 0.05 by ANOVA). (e,f) Low power photomicrographs showing no GDNF-induced alteration of nigral TH expression. TH immunoreactivity in animal injected in the (e) SN with rAAV2/5-GDNF, and (f) rAAV2/5-GFP f. Bars in e,f = 1 mm. ANOVA, analysis of variance; DA, dopamine; GDNF, glial cell line–derived neurotrophic factor; GFP, green fluorescent protein.

examine directly whether nigrostriatal GDNF overexpression accounted for the observed effects on body weight. Furthermore, we chose to use F344 X BN animals because this model has shown increased adiposity and leptin resistance without increased feeding.31 Our results show that GDNF overexpression either in the hypothalamus or in the SN significantly reduces body weight in these aged leptin-resistant rats. Furthermore, nigrostriatal GDNF overexpression causes much more pronounced loss of body weight over time. Altered DA signaling (particularly in the ventral tegmental area-NAcc) is believed to be vital to feeding behavior,32–34 but no consistent alterations in DA neurotransmission in the hypothalamus or striatum were found in this study. However, DA content in the NAcc was drastically elevated in animals overexpressing GDNF via the hypothalamic injection. In addition, nigrostriatal GDNF overexpression caused a significant decrease in FI lasting for several weeks following the viral injection, including a significant difference between the two GDNF-treated groups. The hyp-GDNF group also displayed a transient reduction in FI; however, cumulative FI posthypothalamic injection was not different from that of the control. It is possible 986

that this transient FI behavior corresponds to the timing and onset of transgene expression which would reach maximum levels after ~4 weeks,35 whereas the return to baseline behavior may be due to compensatory mechanisms whereby this effect of the transgene is diminished over time.36 In contrast, the progressive weight loss in both the hyp-GDNF and SN-GDNF groups was maintained when compared to the respective control even after feeding returned to normal. Therefore, as we reported previously,15 anorexia does not fully account for the weight loss seen in this study. Long-term overexpression of GDNF has been shown to lead to significantly lower levels of TH in transduced areas,24 however, in this study DA synthesis and turnover were unaffected in all GDNF-treated groups over the survival time studied here. In fact, short-term GDNF overexpression enhances DA turnover and synthesis,21,29 and it has been hypothesized that the observed TH downregulation is an adaptation to this phenomenon to readjust neuronal activity to within normal range.24 Thus, it is also possible that increased striatal DA early after the onset of GDNF expression may have accounted for the transiently reduced FI. However, transient fluctuations in FI is www.moleculartherapy.org vol. 17 no. 6 june 2009


Adrenal gland TH protein 180

140

160

120

140

100 80 60

100 80 60 40

20

20 Hyp-GFP

Hyp-GDNF

SN-GFP

0

SN-GDNF

d

Hypothalamus TH protein 140

*

120

Hyp-GDNF

SN-GFP

SN-GDNF

Hypothalamus AT1 protein 140

100 Arbitrary units

80 60 40 20 0

Hyp-GFP

120

100 Arbitrary units

120

40

0

c

b

Adrenal gland TH mRNA 160

Arbitrary units

Arbitrary units

a


80 60 40 20

Hyp-GFP

Hyp-GDNF

SN-GFP

e

0

SN-GDNF

Hyp-GDNF

SN-GFP

SN-GDNF

Adrenal gland NPY mRNA 300

*

250 Arbitrary units

Hyp-GFP

*

200 150 100 50 0

Hyp-GFP

Hyp-GDNF

SN-GFP

SN-GDNF

Figure 7 Protein and mRNA analysis. Measurement of (a) TH mRNA levels and (b) protein levels in the adrenal gland indicated no significant difference among the various groups. (c) Evaluation of hypothalamic TH protein content indicated a slight but significant (P = 0.03 by ANOVA) increase in the hyp-GFP (n = 6) group as compared to the SN-GFP (n = 6) group. (d) Levels of AT1 protein in the hypothalamus was roughly equal among all groups except from the hyp-GDNF group, this result was not significant, however (P = 0.1 by ANOVA). (e) The SN-GDNF (n = 6) group displayed dramatically elevated levels (twofold increase) of NPY mRNA in the adrenal gland when compared to the GFP control (n = 6) and the hyp-GFP group (n = 6) (P < 0.0001 by ANOVA). ANOVA, analysis of variance; GDNF, glial cell line–derived neurotrophic factor; GFP, green fluorescent protein; mRNA, messenger RNA; NPY, neuropeptide Y; TH, tyrosine hydroxylase. Table 1 White and brown adipose tissue weights in grams (SEM) WAT

PWAT

RTWAT

EWAT

BAT

SOL

LV

RV

Heart

Hyp-GFP

Treatment

29.43 (1.32)

2.16 (0.16)

13.25 (0.75)

14.02 (0.71)

0.71 (0.06)

0.34 (0.01)

0.94 (0.04)

0.22 (0.01)

1.16 (0.04)

Hyp-GDNF

17.22 (2.01)

1.06 (0.16)

7.72 (1.19)

8.44 (0.70)

0.52 (0.03)

0.34 (0.02)

0.93 (0.03)

0.17 (0.00)

1.11 (0.04)

SN-GFP

25.70 (1.42)

1.71 (0.12)

11.70 (0.93)

12.30 (0.68)

0.67 (0.06)

0.31 (0.03)

0.94 (0.02)

0.20 (0.01)

1.14 (0.02)

SN-GDNF

14.74 (2.31)

1.08 (0.17)

5.63 (1.08)

8.03 (1.10)

0.41 (0.03)

0.31 (0.03)

0.92 (0.01)

0.20 (0.01)

1.12 (0.01)

Significance

b,c

b,c,d

b,c

b,c

b,c

a,b,d

Abbreviations: BAT, brown adipose tissue; EWAT, epididymal white adipose tissue; GDNF, glial cell line–derived neurotrophic factor; GFP, green fluorescent protein; LV, left ventricle; PWAT, parametrial white adipose tissue; RTWAT, retroperitoneal white adipose tissue; RV, right ventricle; SOL, soleus; WAT, white adipose tissue. Values represent the mean with the SEM in parenthesis. Significance is indicated by a: hyp-GDNF versus SN-GDNF, b: hyp-GFP versus hyp-GDNF, c: SN-GFP versus SN-GDNF, and d: hyp-GFP versus SN-GFP.

also a common result of the administration of anorexic drugs, whereas baseline FI levels are restored through multiple and redundant orexigenic pathways.37 Molecular Therapy vol. 17 no. 6 june 2009

Although we did not observe any differences in hypothalamic or striatal DA levels between the different experimental groups, we did observe significantly higher NAcc DA as a result 987


of hypothalamic GDNF overexpression than as a result of nigrostriatal GDNF overexpression (Figure 6c). Therefore, although it is possible that increases in NAcc DA may contribute to the observed weight loss in aged animals, the fact that the magnitude of the NAcc DA increase is not correlated to the magnitude of weight loss suggests other factors must contribute as well. In addition, elevated serum levels of NE were observed and are most likely derived from the adrenal gland. Although we have not assessed serum GDNF levels, it is unlikely that GDNF overexpressed in the brain can escape to the periphery to affect the adrenal gland directly. Thus, although a direct effect of GDNF on the adrenal gland in this study cannot be ruled out, the most likely explanation is that GDNF indirectly affected the adrenal gland due to GDNF-mediated activation of hypothalamic nuclei. Taken together, the observations of the nigrostriatal dopaminergic system may explain some of the weight loss seen in GDNFtreated animals. These findings add to the idea that the striatum, in fact, is important in maintenance of body weight, although the ultimate biological mechanism is not understood. However, our data also point to the fact that GDNF-mediated activation of various hypothalamic areas played a role in the observed weight loss. Our results do not completely clarify the biological mechanism of SN-GDNF-induced weight loss. However, hypothalamic GDNF levels were drastically increased regardless of injection site indicating that both injection paradigms had the potential to modulate hypothalamic function in addition to that of the NAcc. The dopaminergic axons traveling via the MFB are unmyelinated which may enable release of GDNF throughout the length of the nigrostriatal tract. Indeed, Figure 1 clearly demonstrates GDNF distributed distal to intracellular GFP when both transgenes were expressed from a single vector. Thus, nigral rAAV2/5-GDNF injections resulted in clear hypothalamic localization of released GDNF (Figure 1c–e,g,h). Moreover, it is clear that rAAV2/5-GDNF had a biological effect because ERK phosphorylation was induced in both the SN and the hypothalamus in the nigral rAAV2/5-GDNFinjected animals and in the hypothalamus of the hyp-GDNF-injected animals, indicating hypothalamic GDNF receptor binding and intracellular signal activation regardless of the injection site. Significantly, no ERK phosphorylation was observed in either the striatum or the NAcc strongly suggesting that GDNF overexpression either in the nigrostriatal tract or in the hypothalamus exerted its effects both in the SN and in the hypothalamus in the SN-GDNF group and exclusively in the hypothalamus in the hyp-GDNF group. Thus, it is clear that hypothalamic GDNF overexpression which causes ERK phosphorylation in the PVH is sufficient to induce weight loss in obese rats. Nigral rAAV injections lead to GDNF overexpression in both the SN and the hypothalamus, which results in ERK phosphorylation in SN and specifically in the MPD division of the hypothalamus, and induce a much greater magnitude of weight loss in these rats. Therefore, these data suggest that a synergistic effect of simultaneous GDNF overexpression in the hypothalamus and SN is likely. On one hand, increased GDNF in the basal ganglia may have modulated catecholaminergic function of NAcc triggering downstream effects which ultimately resulted in a transient reduction in FI and subsequent weight loss. On the other hand, increased hypothalamic stimulation due to GDNF, which 988


was markedly enhanced in both injection groups, resulted in increased adrenal function, subsequently causing increased sympathetic tone and fat metabolism, thereby leading to a significant reduction in fat deposits throughout the body. It has been shown that phosphorylation of ERK is an important signaling event in the activation of the CRH releasing neurons of the MPD and the subsequent enhancement of hypothalamic– pituitary–adrenal axis.23 It is unclear whether SN-GDNF induced ERK activation is due to direct action of GDNF at the level of the hypothalamus. We observed a nonsignificant 50% increase of hypothalamic NE exclusively in the SN-GDNF group (P < 0.07), and it has been shown that direct application of NE in the hypothalamus results in ERK phosphorylation in the MPD.23 It is feasible that nigral overexpression of GDNF acted on noradrenergic efferents projecting to the hypothalamus through the midbrain such as projections from locus coeruleus.23 One potential confounding effect in this study was the minor effect on DA in the SN-GFP group where DA levels were significantly lower in the striatum and NAcc when compared to the hypothalamic GFP group (Figure 6a,c). Although most studies have not noted toxicity due to vector-mediated GFP overexpression, in vitro results38,39 and some animal studies40,41 have indicated that very high levels of GFP expression may be detrimental to cells and ultimately cause death. A mild rAAV-GFP effect on nigral DA neurons has been observed previously.40 However, in this study, obvious nigral DA neuron loss neither occurred in SN-GFP animals (Figure 6e,f) nor were their striatal DA levels altered significantly directly compared to SN-GDNF-treated rats (Figure 6a). In summary, because there was no effect of nigral GFP expression on body weight and no p-ERK staining was observed in SN-GFP rats (Figure 5g,h), we do not believe that the observed nigral GDNF overexpression induction of hypothalamic ERK phosphorylation in CRH+ neurons is due to a GFP lesion-induced reduction of ERK phosphorylation in control animals. GDNF and GDNF family members are currently under intensive evaluation for a gene therapeutic strategy using rAAV vectors to treat PD.42 In PD clinical trials where GDNF protein was delivered via mechanical pumps, weight loss was a reported side effect.42 Indeed, the liability of GDNF administration to cause weight loss was the impetus for our previous study15 and the current study. Thus, the finding that nigral GDNF overexpression induces robust weight loss might be a concern for rAAV-GDNF treatment of PD. For example, in an extremely promising study in rhesus monkeys, unilateral delivery of lentivirus encoding GDNF to both striatum and SN resulted in remarkable clinical efficacy.43 The present results clearly suggest that direct injections in the SN are not advisable because weight loss in PD patients would be a confounding effect at the very least. Along these lines, there are data indicating that vector-mediated GDNF expression exclusively in the striatum is also efficacious.21,42,44 Recombinant AAV serotypes other than rAAV2 do display significant retrograde transport to SN.20 However, PD patients have markedly reduced striatal DA innervation which would greatly limit SN transduction from striatal injections in PD. Moreover, we have observed that there is a threshold level of nigral GDNF expression needed to induce weight loss in rats (F.P. Manfredsson and R.J. Mandel, unpublished results). Thus, the data reported here suggest that any www.moleculartherapy.org vol. 17 no. 6 june 2009


rAAV-GDNF clinical development for PD should focus on striatal injection sites such as the postcommissural putamen.42 Therefore, these data should not be interpreted to preclude further clinical development of rAAV-GDNF for treatment of PD. Another clinical implication of our data may be the potential use of the nigral rAAV-GDNF injection strategy as an obesity treatment. Thus far, we have not been able to document any negative side effects other than evidence of increased metabolism and weight loss in obese rats overexpressing GDNF bilaterally in the nigrostriatal tract. This lack of obvious untoward GDNF-induced effects by no means indicates that there are no negative consequences of bilateral nigrostriatal GDNF overexpression, and much more study is required before concluding that this strategy might be a safe treatment for obesity. In conclusion, the data reported here extend our knowledge of the multiple effects of GDNF overexpression in the brain that need to be taken into account when attempting this paradigm in humans.

Materials And Methods Animals. Two Sprague-Dawley rats (Harlan, Indianapolis, IN) were used

for a preliminary anatomical experiment to determine the distribution of intracellular transgene (GFP) and secreted GDNF in the nigrostriatal tract. For the main study, male 25-month-old F344 X BN rats were obtained from Harlan Sprague-Dawley (Indianapolis, IN). Upon arrival the rats weighed 525 ± 4 g and were quarantined for 1 week before any testing. Animals were cared for in accordance with the principles of the Guide to the Care and Use of Experimental Animals. Rats were housed with a 12:12 hours light:dark cycle (0700–1900 hours), and food and water were available ad libitum. Body weight and food consumption were recorded daily.

rAAV2/5-production. The AAV backbone was identical to that used previously.15 Briefly, the GDNF complementary DNA was under the control of the cytomegalovirus/chicken β-actin promoter hybrid, followed immediately downstream by the woodchuck hepatitis post-transcriptional regulatory element. The entire transcription cassette was flanked by AAV2 terminal repeats. The GFP vector consisted of the humanized enhanced GFP complementary DNA downstream of the same promoter as in the GDNF plasmid. The virus used to evaluate GDNF transduction properties consisted of the same GDNF cassette used in the main study followed downstream by the herpes simplex virus thymidine kinase promoter controlling the expression of enhanced GFP. The viruses were created by the co-transfection of the AAV plasmid with the pXYZ5 helper plasmid which carries the AAV2 rep and AAV5 cap genes in addition to adenovirus helper genes. The virus was subsequently purified by iodixanol step gradients and sepharose Q column chromatography as described previously.45 Vector titers were determined using dot blot assays as described.45 rAAV2/5 GDNF and rAAV2/5 GFP were 3.8 × 1012 genome copies/ml (gc/ml) and 4.5 × 1013 gc/ml, respectively. The rAAV2/5 GDNF/ enhanced GFP virus titer was 2.5 × 1012 gc/ml. The virus stock was at least 99% pure as judged by silver-stained sodium dodecyl sulfate acrylamide gel fractionation. Intracerebral administration of rAAV2/5. All surgical procedures were

performed using aseptic techniques and an isoflurane gas anesthesia machine. Following anesthesia, the incision site was pretreated with a subcutaneous injection of Marcaine, and the rat was thereafter placed in a stereotaxic frame (Kopf Instruments, Tujunga, CA) and maintained continuously under isoflurane during the injection procedure. Injections were performed with a 10-µl Hamilton syringe fitted with a pulled glass micropipette with an opening of ~ 60–80 µm. The speed of the injection

Molecular Therapy vol. 17 no. 6 june 2009


was accurately controlled by an infusion pump that pushes a piston which in turn depresses the Hamilton syringe. Each animal in the study groups received one injection in both hemispheres in either the hypothalamus or the SN, whereas the group used to evaluate the GDNF transduction pattern in nigral injections only received a unilateral injection in the left hemisphere (n = 2). The coordinates for hypothalamic injections were anterior–posterior −1.8 mm, from bregma, medial–lateral ± 1.0 mm from bregma, and −9.0 mm dorsoventral from the dura. The total injection volume was 0.25 µl at a rate of 0.125 µl/minute. The coordinates for the nigral injections were anterior–posterior −5.4 mm, medial–lateral ± 2.0 mm, and dorsoventral −7.2 mm from dura. The injection rate was 0.5 µl/minute with a total injected volume of 2 µl. Following all injections, the glass micropipette was left in place for an additional 5 minutes to allow for viral diffusion before being slowly removed from the brain. Motor activity monitoring. For the activity measurements, the animals were divided into three groups, each containing randomly selected animals from each group and rats were placed individually in home cages. Following 1 week of acclimation to single housing and the recording room, the activity was measured for three full days per recording group ~40 days following the viral injections. Activity was measured by total beam breaks over 72 hours using a home-cage photobeam activity system (San Diego Instruments, San Diego, CA). Tissue harvesting and preparation. At the end point of the study, the animals were deeply anesthetized with pentobarbital. Blood samples were collected by heart puncture and serum was harvested by a 10-minute centrifugation in serum separator tubes. Serum samples were immediately treated with 0.5 ml 0.5 mol/l perchloric acid for subsequent highperformance liquid chromatography (HPLC) analysis. The animals were thereafter decapitated and their brains removed. The brains were rapidly removed and placed in a block; the brain was then sectioned coronally at the level of the cerebral peduncles. The region posterior to the cut, containing the midbrain, was immediately placed in ice-cold 4% paraformaldehyde in 0.01 mol/l phosphate-buffered saline (PBS) and postfixed for 24 hours, following fixation the brains were treated as described below. The various nuclei were thereafter dissected from the fresh tissue samples. The hypothalamus was removed by making an incision medial to the piriform lobes and caudal to the optic chiasm to a depth of 2–3 mm. The striatum and NAcc were dissected by visual guidance from the remainder of the sample. The collected samples were thereafter homogenized briefly in 0.3 ml 10 mmol/l Tris-HCl, pH 6.8, 2% sodium dodecyl sulfate, and 0.08 µg/ml okadaic acid. Protease inhibitors, 1 mmol/l phenylmethylsulphonyl fluoride, 0.1 mmol/l benzamidine, and 2 µmol/l leupeptin were also present. The homogenate was aliquoted and fractions destined for RNA analysis were immediately frozen at −70 °C, samples for HPLC analysis were treated with 0.5 ml 0.5 mol/l perchloric acid, and fractions to be used for protein analysis were immediately boiled and stored frozen at −70 °C. Furthermore, brown adipose tissue, and perirenal and retroperitoneal white adipose tissue were excised and stored at −70 °C. One animal from each group was excluded from biochemical evaluation, instead, following deep anesthesia the animal was perfused through the ascending aorta with sterile Tyrode’s solution, followed by 350 ml of ice-cold 4% paraformaldehyde in 0.01 mol/l PBS buffer. Brains were rapidly removed and postfixed for 4 hours in the same solution, and then transferred to a 30% sucrose solution in 0.01 mol/l PBS solution. Brains destined for immunohistochemistry were cut into 40 μm thick sections using a freezing stage sliding microtome. GDNF enzyme-linked immunosorbent assay. Tissue samples from the dis-

sections were analyzed using a GDNF Emaxkit (Promega, Madison, WI). Tissues were homogenized in 100 µl buffer according to the kit protocol, using a Tissue-Tearor roto/stator-type homogenizer (BioSpec Products,

989



Bartlesville, OK). The hypothalamus was assayed at a 1:500 dilution, striatum at 1:50 or 1:100, and NAcc at 1:9 dilutions, and were read with a VersaMax microplate reader (Molecular Devices, Sunnyvale, CA). Raw plate values were adjusted for tissue weight. Immunohistochemistry. Floating sections were washed with 0.01 mol/l PBS and then treated for 15 minutes with 0.5% H2O2 + 10% methanol in 0.01 mol/l PBS. The sections were then treated with 3% serum + 0.1% Triton X-100 in 0.01 mol/l PBS, and then incubated overnight at room temperature with either a 1:2,000 dilution of a mouse anti-TH antibody, a 1:2,000 rabbit anti-GFP antibody (Chemicon, Temecula, CA), a 1:1,000 dilution of goat anti-human GDNF (R&D Systems, Minneapolis, MN), a 1:100 dilution of rabbit anti-phospho-p44/42 MAP kinase (p-ERK) (Cell Signaling Technology, Danvers, MA), or a 1:800 dilution of mouse antioxytocin (courtesy of Harold Gainer46). Following the incubation, the tissue was washed and incubated for 2 hours at room temperature with an appropriate secondary antibody directed against the species in which the primary antibody was raised (biotinylated horse anti-mouse and goat anti-rabbit, AMCA conjugated horse anti-mouse and Texas red conjugated goat anti-rabbit (Vector Laboratories, Burlingame, CA), biotinylated donkey anti-goat (Santa Cruz Biotechnology, Santa Cruz, CA), Alexa Fluor 488 donkey anti-rabbit and Alexa Fluor 594 donkey anti-goat (Invitrogen, Carlsbad, CA). The reactions were visualized using a avidin–biotin peroxidase complex (Vector Laboratories) followed by incubation with NovaRED substrate (Vector Laboratories). Sections were mounted on subbed slides, dehydrated in ascending alcohol concentrations, cleared in xylene, and coverslipped with Permount. Fluorescently stained sections were mounted and coverslipped in fluorescent mounting media without dehydration. Western analysis. TH and AT1 protein levels were determined using west-

ern blot as described previously.47 Briefly, an equal amount of protein for each sample was separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred onto a polyvinylidene difluoride membrane and blocked with 5% skimmed milk in Tris-buffered saline containing 0.1% Tween 20. Blots were incubated overnight at 4 °C with anti-TH polyclonal antibody (Pel-Freez Biologicals, Rogers, AR) or anti-AT1 antibodies (Abcam, Cambridge, MA). The bound antibodies were detected by chemiluminescence. HPLC. Brain tissues were assayed for NE, DA, and DOPAC, and serum samples were analyzed for NE content, by HPLC as described previously.48 Briefly, samples were stored at −80 °C until HPLC analysis. Samples were prepared for analysis by adding 450 μl of 0.1 N perchloric acid and 50 μl (750 ng/ml) of dihydroxybenzylamine and used as the internal standard to correct for changes in tissue concentration due to sample preparation. The sample was allowed to thaw then thoroughly homogenized using a microhomogenizer for 10–20 seconds. An aliquot equivalent to ~3 mg sample was thereafter transferred to a separate tube containing 0.1 N perchloric acid to yield a total volume of 1 ml. The diluted sample was centrifuged at 14,000 rpm for 12 minutes at 4 °C. The supernatant was then filtered through a 0.2-μm nylon syringe filter, and collected into an HPLC vial. NE, DA, and DOPAC levels were analyzed on a Beckman Gold System using a C18 Waters Symmetry column (3.9 mm × 15 cm), and an ESA Coulochem electrochemical detector equipped with a sensitive 5011A analytical cell. The detector settings were E1 = −75 mV, and E2 = 300 mV. The mobile phase was composed as follows: 8.2 mmol/l citric acid, 8.5 mmol/l sodium phosphate monobasic, 0.25 mmol/l ethylenediaminetetraacetic acid, 0.30 mmol/l sodium octyl sulfate, and 7.0% acetonitrile, pH adjusted to 3.5 then filtered through a 0.2-μm filter membrane. The flow rate was set at 1.5 ml/minute with a 30 μl injection volume. Relative quantitative reverse transcriptase–PCR. TH and neuropeptide Y messenger RNA levels were determined using northern blot analysis as described previously.47 Briefly, total cellular RNA was extracted using

990

ri-reagent (Sigma, St Louis, MO), and the isolated RNA was quantified by T spectrophotometry. Membranes were hybridized with 32P random primer-generated probes. After hybridization, the membranes were washed and exposed to phospho screen for 24 hours using Phospho Imager (Molecular Dynamics, Sunnyvale, CA). Screens were scanned and analyzed using Image Quant software (Molecular Dynamics). Statistical analysis. All differences between groups were assessed using

analysis of variance. The body weight and FI data were analyzed via repeated measures analysis. Post hoc differences were evaluated using the Bonferroni/Dunn post hoc test. Individual contrasts in repeated measures analysis of variance were only undertaken if there was a significant GROUP × TIME interaction.

Acknowledgments This work was supported by NINDS grant 5P01NS036302-10 (R.J.M. and N.M.) and by the Medical Research Service of the Department of Veterans Affairs (N.T. and P.J.S.). We thank Izzie Williams and Bobbi Johnson for technical assistance. N.M. is an inventor of patents related to recombinant AAV technology and owns equity in a gene therapy company that is commercializing AAV for gene therapy applications.

References

1. Lin, LF, Doherty, DH, Lile, JD, Bektesh, S and Collins, F (1993). GDNF: a glial cell line-derived neurotrophic factor for midbrain dopaminergic neurons. Science 260: 1130–1132. 2. Pascual, A, Hidalgo-Figueroa, M, Piruat, JI, Pintado, CO, Gomez-Diaz, R and LopezBarneo, J (2008). Absolute requirement of GDNF for adult catecholaminergic neuron survival. Nat Neurosci 11: 755–761. 3. Bespalov, MM and Saarma, M (2007). GDNF family receptor complexes are emerging drug targets. Trends Pharmacol Sci 28: 68–74. 4. Jing, S, Wen, D, Yu, Y, Holst, PL, Luo, Y, Fang, M, et al. (1996). GDNF-induced activation of the ret protein tyrosine kinase is mediated by GDNFR-α, a novel receptor for GDNF. Cell 85: 1113–1124. 5. Treanor, JJ, Goodman, L, de Sauvage, F, Stone, DM, Poulsen, KT, Beck, CD, et al. (1996). Characterization of a multicomponent receptor for GDNF. Nature 382: 80–83. 6. Trupp, M, Scott, R, Whittemore, SR and Ibanez, CF (1999). Ret-dependent and -independent mechanisms of glial cell line-derived neurotrophic factor signaling in neuronal cells. J Biol Chem 274: 20885–20894. 7. Mograbi, B, Bocciardi, R, Bourget, I, Busca, R, Rochet, N, Farahi-Far, D et al. (2001). Glial cell line-derived neurotrophic factor-stimulated phosphatidylinositol 3-kinase and Akt activities exert opposing effects on the ERK pathway: importance for the rescue of neuroectodermic cells. J Biol Chem 276: 45307–45319. 8. Zigmond, MJ (2006). Triggering endogenous neuroprotective mechanisms in Parkinson’s disease: studies with a cellular model. J Neural Transm Suppl 439–442. 9. Golden, JP, DeMaro, JA, Osborne, PA, Milbrandt, J and Johnson, EM Jr (1999). Expression of neurturin, GDNF, and GDNF family-receptor mRNA in the developing and mature mouse. Exp Neurol 158: 504–528. 10. Zhang, Z, Miyoshi, Y, Lapchak, PA, Collins, F, Hilt, D, Lebel, C, et al. (1997). Dose response to intraventricular glial cell line-derived neurotrophic factor administration in parkinsonian monkeys. J Pharmacol Exp Ther 282: 1396–1401. 11. Kordower, JH, Palfi, S, Chen, EY, Ma, SY, Sendera, T, Cochran, EJ et al. (1999). Clinicopathological findings following intraventricular glial-derived neurotrophic factor treatment in a patient with Parkinson’s disease. Ann Neurol 46: 419–424. 12. Nutt, JG, Burchiel, KJ, Comella, CL, Jankovic, J, Lang, AE, Laws, ER Jr et al. (2003). Randomized, double-blind trial of glial cell line-derived neurotrophic factor (GDNF) in PD. Neurology 60: 69–73. 13. Patel, NK, Bunnage, M, Plaha, P, Svendsen, CN, Heywood, P and Gill, SS (2005). Intraputamenal infusion of glial cell line-derived neurotrophic factor in PD: a two-year outcome study. Ann Neurol 57: 298–302. 14. Slevin, JT, Gash, DM, Smith, CD, Gerhardt, GA, Kryscio, R, Chebrolu, H, et al. (2007). Unilateral intraputamenal glial cell line-derived neurotrophic factor in patients with Parkinson disease: response to 1 year of treatment and 1 year of withdrawal. J Neurosurg 106: 614–620. 15. Tumer, N, Scarpace, PJ, Dogan, MD, Broxson, CS, Matheny, M, Yurek, DM, et al. (2006). Hypothalamic rAAV-mediated GDNF gene delivery ameliorates age-related obesity. Neurobiol Aging 27: 459–470. 16. Lindvall, O and Bjorklund, A (1978). Anatomy of the dopaminergic neuron systems in the rat brain. Adv Biochem Psychopharmacol 19: 1–23. 17. Ungerstedt, U (1971). Adipsia and aphagia after 6-hydroxydopamine induced degeneration of the nigro-striatal dopamine system. Acta Physiol Scand Suppl 367: 95–122. 18. Sotak, BN, Hnasko, TS, Robinson, S, Kremer, EJ and Palmiter, RD (2005). Dysregulation of dopamine signaling in the dorsal striatum inhibits feeding. Brain Res 1061: 88–96. 19. Szczypka, MS, Kwok, K, Brot, MD, Marck, BT, Matsumoto, AM, Donahue, BA, et al. (2001). Dopamine production in the caudate putamen restores feeding in dopamine‑deficient mice. Neuron 30: 819–828. 20. Burger, C, Gorbatyuk, OS, Velardo, MJ, Peden, CS, Williams, P, Zolotukhin, S, et al. (2004). Recombinant AAV viral vectors pseudotyped with viral capsids from serotypes 1, 2, and 5 display differential efficiency and cell tropism after delivery to different regions of the central nervous system. Mol Ther 10: 302–317.

www.moleculartherapy.org vol. 17 no. 6 june 2009


21. Kirik, D, Rosenblad, C, Bjorklund, A and Mandel, RJ (2000). Long-term rAAV-mediated gene transfer of GDNF in the rat Parkinson’s model: intrastriatal but not intranigral transduction promotes functional regeneration in the lesioned nigrostriatal system. J Neurosci 20: 4686–4700. 22. Gorbatyuk, OS, Milner, TA, Wang, G, Regunathan, S and Reis, DJ (2001). Localization of agmatine in vasopressin and oxytocin neurons of the rat hypothalamic paraventricular and supraoptic nuclei. Exp Neurol 171: 235–245. 23. Khan, AM, Ponzio, TA, Sanchez-Watts, G, Stanley, BG, Hatton, GI and Watts, AG (2007). Catecholaminergic control of mitogen-activated protein kinase signaling in paraventricular neuroendocrine neurons in vivo and in vitro: a proposed role during glycemic challenges. J Neurosci 27: 7344–7360. 24. Georgievska, B, Kirik, D and Bjorklund, A (2002). Aberrant sprouting and downregulation of tyrosine hydroxylase in lesioned nigrostriatal dopamine neurons induced by long-lasting overexpression of glial cell line derived neurotrophic factor in the striatum by lentiviral gene transfer. Exp Neurol 177: 461–474. 25. Aoi, M, Date, I, Tomita, S and Ohmoto, T (2000). The effect of intrastriatal single injection of GDNF on the nigrostriatal dopaminergic system in hemiparkinsonian rats: behavioral and histological studies using two different dosages. Neurosci Res 36: 319–325. 26. Hoane, MR, Gulwadi, AG, Morrison, S, Hovanesian, G, Lindner, MD and Tao, W (1999). Differential in vivo effects of neurturin and glial cell-line-derived neurotrophic factor. Exp Neurol 160: 235–243. 27. Lapchak, PA, Jiao, S, Collins, F and Miller, PJ (1997). Glial cell line-derived neurotrophic factor: distribution and pharmacology in the rat following a bolus intraventricular injection. Brain Res 747: 92–102. 28. Gash, DM, Zhang, Z, Cass, WA, Ovadia, A, Simmerman, L, Martin, D, et al. (1995). Morphological and functional effects of intranigrally administered GDNF in normal rhesus monkeys. J Comp Neurol 363: 345–358. 29. Horger, BA, Nishimura, MC, Armanini, MP, Wang, LC, Poulsen, KT, Rosenblad, C, et al. (1998). Neurturin exerts potent actions on survival and function of midbrain dopaminergic neurons. J Neurosci 18: 4929–4937. 30. Palmiter, RD (2007). Is dopamine a physiologically relevant mediator of feeding behavior? Trends Neurosci 30: 375–381. 31. Li, H, Matheny, M, Nicolson, M, Tumer, N and Scarpace, PJ (1997). Leptin gene expression increases with age independent of increasing adiposity in rats. Diabetes 46: 2035–2039. 32. Bassareo, V and Di Chiara, G (1999). Differential responsiveness of dopamine transmission to food-stimuli in nucleus accumbens shell/core compartments. Neuroscience 89: 637–641. 33. Hernandez, L and Hoebel, BG (1988). Feeding and hypothalamic stimulation increase dopamine turnover in the accumbens. Physiol Behav 44: 599–606. 34. Martel, P and Fantino, M (1996). Mesolimbic dopaminergic system activity as a function of food reward: a microdialysis study. Pharmacol Biochem Behav 53: 221–226.

Molecular Therapy vol. 17 no. 6 june 2009


35. Reimsnider, S, Manfredsson, FP, Muzyczka, N and Mandel, RJ (2007). Time course of transgene expression after intrastriatal pseudotyped rAAV2/1, rAAV2/2, rAAV2/5, and rAAV2/8 transduction in the rat. Mol Ther 15: 1504–1511. 36. Rosenblad, C, Georgievska, B and Kirik, D (2003). Long-term striatal overexpression of GDNF selectively downregulates tyrosine hydroxylase in the intact nigrostriatal dopamine system. Eur J Neurosci 17: 260–270. 37. Morton, GJ, Cummings, DE, Baskin, DG, Barsh, GS and Schwartz, MW (2006). Central nervous system control of food intake and body weight. Nature 443: 289–295. 38. Liu, HS, Jan, MS, Chou, CK, Chen, PH and Ke, NJ (1999). Is green fluorescent protein toxic to the living cells? Biochem Biophys Res Commun 260: 712–717. 39. Detrait, ER, Bowers, WJ, Halterman, MW, Giuliano, RE, Bennice, L, Federoff, HJ, et al. (2002). Reporter gene transfer induces apoptosis in primary cortical neurons. Mol Ther 5: 723–730. 40. Kirik, D, Rosenblad, C, Burger, C, Lundberg, C, Johansen, TE, Muzyczka, N, et al. (2002). Parkinson-like neurodegeneration induced by targeted overexpression of α-synuclein in the nigrostriatal system. J Neurosci 22: 2780–2791. 41. Klein, RL, Dayton, RD, Leidenheimer, NJ, Jansen, K, Golde, TE and Zweig, RM (2006). Efficient neuronal gene transfer with AAV8 leads to neurotoxic levels of tau or green fluorescent proteins. Mol Ther 13: 517–527. 42. Isacson, O and Kordower, JH (2008). Future of cell and gene therapies for Parkinson’s disease. Ann Neurol 64: S122–S138. 43. Kordower, JH, Emborg, ME, Bloch, J, Ma, SY, Chu, Y, Leventhal, L, et al. (2000). Neurodegeneration prevented by lentiviral vector delivery of GDNF in primate models of Parkinson’s disease. Science 290: 767–773. 44. Eslamboli, A, Georgievska, B, Ridley, RM, Baker, HF, Muzyczka, N, Burger, C, et al. (2005). Continuous low-level glial cell line-derived neurotrophic factor delivery using recombinant adeno-associated viral vectors provides neuroprotection and induces behavioral recovery in a primate model of Parkinson’s disease. J Neurosci 25: 769–777. 45. Zolotukhin, S, Potter, M, Zolotukhin, I, Sakai, Y, Loiler, S, Fraites, TJ Jr et al. (2002). Production and purification of serotype 1, 2, and 5 recombinant adeno-associated viral vectors. Methods 28: 158–167. 46. Ben-Barak, Y, Russell, JT, Whitnall, M, Ozato, K and Gainer, H (1984). Phylogenetic cross-reactivities of monoclonal antibodies produced against rat neurophysin. Cell Mol Neurobiol 4: 339–349. 47. Erdem, SR, Demirel, HA, Broxson, CS, Nankova, BB, Sabban, EL and Tumer, N (2002). Effect of exercise on mRNA expression of select adrenal medullary catecholamine biosynthetic enzymes. J Appl Physiol 93: 463–468. 48. Manfredsson, FP, Burger, C, Sullivan, LF, Muzyczka, N, Lewin, AS and Mandel, RJ (2007). rAAV-mediated nigral human parkin over-expression partially ameliorates motor deficits via enhanced dopamine neurotransmission in a rat model of Parkinson’s disease. Exp Neurol 207: 289–301. 49. Swanson, LW (1999). Brain Maps: Structure of the Rat Brain, 2nd edn. Elsevier: Amsterdam.

991