A Food Restriction Protocol that Increases Drug Reward Decreases TrkB in the Ventral Tegmental Area, with No Effect on BDNF or TrkB Protein Levels in Dopaminergic Forebrain Regions

Abstract

Food restriction (FR) decreases BDNF expression in hypothalamic and hindbrain regions that regulate feeding and metabolic efficiency, while increasing expression in hippocampal and neocortical regions. Drugs of abuse alter BDNF expression within the mesocorticolimbic dopamine (DA) pathway, and modifications of BDNF expression within this pathway alter drug-directed behavior. Although FR produces a variety of striatal neuroadaptations and potentiates the rewarding effects of abused drugs, the effects of FR on BDNF expression and function within the DA pathway are unknown. The primary purpose of the present study was to examine the effect of FR on protein levels of BDNF and its TrkB receptor in component structures of the mesocorticolimbic pathway. Three to four weeks of FR, with stabilization of rats at 80% of initial body weight, did not alter BDNF or TrkB levels in nucleus accumbens, caudate-putamen, or medial prefrontal cortex. However, FR decreased TrkB levels in the ventral tegmental area (VTA), without change in levels of BDNF protein or mRNA. The finding that FR also decreased TrkB levels in substantia nigra, with elevation of BDNF protein, suggests that decreased TrkB in VTA could be a residual effect of increased BDNF during an earlier phase of FR. Voltage-clamp recordings in VTA DA neurons indicated decreased glutamate receptor transmission. These data might predict lower average firing rates in FR relative to ad libitum fed subjects, which would be consistent with previous evidence of decreased striatal DA transmission and upregulation of postsynaptic DA receptor signaling. However, FR subjects also displayed elevated VTA levels of phospho-ERK1/2, which is an established mediator of synaptic plasticity. Because VTA neurons are heterogeneous with regard to neurochemistry, function and target projections, the relationship(s) between the three changes observed in VTA, and their involvement in the augmented striatal and behavioral responsiveness of FR subjects to drugs of abuse, remains speculative.

BDNF belongs to the family of neurotrophins that have well established roles in neurodevelopmental processes. There is increasing evidence that BDNF, acting as a classical neurotrophin and/or excitatory transmitter with rapid synaptic effects (Thoenen, 2000; Rose et al., 2004), also modulates an array of behavioral functions in the mature CNS including mood (Eisch et al, 2003; Nestler and Carlezon, 2006), drug addiction (Horger et al, 1999; Pierce and Bari, 2001; Lu et al., 2004; Graham et al, 2007), learning (Rattiner et al., 2004; Rademacher et al., 2006), and ingestive behavior (Lebrun et al., 2006; Wang et al, 2007a,b). Several of the adaptive functional changes associated with food deprivation are mediated or modulated by decreased BDNF activity. For example, fasting is accompanied by decreases in BDNF protein and mRNA levels in hypothalamic and hindbrain nuclei that regulate energy homeostasis (Xu et al., 2003; Bariohay et al., 2005), and microinjection of BDNF into medial hypothalamic nuclei (Wang et al., 2007a,b) or the dorsal vagal complex (Bariohay et al., 2005) decreases food intake. Complementing these observations, mice that are heterozygous for targeted disruption of BDNF exhibit a chronic hyperphagia and obesity (Fox and Byerly, 2004).

Among the several members of the tyrosine receptor kinase family that bind neurotrophins, BDNF binds with high affinity to TrkB (Squinto et al., 1991; Huang and Reichert, 2003); only the full-length isoform (TrkB-FL) has intracellular tyrosine kinase activity. Mutant mice that express very low levels of TrkB-FL throughout the brain, exhibit hyperphagia and obesity (Xu et al., 2003). Moreover, unlike control mice which decrease food intake when switched from a low fat to moderate fat diet, the TrkB “hypomorphs” increase intake. Overall, BDNF and TrkB signaling appear to be important regulators of ingestive behavior and metabolic function.

In contrast to the food deprivation-induced decreases in BDNF observed in brain regions that regulate energy homeostasis, long-term food restriction increases BDNF protein levels in neocortical, hippocampal and dorsal striatal regions (Lee et al., 2000; Duan et al., 2001). These latter responses have been implicated in the neuroprotective effects of food restriction and are proposed to mediate the cognitive-enhancing effects. It is therefore possible that opposite and neuroanatomically distributed changes in BDNF function during periods of negative energy balance modulate different adaptive functional responses.

Chronic food restriction (FR) is associated with increased behavioral sensitivity to drugs of abuse, as documented in self-administration (Carroll and Meisch, 1984), conditioned place preference (e.g., Bell et al., 1997; Stuber et al., 2002; Cabib et al., 2000; Liu et al., 2011), electrical brain stimulation reward (e.g., Cabeza de Vaca and Carr, 1998; Cabeza de Vaca et al., 2004), and motor activity assays (e.g., Deroche et al., 1993; Carr et al., 2003). These phenomena are likely to reflect neuroadaptations that otherwise facilitate foraging, incentive motivation, and reward-related learning during periods of food scarcity. Ventral tegmental area (VTA) dopamine (DA) neurons that innervate nucleus accumbens (NAc) are fundamentally involved in the unconditioned and conditioned rewarding effects of most abused drugs and express high levels of mRNA for BDNF and TrkB (Merlio et al., 1992; Conner et al., 1997; Furukawa et al., 1998), and both protein products coexpress with tyrosine hydroxylase (Seroogy et al., 1994; Hoover et al, 2007). NAc expresses relatively little BDNF mRNA but has high levels of BDNF protein (Conner et al., 1997), most of which is of VTA and medial prefrontal cortical (mPFC) origin (Altar et al., 1997; McGinty et al., 2010). While microinfusions of BDNF in VTA, NAc, and mPFC have been shown to modify behavioral responsiveness to drugs of abuse (Horger et al., 1999; Lu et al, 2004; Graham et al., 2007; Berglind et al., 2007; McGinty et al., 2010), effects of FR on BDNF and TrkB in this system, and the possible modulatory effects of altered BDNF and TrkB function on drug reward under the condition of FR, have not been investigated.

In the light of increasing evidence that changes in BDNF function may mediate a variety of adaptive responses to FR, we introduced subjects to a FR regimen that enhances rewarding (Cabeza de Vaca and Carr, 1998; Cabeza de Vaca et al. 2004), place preference conditioning (Liu et al., 2011) and motor-activating effects of abused drugs (Cabeza de Vaca and Carr, 1998; Carr et al., 2003; Liu et al., 2011), then investigated the effects on BDNF and TrkB in several DA-related brain regions. In an initial experiment, an enzyme immunoassay was used to measure BDNF protein levels in mPFC, NAc, caudate-putamen (CPu), VTA and substantia nigra (SN) of FR and ad libitum fed (AL) rats. In a second experiment, Western immunoblotting was used to measure TrkB protein levels in these same forebrain and midbrain regions. In a third experiment, real-time RT PCR was used to measure BDNF mRNA levels in the VTA. In a fourth experiment, and as follow-up to an observation of decreased TrkB protein in the VTA of FR rats, levels of activated signaling molecules downstream of TrkB were measured. Binding of BDNF to TrkB activates several intracellular signaling pathways, with the mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) and phosphoinositide 3-kinase (PI3K)/Akt pathways clearly implicated in neurotrophic effects (Numakawa et al., 2010). In a fifth experiment, voltage-clamp recording was used to measure evoked excitatory postsynaptic currents (EPSCs) in VTA neurons in midbrain slices obtained from AL and FR rats.

EXPERIMENTAL PROCEDURES

Subjects and food restriction

All subjects were mature male Sprague-Dawley rats (Taconic Farms, Germantown, NY) weighing 350–400 g at the start of the experiment. Rats were housed, under a 12 h light:dark photoperiod with lights on at 0700 h, in a central animal facility in individual plastic cages with bedding and free access to water and standard lab pellets (Laboratory Rodent Diet #5001, Lab Diet) except when restricted feeding conditions applied (see below). Experimental procedures were approved by the Institutional Animal Care and Use Committee at the New York University School of Medicine and were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23) revised 1996. All efforts were made to minimize the number of animals used and their suffering.

Several days after their arrival in the central animal facility, half of the rats in Experiments 1–4 were switched to a restricted feeding regimen whereby a single 10 g meal of Purina rat chow was delivered at approximately 1700 h each day. These rats continued to have ad libitum access to water. Once body weight had declined by 20% (approximately 15 days) daily food allotments were titrated to maintain body weight at this value for an additional 7–14 days prior to sacrifice. This feeding regimen and time of analysis were the same as used in the prior studies where behavioral, immediate-early gene and striatal cell-signaling responses to psychostimulant and direct DA receptor agonist challenge were found to be augmented in FR relative to AL rats (e.g., Cabeza de Vaca and Carr, 1998; Carr et al., 2003; Haberny et al., 2004). For all experiments, rats were briefly narcotized with CO₂, decapitated, and brains were rapidly removed and immediately frozen in powdered dry ice. As in the prior studies, sacrifice occurred during the light phase 4–5 h before the FR subjects were to receive their scheduled daily meal. Subjects were maintained in a quiet location within their home cages in the period preceding sacrifice.

Tissue sampling

Five hundred micrometers sections were cut using an IEC Minotome cryostat. Under an Olympus dissecting microscope, different brain regions (mPFC, NAc, CPu, VTA, and SN) were micropunched from a series of 500 μm frozen coronal sections. For the RT-PCR experiment, all equipment and surfaces were treated with RNase Zap (Ambion, Austin, TX) to minimize RNase contamination.

BDNF immunoassay

Brain-derived neurotrophic factor protein levels were determined using the commercially available BDNF Emax^® ImmunoAssay Systems (Promega, Madison, WI). The ELISAs were performed according to the manufacturer’s protocol. Tissue samples were homogenized in lysis buffer, followed by centrifugation and protein determination using BCA reagent kit as described by the manufacturer (Pierce, Rockford, IL). Aliquots of supernatants were stored at −80°C. On the day of the immunoassay, aliquots were thawed, and added to a 96-well immunoplate precoated with human BDNF specific monoclonal antibody. The plate was incubated at room temperature for 2 h with shaking. Anti-BDNF monoclonal antibody was used as the capture Ab and anti-BDNF pAb was used as reporter Ab. After washing, the amount of specifically bound pAb was detected using a species-specific anti-IgY antibody conjugated to horseradish peroxidase as a tertiary reactant. The unbound conjugate was removed by washing followed by incubation with a chromogenic substrate. Absorbance of samples was measured at 450 nm and read using OPTImax microplate reader (Molecular Devices Corp., Sunnyvale, CA). All samples were assayed in duplicate. The readings were normalized to the amount of protein applied. Samples of forebrain and midbrain regions were obtained from 4 and 6 animals per diet group, respectively. Results obtained in each brain region were analyzed as separate experiments using Student’s t-test.

Lysate preparation and Western blotting

Tissue samples were homogenized in 10 volumes of 50 mM Tris–HCl, pH 7.5 containing 50 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1 mM Na₃VO₄, 40 mM β-glycerophosphate, 50 mM NaF and 5 mM Na₄P₂O₇, 1% Tx-100, 0.5 μM okadaic acid, 0.5% sodium deoxycholate, and 0.1% SDS, followed by centrifugation and protein determination using BCA reagent kit as described by the manufacturer (Pierce). Supernatants were mixed with 5× SDS-PAGE sample buffer, boiled for 5 min, cooled on ice, and kept at −80 °C until future use.

Protein (10 μg per lane) was separated by electrophoresis on precast 8% polyacrylamide gels (Bio-Rad). Precision Plus protein standard molecular weight markers (Bio-Rad) were also loaded to assure complete electrophoretic transfer and to estimate the size of bands of interest. The gels were transferred to nitrocellulose membrane (Osmonics, Minnetonka, MN) for 1 h, with a constant voltage of 100 V. Membranes were blocked for 1 h at room temperature with blocking buffer, 5% non-fat dry milk in 50 mM Tris–HCl, pH 7.5 containing 150 mM NaCl and 0.1% Tween 20 (TBS-T), then probed overnight at 4 °C using primary antibodies for target proteins or 1h at room temperature using primary antibody for α-tubulin. Antibodies used included mouse monoclonal anti-TrkB (1:2000; BD Biosciences, San Jose, CA), mouse monoclonal anti-phospho-(Thr202/Tyr204)-p44/42 ERK1/2 (1:1000; Cell Signaling, Beverly, MA), rabbit polyclonal anti-p42/44 ERK1/2 (1:2000; Cell Signaling, Beverly, MA), rabbit polyclonal anti-phospho-Ser473 Akt (1:1000; Cell Signaling, Beverly, MA), rabbit polyclonal anti-Akt (1:10000; Cell Signaling, Beveraly, MA), and mouse monoclonal anti-α-tubulin (1: 5,000; T6199, Sigma-Aldrich, St. Louis, MO).

After probing with primary antibodies and washing with TBS-T buffer (3 × 5 min), membranes were incubated with horseradish peroxidase conjugated anti-mouse IgG (Cell Signaling). Proteins were visualized using a chemiluminescence ECL kit (Pierce). Densitometric analysis of the bands was performed using the MCID imaging system (St. Catherines, Ontario, Canada). Total TrkB values were normalized to tubulin, p-ERK 1 and p-ERK 2 values were normalized to total ERK 1 and ERK 2, and p-Akt was normalized to total Akt. Target protein levels were expressed as percentage of the normalized control, which was the AL group. For measurements of TrkB, samples of striatal and midbrain regions were obtained from 6 animals per diet group, and 4 animals per diet group for mPFC. For measurements of signaling proteins, VTA samples were obtained from 6 animals per diet group. Results of each experiment were analyzed by Student’s t-test.

RNA extraction and quantitative real-time RT PCR

Total RNA was extracted from individual VTA tissue samples using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Sample RNA concentrations were determined by spectrophotometry.

Specific mRNA levels in each sample were measured on the iCycler^™ (Bio-Rad, Hercules, CA) in a final volume of 25 μl. Each reaction was performed using reagents from the one-step SYBR Green Quantitative RT-PCR kit (Sigma-Aldrich, St. Louis, MO) according to the standard protocol provided. Amplification consisted of 40 cycles of denaturation at 95 °C for 30 s, annealing at 56 °C for 30 s, and extension at 72 °C for 30 s, following one step of activation (pre-soak) at 48 °C for 30 min and initial denaturation at 94°C for 3 min. Fluorescence signals were monitored sequentially for each sample tube once per cycle at the end of extension. An external standard RNA concentration curve for each primer pair was generated using pooled RNA samples and verified by agarose gel electrophoresis. To evaluate RNA purity, each sample was tested without addition of reverse transcriptase and found to contain no genomic DNA above the threshold detection level. For each experiment, specificity of RT-PCR products was confirmed by analysis of melting curves produced by iCycler^™ showing the presence of a single species of DNA product per primer pair, and by agarose gel electrophoresis revealing single bands of the predicted molecular weight for each product. Table 1 details the sequence of primer pairs (purchased from GeneLink, Hawthorne, NY) and the sources for their design.

Table 1.

Sequences of primer pairs used for real-time RT-PCR with the sources for their design

β-actin	5′ GTC GTA CCA CTG GCA TTG TG 3′
	5′ GCC ATC TCT TGC TCG AAG TC 3′
	(Spangler et al., 2004)
BDNF	5′ AGC CTC CTC TGC TCT TTC TGC TGG A 3′
	5′ CTT TTG TCT ATG CCC CTG CAG CCT T 3′
	(Fang et al., 2000)

Sample mRNA levels for the BDNF gene were averaged from two separate experiments, each performed in duplicate. To correct for minor variability among samples, each subject’s average BDNF expression was normalized to its expression level of the housekeeping gene β-actin. VTA samples were obtained from 6 animals per diet group, relative expression levels for each subject were normalized to the corresponding mean of the AL group, and fold changes in induction were calculated and compared between groups using Student’s t-test.

Visualized whole-cell recording

Rats were deeply anesthetized with 50 mg/kg pentobarbital (i.p.) then perfused transcardially with ice-cold (~0 °C) modified artificial (aCSF) (in mM): 225 sucrose; 2.5 KCl; 0.5 CaCl₂; 7 MgCl₂; 28 NaHCO₃; 1.25 NaH₂PO₄; 7 glucose; 1 ascorbate; and 3 pyruvate (Koós and Tepper, 1999). After perfusion, the brain was removed into ice-cold modified aCSF for 1–2 min, then blocked and coronal midbrain slices (300 μm thickness) prepared using a Leica VT1200S vibrating blade microtome (Leica Microsystems, Bannockburn, IL). Slices were then transferred to a holding chamber for 30 min at 34°C, then slowly returned to room temperature in media containing (in mM): 125 NaCl; 2.5 KCl; 2 CaCl₂; 1 MgCl₂; 25 NaHCO₃; 1.25 NaH₂PO₄; 25 glucose; 1 ascorbate; and 3 pyruvate; pH 7.3–7.4, equilibrated with 95% O₂/5% CO₂ (Koós and Tepper, 1999; Avshalumov et al., 2005).

For whole-cell recording, slices were transferred to a submersion chamber (Warner Instruments LLC, Holliston, MA) maintained at 32 °C and superfused at 1.2 mL/min with bicarbonate-buffered aCSF containing (in mM): 124 NaCl; 3.7 KCl; 26 NaHCO₃; 1.3 MgSO₂; KH₂PO₄; 10 glucose; 2.4 CaCl₂ and saturated with 95% O₂/5% CO₂ (Avshalumov et al., 2005). Neurons were visualized using infrared differential interference contrast (IR-DIC) Olympus BX51WI microscope equipped with a 40× water-immersion objective (Olympus America, Center Valley, PA). Pipettes were made from 1.5 mm outer diameter borosilicate capillary tubing (World Precision Instruments, Sarasota, FL) using a Sutter P-97 Flaming/Brown micropipette puller (Sutter Instrument Company, Novato, CA). The pipette backfill solution for voltage-clamp recording contained (in mM): 117 Cs-methanesulfonate, 2 MgCl₂; 0.2 CaCl₂; 2 EGTA; 20 HEPES; 5 TEA-Cl; 3 Na₂-ATP; and 0.2 GTP; pH 7.3 (modified from Akopian and Witkovsky 2001; Ungless et al. 2001). Data were acquired using an Axopatch 200B amplifier and Digidata board 1322A controlled by Clampex 9.0; analysis will use ClampFit (Molecular Devices), including correction of liquid junction potential. Recorded cells in the VTA were identified as dopaminergic (DAergic) neurons by the presence of I_h current when membrane voltage was stepped in the hyperpolarizing direction to −140 mV in 10 mV increments (e.g., Avshalumov et al., 2005); this identification procedure was conducted immediately after a cell was patched, before Cs⁺ and TEA⁺ in the backfill solution infiltrated the cell.

Components of all physiological solutions were obtained from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA), as was picrotoxin, a GABA_A receptor blocker. All drugs used were water soluble and were prepared as aqueous stock solutions or dissolved directly in aCSF immediately before use.

RESULTS

Brain regional BDNF protein levels

Measurement of BDNF protein levels by ELISA indicated no significant differences between AL and FR rats in NAc (t(6)=0.3, p>.10), CPu (t(6)=1.8, p>.10), mPFC (t(6)=0.6, p>.10) or VTA (t(10)=0.7, p>.10). However, levels were higher in SN of FR relative to AL rats (t(10)=2.47, p<.05; Figure 1).

Figure 1.

Effect of chronic food restriction (FR) on brain regional BDNF protein levels. Levels of BDNF protein (mean ± S.E.M.), as determined by ELISA, expressed as percentage of control ad libitum (AL) fed group, in nucleus accumbens (NAc), caudate-putamen (CPu), medial prefrontal cortex (mPFC), ventral tegmental area (VTA) and substantia nigra (SN). (forebrain: n_AL = 4; n_FR = 4; midbrain: n_AL = 6; n_FR = 6), *p<.05

Brain regional TrkB protein levels

Semi-quantitative measurements of TrkB protein levels by Western analysis indicated that in the forebrain regions examined there were no significant differences between AL and FR rats in NAc (t(10)=0.6, p>.10), CPu (t(10)=0.57, p>.10), or mPFC (t(6)=0.27, p>.10; Figure 2). However, in midbrain regions FR subjects displayed decreased levels of TrkB protein in both the VTA (t(10)=2.21, p=.05) and SN (t(10)=3.48, p<.01) relative to AL subjects (Figure 3).

Figure 2.

Mean ± SEM ratio of full-length neurotrophin receptor TrkB protein/tubulin in nucleus accumbens (NAc), caudate-putamen (CPu) and medial prefrontal cortex (mPFC) of food-restricted (FR) rats determined by Western blot and expressed in comparison to the normalized control ad libitum (AL) fed group. Graphed results are displayed with representative immunoblots. (NAc and CPu: n_AL = 6; n_FR = 6; mPFC: n_AL = 4; n_FR = 4)

Figure 3.

Mean ± SEM ratio of full-length neurotrophin receptor TrkB protein/tubulin in ventral tegmental area (VTA) and substantia nigra (SN) of food-restricted (FR) rats determined by Western blot and expressed in comparison to the normalized control ad libitum (AL) fed group. Graphed results are displayed with representative immunoblots. ( n_AL = 6; n_FR = 6), *p<.05

BDNF mRNA in VTA

BDNF mRNA as measured in extracts of micropunched VTA using real-time RT PCR revealed no difference between AL and FR subjects (t(10)=0.27; data not shown).

ERK 1/2 and Akt signaling in VTA

Results of Western analyses indicated that under basal conditions, FR subjects displayed increased phosphorylation of ERK 1/2 (t(10)=2.52, p<.05) and a trend toward decreased phosphorylation of Akt(t(10)=1.78, p>.05) in VTA (Figure 4).

Figure 4.

Effect of chronic food restriction on phosphorylation of ERK 1/2 and Akt under basal conditions in ad libitum (AL) fed and food-restricted (FR) rats. Lysates were immunoblotted with anti-phospho p44/42 MAPK, anti-p44/42 MAPK, or anti-phospho-Ser473-Akt antibodies. Following densitometry, intensities of bands corresponding to phospho-proteins for each subject were divided by the intensities of the corresponding total protein (ERK or Akt) bands to correct for small differences in protein loading. Results (mean ± S.E.M.) are expressed in comparison to the normalized control, which was defined as the ad libitum fed group injected with vehicle. Graphed results are displayed with representative immunoblots. (n_AL = 6; n_FR = 6), *p ≤.05

Voltage-clamp recording of EPSCs in VTA DAergic neurons

Possible alterations in the responsiveness of VTA DAergic neurons to excitatory glutamate input after FR were examined using voltage-clamp recording in midbrain slices. Picrotoxin (100 μM; Johnson and North 1992) was included in all media to block concurrent activation of GABA_A receptors. It should be noted that slow GABA_B-receptor-and D2 autoreceptor-dependent currents do not interfere with fast EPSCs (Bonci and Malenka, 1999; Beckstead, 2004). In the presence of picrotoxin, local stimulation elicited robust excitatory postsynaptic currents (EPSCs) in FR and control DA cells held at −70 mV (Fig. 5A), at which EPSCs are exclusively AMPAR-dependent (Schilstrom et al., 2006). Under these conditions, EPSCs were lower, albeit not significantly, in VTA neurons from FR vs. AL rats (n = 4–5 neurons from 2–3 rats per group). When the holding potential was raised to +40 mV to reveal NMDA receptor contributions by removing the Mg²⁺ block of the receptor-gated channel (Bonci and Malenka, 1999), however, EPSC amplitude was a significant 50% lower in DAergic neurons from FR animals than AL controls (p < 0.05) (Fig. 5B), confirming decreased NMDAR and/or AMPAR transmission.

Figure 5.

Decreased glutamatergic excitability of VTA DA neurons in FR (food-restricted) vs. AL (ad libitum fed) controls. A) Representative voltage-clamp EPSC records and mean EPSC amplitude in presumed VTA DA neurons at a holding potential of −70 mV, reflecting primarily AMPA receptor-mediated current (negative deflection indicates inward current). B) Representative EPSCs in VTA DA neurons with +40 mV holding potential to remove Mg2+ block of NMDARs (positive deflection is outward current); mean AMPA + NMDA EPSCs were lower in FR than AL (n = 4–5 cells from 2–3 rats per group; *p<.05).

DISCUSSION

Recent findings indicating that changes in brain regional BDNF mediate adaptive responses to negative energy balance and also regulate behavioral responsiveness to drugs of abuse provided the rationale for investigating BDNF and its TrkB receptor in reward-related brain regions of rats that were food-restricted in a manner that potentiates drug reward. A spatial component determining distinct consequences of altered BDNF levels on behavioral responses to abused drugs is apparent, with opponent effects being exerted by mPFC and the mesoaccumbens DA pathway. Chronic cocaine decreases levels of BDNF protein in PFC (Fumagalli et al., 2007), which likely facilitates drug-directed behavior, given that decreased BDNF protein expression in mPFC increases cocaine self-administration (Sadri-Vakili et al., 2008) and microinjection of BDNF decreases cocaine-seeking (Berglind et al., 2007). On the other hand, though basal levels of BDNF mRNA in NAc are extremely low, acute and chronic cocaine increase BDNF mRNA and protein in NAc (Narita et al., 2003; Filip et al., 2006; Graham et al., 2007), with a significant portion of the latter attributed to transport from VTA and activity-dependent release (Altar and DiStefano, 1998; Krishnan et al., 2007). The involvement of NAc BDNF/TrkB function in drug abuse and addiction is further supported by findings that microinjection of BDNF into NAc increases cocaine self-administration (Graham et al., 2007) while microinjection of BDNF or TrkB antiserum into NAc decreases self-administration, cocaine-seeking during withdrawal, and methamphetamine-induced hyperlocomotion (Graham et al., 2007; Narita et al., 2003).

Thus, although alterations of BDNF/TrkB in mPFC and NAc during FR would have provided a tractable new line of approach to illuminate the effect of FR on drug-directed behavior, the present results indicate that FR did not alter levels of BDNF or TrkB protein in mPFC, NAc, or CPu, despite the efficacy of the FR regimen to enhance drug-induced behavioral responses linked to numerous pre- and postsynaptic neuroadaptations in striatum (Carr et al., 2003; Haberny et al., 2004; Zhen et al., 2006; Carr et al., 2010). The lack of effect of FR on BDNF/TrkB in these forebrain regions contrasts with the prior finding that FR increased BDNF protein levels in neocortex and striatum (Lee et al. 2000; Duan et al., 2001). The discrepant observations can most likely be attributed to differences in the FR regimens employed. In the prior studies, subjects were maintained on an every other day (EOD) protocol, which interdigitated days of free feeding with days of total food deprivation for a period of 3 months. In the present study, subjects were fed a limited amount of food every day and assays were conducted after 3–4 weeks of FR. The ability of this FR protocol to increase the reward magnitude and reinforcing efficacy of abused drugs without upregulation of BDNF in key forebrain regions, casts doubt on involvement of forebrain BDNF in the augmentation of drug reward and drug-seeking by FR. Of course, the possibility of altered responsiveness of BDNF/TrkB to acute or repeated drug challenge in FR subjects cannot be ruled out.

Additionally, FR caused a decrease in TrkB protein levels in VTA, which would in fact imply decreased BDNF signaling. The high expression of TrkB in basal midbrain DA cells makes it likely that this reflects decreased TrkB in DA neurons, with the caveat that the methods used cannot distinguish cell type. The source of BDNF that normally interacts with VTA TrkB is believed to be primarily the VTA DA neurons themselves, which synthesize and locally release BDNF (Pu et al., 2006). In cultured cortical and hippocampal neurons, BDNF increases the synthesis and synaptic delivery of NMDA and AMPA receptors (Caldeira et al., 2007; Nakata and Nakamura, 2007). Consequently, we assessed BDNF mRNA and excitatory synaptic transmission in VTA DA neurons to provide evidence consistent with TrkB downregulation during FR. The RT PCR experiment revealed no alteration in BDNF synthetic activity; however, consistent with TrkB downregulation, voltage-clamp data yielded evidence of decreased glutamate receptor transmission in the DA neurons of FR subjects. These data might predict lower average firing rates in FR than AL, which would contribute to the low basal extracellular concentrations of DA observed in NAc of FR subjects (Pothos et al., 1995), the compensatory decrease in V_max of the striatal DA transporter (Zhen et al., 2006), and upregulation of cellular and behavioral responses downstream of NAc DA receptor stimulation (Carr, 2007).

Although a neurophysiological change consistent with TrkB downregulation was observed, the downregulating stimulus is unknown. One speculative scheme is that FR leads to an initial increase in midbrain BDNF that leads to a downregulation of TrkB, which is still seen after 3 weeks of FR as a residual compensatory response. The plausibility of this explanation gains some support from the finding that in SN, TrkB was downregulated and BDNF protein levels were (still) elevated. A short time course of increased BDNF activity in VTA may be adaptive because increased excitatory synaptic transmission would facilitate behavioral activation, which during early FR could promote foraging and food procurement but after a certain duration of FR, particularly in an unchanging environment, may increase counterproductive energy expenditure. Moreover, analogous to the case of chronic drug treatment, the neuroadaptations that underlie enduring changes in goal-directed behavior and reward may be based primarily in NAc, though dependent upon antecedent events in VTA (Kauer and Malenka, 2007). Among other possible mechanisms of TrkB downregulation it is important to consider that VTA DA neurons express receptors for endocrine hormones that are affected by FR, including leptin and insulin (Figlewicz et al., 2003) and insulin-like growth factor (IGF-1; Bondy et al., 1992; Kar et al., 1993), as well as feeding-related peptides including ghrelin (Abizaid et al., 2006) and orexin (Malherbe et al., 2009). All of these signals are known to modify DA neuronal function and may be involved in regulatory cross-talk.

Of particular interest is IGF-1. FR decreases expression and secretion of hepatic IGF-1 and decreases serum and brain concentrations (Olchovsky et al., 1993; LaPaglia et al., 1998; Sohlstrom et al., 1998; Bondy et al., 2004). The FR protocol used in the present set of experiments decreased serum IGF-1 concentrations by more than 50% (Liu and Carr, unpublished). In some cell types, IGF-1 activates the PI3K-Akt pathway, transiently activates the Ras-Raf-MEK-ERK 1/2 pathway, but then exerts a prolonged inhibition of ERK 1/2 signaling via Akt-mediated phosphorylation of Raf-1 on Ser259 (Moelling et al., 2002). Consequently, a sustained decrease in physiological IGF-1 stimulation may decrease phospho-Akt and secondarily increase phospho-ERK. Our finding of a trend toward decreased phospho-Akt and a significant increase in phospho-ERK in VTA of FR subjects may be consistent with a role of low IGF-1 tone in VTA. Importantly, IGF-1 has a strong positive interaction with BDNF/TrkB at the cellular level (Ding et al., 2006; Johnson-Farley et al., 2006; Chen et al., 2007), and can regulate TrkB expression (McCusker et al., 2006). Thus, IGF-1 deficiency in VTA could be involved in the downregulation of TrkB and the upregulation of ERK phosphorylation.

The findings obtained in FR subjects strongly suggest altered functional properties of VTA DA neurons. However, the relationship(s) between any two or all three of the observed changes remains speculative, in part because VTA is a heterogeneous tissue with up to 35% GABAergic neurons and DA neuronal subpopulations that differ in function (Matsumoto and Hikosaka, 2009) and projection patterns (Ikemoto, 2006; Lammel et al., 2008). Yet, the plausible involvement of TrkB downregulation in the enhanced behavioral and striatal responses of FR subjects to drugs of abuse receives additional support from the recent demonstration that 30-min after ingestion of a palatable high fat meal, mice displayed elevation of TrkB mRNA in VTA (Cordeira et al., 2010). In addition, central BDNF depletion decreased evoked DA release in NAc shell and CPu. These results suggest reciprocal effects of FR and palatable meal ingestion on VTA DA neuronal TrkB, with corresponding modulatory effects on excitatory synaptic transmission and striatal DA release.

Highlights.

• Food restriction did not alter BDNF or TrkB protein levels in dopaminergic forebrain regions

• Food restriction decreased TrkB levels in ventral midbrain

• Food restriction decreased glutamate receptor transmission in dopamine neurons

• Food restriction increased levels of pERK 1/2 in ventral tegmental area

• Observed changes in midbrain may enhance drug reward in food-restricted rats

ABBREVIATIONS

DA

dopamine

BDNF

brain-derived neurotrophic factor

TrkB

tropomyosin receptor kinase B

ERK

extracellular signal-related kinase

NAc

nucleus accumbens

mPFC

medial prefrontal cortex

FR

food-restriction, food-restricted

AL

ad libitum fed

VTA

ventral tegmental area

EPSC

excitatory postsynaptic current