Aging Impacts Motivation: 30% Drop in Reward-Seeking Behavior

Chronic social isolation and aging differentially impair motivational behaviors

Voluntary wheel-running activity (VWR) is innately rewarding to rodents even with no other extrinsic reward [9, 10]. Therefore, to test whether age impacts this endogenously motivated behavior, we compared VWR activities between young and old mice under grouped or socially isolated housing conditions. Old mice ran much less than young mice during the dark time, as reported previously [11, 12]. However, social isolation substantially reduced dark-time wheel-running in both young and old mice (Fig. 1a, b). Most VWR reduction occurred in the early dark time when mice reached their daily peaks of running (Extended Data Fig. 1a, b), and the effect of social isolation was greater in old mice, compared to young mice (40.3 vs. 25.7% reduction). Although VWR is a complex behavior that can be affected by both physiological (e.g., reduced skeletal muscle strength and energy metabolism) and motivational changes, these results suggest that both aging and social isolation undermine motivation.

To directly examine whether aging and social isolation reduce motivation levels in mice, we next conducted an operant progressive ratio (PR) test to measure the amount of effort that animals are willing to expend to gain a palatable food reward. We habituated mice to sugar pellets in home cages two days before training and trained under fixed ratio-1 and fixed ratio-3 for five days each (Extended Data Fig. 1c). Following this task learning, we next tested the mice under a progressive ratio (PR) for one day (Fig. 1c). In the PR test, socially isolated young mice performed significantly fewer correct nose-pokes than group-housed controls (Fig. 1d). Interestingly, group-housed old mice showed significantly lower correct nose-pokes than group-housed young mice, but very similar to those of the socially isolated young mice (Fig. 1d). Socially isolated old mice tend to show further reduction in correct nose-pokes (Fig. 1d). Breakpoint analysis, which determines the maximum number of nose-pokes that an animal achieves to gain a single reward [13], further demonstrates consistent reduction in effort to achieve rewards in old and socially isolated mice (Fig. 1e). In both socially isolated groups, there were tendencies of fewer entries into the food tray (Extended Data Fig. 1d). These results clearly suggest that reduced motivation is a shared outcome of social isolation and aging, and old mice seemingly responded less to the stress of social isolation perhaps due to the already reduced effort induced by aging.

Both social isolation and aging induce depressive- and anxiety-like behaviors

Apathy and anhedonia are often comorbid with decreased motivation [14]. To determine whether normal reward processing is impacted by social isolation and aging, we next conducted the sucrose preference test (SPT). All groups of mice manifested robust preference to sucrose solution over water in the sucrose preference test (Extended Data Fig. 1e), indicating no obvious anhedonia. Nonetheless, the amount of sucrose solution consumed was selectively reduced by social isolation and aging (Fig. 1f), although the extent of reduction by social isolation did not reach statistical significance in either young or old mice. Reduction in energy demand was an unlikely explanation because daily food intake was not different for any group on the test day (Extended Data Fig. 1f). Notably, the results from the SPT showed very similar patterns to those from the PR test, supporting the notion that social isolation and aging both reduce motivated behavior in mice.

Finally, to determine whether approach-avoidance conflicts that underly some features of an anxiety-like state might be altered by social isolation and aging in young and old mice, we tested mice in the locomotor activity/open field test (LA/OFT) and elevated plus maze (EPM). Socially isolated young mice spent less time exploring the center area than the group-housed controls, whereas both group-housed and socially isolated old mice exhibited similarly lowered center time and distance traveled (Fig. 1g and Extended Data Fig. 2a, b). Contrarily, socially isolated young mice showed significant increases in time and distance traveled to explore the peripheral area (Extended Data Fig. 2a, b). Additionally, in the EPM, social isolation increased distance traveled and percentages of moving times in the close arm in both young and old mice, as we observed previously [15], whereas there were no differences between young and old mice (Extended Data Fig. 2c–e). We also conducted the hole-board olfactory preference test and the social interaction test to assess other types of motivation-related social behavior between young and old mice in both group-housed and socially isolated conditions (Extended Data Fig. 3). Social isolation had no effect on either assay, except for a trend for isolated young mice to seek more social cues. Therefore, when motivation was reduced by social isolation and aging, certain anxiety- or depression-related behavioral states were also present. Together, these results suggest that both aging and social isolation reduce motivation in mice, but perhaps not additively.

Aging impairs dark time-dependent enhancement of reward-seeking motivation

To determine the extent to which age alters motivation, we tested a new cohort of young and old mice across different testing conditions. Remarkably, we found that when food-deprived and tested during the dark time (Fig. 2a), differences in correct nose-pokes detected in the PR test between young and old mice were greatly enhanced (Fig. 2b). Similar results were obtained for break points (Fig. 2c). These results suggest that the reward-seeking motivation of young mice is enhanced by fasting and heightened during the dark time, whereas old mice fail to show such dark time-dependent enhancement of reward-seeking motivation, confirming that aging is a critical contributor to reduced motivation.

BDNF mRNA and protein levels are decreased specifically in the ventral tegmental area (VTA) during aging

Extensive literature has defined the role of the mesolimbic dopamine (DA) system in regulating reward and motivation [16, 17]. Two major DA populations are identified in the ventral tegmental area (VTA) and the substantia nigra (SN) that locate adjacently in the ventral midbrain [18]. To determine the mechanism for this age-associated decline in motivation, we examined mRNA expression levels of characteristic functional genes in these two regions by collecting tissue samples from young and old brain sections with laser-capture microdissection and conducting quantitative RT-PCR. We found that expression of brain-derived neurotrophic factor (Bdnf) was significantly decreased in the aged VTA (Fig. 3a). Interestingly, genes related to DA turnover did not show any significant changes. The expression of Gria1, which encodes the GluA1 AMPA glutamate receptor subunit, tended to decrease in aged VTA (Fig. 3a). On the other hand, no genes examined in the SN were significantly altered during aging, except for a trend of increase in the expression of Maoa, which can be indicative for augmented DA metabolism and oxidative stress in this region (Fig. 3b) [19, 20]. We also confirmed this age-associated decrease in Bdnf expression by fluorescent in situ hybridization. Lower Bdnf signals were detected across eight consecutive bregma levels throughout the aged VTA, compared to the young VTA (Fig. 3c–d), whereas the expression of Slc6a3, a DA neuronal marker, did not show any significant differences between young and old mice (Extended Data Fig. 4a, b). Age-associated decreased Bdnf expression was present throughout the VTA in both Slc6a3-positive and Slc6a3-negative cells (Extended Data Fig. 4c–f), suggesting widespread downregulation of Bdnf throughout the VTA. Whereas aging selectively decreased Bdnf expression in the VTA, social isolation did not decrease Bdnf expression in the VTA (Extended Data Fig. 4g, left panel). On the other hand, social isolation tended to decrease Bdnf expression in the dorsal raphe nucleus (DRN) at least in young mice (Extended Data Fig. 4g, right panel). Given that the activity of DRN DA neurons is important to elicit a loneliness-like state and promote social preference [21, 22], a commonality between aging and social isolation appears to be decreased Bdnf expression, but in different areas of the brain. Following from the decrease in Bdnf mRNA, we further observed that matured BDNF (mBDNF) protein levels were moderately reduced in the aged VTA (Fig. 3f, g), suggesting a potential involvement of Bdnf in age-associated reduced motivation.

VTA-specific Bdnf knockdown in young mice recapitulates the reduced motivation observed in aged mice

Following the observation that VTA Bdnf was decreased in aged mice, we attempted to restore VTA Bdnf expression in old mice to determine if their motivation could be rescued. To do so, we used a Bdnf-expressing lentivirus. This approach yielded significant BDNF protein overexpression in primary mesencephalon neuronal culture, but it failed to increase BDNF protein levels in vivo so that PR performances were not improved in old mice (Extended Data Fig. 5). We suspect that this could be due to certain auto-regulatory mechanisms that suppress Bdnf transcription and/or translation in the VTA neurons in response to overexpression [23,24,25,26]. It is also possible that the lentiviral transgene was silenced in the host cells over time [27,28,29]. We next sought to determine whether decreasing VTA Bdnf expression in young mice can recapitulate the motivational dysfunction observed in aged mice. To do so, we knocked down Bdnf expression in the VTA by stereotactically injecting a Bdnf shRNA-expressing lentivirus into the VTA of young mice (Fig. 4a). Bdnf mRNA levels were reduced by ~30% (Fig. 4b, c), and the mBDNF protein levels were also reduced to a similar extent to aged mice in the VTA, but not in the downstream nucleus accumbens (NAc) (Fig. 4d–f). Two months after surgery, the VTA-specific Bdnf-knockdown (VTA-Bdnf-KD) mice were tested for the extended PR paradigm with food deprivation during light and dark times. Importantly, VTA-Bdnf-KD mice exhibited significant reduction in correct nose-pokes and break points with food deprivation during both light and dark times, successfully recapitulating motivational phenotypes observed in aged mice (Fig. 4g, h, and Extended Data Fig. 6a, b). On the other hand, VTA-Bdnf-KD mice did not alter behaviors in the VWR and OFT (Fig. 4i–l, and Extended Data Fig. 6c, d), supporting a selective role of Bdnf in the VTA in reward-seeking motivational behavior. Prior work has shown that VTA BDNF can play a role in regulating VTA dopamine neuron excitability [30, 31]. To determine whether the downregulation of VTA Bdnf in aged mice altered excitability in these VTA neurons, we examined the intrinsic neuronal electrical properties of both putative dopaminergic and non-dopaminergic neurons in young and aged mice. For these purposes, we operationally defined these neurons by the expression of hyperpolarization-activated cation current (Ih) [32] (Extended Data Fig. 7a, b). No significant differences were found between young and aged mice, suggesting that baseline neuronal excitability in the VTA does not drive the decreased motivation through aging (Extended Data Fig. 7c–f). Taken together, these results demonstrate that specific decreases in Bdnf expression in the aged VTA contribute, at least in part, to age-associated reduced motivation in mice.

All animal procedures, including the ones on live animals, were approved by the Washington University Institutional Animal Care and Use Committee (WU IACUC) and were in accordance with National Institutes of Health guidelines. The animal protocol numbers approved by WU IACUC were 22-0007 and 24-0384 for the Imai lab and 20-0139 and 23-0261 for the McCall lab. Unless otherwise noted, all mice were housed in groups of 2–5, fed ad libitum on a standard chow diet (PicoLab Rodent Diet 20-5053; LabDiet, St. Louis, MO) and autoclaved water in a temperature/humidity-controlled holding room within a barrier facility, maintained on a 12:12-h light/dark cycle. Food, beddings (Aspen Chip, Northeastern Products Corp.), and nesting material (Nestlets, Ancare) were changed once per week; water bottles were changed as needed. Mice were labeled on their tails with sharpie markers for easy identification and monitored constantly for their health status. All handling, cage changing, measuring procedures, and behavior tests were conducted by the same experimenters throughout the study. Only male mice were used.

All aged mice and the corresponding young groups were obtained from the NIH aged rodent colony (C57BL/6JN) or our in-house C57BL/6J colony. Due to the limited availability of aged mice, this study only used male mice to limit the number of potential biological variables to remain sufficiently powered. All group-housed cages were housed only with original cage mates from weaning. Mice used for Bdnf knock-down analysis were purchased from Jackson Laboratories (C57BL/6J) at 12-week-old and acclimated to the animal facility for ~2 months before surgeries; upon arrival, mice were randomly assigned into cages of 4/5 and maintained until sacrifice.

Social isolation procedure

All mice were originally group-housed in cages of 3–5 cage mates. After acclimation to the holding rooms, cages containing around half of the animals were separated into individual cages and maintained isolated until sacrifice; the rest of the control cages remained unchanged across the study. All cages within a certain cohort were placed in proximity on the same rack to control for traffic, noise levels, and light exposure.

All behavioral tests were performed within a sound-attenuated room maintained at 20–26 °C after at least one week of acclimation to the local holding room. All tests were performed by female experimenters. Unless otherwise noted, all mice in the same cohort underwent the same test/procedure on the same day(s) from ZT3 to ZT9; if a test/procedure cannot be finished within one day, the cohort will be divided into sub cohorts by counterbalancing grouping, home cages, and other factors. For all experiments performed outside of the holding room, mice were brought into the corresponding behavior space to habituate for at least 20 min before any test/procedure started. All behavioral apparatuses were cleaned with 2% chlorhexidine in between subjects.

Operant progressive-ratio test

To expose the animals to the reward, mice fed ad libitum were given 3 g/mouse of the reward sugar pellet (Dustless Precision Pellet, F0071, Bio-Serv) on the bedding of their home cages when cages were changed at ZT12. Consumption of the pellets was confirmed the next day. One day after the exposure, mice were gently placed in a mouse operant-conditioning box (15.24 × 13.34 × 12.7 cm) within a sound attenuating cubicle (ENV-307A & ENV-020M, Med Associates Inc., Fairfax, VT). The box was equipped with a stainless steel grid floor, two nose holes (active and inactive) flanking a trough pellet receptacle (food tray) connected to a pellet dispenser on the bottom of the right-hand side wall; a house light was installed at the top center of the left-hand side wall. Pokes into the ports are monitored by infrared light beam sensors. Upon starting of a fixed-ratio 1 (FR1) session, both nose holes present a cue light; one nosepoke into the active hole results in a sugar pellet (20 mg) dispensed to the food tray, along with cue lights turned off and house light turned on for 2 s, during which no action triggers event (time-out period); poking the inactive hole has no consequence. Each session lasted 60 min. In between trials, boxes were cleaned with 2% chlorhexidine, and leftover pellets in the dispenser were replaced with fresh and intact ones. All trials took place from ZT3 to ZT8, unless otherwise specified, and each animal was subjected to the sessions at the same time across days. In FR1, mice were trained to discriminate the active hole from the inactive one and receive rewards from the food tray. Following 5 consecutive days of FR1 sessions, mice were moved on to another 5 consecutive days of 60-min fixed-ratio 3 (FR3) sessions, in which three active nosepokes results in one reward delivery and 20 s time-out period. Mice were conditioned to repeatedly perform the correct operant behavior until receipt of the reward during FR3. After completion of the training, motivation levels of the mice were assessed using a progressive-ratio (PR) schedule of reinforcement following the equation: response ratio = 5ereward number/5 – 5 (rounded to the nearest integer), resulting in an exponential increase of the number of correct pokes required for each subsequent reward: 1, 2, 4, 7, 9, 12, 15, 20, 25, 32, 40, 50, 62, 77, 95, 118, 145, 178, 219, 268, 328, 402… [49]. All events were programed, and actions were recorded using MED-PC software (Med Associates) controlled by a PC. The number of active nosepokes counted toward rewards during the 60-min session was calculated as the readout of motivation; A break point (BP) was calculated by taking the bigger number of the last completed ratio versus the number of correct nosepokes after the last reward [13]. For the extended PR paradigm, following FR1 and FR3 training, two PR sessions were conducted on consecutive days, and the averaged readouts served as the PR baseline. Mice then underwent a break day with no session, followed by a third PR test (light fed) under the same conditions (ad libitum in the light time); after another break day and overnight fasting from ZT12, a fourth PR test was conducted (light fasted); lastly, two days after, the mice were fasted from ZT0, and a fifth PR test under fasting was conducted during ZT15-20 under dim red lighting (~10 Lux).

Sucrose preference test