Vortioxetine Reverses Impairment of Visuospatial Memory and Cognitive Flexibility Induced by Degarelix as a Model of Androgen Deprivation Therapy in Rats

Abstract

Introduction

Androgen deprivation therapy (ADT) is a mainstay treatment for prostate cancer, but many patients experience cognitive impairment in domains mediated by the medial prefrontal cortex (mPFC) and hippocampus. Prostate cancer typically occurs in older patients (>65 years). As age is often accompanied by cognitive decline, it may impact the efficacy of any treatment aimed at restoring cognitive impairment induced by ADT. Vortioxetine, a multimodal antidepressant that improves cognition in depression, has been shown to be efficacious in elderly patients. Therefore, vortioxetine may improve cognition in older patients who experience cognitive decline after ADT.

Methods

Young (3 months) and middle-aged (13 months) rats were used to investigate the influence of age on treating ADT-induced cognitive decline. As our previous studies used surgical castration, we tested if vortioxetine would reverse cognitive deficits associated with more translationally relevant chemical castration using degarelix. Vortioxetine was given in the diet for 21 days. Animals underwent behavioral testing to assess visuospatial memory mediated by the hippocampus and cognitive flexibility mediated by the mPFC. We also investigated changes in afferent-evoked responses in these regions in middle-aged rats.

Results

Degarelix induced impairments in both visuospatial memory and cognitive flexibility that were reversed by vortioxetine. Vortioxetine also rescued afferent-evoked responses in the mPFC and hippocampus. However, modest age-related reductions in baseline visuospatial memory limited our ability to detect further decreases induced by degarelix in middle-aged rats due to a floor effect.

Conclusion

These results suggest that vortioxetine may be a treatment option for older prostate cancer patients who experience cognitive decline after ADT.

Introduction

Cognitive impairment associated with androgen deprivation therapy (ADT) for the treatment of prostate cancer has been shown to be long-lasting and to significantly disrupt the quality of life for patients and their caregivers [1, 2]. Deficits have been observed in cognitive domains such as learning, memory, and executive function [3, 4], which are all important for optimal functioning in daily life tasks. Risk factors for prostate cancer include race and heredity [5–7], but disease progression can also be influenced by exogenous factors, such as diet and physical activity, among others [8]. Another significant risk factor for prostate cancer and ADT-induced cognitive decline is age. Prostate cancer primarily affects men above the age of 65 [9], and there is a clear relationship between age and diminished cognition [10]. In males, age is also accompanied by decreased testosterone levels, which can further exacerbate changes in cognition [11]. Thus, ADT could further increase vulnerability in brain regions that mediate executive function and visuospatial learning in aging patients. In these studies, we investigated the potential influence of age on ADT-induced cognitive decline on the attentional set-shifting test (AST) and Novel Object Location (NOL) test as readouts of function of the medial prefrontal cortex (mPFC) and hippocampus, respectively. These two regions have been identified in clinical studies to be disrupted after ADT [12, 13].

We have previously shown that surgical castration induced cognitive set-shifting deficits in young adult rats [14, 15]. Further, we have shown previously that the multimodal antidepressant vortioxetine reversed these impairments. Vortioxetine is FDA-approved for the treatment of depression. In addition to blocking the serotonin transporter, vortioxetine acts at multiple pre- and post-synaptic serotonin receptors [16], which are thought to contribute to its unique effects on cognition. In direct comparison, vortioxetine effectively improved cognitive impairments associated with depression, whereas other antidepressants, including other selective-serotonin reuptake inhibitors such as duloxetine, did not [17–19].

The aims of the current study were first to enhance the translational relevance of our model by using the gonadotropin-releasing hormone antagonist, degarelix, to test ADT-induced cognitive impairment after chemical castration, which is used more commonly in the clinic than is surgical castration. Second, to investigate the potential influence of age, we used middle-aged rats to replicate previous results obtained in young rats. We then investigated circuit-level changes in middle-aged rats after degarelix and vortioxetine treatment as potential mediators of cognitive impairment induced by ADT and the potential therapeutic efficacy of vortioxetine in middle-aged rats. The results from these experiments will provide insight into the brain regions compromised by ADT and will investigate vortioxetine as a potential therapeutic option to alleviate cognitive decline associated with ADT, in particular, for older prostate cancer patients.

Materials and Methods

Animals

A total of 207 rats were used in these experiments (69 young adult, 138 middle-aged). Young adult male Sprague-Dawley rats were obtained from Envigo (USA) at 225–249 g upon arrival (∼2 months old). They were initially group-housed (3/cage) for approximately 1 week before entering into experimental procedures. Middle-aged male Sprague-Dawley rats were obtained from Envigo at approximately 3 months of age and group-housed in large polycarbonate cages (Techniplast, 610 × 435 × 215 mm) until 12–13 months of age. All rats were separated into single housing after receiving degarelix or 5% mannitol treatment (below). Housing lights were maintained on a 12/12 h light cycle (lights on at 07:00 h), and rats had access to food and water ad libitum until the food restriction period. Young rats were tested at 3–4 months of age, and middle-aged animals were tested between ages 12–14 months as this is an age at which mild cognitive impairments begin to emerge in rats [20]. There were no differences in body weight between any conditions within age group at the time of testing (data not shown). Behavioral testing was performed during the light phase, between the hours of 9:00 and 17:00. Separate cohorts were used for behavioral and electrophysiological experiments. All procedures were approved by the University of Texas Health Science Center at San Antonio Institutional Animal Care and Use Committee, approval #20170174AR, and complied with the National Institute of Health guidelines. The experimental timeline began on day 0 with degarelix or vehicle administration, after which the animals were single-housed. On day 11, the animals began receiving drug-infused or control diet. The animals were handled for at least 5 min/day for 2 days prior to behavioral testing (see timeline in Fig. 1).

Fig. 1.

Experimental timeline. Degarelix injection or surgical castration was administered on day 0.

Drug Treatment

Vortioxetine (provided by H. Lundbeck A/S) was incorporated into standard Purina #5001 chow by Research Diets, Inc. The base chow served as the control diet. Drug-infused diet contained 0.6 g/kg of vortioxetine. Rats receiving either drug or control diet had free access to food and water for 14 days before animals were food-restricted for 7 consecutive days in preparation for the AST. During food restriction, the animals consumed 12 g of food per day, which corresponded to about 12 mg/kg/day of drug for a 600 g rat and 24 mg/kg/day for a 300 g rat. After chemical castration, the animals were given 8–11 days for sufficient testosterone decline before initiating vortioxetine or control diet. In total, the animals received drug for 21 days. In previous studies, this drug treatment protocol has been shown to be effective in reversing stress-induced cognitive impairment in rats, with 60–95% occupancy of relevant drug receptor targets in the brain [21].

Degarelix Administration and Plasma Testosterone Measurements

Degarelix was prepared in 5% mannitol dissolved in sterile water. Initially, a dose of degarelix was determined that would sufficiently reduce testosterone for the full duration of the experimental timeline. The animals received either vehicle or degarelix on day 0. To administer degarelix, the animals were briefly anesthetized with isoflurane (5–10 min). The final dose of degarelix was determined to be 3 mg/kg, administered subcutaneously as a single depot injection (0.1 mL/kg) in the shoulder region at a concentration of 30 mg/mL. At this time, blood samples were taken via the tail vein. After all experimental testing was complete, a second sample was taken to ensure degarelix sufficiently maintained reduced testosterone. Plasma samples were prepared for ELISA (IBL Inc., Minneapolis, MN, Kit #IB79174, interassay coefficient of variability = 3.75%, intraassay coefficient of variability = 2.54%) by treating with 25 µL of 0.1 m EDTA per 1 mL of blood and were then centrifuged (4°C for 10 min at 2,000 rpm). Degarelix-treated animals had testosterone levels significantly reduced in comparison to vehicle controls. Detection limit of the assay was 0.066 ng/mL. Levels in degarelix-treated rats were on average below detection, whereas levels in vehicle-treated rats were 1.94 ± 0.83 ng/mL and 1.37 ± 0.46 ng/mL for young and middle-aged animals, respectively. Animals were removed from the study if testosterone levels did not decline in the final sample compared to baseline. This resulted in the exclusion of two young rats from behavioral analysis and one middle-aged rat from electrophysiological analysis.

Open Field Test/NOL Habituation

The open field test (OFT), which also served as habituation for the NOL test, was conducted in a white square arena (65 × 65 × 42 cm), marked by lines on the floor into a grid of 36 squares of 10 × 10 cm. Blue horizontal stripes on one wall served as a spatial orientation cue in the NOL test. The animals started the OFT in the center, facing this wall. Habituation was 20 min in duration, and the first 5 min were used for analysis of the OFT. Line crosses were scored as a measure of locomotion. The time in the center of the arena (i.e., in squares not adjacent to a wall) was used to measure anxiety-like behavior.

NOL Test

Adapted from Barker and Warburton [22], the NOL test was performed 24 h after habituation, as previously described [15]. Testing occurred in the same arena as the OFT. Identical test objects consisted of Lego^® figures (9.5 × 5 × 9.5 cm) made of blue, green, and yellow blocks. All testing was performed in low light at approximately 10 lux. The sample and test phases were recorded by an overhead GoPro^® camera. Videos were analyzed by an experimenter blinded to the conditions. The sample phase started with two objects placed in the two adjacent corners opposite the striped wall, 10 cm from the walls. Animals were given 3 min to interact with both objects, after which they were returned to their home cage for 5 min. After this delay, one object was moved directly across the arena to a corner adjacent to the striped wall, and the animal was returned to the test arena. The test phase was also 3 min, during which the interaction time with each object was measured, defined as the rat touching, sniffing, or facing the object within 2 cm. Between sessions and sample/test, objects were cleaned with 70% ethanol and then with water and completely dried. The discrimination ratio (DR=(Tn-Tf)/(Tn+Tf)) was used to assess relative interaction time with the objects in the familiar (Tf) and novel locations (Tn). The animals were excluded if they spent less than 5 s total with both objects or failed to interact with one of the objects in the sample phase. The animals were also excluded if they failed to interact with at least one object during the test phase. These criteria resulted in 7 middle-aged and 3 young rats being excluded from the analysis.

Attentional Set-Shifting Test

The AST was performed as described previously [15, 23]. One half of a Honey Nut Cheerio (General Mills, Minneapolis, MN) was used as reward. Two terra cotta pots were differentiated by the textured medium that filled each pot and an odor applied to the rim. The procedure took 3 days beginning on day 30, which consisted of habituation, training, and testing.

• 1.Habituation: on day 30, the rats learned to find the Cheerio reward in unscented pots filled with increasing amounts of sawdust, first in their home cage and then in the testing arena.
• 2.Training: on day 31, training began with baited pots placed on both sides of the arena. They then learned to discriminate a correct cue based on the odor (lemon or rosewood) using sawdust-filled pots. Once the animals mastered the odor contingency (6 correct choices in a row), they learned to discriminate based on the digging medium (paper vs. felt) using unscented pots.
• 3.Testing: on day 32, the animals learned to locate the food reward based on a simple discrimination of cues in one of the stimulus dimensions. The animals were counterbalanced at testing, half beginning with odor and half with the medium as the salient cue. In the first 5 tasks, the rats formed a “cognitive set,” in which they learn that one stimulus dimension, regardless of how the cues have changed, will signal the location of the reward. In the extradimensional (ED) set-shift stage, the rats must abandon this cognitive set to learn that the previously irrelevant dimension now signals the location of the Cheerio. This set-shift is mediated by the mPFC. The animals were excluded if they failed to make a choice within 20 min on 3 consecutive trials in a single task.

In vivo-Evoked Local Field Potentials

A separate cohort of middle-aged rats was used for electrophysiological recording, in groups defined by degarelix and vortioxetine treatment. Recording occurred on day 32, equivalent to the day of testing on the AST, as indicated in the timeline. After the experiment, electrode placement was confirmed histologically, and in cases where electrodes were located outside the target region, the animals were excluded (n = 7). In vivo, evoked local field potentials were recorded in the ventral hippocampus (vHipp)-mPFC projection and in the Schaeffer collateral (SC) projection to the CA1 region of the dorsal hippocampus (dHipp) as previously described [24]. To conserve animal numbers, recordings were made in both regions in the same animals when possible. To counterbalance, half of the animals started with mPFC recordings and half started with dHipp recordings. The rats were anesthetized (chloral hydrate 400 mg/kg, i.p., supplemented 10% as needed through the duration of recording), placed in a stereotaxic apparatus, and a heating pad maintained body temperature at 37°C. Coordinates were adjusted, if needed, based on measurement of optimal evoked response. This resulted in +0.2 mm being added to the anterior/posterior coordinates for the larger, middle-aged rats. All electrode placements were verified histologically following the procedure. For mPFC recordings, an insulated stainless steel bipolar twisted stimulating electrode was positioned in the vHipp (AP: −6.0–6.2, ML: +5.4, DV: −7.5 mm), and a tungsten recording electrode was placed in the ipsilateral mPFC (AP +3.0–3.2, ML +0.6, DV –3.5 mm). For recordings in the dHipp, the recording electrode was placed in the CA1 region (AP –3.8–4.0, ML +2.0, DV –3.0 mm) and the stimulating electrode was placed in the SCs at a 30° lateral angle (coordinates from bregma: AP –3.8–4.0, ML +5.0, DV: 4.0–4.7 mm). Signal was filtered (low cutoff 0.3 Hz, high cutoff 1,000 Hz, sampling 2,000 Hz) and digitized using PowerLab (ADInstruments). Recordings began after a 15-min equilibration period. The amplitude of the first peak of the evoked response (latency of the mPFC response was ∼20 ms, and in CA1 it was ∼5 ms) was measured at stimulation steps of 100 µA from 100–600 µA for mPFC and 100–800 µA for dHipp (0.1 Hz, 260 µs pulse width) to create a stimulus-response curve, which was compared across groups.

Statistical Analyses

Behavioral results (OFT, NOL, and ED set-shifting) in each age group were analyzed with two-way ANOVA (ADT × vortioxetine) using GraphPad Prism 9 (San Diego, CA, USA). All data were presented as means ± SEM. Holm-Sidak pairwise comparisons were used to detect differences between experimental conditions. In the AST, a three-way ANOVA (task × ADT × vortioxetine) was used to test for any changes in performance on the tasks preceding ED that could have accounted for nonspecific differences in set-shifting. Differences in baseline behavior on the NOL test in young versus middle-aged controls were assessed using an unpaired t-test. Significance was determined at p < 0.05. Within the NOL test, animals were excluded from analysis if they failed to interact with both objects during the sample phase or at least one object during testing. The animals were excluded from AST if they failed to make a choice within 20 min on 3 consecutive trials. In the electrophysiological experiments, stimulus-response curves were analyzed and compared by the least sum-of-squares F test.

Results

Experiment 1: Determination of Degarelix Dose in Young and Middle-Aged Rats

This experiment used 26 young and 26 middle-aged rats. A group of surgically castrated rats was used for comparison to find an effective dose of degarelix that reduced testosterone to castration levels. The control animals received either a sham surgery or 5% mannitol. Young rats were treated with 2 mg/kg and 3 mg/kg doses, and blood samples were taken on days 0 (prior to injection of degarelix), 7, 14, 21, and 32. Day 32 corresponded to the day of testing in subsequent behavioral experiments (Fig. 1). While 2 mg/kg significantly reduced testosterone (Fig. 2a), it did not reduce testosterone to castration levels, whereas the 3 mg/kg dose was comparable to surgical castration (Fig. 2b). In middle-aged rats, 3 mg/kg degarelix also depleted testosterone to castration levels (Fig. 2c). Therefore, the 3 mg/kg dose was used for all subsequent experiments as this dose comparably reduced testosterone to levels obtained by surgical castration in both duration and efficacy.

Fig. 2.

Degarelix (3 mg/kg) reduced testosterone to castration levels for the duration of the experimental timeline. a Degarelix reduced testosterone in a dose-dependent manner. 3 mg/kg was sufficient to completely reduce testosterone to <10% of baseline levels for the duration of the time required for the experiments. b 3 mg/kg degarelix significantly reduced testosterone to castration levels on day 32 in comparison to control animals. Planned comparison of intact animals to castrated animals was made using a t test that revealed a significant decrease in testosterone in sham versus surgical castration and in vehicle versus degarelix-treated animals (****p < 0.0001). There was not a significant difference between surgically and chemically castrated animals (p = 0.80, n = 6–7/group). c Degarelix (3 mg/kg) also reduced testosterone in middle-aged rats for the duration of time required for the experiments, and there were no differences between surgical and chemical castration (p = 0.99). After both surgical and chemical castration, testosterone levels were significantly decreased compared to their respective controls (***p < 0.001 for both groups).

Experiment 2: Measures of Anxiety-Like Behavior and Locomotion in Young and Middle-Aged Rats after Degarelix and Vortioxetine

For this experiment, 43 young rats were used. In the open field test, there were no significant differences between groups in line crosses in any comparison (Fig. 3a, F[1, 39] = 0.59, main effect of ADT p = 0.18, main effect of vortioxetine p = 0.91, interaction effect p = 0.45). Similarly, there were no differences in time in the center of the arena (Fig. 3b, F[1, 39] = 0.028, main effect of ADT p = 0.30, main effect of vortioxetine p = 0.68, interaction effect p = 0.87). In 78 middle-aged rats, there were also no differences in line crosses (Fig. 3c, F[1, 64] = 0.29, main effect of ADT p = 0.47, main effect of vortioxetine p = 0.69, interaction effect p = 0.59), nor time in the center of the arena (Fig. 3d, F[1, 64] = 0.30, main effect of ADT p = 0.21, main effect of vortioxetine p = 0.54, interaction effect p = 0.59). These results indicate that degarelix and vortioxetine treatment do not increase anxiety or alter locomotion in young or middle-aged animals, eliminating these as potential confounds to interpretation of subsequent behavioral analyses. The same rats tested in the OFT were used for behavioral testing on the novel object and set-shifting tests in experiments 3 and 4.

Fig. 3.

There were no differences in locomotion or anxiety-like behavior after ADT in young and middle-aged rats (young n = 8–14/group; aged n = 11–22/group). a In young rats, there were no differences between groups in line crosses on the open field test, a measure of locomotion. b There were no differences in time in the center of the arena, a readout of anxiety-like behavior. c In middle-aged rats, locomotion was not affected by degarelix or vortioxetine treatment. d There were no effects on anxiety-like behavior in middle-aged rats.

Experiment 3: Effects of Degarelix and Vortioxetine on Visuospatial Cognition in Young and Middle-Aged Rats

In young rats, two-way ANOVA showed a significant main effect of degarelix (Fig. 4a, F[1, 36] = 4.35, p < 0.05) and an interaction of degarelix × vortioxetine (F[1, 36] = 14.51, p < 0.001). There was no main effect of vortioxetine (F[1, 36] = 0.80, p = 0.38). Holm-Sidak post hoc comparisons revealed that animals treated with degarelix and fed control diet had a significant deficit in visuospatial memory on the NOL test in comparison to vehicle controls (p < 0.001). Vortioxetine reversed this deficit in degarelix-treated animals (p < 0.05). There was a moderate age-related decline in baseline performance in middle-aged control rats compared to young controls (Fig. 4b, t[28] = 1.736, p = 0.09). Although this decrease was not significant, it was sufficient to compress the range between baseline and 0 in the middle-aged rats (i.e., a floor effect), limiting our ability to detect any further decrease due to degarelix. Although the pattern and direction of change appeared similar to those in young rats, there were no significant differences between groups in the middle-aged rats (Fig. 4c, F[1, 58] = 0.56, main effect of ADT p = 0.60, main effect of vortioxetine p = 0.16, interaction effect p = 0.46). These results support the hypothesis that chemical castration with degarelix impairs hippocampal-based cognition and that vortioxetine ameliorates these deficits after chronic treatment.

Fig. 4.

Vortioxetine reversed hippocampal-mediated visuospatial cognitive impairments after degarelix treatment (young n = 7–13; aged n = 10–19/group). a In young rats in the NOL test, degarelix treatment decreased the DR, indicating an impairment in visuospatial memory, compared to intact-control rats (***p < 0.001). Degarelix-treated rats that received vortioxetine had a DR comparable to intact rats (*p < 0.05 compared to degarelix-control diet). bt-test revealed a trend for moderate decrease in baseline performance of middle-aged control rats compared to young rats (n = 13–17/group, p = 0.09). c Although the pattern and direction of change appeared similar to those in young rats, due to the moderate age-related reduction in baseline, there were no significant effects of degarelix or vortioxetine in the middle-aged rats (p = 0.46).

Experiment 4: Effects of Degarelix and Vortioxetine on Cognitive Flexibility in Young and Middle-Aged Rats

In young rats, two-way ANOVA revealed a significant main effect of ADT (F[1, 40] = 11.69, p < 0.01), a main effect of vortioxetine (F(1, 40) = 5.56, p < 0.05) and an interaction effect (ADT × vortioxetine, F(1, 40) = 4.77; p < 0.05). Post hoc comparisons identified a significant increase in trials to criterion in degarelix-treated rats fed with control diet (Fig. 5a, p < 0.0001), which was reversed by chronic vortioxetine treatment (p < 0.05). There were no significant effects in any task preceding the ED set-shifting task that might have accounted for effects on set-shifting (F[4, 156] = 0.66, p = 0.62, data not shown). Similarly, in middle-aged rats, there were significant main effects of ADT (F[1, 51] = 18.13, p < 0.0001) and vortioxetine (F[1, 51] = 5.54, p < 0.05), and an interaction of ADT × vortioxetine (F[1, 51] = 4.10, p < 0.05). Post hoc comparisons identified a significant deficit in degarelix-treated rats fed with control diet (Fig. 5b, p < 0.0001), and this was reversed with vortioxetine (p < 0.01). There were no effects in tasks preceding ED in the middle-aged rats (F[4, 204] = 1.34, p = 0.26, data not shown). The results from the AST support the hypothesis that degarelix significantly alters cognitive flexibility in rodents, an executive function also affected in ADT patients, and that vortioxetine is efficacious in reversing these impairments in both middle-aged and young rats.

Fig. 5.

Vortioxetine reversed impairments in cognitive flexibility mediated by the medial prefrontal cortex after degarelix treatment. a In young rats on the set-shifting task, degarelix increased trials to criterion compared to intact-control rats (***p < 0.001), indicating an impairment in cognitive flexibility. Vortioxetine reversed this deficit in degarelix-treated rats (*p < 0.05 compared to degarelix-control diet, n = 9–14/group). b In middle-aged rats on the set-shifting task, degarelix increased trials to criterion compared to intact-control rats (***p < 0.0001). This deficit in cognitive flexibility was reversed by vortioxetine (**p < 0.01 compared to degarelix-control diet, n = 10–17/group).

Experiment 5: Effects of Degarelix and Vortioxetine on Afferent-Evoked Responsivity of the mPFC and Hippocampus in Middle-Aged Rats

Evoked potentials were recorded in 34 middle-aged rats. There was a significant difference in the responses evoked in the CA1 region by stimulating the SCs (Fig. 6a, F[9, 284] = 10.44, p < 0.0001). Planned comparisons revealed that afferent-evoked responses were significantly reduced in degarelix-treated rats fed with control diet compared to vehicle control rats (F[7, 105] = 8.60, p < 0.0001). Vortioxetine reversed this decrease (F[7, 133] = 5.02, p < 0.0001). There was also a significant difference in responses evoked in the mPFC by stimulating in the vHipp (Fig. 6c, F[9, 180] = 4.92, p < 0.0001). Responses in degarelix-treated rats fed with control diet were significantly reduced compared to vehicle control rats (F[5, 65] = 4.96, p < 0.001). Again, vortioxetine restored the attenuated response in degarelix-treated animals to levels comparable to vehicle controls (F[5, 75] = 3.02, p < 0.05). Electrode placements and examples of afferent-evoked responses are displayed in Figure 6b for CA1 and Figure 6d for mPFC. These results support the hypothesis that ADT disrupts the circuitry that underlies visuospatial memory and cognitive flexibility and that chronic vortioxetine treatment is able to ameliorate these changes.

Fig. 6.

Degarelix treatment attenuated afferent-evoked responsivity in the dHipp and mPFC of middle-aged rats. a There was a significant decrease in responsivity in the SC to CA1 pathway in degarelix-treated rats compared to vehicle controls (****p < 0.0001, n = 8 vehicle controls vs. n = 9 degarelix-controls). This was reversed in degarelix-vortioxetine rats (p < 0.0001, n = 12 degarelix-vortioxetine). b Left: placement of stimulating electrodes in SCs and recording electrodes in CA1 were verified histologically. Right: a sample trace recorded after stimulating at 800 µA. Arrow indicates stimulation artifact. c Degarelix treatment attenuated responsivity in the vHipp-mPFC pathway (***p < 0.001, n = 8 vehicle controls vs. n = 7 degarelix-controls). This was reversed by vortioxetine treatment (*p < 0.05, n = 10 degarelix-vortioxetine). d Left: recording electrode placement in mPFC (top) and stimulating electrodes in vHipp (bottom). Right: a sample trace recorded after stimulating at 600 µA. Arrow indicates stimulation artifact.

Discussion

Age is a significant risk factor for prostate cancer as most patients are over the age of 65 [5–7, 9]. So, any effective treatment for ADT-induced cognitive impairment must be tested in older subjects as age is accompanied by reductions in cognition. In addition, age-related decreases in testosterone can exacerbate cognitive changes [10, 11]. In the context of cancer, results of a human imaging study showed increased signs of aging within the brain after ADT [25]. Increased age is a significant risk factor for dementia, and clinical studies have shown that men who undergo ADT are predisposed to subsequent diagnoses of Alzheimer’s disease and dementia [26, 27]. Therefore, it is possible that ADT exacerbates processes that lead to age-related neurocognitive decline. Given this, it is necessary to demonstrate the efficacy of vortioxetine in reversing ADT-induced cognitive impairment in middle-aged rats, as well as young rats if it is to be considered as a viable potential treatment for cognitive impairment associated with ADT in prostate cancer patients. In this regard, the results of clinical trials in elderly depressed patients are promising as vortioxetine was more effective in improving cognition in elderly depressed patients than duloxetine [17]. Another study found that vortioxetine improved cognition on a battery of cognitive tests, including attention and visuospatial memory tasks, in comparison to other antidepressants (escitalopram, paroxetine, bupropion, venlafaxine, and sertraline) in elderly patients with Alzheimer’s disease [28].

To investigate the potential influence of age in the present study, we used middle-aged rats, approximately 12–13 months old. At this age, mild cognitive changes just begin to emerge in rats [20]. We can, therefore, investigate the potential influence of age on cognition after ADT treatment without the more severe cognitive impairment associated with advanced age that may confound the results of this study. Consistent with this, there was no effect of age on baseline performance on the AST, and the cognitive impairment induced by degarelix treatment on the AST was comparable in the two age groups as was the efficacy of vortioxetine in restoring set-shifting impaired by ADT. There was, however, a modest age-related decrease in visuospatial recognition on the NOL test in control middle-aged rats compared to young controls, perhaps reflecting the early stages of mild cognitive impairment expected at this age. Although the age-related decline in visuospatial cognition was not significant, the reduction in the baseline discrimination ratio in the middle-aged rats effectively compressed the range between baseline and the floor of 0 (which indicates a complete absence of visuospatial recognition memory). This constricted the possible effect size, making it more difficult to detect any further reduction in DR attributable to degarelix treatment. Although the pattern and direction of change seen in DR after degarelix treatment in middle-aged rats appeared similar to those in young animals, they were not significant, and the limited range would have required considerably more animals per group to achieve statistical power. Due to the time and cost of aging the rats in-house, this was not feasible. The fact that there were no such baseline differences in middle-aged rats on the set-shifting task further suggests that the hippocampus may be especially sensitive to age-related cognitive decline. While we did not observe an exacerbation of ADT-induced impairments in the 12–13 month-old middle-aged rats, our results suggest that at least in some cognitive domains, e.g., hippocampal-mediated visuospatial cognition, age-related impairments may be additive with impairments induced by ADT. Further, these results do not rule out the possibility that age-related effects may accelerate or increase vulnerability to ADT-induced impairment as age increases.

In the in vivo electrophysiology experiments, we found that degarelix treatment reduced responsivity in the afferent projection from the vHipp to the mPFC, and in the SC-to-CA1 pathway in the dHipp. Vortioxetine treatment reversed the ADT-induced reduction in afferent-evoked response in both pathways in middle-aged rats. Cognition is dependent on functional plasticity, including in the circuitry between the mPFC and hippocampus. In our previous study, we found that vortioxetine reversed the attenuated responsivity induced in mPFC- and hippocampus-related circuitry after surgical castration in young rats [15]. The fact that vortioxetine was also effective in reversing ADT-induced changes in evoked local field potentials of middle-aged rats is important to informing the potential utility of vortioxetine to improve cognition impaired by ADT in prostate cancer patients. These results further indicate that the mechanisms associated with ADT-induced cognitive decline in middle-aged rats are similar to those in young rats, and are still responsive to vortioxetine as a potential therapeutic intervention. The fact that the behavioral measure on the novel object task reflecting the hippocampal function in middle-aged rats was less sensitive to the effects of both degarelix and vortioxetine than the electrophysiological afferent-evoked response measured in the hippocampus is most likely attributable to the floor effect caused by the modest age-related decline in the discrimination ratio discussed above.

In our previous studies, we reported results using young adult male Sprague-Dawley rats [14, 15]. Surgical castration was used to isolate the cognitive effects of testosterone depletion from potentially confounding pharmacological effects of drugs used for chemical castration. In the current study, we sought to expand the translational relevance of these findings by investigating if vortioxetine could also ameliorate changes associated with the gonadotropin-releasing hormone antagonist degarelix as a more clinically relevant model of ADT. Due to the relatively short behavioral timeline of our experiments, we wanted to use a method of chemical castration that would reliably and quickly deplete testosterone levels, comparably to surgical castration, and maintain those levels for the duration of the study. We also wanted to avoid using a drug (e.g., a GnRH agonist) that would produce an initial “androgen flare.” Thus, we chose to use the GnRH antagonist, degarelix, as this approach avoids the androgen flare and reduces testosterone rapidly [29]. Using degarelix in young rats, we found that 3 mg/kg reduced testosterone to similar plasma concentrations as surgical castration, while 2 mg/kg was not sufficient to completely deplete androgen. Similarly, in middle-aged rats, 3 mg/kg degarelix sufficiently depleted and maintained testosterone at castration levels. Therefore, this dose of degarelix was used in both age groups for subsequent behavioral and electrophysiological experiments.

One caveat to these studies is that the dose of vortioxetine administered in the diet was unintentionally different in young and middle-aged rats. In young adult rats, we estimated the daily food intake to be approximately 21 g for a 300 g rat. This corresponds to a free-feeding dose of 42 mg/kg/day. During the 1-week food restriction (12 g food/day), young rats received a dose of 24 mg/kg/day vortioxetine. This dose regimen has been shown to be fully effective at restoring reversal learning compromised by serotonin depletion or chronic cold stress [21] and to produce fractional occupancy of relevant targets in the brain (e.g., 5-HT1B receptor, serotonin transporter, 5-HT3 receptor) of approximately 50% to >90% for low and high affinity targets, respectively [21, 30]. This was also determined to be equivalent to clinically effective doses based on similar occupancy levels in humans [31, 32]. Acute administration of lower doses of 3–10 mg/kg vortioxetine was also effective behaviorally in rats, and produced fractional occupancy at high affinity targets, similar to chronic dietary treatment at 24 mg/kg/day [30, 33]. Thus, we are confident the dietary administration of vortioxetine produced a clinically relevant dose in young adult rats in this study, and was effective in restoring behavior in both the AST and NOL tests, and afferent-evoked responses in dHipp after androgen deprivation. However, we found that middle-aged rats consumed similar amounts of food per day compared to young rats (approximately 21–22 g/day), despite their larger average body weight (about 575 g). This resulted in a free-feeding dose of 23 mg/kg/day, comparable to the dose in young rats during the period of food restriction. Further, it was necessary to food restrict the middle-aged rats to the same daily amount as the young rats (12 g) to sustain behavior for food reward on the AST. This resulted in a dose during food restriction of approximately 12.5 mg/kg/day. Even though this dose acutely produces brain occupancy within a clinically relevant range, it is possible this was not a fully effective dose in middle-aged rats, perhaps accounting in part for the lack of significant effect on the NOL test in this age group. However, this dose was fully effective in restoring both set-shifting behavior and afferent-evoked responses in dHipp in middle-aged rats. Moreover, vortioxetine did increase DR on the NOL test after ADT, albeit not significantly. A more likely explanation is that the lack of significant effect of vortioxetine on the NOL in middle-aged rats is attributable to the same floor effect discussed above due to an age-related decrease in baseline that made it difficult to detect a significant change after ADT, and as a result, a significant restoration of behavior after vortioxetine.

Having established the utility of our model of ADT and the effectiveness of vortioxetine in reversing ADT-induced cognitive impairment in healthy rats, the next step in future experiments will be to investigate the potential influence of cancer pathophysiology and its contribution to cognitive decline, particularly in middle-aged rats. In cancer, inflammatory signaling is a prominent feature of the tumor microenvironment, which ultimately promotes cancer growth [34]. Inflammation is also a hallmark of aging [35]. Other metabolic processes such as oxidative stress have been shown to be exacerbated in men with low testosterone [36]. Few studies have incorporated cancer into rodent models to investigate mechanisms underlying cognitive decline associated with cancer treatment, or to target those mechanisms to improve cognition without compromising the anticancer efficacy of the primary treatment. An important future direction will be to include cancer in our middle-aged model of ADT to investigate the potential contribution of inflammation and other components of cancer pathophysiology that may also be detrimental to cognition [37]. Clinical studies of cancer in older patients have shown that cognitive impairments can negatively impact the treatment outcome [38]. Therefore, new treatment approaches that can mitigate these effects may not only improve the quality of life after cancer treatment but may even improve the treatment outcome itself.

In conclusion, the results of these studies in rats identify circuit-related mechanisms that may underlie the cognitive impairment associated with ADT that compromises the quality of life for many prostate cancer survivors. We found that ADT, modeled using degarelix in middle-aged and young adult rats, induced cognitive impairments in visuospatial memory and cognitive flexibility. These deficits were ameliorated by chronic vortioxetine treatment in both age groups. Similarly, vortioxetine reversed the attenuation of afferent-evoked responses in the circuitry that mediate the cognitive domains studied in our behavioral analyses in middle-aged rats. These results further highlight the potential utility of vortioxetine in mitigating the cognitive effects of ADT, specifically in an aged population, which may be more vulnerable to changes in cognitive performance after cancer treatment. Survivorship is unfortunately an understudied area of cancer research, and negative changes in cognition are a major concern for all cancer survivors [39]. Developing novel therapies or even repurposing currently approved therapies such as vortioxetine to alleviate the cognitive decline associated with cancer therapy without compromising the anticancer efficacy of the primary treatment could significantly improve the quality of life for cancer survivors, and may even improve survival itself.

Acknowledgments

We thank Teresa L. Johnson-Pais, Ph.D., Jonathan Gelfond, M.D., Ph.D., Robin Leach, Ph.D., Chethan Ramamurthy, M.D., and Ian Thompson, M.D., for their expertise and insight.

Statement of Ethics

The Institutional Animal Care and Use Committee (IACUC) of the University of Texas Health, San Antonio, reviewed and approved all animal procedures, approval #20170174AR. Experiments were compliant with ethical standards of the National Institutes of Health as specified in the Guide for the Care and Use of Laboratory Animals.

Conflict of Interest Statement

Vortioxetine was provided by H. Lundbeck A/S, which had no input into the design or conduct of the study, analysis, or interpretation of the data and no role in the decision to publish or in the writing of the manuscript.

Funding Sources

This work was supported by research grant R01 CA224672 from the National Cancer Institute, National Institutes of Health; by research grant RP180055 from the Cancer Prevention and Research Institute of Texas (CPRIT); by CPRIT training grant RP170345; and by pilot funding from the Mays Cancer Center, UT Health San Antonio. In-kind support was provided by H. Lundbeck A/S, which generously provided the drug-containing chow and control chow.

Author Contributions

A.V. conceptualized and designed the study, contributed to acquisition, analysis, and interpretation of the data, and drafted and revised the manuscript; A.A. and K.T. contributed to acquisition, analysis, and interpretation of the data and reviewed the manuscript; and D.M. conceptualized and designed the study, coordinated and supervised acquisition, analysis, and interpretation of the data, and drafted and revised the manuscript.

Funding Statement

This work was supported by research grant R01 CA224672 from the National Cancer Institute, National Institutes of Health; by research grant RP180055 from the Cancer Prevention and Research Institute of Texas (CPRIT); by CPRIT training grant RP170345; and by pilot funding from the Mays Cancer Center, UT Health San Antonio. In-kind support was provided by H. Lundbeck A/S, which generously provided the drug-containing chow and control chow.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.