[¹⁸F]FDG PET metabolic patterns of the rapid-acting antidepressant effects of NLX-101, a 5-HT_1A receptor biased agonist

Abstract

Rapid-acting antidepressants (RAADs) such as ketamine are currently under development. The aim of this study is to characterize the neural circuits affected by ketamine and NLX-101, a selective 5-HT_1A receptor biased agonist which has shown promising effects, by using [¹⁸F]FDG PET imaging in rats that had received chronic administration of corticosterone (CORT), a model of anxiety-depression. In a longitudinal study, regional changes in brain activity were investigated in 24 selected CORT rats. Each animal underwent PET scans in 3 conditions, i.e. with ketamine (10 mg/kg), NLX-101 (0.16 mg/kg) or saline on day 0 and five days later to assess sustained effects. The anxious-depressive phenotype produced by CORT was supported by behavioural and biological observations. Changes in [¹⁸F]FDG uptake were determined using voxel-based and region of interest analyses. Metabolic connectivity analysis was also performed to investigate the acute and delayed effects of the treatments. Voxel-based and region-of-interest analyses showed marked hypometabolism in regions implicated in depression, particularly cingulate cortex (−7%) and lateral septum (−9%) as well as the striatum (−10%). Acute effects of NLX-101 and ketamine were observed in the lateral septum, resulting in an increase in brain glucose metabolism (p < 0,05). Interestingly, connectivity analyses also showed effects of NLX-101 in the frontal cortex, the thalamus and amygdala (p < 0.05), suggesting that the two molecules converge on common brain regions. This study is the first to show brain activation patterns of RAADs in a CORT rat model by functional PET imaging. NLX-101 appears to exert antidepressant effects by preferentially activating postsynaptic 5-HT_1A heteroreceptors in primary regions common to ketamine. These results support investigation of cortical 5-HT_1A receptors as a target for new generation biased agonist antidepressants.

Highlights

• [¹⁸F]FDG-PET imaging was used in rats which had received chronic corticosterone administration.

• Corticosterone treated rats exhibited decreased glucose consumption in multiple brain regions.

• NLX-101, a 5-HT_1A receptor biased agonist, and ketamine modified glucose consumption in cortex and lateral septum.

• These brain regions may contribute to the RAAD-like properties of the compounds.

Introduction

Depression constitutes a substantial public health concern, exerting a considerable toll on healthcare systems, economies, and society at large. It is estimated that one in five individuals will experience depressive disorders during their lifetime [1], and approximately one third of these individuals are resistant to conventional antidepressants, e.g. monoamine reuptake inhibitors [2]. Therapeutic options for patients with treatment-resistant severe depression (TRD) are limited, with few alternatives to existing treatments such as ketamine, which can have significant side effects [3]. Therefore, the development of new, rapidly acting antidepressant (RAAD) medications is a crucial priority [4].

A substantial body of evidence from numerous studies has demonstrated the role of 5-HT_1A receptors in the pathogenesis of depression. In particular, molecules that directly target these receptors have demonstrated antidepressant-like effects in animal models, suggesting greater efficacy compared with selective serotonin reuptake inhibitors, SSRIs [5, 6]. As a leading RAAD, ketamine has also demonstrated efficacy in modulating the serotonergic system by stimulating the release of 5-HT in the prefrontal cortex and activating post-synaptic 5-HT_1A receptors [7, 8]. Such effects could explain the prolonged action of ketamine, which is attenuated if 5-HT_1A receptors are blocked [9, 10] or if serotonin is reduced [11].

These observations suggest that the development of selective 5-HT_1A receptor biased agonists targeting of specific brain regions, including cortical areas, would achieve a RAAD therapeutic effect similar to that of ketamine, while avoiding the occurrence of the latter’s side effects [12]. A number of such biased agonists of 5-HT_1A receptors are currently under investigation, including the molecule NLX-101 [13], and preclinical studies have demonstrated that these novel compounds have rapid and sustained antidepressant effects in anxiety-depressive animal models, indicating their potential for treating depression [14–17].

The objective of the present project is to elucidate the role of 5-HT_1A receptors in the pathophysiology of depression and the mechanism of action of RAADs like ketamine and NLX-101 using a translational functional imaging approach. Neuroimaging techniques are indispensable for mapping brain activity in vivo, both in rodents and humans. A widely utilized method for large-scale assessment of cerebral activity is the measurement of brain glucose metabolism with [¹⁸F]-fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET). The uptake of [¹⁸F]FDG is highly correlated with neuronal activity and has been extensively employed to investigate the effects of psychoactive drugs and their underlying mechanisms of action [18, 19].

Here, we investigated the central effects of subanesthetic doses of ketamine and therapeutic-like doses of NLX-101 in a rat anxiety-depressive model induced by chronic corticosterone administration (CORT rat) [20]. The effects of the compounds were assessed by [¹⁸F]FDG PET imaging. The results provide further support for the concept of targeting 5-HT_1A receptors with selective biased agonists, with the aim of developing new treatments for TRD.

Materials and methods

Animals

A total of forty-eight adult male Sprague-Dawley rats (Charles River Laboratories, France) with a weight range of 200–225 g were utilized in the present study. Unless otherwise stated, the rats were housed in groups of five per cage for the CORT group and two for the control group. The housing conditions were standardized, with a temperature of 22 ± 2 °C and humidity of 50%, and a 12/12 h light/dark cycle, with light on at 8.00 am. The rats were provided with food (Teklad 18% protein rodent diet, Envigo, USA) and tap water ad libitum. All experiments were conducted in accordance with the European guidelines for the care and use of laboratory animals (Directive 2010/63/EU) and were approved by the University of Lyon review board (CELYNE).

Corticosterone induced rat depression model (CORT model)

Following an initial acclimatization period of five days, the rats were randomly assigned to two groups: a control group (n = 8) and a CORT group (n = 40). The CORT group received daily subcutaneous injections of corticosterone (40 mg/kg; s.c.) between 9:00 and 11:00 am for 21 consecutive days, based on previous studies [21, 22] to induce an anxiety-depression phenotype and conduct the behavioral tests. Corticosterone (11β,21-dihydroxy-4-pregnene-3,20-dione) solution (Sigma-Aldrich) was prepared after initial solubilization in a 20% Tween 80 parent solution, followed by dilution with 0.9% NaCl, leading in a concentration of 2% at volume of 1 mL/kg, which equates to a dose of 40 mg/kg.

Behavioral tests and animal selection

To select only twenty-four CORT rats out of the forty for PET imaging (n = 8 per group), behavioral tests were performed to assess the depth of the depressive phenotype. A scoring system was devised for each test, with a separate score assigned to each rat. Briefly, a score from 1 to 12 was assigned for each test: the lowest score was considered non-anxio-depressive, and the highest score highly anxio-depressive. For example, the maximum score of 12 corresponded to the longest total immobility in the FST, the lowest sucrose preference, the least weight gain or the longest immobility time in the O-maze test.The four rats with the lowest scores on each test were excluded from the imaging study. The control group (n = 8) was also subjected to the aforementioned tests.

Forced Swim Test (FST)

The objective of this test was to evaluate the animal’s resignation by determining their immobility time (a measure of depression-like behavior) in a container filled with water and according to a precise protocol (Porsolt et al. [23–26]. On test day (Day 21), the rats were placed in the cylinder for a period of five minutes (between 1 and 4 pm). The water was replaced between each trial. The test sessions were recorded and analyzed using video analysis software (Boris®) by two independent operators, who also measured the total immobility time and time of first immobility.

0-maze test

The test was conducted on the 22nd day of the experiment and was designed to assess anxiety and locomotion [27]. The apparatus comprised a cylindrical platform with two closed arms and two open arms, situated at a height of 50 cm above the floor. The animals were placed in the same location on the platform in an isolated room for a period of five minutes. Each animal was filmed and the software (EthoVisionXT, Noldus®) analyzed its behavior and movements.

Sucrose preference test

A sucrose preference test (SPT) was conducted for the purpose of assessing anhedonia, with the measurement of low sucrose preference serving as the primary indicator [28, 29]. The animals were placed in individual cages for a period of five days (from Day 29), after four days of adaptation. The first two days of the study involved the consumption of two bottles of plain water. On the third day, the subjects were provided with two bottles of water containing 1% sucrose. On the fourth day, the subjects consumed one bottle of plain water and one bottle of water containing 1% sucrose. On the fifth day, which constituted the actual test phase, the subjects consumed one bottle of plain water and one bottle of water containing 1% sucrose for a period of 12 h.

Body weight

The animals were weighed twice a week since anxio-depressive phenotype presents decreased body weight [30].

PET imaging procedures and data analyses

PET/CT protocol and image acquisitions

As previously stated, following the selection of anxio-depressive CORT-injected animals, a total of 32 rats were scanned to evaluate the effect of the studied molecules (8 control rats, 8 CORT saline rats, 8 CORT ketamine rats, 8 CORT NLX-101 rats).

A longitudinal approach was employed, with each CORT animal undergoing three different [¹⁸F]FDG PET/CT acquisitions. The first was conducted as the baseline (day −2, D-2), the second 48 h later (day 0, D0), and the third with saline administration 5 days after the administration of the molecule of interest (ketamine, NLX-101, NaCl).

Prior to each scan, the rats were fasted for a period of four hours. Fifteen minutes following the intraperitoneal injection of 0.5 mL of NLX-101, ketamine, or saline, catheterization of the caudal vein was conducted, and [¹⁸F]FDG (37.2 ± 2.7 kBq/g) was administered under conscious conditions. Subsequently, the rats were anaesthetized with constant isoflurane inhalation (4% for induction, then 2% during acquisition, 1 L/min) and positioned within the PET-CT imaging scanner (INVEON®, Siemens). The PET scans were acquired in list mode for a period of 30 min, followed by a computed tomography (CT) scan to correct for tissue attenuation. The static images resulting from the 30-min acquisitions were reconstructed with the application of attenuation and scatter corrections. The reconstructed volume comprised 159 slices of 128 × 128 voxels, encompassing a bounding box of 49.7 × 49.7 × 126 mm³ with a voxel size of 0.388 × 0.388 × 0.796 mm³, generated using OSEM 3D.

PET data analyses

The data acquisition and processing were conducted using the INVEON Research Workplace (IRW®, Siemens) software and the statistical parametric mapping software SPM12®. The images were subjected to three distinct analytical processes, namely a voxel-based analysis, a regions of interest (ROIs) analysis and a metabolic connectivity analysis.

Voxel-based analyses (VBA)

The individual PET images were realigned and spatially normalized based on a previously constructed CT template co-registered on an anatomical MRI template. Subsequently, each PET volume was spatially normalized and smoothed using an isotropic Gaussian filter with a resolution of 1 × 1 × 1 mm. Local voxel activities were normalized to the mean uptake value of the entire brain. A voxel-to-voxel statistical analysis was conducted for each group, employing a variance analysis within a generalized linear model, with the objective of comparing mean uptakes between the different conditions. The resulting maps were those of baseline (D-2) or D5 for the activation of each molecule, and of baseline (D-2) or D5 – D0 for the inhibition of each molecule, in the case of the CORT rats. A significant threshold was set at p < 0.005 uncorrected at the voxel level, and a thresholding of k = 700 was applied, corresponding to PET scanner resolution. A second analysis was conducted to compare the metabolic profiles of CORT rats (saline at D0) and control rats under saline conditions using an unpaired Student’s t-test.

Regions of interest (ROIs) analyses

Brain ROIs were already defined in the Lancelot rat brain atlas [31] on the corresponding MRI template. Mean uptake ratios were extracted in the different ROIs, and a two-way mixed ANOVA for repeated measures (factors: treatment and ROI) with Tukey’s post-hoc tests (corrected for multiple comparisons) were performed using GraphPad Prism 8.0 software to compare the [¹⁸F]FDG uptake ratios between the different conditions. Additionally, a standard uptake value (SUV) analysis was conducted, i.e. normalizing the radioactivity to the animal’s weight and the injected dose.

To minimize the variability and account for potential confounding factors such as changes in peripheral glucose metabolism, the standardized uptake values in each scan were normalized to the mean radioactivity in the whole brain in order to ensure a consistent comparison of regional effects. However, it is important to note that this approach may underestimate global changes in brain metabolism, which is why we also measured the mean standard uptake values (SUVs).

Metabolic connectivity analyses

For each day (D-2, D0, D5), the Pearson correlation coefficient was employed to evaluate the pairwise regional correlations across subjects. To facilitate statistical comparisons of the correlation matrices across the different days, a permutation testing method was employed, as previously described by [32] using Matlab (R2023a, Mathworks). For each comparison (D-2 vs D0, D-2 vs D5, D0 vs D5), the PET images of the corresponding groups were randomly permuted to create pseudo-random groups, which were reassigned 10,000 times and for which a correlation matrix was calculated. For each pair of regions, the p-value was determined by comparing the actual observed Z score difference (obtained with the original groups) with the distribution obtained from the permuted data. The p-values were then corrected for multiple comparisons using the false discovery rate method with a 0.1 threshold, to yield adjusted q-values.

Treatments

All animals received an intraperitoneal injection of ketamine or NLX-101 at D0 in a volume of 0.5 mL. Ketamine (100 mg/mL, Virbac) was diluted to 1/10 in 0.9% saline solution at a subanaesthetic dose of 10 mg/kg. The compound NLX-101 (also known as F15599; 3-chloro-4-fluorophenyl)-[4-fluoro-4-[[(5-methylpyrimidin-2-yl)methylamino]methyl]piperidin-1-yl]methanone, fumarate salt (Neurolixis) was diluted in 0.9% NaCl solution at a concentration of 0.5 mg/mL and administered at the dose of 0.16 mg/kg, which is known to be pharmacologically active [15].

Tissue analyses

Blood samples

On day 48, blood samples were collected in EDTA tubes from the retro-orbital sinus between the hours of 9 and 11 a.m. After centrifugation, they were frozen at −80 °C in advance of analysis. Corticosterone levels were quantified using an enzyme-linked immunosorbent assay (ELISA) method (Enzo Life Sciences) and according to the manufacturer protocol. After colorimetric reaction, the optical densities of the samples were compared to those of previously prepared standards.

Adrenal gland sampling

Immediately following the collection of blood samples, the animals were euthanized and the adrenal glands were excised and weighed on a precision scale. Due to the considerable discrepancy in body weight between the CORT group and the control group, the total adrenal gland mass was expressed as a percentage of the animal’s body weight.

Statistical analyses

The results obtained from the behavioral tests and assays were subjected to a series of U-Mann-Whitney tests, a non-parametric statistical test that is employed for the analysis of independent values derived from small samples. With regard to the SPT, the data were subjected to analysis using a non-parametric Kruskal-Wallis test, followed by a Dunn’s post-hoc multiple comparison test. A p-value of less than 0.05 was considered to indicate a statistically significant result.

Results

Characterizations of the CORT- model

Behavioral tests and animal selection

The Fig. 1A presents the body weight change curve of rats treated with chronic subcutaneous corticosterone, with the curve of the control rats serving as reference. Prior to the commencement of the stress procedure (baseline), the body weights of all animal groups were similar (mean CORT = 226 g and mean control rats = 229 g, not statistically significant). However, the progression of the curves differed significantly between the start and end of the study (from Day 2 to Day 48), with a lower weight increase for the CORT group versus for the control group (p < 0.0001).

Fig. 1. Characterization of chronic corticosterone-induced depression in rats treated with a minimum of 21 days of subcutaneous corticosterone at 40 mg/kg (compared to saline-treated control rats).

Body weight gain (A), Forced swim test on day 21 (B), Anxiety-related behavioral parameters of all rats were assessed in the 0-maze on day 22 (C), 12-h sucrose preference test (D) and concentrations of corticosterone in plasma sampled at day 48 sacrificed during on morning (between 09 and 10 am) (E). Data are expressed as means ± SEM; n = 40 in the CORT-induced rats group vs n = 8 in the control group; ***p < 0.001; ****p < 0.0001; Mann-Whitney tests.

Behavioral changes were observed in the FST. The induction of an anxiety-depressive phenotype in rats was shown by a reduction of the time of onset of immobility, with a mean of 52 ± 8 s versus 135 ± 17 s for the control group (p = 0.0004). In the CORT group, there is a notable increase in immobility, with a mean of 80 s compared to 18 s in the vehicle-exposed rats (p = 0.0004). (Fig. 1B). The 0-Maze test revealed that the CORT model elicited an increase in immobility within the open arms, as illustrated in Fig. 1C: 26 vs 6 s (p < 0.0001). Conversely, no significant difference is observed in the time spent in the closed arms (p = 0.6986). The CORT model induced no significant difference on sucrose at 12 h compared to controls (Fig. 1D: x̄ = 81,2%; p = 0.9841).

These experiments, conducted on a total of 40 CORT rats, permitted the selection of the 24 animals that exhibited the highest anxiety-depression scores for the following [¹⁸F]FDG PET imaging sessions.

Post-mortem analyses

The plasma concentrations of corticosterone were significantly different between the CORT group and the control group (mean = 2.51 × 105 pg/mL vs mean = 2.91 × 104 pg/mL, p < 0.0001; Fig. 1E). The adrenal gland mass normalized to the animal’s weight was significantly higher in the CORT group (x̄ = 3.6. 10⁻⁴%) compared to the control group (x̄ = 2.0. 10⁻⁴%), with a p-value of less than 0.0001 (Fig. 1E). When expressed in absolute mass, the adrenal gland mass of the CORT s.c. rats was significantly lower than that of the vehicle-exposed rats (14 vs 56 mg; p < 0.0001).

PET/CT imaging

PET metabolic profiles of CORT-model

The voxel-based analyses of [¹⁸F]FDG uptake ratios showed a reduction in metabolic activity in CORT anxiety-depressive animals in comparison to control animal group (n = 8/8) (Fig. 2A). The pattern of CORT rats was modified in frontal cortex, cingulate cortex, hippocampus, lateral septum, and striatum. Following statistical analysis (U-Mann-Whitney test), the uptake of [¹⁸F]FDG expressed in SUV was confirmed to be significantly lower than that of the control rats in the total brain (2.49 ± 0.17 vs 3.26 ± 0.13; p = 0.0030; Fig. 2B). Finally, the ROIs analysis confirmed a significant decrease in [¹⁸F]FDG uptake ratios in the CORT rats at the level of the striatum (−10%), cingulate cortex (−7%), and lateral septum (−9%) (Fig. 2B).

Fig. 2. Comparison of basal metabolic profile using [¹⁸F]FDG PET scans between chronic corticosterone-induced depression rats versus control rats.

A Voxel-to-voxel statistical comparisons of [¹⁸F]FDG uptake ratio between chronic injection of CORT (40 mg/kg, s.c, for 21 days) and saline baseline in rats (n = 8). T scores are represented in color scales (significant decreases in blue; p < 0.005, Student’s t test, Resolution threshold k = 700). Coronal sections are from +4 to −13 mm with respect to Bregma. B [¹⁸F]FDG mean SUVs (on the left) and [¹⁸F]FDG uptake ratio (on the right) comparisons at D0 in several ROIs after saline injection in CORT-induced depression rats (n = 8) and saline baseline in control rats (n = 8). Bars are the mean + s.e.m.; * p < 0.05, ** p <  0.01, *** p < 0.001, **** p  <  0.0001, Holm-Sidak’s multiple comparison test following a two-factor ANOVAs. SUV analysis: region factor: F(2.112, 29.57) = 59.53, p < 0.0001; treatment factor: F(1, 14) = 11.81, p = 0.0040; interaction factor: F(12, 168) = 6.035, p < 0.0001. ROI analysis : region factor: F(2.073,29.02) = 60.11, p < 0.0001; treatment factor: F(1, 14) = 12.26, p = 0.0035; interaction factor: F(11,154) = 5.898, p < 0.0001. C Effects of chronic injection of CORT (n = 8) on metabolic connectivity assessed by [¹⁸F]FDG PET in comparison with saline baseline (n = 8). Pairwise connectivity matrices for CORT-induced rat model and control group, obtained by correlation analysis of individual uptake ratio values between different ROIs. Thal Thalamus, Hip Hippocampus, Amy Amygdala, Cereb Cerebellum, Cing Cingulate cortex, Front Frontal cortex.

Metabolic connectivity analyses

The metabolic activities of regions of interest were statistically correlated with each other and changes in pairwise regional correlations between CORT rats and control rats were evaluated. As illustrated in Fig. 2C, the regional correlation matrices exhibit notable differences between both groups (significant differences after permutation and multiple comparison correction are indicated by stars) with a significant decrease in metabolic connectivity between the thalamus and the striatum (q = 0.0945) in CORT rats compared to controls. Similarly, there was a significant negative correlation between the [¹⁸F]FDG uptake ratios of the hippocampus and the striatum (q = 0.045) and between the lateral septum and thalamus (q = 0.071).

Metabolic effects of NLX-101 versus ketamine in the CORT model

Acute effects

PET data analyses

The voxel-based statistical analysis images of [¹⁸F]FDG uptake ratios following ketamine and NLX-101 administration in CORT rats are presented in Fig. 3A (n = 8 in each group). An acute subanesthetic dose (10 mg/kg) of ketamine induces an increase in glucose metabolism, primarily in the lateral septum, in comparison to a saline injection. Similarly, acute administration of NLX-101 at an antidepressant dose (0.16 mg/kg) also induces increased metabolism in the lateral septum compared to a NaCl injection. Additionally, an increase in glucose uptake was observed in the striatum and cerebellum.

Fig. 3. Effects of ketamine compared to the candidate RAAD, NLX-101, on brain metabolic profile in a CORT-induced depression rat model using [¹⁸F]FDG PET scans.

A Voxel-to-voxel statistical comparisons of [¹⁸F]FDG uptake ratio between acute injection of ketamine at subanaesthetic dose and saline (n = 8) on the left and NLX-101 and NaCl (n = 8), on the right, in CORT-induced depression rats. T scores are represented in colour scales (significant decreases in blue and increase in red, p < 0.005) (B) [¹⁸F]FDG mean SUVs (on the left) and [¹⁸F]FDG uptake ratio (on the right) comparisons at D0 in several ROIs after acute ketamine injection, NLX-101 or saline in CORT-induced depression rats (n = 8). Bars are the mean + s.e.m.; * p < 0.05, *** p < 0.001, Holm-Sidak’s multiple comparison test following a two-factor ANOVAs. SUV analysis: region factor: F(2.199, 46.18) = 88.10, p < 0.0001; treatment factor: F(2, 21) = 0.7737, p = 0.4740, interaction factor: F(14, 147) = 2.070, p < 0.0166.ROI analysis: region factor: F(8168) = 137.0, p < 0.0001; treatment factor: F(2, 21) = 1.491, p = 0.2480; interaction factor: F(16,168) = 1.693, p = 0.0521. C Effects of chronic injection of CORT (n = 8) on metabolic connectivity assessed by [¹⁸F]FDG PET in comparison with saline baseline (n = 8) on the left and effects of ketamine (n = 8) and NLX-101 (n = 8) on metabolic connectivity in CORT-induced depression rats assessed by [¹⁸F]FDG PET in comparison with saline CORT group (n = 8). Pairwise connectivity matrices for the CORT-induced rat model injected with ketamine and NLX-101 and the CORT saline group, obtained by correlation analysis of individual uptake ratio values between different regions of interest (ROIs). Thal Thalamus, Hip Hippocampus, Amy Amygdala, Cereb Cerebellum, Cing Cingulate cortex, Front Frontal cortex.

To further support this analysis, ROIs analyses revealed a significant increase in [¹⁸F]FDG uptake ratios in the lateral septum following ketamine administration (p = 0.0382) compared to the CORT group injected with NaCl. In the frontal cortex, a significant decrease was noted between NLX-101 and NaCl (p = 0.0391). In other brain regions, the effects of ketamine and NLX-101 (compared to NaCl) were similar, with a non-significant increase of glucose metabolism in the cingulate cortex (+2.34% et + 2.3%) and the striatum (+1.4% et + 0.62%) and a decrease in the thalamus (−1.97% et −1.17%) and raphe (−1.66% et −2.24%) (Fig. 3B).

Metabolic connectivity

The Fig. 3C illustrates that the regional correlation matrices are distinct following the administration of the three molecules at D0 (the significant changes after multiple comparison correction are indicated by stars on the corresponding pairs of regions). A significant increase in metabolic connectivity between the thalamus and the amygdala (q = 0.058) was observed following NLX-101 administration when compared to the CORT group treated with NaCl whereas the agonist reduces the correlation between the frontal cortex and the thalamus (q = 0.063). With regard to ketamine, no notable alteration in metabolic connectivity was discerned in comparison to the CORT group that had been administered NaCl.

Delayed effects

The VBA and ROIs analyses revealed no statistically significant differences between Day 0 and Day 5 (data not shown). The administration of ketamine/NLX-101/NaCl did not result in any significant alteration in brain glucose metabolism five days later.

Discussion

The objective of this study was to investigate the neural regions and circuits associated with depression using a corticosterone-induced rodent model of anxiety-depression. Additionally, the study aimed to assess the antidepressant effects of NLX-101 as a potential RAAD medication, employing microPET imaging with [¹⁸F]FDG.

Validity and limits of the CORT model

While no single animal model of anxiety-depression is considered as a gold standard, the chronic corticosterone administration is a promising option since elevated levels of glucocorticoids mimic chronic stress [21, 26, 33]. The CORT model offers the advantage of high reproducibility [22] and we observed that CORT rats exhibited higher levels of anxiety-depressive behaviors when compared to control rats, confirming its strong face validity. Additionally, CORT animals exhibited notable weight loss, elevated anxiety, and augmented despair-like behaviors as previously described [20, 21, 26, 34]. Elevated plasma corticosterone levels, changes in adrenal gland volumes and alterations in body weight are attributable to corticosterone inhibition of the hypothalamo-pituitary axis [30], observations which resonate with the adrenal insufficiency which is also observed in patients with depressive disorders [35].

Some limitations of the present CORT model are that the sucrose preference test for anhedonia was unchanged between both groups and the O-maze results suggest an anxiolytic effect. Nevertheless, similar results have already been reported in the literature [36, 37] and are likely attributed to several potential confounding factors, such as differences in behavioral measurement methodology, strain differences, inter-individual variability or the number of animals used [22]. In our case, the O-maze was used, which may differ from the Open Field Test (OFT) or the Elevated Plus Maze (EPM). Additionally, for the SPT, a 1% sucrose concentration was used, whereas some studies have used 5% [20].

Another limitation is that only male rats were used, a choice made to maintain consistency with previous studies. Furthermore, the CORT model does not address TRD, a trait which presents a significant challenge to the development of new RAADs.

Cerebral glucose metabolism in the CORT model

Changes in glucose uptake ratios, as observed through [¹⁸F]FDG PET imaging, are indicative of the dynamics of cerebral activity [38]. The PET metabolic profile measured in the anxiety-depressive CORT model is consistent with previous data in a depression model of rats [36] and in depressed subjects [39]. Our study corroborates the existence of a significant hypometabolism in various brain regions associated with depression [40], including disruptions in interconnected structures forming a simplified circuit, which has been modelled as the fronto-limbic network [41, 42]. In particular, a reduction in glucose uptake has been observed in the striatum, which is involved in the reward system, as well as in cortical regions such as the prefrontal cortex and the cingulate cortex, both of which are involved in motivation. A comparable reduction is observed in limbic regions, including the amygdala and hippocampus, which are also integral to the regulation of stress. Furthermore, the lateral septum, which plays a pivotal role in emotional regulation, motivation, and social behavior, has been identified as a potential target for future pharmacological interventions in depression [43, 44].

Nevertheless, it can be argued that the overall brain hypometabolism induced by corticosterone may be attributed not only to depressive pathology itself [45] but also to systemic biochemical changes involving blood glucose levels [46]. Furthermore, prolonged corticosteroid treatment induces structural alterations in the brains of rodents, including a reduction in hippocampal volume and an increase in amygdala volume. In the present study, the use of a global brain ratio in analysis served to mitigate this issue. For example, increased [¹⁸F]FDG uptake is observed in the thalamus, which typically exhibits hyperactivity in major depressive disorder [39]. However, hypermetabolism was also observed in the frontal cortex, despite this region typically being associated with hypofunction in depression [47, 48]. This discrepancy may be attributed to the relatively large size of the ‘frontal cortex’ region of interest in the atlas employed.

The neuroimaging pattern of RAAD drug candidates

This study is the first to visualize the in vivo effect of RAAD-type molecules on a CORT-induced anxiety-depressive model using [¹⁸F]FDG microPET imaging. It is noteworthy that the acute effects of NLX-101 and ketamine appear to be mediated by two principal regions. In particular, the administration of both molecules was associated with a notable increase in glucose metabolism in the lateral septum. Papp et al. [12] postulated that the effects of ketamine and NLX-101 are mediated by convergent signalling at cortical GABAergic interneurons. The lateral septum, densely populated by GABAergic neurons and 5-HT_1A receptors [5, 49] may therefore plausibly be involved. Therefore, the inhibition of GABA interneuron activity by ketamine and by NLX-101 would both result in disinhibition of glutamatergic pyramidal neurons and increased glutamate release, due to antagonism of NMDA receptors and agonism at 5-HT_1A receptors, respectively.

Interestingly, both compounds also had effects in the frontal cortex, where NLX-101 has been shown to reduce glucose uptake, whereas ketamine enhanced it. It can therefore be surmised that this structure plays a pivotal role in the mechanism of action of both these RAADs. Indeed, the antidepressant effects of both ketamine and NLX-101 require activation of 5-HT_1A receptors in frontal cortex [7, 10, 50] in accordance with a hypothesis that has been previously proposed in the literature [51, 52].

Furthermore, the investigation of metabolic connectivity contributed to the advancement of this research by elucidating disturbances in neural circuits in CORT-treated rats, particularly within the limbic system, including the thalamus, striatum, hippocampus, and lateral septum. The administration of RAAD molecules (ketamine and NLX-101) resulted in alterations to the observed connectivity profile, which appeared to leave distinct signatures. Of particular interest is the effect of NLX-101 on the connectivity between the frontal cortex and the thalamus, regions that are significantly impacted in depression.

The 48-h and 5-day timepoints were chosen to match a previous study conducted on healthy rats [53] in which the authors demonstrated a persistence of ketamine’s effects up to 5 days (beyond the molecule’s half-life of 2–3 h), in line with findings in humans [54]. Unfortunately, the long-term antidepressant effects of RAADs are believed to result from modifications in neuronal plasticity [17, 52, 55, 56], but no long-term metabolic alterations were identified in our animal model. Further investigation is therefore warranted to identify suitable animal model(s) that will allow study of RAAD drug candidates by brain imaging techniques.

Conclusion

This study represents a significant advance in the understanding of metabolic brain patterns of rapid-acting antidepressant candidates through functional PET imaging in a CORT rat model. These results highlight the potential of metabolic PET imaging in detecting the functional activation of brain regions linked to therapeutic effets and their translational value allowing the same approaches to be considered in humans. In particular, the findings highlight the critical role of the frontal cortex and lateral septum in mediating the antidepressant effects of ketamine, through the blocking of NMDA receptors, and of the biased agonist NLX-101, through the activation of postsynaptic 5-HT_1A heteroreceptors. The pattern of activation differs between the two drugs, thus providing an alternative to the current ketamine strategy. Thus, the development of biased agonists for 5-HT_1A receptors offers a more precise targeting of responses in specific brain regions, and potentially enhanced therapeutic benefits while minimizing side effects. Looking ahead, even minor structural variations among biased agonists can result in significant differences in vitro and in vivo, leading to a range of potential therapeutic applications [57]. A number of new molecules are currently under investigation, including NLX-204, which may offer distinct properties and potentially superior therapeutic profiles [58].

Author contributions

SC carried out the experiments with behavioral tests and PET imaging, performed data analysis, participated in the study design, and wrote the manuscript. EL coordinated the study, carried out the experiments and participated in its design. SB carried out the experiments with behavioral tests and samples collection, and participated in the study design. CB carried out the experiments with PET imaging. BV performed metabolic connectivity data analysis and reviewed the manuscript. AF carried out corticosterone dosages. ANT provided NLX-101 and reviewed the manuscript. LZ initiated the study, participated in its design and reviewed the manuscript. All authors have read and approved the final manuscript.

Funding

Dr. Adrian Newman-Tancredi is an employee and stockholder of Neurolixis. The other authors report no conflict of interest and have nothing to disclose.

Data availability

Supplementary data related to this study can be provided upon reasonable request to the authors.

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

All experiments were conducted in accordance with the European guidelines for the care and use of laboratory animals (Directive 2010/63/EU) and were approved by the University of Lyon review board (CELYNE) with the registration number APAFIS#29967.