The decanoate esters of nandrolone, testosterone, and trenbolone induce steroid specific memory impairment and somatic effects in the male rat

Long-term use of anabolic androgenic steroids (AAS) in supratherapeutic doses is associated with severe adverse effects, including physical, mental, and behavioral alterations. When used for recreational purposes several AAS are often combined


Introduction
The illicit use of anabolic androgenic steroids (AAS) for recreational purposes is a problem in many communities world-wide (Andreasson and Henning, 2019;Hearne et al., 2021;Kanayama and Pope, 2018;Sagoe et al., 2014), but is most apparent among male weightlifters.When used for its muscle enhancing properties, AAS are often used in supratherapeutic doses and for a long period of time.This type of usage is associated with serious physical health problems and somatic effects, reviewed in (Goldman and Basaria, 2018), as well as psychological effects, e.g.altered behavior and neuropsychiatric problems (Hauger et al., 2021;Ip et al., 2012;Kanayama et al., 2008;Lindqvist Bagge et al., 2017;Piacentino et al., 2015;Scarth et al., 2022a;Scarth et al., 2022b).Despite the fact that only a minority of the AAS-users develops severe altered behavior (Kanayama et al., 2010), further studies on AAS-induced behavioral alterations are of importance, especially the longterm effects on the brain.Moreover, the AAS-induced impact on the brain and body is suggested to be AAS-specific, but the rationale behind the adverse effects is still not clear and needs further investigation.
Cognitive problems and altered executive function are reported among long-term AAS-users (Bjørnebekk et al., 2019;Hauger et al., 2019;Hauger et al., 2020;Heffernan et al., 2015;Vaskinn et al., 2020), and the deficits seem to be associated with total life-time exposure to AAS (Kanayama et al., 2013).The impairment of cognitive function caused by AAS is associated with certain morphological and functional changes in the brain, and brain areas involved in cognitive control seem to be particularly affected (Bjornebekk et al., 2017;Bjørnebekk et al., 2019;Kaufman et al., 2015;Westlye et al., 2017).Furthermore, longterm AAS exposure is reported to be associated with deviant brain aging (Bjørnebekk et al., 2021) and long-term use of high-dose AAS is suggested as a risk-factor for developing neurodegenerative disorders (Kaufman et al., 2019;Pope et al., 2017).Therefore, increased awareness of the psychological side-effects caused by AAS is important in order to provide better treatment and healthcare for these patients.
To investigate AAS-specific effects in human users is difficult, as several AAS compounds often are combined and polydrug abuse is common among these individuals.Because of this there is a need of experimental animal studies, in order to investigate possible AASspecific effects and the underlying mechanisms causing these problems.Studies in rodents have demonstrated AAS-induced changes in a wide-range of neurosignaling pathways (Carrillo et al., 2011;Gronbladh et al., 2014;Hallberg, 2011;Kindlundh et al., 2002;Le Greves et al., 1997;Lindqvist et al., 2002;Nyberg and Hallberg, 2012;Ricci et al., 2012;Tucci et al., 2012) and several in vitro studies report AAS to have neurotoxic effects by altering neurite development (Zelleroth et al., 2021), inducing cell death (Basile et al., 2013;Estrada et al., 2006;Zelleroth et al., 2019), and increasing cellular vulnerability to other toxic insults (Caraci et al., 2011;Orlando et al., 2007).Furthermore, recent findings indicate that long-term heavy AAS use contributes to increased neuroinflammation, oxidative stress, and impaired neuroplasticity, which may correspond to the changes in mood and cognition reported by AAS-users (Damião et al., 2021;El-Shamarka et al., 2020b;Moradpour et al., 2019;Scarth and Bjørnebekk, 2021).
Nandrolone and testosterone are two of the most common AAS used for recreational purposes (Börjesson et al., 2020), and have both previously been shown to affect cognitive abilities and executive functions in different types of behavioral tests when studied in rodents (Ahmed and El-Awdan, 2015;El-Shamarka et al., 2020a;Kouvelas et al., 2008;Magnusson et al., 2009;Pieretti et al., 2013;Tanehkar et al., 2013;Wallin et al., 2015;Wallin-Miller et al., 2018;Wood and Serpa, 2020).Trenbolone is another AAS especially used by heavy AAS-users, which have reported this compound to cause particular physical and psychological harms (Piatkowski et al., 2023;Underwood, 2022).However, there are few studies in the literature investigating trenbolone, especially regarding the effects on behavior when used for recreational purposes.The aim of the present study was to investigate steroid specific impact and differences between nandrolone decanoate, testosterone decanoate, and trenbolone decanoate when administered to the male rat in a supratherapeutic dose (15 mg/kg) mimicking heavy AAS use.To evaluate and compare the impact of the different AAS on learning and memory the novel object recognition test (NOR-test) was used.Furthermore, the multivariate concentric square field test (MCSF-test) was used to investigate AAS-induced effects on behavioral profile, including general activity, exploratory behavior, shelter-seeking behavior, risk-assessment, and risk-taking behavior.Moreover, body weight gain, organ weights, and plasma corticosterone levels were measured to determine possible AAS-specific somatic effects.

Animals
The male Wistar rats (n = 48) included in the study were obtained from Envigo (Netherlands) and arrived to the animal facilities at five weeks of age.The rats were housed under standardized housing conditions (i.e., 20-24 • C and 45-65 % humidity) in groups of three rats per cage (cage type IV, 59 × 38 × 20 cm), which had elevated cage lids and bedding consisting of wood-chip, paper sheets, and a wooden house as enrichment.Before the start of any experimental procedures the rats were allowed two weeks of acclimatization in order to adapt to the new environment and the reversed 12 h light/dark cycle, which was separated with a dusk/dawn shift with the lights off at 07:00 h.Also, prior to start of the experiment the main experimenter handled all animals to allow the rats to habituate to the different situations the rats will be exposed to e.g., being held, carried, weighed, and transported to test rooms.The rats were monitored daily and had access to pelleted food (type R36, Lantmännen) and tap water in their home cages ad libitum.In addition, animal health was monitored by weight (every third day), as well as posture, eye, and fur condition daily throughout the study.Housing rooms and test rooms were kept at similar conditions and all rooms had masking background noise.All experimental procedures including animals in the present study followed Swedish Legislation on Animal Experimentation (Animal Welfare Act SFS 1998:56) and the European Union Directive on the Protection of Animals Used for Scientific Purposes (2010/63/EU) and were approved by the local animal ethics committee in Uppsala (5.8.18-02249/2017).

AAS treatment
The AAS, nandrolone, testosterone and trenbolone, were purchased from Sigma Aldrich.The esterification of the AAS was conducted as described below in Section 2.3.The AAS were dissolved in peanut oil (50 mg/mL, Sigma Aldrich).The rats were randomized into four treatment groups, with 12 rats per group, either receiving nandrolone decanoate, testosterone decanoate, trenbolone decanoate, or peanut oil (controls).To mimic heavy AAS-use, the rats received subcutaneous injections of 15 mg/kg of the selected steroid on their upper back, with a maximum volume of 100 μL, every third day for a total of 24 days.This dose has previously been used in studies investigating behavioral effects of AAS (Ahmed and El-Awdan, 2015;El-Shamarka et al., 2020a;Gronbladh et al., 2013;Lindqvist et al., 2002;Magnusson et al., 2009;Steensland et al., 2005;Tanehkar et al., 2013;Zelleroth et al., 2022) and was selected to be able to compare results to previous findings.The injection site was varied to prevent skin irritation.

Novel object recognition test (NOR)
The NOR-test was first described by Ennaceur and Delacour (Ennaceur and Delacour, 1988) where recognition memory was studied by investigating the ability of the rat to recognize and remember different objects.The test procedure used in the present study is a modified version of the NOR-test described by Feltmann et al., (Feltmann et al., 2015).The NOR-arena consisted of a rectangular box (80 × 30 × 50 cm) with an open roof, black solid walls and visual cues of different blackand-white patterns on the short sides.The floor was primed with bedding material from the home cages and the illumination in the arena was 9 lx.The objects used for discrimination were a Lego brick cube and a ceramic cup in similar size.The objects were placed at the short ends of the arena using adhesive tape or weights, in order for the rats not to be able to move the objects.The NOR-test was performed when the rats were 9 weeks of age over three consecutive days (experimental day 17-19), including habituation (day 1), training (day 2), and testing (day 3).An illustration of the NOR-arena and test set-up is shown in Fig. 1.For all NOR-sessions the rats were transferred to the test room in a transportation-bucket and placed in the middle of the NOR-arena facing the wall.During the habituation session the rats were allowed to get familiar with the arena for 20 min.The training session was performed 24 h later and two identical objects were placed in opposite ends of the arena.The rat was allowed to explore the arena and the two objects for 15 min.Another 24 h later the test session was performed.Then, the object the rat spent the least amount of time exploring during training was replaced by a novel object.The rat was allowed to explore the arena with the novel object and the familiar object for 5 min.Between trial sessions fecal boli was removed and the objects were cleaned with 70 % ethanol.The object combinations were counterbalanced within treatment groups.The EthoVision system version 15 (Noldus Information Technology) was used to record the NOR-sessions for each animal and the time the rats spent exploring the objects (training and testing) was manually scored together with midline crossings (during training) as a measure of locomotor activity.Exploration of the objects was defined as sniffing, licking, and biting.Sitting, climbing or leaning against the objects without active exploration was not included.Because of the innate preference of novelty rats will spend more time exploring the novel object compared to the familiar object during the test-session, indicating remembrance of the familiar object.Data obtained from the testing session was expressed as discrimination ratio (exploration of the novel object divided by the sum of exploration of both objects).

Multivariate Concentric Square Field™ test (MCSF)
The MCSF-test was developed by Meyerson andcolleagues in 2006 (Meyerson et al., 2006).The test is based on forced exploration and allows observation of behaviors associated with shelter seeking, general activity, exploratory behavior, risk assessment, and risk-taking.In the multivariate test situation, the rat has the free choice to visit areas with different qualities contained in the same arena and trial session, and thereby generates a behavioral profile rather than single parameters.The MCSF-test has been described in detail by Meyerson et al., (Meyerson et al., 2006).Briefly, the arena provides different areas for the rat to explore including elevated, sheltered, and open areas, a holeboard device, and areas of different illumination, illustrated in Fig. 2. The arena consists of an outer square field (100 × 100 cm) with black solid walls on three sides and a transparent wall on the fourth.An inner square field (70 × 70 cm) separates the center from the peripheral Fig. 1.Experimental set up of the novel object recognition (NOR) test.The rats were allowed to habituate to the NOR arena for 20 min, after 24 h delay the training session started and the rats were presented with two identical objects for 15 min, after another 24 h delay the rats were tested for discrimination of the familiar and a novel object during the 5 min testing session.S. Zelleroth et al. corridors, which are to be reached through openings in the internal walls.The open area in the middle of the center is called central circle (25 cm in diameter and 10 lx).Enclosed at the end of one corridor is a covered shelter located, called the dark corner room.From the two other corridors, the hole-board device (hurdle) is accessible, which is elevated 10 cm from the floor and consists of a platform with two holes and a photocell to register the number of nose pokes.Along the side with the transparent wall a stainless-steel wire-mesh construction is located.The inclined beginning of the construction, called the slope, leads up to the brightly illuminated (100 lx) bridge.When the rats entered the bridge, they were subjected to an aversive air-puff.Using the air-puff as a threatening stimulus in the MCSF has previously been used to evaluate avoidance learning and memory, as rats exposed to the air-puff avoid the bridge (Karlsson et al., 2009).In the present study, the air-puff was included with the intention to increase the risk to enter the bridge.All rats were tested at 10 weeks of age (experimental day 23) during the dark phase of the light/dark cycle, in mixed order to avoid time-and order bias.The rats were transferred from the home cage to the test room in a transportation-bucket and placed in the center of the MCSF facing the wall without openings.Recordings started immediately and continued for 20 min.Each trial session was observed from an adjacent room were the observer manually performed the air puff when the rat entered the bridge.After each trial session, the floor of the MCSF apparatus was wiped with 10 % (v/v) ethanol, and allowed to dry before the next rat was placed in the arena.The EthoVision system version 15 (Noldus Information Technology) was used to automatically score distance moved (total, cm) and velocity (mean, cm/s) during the MCSFtest.Direct observation was used to score rearing and grooming.After each trial, the number of urinations, fecal boli, and nose-pokes in the hurdle were noted.EthoVision was used for scoring of the zone measures, latency (L, s) to first visiting a zone, frequency (F) of visits in specific zone, and duration (D, s) of total time spent in specific zone.In addition, following parameters were derived for analysis; latency to leave center (L leave), the mean duration per visit in specific zone (D/F), sum of all frequencies (total activity), duration and D/F, respectively, for the corridors (total corridors), the frequency and duration of risk/shelter indices ((F bridge -F dark corner room)/(F bridge + F dark corner room) and (D bridge -D dark corner room)/(D bridge + D dark corner room)), and the slope/bridge interval ((L slope -L bridge)/L slope).

Trend analysis
A statistical method, called trend analysis, is described to be used in order to measure and evaluate variation in behavioral traits monitored in the MCSF-test (Meyerson et al., 2013).This statistical approach is based on the fact that rats may display different behavioral strategies dependent on their personality traits.In the MCSF-test, each parameter is associated with a specific type of behavior and when the parameters that correlates are grouped together, they form a functional category i.e., general activity, exploratory behavior, risk assessment, risk-taking, and shelter seeking behavior.The trend analysis is a rank-order procedure where the rats are ranked against each other in each parameter.The animal with the lowest score gets the lowest rank and the animal with the highest score is given the highest rank.The rank values are then summed together so each rat gets a summed rank value for each functional category.Thus, the group-wise comparison is based on the relative position of the rat within the entire population.Following parameters were selected for the different functional categories; general activity (total activity, distance arena, F center, F and F/D total corridors), exploratory activity (D total corridors, D center, D hurdle, rearing, and nose pokes), risk assessment (D, F and D/F slope), risk-taking (D, F and D/F for bridge and central circle), and shelter seeking (D, F and D/F dark corner room).The parameters that are negatively related to the functional category (D center, D and D/F total corridors) were inversely ranked before the rank values were summed.

Tissue collection
The rats were euthanized by decapitation the day after last behavioral testing (experimental day 24).Blood was collected in lithium/ heparin tubes (Sarstedt) and centrifuged at 3000 rpm for 10 min at 4 • C to obtain blood plasma.Peripheral organs i.e., liver, right kidney, right testicle, heart, and thymus gland were dissected and weighed before they were put on dry ice.All samples were stored in − 80 • C until further biochemical analysis.

Corticosterone plasma concentrations
The commercial corticosterone ELISA kit (ab108821, Abcam), was used to measure corticosterone levels in blood plasma obtained from the rats.The assay was performed in accordance with the instructions from the manufacturer with minor modifications.Briefly, all reagents were equilibrated to room temperature and prepared according to the provided protocol.A 100-fold dilution of the plasma samples was recommended and all standards, controls and samples were assayed in duplicates.The mean absorbance was detected at 450 nm and 570 nm using the FLUOstar Omega spectrophotometer (BMG Labtech).

Statistical analysis
The data was managed using Microsoft Excel version 16.78 (Microsoft) and statistical calculations were performed using GraphPad software (Prism 9.0).Measurements of body weight and organ weights, together with the measurements of plasma-corticosterone levels were normally distributed according to the Shapiro-Wilk's normality test.Therefore, body weight development was analyzed using two-way repeated measurement analysis of variance (ANOVA), with time and treatment as factors to account for differences.When an overall effect was found, further group-wise comparison was made with Dunnett's post hoc test.Differences in organ weights and corticosterone levels between treatment groups were analyzed with one-way ANOVA, followed by Dunnett's post hoc test.The parameters obtained from the MCSF test together with the NOR data did not follow a normal distribution according to the Shapiro-Wilk's normality test.Therefore, the Kruskal-Wallis test was used to analyze the MCSF parameters, discrimination ratio, and midline crossing.Two-way ANOVA was used to analyze preference (exploration time) of objects, using objects and treatment as factors to account for differences.Repeated measurements two-way ANOVA was used to analyze preference of side in the NOR arena, using sides (dots or lines) and treatment to account for differences when appropriate.When an overall effect was found, further group-wise analysis was made using Dunn's multiple comparison test.Furthermore, the discrimination ratio was investigated across the test trial minutes (1-5 min).If a rat did not explore any of the two objects during the specific minute of interest the rat was excluded from the analysis.As the repeated measurements ANOVA cannot handle missing values, the result was statistically analyzed using the mixed-effects model with the Geisser-Greenhose correction using "trial minutes" and "treatment" as factors.The results from the trend analysis were statistically analyzed using one-way ANOVA.A p-value <0.05 was considered statistically significant and effect size eta-squared (η 2 ) was calculated for all significant parameters identified.

Multivariate data analysis
The MCSF-test generates large data sets with several dependent variables, i.e., multivariate data.Therefore, Meyerson and colleagues suggest multivariate data analysis to be performed in combination with the trend analysis and traditional statistics in order to extract additional information, define relationships, and to get an overview of the results obtained in the MCSF-test (Meyerson et al., 2006;Meyerson et al., 2013).Multivariate data analysis was performed using SIMCA 15 (Umetrics AB, Umeå, Sweden) as a complement to the conventional statistic, to summarize, and visualize the data obtained from the MCSFtest.The principal component analysis (PCA) is an unsupervised method that examines the relationship between several X:s (parameters from the MCSF).The result from the PCA is presented in two-dimensional score and loading plots generated from the two first components.The score plot illustrates a summary of possible relationships between the individuals and enables interpretation of the relatedness between individuals and identification of groups and outliers.The loading plot identifies study parameters important for creating these relationships and enables grouping of closely related parameters.Graphical interpretation is made by comparing the spatial relationship of two points by connecting them with a straight line and the origo.The angle formed explains the relationship between the points, where a straight line through origo implies negative correlation, a 90 • angle indicate the points to be independent and if the points form a narrow angle they correlate in a positive manner.In the PCA, all MCSF-parameters, except latencies, obtained from the MCSF test were included.The autofit option was used to create the models resulting in the maximal number of significant components.

Results
The present study examined the impact of nandrolone decanoate, testosterone decanoate, and trenbolone decanoate on the male rat by evaluating the effects on recognition memory, behavioral profile, body weight development, organ weights, and plasma corticosterone levels.The findings revealed diverse AAS-induced effects indicating the specific type of AAS to be of importance both for the physiological and psychological outcome when used in a supratherapeutic dose.

Body weight gain
The AAS-treatments caused a reduction in body weight development (Fig. 3) indicated by the two-way ANOVA as an effect of treatment was identified (F (3, 44) = 6.8; p = 0.0007; η 2 = 0.06).The post hoc test was used to investigate differences between the treatment groups at each specific time point (experimental day), which revealed a significant difference in weight gain between nandrolone decanoate treated rats and the control at day 13 (Dunnett's test; p = 0.0075).Thereafter, this significant difference lasted throughout the study (experimental day 13-24), indicating nandrolone decanoate to cause slower weight development.Furthermore, a significant difference between trenbolone decanoate and control was found as from day 19 (Dunnett's test; p = 0.0039), and at the end of the study (day 24) a significant difference between testosterone decanoate treated rats and control (Dunnett's test; p = 0.036) was found as well.All rats significantly increased their body weight gain throughout the study as an effect of time was identified (F (8, 352) = 535.5;p < 0.0001; η 2 = 0.72), and there were no differences in body weight between the groups at the beginning of the study (data not shown).

Organ weights and corticosterone levels
Specific AAS-induced differences on organ weights were found.A significant difference in the thymus weight was observed between the treatment groups (F (3, 44) = 131.3;p < 0.0001; η 2 = 0.90), and the post hoc test revealed all three AAS (Dunnett's test; p < 0.0001) to significantly reduce the weight of the thymus in comparison with the control (Fig. 4a).In addition, an overall effect on liver weight was identified (F (3, 43) = 6.156; p = 0.0014; η 2 = 0.3), and nandrolone decanoate (Dunnett's test; p = 0.0015) was shown to significantly reduce the weight of the liver in comparison to the control (Fig. 4b).Furthermore, a significant difference in the kidney weight was observed (F (3, 44) = 91.85;p < 0.0001; η 2 = 0.86), and the post hoc test revealed an increase of the weight of the kidney to be caused by nandrolone decanoate (Dunnett's test; p < 0.0001) and trenbolone decanoate (Dunnett's test; p < 0.0001) in comparison to the control (Fig. 4c).Also, an overall effect on testicle weight was found (F (3, 43 = 14.17; p < 0.0001; η 2 = 0.50), an effect demonstrated to be caused by testosterone decanoate (Dunnett's test; p = 0.0008) which significantly reduced the testicular weight (Fig. 4d).Moreover, the weight of the heart (F (3, 44) = 1.6; p = 0.2) and the corticosterone plasma levels (F (3, 42) = 0.26; p = 0.85) were not affected by the AAS treatments (Fig. 4e-f).In the statistical analysis one liver, one testis, and two plasma samples are missing due to technical issues during the dissection.

Recognition memory (NOR-test)
The AAS-induced effects on recognition memory were studied in the NOR-test, where an overall significant difference was identified between the treatment groups using the Kruskal-Wallis test (H = 8.94; p = 0.03; η 2 = 0.1).Following post hoc test revealed nandrolone decanoate treated rats to have significant lower discrimination ratio compared to the control (Dunnett's test; p = 0.02), indicating reduced remembrance of the familiar object and impaired recognition memory (Fig. 5a).No significant effect on recognition memory was found to be caused by testosterone decanoate or trenbolone decanoate.However, six of the 12 rats treated with testosterone decanoate demonstrated a discrimination ratio close to 0.5 (all below 0.54) which indicate that half of the rats treated with testosterone did not remember the familiar object.This indicates a possible negative effect on memory to be caused by repeated treatment of high-dose testosterone, even though no significant difference compared to the control group could be demonstrated in the present study.Surprisingly, when the discrimination ratio was analyzed across the test trial minutes (1-5 min) no significant differences between the trial minutes (F (3.398, 142.7) = 1.182; p = 0.32) or between the treatment groups (F (3, 43) = 0.1676; p = 0.92) could be detected (data not shown).During the training session, no preference of the objects (time exploring the Lego or cup) was found (F (1, 40) = 0.58; p = 0.44) Fig. 3. Body weight gain.Body weight measurements were obtained every third day throughout the study (experimental day 1-24) from rats treated with selected AAS (nandrolone decanoate, testosterone decanoate, or trenbolone decanoate) or control (peanut oil).The data is presented as percent of initial body weight with mean ± SD, n = 12 rats per group.Statistical testing was performed using a two-way repeated measurements ANOVA, followed by Dunnett's multiple comparison test.A p-value <0.05 was considered statistically significant (* indicates difference in comparison to the control group at that day, * < 0.05, ** < 0.01, and **** < 0.0001).and there was no difference between treatments (F (3, 40) = 0.40); p = 0.76).Furthermore, there was no significant preference of side in the arena (F (1, 44) = 2.10; p = 0.15), and no differences found between the treatment groups (F(3, 44) = 0.55; p = 0.65).Moreover, there was no difference in number of midline crossings in the arena between the treatment groups (H = 2.95; p = 0.4), indicating the effect on recognition memory not to be caused by altered exploratory behavior or motor function (Fig. 5b-d).

Discussion
In the present study AAS-specific cognitive and somatic effects were demonstrated to be caused by the structurally different AAS; nandrolone decanoate, testosterone decanoate, and trenbolone decanoate.Nandrolone decanoate was identified to cause more harm in comparison to testosterone decanoate and trenbolone decanoate, based on the fact that nandrolone decanoate caused the most prominent impact on body weight development, affected the weight of multiple organs, and was the only AAS causing impaired recognition memory in male rats.
Rats treated with nandrolone decanoate demonstrated impaired performance in the NOR-test, as they spent an equal amount of time exploring the novel and familiar object, indicating altered recognition memory.However, due to the experimental design and treatment regimen used it is not possible to conclude whether the effect is due to impaired memory acquisition, consolidation, or recall.Nevertheless, the effect on memory does not seem to be caused by altered exploratory behavior or motor function, as no effects were found in regard to preference for objects, preference of side, motor activity, or exploratory time.This was further supported by the fact that the behavioral profiles of the rats were not affected by the AAS-treatment, as indicated by the lack of separation between the groups in the PCA and as no difference in the trend analysis of the MCSF-test was detected.
The nandrolone-induced effect on memory is in line with previous literature demonstrating the same dose of nandrolone (15 mg/kg), and by using similar administration schedules as the present study, to alter cognitive function in the Morris Water Maze (MWM) task (Ahmed and El-Awdan, 2015;El-Shamarka et al., 2020b;Gronbladh et al., 2013;Magnusson et al., 2009;Tanehkar et al., 2013).Furthermore, a lower dose of nandrolone decanoate (5 mg/kg) is also demonstrated to impair the performance in the MWM (Pieretti et al., 2013).In addition, daily injections of 15 mg/kg nandrolone decanoate are shown to impair olfactory social memory in rats via activation of the androgen receptor (Kouvelas et al., 2008).The previous results, together with the present findings, suggest nandrolone to impair several types of memory functions, including spatial learning and memory, social memory, and episodic-like memory.
The NOR-test can be adapted in order to study different types of effects e.g.memory enhancing effects, memory impairments, short-term memory and long-term memory.Interestingly, daily injections of 15 mg/kg of nandrolone decanoate has previously been shown not to affect short-term memory of rats tested in the NOR-test (Struik et al., 2017).Our results, together with previous findings suggest nandrolone  decanoate to impair long-term recognition memory, but not short-term recognition memory.The mechanism behind this nandrolone-induced effect on recognition memory is not clear.However, the dopaminergic system is known to be involved in the regulation of cognitive control, and it has been suggested that dopamine release in the hippocampus is necessary for signaling novelty and needed for memory storage of the novel information (Lisman and Grace, 2005).In addition, local activation of the D 1/5 receptor in the hippocampus, as well as the medial prefrontal cortex, is shown to potentiate recognition memory in rodents, while a local injection of a D 1/5 receptor antagonist prevents the formation of episodic-like memory (De Bundel et al., 2013).Furthermore, a recent study demonstrates increased release of dopamine and norepinephrine in the rat hippocampus as a response to novelty when studied in an adapted version of the NOR-test (Titulaer et al., 2021).These findings further support the suggestion of dopamine to signal novelty within the hippocampus and to be involved in the facilitation of memories.The literature describing the effects of AAS on monoaminergic systems of the brain is quite limited, but there are some reports indicating AAS to affect the dopamine system in the rat brain.For example, an early study indicates long-term treatment (5 mg/kg) of both nandrolone and testosterone to increase dopaminergic metabolism in rat brain (Thiblin et al., 1999).Furthermore, nandrolone (15 mg/kg) is previously demonstrated to affect the density of D 1 -like receptors and D 2 -like receptors in the rat brain (Kindlundh et al., 2001), as well as cause an increase in dopamine transporter density in the striatum (Kindlundh et al., 2004).Interestingly, long-term high-dose AAS-use is demonstrated to cause structural changes in brain morphology, including reduced volume of total gray matter and cortical regions (Bjornebekk et al., 2017) and the pre-frontal cortex of dependent users seem to be particularly affected (Hauger et al., 2019).In addition, these individuals exhibit poorer executive function and suffer from psychological distress, as compared to non-dependent users who instead display impaired impulse inhibition and working memory (Hauger et al., 2020).Furthermore, a recent study in rats demonstrated nandrolone to increase inflammatory markers and oxidative stress in both the hippocampus and pre-frontal cortex (El-Shamarka et al., 2020a).Additionally, high-dose AAS treatment is suggested to affect neuroplasticity in limbic regions, including the hippocampus, of mice (Damião et al., 2021), and we have previously demonstrated nandrolone, testosterone, and trenbolone to cause cell death and reduction in neurite length using rat cortical cell cultures (Zelleroth et al., 2022;Zelleroth et al., 2019).
In the present study testosterone decanoate and trenbolone decanoate did not significantly impair recognition memory.However, the result indicated a possible testosterone-induced reduction in recognition memory, as half of the animals treated with testosterone displayed a discrimination ratio close to 0.5.Interestingly, testosterone is suggested to have both neurotoxic and neuroprotective effects (Hammond et al., 2001) and a previous study has shown testosterone to enhance memory function in rats when tested in the NOR-test (Frye and Lacey, 2001).A possible explanation is that low concentrations of testosterone have neuroprotective effects, while high concentrations are neurotoxic (Orlando et al., 2007).Furthermore, many of the psychological effects of AAS are suggested to be idiosyncratic (Kanayama et al., 2010) and it is therefore possible that AAS-users respond differently to AAS based on their personality and individual vulnerability.The individual variation in the testosterone decanoate treated group could be an indication of this, which would explain why some of the rats do remember the familiar object while others do not.Similarly, relatively large individual variation was seen in the behavioral profiles obtained from the MCSFtest in all treatment groups.These individual differences probably reflect the variation within the general population.However, this is only a speculation and further studies are needed to confirm this.Therefore, in order to investigate the effects of AAS within a group of individuals with a certain type of personality or specific vulnerability, it would be interesting to design a study where the behavior of the rat is characterized before the start of AAS treatment.
Furthermore, long-term treatment with lower doses of testosterone (2 mg/kg and 7.5 mg/kg) has previously been shown not to affect learning and memory in rats (Clark et al., 1995;Wood and Serpa, 2020), but may however impair cognitive flexibility, as treatment with testosterone (7.5 mg/kg) alters decision-making in the operant discounting task (Wallin et al., 2015) and in the rodent version of the Iowa gambling task (Wallin-Miller et al., 2018), as well as the biconditional task performance (Wood and Serpa, 2020).Moreover, we were surprised by the fact that trenbolone decanoate did not cause a similar effect on memory as compared to nandrolone decanoate, as we seen similar neurotoxic effects between these AAS in our previous in vitro studies (Zelleroth et al., 2019;Zelleroth et al., 2021).Trenbolone is a nandrolone derivative and displays similar features as nandrolone, such as a high anabolic: androgenic ratio (Yarrow et al., 2010) and not being converted to estrogen by aromatase (Donaldson et al., 1981;Quinn et al., 2007).Moreover, trenbolone is suggested to cause neurodegeneration (Ma and Liu, 2015), as well as being mentioned by AAS users to cause particular harm, although, reported features that seem to be particularly affected are increased aggression, violent behavior, and issues with impulsivity regulation (Piatkowski et al., 2023;Scarth and Bjørnebekk, 2021;Underwood, 2022).
Nandrolone decanoate had the most prominent impact on body weight development, which was expected as this effect has been described previously both by us (Gronbladh et al., 2013;Zelleroth et al., 2022) and by others (Lindblom et al., 2003), and has been suggested as a result of a reduction in food intake (Lindblom et al., 2003).Interestingly, a similar effect was found on body weight development following treatment with trenbolone decanoate, although, detected at a later timepoint during the study.The delayed effect could possibly be explained by the pharmacological differences between nandrolone and trenbolone (Bauer et al., 2000;Death et al., 2004;Durhan et al., 2006;Scippo et al., 2002;Spranger and Metzler, 1991).Alternatively, the supratherapeutic dose used may increase the chance of affecting other receptor systems, e. Fig. 6.The MCSF trend analysis.The trend analysis is a rank-order procedure where the rats are ranked against each other in each descriptive parameter.The descriptive parameters are associated with a specific type of behavior and when correlated descriptive parameters are grouped together, they form a functional category i.e., general activity, exploratory behavior, risk assessment, risktaking, and shelter seeking behavior.Individual values are presented together with the mean, boxes indicate 5-95 percentile, and error bars show min to max, n = 11-12 rats per group.Statistical analysis was performed using the one-way ANOVA followed by Dunnett's multiple comparison test and p < 0.05 was considered statistically significant.
g. trenbolone has high affinity to the progesterone receptor (Bauer et al., 2000), suggesting a compensatory mechanism of action.Moreover, at the end of the study, a significant difference in weight gain was found between testosterone decanoate and control as well.This is something we did not observe in previous study (Zelleroth et al., 2022), however, the different results could probably be explained by the extended study time (24 days compared to 21 days) or possibly by the difference in esters used (testosterone decanoate compared to testosterone undecanoate).In general, the longer ester-chain length, the more slowly the preparation is released into circulation which increases the duration of action (Kicman, 2008).
In addition to the altered body weight gain, the AAS-specific effects on organ weights further highlight the diverse impact caused by different AAS.The reduction of the thymus gland confirms the systemic effect of the AAS investigated, as it is well known that steroids induce thymus atrophy (Grönbladh et al., 2013;Hince et al., 2008;Johansson et al., 2000).Interestingly, the effect on thymus could also indicate an AAS-induced effect on immune function.The long-term effects of AAS on the immune system are not clearly elucidated.However, the literature suggests AAS use to influence immunological function, and that the effects are dependent on the dose and type of AAS used, reviewed in (Marshall-Gradisnik et al., 2009).Moreover, AAS-induced alterations in both physiology and morphology of the liver and kidneys have previously been described (Herlitz et al., 2010;Modlinski and Fields, 2006).Hepatic adverse effects are one of the most common side-effects caused by AAS (Albano et al., 2021), but is primarily associated with 17αalkylated AAS (Goldman and Basaria, 2018).However, in the present study nandrolone decanoate caused a reduction in liver weight.Supratherapeutic doses of nandrolone decanoate have previously been demonstrated to affect liver morphology, enzymes, and proteins in rats (Vieira et al., 2008).In addition, the kidney weight was increased by nandrolone decanoate and trenbolone decanoate, but not testosterone decanoate.Nandrolone decanoate is previously shown to cause cell growth and hypertrophy in the kidney (Brasil et al., 2015), whereas testosterone only affect the kidney size when combined with physical training (Bento-Silva et al., 2010).Furthermore, testosterone decanoate was the only AAS investigated that affected testicular weight.Testicular atrophy is a well-known problem in long-term AAS-users (Duca et al., 2019).In general, AAS are thought to inhibit the endogenous testosterone production through negative feedback of the HPG-axis, resulting in low levels of gonadotropins and endogenous testosterone causing hypogonadism in men (Christou et al., 2017).
In conclusion, nandrolone decanoate, testosterone decanoate, and trenbolone decanoate induce diverse somatic effects and affect recognition memory differently.Nandrolone decanoate was identified to cause the most harm as it affected not only body weight gain, as well as the thymus, liver, and kidney weights, but also caused an impaired memory in the male rats.Furthermore, the overall behavioral profile and stress hormone levels were not affected.To summarize, we emphasize the importance of increased awareness due to the diverse effects induced by different AAS compounds.

Fig. 2 .
Fig. 2. Illustration of the MCSF arena.The arena (1 × 1 m) is divided into zones by walls or imagined boundaries (dashed lines).Zones in the arena: center, central circle, corridors, dark corner room, hurdle, slope, and bridge.The aversive air-puff is located in the entrance of the bridge.

Fig. 4 .
Fig. 4. Organ weights and corticosterone levels.Weights of A) thymus, B) liver, C) kidney, D) testis, E) heart, and F) plasma concentration of corticosterone obtained from rats receiving either nandrolone decanoate, testosterone decanoate, trenbolone decanoate, or control (peanut oil).Statistical analyses were made using one-way ANOVA, followed by Dunnett's post hoc test when appropriate.A p-value <0.05 was considered statistically significant (* indicates differences in comparison to the control group, ** < 0.01, *** < 0.001, and **** < 0.0001).Organ weights are presented as percent of body weight, and corticosterone levels are presented as percent of control, with individual values together with mean ± SD, n = 10-12 samples per group.Abbreviations: control (C), nandrolone decanoate (ND), testosterone decanoate (TD), and trenbolone decanoate (TrD).

Fig. 5 .
Fig. 5. Novel object recognition.A) discrimination ratio [(time exploring novel object) / (time exploring both objects)] obtained during the test session of the NORtest.B) midline crossings, C) exploration of objects, D) preference of side (exploration time on either left or right side of the arena), during the training session of the NOR-test.Statistical calculations were made with the Kruskal-Wallis test, two-way ANOVA, or repeated measurement two-way ANOVA when appropriate.When an overall effect was found, further group-wise analysis was made using Dunn's multiple comparison test.Individual data points are presented together with median and interquartile range, n = 12 rats per group.A p-value <0.05 was considered statistically significant (* indicates difference in comparison to the control group).Abbreviations: novel object recognition (NOR), control (C), nandrolone decanoate (ND), testosterone decanoate (TD), and trenbolone decanoate (TrD).

Table 1
Behavioral parameters obtained during the 20-min trial of the MCSF-test.