DOI: 10.31038/EDMJ.2023731

Abstract

Aims: Accelerated cognitive decline frequently complicates traumatic brain injury. Obesity and type 2 diabetes mellitus drive peripheral inflammation which may accelerate traumatic brain injury-associated neurodegeneration. The Zucker rat harbors G-protein coupled receptor agonist IgG autoantibodies and in vitro neurotoxicity caused by these autoantibodies was prevented by a novel synthetic fragment of the serotonin 2A receptor. The aim of the present study was to test whether genetic obesity manifested in Zucker diabetic fatty rat is associated with greater spatial memory impairment before and after mild traumatic brain injury compared to Zucker lean rats. Furthermore, we investigated whether these neurodegenerative complications can be lessened by administration of a novel putative neuroprotective peptide comprised of a fragment of the second extracellular loop of the serotonin 2A receptor.

Methods: Age-matched lean and fatty diabetic Zucker rats were tested in the Morris water maze (spatial memory) prior to receiving a sham-injury or lateral fluid percussion (LFP) mild traumatic brain injury. Behavioral testing was repeated at 1-week, 1-month, and 3-month intervals following injury. A synthetic peptide consisting of a portion of the 5-hydroxytryptamine (serotonin) 2A receptor (2 mg/kg) (vehicle, or an inactive scrambled version of the peptide (2 mg/kg)) was administered via intraperitoneal route every other day for 7 days after sham or LFP injury to lean rats or 7 days before and after sham or LFP injury to fatty rats.

Results: Mild traumatic brain injury impaired recall of spatial memory in fatty and lean rats. Zucker fatty rats subjected to sham-injury or mild TBI experienced a significantly greater longitudinal decline in recall of spatial memory compared to lean Zucker rats. A synthetic peptide fragment of the 5-hydroxytryptamine 2A receptor significantly enhanced acquisition of spatial learning and it appeared to strengthen recall of spatial learning (one-week) after sham injury in Zucker rats.

Conclusions: These data suggest that the Zucker diabetic fatty rat is a suitable animal model to investigate the role of metabolic factor(s) in accelerated cognitive decline. A novel synthetic peptide comprised of a fragment of the second extracellular loop of the human serotonin 2A receptor appeared to have neuroprotective effects on both acquisition and recall of spatial memory in subsets of Zucker rats, with relatively greater benefit in sham-injured, lean Zucker rats.

Introduction

Cognitive dysfunction increases substantially following Traumatic Brain Injury [1] (TBI) contributing to substantial morbidity and mortality in affected persons. Peripheral and central inflammation drive neurodegeneration via activation of innate and adaptive immune mechanisms which may target (in part) neurovascular antigens released during traumatic brain injury. Owing to a global epidemic of obesity and type 2 diabetes mellitus [2], lifetime TBI-sufferers having metabolic derangements associated with peripheral inflammation may be at increased risk for experiencing certain neurodegenerative complications [3].

The Zucker diabetic fatty rat (ZDF) is a model of morbid obesity, type 2 diabetes mellitus and hypertension [4] which exhibits high level of innate immunity, i.e. pro-inflammatory cytokines [5]. Previously, we reported spontaneously-occurring neurotoxic agonist autoantibodies targeting the 5-hydroxytryptamine (serotonin) 2A receptor in plasma from both ZDF rat and Zucker lean rat (ZLR) sub-strains with the ZDF rats having persistently high level at 5 months of age [6]. The serotonin 2A receptor is a known treatment target in major depressive disorder and Parkinson’s disease, two neurodegenerative complications following TBI. Here we tested for changing in the acquisition and/or recall of spatial learning in substrains of Zucker fatty and lean rats following mild TBI (induced by lateral fluid percussion) or sham injury.

In addition, a putative neuroprotective peptide comprised of a 5-hydroxytryptamine (5HT) 2A receptor second extracellular loop region fragment was effective in preventing in vitro neurotoxicity caused by autoantibodies targeting the serotonin 2A receptor from both ZDF rat plasma [6] and human neurovascular disorders plasma [7]. Therefore the efficacy of this peptide in modulating TBI- (or sham-injury associated spatial memory impairments in fatty and lean Zucker rats was tested. The mechanism of the action of the 5HT2AR peptide fragment was explored using alanine substitution of key functional amino acid residues.

Materials and Methods

Peptides

A linear synthetic peptide, SCLLADDN (SN..8 or “P4”) having a sequence identical to that of a fragment of the second extracellular loop region of the human 5-hydroxytryptamine 2A receptor was synthesized at Lifetein, Inc (Hillsborough, NJ) and had ≥95% purity. Substitutions of SCLLADDN containing a single alanine amino acid replacement, e.g. SALLADDN, SCLLADAN were synthesized at Lifetein, Inc (Hillsborough, NJ) and had purity of ≥95%. A “scrambled” peptide containing the same amino acids as SCLLADDN had a sequence of LASNDCLD (LD.8) and a purity of 96.37%, MW 849.91. The lyophilized peptides were stored (in the presence of dessicant) at −40 degrees C prior to use. On the day of an experiment, an aliquot of lyophilized peptide was reconstituted in sterile saline at the indicated concentration. Reconstituted peptide(s) were prepared fresh before each experiment.

Animals

All procedures were conducted according to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of the Veterans Affairs Medical Center (East Orange, New Jersey). Male ZDF (fa/fa) and lean (+/?) Zucker rats were obtained from Charles River Laboratories (Kingston, NY) at approximately 6-7 weeks of age. All rats were single housed upon arrival, with modest enrichment (a PVC tube). Rats were provided ad libitum access to food and water and maintained in a 12 h light/dark cycle with lights on at 0630. All procedures occurred during the light phase of the circadian cycle and a timeline of procedures is shown in Figure 1. Results of the open field test (OFT) for anxiety-like behavior and sucrose preference test results (a measure of anhedonic behavior) are not reported here.

Figure 1: Timeline of treatments, exposures and behavioral assessments in Zucker rats

Injections

Peptide (SN..8 or LD.8) was dissolved in sterile saline (2 mg/kg) and SN.8 peptide vs. vehicle (sterile saline) or LD..8 peptide intraperitoneal (IP) injections were administered every other day according to the schedule shown in Figure 1. Peptide (SN.8) (vs. saline or LD..8) injections commenced 1-week prior to surgery/injury in ZDF rats to prevent (excess) acute post-injury mortality in ZDF rats thought to be related (in part) to pre-injury moderate hypertension. The peptide SN..8 was previously reported to substantially lower blood pressure in the ZDF rat [8].

Surgeries/Injury

All animals underwent a surgical procedure to attach a luer-lock connector to a craniectomy, positioned either on the left or right parietal bone plate. Rats were anesthetized with 1-4% isoflurane (5 L/min O₂ for 60 s) and transferred to a stereotactic instrument with 1-4% isoflurane mixed with 1-1.5 L/min O₂ delivered via a nosecone to maintain anesthesia throughout the surgery. A craniectomy (2 mm in diameter, centered at 3 mm posterior to bregma and 3.5 mm lateral to midline) was made in the skull and super glue was used to secure a luer-lock connector around the edge of the craniectomy. A plastic cylinder (cut from a 10 mL syringe) was placed on the skull surrounding the luer-lock connector; this was placed to provide stability and protect the area around the craniectomy. Dental cement was used to fill the space in between the connector and the plastic cylinder. A small amount of sterile saline was deposited into the connector along with a small piece of Kimwipe to keep the dura clean of debris.

One day following implantation of the surgical hub, fluid percussion injury is achieved with a device using a computer-controlled voice-coil to deliver defined pressure waves to the dura through a water piston [9]. Water is first placed in the surgical cap to prevent any air displacing the pressure waves. Rats were anesthetized with isoflurane in an induction chamber followed by fluid percussion injury or sham. The injury procedure occurs one day after surgery to prevent the dura from drying out and thereby reducing the injury of the brain tissue. Mild to moderate severity TBI will occur at pressure amplitudes of 20-30 psi (±3 psi) and rates of rise between 10-20 ms. To reduce the effects of anesthesia as much as possible, rats are connected to the device by the syringe hub and allowed to recover from anesthesia until they respond lightly to a foot pinch before injury occurs. Therefore, the rats were lightly anesthetized at the time of injury. Sham rats are treated the same way as injured rats including isoflurane anesthesia and connection of the device to the rat, but the pressure wave was not delivered. The animal is monitored for startle in response to the pressure wave and then placed on its back in a recovery cage and evaluated for apnea. Latency for the righting reflex to return is recorded and used as an index of traumatic unconsciousness.

Behavioral Tests

Morris Water Maze

The Morris water maze is a test of short-term recall of a spatial learning task. Zucker lean and fatty rats subjected to sham vs. traumatic brain injury were tested for the ability to locate a submerged platform in a swimming pool. Sample phase corresponds to acquisition of learning; the choice phase measures recall of newly-acquired spatial learning task. Distance is a measure of how far the rats swam before locating the submerged swim platform. Longer distance (choice phase) is indicative of impairment of recall of short-term spatial learning acquired in the sample phase. Path efficiency (efficiency) is the ratio of the shortest path to the platform vs. the actual swim path taken by the animal. Maximum path efficiency has a value of 1.0 and is a measure of recall of swim platform location.

Delayed Match to Position in Morris Water Maze

Before injury, the animals underwent extensive training in the Morris water maze apparatus. The delayed match to position (DMTP) test is a short-term recall of a spatial memory task. Rats were tested for the ability to locate a submerged platform in a swimming pool. The swimming pool was made opaque by non-toxic paint (Crayola washable paint, 16 oz, white 54-2016-053). Submerged in the pool was a platform on which the rat could escape from swimming, with the surface of the escape platform approximately 1 cm below the surface of the water. During the first acclimation session, the rat was initially placed on the platform and allowed to remain on the platform for 10 – 20 seconds. In subsequent training trials, the rat was placed further from the escape platform so that the rats learned that there was a place to escape from swimming, learned to climb onto the escape platform, and acclimated to swimming in the pool. During this training phase, the escape platform was in the same location and rats were started twice from 3 distinct locations (6 trials/session). Rats were trained for 4 days with one session per day. Each trial lasted a maximum of 60 seconds and if the rat had not located the platform during that time the rat was led to the platform using a wooden stick. The rat was then allowed to climb onto the platform where they remained for 10-20 seconds. After this, the rat was dried and moved to a holding cage with a heater if necessary.

One day after the 4th training session, a single probe trial was given in which the platform was removed from the pool and the rat was allowed to swim from one of the start locations for 60 seconds, after which the rat was removed from the pool, dried, and placed back in the home cage. For both training and probe sessions, the swim path of the rat was recorded and analyzed for multiple measures to assess spatial learning using AnyMaze (AnyMaze version 6.21 (64-bit) (Stoelting Co., 620 Wheat Lange, Wood Dale, IL 60191 USA).

Testing of spatial working memory commenced approximately 2-3 days after the probe trial using a DMTP procedure that consisted of 6 trials per rat with each trial having two phases, a sample phase and a choice phase. At the beginning of each trial, an animal entered sample phase by being placed in the water at a novel start position and has 60 seconds maximum swim time to find the platform. After either successfully finding the platform or being led to it after the 60 seconds swim time, the rat was left on the platform for 15 seconds, and then returned to its home cage for 30 seconds. That same animal then began choice phase, was placed back in the water at the same location as in sample phase and had 60 seconds maximum swim time to locate the platform. Each trial has a novel combination of start and platform positions. All rats were tested at the same start-platform positions. The pool was divided by AnyMaze into three equal-area zones, covering 120° of circumference each. At the beginning of the sample phase, each rat was placed in the water facing the perimeter of the tank in a selected zone. The escape platform was in a different zone than that of the start position and remained in that zone for both phases. The rat was allowed a maximum of 60 seconds to find the platform. After either successfully finding the platform or being led to it after 60 seconds, the rat was left on the platform for 10-20 seconds, and then returned to its home cage for approximately 30 seconds. The same animal then began the choice phase, where it was placed back in the water at the same start location as in sample phase and allowed 60 seconds to locate the platform. On each subsequent trial, the starting location and the platform location were changed across trials, not within trials, within the session. Approximately 1 hour separated each trial.

The swim path was recorded for later analysis. One to two days of working memory testing was required. Swim Distance was used as a measure of how far the rats swam before locating the submerged swim platform.

Statistics

Student’s t-test for single comparisons, two-way ANOVA for differences between two or more factors (strain, injury, drug) and their interaction, and repeated measures ANOVA for differences over time (pre-injury, 1-week, 1-month and 3-months post injury). The post hoc Bonferroni test was employed in all ANOVA. All statistical analyses were conducted using the SPSS software. A P-value <0.05 was considered significant and values are expressed as means ± SEM.

Results

Morris Water Maze – A Test of Spatial Memory

In fatty and lean Zucker rats exposed to TBI, distance (needed to locate swim platform) in the choice phase (CP) significantly exceeded that of sham-injured Zucker rats (Figure 2A and 2B) (F(1,12)=8.314, p=0.014) indicative of TBI-associated spatial memory impairment. In a repeated measures ANOVA, there was a trend of a significant time x injury interaction, F(3,36)=2.401, p=0.084, which did not reach statistical significance. Path efficiency was significantly lower in fatty vs. lean Zucker rats at all timepoints post-baseline (Figure 3) including TBI and sham-injured fatty rats. These data suggest factors associated with obesity, hypertension and type 2 diabetes likely contribute to accelerated decline in cognitive function (spatial memory recall) in fatty vs. age-matched lean Zucker rats.

Figure 2: Delayed match to position in the Morris water maze: distance (meters) as a function of time before and after injury in Zucker lean and fatty rats. Results are mean ± SEM. Distance was determined by Any Maze software as described in Methods. Fatty: sham (N=6), tbi (N=6) Lean: Sham (N=9), tbi (N=8).

Figure 3: Delayed match to position in the Morris water maze: path efficiency as a function of time before and after injury in Zucker lean and fatty rats. Results are mean ± SEM. Path efficiency was determined by Any Maze software as described in Methods. Fatty (N=12); Lean (N=17) rats.

Effect of Short-term Peptide Administration on Spatial Memory

Peptide (SCLLADDN=SN..8) was administered to fatty rats (every other day) for seven days before and after injury and in lean rats for seven days after injury. The fatty rats received three additional doses of SN..8 or saline (before injury) because of prior work showing that SN..8 significantly lowered blood pressure in ZDF rats and appeared cardioprotective [8]. Zucker diabetic fatty (vs. lean) rats experienced a disproportionate excess mortality associated with TBI which was reduced by pre-treatment with SN..8 (Grinberg M, Burton J, Pang K, Zimering MB, unpublished observations). Path efficiency difference is defined as [choice phase efficiency – sample phase efficiency] and is a measure of the strength of spatial memory recall. One-week post-injury, SN..8 peptide-treated, sham-injured fatty and lean Zucker rats displayed significantly larger gain in path efficiency difference compared to saline-treated, sham-injured fatty and lean rats (F(1,10)=5.777, p=0.037) (Figure 4). There was a significant (drug x injury) interaction F(2,20)=4.442, P=0.025. Post-hoc Bonferroni test showed significantly higher (one-week) path efficiency difference in Sham, SN..8-treated vs. Sham, Saline-treated Zucker rats (P_bonf=0.039) (Figure 4). Our sample size (N=29) was only powered to detect two-way, but not three-way interactions. There was no significant (strain x drug) interaction F(2,23)=0.584, P=0.566. There was no significant (strain x injury) interaction F(1,25)=0.328, P=0.572.

Figure 4: Delayed match to position in the Morris water maze: Path efficiency difference at 1-week post injury in Zucker fatty and lean rats. Path efficiency difference was calculated as described in Methods and is a measure of the strength of recall of spatial learning. Results are mean ± SEM. Sham (N=9); TBI (N=11). Sham, saline (N=5), Sham, P4 (N=4), TBI, saline (N=4), TBI, P4 (N=7). P4= SN..8 peptide.

Cohorts 1 and 2 combined sample size (N=29 rats) lacked sufficient power to detect statistically significant differences in behavior in peptide- vs. saline-treated rats analyzed at repeated intervals following injury. Next we tested three additional cohorts of Zucker fatty (N=22) and lean rats (N=12), total N=34 rats in the Morris water maze using an identical (2 mg/kg) concentration of scrambled peptide (LD..8) as control for SN..8 peptide treatment. Duration to locate the swim platform in the choice phase (CP) was an outcome measure of (recall of spatial learning), in a repeated measures ANOVA (preinjury, 1-week, and 1-month post-injury) which demonstrated a main effect of time, F(2, 56)=6.214, P=0.004 (Figure 5). There was a main effect of strain F(1,28)=7.797, P=0.009 with leans requiring less time than fatty Zucker rats to locate the platform. There was a main effect of injury F(1,28)=11.300, P=0.002 with sham-injured rats requiring less time than TBI rats to locate the platform (Figure 5). There was a significant time x injury interaction F(2,56)=9.584, P < 0.001, post-hoc: sham, 1 week x tbi, 1 week, P bonf <0.001 & sham, 1 month x tbi, 1 month, P bonf=.013.

Figure 5: Delayed match to position in the Morris water maze: duration in the choice phase CP (recall of spatial learning) as a function of time before and after injury in Zucker lean (N=12) and fatty rats (N=22). Results are mean ± SEM. Duration (seconds) was determined by Any Maze software as described in Methods.

Duration (to locate swim platform) in the Sample Phase (SP) was a measure of (spatial learning acquisition). A repeat measures ANOVA of duration in the sample phase demonstrated main effects of strain (F(1,28)=12.613, P=0.001 with leans having reduced duration compared to Zucker fatty rats; and injury F(1, 28)=9.335, P=0.005 with sham-injured rats having reduced duration compared to TBI (Figure 6A). There was a significant injury x drug interaction F(1,28)=7.788, p=0.009 with sham, SN..8-treated animals having shorter duration (to locate platform) than sham, scrambled-LD..8-treated rats (P bonf=0.001), (Figure 6B). There was no significant difference in post-hoc tests of TBI, SN..8 vs. TBI, LD..8-treated rats (Figure 6B). There was a significant (strain x drug) interaction F(1,28)=4.589; P=0.041 (Figure 6C). In post-hoc testing, fatty P4 had significantly higher SP duration than lean P4; P _bonf=0.002; and fatty, scrambled had significantly higher SP duration than lean P4; P_bonf=0.021 (Figure 6C). There was no significant (strain x injury) interaction F(1,25)=0.328; P=0.572. Sample size limitations prevented testing for a strain x drug x injury interaction.

Figure 6: Delayed match to position in the Morris water maze: duration in the sample phase SP (acquisition of spatial learning) as a function of time, A) and injury or B) peptide (P4= SN..8, scrambled = LD..8) treatment group or C) strain before and after injury in Zucker lean (N=12) and fatty rats (N=22). Results are mean ± SEM. Duration was determined by Any Maze software as described in Methods.

In vitro Neuroprotection by SN.8 in Mouse Neuroblastoma N2A cells

The SN..8 peptide dose-dependently inhibited Zucker lean heterozygote rat Ig-induced neurite retraction in vitro exhibiting an IC₅₀ of ~ 9 mg/L (Figure 7).

Figure 7: Dose-dependent inhibition of N2A neurite retraction from 100 nanomolar concentration of Zucker heterozygous lean by SN..8. Neurite retraction assay were performed as described in Methods. Results are mean ± SEM.

Mechanism of Action of SCLLADDN (SN….8) Peptide

The precise mechanism of SN..8 peptide’s in vitro and in vivo neuroprotective action is unknown. We had proposed that the epitope-specific SN..8 peptide may act as a ‘decoy’ which directly binds 5-HT2AR agonist IgG preventing receptor activation. Although we can’t exclude this possibility, we now provide evidence for an additional possible mechanism. Using linear synthetic peptides containing an alanine substitution for a highly conserved cysteine at EL2.50 (Ballesteros-Weinstein residue numbering system [10]) SALLADDN or for an aspartic acid residue at position EL2.55 (SCLLADAN) we compared the in vitro neuroprotective effects (Gq11/IP-mediated neurite retraction) of mutant vs. ‘wild-type’ SCLLADDN peptide. A 50-100 nanomolar concentration of Zucker heterozygous lean rat IgG caused 50% acute N2A neurite retraction which was completely prevented by co-incubation with twenty mg/mL concentration of wild-type SN..8 (Figure 8). Pre-incubation of IgG with twenty mg/mL concentration of either ‘C to A’ or ‘D to A’ mutant peptides had no inhibitory effect on Zucker rat IgG-induced neurite retraction. Although (20 ug/mL) wild-type peptide alone had no effect on N2A neurite retraction, we found unexpectedly that incubation of N2A cells with a twenty mg/mL concentration of either ‘C to A’ or ‘D to A’ mutant peptides alone caused transient, reversible (after 5 minutes), Gq11/IP-mediated N2A neurite retraction (not shown in Figure 8).

Figure 8: Effect of targeted ‘C to A’ or ‘D to A’ amino acid substitutions in ‘wild-type’ SCLLADDN peptide on the resulting mutant peptides’ (SALLADDN) or (SCLLADAN) ability to prevent acute N2A neurite retraction in the presence of 50-100 nanomolar concentration of Zucker heterozygous lean rat IgG. Results are mean ± SEM of three experiments. Similar results were obtained in experiments using the IgG fraction of plasma from four different middle-aged human TBI patients. Neurite retraction assay in mouse N2A neuroblastoma cells was carried out as described in Methods.

The wild-type SCLLADDN (SN..8) is identical to a subregion of the second extracellular loop (ECL) of the human serotonin 2A receptor reported by Wacker et al [11] to function as a ‘lid’ modulating the ingress and egress of ligands into and out of the 5HT2(B, or A) receptor’s orthosteric binding pocket (OBP). For example, closure of the lid prevented egress of the hallucinogenic ligand lysergic acid diethylamine (LSD) from the OBP causing an unusually long off-reaction time of LSD at the 5HT2B and 5HT2A receptors [11]. Mobility of the lid peptide is dependent on its specific amino acid residues which in turn determines the kinetics of the open and closed receptor conformations [11]. The lid function is conserved among many different aminergic GPCRs [12] and a highly conserved cysteine residue EL2.50 involved in intrachain disulfide bonding normally prevents constitutive aminergic GPCR activation [12]. The observation that C to A (SCLLADDN to SALLADDN) or D to A (SCLLADDN to SCLLADAN) single amino acid residue substitution(s) each led to transient constitutive receptor Gq11/IP activation (in the presence of serotonin in the culture medium) is of interest since it suggests a possible direct effect of exogenously administered short lid peptides on stabilizing an active vs. inactive conformation of the receptor. We speculate that cysteine-containing ‘wild-type’ SN..8 (SCLLADDN) may spontaneously form aggregates important in antigen-antibody binding or which interfere with hydrogen bonding between GPCR transmembrane helices [13] required for receptor activation. The Asp at position EL 2.54, shown in bold SCLLADDN in the wild-type peptide, is also highly conserved among family A GPCRs [13] and it mediates hydrogen bonding between helix 2 and helix 7 underlying receptor activation [13]. Replacement of the adjacent Asp (EL2.55) with Ala to form SCLLADAN may have led to a more highly mobile lid peptide which mediated transient receptor activation by increasing the probability of hydrogen bonding between helices 2 and 7. To our knowledge, these are the first data suggesting that small peptide mimics of an aminergic GPCR receptor ‘lid’ region in which key amino acid residues (Cys EL2.50 or Asp 2.55) have been replaced by alanine not only mediate transient receptor activation, but also abrogate neuroprotection associated with ‘wild type’ SN…8 lid peptide.

Discussion

Obese type 2 diabetes mellitus and hypertension are associated with accelerated age-related cognitive decline in older adults [14], and in a recent study of lifelong TBI-sufferers, (obesity and hypertension) were significant predictors of an increased hazard rate for the occurrence of major depressive disorder, Parkinson’s disease or dementia [3]. The Zucker fatty rat is a widely-used genetic model of morbid obesity [4]. Morbid obesity promotes peripheral inflammation associated with increased risk for neuropsychiatric and neurodegenerative disorders in humans [15], yet there have been relatively few prior neurobehavioral studies in the Zucker fatty rat strain which might shed light on underlying mechanisms.

Here we report that the Zucker diabetic hypertensive fatty rat experienced significantly greater spontaneous decline in spatial memory compared to age-matched lean Zucker rats.

It is not clear to what extent diabetes or hypertension (in Zucker fatty rats) may have affected the rate of decline in recall of spatial memory. In human studies, older adult type 2 diabetes populations experienced increased rate(s) of decline in executive function and processing speed; and hypertension, hypertriglyceridemia, and diabetes duration, (but not glucose level per se) was each a significant predictor of accelerated cognitive decline [14]. Severe hypertriglyceridemia which paralleled the development of morbid obesity in Zucker fatty rat may have contributed in part to observed sub-strain behavioral differences. In a prior report, male albino rats on a high-fat (vs. normal diet) developed obesity, and elevated lipid profile and had worse performance (on spatial memory task) in the Morris water maze [16] suggesting a role for metabolic factors in the Zucker fatty rat spatial memory impairment.

It is unclear how the SN..8 serotonin 2A receptor peptide fragment mediated its apparent in vivo neuroprotective effects, but one possibility may involve hippocampal neurogenesis. Neurogenesis occurs in the dentate gyrus (DG) region of the hippocampus in adult mammals, and it affects acquisition and recall of spatial learning [17]. Hippocampal neural progenitor cells (NPC) develop in a unique vascular niche [18] exposed to the general circulation. Circulating serotonin 2A receptor modulatory Ig (from morbidly-obese, and adult diabetic patients and in Zucker rats) not only had anti-endothelial effects [19] which could adversely impact neurogenesis, but also adversely affected the maturation, survival and electrical excitability of rat dentate gyrus NPC in vitro [20]. Serotonergic receptors expressed in the hippocampus have a well-established role in the regulation of mood and emotional disorders [21]. In addition, systemic administration of 5-HT2A receptor agonists (psilocin, or TCB-2) in rodents significantly impaired spatial memory recall in the Morris water maze (MWM) [22]. Taken together, it is possible that SN..8 peptide administration could act at the level of the dentate gyrus region of the hippocampus in modulating neurogenesis important in both acquisition and recall of spatial memory.

Another major hormone system which influences dentate gyrus neurogenesis are glucocorticoid (GR) and Mineralocorticoid Receptors (MR). Both receptors are highly expressed in hippocampal brain regions [23] and participate in the complex regulation of the effects of stress on the hypothalamic/pituitary/adrenal axis. It is possible that stress associated with mild TBI (LFP) causes greater GR activation in the hippocampus leading to stronger suppression of neurogenesis compared to sham-injury. An SN.8 peptide which targets the serotonin 2AR may have little effect on GR-induced suppression of neurogenesis. A future study of serotonergic and glucocorticoid receptor expression in Zucker rat brain following sham vs. TBI is needed to clarify the underlying mechanisms. A possible role for DG neurogenesis in mediating the putative neuroprotective effects of a serotonin 2A receptor peptide on acquisition and recall of spatial learning also requires more direct study.

Our data that a 5HT2AR-specific lid peptide SCLLADDN (SN..8) significantly enhanced acquisition of spatial memory and strength of spatial memory recall (1-week post-injury in sham-injured rats) is consistent with the possibility that the receptor peptide may prevent 5HT2A receptor activation by endogenous 5-HT2AR ligand agonists including possibly circulating 5HT2AR agonist IgG.

Taken together, these data suggest that a novel serotonin 2A receptor peptide may have neuroprotective effects in sham-injured Zucker rats under conditions in which hippocampal neurogenesis is not already strongly suppressed by other hormonal factors principally, increased hippocampal GR and possibly MR activity in obese, hypertensive Zucker rats [24,25].

In summary, systemic administration of a small peptide mimic of the 5HT2A receptor lid peptide appeared to confer neuroprotective effects on acquisition and strengthening of recall of spatial learning tasks in sham-injured Zucker rats. Larger studies are needed to confirm these preliminary results and further test for any possible strain differences. Injections of both SN..8 and LD..8 peptides were well-tolerated and not associated with chronic pain, or local tissue injury. TBI was associated with greater impairments in spatial memory recall, and the neuroprotective peptide appeared less likely to significantly modulate the behavioral impairment(s) in brain-injured Zucker rats, especially fatty rats susceptible to the additional adverse cognitive effects of abnormal metabolic factors.

Acknowledgements

This work was supported in part by a grant CBIR 22 PIL022 from the New Jersey Commission on Brain Injury Research (Trenton, NJ) to MBZ, and by grants from the Technology Transfer Program/BLRD, Office of Research and Development, Department of Veterans Affairs (Washington, DC) to MBZ. The opinions expressed herein are solely those of the authors and do not reflect the official position of the US Government.

References

1. Walker KR, Tesco G (2013) Molecular mechanisms of cognitive dysfunction following traumatic brain injury. Front Aging Neurosci 5: 29. [crossref]
2. Jaacks LM, Vandevijvere S, Pan A, McGowan CJ, Wallace C, et al. (2019) The obesity transition: stages of the global epidemic. The lancet. Diabetes & endocrinology 7: 231-240. [crossref]
3. Zimering MB, Patel D, Bahn G (2019) Type 2 Diabetes Predicts Increased Risk of Neurodegenerative Complications in Veterans Suffering Traumatic Brain Injury. J Endocrinol Diabetes 6: 137. [crossref]
4. Kurtz TW, Morris RC, Pershadsingh HA (1989) The Zucker fatty rat as a genetic model of obesity and hypertension. Hypertension 3: 896-901. [crossref]
5. Shibata T, Takeuchi S, Yokota S, Kakimoto K, Yonemori F, et al. (2000) Effects of peroxisome proliferator-activated receptor-alpha and -gamma agonist, JTT-501, on diabetic complications in Zucker diabetic fatty rats. Br J Pharmacol 130: 495-504. [crossref]
6. Zimering MB, Grinberg M, Burton J, Pang K (2020) Circulating Agonist Autoantibody to 5-Hydroxytryptamine 2A Receptor in Lean and Diabetic Fatty Zucker Rat Strains. Endocrinol Diabetes Metab J 4: 413. [crossref]
7. Zimering MB (2019) Autoantibodies in Type-2 Diabetes having Neurovascular Complications Bind to the Second Extracellular Loop of the 5-Hydroxytryptamine 2A Receptor. Endocrinol Diabetes Metab J 3: 118. [crossref]
8. Zimering MB (2021) A serotonin 2A receptor decoy peptide potently lowers blood pressure in male Zucker diabetic fatty rats. Endo Diab Metab J 5: 1-13. [crossref]
9. Wahab RA, Neuberger EJ, Lyeth BG, Santhakumar V, Pfister BJ (2015) Fluid percussion injury device for the precise control of injury parameters. J Neurosci Methods 248: 16-26. [crossref]
10. Ballesteros JA, Weinstein H (1995) Integrated methods for the construction of three-dimensional models and computational probing of structure-function relations in G protein-coupled receptors. Methods Neurosci 25: 366-428.
11. Wacker D, Wang S, McCorvy JD, Betz RM, Venkatakrishnan AJ, et al. (2017) Crystal Structure of an LSD-Bound Human Serotonin Receptor. Cell 168: 377-389.e12. [crossref]
12. Michino M, Beuming T, Donthamsetti P, Newman AH, Javitch JA, et al. (2015) What can crystal structures of aminergic receptors tell us about designing subtype-selective ligands?. Pharmacol Rev 67: 198-213. [crossref]
13. Sanchez-Reyes OB, Cooke ALG, Tranter DB, Rashid D, Eilers M, et al. (2017) G Protein-Coupled Receptors Contain Two Conserved Packing Clusters. Biophys J 112: 2315-2326. [crossref]
14. Zimering MB, Knight J, Ge L, Bahn G (2016) VADT Investigators. Predictors of Cognitive Decline in Older Adult Type 2 Diabetes from the Veterans Affairs Diabetes Trial. Front Endocrinol (Lausanne) 7: 123. [crossref]
15. Castanon N, Lasselin J, Capuron L (2014) Neuropsychiatric comorbidity in obesity: role of inflammatory processes. Front Endocrinol (Lausanne) 5: 74. [crossref]
16. Wahid RM, Samy W, El-Sayed SF (2021) Cognitive impairment in obese rat model: role of glial cells. Int J Obes (Lond) 45: 2191-2196.
17. Dupret D, Revest JM, Koehl M, Ichas F, De Giorgi F, et al. (2008) Spatial relational memory requires hippocampal adult neurogenesis. PLoS One 3: e1959. [crossref]
18. Palmer TD, Willhoite AR, Gage FH (2000) Vascular niche for adult hippocampal neurogenesis. J Comp Neurol 425: 479-494. [crossref]
19. Zimering MB, Alder J, Pan Z, Donnelly RJ (2011) Anti-endothelial and anti-neuronal effects from auto-antibodies in subsets of adult diabetes having a cluster of microvascular complications. Diabetes Res Clin Pract 93: 95-105. [crossref]
20. Zimering MB, Behnke JA, Thakker-Varia S, Alder J (2015) Autoantibodies in Human Diabetic Depression Inhibit Adult Neural Progenitor Cells In vitro and Induce Depressive-Like Behavior in Rodents. J Endocrinol Diabetes 2. [crossref]
21. Bohn LM, Schmid CL (2010) Serotonin receptor signaling and regulation via β -arrestins. Crit Rev Biochem Mol Biol 45: 555-566. [crossref]
22. Zhang G, Stackman RW Jr (2015) The role of serotonin 5-HT2A receptors in memory and cognition. Front Pharmacol 6: 225. [crossref]
23. Van Eekelen JA, Jiang W, De Kloet ER, Bohn MC (1988) Distribution of the mineralocorticoid and the glucocorticoid receptor mRNAs in the rat hippocampus. J Neurosci Res 21: 88-94. [crossref]
24. Wong EY, Herbert J (2005) Roles of mineralocorticoid and glucocorticoid receptors in the regulation of progenitor proliferation in the adult hippocampus. Eur J Neurosci 22: 785-792. [crossref]
25. Fredersdorf S, Endemann DH, Luchner A, Heitzmann D, Ulucan C, et al. (2009) Increased aldosterone levels in a model of type 2 diabetes mellitus. Exp Clin Endocrinol Diabetes 117: 15-20. [crossref]

DOI: 10.31038/EDMJ.2023724

Introduction

Anxiety and depression increase substantially after traumatic brain injury (TBI) in humans. They are among the most common early neuropsychiatric complications of TBI [1] contributing to substantial morbidity and mortality. Owing to the current global epidemic of obesity [2], a possible role for obesity in the development of anxiety disorders has received increased attention [3]. In a recent study of TBI in humans, obesity was a significant predictor of an increased risk for the later occurrence of a composite neurodegeneration endpoint comprised of major depression and suicide, Parkinson’s disease or dementia [4]. Yet few prior studies in have explored the relationship between obesity, TBI and anxiety disorders. Here we used the Zucker fatty rat strain, a genetic model of obesity harboring homozygous mutations in the leptin receptor gene (fa/fa) [5] and Zucker lean (Fa/fa/?) rats to test a commonly-held notion that obesity is a risk factor for anxiety-like disorders. We also tested whether mild traumatic injury (induced by lateral fluid percussion-LFP) (vs. sham injury) modifies anxiety-like behavior in obese vs. lean Zucker rats.

Obesity in rats age (8-12 weeks old) was recently reported to be associated with heightened anxiety-like behavior compared to age-matched obese-resistant genetic strains [6]. Since anxiety-like behavior was not correlated with body weight (in the study) the authors concluded that acquired (vs. genetic factors) may play a less important role in the etiology of anxiety [6]. In the current study, we tested rats for anxiety-like behavior at different times in the developmental cycle to further evaluate a possible role for chronic acquired factors (related to obesity and diabetes) in anxiety-like behavior. Current drug therapy for anxiety disorders modulates serotoninergic signaling at central synapses in the cerebral cortex, hippocampus and other brain regions. We also conducted an exploratory test of whether systemic administration of a novel, putative neuroprotective synthetic peptide fragment of the human serotonin 2A receptor (SN..8) may modulate anxiety-like behavior in Zucker fatty (vs. lean) rats exposed to mild TBI (vs. sham injury).

Materials and Methods

Peptides

A linear synthetic peptide, SCLLADDN (SN..8) having a sequence identical to that of a fragment of the second extracellular loop region of the human 5-hydroxytryptamine 2A receptor was synthesized at Lifetein, Inc (Hillsborough, NJ) and had > 95% purity. A scrambled peptide containing the same amino acids as SCLLADDN had a sequence of LASNDCLD (LD.8) and a purity of 96.37%, MW 849.91. The lyophilized peptides were stored (in the presence of dessicant) at −40 degrees C prior to use. On the day of an experiment, an aliquot of lyophilized peptide was reconstituted in sterile saline at the indicated concentration. Reconstituted peptide(s) were prepared fresh before each experiment.

Animals

All procedures were conducted according to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of the Veterans Affairs Medical Center (East Orange, New Jersey). Male ZDF (fa/fa) and lean (+/?) Zucker rats were obtained from Charles River Laboratories (Kingston, NY) at approximately 6-7 weeks of age. All rats were single housed upon arrival, with modest enrichment (a PVC tube). Rats were provided ad libitum access to food and water and maintained in a 12 h light/dark cycle with lights on at 0630. All procedures occurred during the light phase of the circadian cycle.

Injections

Peptide (SN..8 or LD.8) was dissolved in sterile saline (2 mg/kg) and SN.8 peptide vs. vehicle (sterile saline) or LD..8 peptide intraperitoneal (IP) injections were ad-ministered every other day. Peptide (SN.8) (vs. saline or LD..8) injections commenced 1-week prior to surgery/injury in ZDF rats to minimize the risk of excess acute post-injury mortality in ZDF rats thought to be related (in part) to pre-injury moderate hypertension. The peptide SN..8 was previously reported to substantially lower blood pressure in the ZDF rat [7].

Surgeries/Injury

All animals underwent a surgical procedure to attach a luer-lock connector to a craniectomy, positioned either on the left or right parietal bone plate. Rats were anesthetized with 1-4% isoflurane (5 L/min O2 for 60s) and transferred to a stereotactic instrument with 1-4% isoflurane mixed with 1-1.5 L/min O2 delivered via a nosecone to maintain anesthesia throughout the surgery. A craniectomy (2mm in diameter, centered at 3mm posterior to bregma and 3.5mm lateral to midline) was made in the skull and super glue was used to secure a luer-lock connector around the edge of the craniectomy. A plastic cylinder (cut from a 10mL syringe) was placed on the skull surrounding the luer-lock connector; this was placed to provide stability and protect the area around the craniectomy. Dental cement was used to fill the space in between the connector and the plastic cylinder. A small amount of sterile saline was deposited into the connector along with a small piece of Kimwipe to keep the dura clean of debris.

One day following implantation of the surgical hub, fluid percussion injury is achieved with a device using a computer-controlled voice-coil to deliver defined pressure waves to the dura through a water piston [8]. Water is first placed in the surgical cap to prevent any air displacing the pressure waves. Rats were anesthetized with isoflurane in an induction chamber followed by fluid percussion injury or sham. The injury procedure occurs one day after surgery to prevent the dura from drying out and thereby reducing the injury of the brain tissue. Mild to moderate severity TBI will occur at pressure amplitudes of 20-30 psi (±3 psi) and rates of rise between 10-20 milliseconds. To reduce the effects of anesthesia as much as possible, rats are connected to the device by the syringe hub and allowed to recover from anesthesia until they respond lightly to a foot pinch before injury occurs. Therefore, the rats were lightly anesthetized at the time of injury. Sham rats are treated the same way as injured rats including isoflurane anesthesia and connection of the device to the rat, but the pressure wave was not delivered. The animal is monitored for startle in response to the pressure wave and then placed on its back in a recovery cage and evaluated for apnea. Latency for the righting reflex to return is recorded and used as an index of traumatic unconsciousness.

Behavioral Tests

Open Field Test. Rats were placed in the center of a round, gray, open field apparatus (diameter of 74 cm, height of 51 cm). A light (120 W) was located 135 cm directly above the center of the apparatus which had a measured light intensity, LT300-NIST Light Meter (Extech Instruments, Knoxville, TN) of 400-500 lux. The open field apparatus was placed in a novel environment. The path of the rat in the open field was recorded, digitized, analyzed, and stored on a computer using AnyMaze version 6.21 (64-bit) (Stoelting Co., 620 Wheat Lange, Wood Dale, IL 60191 USA). For analysis, the open field was separated into two zones, the center, and the periphery. The center zone was measured to have the area of the total field (diameter of approximately 52.32 cm). Performance in the open field was scored during a 3-minute time window before returning the animal to its home cage. The apparatus was wiped with 70% ethanol solution between testing of each rat. Increased mobility in the center zone, defined as distance, speed, and time in motion, was used as an indicator of reduced anxiety, whereas staying in the peripheral zone, was used as an indicator of increased anxiety [9].

Statistical Analysis

Statistical analysis was performed using the Student’s t-test for single comparisons and one-way ANOVA for differences between more than two groups followed by post hoc Bonferroni tests. All statistical analyses were conducted using the SPSS software. A p-value <0.05 was considered significant and values are expressed as means ± SEM.

Results

Open Field Test – Bright Light

Twenty-two rats were initially tested 1.5 months post-injury (i.e. at 18=19 weeks of age). An ANOVA demonstrated a main effect of strain, P=0.037 (Figure 1). Sham-injured Zucker fatty rats spent more time in center (mean 82 ± 10 sec) compared to sham-injured Zucker lean rats (mean 55 ± 7 sec) (Figure 1A and 1B). The preliminary study of anxiety-like behavior was only powered to test for main effects. Still it suggested that Zucker fatty rats display reduced anxiety-like behavior when tested later in the developmental cycle.

Figure 1: Center time (sec) in bright light in A) lean or B) fatty Zucker rats after sham- vs. mild TBI. A) Lean: Sham (N=7), TBI (N=6). B) Fatty: Sham (N=5), TBI (N=4).

Open Field Test – Under Dark Conditions (Mobility Test)

General mobility was assessed in the open field test performed under dark conditions. There were no significant differences in time in the center by strain (fatty vs. lean) or injury (sham vs. TBI) (Figure 2). These results suggest that the differences observed under bright light conditions were likely attributable to anxiety-like behavior.

Figure 2: Open field test under dark conditions, 1 month post-injury in Zucker lean and fatty rats. Results are mean ± SEM. Lean: Sham (N=7), TBI (N=6). Fatty: Sham (N=5), TBI (N=4).

Next, in a different cohort of Zucker fatty (N=22) and lean (N=12) rats, the open field test (under bright light) was performed before injury (11-12 weeks of age), 1-week after injury (15-16 weeks of age) and 1-month after injury (18-19 weeks of age). In a repeated measures ANOVA of center time at (preinjury, 1week and 1 month post-injury), there was a significant main effect of strain (F(1,26)=5.639, p=0.025) and a significant interaction effect of (time x strain) F(2,52)=8.248, p<0.001 (Figure 3). The post-hoc tests showed significant differences between fatty, preinjury vs. fatty, 1week post-injury (P_bonf=0.002) and between fatty, 1week post-injury vs. lean, 1week post-injury (P_bonf=0.023) (Figure 3). The fatty rats spent significantly more time in the center zone after injury compared to preinjury, and spent significantly more time in the center zone after injury than leans rats (Figure 3). There was no significant (strain x injury) interaction.

Figure 3: Center time: significant (strain x time) interaction. Each point is mean ± SEM

Mean difference (fatty vs. lean Zucker rats) in center time (Figure 4) underwent a ‘reversal’ after 11-12 weeks and before 15-16 weeks of age. At the earlier timepoint fatty rats had higher anxiety and later underwent a shift to reduced anxiety-like behavior (vs. lean rats). Zucker fatty rats acquired body weight significantly more rapidly than lean rats between 7-12.5 weeks of age (Figure 5), however injury (sham or TBI) which occurred at 13 weeks of age (Figure 6, arrowhead) contributed to a temporary ‘fall off’ in the normal trajectory of weight gain in Zucker fatty rats (Figure 5). Severe hyperglycemia in Zucker fatty rats (16 weeks and older) may have also contributed to a slowing in the relative trajectory of weight gain in fatty vs. lean (normoglycemic) Zucker rats (Figure 5).

Figure 4: Developmental ‘onset’ of reduced anxiety-like behavior in fatty vs. lean Zucker rats. Data are mean ± SEM.

Figure 5: Change in body weight in Zucker lean and fatty rats across development and injury (arrowhead). Arrows signify pre-injury, 1-week post and 1-month post-injury timepoints. Each point is mean ± SEM.

Overall mean speed, and exploratory behavior (distance traveled) were unexpectedly significantly higher in fatty vs. lean Zucker rats (Figure 6A and 6B) and immobility time was significantly reduced in fatty vs. lean Zucker rats (Figure 6C), assessed 1-week after injury. Mean distance traveled (Figure 7A) was significantly higher in Zucker fatty (vs. lean rats), before and 1-week post injury. Total immobility time was significantly reduced, before injury, 1 week, and 1 month post-injury, in Zucker fatty vs. lean rats (Figure 7B). Taken together, center time in the open-field test under bright light conditions (a measure of anxiety-like behavior) was unexpectedly decreased in Zucker fatty (vs. lean) rats at 15-16 weeks of age and older and could not be accounted for by strain differences in general mobility, or exploratory behavior.

Figure 6: One week after injury: A) Mean speed and B) distance traveled were both significantly higher in fatty vs. lean Zucker rats; C) total time immobile was significantly higher in lean vs. fatty Zucker rats.

Figure 7: A) Zucker fatty (vs. lean) rats traveled significantly greater distance before and 1-week after injury B) Zucker lean (vs. fatty) rats had significantly greater immobility at all three timepoints before and after injury. * P < 0.05.

A novel small peptide medication (SN..8) comprised of a fragment of the second extracellular loop of the human 5HT2AR prevented 5HT2AR activation on mouse neuroblastoma cells in vitro. Unlike direct antagonist 5HT2AR medications which promote significant weight gain, chronic administration (for 13 weeks or longer) of SN..8 (vs. an inactive scrambled version of the peptide – LD..8) did not cause significant weight gain in Zucker rats [7]. We next tested whether systemic (intraperitoneal) administration of SN..8 (vs. LD.8) may affect anxiety-like behavior in (sham-injury vs. TBI) Zucker rats.

There were no significant differences in center time between fatty vs. lean Zucker rats subjected to (sham- vs. TBI injury) or (SN..8 vs. scrambled peptide) treatment, i.e. 2 mg/kg, IP every other day when assessed either 1 -week after injury (Figure 8) or 1-month after injury (Figure 9). In repeated measures ANOVA of center time, there was no significant (drug x injury) or (drug x strain) interaction (data not shown). Still, the sample sizes were small and had reduced power to detect possible statistically significant difference(s) in drug (SN.. 8 vs. LD.8) or injury (sham vs. TBI) effects on anxiety-like behavior.

Figure 8: SN..8 (vs. scrambled peptide) treatment did not significantly modify time in center in lean or fatty rats evaluated a A) 1 week after sham injury or B) 1 week after mild TBI. Data are mean ± SEM.

Figure 9: SN..8 (vs. scrambled peptide) treatment did not significantly modify time in center in lean or fatty rats evaluated a A) 1 month after sham injury or B) 1 month after mild TBI. Data are mean ± SEM.

Discussion

Obesity has reached epidemic proportions in the United States and other parts of the world [2]. Anxiety and depressive disorders rank very high among conditions contributing to morbidity and mortality. Understanding biological links between obesity and anxiety or depressive disorders is of considerable public health importance. Among lifelong TBI-sufferers, obesity caused in part by a sedentary lifestyle and medications useful in treatment-refractory depression is both common and appeared to increase the risk for later occurrence of a composite neurodegenerative disease outcome including severe depression [4]. In a prior study that compared rats fed a standard diet vs. high-fat Western diet (leading to obesity), obesity was reported to increased anxiety-like behavior in the high-fat diet fed rats [10]. One proposed mechanism linking obesity to anxiety is increased inflammation [3]. In a different study that compared obese-prone vs. obese-resistant rat strains at 8-12 weeks of age, the Zucker fatty rat had lower exploratory activity, lower general mobility and decreased center time in the open field test (i.e. increased anxiety-like behavior) compared to several different obese-resistant rat strains, but not including Zucker lean rats [6]. In the present study, we compared obese and lean Zucker rats having similar overall genetic background, but differing at the leptin receptor locus (i.e. fa/fa vs. Fa/?). At 8-12 weeks of age, anxiety-like behavior was significant increased in fatty vs. lean Zucker rats perhaps consistent with prior reports of a role for inflammation and obesity [6,10]. However, when the open field test was conducted at later developmental age(s) in the present study, anxiety-like behavior was significantly reduced in fatty vs. Zucker rats. This may be indicative of development change in one or more receptors in the Zucker fatty rat brain important in mediating anxiety-like behavior.

The serotonin 2A receptor and the serotonin 1A receptor play opposing roles in the regulation of anxiety and depression [11]. Increased activity in the 5HT2AR in certain brain regions is associated with increased anxiety and depression [12] and several existing FDA-approved 5HT2AR antagonist medications are effective in treatment-resistant depression. On the other hand, sustained activity in the 5HT1A receptor underlies (in part) the anxiolytic and anti-depressant effect of SSRI medications [13].

Both 5-HT2AR and mineralocorticoid receptor (MR) are expressed in cortical brain regions and have a role in anxiety-like behavior. For example, cortical expression of 5HT2A, -B, -C receptors was identical in ‘high anxiety’ Lewis vs. ‘low anxiety’ SHR rat strains [14], however, cortical Gq11/IP accumulation (via 5HT2A, B or C receptors) was substantially higher in the ‘high anxiety’ Lewis rat [15]. Rozeboom et al. [16] reported that transgenic mice harboring chronic increased forebrain mineralocorticoid receptor (MR) expression displayed reduced anxiety-like behavior (vs. wild-type mice). Forebrain MR overexpression led to increased CA1 hippocampal expression of the 5HT1A receptor (important in mediating anxiolysis) and reduced hippocampal expression of the glucocorticoid receptor (important in mediating the stress response). Zucker fatty (vs. lean) rat were reported to have two-fold higher level of plasma aldosterone [17] (at 25 weeks of age) owing in part to obesity-associated hypertension, and to activation of the renin-angiotensin-aldosterone as a result of severe insulin resistance, hyperglycemia and insulin deficiency [18]. Taken together, morbid obesity- and diabetes-associated hormonal changes may cause increased brain MR activity (in the Zucker fatty rat) resulting in increased hippocampal expression of the 5HT1A receptor and decreased glucocorticoid expression – both mediating anxiolysis. Future study can directly test for changes in the hippocampal expression level of the 5HT1A and glucocorticoid receptor in Zucker fatty vs. lean rat brain, and its developmental onset and possible modulation by sham-injury vs. TBI.

A limitation of the present study is that subgroups of rats treated with SN..8 (vs. LD..8) and subjected to sham-injury vs. TBI may have been too small in their number(s) to detect significant differences in anxiety-like behavior. Zucker diabetic fatty rats (ZDF) spontaneously harbor 5-HT2AR agonist autoantibodies causing Gq11/IP3 accumulation [19] and the ZDF Ig mediated neurotoxicity in vitro was nearly completely prevented by incubation (of neuroblastoma cells) with SN..8 [19]. If the circulating 5-HT2AR agonist Ig were able to access (anxiogenic) cortical 5HT2A receptors, e.g. following mild TBI and disruption of the blood brain barrier, SN..8 might prevent Ig-induced 5HT2AR activation. A larger study is needed to test whether preventing cortical 5HT2AR activation by Ig, with SN..8 (vs. LD..8) may have an anxiolytic effect in lean vs. fatty Zucker rats.

Acknowledgements

This work was supported in part by a grant CBIR 22 PIL022 from the New Jersey Commission on Brain Injury Research (Trenton, NJ) to MBZ, and by grants from the Technology Transfer Program/BLRD, Office of Research and Development, Department of Veterans Affairs (Washington, DC) to MBZ. The opinions expressed herein are solely those of the authors and do not reflect the official position of the US Government.

References

1. Howlett JR, Nelson LD, Stein MB (2022) Mental health consequences of traumatic brain injury. Biol Psychiatry 91: 413-422. [crossref]
2. Jaacks LM, Vandevijvere S, Pan A, McGowan CJ, Wallace C, et al. (2019) The obesity transition: stages of the global epidemic. The lancet. Diabetes & endocrinology 7: 231-240. [crossref]
3. Castanon, Nathalie, Julie Lasselin, Lucile Capuron (2014) Neuropsychiatric comorbidity in obesity: role of inflammatory processes. Frontiers in endocrinology 5: 74. [crossref]
4. Zimering MB, Patel D, Bahn G (2019) Type 2 Diabetes Predicts Increased Risk of Neurodegenerative Complications in Veterans Suffering Traumatic Brain Injury. J Endocrinol Diabetes 6: 137. [crossref]
5. Kurtz TW, Morris RC, Pershadsingh HA (1989) The Zucker fatty rat as a genetic model of obesity and hypertension. Hypertension 3: 896-901. [crossref]
6. Vogel H, Kraemer M, Rabasa C, Askevik K, Adan RAH, et al. (2017) Genetic predisposition to obesity affects behavioural traits including food reward and anxiety-like behaviour in rats. Behav Brain Res 328: 95-104. [crossref]
7. Zimering MB (2021) A serotonin 2A receptor decoy peptide potently lowers blood pressure in male Zucker diabetic fatty rats. Endo Diab Metab J 5: 1-13. [crossref]
8. Wahab RA, Neuberger EJ, Lyeth BG, Santhakumar V, Pfister BJ (2015) Fluid percussion injury device for the precise control of injury parameters. J Neurosci Methods 248: 16-26. [crossref]
9. Walsh RN, Cummins RA (1976) The open-field test: A critical review. Psychological Bulletin 83: 482-504. [crossref]
10. André C, Dinel AL, Ferreira G, Layé S, Castanon N (2014) Diet-induced obesity progressively alters cognition, anxiety-like behavior and lipopolysaccharide-induced depressive-like behavior: focus on brain indoleamine 2,3-dioxygenase activation. Brain Behav Immun 41: 10-21. [crossref]
11. Kandel Eric R, James H Schwartz, Thomas M Jessell, Steven Siegelbaum, A James Hudspeth, et al. (2000) Principles of neural science 4.
12. Celada P, Puig M, Amargós-Bosch M, Adell A, Artigas F (2004) The therapeutic role of 5-HT1A and 5-HT2A receptors in depression. J Psychiatry Neurosci 29: 252-265. [crossref]
13. Gordon JA, Hen R (2004) The serotonergic system and anxiety. Neuromolecular Med 5: 27-40. [crossref]
14. Berton O, Aguerre S, Sarrieau A, Mormede P, Chaouloff F (1998) Differential effects of social stress on central serotonergic activity and emotional reactivity in Lewis and spontaneously hypertensive rats. Neuroscience 82: 147-159. [crossref]
15. Gauffre JC, Aguerre S, Mormède P, Chaouloff F (1997) Cortical [3H] ketanserin binding and 5-HT2A receptor-mediated inositol phosphate production in the spontaneously hypertensive rat and Lewis rat strains. Neurosci Lett 236: 112-116. [crossref]
16. Rozeboom AM, Akil H, Seasholtz AF (2007) Mineralocorticoid receptor overexpression in forebrain decreases anxiety-like behavior and alters the stress response in mice. Proc Natl Acad Sci USA 104: 4688-4893. [crossref]
17. Fredersdorf S, Endemann DH, Luchner A, Heitzmann D, Ulucan C, et al. (2009) Increased aldosterone levels in a model of type 2 diabetes mellitus. Exp Clin Endocrinol Diabetes 117: 15-20. [crossref]
18. Miller Judith A (1999) “Impact of hyperglycemia on the renin angiotensin system in early human type 1 diabetes mellitus.” Journal of the American Society of Nephrology 10: 1778-1785. [crossref]
19. Zimering MB, Grinberg M, Burton J, Pang K (2020) Circulating Agonist Autoantibody to 5-Hydroxytryptamine 2A Receptor in Lean and Diabetic Fatty Zucker Rat Strains. Endocrinol Diabetes Metab J 4: 413. [crossref]

DOI: 10.31038/EDMJ.2023723

Abstract

Objective: Among experimental animal models, horses are the most adaptable to exercise and this ability has been extensively studied. Research on equine exercise physiology is mostly focused on genetics, and few integrated studies have focused on equine metabolomics. This study were conducted to analyze metabolites in plasma, urine, and sweat samples collected from Jeju pony and thoroughbred horses before and after exercise. In this study, we analyze the various equine samples using NMR (nuclear magnetic resonance) spectroscopy.

Methods: ¹H NMR spectroscopy analysis were conducted with equine plasma, urine, and sweat samples collected from Jeju pony and thoroughbred horses before and after exercise. Relative metabolite levels between three types of were compared under exercise stimuli and by breeds.

Results: A total 26, 39, and 36 metabolites were identified in each of plasma, sweat, and urine samples, respectively, of both thoroughbred and Jeju pony. A total 3, 12, 15 metabolites were exclusively detected in plasma, sweat, and urine samples, respectively, and 15 metabolites were detected in all samples at the same time. In addition, total 8 and 5 metabolites were detected after exercise in plasma and urine samples. Additionally, we obtained 16, 6, and 30 metabolites in plasma, urine, and sweat by breeds.

Keywords

Horse, Thoroughbred, Korean native horse, Jeju pony, Metabolites, Nuclear magnetic resonance spectroscopy

Introduction

Horses are the most adaptable experimental animal models to exercise and, as such, are the most suitable for studying its effects. Moreover, studies focused on exercise physiology in horses can provide valuable basic information for understanding underlying mechanisms associated with exercise in humans. For this reason, further research on exercise physiology is necessary [1]. However, although studies focused on improving the athletic performance of horses have not had much success, economic trait -related genes have recently received greater attention [2-4]. In addition, equine tissue derived cells are being used in studies on the functional validation of these genes [5,6].

In recent years, multivariate analyses, so called multi-omics (genomics, epigenomics, transcriptomics, metabolomics, and proteomics) have been used to explain the biological mechanisms in numerous animals. Metabolites are the final biological products of cellular processes in cells, tissues, organs, or organisms [7]. Quantification of metabolomes can explain several biological phenomena along with other omics studies.

Exercise has a powerful effect on the body metabolism [8]. Repetitive and unilateral contraction of muscles associated with frequent exercise training is a suitably strong stimulus of physiologic function. During exercise, blood-borne glucose, creatine phosphate, glycogen, free fatty acids, and lactate which is known as external molecular substrates were used to produce ATP in muscle. The importance of these external molecular substrates in exercise metabolism is mostly affected by exercise rate and duration, but can also be affected by the type of exercise itself, as well as diet and environmental factors [9]. In addition, other energy mechanisms may be needed depending on the degree and duration of exercise [10].

In a previous study, we investigated a metabolic mechanism activated by physical activity using ¹H nuclear magnetic resonance (NMR) spectroscopy in thoroughbred horses [11]. Specifically, we profiled exercise specific metabolome in muscles and plasma. However, the metabolic mechanism during exercise has not yet been analyzed in urine and sweat, which are much easier to collect than plasma and muscle.

In this study, the metabolite profiling of the sweat, plasma, and urine in various equine breeds under exercise stimulus was analyzed by ¹H NMR spectroscopy. Based on the results, commonly or specifically released metabolites were identified from various equine biopsied specimen. Subsequently, metabolic pathways associated with obtained metabolites were investigated. The present study could contribute to a better understanding of metabolic fluctuations caused by exercise in thoroughbred and Jeju pony.

Materials and Methods

Animals

In this study, samples were gathered from five Thoroughbred and five Jeju pony. The study design was approved by the Pusan National University-Institutional Animal Care and Use Committee (Approval Number: PNU-2015-0864).

Horse Sampling

Jeju pony and Thoroughbred horse samples of sweat, plasma, and urine were collected in a stable setting and following exercise (30 min). A 15 mL syringe was used to obtain blood samples, which were then transferred to heparin tubes and centrifuged at 5,000 rpm for 15 min to extract the plasma. Sweat samples were obtained only after exercise. In case of urine, obtained sample was centrifuged to remove solids. Supernatant of centrifuged urine samples was added to a 1.5 ml tube which, is containing D2O (deuterium oxide) solution, DSS( dextran sulphate sodium), and 10 mM imidazole. In addition, 0.42% sodium azide was added. Obtained plasma samples and sweat samples were stored at -20°C, and urine samples were stored at -70°C until conducting NMR spectroscopy.

Nuclear Magnetic Resonance Spectroscopy

Plasma, urine, and sweat samples were ¹H NMR spectroscopy analyzed. Briefly, plasma, urine, and sweat samples were used with D2O containing the reference material TSP (trimethylsilylpropionate) before NMR measurement. We conducted high-resolution magic angle spinning NMR for plasma samples, with a spinning rate of 2,050 Hz. Water peak and macromolecular peak signals were removed using the Carr-Purcell-Meiboom-Gill pulse sequence for analysis of plasma, sweat, and urine samples. Used to eliminate signals from water peaks and macromolecular peaks. Measured spectrum data were optimized by Chenomx NMR Suite 7.1 (Chenomx Inc., Edmonton, AB, Canada), and statistical analysis were conducted by SIMCAp+12.0 (Umetrics, Umea, Sweden) software. In this study, we measured the absolute concentrations of the metabolites in various equine samples. Relative concentrations were determine, and amount of metabolites present in the samples were calculated by multivariate statistical analysis method.

Statistical Analysis

A T-test and One-way ANOVA analysis of variance was conducted to determine significance levels. Data were shown as mean ± standard deviation of mean. One-way ANOVA  analysis of variance followed by Duncan multiple test was used to compare before and after exercise training results and used for each sample of thoroughbreds and Jeju pony.

Results

Comparison of Metabolic Patterns in Thoroughbred and Jeju Pony Before and After Exercise

In our previous study, we conducted ¹H NMR spectroscopy analysis with various equine tissue samples (plasma, muscle, and urine) following exercise [11]. In this study, we obtained plasma, urine, sweat samples from both thoroughbred and Jeju pony following exercise, as well as before exercise, and conducted ¹H NMR spectroscopy analysis (Figure 1A). We obtained a very large quantity of metabolomics data. A total of 26, 39, and 36 metabolites were identified in plasma (Figure 1B), sweat (Figure 1C), and urine samples (Figure 1D), respectively. To assess which metabolites were significantly released after exercise, we compared samples obtained before and after exercise in thoroughbred and Jeju pony. Glutamate, glutamine, glutathione, lactate, and pyruvate were detected in the Jeju pony plasma samples and betaine, citrate, glucose, glutamate, glutamine, glutathione, histidine, isoleucine, leucine, phenylalanine, proline, and valine were significantly released in plasma of thoroughbred horses (Supplementary Table 1). In urine samples, trimethylamine were identified in Jeju pony and 2-oxovalerate, 3-aminoisobutyrate, alanine, citrulline, glucose, glutamine, glutarate, methylsuccinate, N-isovaleroylglycine, N-phenylacetylglycine, proline, pyruvate, taurine, threonine, tryptophan, and urea were significantly released in thoroughbred horses (Supplementary Table 2). Notably, sweat samples were difficult to collect before exercise; as such, only those collected after exercise were used (Supplementary Table 3). In addition, we analyzed metabolites that were specifically released in each tissue (Table 1). A total of 3, 12, and 15 metabolites were identified in plasma, sweat, and urine, respectively.

Figure 1: Venn diagram showing shared and unique metabolites (A), and heatmap analysis of the differentially expressed metabolites (B-D) in the plasma, sweat, and urine. Red and green shadings represent higher and lower relative expression levels, respectively.

Table 1: Tissue specific metabolites in both of Thoroughbred and jeju pony

Clustering	Total	Metabolites
Plasma Only	3	Glutathione, Malonate, Ornithine
Sweat Only	12	2-Hydroxybutyrate, Acetoin, Choline, Formate, Fumarate, Glycerate, Homoserine, Mannose, N-Methylhydantoin, Phenylacetate, Pyroglutamate, Urocanate
Urine Only	15	2-Oxovalerate, 3-Aminoisobutyrate, 3-Hydroxyisovalerate, Acetoacetate, Citrulline, Dimethylamine, Glutarate, Hippurate, Methylsuccinate, N-Isovaleroylglycine, N-Phenylacetylglycine, Succinate, Trimethylamine, Trimethylamine N-oxide, Tryptophan
Plasma and Sweat	22	Acetate, Alanine, Betaine, Citrate, Creatine, Glucose, Glutamate, Glycerol, Glycine, Histidine, Isoleucine, Lactate, Leucine, Lysine, Proline, Pyruvate, Serine, Threonine, Tyrosine, Valine, myo-Inositol
Sweat and Urine	20	Acetate, Alanine, Arginine, Benzoate, Creatine, Creatinine, Glucose, Glycine, Isoleucine, Lactate, Leucine, Phenylalanine, Proline, Pyruvate, Taurine, Threonine, Tyrosine, Urea, Valine, myo-Inositol
Plasma and Urine	16	Acetate, Alanine, Creatine, Glucose, Glutamine, Glycine, Isoleucine, Lactate, Leucine, Phenylalanine, Proline, Pyruvate, Threonine, Tyrosine, Valine, myo-Inositol
Plasma, Sweat, and Urine	15	Acetate, Alanine, Creatine, Glucose, Glycine, Isoleucine, Lactate, Leucine, Phenylalanine, Proline, Pyruvate, Threonine, Tyrosine, Valine, Myo-Inositol

Metabolite Set Enrichment Analyses Based on Exercise Status

Enrichment analyses of the overlapped metabolites among plasma, urine, and sweat were conducted by MetaboAnalyst 5.0 [12], and total 41 pathways were identified (Table 2). Among various pathways, the glucose-alanine cycle, glycine and serine metabolism, and alanine metabolism were the most significantly expressed after exercise.

Table 2: Enriched metabolite pathway among plasma, urine and sweat

	Total	Expected	Hits	Raw p	Holm p	FDR
Glucose-Alanine Cycle	13	0.19	3	0.000667	0.0654	0.05
Glycine and Serine Metabolism	59	0.864	5	0.00103	0.1	0.05
Alanine Metabolism	17	0.249	3	0.00153	0.147	0.05
Gluconeogenesis	35	0.513	3	0.0126	1	0.308
Pyruvate Metabolism	48	0.703	3	0.0296	1	0.426
Glutamate Metabolism	49	0.718	3	0.0312	1	0.426
Glutathione Metabolism	21	0.308	2	0.0358	1	0.426
Arginine and Proline Metabolism	53	0.776	3	0.0383	1	0.426
Transfer of Acetyl Groups into Mitochondria	22	0.322	2	0.0391	1	0.426
Warburg Effect	58	0.85	3	0.0483	1	0.429
Glycolysis	25	0.366	2	0.0495	1	0.429
Valine, Leucine and Isoleucine Degradation	60	0.879	3	0.0526	1	0.429
Phenylalanine and Tyrosine Metabolism	28	0.41	2	0.0608	1	0.453
Urea Cycle	29	0.425	2	0.0647	1	0.453
Ammonia Recycling	32	0.469	2	0.0771	1	0.499
Amino Sugar Metabolism	33	0.483	2	0.0814	1	0.499
Galactose Metabolism	38	0.557	2	0.104	1	0.599
Lactose Degradation	9	0.132	1	0.125	1	0.68
Pyruvaldehyde Degradation	10	0.146	1	0.138	1	0.711
Thyroid hormone synthesis	13	0.19	1	0.176	1	0.86
Phosphatidylinositol Phosphate Metabolism	17	0.249	1	0.223	1	1
Ethanol Degradation	19	0.278	1	0.246	1	1
Catecholamine Biosynthesis	20	0.293	1	0.258	1	1
Lactose Synthesis	20	0.293	1	0.258	1	1
Threonine and 2-Oxobutanoate Degradation	20	0.293	1	0.258	1	1
Carnitine Synthesis	22	0.322	1	0.28	1	1
Cysteine Metabolism	26	0.381	1	0.322	1	1
Inositol Phosphate Metabolism	26	0.381	1	0.322	1	1
Selenoamino Acid Metabolism	28	0.41	1	0.342	1	1
Citric Acid Cycle	32	0.469	1	0.381	1	1
Inositol Metabolism	33	0.483	1	0.39	1	1
Aspartate Metabolism	35	0.513	1	0.409	1	1
Fatty Acid Biosynthesis	35	0.513	1	0.409	1	1
Porphyrin Metabolism	40	0.586	1	0.452	1	1
Sphingolipid Metabolism	40	0.586	1	0.452	1	1
Propanoate Metabolism	42	0.615	1	0.469	1	1
Methionine Metabolism	43	0.63	1	0.477	1	1
Tryptophan Metabolism	60	0.879	1	0.598	1	1
Bile Acid Biosynthesis	65	0.952	1	0.629	1	1
Tyrosine Metabolism	72	1.05	1	0.668	1	1
Purine Metabolism	74	1.08	1	0.678	1	1

Differentially Released Metabolites that Responded to Exercise in Plasma and Urine

A total of 15 metabolites, including acetate, alanine, and creatine, were observed in all sample types (plasma, urine, and sweat) (Table 1). For these metabolites, release pattern analysis after exercise was conducted in plasma and urine (Table 3). Lactate and pyruvate were significantly identified in the plasma of Jeju pony (Figure 2A) and six metabolites (glucose, isoleucine, leucine, phenylalanine, proline, and valine) were significantly identified in the thoroughbreds plasma samples (Figure 2B). In thoroughbred horses, most metabolites doubled after exercise, with glucose showing the biggest increase. Interestingly, metabolites that significantly increased after exercise in Jeju pony were showed a decreasing trend in thoroughbred horses after exercise. In addition, metabolic analysis was conducted in urine samples after exercise (Figure 3). In contrast with the plasma results, significant changes in the release of metabolites in urine were only found in the samples from thoroughbred horses. Alanine, glucose, proline, pyruvate, and threonine were significantly identified after exercise.

Table 3: Expression pattern of plasma metabolites overlapped among plasma, urine and sweat

Metabolites	Jeju Horse				Thoroughbreds
	Before (Mean ±SE) mM	After (Mean ±SE) mM	p value		Before (Mean ±SE) mM	After (Mean ±SE) mM	p value
Acetate	13.30 ±1.77	16.44 ±1.49	0.259	17.10±3.49		15.20±1.38	0.633
Alanine	15.26 ±1.35	17.09 ±1.82	0.49	16.39±2.46		10.78±3.18	0.248
Creatine	3.53 ±0.32	3.52 ±0.22	0.979	2.75 ± 0.41		1.79 ± 0.43	0.186
Glucose	118.24 ± 9.10	98.81 ± 5.52	0.141	104.00 ± 15.69		51.79 ± 9.07	0.0352**
Glycine	26.46 ± 2.95	23.95 ± 2.58	0.583	31.44 ± 7.79		14.75 ± 2.62	0.107
Isoleucine	2.30 ± 0.19	2.00 ± 0.24	0.4	2.46 ± 0.40		1.18 ± 0.30	0.0512*
Lactate	20.31 ± 2.14	32.04 ± 1.57	0.00418***	18.22 ± 3.18		15.94 ± 4.46	0.719
Leucine	7.33 ± 0.22	6.62 ± 0.46	0.243	7.53 ± 1.09		3.79 ± 0.78	0.0372**
Phenylalanine	2.02 ± 0.10	2.06 ± 0.20	0.895	2.30 ± 0.30		1.23 ± 0.27	0.0436**
Proline	8.38 ± 0.73	8.76 ± 0.47	0.704	11.48 ± 1.95		5.41 ± 1.42	0.0549*
Pyruvate	0.87 ± 0.12	1.40 ± 0.15	0.0388**	0.94 ± 0.16		0.64 ± 0.11	0.21
Threonine	11.06 ± 1.94	10.60 ± 1.84	0.88	16.51 ± 2.96		9.29 ± 3.14	0.172
Tyrosine	2.75 ± 0.27	2.51 ± 0.34	0.635	3.54 ± 0.46		2.25 ± 0.57	0.133
Valine	8.68 ± 0.75	8.27 ± 1.05	0.781	11.32 ± 2.17		5.45 ± 0.79	0.0527*
Myo-Inositol	2.73 ± 0.29	2.45 ± 0.15	0.475	3.38 ± 0.53		2.35 ± 0.94	0.418

Figure 2: Significant difference of metabolites in plasma by exercise in Jeju pony (A) and Thoroughbreds (B). *p<0. 1, **p<0.05, ***p<0.01, ****p<0.001. All values expressed in mM as mean ± SD.

Figure 3: Significant difference of metabolites in urine by exercise in Thoroughbreds. *p<0. 1, **p<0.05, ***p<0.01, ****p<0.001. All values expressed in mM as mean ± SD.

Comparison of Metabolites between Equine Breeds (Thoroughbred and Jeju Pony)

In addition, we compared the metabolites between thoroughbred and Jeju pony under exercise stimuli (Tables 4-7). A greater difference was found between the metabolites released by the two breeds after exercise than before exercise in all sample types. Citrate and histidine were significantly released before exercise (Figure 4A), and 16 metabolites, including betaine and citrate, were significantly released after exercise in plasma in both breeds (Figure 4B). Among them, citrate values tripled in samples collected after exercise in both breeds and betaine and pyruvate showed largest difference between species (Figure 4B). In urine samples, six metabolites, including creatine and creatinine, showed significant differences between breeds (Figure 5). The release of taurine and myo-inositol was significantly different by more than 3.5-fold between breeds before exercise (Figure 5A) and five metabolites (creatine, creatinine, trimethylamine N-oxide, urea, and myo-inositol) were significantly different after exercise (Figure 5B). Creatine, urea, and myo-inositol more than doubled their values in urine samples after exercise (Figure 5). Although sweat samples were difficult to collect before exercise, we still analyzed sweat metabolite patterns after exercise. Among 39 metabolites, 30, including 2-hydroxybutyrate, showed significant differences between species (Table 6). Interestingly, most detected metabolites had a higher value in Jeju pony than in thoroughbred horses.

Table 4: Expression pattern of urine metabolites overlapped among plasma, urine and sweat

Metabolites	Jeju Horse				Thoroughbreds
	Before (Mean ± SE) mM	After (Mean ± SE) mM	p value		Before (Mean ± SE) mM	After (Mean ± SE) mM	p value
Acetate	0.56 ± 0.14	0.41 ± 0.10	0.366	0.38 ± 0.12		0.43 ± 0.09	0.778
Alanine	0.04 ± 0.01	0.06 ± 0.02	0.573	0.03 ± 0.01		0.05 ± 0.01	0.0707*
Creatine	0.43 ± 0.32	0.09 ± 0.02	0.331	0.11 ± 0.04		0.17 ± 0.03	0.285
Glucose	0.43 ± 0.08	0.61 ± 0.10	0.285	0.27 ± 0.05		0.64 ± 0.09	0.00719***
Glycine	6.45 ± 4.42	0.39 ± 0.10	0.207	0.15 ± 0.04		0.23 ± 0.08	0.388
Isoleucine	0.06 ± 0.01	0.08 ± 0.01	0.165	0.05 ± 0.01		0.11 ± 0.03	0.107
Lactate	0.12 ± 0.04	0.13 ± 0.02	0.92	0.07 ± 0.02		0.13 ± 0.03	0.108
Leucine	0.08 ± 0.01	0.12 ± 0.02	0.308	0.10 ± 0.03		0.16 ± 0.03	0.16
Phenylalanine	0.48 ± 0.10	0.43 ± 0.08	0.793	0.31 ± 0.10		0.54 ± 0.11	0.158
Proline	0.49 ± 0.08	0.68 ± 0.17	0.2	0.32 ± 0.08		0.60 ± 0.10	0.076*
Pyruvate	0.11 ± 0.02	0.14 ± 0.05	0.71	0.08 ± 0.02		0.15 ± 0.02	0.0676*
Threonine	0.34 ± 0.10	0.19 ± 0.04	0.244	0.16 ± 0.04		0.29 ± 0.05	0.0783*
Tyrosine	0.48 ± 0.13	0.50 ± 0.11	0.924	0.36 ± 0.10		0.63 ± 0.11	0.114
Valine	0.06 ± 0.01	0.08 ± 0.01	0.349	0.07 ± 0.02		0.12 ± 0.03	0.177
Myo-Inositol	1.00 ± 0.19	1.15 ± 0.16	0.288	0.27 ± 0.06		0.41 ± 0.06	0.166

Table 5: Metabolite comparison between Thoroughbreds and jeju pony in plasma

Metabolites	Before (Mean ± SE) mM					After (Mean ± SE) mM
	TH	JH	p value			TH		JH	p value
Acetate	17.10 ± 3.49	13.30 ± 1.77	0.409		15.02 ± 1.38			16.44 ± 1.49	0.55
Alanine	16.39 ± 2.46	15.26 ± 1.35	0.729		10.78 ± 3.18			17.09 ± 1.82	0.162
Betaine	3.73 ± 1.06	3.59 ± 0.17	0.908		1.41 ± 0.34			3.17 ± 0.16	0.003***
Citrate	2.99 ± 0.36	4.09 ± 0.33	0.081**		1.48 ± 0.33			4.51 ± 0.37	0.0006****
Creatine	2.75 ± 0.41	3.53 ± 0.32	0.213		1.79 ± 0.43			3.52 ± 0.22	0.012**
Glucose	104.00 ± 15.69	118.24 ± 9.10	0.502		51.79 ± 9.70			98.81 ± 5.52	0.005***
Glutamate	9.08 ± 1.46	11.01 ± 0.67	0.315		4.07 ± 0.67			7.18 ± 0.71	0.022**
Glutamine	10.07 ± 1.89	9.98 ± 0.98	0.971		3.88 ± 0.50			7.61 ± 0.49	0.001***
Glutathione	24.70 ± 4.59	21.68 ± 1.19	0.584		12.90 ± 3.15			17.66 ± 1.52	0.259
Glycerol	3.28 ± 0.50	4.39 ± 0.27	0.118		2.64 ± 0.94			3.79 ± 0.15	0.311
Glycine	31.44 ± 7.79	26.46 ± 2.95	0.607		14.75 ± 2.62			23.95 ± 2.58	0.06*
Histidine	13.92 ± 2.01	8.56 ± 0.92	0.0621**		6.57 ± 1.55			8.70 ± 0.97	0.329
Isoleucine	2.46 ± 0.40	2.30 ± 0.19	0.759		1.18 ± 0.30			2.00 ± 0.24	0.089*
Lactate	18.22 ± 3.18	20.31 ± 2.14	0.639		15.94 ± 4.46			32.04 ± 1.57	0.016**
Leucine	7.53 ± 1.09	7.33 ± 0.22	0.877		3.79 ± 0.78			6.62 ± 0.46	0.024**
Lysine	59.40 ± 17.18	19.49 ± 9.91		0.11	39.78 ± 16.49		16.19 ± 8.54		0.289
Malonate	3.22 ± 0.42	3.98 ± 0.45		0.302	1.82 ± 0.58		3.19 ± 0.28		0.094*
Ornithine	6.46 ± 2.44	10.88 ± 3.31		0.364	3.35 ± 0.82		9.22 ± 3.22		0.152
Phenylalanine	2.30 ± 0.30	2.02 ± 0.10		0.455	1.23 ± 0.27		2.06 ± 0.20		0.056*
Proline	11.48 ± 1.95	8.38 ± 0.73		0.219	5.41 ± 1.42		8.76 ± 0.47		0.081*
Pyruvate	0.94 ± 0.16	0.87 ± 0.12		0.775	0.64 ± 0.11		1.40 ± 0.15		0.00625***
Serine	18.34 ± 3.74	15.52 ± 0.96		0.532	10.33 ± 3.30		14.46 ± 1.86		0.357
Threonine	16.51 ± 2.96	11.06 ± 1.94		0.205	9.29 ± 3.14		10.60 ± 1.84		0.756
Tyrosine	3.54 ± 0.46	2.75 ± 0.27		0.222	2.25 ± 0.51		2.51 ± 0.34		0.718
Valine	11.32 ± 2.17	8.68 ± 0.75		0.335	5.45 ± 0.79		8.27 ± 1.05		0.0917*
myo-Inositol	3.38 ± 0.53	2.73 ± 0.29		0.362	2.35 ± 0.94		2.45 ± 0.15		0.927

Table 6: Metabolite comparison between Thoroughbreds and jeju pony in Urine

Metabolites	Before (Mean ± SE) mM			After (Mean ± SE) mM
	TH	JH	p value	TH		JH	p value
2-Oxovalerate	0.09 ± 0.02	0.12 ± 0.03	0.378	0.22 ± 0.05		0.14 ± 0.03	0.338
3-Aminoisobutyrate	0.28 ± 0.07	0.29 ± 0.07	0.882	0.48 ± 0.08		0.41 ± 0.12	0.851
3-Hydroxyisovalerate	0.06 ± 0.01	0.04 ± 0.00	0.34	0.10 ± 0.02		0.08 ± 0.04	0.971
Acetate	0.38 ± 0.12	0.56 ± 0.14	0.362	0.43 ± 0.09		0.41 ± 0.1	0.946
Acetoacetate	0.17 ± 0.05	0.18 ± 0.04	0.949	0.47 ± 0.20		0.21 ± 0.04	0.231
Alanine	0.03 ± 0.01	0.04 ± 0.01	0.103	0.05 ± 0.01		0.06 ± 0.02	0.628
Arginine	0.29 ± 0.10	0.43 ± 0.12	0.397	0.32 ± 0.08		0.60 ± 0.2	0.246
Benzoate	0.07 ± 0.02	5.73 ± 3.49	0.143	0.05 ± 0.01		0.06 ± 0.01	0.937
Citrulline	0.30 ± 0.07	0.41 ± 0.07	0.298	0.58 ± 0.10		0.61 ± 0.15	0.735
Creatine	0.11 ± 0.04	0.43 ± 0.32	0.342	0.17 ± 0.03		0.09 ± 0.02	0.0719*
Creatinine	11.73 ± 3.30	11.90 ± 2.37	0.968	17.00 ± 2.32		8.84 ± 2.61	0.0733*
Dimethylamine	0.13 ± 0.03	0.18 ± 0.03	0.272	0.16 ± 0.03		0.20 ± 0.03	0.34
Glucose	0.27 ± 0.05	0.43 ± 0.08	0.146	0.64 ± 0.09		0.61 ± 0.1	0.928
Glutamine	0.37 ± 0.09	0.54 ± 0.08	0.217	0.70 ± 0.13		0.53 ± 0.1	0.502
Glutarate	0.08 ± 0.02	0.12 ± 0.02	0.299	0.19 ± 0.05		0.13 ± 0.03	0.585
Glycine	0.15 ± 0.04	6.45 ± 4.42	0.191	0.23 ± 0.08	0.39 ± 0.1		0.205
Hippurate	26.02 ± 8.77	19.32 ± 3.47	0.498	53.58 ± 15.38	35.04 ± 7.85		0.381
Isoleucine	0.05 ± 0.01	0.06 ± 0.01	0.814	0.11 ± 0.03	0.08 ± 0.01		0.313
Lactate	0.07 ± 0.02	0.12 ± 0.04	0.266	0.13 ± 0.03	0.13 ± 0.02		0.976
Leucine	0.10 ± 0.03	0.08 ± 0.01	0.553	0.16 ± 0.03	0.12 ± 0.02		0.405
Methylsuccinate	0.14 ± 0.04	0.15 ± 0.02	0.814	0.30 ± 0.07	0.21 ± 0.05		0.474
N-Isovaleroylglycine	0.10 ± 0.02	0.13 ± 0.03	0.502	0.16 ± 0.01	0.14 ± 0.03		0.831
N-Phenylacetylglycine	5.94 ± 1.38	7.40 ± 1.50	0.494	10.96 ± 1.83	8.63 ± 1.34		0.524
Phenylalanine	0.31 ± 0.10	0.48 ± 0.10	0.264	0.54 ± 0.11	0.43 ± 0.08		0.687
Proline	0.32 ± 0.08	0.49 ± 0.08	0.198	0.60 ± 0.10	0.68 ± 0.17		0.466
Pyruvate	0.08 ± 0.02	0.11 ± 0.02	0.44	0.15 ± 0.02	0.14 ± 0.05		0.998
Succinate	0.03 ± 0.01	0.04 ± 0.01	0.197	0.05 ± 0.01		0.03 ± 0.01	0.419
Taurine	0.23 ± 0.06	0.86 ± 0.23	0.0312**	0.78 ± 0.17		1.16 ± 0.41	0.738
Threonine	0.16 ± 0.04	0.34 ± 0.10	0.127	0.29 ± 0.05		0.19 ± 0.04	0.3
Trimethylamine	0.03 ± 0.01	0.04 ± 0.01	0.138	0.04 ± 0.00		0.02 ± 0	0.198
Trimethylamine N-oxide	0.15 ± 0.05	0.20 ± 0.06	0.553	0.11 ± 0.02		0.22 ± 0.03	0.025**
Tryptophan	0.26 ± 0.07	0.29 ± 0.03	0.69	0.50 ± 0.09		0.47 ± 0.08	0.973
Tyrosine	0.36 ± 0.10	0.48 ± 0.13	0.512	0.63 ± 0.11		0.50 ± 0.11	0.636
Urea	84.58 ± 16.17	79.19 ± 9.54	0.782	173.25 ± 11.44		103.46 ± 9.89	0.00177***
Valine	0.07 ± 0.02	0.06 ± 0.01	0.842	0.12 ± 0.03		0.08 ± 0.01	0.322
Myo-Inositol	0.27 ± 0.06	1.00 ± 0.19	0.0062***	0.41 ± 0.06		1.15 ± 0.16	0.00639***

Table 7: Metabolite comparison between Thoroughbreds and jeju pony in Sweat

Metabolites	Before (Mean ± SE) mM
	TH	JH	p value
2-Hydroxybutyrate	0.05 ± 0.01	0.28 ± 0.08	0.0505*
Acetate	1.06 ± 0.42	2.15 ± 0.29	0.331
Acetoin	0.00 ± 0.00	0.05 ± 0.00	1.07e-05****
Alanine	0.14 ± 0.05	1.40 ± 0.37	0.0281**
Arginine	0.06 ± 0.02	1.06 ± 0.48	0.108
Benzoate	0.03 ± 0.01	0.26 ± 0.10	0.095*
Betaine	0.01 ± 0.00	0.06 ± 0.02	0.0602*
Choline	0.00 ± 0.00	0.01 ± 0.00	0.00326***
Citrate	0.49 ± 0.06	8.16 ± 3.08	0.0693*
Creatine	0.03 ± 0.01	0.18 ± 0.01	0.000221****
Creatinine	0.07 ± 0.03	0.12 ± 0.04	0.532
Formate	0.18 ± 0.04	0.49 ± 0.02	0.00646***
Fumarate	0.01 ± 0.00	0.05 ± 0.01	0.0561*
Glucose	0.17 ± 0.06	1.55 ± 0.36	0.0223**
Glutamate	0.08 ± 0.01	0.45 ± 0.08	0.00996***
Glycerate	0.07 ± 0.02	0.31 ± 0.02	0.00222***
Glycerol	0.10 ± 0.02	1.18 ± 0.20	0.00578***
Glycine	0.14 ± 0.05	1.70 ± 0.47	0.0299**
Histidine	0.02 ± 0.00	0.63 ± 0.41	0.215
Homoserine	0.06 ± 0.01	0.33 ± 0.12	0.114
Isoleucine	0.03 ± 0.01	0.20 ± 0.06	0.0457**
Lactate	0.43 ± 0.14	5.90 ± 1.92	0.0519*
Leucine	0.04 ± 0.01	0.23 ± 0.06	0.0324**
Lysine	0.02 ± 0.01	0.19 ± 0.06	0.057*
Mannose	0.10 ± 0.03	0.16 ± 0.01	0.0367**
N-Methylhydantoin	0.00 ± 0.00	0.03 ± 0.01	0.0339**
Phenylacetate	0.02 ± 0.00	0.07 ± 0.03	0.129
Phenylalanine	0.02 ± 0.00	0.21 ± 0.06	0.0362**
Proline	0.06 ± 0.01	0.19 ± 0.04	0.0648*
Pyroglutamate	0.15 ± 0.05	1.61 ± 0.58	0.0639*
Pyruvate	0.10 ± 0.03	0.84 ± 0.01	1.44e-05****
Serine	0.18 ± 0.06	2.77 ± 1.41	0.138
Taurine	0.02 ± 0.00	0.08 ± 0.01	0.000513****
Threonine	0.04 ± 0.01	0.53 ± 0.27	0.15
Tyrosine	0.02 ± 0.00	0.13 ± 0.04	0.0503*
Urea	10.05 ± 1.98	7.93 ± 3.51	0.894
Urocanate	0.04 ± 0.01	0.24 ± 0.06	0.0283**
Valine	0.04 ± 0.01	0.26 ± 0.08	0.0506*
Myo-Inositol	0.06 ± 0.01	0.25 ± 0.04	0.0134**

Figure 4: Significant difference of metabolites in plasma between breeds (Jeju pony and Thoroughbreds) before (A) and After exercise (B). *p<0. 1, **p<0.05, ***p<0.01, ****p<0.001. All values expressed in mM as mean ± SD.

Figure 5: Significant difference of metabolites in urine between breeds (Jeju pony and Thoroughbreds) before (A) and after exercise (B). *p<0. 1, **p<0.05, ***p<0.01, ****p<0.001. All values expressed in mM as mean ± SD.

Discussion

Almost 60 million horses currently exist on the planet. In addition to providing important services such as transport, meat, leather, and ploughing force and in the majority of developing countries, horses are mainly used for sports and leisure activities in most developed countries [13]. Therefore, as one of their most important economic traits, most research conducted in horses focuses on improving their athletic abilities [14,15]. However, although their physical and physiological adaptations receive much attention [16], targeted genes and metabolites or underlying mechanisms associated with exercise are still understudied.

The advances in metabolic analysis technology that have been carried out allow the assessment of the physiological state of individuals [17] and prediction of their condition [18]. Therefore, metabolomics demonstrates various biological responses to environmental influences, genetic, transcriptomic, and proteomic, [19-21]. Because of these advantages, metabolic analysis is widely used to explore metabolic patterns [22] or to discover new biomarkers through physical changes associated with diseases or environmental changes [21,23].

Although previous studies have investigated the metabolic changes caused by exercise, most only analyzed skeletal muscle [24] and were further limited by their small sample size and little expansive metabolite platform [25]. Previous metabolic studies on exercise mainly focus on the effect of exercise in various tissues [26,27], and studies on the discovery of biomarkers, which are affected by the athletic ability of individuals, are relatively poorly performed. Jang et al., 2017, the basis of this study, conducted a metabolic analysis in skeletal muscle, plasma, and urine samples after exercise [11]. In this study, we performed a metabolic analysis in plasma, urine, and sweat samples of thoroughbred and Jeju pony by exercise. In addition, we demonstrated the influence of exercise and breed in metabolite levels. We obtained a large amount of metabolite data that were released after exercise. Among 15 metabolites that were commonly detected in plasma, urine, and sweat, the levels of lactate, pyruvate, glucose, isoleucine, leucine, phenylalanine, proline, and valine showed significantly changes after exercise in plasma samples (Figure 2), and the levels of alanine, glucose, proline, pyruvate, and threonine had significantly changed after exercise in urine samples (Figure 3). These results are in line with those of previous studies [11]. The metabolites observed in samples collected after exercise were all associated with the tricarboxylic acid (TCA) cycle, with some being intermediate products. Alanine, aspartate, and glutamate metabolism and aminoacyl-tRNA and arginine biosynthesis related metabolic pathways are activated by acute exercise [28]. These results suggest that several metabolic pathways that utilize skeletal muscle substrate are regulated after exercise, and previous studies reported that this occurs in various tissues [29,30].

During exercise, muscle glycogen, its main source of energy, is altered to glucose and subsequently to pyruvate via glycolysis [31]. The pyruvate converted by glycolysis can enter TCA and glucose-alanine cycles or be converted to lactate [32]. During aerobic exercise, muscle glycogen can be used to produce ATP through glycolysis; however, when anaerobic exercise like a sprint is conducted, the muscles cannot use oxygen for glycolysis [33]. Therefore, muscle glycogen (glucose) is altered to lactate through anaerobic glycolysis [33]. Then, the lactate is released to the bloodstream and transferred to the kidneys and liver [34]. In the liver, lactate is altered to pyruvate through gluconeogenesis [35]. In addition, when amino acids are used for energy in extrahepatic tissues, pyruvate derived from the glycolysis is used as an amino group receptor to form alanine, a non-essential amino acid [36]. The produced alanine is transferred to the liver through the bloodstream and converted to either pyruvate for gluconeogenesis via the glycose–alanine cycle or to glutamate, which then goes through the urea cycle. Collectively, the detected metabolites in equine plasma and urine including glucose, alanine, and lactate were altered to pyruvate and used for energy production. Therefore, the metabolites discovered in this study can be used as a reasonable indicator to measure athletic ability and exercise fatigue.

In conclusion, we compared metabolite presence between thoroughbreds and Jeju pony after exercise and analyzed enriched metabolic pathways of commonly detected metabolites in all samples (plasma, urine, and sweat). Our results could help improve our understanding of exercise fatigue and find regulation markers for fatigue reduction. Further research is necessary to combine these results with other omics data and reveal the function of metabolic markers.

Declarations

Ethics Approval and Consent to Participate

All animal procedures used in the study were conducted in compliance with international standards and were approved by the Institutional Animal Care and Use Committee of Pusan National University (Approval Number: PNU-2015-0864).

Competing Interests

The authors declare that they have no competing interests.

Acknowledgments

This work was supported by a 2-Year Research Grant from the Pusan National University.

Author’s Contribution

The research was conceptualized by Park JW, Cho BW and further edition was done by all the authors.  Data was curated by Park JW, Kim KH, and analyzed by Park JW, Lee SI, Sang SS. All authors have participated on data interpretation. The draft of the manuscript was written by Park JW and Kim KH, and the final form was edited by Lee SI, Sang SS, and Cho BW. All authors have contributed by interpretation, analysis, critical discussion.

References

1. Kim H, Lee T, Park W, Lee JW, Kim J, et al. (2013) Peeling back the evolutionary layers of molecular mechanisms responsive to exercise-stress in the skeletal muscle of the racing horse. DNA Res 20: 287-298. [crossref]
2. PARK Jeong-Woong, Choi JY, Hong SA, Kim NY, Do KT, et al. (2017) Exercise induced upregulation of glutamate-cysteine ligase catalytic subunit and glutamate-cysteine ligase modifier subunit gene expression in Thoroughbred horses. Asian-Australasian Journal of animal sciences 30: 728-735. [crossref]
3. CHO Hyun-Woo, Shin S, Park JW, Choi JY, Kim NY, et al. (2015) Molecular characterization and expression analysis of the peroxisome proliferator activated receptor delta (PPARδ) gene before and after exercise in horse. Asian-Australasian Journal of Animal Sciences 28: 697. [crossref]
4. KHUMMUANG Saichit, Lee HG, Joo SS, Park JW, Choi JY, et al. (2020) Comparison for immunophysiological responses of Jeju and Thoroughbred horses after exercise. Asian-Australasian Journal of Animal Sciences 33: 424-435. [crossref]
5. PARK Jeong-Woong, Kim KH, Choi JK, Park TS, Song KD, et al. (2021) Regulation of toll-like receptors expression in muscle cells by exercise-induced stress. Animal Bioscience 34: 1590. [crossref]
6. LEE Hyo Gun, Choi JY, Park JW, Park TS, Song KD, et al. (2019) Effects of exercise on myokine gene expression in horse skeletal muscles. Asian-Australasian Journal of Animal Sciences 32: 350-356. [crossref]
7. Jordan Kate W, Nordenstam J, Lauwers GY, Rothenberger DA, Alavi K, et al. (2009) “Metabolomic characterization of human rectal adenocarcinoma with intact tissue magnetic resonance spectroscopy.” Diseases of the colon and rectum 52: 520. [crossref]
8. Hargreaves Mark (2000) “Skeletal muscle metabolism during exercise in humans.” Clinical and Experimental Pharmacology and Physiology 27: 225-228. [crossref]
9. Moghetti Paolo, Bacchi E, Brangani C, Donà S, Negri C (2016) “Metabolic effects of exercise.” Sports Endocrinology 47: 44-57. [crossref]
10. Jang Hyun-Jun, Kim DM, Kim KB, Park JW, Choi JY, et al. (2017) “Analysis of metabolomic patterns in thoroughbreds before and after exercise.” Asian-Australasian journal of animal sciences 30: 1633-1642. [crossref]
11. PANG Zhiqiang, Zhou G, Ewald J, Chang L, Hacariz O, et al. (2022) Using MetaboAnalyst 5.0 for LC–HRMS spectra processing, multi-omics integration and covariate adjustment of global metabolomics data. Nature Protocols 17: 1735-1761. [crossref]
12. ORLANDO Ludovic (2020) The evolutionary and historical foundation of the modern horse: Lessons from ancient genomics. Annual Review of Genetics 54: 563-581. [crossref]
13. Wilkin Tessa, Anna Baoutina, Natasha Hamilton (2017) “Equine performance genes and the future of doping in horseracing.” Drug testing and analysis 9: 1456-1471. [crossref]
14. Arfuso Francesca, Assenza A, Fazio F, Rizzo M, Giannetto C, et al. (2019) “Dynamic change of serum levels of some branched-chain amino acids and tryptophan in athletic horses after different physical exercises.” Journal of equine veterinary science 77: 12-16. [crossref]
15. Arfuso Francesca, Giudice E, Panzera M, Rizzo M, Fazio F, et al. (2022) “Interleukin-1Ra (Il-1Ra) and serum cortisol level relationship in horse as dynamic adaptive response during physical exercise.” Veterinary Immunology and Immunopathology 243: 110368. [crossref]
16. SABATINE Marc S, Liu E, Morrow DA, Heller E, McCarroll R, et al. (2005) Metabolomic identification of novel biomarkers of myocardial ischemia. Circulation 112: 3868-3875. [crossref]
17. SHAH Svati H, Bain JR, Muehlbauer MJ, Stevens RD, Crosslin DR, et al. (2010) Association of a peripheral blood metabolic profile with coronary artery disease and risk of subsequent cardiovascular events. Circulation: Cardiovascular Genetics 3: 207-214. [crossref]
18. GERMAN J Bruce, HAMMOCK Bruce D, WATKINS Steven M (2005) Metabolomics: building on a century of biochemistry to guide human health. Metabolomics 1: 3-9. [crossref]
19. OREŠIČ Matej, VIDAL-PUIG Antonio, HÄNNINEN Virve (2006) Metabolomic approaches to phenotype characterization and applications to complex diseases. Expert review of molecular diagnostics 6: 575-585. [crossref]
20. MOORE Rowan E, Kirwan J, Doherty MK, Whitfield PD (2007) Biomarker discovery in animal health and disease: the application of post-genomic technologies. Biomarker insights 2: 185-196. [crossref]
21. KELL Douglas B (2004) Metabolomics and systems biology: making sense of the soup. Current opinion in microbiology 7: 296-307. [crossref]
22. WHITFIELD, Phillip David, et al. (2005) Metabolomics as a diagnostic tool for hepatology: validation in a naturally occurring canine model. Metabolomics 1: 215-225.
23. HUFFMAN Kim M, Koves TR, Hubal MJ, Abouassi H, Beri N, et al. (2014) Metabolite signatures of exercise training in human skeletal muscle relate to mitochondrial remodelling and cardiometabolic fitness. Diabetologia 57: 2282-2295. [crossref]
24. HUFFMAN Kim M, Slentz CA, Bateman LA, Thompson D, Muehlbauer MJ, et al. (2011) Exercise-induced changes in metabolic intermediates, hormones, and inflammatory markers associated with improvements in insulin sensitivity. Diabetes Care 34: 174-176. [crossref]
25. BRENNAN Andrea M, Benson M, Morningstar J, Herzig M, Robbins J, et al. (2018) Plasma metabolite profiles in response to chronic exercise. Medicine and science in sports and exercise 50: 1480-1486. [crossref]
26. BRENNAN Andrea M, Tchernof A, Gerszten RE, Cowan TE, Ross R (2018) Depot-specific adipose tissue metabolite profiles and corresponding changes following aerobic exercise. Frontiers in Endocrinology 9: 759. [crossref]
27. TABONE Mariangela, et al. (2021) The effect of acute moderate-intensity exercise on the serum and fecal metabolomes and the gut microbiota of cross-country endurance athletes. Scientific reports 11: 1-12.
28. MANAF Faizal A, Lawler NG, Peiffer JJ, Maker GL, Boyce MC, et al. (2018) Characterizing the plasma metabolome during and following a maximal exercise cycling test. Journal of Applied Physiology 125: 1193-1203. [crossref]
29. HEINONEN Ilkka, Kalliokoski KK, Hannukainen JC, Duncker DJ, Nuutila P, et al. (2014) Organ-specific physiological responses to acute physical exercise and long-term training in humans. Physiology 29: 421-436. [crossref]
30. SPRIET Lawrence L (2002) Regulation of skeletal muscle fat oxidation during exercise in humans. Medicine and science in sports and exercise9: 1477-1484. [crossref]
31. ISHIKURA, Keisuke; RA, Song-Gyu, OHMORI, Hajime (2013) Exercise-induced changes in amino acid levels in skeletal muscle and plasma. The Journal of Physical Fitness and Sports Medicine3: 301-310.
32. HIGGINS, Chris. Lactate and lactic acidosis. 2007. [crossref]
33. YANG Woo-Hwi, Park H, Grau M, Heine O (2020) Decreased blood glucose and lactate: is a useful indicator of recovery ability in athletes? International Journal of Environmental Research and Public Health15: 5470. [crossref]
34. CHIANG J (2014) Liver physiology: Metabolism and detoxification.
35. EXTON JH, Mallette LE, Jefferson LS, Wong EH, Friedmann N, et al. (1970) The hormonal control of hepatic gluconeogenesis. In: Proceedings of the 1969 Laurentian Hormone Conference. Academic Press 411-461. [crossref]
36. JOSE Caroline, Melser S, Benard G, Rossignol R (2013) Mitoplasticity: adaptation biology of the mitochondrion to the cellular redox state in physiology and carcinogenesis. Antioxidants & Redox Signaling7: 808-849. [crossref]