Paraxanthine enhances memory and neuroplasticity more than caffeine in rats

Abstract

Paraxanthine (PXN) is the main metabolite of caffeine (CAF). PXN supplementation has been shown to increase measures of cognition, memory, reasoning, response time, and sustained attention; however, no preclinical study has compared the effects of PXN with those of CAF. The aim of this study was to compare the effects of PXN and CAF on memory and related biomarkers in rats. The effects of two different doses of PXN (PXN LOW, PXN HIGH), CAF (CAF HIGH), and a control group on cognition (escape latency in the Morris water maze test), neurotransmitters (acetylcholine, dopamine, and gamma-aminobutyric acid), and neurochemicals (BDNF, catalase, glutathione, and cyclic GMP) were analyzed from whole brain samples in young (8 weeks old) and aged (16 months old) rats. Compared to the control group, escape latency improved in PXN LOW, PXN HIGH, and CAF HIGH (all P < 0.05) in young animals, and in PXN HIGH and CAF HIGH in older animals (P < 0.001). PXN HIGH improved escape latency compared to CAF HIGH in both young (P < 0.001) and old animals (P = 0.003). BDNF levels increased in PXN LOW, PXN HIGH, and CAF HIGH (all P < 0.001), with PXN HIGH increasing BDNF to a greater extent compared to CAF HIGH (P = 0.03). PXN HIGH also significantly increased BDNF levels compared to PXN LOW (P < 0.001). All other neurotransmitters and neurochemicals significantly increased in the PXN HIGH and CAF HIGH groups compared to the control. In conclusion, PXN showed greater improvements in cognition and BDNF levels compared to CAF, further substantiating PXN as a nootropic with greater benefits compared to CAF.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00221-024-06954-0.

Introduction

Caffeine (CAF) is one of the most consumed naturally occurring nootropics and exercise-related ergogenic aids (Guest et al. 2021). However, studies on CAF supplementation have shown substantial variability in outcomes, with some studies reporting no benefits or even a worsening of performance in certain subjects. Individuals with a homozygous A allele of the CYP1A2 gene tend to produce more cytochrome P450, an enzyme responsible for about 95% of caffeine metabolism, and consequently metabolize CAF more quickly. Fast metabolizers of CAF experience greater ergogenic outcomes in most studies (Guest et al. 2018; Minaei et al. 2022; Womack et al. 2012), but not all (Pataky et al. 2016), indicating that the metabolic products of CAF are the actual active ingredients.

Paraxanthine (1,7-dimethylxanthine, PXN) is the main metabolite of CAF, accounting for 70–72% of ingested CAF and 85% of the methylxanthine metabolic by-products (Stavric 1988). In comparison to CAF, PXN has a shorter half-life (Lelo et al. 1986), is less toxic (Purpura et al. 2021), and is less anxiogenic (Benowitz et al. 1995). Like CAF, PXN is a central nervous system stimulant; however, compared to CAF, PXN has higher binding potency for adenosine A1 and A2A receptors and produces more substantial locomotor activation effects (Orrú et al. 2013). PXN inhibits phosphodiesterase 9 (PDE9), which terminates nitric oxide (NO) neurotransmission by metabolizing cyclic guanosine monophosphate (cGMP) back to GMP. Through PDE9 inhibition, PXN potentiates NO neurotransmission, an effect that is absent with CAF (Orrú et al. 2013; Ferré et al. 2013).

Similar to CAF (Canas et al. 2009; Dall’Igna et al. 2003), PXN has been shown to have protective effects on dopaminergic neurons and is reported to protect against neurodegeneration and the loss of synaptic function (Guerreiro et al. 2008). Epidemiological studies suggest that higher intake of CAF reduces the risk of Alzheimer’s disease (Eskelinen and Kivipelto 2010). Preclinical studies indicate that reduced β-amyloid production is the likely mechanism behind caffeine’s protective effects on cognition (Arendash et al. 2006). CAF has been shown to improve memory by increasing brain-derived neurotrophic factor (BDNF) levels in mice (Sallaberry et al. 2013).

Beyond these mechanisms, broader evidence indicates that the impact of CAF on memory is intricately linked to A2A receptor-mediated control of synaptic dysfunction. This association has been observed in various models, including non-amyloid models of Alzheimer’s disease (Viana da Silva et al. 2016; Espinosa et al. 2013; Gonçalves et al. 2019; Laurent et al. 2014), as well as in animal models not characterized by altered amyloid metabolism, such as instances of repeated stress (Kaster et al. 2015), convulsions (Cognato et al. 2010), diabetic encephalopathy (Duarte et al. 2012), and aging (Temido-Ferreira et al., 2022). Thus, the neuroprotective and memory-preserving effects of CAF have been consistently associated with the correction of synaptic dysfunction. However, a direct comparison between CAF and PXN regarding their impact on cognition and associated mechanisms is lacking.

Aging is accompanied by a range of physiological and neurological changes that can significantly impact cognitive function, including declines in synaptic plasticity (deToledo-Morrell et al., 1988), neurogenesis (Culig et al. 2022), and alterations in neurotransmitter systems (Mora et al. 2007)—all of which may impair learning and memory. Clinical studies in young (Lieberman et al. 1987; Kaplan et al. 1997) and older adults (Ryan et al. 2002; Rees et al. 1999) generally demonstrate that CAF enhances attention, alertness, and short-term memory at dosages ranging from approximately 32 to 500 mg. PXN, a less-studied metabolite of CAF, has distinct pharmacokinetic and pharmacodynamic properties (Lelo et al. 1986; Benowitz et al. 1995), increasing dopaminergic activity (Guerreiro et al. 2008) and potentially improving memory consolidation more effectively than caffeine. However, the extent of these benefits across different age groups, particularly regarding age-related cognitive decline, remains to be explored.

We recently reported that acute ingestion of PXN enhanced memory, reaction time, and attention for up to six hours in healthy adults (Yoo et al. 2021), and that acute ingestion of as little as 50 mg of PXN for seven days enhanced measures of cognition, memory, reasoning, response time, and sustained attention (Xing et al. 2021). However, no preclinical study has directly compared the cognitive effects of PXN and CAF supplementation. The primary aims of this study were to determine: (1) whether PXN ingestion affects cognitive function in young and old rats; (2) whether PXN is more effective than CAF; and (3) whether CAF and PXN influence the same mechanistic targets known to enhance cognition: neurotransmitter levels, β-amyloid production, BDNF, and antioxidant capacity.

Materials and methods

Animals and study design

Sixty-four male Swiss Albino rats, aged 8 weeks (n = 32; 8 per group) and 16 months (n = 32; 8 per group), were housed in an animal room at a constant temperature (22 ± 3 °C) and humidity (30–70%) under a 12:12 h light-dark cycle, with a standard laboratory diet (Purina 5L79, Rat and Mouse 18% protein; PMI Nutrition International, Brentwood, MO, USA). Animals were housed in standard polypropylene cages (4 animals per cage) with a stainless-steel top grill, containing pelleted food and drinking water. Sterile corncob was used as bedding material and was changed daily. Reverse osmosis-purified water was provided ad libitum. All procedures involving animals were conducted humanely and were performed by or under the direction of trained personnel. The study protocol was reviewed and approved by the Institutional Animal Ethical Committee (IAEC) of Radiant Research Services Pvt. Ltd (Bangalore, India).

To compare the effects of paraxanthine against caffeine and the impact of age, young 8-week-old (YOUNG) and aged 16-month-old (OLD) rats were evenly divided into four groups (n = 8 per group): (i) vehicle control (CON), (ii) low-dose paraxanthine (PXN LOW), (iii) high-dose paraxanthine (PXN HIGH), and (iv) high-dose caffeine (CAF HIGH). The doses administered to the rats were calculated using the US Food and Drug Administration guidelines for human equivalence doses (HED), assuming a human weight of 60 kg (Nair et al., 2016). The following HEDs were used in this study: 25 mg/d paraxanthine (PXN LOW; rat dose: 2.57 mg/kg/d), 100 mg/d paraxanthine (PXN HIGH; rat dose: 10.28 mg/kg/d), and 100 mg/d of caffeine (CAF; rat dose: 10.28 mg/kg/d). Paraxanthine (enfinity^®) was supplied by Ingenious Ingredients L.P. (Lewisville, TX, USA), and caffeine in anhydrous form was supplied by NNB Nutrition, China. Previous studies at a dosage of up to 10 mg/kg caffeine showed no anxiolytic or anxiogenic effects in an open-field test and an elevated plus-maze test (Bhattacharya et al. 1997), nor did it produce either rewarding or aversive effects in a conditioned place preference test (Brockwell et al. 1991). However, it did have a significant facilitative effect on fear extinction (Ozawa et al. 2022).

The prescribed doses were prepared daily by taking the required quantity of the test item in a mortar and pestle. The doses were triturated, and an adequate quantity of vehicle was added, mixed well, and transferred to a volumetric flask. An additional quantity of vehicle was added to the beaker, rinsed, and transferred to the volumetric flask. The required volume was made up by adding sufficient vehicle to the volumetric flask, mixed well, and then transferred to labeled beakers with magnetic beads. Homogeneity of the test item formulations during dose administration was maintained by continuous stirring using a magnetic stirrer. The amount of the test item and the volume of the formulations prepared varied depending on the requirements and/or body weight of the animals. A total of 0.5% carboxymethyl cellulose sodium was used as the CON, and the test item formulations were prepared daily.

Dosing of all test materials was conducted via oral gavage using disposable polypropylene syringes with sterilized stainless steel gavage tubes. The test item formulation was administered once each day by oral route for 11 consecutive days between 9:00 and 11:00 h. The dose volume administered to each rat was 10 mL/kg/day. The dose volume was calculated for individual animals on the first day of treatment and was re-calculated according to the most recent body weights recorded during the treatment period. Body weight was recorded at baseline and on days 5, 8, and 15.

Morris water maze test

The Morris water maze test is widely used to study spatial memory and learning (Nunez 2008) and was adapted from Morris (Morris 1984). The Morris water maze pool (Orchid Scientific & Innovative India Pvt Ltd) consisted of a circular pool with an inner diameter of 183 cm and an outer diameter of 76 cm in height. The pool was located in a room with geometric shapes on the walls serving as spatial cues. A platform was placed in one quadrant of the pool, submerged 1 cm below the water surface. The water in the pool (25 ± 1 °C) was mixed with non-toxic black paint to prevent the animals from seeing the platform, encouraging them to rely on external maze cues. All animals were trained to locate the submerged platform in a constant location. As the animals became more familiar with the task, they were able to find the platform more quickly, highlighting their spatial memory and learning abilities.

Rats were placed in the water, and escape latency—the time (in seconds) taken to recognize the submerged platform in the designated quadrant—was measured. During training, each rat started at one of four starting points (north, south, east, or west) and was allowed to swim until it located the platform or until 60 s had elapsed. If the rat left the platform within 15 s, the measurement continued. Prior to any supplementation, all animals were subjected to the Morris water maze test to familiarize them with the task. Familiarization was completed over four consecutive days (Days 1–4), with each day providing exposure to each of the four starting points: north, south, east, and west. After the initial training, a probe test was conducted by hiding the platform to ensure that the animals’ performance was truly dependent on spatial memory. Once familiarization was complete, the animals were treated for 10 consecutive days (Days 5–14) in a randomized, placebo-controlled manner. On Day 15, 30 min after administration of the treatments, the animals performed the Morris water maze test. Escape latency was recorded, and the mean of the four starting points was used for data analysis.

Sample collection and neurochemical analysis

On Day 16, after administration of the final treatment, all animals were euthanized using 95% CO2, and the brain was excised and weighed. The brain tissues were rinsed with ice-cold PBS (pH 7.4) to thoroughly remove excess blood and then weighed before homogenization. A portion of brain tissue (100 mg) was taken in a 900-microliter volume of PBS (1:9 ratio) in a glass homogenizer on ice and homogenized at 2,000 rpm for 5 min. After homogenization, the supernatant was collected for analysis using standard ELISA assay kits according to the manufacturer’s recommendations for clinical biochemical markers, including acetylcholine (BT LAB, China; Lot No.: E0698Mo), dopamine (BT LAB, China; Lot No.: E0667Mo), brain-derived neurotrophic factor (BDNF; Elabscience, USA; Lot No.: E-EL-R1235), β-amyloid (1–40) (Elabscience, USA; Lot No.: E-EL-R3030), catalase (Elabscience, USA; Lot No.: E-BC-K031), glutathione (Elabscience, USA; Lot No.: KTE101106), gamma-aminobutyric acid (GABA; Fine Test, China; Lot No.: ER1707), and cyclic GMP (Fine Test, China; Lot No.: ER0831).

Statistical analysis

Brain weight, escape latency, and biochemical markers were analyzed using a two-way analysis of variance (ANOVA), with group and age specified as between-subjects factors. Body mass data were analyzed separately for young and aged individuals due to the known differences in body mass between these groups, using two-way ANOVA with group as a between-subjects factor and time (day) as a within-subjects factor. Statistically significant effects were followed up with pairwise comparisons using the Tukey adjustment to account for multiple comparisons. Due to the large number of ANOVA tests, P-values for main effects and interactions were corrected using the false discovery rate method. Partial eta-squared effect sizes were calculated for all ANOVA tests. Homogeneity of variances was tested using Levene’s test, and normality of residuals was examined through visual inspection of quantile-quantile plots. For models with within-subject factors (i.e., the body mass models), the Greenhouse-Geisser correction was applied if sphericity violations occurred. After P-value adjustments, statistical significance was accepted at P < 0.05. Data analysis was performed using R (v. 4.2.1). Values are presented as means ± SD (Table.1).

Table 1.

Summary of statistical results

Variable	ANOVA Term	DF (numerator)	DF (denominator)	F-value	P-value (adjusted)	Effect Size
Escape Latency	Group	3	56	931.00	< 0.001*	0.98
	Age	1	56	164.79	< 0.001*	0.75
	Group x Age	3	56	8.00	< 0.001*	0.30
Body Weight (Young)	Group	3	28	1.01	0.60	0.10
	Day	2.4	68.4	9053.44	< 0.001*	1.00
	Group x Day	7.3	68.4	0.57	0.90	0.06
Body Weight (Old)	Group	3	28	0.96	0.60	0.09
	Day	1.5	43.2	1258.21	< 0.001*	0.98
	Group x Day	4.6	43.2	0.58	0.85	0.06
Brain Weight	Group	3	56	0.33	0.90	0.02
	Age	1	56	443.26	< 0.001*	0.89
	Group x Age	3	56	0.17	0.94	0.01
Acetylcholine	Group	3	56	33.96	< 0.001*	0.65
	Age	1	56	86.44	< 0.001*	0.61
	Group x Age	3	56	0.31	0.90	0.02
Dopamine	Group	3	56	104.68	< 0.001*	0.85
	Age	1	56	109.79	< 0.001*	0.66
	Group x Age	3	56	0.46	0.85	0.02
BDNF	Group	3	56	72.74	< 0.001*	0.80
	Age	1	56	58.46	< 0.001*	0.51
	Group x Age	3	56	0.93	0.60	0.05
Amyloid Beta	Group	3	56	65.39	< 0.001*	0.78
	Age	1	56	381.70	< 0.001*	0.87
	Group x Age	3	56	0.55	0.84	0.03
Catalase	Group	3	56	166.94	< 0.001*	0.90
	Age	1	56	17.34	< 0.001*	0.24
	Group x Age	3	56	0.64	0.79	0.03
Glutathione	Group	3	56	200.02	< 0.001*	0.91
	Age	1	56	132.44	< 0.001*	0.70
	Group x Age	3	56	12.98	< 0.001*	0.41
GABA	Group	3	56	34.40	< 0.001*	0.65
	Age	1	56	108.44	< 0.001*	0.66
	Group x Age	3	56	0.13	0.94	0.01
Cyclic GMP	Group	3	56	148.91	< 0.001*	0.89
	Age	1	56	236.52	< 0.001*	0.81
	Group x Age	3	56	0.22	0.94	0.01

*Denotes significant difference

Results

Effect of paraxanthine, caffeine and age on body and brain weight

Body weight increased on days 5, 8, and 15 (time effect and pairwise comparisons; P < 0.001) in both young and old animals (Table 2) and was not affected by supplement status (group effect; P = 0.59).

Table 2.

Effect of paraxanthine and caffeine on body weight (in grams)

	YOUNG				OLD
Day	CON	PXN LOW	PXN HIGH	CAF HIGH	CON	PXN LOW	PXN HIGH	CAF HIGH
0	162.9 ± 1.7	163.7 ± 1.4	164.0 ± 1.5	163.8 ± 1.8	490.6 ± 2.0	491.3 ± 2.1	490.3 ± 2.3	489.6 ± 2.1
5*	168.0 ± 1.7	168.6 ± 1.6	169.2 ± 1.2	168.9 ± 1.8	493.6 ± 2.2	494.4 ± 2.2	493.4 ± 2.5	492.6 ± 2.1
8*	171.0 ± 1.7	171.6 ± 1.6	172.4 ± 1.2	171.9 ± 1.8	495.1 ± 2.1	495.8 ± 2.0	494.7 ± 2.4	493.9 ± 2.1
15*	182.4 ± 1.4	183.0 ± 1.5	184.0 ± 1.9	183.2 ± 1.7	500.0 ± 2.2	500.2 ± 2.6	499.0 ± 2.3	498.7 ± 1.9

*Denotes significant difference from day 0 on pairwise comparisons (P < 0.001 for each comparison). Data presented as mean ± SD for n = 8 in each group. CON, control; PXN LOW, low dose paraxanthine; PXN HIGH, high dose paraxanthine; CAF HIGH, high dose caffeine

Irrespective of supplement status, brain weight was greater in old compared to young animals (age effect; P < 0.001) with no effect of group (P = 0.92; Table 3).

Table 3.

Effect of paraxanthine and caffeine on brain weight (in grams)

YOUNG				OLD*
CON	PXN LOW	PXN HIGH	CAF HIGH	CON	PXN LOW	PXN HIGH	CAF HIGH
1.72 ± 0.03	1.72 ± 0.05	1.72 ± 0.03	1.73 ± 0.03	1.97 ± 0.05	1.97 ± 0.05	1.98 ± 0.07	1.99 ± 0.06

*Denotes main effect of age (P < 0.001) with brain weight is greater in OLD compared to YOUNG. Data presented as mean ± SD for n = 8 in each group. CON, control; PXN LOW, low dose paraxanthine; PXN HIGH, high dose paraxanthine; CAF HIGH, high dose caffeine

Effect of paraxanthine, caffeine and age on escape latency

A group × age interaction was observed for escape latency (P < 0.001; Fig. 1). Post hoc analysis revealed that, in young animals, PXN LOW (P = 0.049), PXN HIGH (P < 0.001), and CAF HIGH (P < 0.001) reduced escape latency by approximately 6–60% compared to control. Escape latency was also reduced in CAF HIGH and PXN HIGH by approximately 57% and 50%, respectively, when compared to PXN LOW (both P < 0.001). Additionally, in young animals, escape latency was further reduced in PXN HIGH compared to CAF HIGH by approximately 14% (P = 0.01).

Fig. 1.

Timeline diagram of the chronological order of manipulations

In old animals, both PXN HIGH (P < 0.001) and CAF HIGH (P < 0.001) reduced escape latency by approximately 40–47% compared to control and by approximately 37–45% compared to PXN LOW. Compared to CAF HIGH, PXN HIGH reduced escape latency in old animals by approximately 12.1% (P = 0.003). No significant differences between control and PXN LOW were observed in aged animals, although a trend was present (P = 0.07). Escape latency in the control group for old animals was significantly higher compared to young animals (P < 0.01).

Effect of paraxanthine, caffeine and age on neurochemicals

Neurochemicals

BDNF, catalase, and cyclic GMP levels were higher in animals supplemented with PXN HIGH and CAF HIGH compared to PXN LOW and control (group effect and pairwise comparisons; all P < 0.001; Fig. 2A-C). PXN LOW also increased BDNF and catalase compared to control (group effect and pairwise comparisons; all P < 0.001). Additionally, PXN HIGH further augmented BDNF (Fig. 2A) to a greater extent compared to CAF HIGH (pairwise comparison; P = 0.03).

Fig. 2.

The impact of age and supplementation on learning and memory assessed as highlighted by reductions in escape latency. *Different from CON within same age group (P < 0.049). #different from PXN LOW within same age group (P < 0.001). ‡Different from CAF within same age group (P < 0.01). P-values: Group<0.001; Age<0.001; Interaction<0.001. Data presented as mean ± SD for n = 8 in each group. CON, control; PXN LOW, low dose paraxanthine; PXN HIGH, high dose paraxanthine; CAF HIGH, high dose caffeine.

Both PXN HIGH and CAF HIGH lowered β-amyloid levels (Fig. 2D) compared to PXN LOW and control (group effect and pairwise comparisons; all P < 0.001). Finally, except for β-amyloid, young animals had higher levels of all neuromodulators compared to old animals (age effects; all P < 0.001).

Neurotransmitters

Levels of acetylcholine, dopamine, GABA, and glutathione were higher in animals supplemented with PXN HIGH and CAF HIGH compared to PXN LOW and control (group effect and pairwise comparisons; all P < 0.001; Fig. 3A-D). In both young and old animals, glutathione (Fig. 3D) was elevated in the PXN HIGH and CAF HIGH groups compared to PXN LOW and control (group × age interaction and pairwise comparisons; P < 0.001). Additionally, in old animals, glutathione levels were greater in PXN LOW compared to control (P = 0.02). Young animals supplemented with PXN HIGH and CAF HIGH had higher glutathione levels compared to old animals (P < 0.01). Finally, young animals exhibited higher levels of all neurotransmitters compared to old animals (group × age interaction and pairwise comparisons; P < 0.001).

Fig. 3.

The impact of age and supplementation on BDNF (A), catalase (B), cyclic GMP (C) and β-amyloid (D). Letters that are different indicate significantly different group values based on pairwise comparisons following significant group effects (P < 0.001). P-values: Group all < 0.001; Age all < 0.001; Interaction < 0.59. Data presented as mean ± SD for n = 8 in each group. BDNF, brain-derived neurotrophic factor; CON, control; PXN LOW, low dose paraxanthine; PXN HIGH, high dose paraxanthine; CAF HIGH, high dose caffeine.

Discussion

We aimed to investigate the effects of paraxanthine and caffeine on learning and memory, as well as to assess neurochemical changes in young and old animals. Our findings show for the first time that both paraxanthine and caffeine improve learning and memory in these age groups. Specifically, in young animals, all three treatments—low-dose paraxanthine, high-dose paraxanthine, and caffeine—enhanced learning and memory compared to control. Furthermore, high-dose paraxanthine was found to be more effective than caffeine in improving these cognitive functions in young animals. In old animals, both high-dose paraxanthine and caffeine significantly reduced escape latency compared to control, while a trend was observed for low-dose paraxanthine. Finally, high-dose paraxanthine elevated BDNF levels to a greater extent than caffeine, regardless of age.

Caffeine is a stimulant that, when consumed acutely in high doses, can enhance brain function, and may re-optimize it with chronic moderate consumption, as shown in multi-omic measurements in both rodents (Paiva et al. 2022) and humans (Picó-Pérez et al. 2023). Its primary metabolite, paraxanthine, has been shown to possess similar (Benowitz et al. 1995) or even superior (Orrú et al. 2013) psychoactive properties compared to caffeine. We recently demonstrated that in young adults, 200 mg of paraxanthine, compared to a placebo, can acutely affect short-term memory, reasoning, and response times to cognitive challenges (Yoo et al. 2021), and may serve as an effective nootropic nutrient at an acute dose as low as 50 mg (Xing et al. 2021). In the present study, we aimed to determine if these acute nootropic effects of paraxanthine in humans translate into chronic learning and memory improvements—by directly comparing it with caffeine—while also examining the impact of age and associated biomarker levels. In humans, voluntary caffeine ingestion can occur multiple times daily, whereas in this study, animals were administered caffeine and paraxanthine once a day to assess the effects of a single dose.

Our findings indicate that both high-dose paraxanthine and caffeine supplementation improved learning and memory in both young and old animals, as evidenced by reductions in escape latency. Memory improvements were more pronounced in young animals supplemented with high-dose paraxanthine compared to caffeine. BDNF, a neurotrophin involved in neuronal plasticity (Leal et al. 2017), is essential for learning and memory processes (Radecki et al. 2005). The increased levels of BDNF in animals supplemented with high-dose paraxanthine and caffeine suggest that these compounds may mediate their positive effects on memory. Our results align with previous studies showing that caffeine increases BDNF expression in the hippocampus (Sallaberry et al. 2013), a brain region crucial for learning and memory (Montkowski and Holsboer 1997; Petzold et al. 2015), as well as in a mouse model of Alzheimer’s Disease (Han et al. 2013). Although our measurements were not region-specific, high-dose paraxanthine augmented BDNF levels more than caffeine, suggesting a dose-dependent effect of paraxanthine, regardless of age. Thus, our data indicate that paraxanthine increases BDNF levels to a greater extent than caffeine, which may underlie its cognitive-enhancing effects in both young and old animals. The observed changes in BDNF may explain why paraxanthine showed reduced errors in cognitive flexibility tests (Berg-Wisconsin Card Sorting Test) compared to caffeine (Yoo et al. 2024).

It is well established that normal aging is associated with declines in learning and memory (Bettio et al. 2017). Our findings indicate that both paraxanthine and caffeine significantly influence memory, independent of age. In young animals, the greatest improvements were observed with high-dose paraxanthine (PXN HIGH), outperforming all other groups (Fig. 4). In older animals, both high-dose paraxanthine and caffeine (CAF HIGH) produced the most significant improvements, consistent with previous research showing age-related declines in learning and memory acquisition in water maze tests (Shukitt-Hale et al. 2004; de Fiebre et al. 2006). As expected, age significantly affected the levels of all neuromodulators, with young animals exhibiting higher levels than older ones. This aligns with previous studies indicating that aging is associated with decreased levels of BDNF (Lommatzsch et al. 2005; Ziegenhorn et al. 2007), catalase (Tsay et al. 2000), and cyclic GMP (Kelly 2018; Domek-Łopacińska and Strosznajder 2010). Importantly, both PXN HIGH and CAF HIGH were found to increase these neuromodulator levels in both age groups.

Fig. 4.

The impact of age and supplementation on (A) acetylcholine, dopamine (B), GABA (C), glutathione (D). Letters that are different indicate significantly different group values based on pairwise comparisons following significant group effects (P < 0.001).*Different from CON within same age group (P < 0.02). #different from PXN LOW within same group (P < 0.001). †YOUNG different from OLD within same treatment group (P < 0.01). P-values: Group all  < 0.001;; Age all  < 0.001; (A-C) Interaction<0.87, (D) Interaction<0.001. Data presented as mean ± SD for n = 8 in each group. CON, control; PXN LOW, low dose paraxanthine; PXN HIGH, high dose paraxanthine; CAF HIGH, high dose caffeine

Additionally, the reduction in β-amyloid (1–40) levels in animals supplemented with PXN HIGH and CAF HIGH suggests a potential mechanism for the cognitive benefits observed. While our study assessed β-amyloid in the brain, other research has linked β-amyloid accumulation to cognitive impairment and dementia (Näslund et al. 1994; Rodrigue et al. 2009) and shown that high doses of caffeine (e.g., 500 mg) can attenuate these levels (Arendash et al. 2006). Conversely, some studies found that four months of caffeine intake did not significantly influence β-amyloid levels in a transgenic mouse model of Alzheimer’s disease (Stazi et al. 2021). The age-related decline in neuromodulator levels, combined with the lack of significant learning and memory effects from PXN LOW in aged animals (though a trend was noted, P = 0.07), may explain the reduced cognitive benefits of both caffeine and paraxanthine in older individuals. Overall, these findings suggest that caffeine and paraxanthine hold therapeutic potential for improving cognitive function, particularly in younger individuals, by enhancing neuromodulation and reducing β-amyloid accumulation.

The present study revealed that both high-dose paraxanthine (PXN HIGH) and high-dose caffeine (CAF HIGH) significantly increased neurotransmitters involved in learning and memory, including acetylcholine, dopamine, and GABA. Acetylcholine is known to be critical for cognitive processes (Hasselmo 2006), and caffeine has been shown to enhance acetylcholine release (Carter et al. 1995) and increase cholinergic neurotransmission (Fabiani et al. 2018). Additionally, dopamine plays a key role in the formation of long-term memories (Rossato et al. 2009; Yamagata et al. 2015), while GABA is involved in the consolidation of these memories (Makkar et al. 2010; Cullen et al. 2014). The modulation of these neurotransmitters by caffeine and paraxanthine may contribute to the observed cognitive improvements. Furthermore, glutathione—an important antioxidant that maintains cellular redox balance (Kalinina et al. 2014) and protects against oxidative damage (Rose et al. 2012)—was found to be elevated in animals supplemented with high doses of paraxanthine and caffeine, irrespective of age. These findings align with previous research indicating that caffeine and its derivatives, including paraxanthine (Matsumura et al. 2023), exert neuroprotective effects by upregulating antioxidant enzymes (Devasagayam et al. 1996) and mitigating oxidative stress-induced cell death (Barcelos et al. 2014).

Notably, the age-related differences in glutathione levels highlight the need for further research to better understand the relationship between age and the effects of paraxanthine and caffeine on glutathione levels and cognition.

Demethylation of caffeine results in the formation of theophylline (1,3-dimethylxanthine), theobromine (3,7-dimethylxanthine), and paraxanthine. The half-life of paraxanthine is approximately 3.1 h, which is shorter than that of caffeine (4.1 h) but significantly different from theophylline (6.2 h) and theobromine (7.2 h) (Nehlig et al., 2018; Lelo et al. 1986). Paraxanthine exhibits lower toxicity than caffeine and is less clastogenic compared to caffeine or theophylline. It is also less harmful in terms of hepatocyte toxicity, less potent as a teratogen than caffeine and theophylline, and less anxiogenic than caffeine (Purpura et al. 2021). Theophylline and theobromine have distinct physiological effects. Theophylline is used as a bronchodilator in asthma treatment (Barnes 2010) and is a more potent adenosine receptor antagonist than caffeine in vitro (Bruns et al. 1983). However, in healthy humans, theophylline does not enhance cognitive function (Fitzpatrick et al. 1992). Theobromine, found in high concentrations in cocoa (Cova et al. 2019), promotes vasodilation and smooth muscle relaxation (Mitchell et al. 2011). Yet, compared to caffeine, clinical studies suggest that theobromine does not significantly influence cognitive processes, such as mood and alertness (Judelson et al. 2013; Mitchell et al. 2011). In summary, while structurally similar dimethylxanthines have different physiological effects, the impact of theophylline and theobromine on cognition is generally less pronounced than that of caffeine and paraxanthine. Direct comparisons between caffeine and its metabolites regarding cognitive effects remain an important area for future research.

The limitations of this study include the absence of measurements for swim speed and distance, as variations in locomotion could influence differences in escape latency. Additionally, we did not assess anxiety levels across the experimental groups. Anxiety is known to impact learning and memory behaviors, and its effects may differ between caffeine and paraxanthine. Moreover, previous research in both mice and rats suggests that the effects of caffeine are more pronounced in preventing memory deterioration rather than enhancing memory in rodents not exposed to potentially harmful stimuli (Cunha et al., 2016). Our findings indicate that control adult rats demonstrated improved learning and memory after caffeine and paraxanthine administration, which may suggest a heightened stress level among these animals. Therefore, future studies should incorporate assessments of anxiety and locomotion to provide a more comprehensive understanding of the effects of these compounds.

Conclusions

In conclusion, our study provides novel evidence that paraxanthine and caffeine supplementation can enhance learning and memory in both young and old animals, with particularly pronounced effects observed in young animals receiving high-dose paraxanthine. Additionally, paraxanthine may boost cognitive function across age groups by increasing BDNF levels. The observed reduction in β-amyloid levels in animals supplemented with high doses of paraxanthine and caffeine offers further mechanistic insights into their cognitive benefits. The age-related decline in neuromodulators, neurotransmitters, and glutathione levels may account for the diminished cognitive benefits of caffeine and paraxanthine in older animals. These findings underscore the therapeutic potential of paraxanthine for improving cognitive function, especially in younger individuals. Future research is needed to better understand the relationship between age and the effects of paraxanthine and caffeine on cognitive function.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: R.J., M.O., M.P., A.G., S.D.W. and K.L.; project management: A.G.; writing—original draft preparation: S.A., and R.J.; writing—review and editing: S.A., M.O., R.J., G.M.T., M.P., S.D.W., K.L. and A.G. The study was conducted by Radiant Research Services Pvt. Ltd. (Bangalore, India). Statistical analysis was performed by G.M.T. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by Ingenious Ingredients, L.P. (Lewisville, TX, USA).

Data availability

The data generated and analyzed for this current study are available from the corresponding author on reasonable request.

Declarations

Conflict of interest

R.J., M.P., S.D.W. and K.L. are researchers and principals of Ingenious Ingredients, the sponsor of the study, and inventors of numerous patent applications for the use of paraxanthine but have been involved in data collection or analysis. R.J., M.P. and S.D.W. serve on the scientific advisory board and K.L. is the CEO of NNB Nutrition, the manufacturer of the paraxanthine used in this study. S.A. is an employee of Iovate Health Sciences, a distributor of paraxanthine, but has not been involved in data collection or analysis. A.G. is employed by Radiant Research Services Pvt. Ltd., and Radiant Research is a CRO who has received funding from Ingenious Ingredients to conduct this study. G.M.T. provides consulting services, including statistical analysis and dietary supplement formulation, through Tinsley Consulting LLC but has no conflicts of interest related to the use of paraxanthine. M.O. declares no competing interests.