Central fatigue, characterized by a reduced ability to maintain cognitive and physical performance, stems from intricate neurobiological (Gandevia, 2001; Davis et al., 2003; Meeusen et al., 2006; Aguiar et al., 2020) and metabolic interactions (Glaister, 2005; Wan et al., 2017). Studies linking aerobic performance, oxidative metabolism, and fatigue have shown that blocking brain adenosine receptors increases resistance to physical fatigue (Davis et al., 2003; Aguiar et al., 2020), highlighting the significant role of adenosine in fatigue signaling (Aguiar et al., 2021). Moreover, by integrating pharmacological research with knockout mouse models, we have previously demonstrated that adenosine A_2A receptors (A_2AR) are crucial for the performance-enhancing effects of caffeine (Aguiar et al., 2020). However, non-toxic doses of caffeine selectively antagonize both adenosine A_2A and A₁ receptors to format information flow within brain neuronal circuits (Lopes et al., 2019) and it remains to be clarified if A₁R also play a role in controlling fatigue.

Additionally, we identified a specific role of A_2AR in the striatum for controlling central fatigue (de Bem Alves et al., 2023). Notably, A_2AR interacts closely with dopamine D₂ receptors (D2R) in this region, forming A_2AR-D₂R heteromers (Ferré et al., 2008; Ferré and Ciruela, 2019), a concept pioneered by Kell Fuxe in the early 90’s (Ferré et al., 1991). These interactions are significant in various conditions, including Parkinson’s disease, schizophrenia, substance abuse, and Attention Deficit Hyperactivity Disorder (ADHD), as reviewed in the literature (Fuxe et al., 2003; Ballesteros-Yáñez et al., 2018; Chen et al., 2023). Consequently, the antagonism of D₂R by haloperidol affects the efficacy of A_2AR antagonists in modulating effort-related behaviors (Salamone et al., 2009; Pardo et al., 2012; Rotolo et al., 2023). While A_2AR-D₂R interactions have been observed in other areas of the brain (Pandolfo et al., 2013; Dremencov et al., 2017; Real et al., 2018), their functional significance remains less understood.

This study aimed to evaluate the potential ergogenic effects of selective antagonists for A₁R and A_2AR, along with the non-selective adenosine receptor antagonist caffeine. Our objective was to determine whether and how these effects are altered by a D₂R antagonist, particularly in the striatum, an area previously shown to influence central fatigue (Davis et al., 2003; Aguiar et al., 2020; de Bem Alves et al., 2023). Additionally, we venture into new territory by examining the impact of adenosine receptors on exercise physiology, specifically focusing on strength performance, which is predominantly dependent on anaerobic metabolism. This approach allows us to expand our understanding of fatigue mechanisms across various physical activities and metabolic requirements.

Figure 1A displays the dose-response curve of haloperidol in the catalepsy test, which prompted selecting the non-cataleptic dose of 250 μg/kg. As expected, systemic caffeine acted both as a psychostimulant and as an ergogenic agent: it increased open field crossings, an effect mitigated by the non-cataleptic dose of haloperidol (F_3,35 = 4.3, η² = 0.27, β = 0.83, p < 0.05, Figure 1B). In the incremental treadmill test (Figure 1C), caffeine boosted running power, an effect that was not affected by haloperidol (F_3,27 = 3.2, η² = 0.25, β = 0.67, p < 0.05, Figure 1D). In the grip strength test, caffeine enhanced grip time (F_3,35 = 17.5, η² = 0.6, β = 0.99, p < 0.05, Figure 1F) and impulse (F_3,34 = 36.4, η² = 0.76, β = 1.0, p < 0.05, Figure 1H) without increasing peak strength (Figure 1E), and haloperidol mitigated these effects. Additionally, caffeine improved the resistance to fatigue in the grip test, a benefit not present in animals also receiving haloperidol (F_30,350 = 4.7, η² = 0.29, β = 1.0, p < 0.05, Figure 1G).

When applied directly in the striatum, caffeine also triggered a psychostimulant and ergogenic response: it increased the number of open field crossings (F_2,12 = 31.1, η² = 0.83, β = 0.99, p < 0.05, Figure 2A), grip time (F_2,26 = 6.4, η² = 0.055, β = 0.99, p < 0.05, Figure 2C), resistance to fatigue (F_8,104 = 4.0, η² = 0.23, β = 0.98, p < 0.05, Figure 2D) and impulse (F_2,12 = 31.1, η² = 0.83, β = 0.99, p < 0.05, Figure 2E) without increasing peak strength (Figure 2B). Strikingly, although haloperidol mitigated the psychostimulant effect of caffeine in the open field, it did not alter the ergogenic properties of intra-striatal caffeine in the treadmill and grip tests.

Since caffeine non-competitively antagonizes adenosine A₁ and A_2A receptors (Fredholm et al., 1999), we exposed animals to DPCPX and SCH58261, which are selective antagonists for these respective receptors. DPCPX administration did not alter any measured behaviors, whereas SCH58261 increased locomotion in the open field; this effect was nullified in the presence of haloperidol (F_3,36 = 11.4, η² = 0.48, β = 0.99, p < 0.05, Figure 3A). Figure 3B shows the increase in running power with increased speed. SCH58261 also significantly improved running power in the treadmill test, a gain blocked by haloperidol (F_3,20 = 9.8, η² = 0.59, β = 0.99, p < 0.05, Figure 3C). While treatments did not affect peak grip strength (Figure 3D), SCH58261 enhanced both grip time (F_3,36 = 8.3, η² = 0.41, β = 0.98, p < 0.05, Figure 3E) and area under the curve (AUC, impulse) (F_3,36 = 6.3, η² = 0.34, β = 0.94, p < 0.05, Figure 3G), with haloperidol diminishing these effects. Additionally, haloperidol reduced resistance to fatigue in SCH58261-treated animals (F_12,144 = 1.9, η² = 0.14, β = 0.9, p < 0.05, Figure 3F).

Lastly, we assessed the intra-striatal effects of SCH58261, foregoing DPCPX due to its lack of systemic effects. Intra-striatal administration of SCH58261 increased locomotion in the open field, an effect that was mitigated by haloperidol (F_2,12 = 37.7, η² = 0.86, β = 1.00, p < 0.05, Figure 4A). While intra-striatal SCH58261 did not affect peak grip strength (Figure 4B), it prolonged grip time (F_2,27 = 5.5, η² = 0.29, β = 0.80, p < 0.05, Figure 4C) and enhanced impulse (F_2,27 = 5.5, η² = 0.28, β = 0.80, p < 0.05, Figure 4E). All animals exhibited a decline in strength over time, with no significant differences between treatments observed (F_6,81 = 2.7, η² = 0.16, β = 0.84, p < 0.05, Figure 4D). Notably, alterations induced by intra-striatal SCH58261 in grip strength were not modified by haloperidol.

All effect sizes were large, except for the resistance to fatigue comparisons following systemic treatment with DPCPX or SCH58261 (Figure 3F), and intra-striatal treatment with caffeine (Figure 2C), which exhibited medium effect sizes. Generally, the statistical power for our analyses was high, exceeding 80%. The only exception was the systemic treatment with caffeine and haloperidol (Figure 1D), which had a statistical power of 67%; however, this did not significantly impact our overall conclusions.

The present study confirmed that the systemic administration of either caffeine and of the selective antagonist of A_2AR, SCH58261 were ergogenic and that these effects were largely reproduced by the direct injection of caffeine or SCH58261 in the striatum, aligning with previous findings (de Bem Alves et al., 2023) and the general role of striatal A_2AR in effort-related behaviors, reviewed in Chen et al. (2023).

The first novelty provided by this study is the re-enforcement of our previous contention that A_2AR are the likely molecular targets operated by caffeine to produce its ergogenic effects (Aguiar et al., 2020). Physiological concentrations of caffeine, achieved in the brain parenchyma after consuming non-toxic doses, influence neuronal network information flow primarily through A₁R and A_2AR antagonism (Lopes et al., 2019), without involving other mechanisms such as phosphodiesterases or ryanodine receptors, which are activated by toxic caffeine doses (Fredholm et al., 1999). Our findings that caffeine’s ergogenic effects are replicated by the selective A_2AR antagonist SCH58261, but not by the A₁R antagonist DPCPX, provide direct evidence that caffeine’s ergogenic effects are specifically mediated through A_2AR rather than A₁R. Additionally, caffeine’s influence on other behavioral responses that rely on striatal information processing, such as locomotion, arousal, or response to psychostimulants, is also mediated by A_2AR rather than A₁R (El Yacoubi et al., 2000a; Chen et al., 2000). Therefore, the specific involvement of A_2AR, rather than A₁R, observed in the acute effects of caffeine on exercise performance, further supports the significant role of striatal A_2AR in governing caffeine’s ergogenic effects (Aguiar et al., 2020). This is especially pertinent given the high density of A_2AR in the striatum compared to other brain regions, which predominantly exhibit A₁R-mediated effects following acute caffeine exposure (Elmenhorst et al., 2012; Kerkhofs et al., 2018).

Another significant conclusion drawn from this study pertains to the interaction between A_2AR and D₂R in managing exercise endurance. This presumed interplay between A_2AR and D₂R is supported by: i) their co-localization within medium spiny neurons of the striatum (Svenningsson et al., 1997; Canals et al., 2003), a region recognized for the ergogenic effects prompted by caffeine blockade (de Bem Alves et al., 2023); ii) their established functional crosstalk (Hillion et al., 2002; Ferré et al., 2016); iii) their dynamic heteromerization, which results in new receptosomes with novel properties (Canals et al., 2003; Fernández-Dueñas et al., 2015); iv) their interplay in controlling motor activity, the impact of psychostimulants, and effort-related behavioral performance (e.g., Pardo et al., 2013; reviewed in Borroto-Escuela et al., 2018; Nunes et al., 2010; Rotolo et al., 2023). Additionally, dopamine signaling has been experimentally linked to exercise and fatigue (reviewed in Meeusen et al., 2021). Exercise induces a hyperdopaminergic state (Greenwood, 2019) and dopamine depletion correlates with mental fatigue (Moeller et al., 2012; Aguiar et al., 2016; Scheffer et al., 2021). Furthermore, polymorphisms within various components of the dopaminergic system have been associated with fatigue (Malyuchenko et al., 2010), L-DOPA has been shown to alleviate physical fatigue (Lou et al., 2003; Scheffer et al., 2021; de Bem Alves and Aguiar, 2024), and exposure to reserpine, which depletes monoamines, serves as an experimental fatigue model (Scheffer et al., 2021; de Bem Alves and Aguiar, 2024).

Despite these findings, the specific dopamine receptors involved in the control of fatigue and their brain locations remain unclear. The traditional view of striatal dopaminergic signaling suggests a primary role for D₁R, but the simplistic dichotomy of behaviors influenced by D₁R and D₂R in medium spiny neurons is being reevaluated (Soares-Cunha et al., 2016). The improvement of performance with increased forced running is linked to dopamine and specifically to the activation of striatal D₁R and extrastriatal D₂R (Toval et al., 2021). Additionally, the dorsolateral prefrontal cortex has been proposed as important during exhaustive exercise (Bigliassi and Filho, 2022). These insights align with our unexpected findings: i) haloperidol reduces the ergogenic effects of systemically administered caffeine or SCH58261; ii) notably, haloperidol does not affect the ergogenic effects when caffeine or SCH58261 is applied directly to the striatum.

These results lead us to conclude that A_2AR-D₂R interactions play a significant role in regulating exercise endurance and fatigue, predominantly outside the striatum. This suggests that adenosine’s central role in exercise endurance and fatigue might involve primarily changes in striatal circuits, but also modifications in other brain areas. This concept aligns with the noted involvement of striatal and extrastriatal circuits in other behaviors affected by A_2AR-D₂R interplay, such as responses to psychostimulants (reviewed in Borroto-Escuela et al., 2018; Chen et al., 2023). However, this conclusion that the extra-striatal A_2AR-D₂R interplay is involved in ergogenic effects should still be regarded as preliminary since it will require confirmation based on direct pharmacological and genetic manipulations of A_2AR and D₂R only in the striatum and the identification of the brain regions where this interaction takes place to control fatigue. Additionally, future studies should also investigate a putative role of dopamine D₁R, which is also abundantly present in the striatum and other brain regions and has also been reported to tightly interact with adenosine A₁R (Ginés et al., 2000) and dopamine D₂R (Rashid et al., 2007; Frederick et al., 2015).

Interestingly, our current findings indicate that the ergogenic effects of caffeine and SCH58261 are distinct from their psychostimulant properties. This observation is supported by our previous research showing that SCH58261 has both psychostimulant and ergogenic effects in male mice, but only ergogenic effects in females, without clear psychostimulant impacts (Aguiar et al., 2020). This adds to the growing body of evidence suggesting a unique role for various A_2AR populations in different cellular locations in modulating diverse behavioral responses (Shen et al., 2008; Yu et al., 2008; Wei et al., 2014).

In summary, our study confirms the critical role of striatal A_2AR in the adenosine modulation of central fatigue and in the ergogenic responses of A_2AR antagonists. Furthermore, we found a dissociation between the psychostimulant and ergogenic effects of caffeine and SCH58261. In parallel, the ergogenic effects of caffeine and of SCH568261 were controlled by D₂R blockade with haloperidol, but this A_2AR-D₂R interplay seems to mostly occurs in extra-striatal circuits. These findings underscore the need for further investigation into the intricate interplay between adenosine and dopamine signaling in striatal and extra-striatal circuits to control central fatigue and other motivational behaviors such as the response to psychostimulants.