During the last couple of decades, there has been an increasing supply of substituted cathinones in the illegal drug market (Vardakou et al., 2011; Deluca et al., 2012; EMCDDA, 2018). Generally, these substances initially gain popularity as legal alternatives to existing amphetamine-type stimulants. As their popularity grows, they are subsequently banned by national governments and once banned, they become available on the illegal drug market next to already controlled substances. One of the better-known substituted cathinones is mephedrone (4-methylmethcathinone, 4-MMC), the usage of which has remained high even after its classification as illegal (Deluca et al., 2012; Papaseit et al., 2020).

4-MMC shows many pharmacological similarities with methamphetamine (METH), with both substances producing rapid and substantial increases in accumbal extracellular dopamine (DA) and serotonin levels in vivo (Kehr et al., 2011). The effects of both METH and 4-MMC are mainly mediated via increased release and blockade of reuptake of DA (Cho and Melega, 2002; Winstock et al., 2011b; Hadlock et al., 2011; Lopez-Arnau et al., 2012; Martinez-Clemente et al., 2012; Korpi et al., 2015). However, long-term neurotoxic effects of mephedrone and METH appear to differ substantially. While METH is known to produce substantial and long-lasting reductions in monoamine levels and other markers of DA neurotoxicity, such findings are generally not replicated with 4-MMC. In fact, several studies have found little evidence of mephedrone neurotoxicity when administered under normal conditions (Angoa-Perez et al., 2012; Baumann et al., 2012; Angoa-Perez et al., 2013; Angoa-Perez et al., 2014; den Hollander et al., 2015; Ciudad-Roberts et al., 2016). Nonetheless, 4-MMC neurotoxicity can be precipitated when the drug is administered under circumstances known to exacerbate stimulant neurotoxicity, such as high ambient temperatures (Hadlock et al., 2011; Lopez-Arnau et al., 2015; Pantano et al., 2017).

In recreational settings, stimulants such as 4-MMC and METH are often used in conjunction with alcohol (ethanol, EtOH) (Winstock et al., 2011a; Papaseit et al., 2020). Alcohol use is a known exacerbating factor for stimulant neurotoxicity and has been shown to increase or precipitate neurotoxic effects of MDMA and METH (Izco et al., 2007; Rodriguez-Arias et al., 2011; Blaker and Yamamoto, 2018; Blaker et al., 2019). This raises the question whether alcohol could also precipitate 4-MMC neurotoxicity. Currently, very little is known about this. One study reported evidence of neurotoxicity when 4-MMC was co-administered with EtOH, but this study also administered the drug at high ambient temperatures, making it difficult to conclusively attribute the effects to EtOH (Ciudad-Roberts et al., 2016). Another study in mice reported EtOH increased 4-MMC-induced conditioned place preference but did not assess toxicity (Ciudad-Roberts et al., 2015). Interestingly, the acute effects of 4-MMC alone or in combination with EtOH was investigated in a clinical study which demonstrated that EtOH increased the cardiovascular effects of the drug as well as self-reported euphoria. However, this study did not investigate long-term neurocognitive effects (Papaseit et al., 2020). Further investigating this question is of importance considering that 4-MMC is often consumed together with EtOH and will help inform both drug users and healthcare providers of potential risks and help guide the development of an evidence-based harm reduction approach.

We previously used manganese-enhanced magnetic resonance imaging (MEMRI) to assess the long-term effects of 4-MMC and METH in rats and showed that METH produced a pattern of widespread neuronal deactivation in monoamine-rich brain areas 2 weeks following a binge-dosing regimen; 4-MMC, conversely, produced an effect that was limited primarily to the parietal cortex, hypothalamus and hippocampus and was characterized by neuronal activation rather than deactivation (den Hollander et al., 2015). MEMRI signal suppression has previously been associated with neurotoxicity in a study where MPTP-induced loss of DA neurons in substantia nigra produced decreased neuronal activation by MEMRI (Weng et al., 2016). The MEMRI method is based on the fact that manganese functions as a Ca²⁺ analogue in vivo; manganese influx into neuronal cells via the Ca²⁺ channels during fast neuronal depolarization and subsequent distribution throughout the neuron and axons thereby represent a measure of mean neuronal activity over time (Bedenk et al., 2018; Svehla et al., 2018).

Here, we employed MEMRI to investigate the long-term effects of EtOH co-administration with 4-MMC or METH on neuronal activity. Additionally, we used Golgi staining to evaluate potential changes in neuronal morphology in regions of interest identified based on MEMRI neuronal activity patterns. We show that EtOH modifies the long-term effects of 4-MMC on neuronal activity, resulting in a similar pattern of deactivation as observed after METH administration.

Here we show that co-administration of EtOH with 4-MMC in a binge-abuse model produces a widespread pattern of neuronal deactivation in monoamine-rich brain areas when assessed with MEMRI two weeks after the last drug administration. The 4-MMC + EtOH deactivation pattern was very similar to the one observed following METH administration. This pattern of widespread deactivation is in contrast with the effect seen when 4-MMC was administered alone. 4-MMC increased, rather than decreased, neuronal activity in an anatomically more limited manner, and occurring in areas not strongly innervated by DA axons. These findings regarding 4-MMC alone were in line with our previous results, indicating good reproducibility of the MEMRI method (den Hollander et al., 2015). However, when examining neuronal morphology in the hippocampus, nucleus accumbens and caudate-putamen, the three regions showing the strongest neuronal deactivation in animals treated with 4-MMC and EtOH, we found no morphological changes to dendrites suggestive of long-term neurotoxic effects.

When administered alone, 4-MMC produces a subtle pattern of long-term neuronal activation in regions like the retrosplenial and primary somatosensory cortex, as well as the dorsal hippocampus, thalamic areas and the superior colliculus. These areas are not strongly innervated by DA, and the observed changes plausibly do not reflect DA neurotoxicity. A potential explanation for the increased neuronal activity is that during 4-MMC intoxication, it has been shown that plasma cortisol levels are elevated (Papaseit et al., 2020). Elevation of cortisol levels in acute stress via the serotonergic effects on the hypothalamo-pituitary-adrenal axis has been shown to activate regions important for memory consolidation, like the hippocampus (Herman and Cullinan, 1997; Kim et al., 2015). Interesting in this regard is that 4-MMC was shown to reduce memory performance 2 weeks after a similar binge-regimen as employed in this study without causing any reductions in monoamine levels that could be indicative of toxicity (den Hollander et al., 2013).

The pattern of widespread neuronal deactivation seen following treatment with 4-MMC in combination with EtOH as well as following METH, either alone or in combination with EtOH, is striking considering the anatomical correlation of the deactivation pattern with brain regions innervated by DA nerve ending and typically implicated in neurotoxicity of amphetamines (Ares-Santos et al., 2014). Nonetheless, even the areas showing the strongest deactivation did not display signs of neurotoxicity as assessed by subsequent Golgi staining and dendritic spine analysis. This raises the question what the exact neurochemical correlates of the MEMRI deactivation pattern are. Co-administration of 4-MMC with EtOH lowers synaptic transmitter function markers for DA and 5-HT, while elevating oxidative stress markers at least in high ambient temperatures (Ciudad-Roberts et al., 2016). It is thus possible that the 4-MMC and EtOH-induced deactivation observed here reflects the long-term functional effects of increased oxidative stress-related adaptation in synaptic terminals. However, it should be noted that deactivation in brain regions largely overlapping with those in the 4-MMC/EtOH-treated group were observed in the group treated with EtOH only. Considering that the EtOH dose employed in this study was moderate it is worth noting that neuronal deactivation as assessed by MEMRI should not be considered as conclusive evidence of neurotoxicity, although such a correlation has been observed previously (Weng et al., 2016). Similarly, neurotoxicity may have occurred in the 4-MMC group in a way that is not reflected in the MEMRI data. The exact neurochemical correlates of altered activation patterns remain to be determined and until then the data must be interpreted with relative caution.

Despite these caveats, our results appear in agreement with previous studies that have generally failed to observe neurotoxicity following administration of 4-MMC at normal ambient temperatures (Angoa-Perez et al., 2012; Baumann et al., 2012; Angoa-Perez et al., 2013; Angoa-Perez et al., 2014; den Hollander et al., 2015; Ciudad-Roberts et al., 2016). Although there are some reports of neurotoxicity as assessed by levels of monoamines, their transporters and tyrosine hydroxylase (Martinez-Clemente et al., 2014; Kaminska et al., 2018), 4-MMC neurotoxicity is primarily observed when administered under exacerbating conditions, specifically increased ambient temperatures (Hadlock et al., 2011; Lopez-Arnau et al., 2015; Pantano et al., 2017). This raises the question as to what pharmacological differences between 4-MMC and METH result in the latter much more readily inducing neurotoxicity in animal models. Anneken et al. (2018) hypothesized that the proposed difference could be due to differences in how METH and 4-MMC access the drug-releasable pool of DA. However, their experiments showed that combining the drugs with l-DOPA, which increases the size of the release pool, did in fact not precipitate any 4-MMC neurotoxicity while enhancing the toxic effects of METH. In a subsequent experiment, the same group investigated whether differences in how the two drugs modulate core body temperatures may be responsible for differences in toxicity using tryptophan hydroxylase 2 knockout mice lacking brain serotonin. The rationale for this experiment was the observation that while both METH and 4-MMC produce hyperthermia, 4-MMC has also been reported to produce a brief hypothermic response. However, they reported that although the knockout mice did not experience the characteristics hypothermic response, no increase in toxicity was observed in the knockouts vs. the wild-types (Anneken et al., 2019). Another possibility is that 4-MMC has a different effect of mitochondria than METH, which has been associated with mitochondrial-dependent mechanisms of toxicity (Shin et al., 2018). However, 4-MMC in fact appears to also affect mitochondria. It has been shown to impair mitochondrial complexes II and IV, collapse mitochondrial membrane potential, induce mitochondrial swelling and lower mitochondrial respiration in vitro (den Hollander et al., 2014; Naserzadeh et al., 2019). Nonetheless, in the case of 4-MMC, these in vitro results do not generally appear to translate into measurable neurotoxicity in vivo. Thus, the exact mechanism(s) underlying the observed differences in toxicity between 4-MMC and METH remain to be elucidated.

In this study, core body temperatures vs. the control group were lowered after EtOH but increased after 4-MMC and METH, in line with results from previous studies (Lomax et al., 1980; Myles and Sabol, 2008; den Hollander et al., 2015). Notably, co-administration of EtOH with 4-MMC and METH neither exacerbated nor blocked the hypothermic effects of these drugs. Since hyperthermia was seen in stimulant-treated groups and not affected by EtOH co-administration, any differences observed between groups were not due to a differing hyperthermic reaction in this study.

Golgi staining followed by spine analysis has been suggested to be an effective and unbiased way to assess neuronal changes at the dendritic spine level (Risher et al., 2014). This has also proven to be the case with stimulant-induced changes in dendritic spines (Robinson and Kolb, 1997; Robinson and Kolb, 1999). However, in the present experiment, no statistically significant alterations were found in the spine analysis in three relevant brain regions. This lack of changes could be due to the limitations in Golgi staining, such as variability of staining intensity, variability in localizing the brain area within thick brain sections or the random nature at which neurons of different types are stained which is inherent to the Golgi method. Stimulant-induced increases in striatal spines have been found in studies in which prolonged psychomotor sensitization has been detected (Li et al., 2004), but our slightly heavier regimen was aimed at finding possible neurotoxicity without sensitization (den Hollander et al., 2015). Another aspect to be taken into account is the fact that changes in spine formation, deformation and maturation can be very fast and bidirectional (Nagerl et al., 2004; Berry and Nedivi, 2017). To our knowledge there has been no previous research to assess the temporal stability of stimulant-induced alterations in spine morphology. It is possible that stimulant-induced alterations in spine morphology are not stable enough to persist following the 2-weeks recovery period after treatments. Furthermore, Golgi-staining stains post-synaptic parts of the synapse whereas the stimulant-induced neurotoxicity has previously been associated with the axonal part of the synapse but at higher doses, from 20 mg/kg and up (Zhu et al., 2006) than used in this experiment. And therefore, our dose of 5 mg/kg METH might not have been sufficient to cause axonal apoptosis which would then cause dendritic spine deformation. Lastly, it is conceivable that the isoflurane anesthesia employed during the MEMRI scanning could have produced alterations or reversals in morphology considering the fact that isoflurane has previously been associated with activation of neurotrophic signaling, but not with spine structure changes (Antila et al., 2017).

A limitation of this study is that only Golgi staining was employed to interpret the neuronal activation patterns observed in MEMRI experiment. The assessments of levels of monoamines or their transporters, as well as other assessments of inflammation and oxidative stress during and following the binge regimen could have provided further insights into potential causative factors. Also, other type of silver staining or terminal deoxynucleotidyl transferase dUTP nick end labeling could be used to assess cell-level neurotoxicity. Moreover, assessments of neurocognitive and neuropsychiatric function such as memory, mood and anxiety tests, could have provided further information about functional effects. These are aspects which will need to be addressed in future studies.

When administered on its own, the effects of 4-MMC on neural activity are subtle, anatomically limited, and likely not reflective of long-term neurochemical perturbations. However, when administered in combination with EtOH, it produces a pattern of widespread neuronal deactivation in brain regions densely innervated by DA neurons similar to what is seen after METH. The deactivated regions nonetheless do not show clear morphological changes indicative of toxicity. This suggests the deactivation pattern is likely due to long-term neurochemical perturbations that are substantial but not so severe as to produce overt neurotoxicity.

In conclusion, this study suggests that binge-treatment with 4-MMC on its own does not substantially alter brain activity, but in combination with EtOH may result in long-term reductions in brain activity. This is important from a harm-reduction perspective, although further studies are needed to outline the exact neurochemical changes responsible for the altered activity patterns.