In this study, we analysed the evolution of body weight and food/water intake, acute gastrointestinal motor function, early development of pain-related behaviours, and lastly the expression of components of the ECS in gastrointestinal and nervous tissues after treatment with a 5 mg/kg dose in adult male rats. In the clinic, cancer patients are usually treated with repeated administrations of chemotherapeutic drugs, and therapeutic approaches directed to reduce or avoid the appearance of chemotherapy-related side effects should be preventively administered some time before, or shortly after the first (and each) administration of chemotherapy. Our goal was to investigate short-term effects induced by a chemotherapy agent shortly after its first administration and immediately before the second one would be given. To do this, we treated our animals with a dose of cisplatin aimed to induce neuropathic pain [20] and analysed the alterations induced by an acute administration of the drug for a week. Moreover, levels of endocannabinoid ligands and related N-acylethanolamines were measured in plasma to investigate the potential usefulness of these parameters as early biomarkers of chemotherapy-induced toxicity.

Cisplatin is known to exert many different toxic effects [21], some of them caused by many antitumour drugs and others quite specific to this drug. In this study, cisplatin significantly reduced body weight one week after treatment, an effect previously shown after acute and repeated treatment with diverse doses of cisplatin [22,23] and caused by many antitumour drugs [24]. One specific effect induced by cisplatin is renal damage [25,26], but we did not appreciate this toxicity simply from the weight and macroscopic appearance of this organ, nor upon microscopic evaluation, probably because of the short, acute treatment used.

Different mechanisms could underlie the reduction in body weight. First, cisplatin reduced food intake, which may account for the reduced size and weight of the stomach and small intestine and its reduced contents at sacrifice (Figure 2B and Table 2). Second, cisplatin-related mucositis and/or diarrhoea could also decrease body weight by decreasing the absorption of nutrients. Indeed, we have previously reported damage to the gastrointestinal mucosa after a single high dose [27] and repeated treatment with cisplatin [28]. However, diarrhoea did not seem to occur in our animals, since no signs were observed either clinically or on the radiographs obtained at different time-points after administration (Figure 2), which is in accordance with a previous study where a higher single dose (6 mg/kg) of cisplatin was administered to the rats [27]. Furthermore, there was a lack of significant alterations in water intake (Figure 2C), which could have been increased if excessive/sudden loss of water occurred through the faeces. A reduction in body weight could also be due to increased locomotor activity, but this was not observed here four days after cisplatin administration, whereas four weekly administrations of 5 mg/kg induced a significant reduction in horizontal and vertical activity in male Sprague Dawley rats [29]. Thus, the most likely explanation for the reduced body weight is reduced food intake.

Our results show a significant reduction in gastric emptying, in concordance with previous studies from our group, following both acute and repeated administration of cisplatin [22,23,27]. These effects may contribute to cisplatin-induced anorexia [30]. Furthermore, cisplatin administration is usually associated with the development of nausea and vomiting in humans [5,31] and with indirect markers of nausea in animals that cannot vomit, such as pica or gastric distension in rats [2,22,23,32]. Indeed, gastric distension was observed both in this study for a single dose of 5 mg/kg (Figure 2E) and previously when a 6 mg/kg of cisplatin was administered [23,27]. Cisplatin is known to induce a massive release of serotonin from enterochromaffin cells, which activates 5-HT3 receptors on vagal afferent fibres and enteric nerves inducing a delay of gastric emptying and an increase in food retention in the fundus of the stomach [27,33,34]. This explains the efficacy of 5-HT3 antagonists as a first-line treatment of acute chemotherapy-induced nausea and vomiting, particularly when cisplatin is used [35]. However, other neurotransmitters may also participate in the development of gastric dysmotility, nausea, and vomiting, particularly at later times [35]. Interestingly, one week after administration, despite no changes in the levels of 2-AG being observed in the antrum, gene levels of its catabolizing enzyme, i.e., MAGL, were significantly lower in cisplatin-treated rats than their saline counterparts. The discordance between the levels of 2-AG and the gene expression of its catabolic enzyme may be due to the fact that the molecular analysis of gastrointestinal tissue was carried out in whole gut wall samples, while the increase in 2-AG induced by a lower expression of MAGL might be present only in the myenteric plexus, the area intrinsically involved in gastric emptying. On the other hand, it could be that alterations in 2-AG levels have occurred due to rapid metabolic modifications occurring postmortem, as suggested by other authors in brain structures [36,37]. Whatever the case may be, this lower expression of MAGL may favour an increase in CB₁ receptor signalling in the antrum and a subsequent decrease in gastric emptying [38,39].

Cisplatin did not alter the transit of the small intestine and caecum during the X-ray session. This agrees with our previous studies showing that acute treatment with 2 mg/kg of cisplatin failed to alter small intestine motility [28]. However, after a weekly administration of this dose of cisplatin (cumulative dose of 10 mg/kg), we observed a slower small intestinal transit, accompanied by important histological and cellular alterations [28], as well as the development of an enteric neuropathy, also observed in the large intestine [28,40]. Interestingly, we also observed a significant increase in the levels of AEA in the ileum of animals treated with cisplatin. We have previously reported that different cannabinoid agonists reduce motor function in the small intestine and that this effect is reversed by CB₁ (but not CB₂) antagonists [38]. Likewise, administration of an FAAH inhibitor also reduced intestinal motility in a CB₁-dependent manner [41]. Therefore, the increase in AEA in the small intestine one week after treatment supports the theory that cisplatin may actually reduce small intestinal motility—if not immediately after its administration, later in time [27,42]—and that this effect might be mediated by increased activity of the intestinal CB₁ receptors through increased AEA signalling.

The radiographic study showed a significant reduction in the colonic contents in cisplatin-treated rats. Although the effect observed in the colon could very likely be due to the delayed gastric emptying, which would decrease the arrival of barium-stained gut contents to the colon, a decrease in the colonic motility per se cannot be discarded. In these regards, one week after repeated cisplatin treatment (cumulative dose of 10 mg/kg), no colonic dysmotility was observed in vivo [28], but colonic contractility was altered in vitro [43], likely due to the mentioned development of an enteric neuropathy [40]. If colonic contractility is also altered one week after a single 5 mg/kg dose of cisplatin is not known, but—if present—such an effect would not likely involve the ECS, since no significant changes in the expression of its components were found here.

The behavioural analysis showed that cisplatin failed to induce statistically significant mechanical and heat hypersensitivity 4 days after treatment. We have previously reported that repeated weekly administration of 2 mg/kg of cisplatin induces significant mechanical hypersensitivity in rats after 4 weeks [22] and after 5 weeks [44]. To the best of our knowledge, only one study analysed the development of pain-related hypersensitivity of a single dose of 5 mg/kg of cisplatin in the rat. In this study, Nayebi and cols. (2012) observed thermal hypersensitivity on day 5 after treatment, although the highest thermal hypersensitivity was observed on day 15 [20]. In contrast, our results in the Hargreaves’ test show a lack of thermal hypersensitivity in our animals 4 days after cisplatin administration. These differences may be due to several factors, such as the different post-treatment time of testing, the different anatomic structures of the animal where the heat stimulus was applied (tail vs. hind paw) or to the different intensity of the heat source. In the present study, the dose used was unable to induce not only thermal but also mechanical hypersensitivity, without changes in locomotor activity, so under our experimental design, the development of chemotherapy-induced neuropathic pain may require a higher dose of cisplatin and/or more time to sensitise the pain pathways.

In this study, we evaluated changes in the ECS of L4 and L5 DRGs. Due to the small size of the tissue, we prioritised the analysis of mRNA expression over the biochemical analysis, so we could not analyse the levels of endocannabinoid ligands and related N-acylethanolamines in this tissue. Khasabova and cols. (2014) [45] did analyse the levels of these molecules after daily administration of cisplatin (1 mg/kg) in mice for a week, and they observed a significant decrease in AEA and 2-AG. These changes were mainly due to alterations in the gene expression levels of their catabolic enzymes (FAAH and MAGL, respectively). However, instead of increasing them, cisplatin significantly reduced mRNA expression of FAAH and MAGL. The authors suggested that a feedback response and/or a direct response to the effect of cisplatin may be responsible for these contradictory results [45]. Guindon and cols. (2013) analysed the effect of a weekly dose (for 3 weeks) of 3 mg/kg of cisplatin on the mRNA expression of cannabinoid receptors and catabolic enzymes in rat DRGs [46]. While we observed a significant increase in the mRNA expression levels of CB₁ in L4 and a trend in the same direction in L5, these authors did not find any significant difference [46]. Curiously, in another report using the spinal nerve ligation model of neuropathic pain (at the level of L5) in rats, a significant increase in the expression levels (both at mRNA and protein levels) of CB₁ was observed in the ipsilateral L4 without changes in L5 [47]. This is interesting since the authors were expecting an increase in this cannabinoid receptor in L5, where the lesion was made, and not L4. On the other hand, they observed a significant increase in the levels of AEA and 2-AG in L5, but not in L4 [47]. The authors interpreted these conflicting results as an adaptation to the hyperexcitability induced by the surgery in the ipsilateral side. This hyperexcitability would induce an increase in CB₁ in the DRGs, in both L4 and L5. However, in the ipsilateral L5, and due to the neuroinflammation observed in this area, there is also an increase in AEA and 2-AG that would reduce the expression of CB₁ in L5, but not in L4, where the levels of AEA and 2-AG were the same as in controls [47]. In view of this, we could suggest that our cisplatin treatment induced a hyperexcitability in sensory neurons of the DRGs, which induced a significant increase in the expression of CB₁. However, as our treatment was less harmful than a direct ligation to the nerve, there was no neuroinflammation and no increase in the levels of AEA and 2-AG of the DRGs, and subsequently no decrease in the already overexpressed CB₁. Obviously, biochemical analysis of these samples would have given us a better perspective of this theory.

A single dose of cisplatin 5 mg/kg failed to induce any change in the levels of endocannabinoid ligands and related N-acylethanolamines in the PFC, PAG, or AMY one week after its administration. In line with this result, an acute treatment with 1 mg/kg did not alter levels of AEA, 2-AG, and PEA in the midbrain of male mice one week after treatment [45]. Remarkably, only one study has analysed the levels of endocannabinoid ligands and related N-acylethanolamines in the nervous system after repeated administration of cisplatin in rats. In this study, from Guindon and cols. (2013) [46], the authors observed that a treatment with a dose of 3 mg/kg/week for 3 weeks induced a significant increase in the levels of AEA and 2-AG in the lumbar spinal cord. The authors suggested that this increase happens as an adaptive central response to the peripheral neuropathy induced by cisplatin [46]. In view of this, we could hypothesise that our single dose, that was inefficient to induce the classic neuropathic pain-related mechanical hypersensitivity, did not elicit a response in the central nervous system (CNS), and hence did not alter the levels of endocannabinoid ligands and related N-acylethanolamines in brain regions involved in the descending pain pathway.

However, despite this lack of effects in the levels of endocannabinoids, we did observe interesting alterations in the gene levels of several elements of the ECS that may suggest that changes in the levels of endocannabinoids do happen, but only shortly after cisplatin administration, tending to return to baseline levels quickly after the initial insult in order to restore homeostasis into the CNS. The molecular analysis showed, for instance, that cisplatin induced a significant increase in the gene expression for MAGL in the PFC and a significant decrease in the AMY. Alterations in the corticoamygdalar pathway have been shown to play a pivotal role not only in pain but also in emotion–pain interactions [48,49,50]. In normal conditions, pyramidal cells (PCs) from the basolateral amygdala (BLA) send excitatory glutamatergic projections to PCs of the PFC that project to the central nucleus of the amygdala (CeA) inducing an inhibition of the outputs in this region. In brief, an activation of the PFC leads to a deactivation of CeA outputs. However, in acute animal models of arthritis, colitis, formalin-induced inflammation, and neuropathic pain, CeA output firing is increased (see [49] for more information) and mediates pain-related emotional responses [50]. The ECS modulates the corticoamygdalar pathway through the CB₁ receptor in inhibitory GABAergic interneurons of the PFC. BLA activation of PCs in the PFC activates the production and release of 2-AG, which will bind to CB₁ in GABAergic interneurons modulating GABA-related inhibition of the PCs [48,50]. Interestingly, in arthritic animals there is a lack of this ECS-mediated control, and the mechanism involved seems to be related to an increase in 2-AG hydrolysis rather than changes in the activity/expression of CB₁ [51]. Our results show that cisplatin induced a significant increase in MAGL in the PFC, without changes in the expression levels of CB₁. This increase in MAGL would induce a decrease of 2-AG that cannot modulate the feedforward inhibition induced by the GABAergic interneurons, decreasing PFC-related firing to the CeA. Since PFC-related control of CeA is through inhibitory interneurons, the decrease in PFC firing would induce an increase in CeA outputs and an increase in pain-related emotional responses (Figure 6). Still, we did not detect any changes in the levels of 2-AG in our biochemical analysis, and we did not evaluate emotionality in these animals, so further analyses are needed to confirm this hypothesis.

Regarding the decrease in MAGL mRNA expression in the AMY, the metabolic activity in PFC and AMY is often inversely correlated in clinical populations with an imbalance in the corticoamygdalar pathway [48]. So, it is plausible that the decreased activity of the ECS in the PFC (shown by an increase in the hydrolytic enzyme MAGL) may be reflected as increased activity (lower levels of MAGL) in the AMY. Despite the fact that no changes in the levels of 2-AG were observed in the AMY, it is possible that the hyperactivation of the ECS is behind the decreased levels of mRNA for CB₁ also observed in the AMY of cisplatin-treated animals, as a way to counteract an excessive activation of CB₁. Anyway, and due to the differential activity of the two nuclei of the AMY (BLA vs. CeA) and the fact that we analysed mRNA expression in the whole AMY, we should be very careful when interpreting these results.

In the PFC, the increased gene expression of MAGL was accompanied by an increased gene expression of CB₂ and PPARα. It is well established that CB₂ expression increases due to neuroinflammatory processes [52] and that a decrease in MAGL and an increase in PPARα signalling have anti-inflammatory and neuroprotective activities in the CNS [53]. Neuroinflammation in the PFC usually leads, when persistent, to cognitive impairments, and due to the pro-inflammatory nature of chemotherapy action, it is well known that chemotherapeutic drugs can induce cognitive deficits in a process known as chemotherapy-induced cognitive impairment or chemobrain [2,54,55]. In fact, it has been observed that a repeated weekly treatment with a 5 mg/kg dose of cisplatin induces cognitive deficits, hippocampal activation of NF-κB and expression of downstream inflammatory mediators in rats [56], and a decreased percentage of alternations in the Y maze, indicative of a poorer hippocampal spatial memory, accompanied by pro-inflammatory and proapoptotic hippocampal markers [29]. In view of this, we could argue that the alterations in the gene expression of CB₂ and PPARα in the PFC of cisplatin-treated rats may be related to acute neuroinflammatory processes that may have induced cognitive deficits in these animals. A future study should analyse the cognitive effects of an acute administration of 5 mg/kg of cisplatin.

Finally, we also measured the levels of endocannabinoid ligands and related N-acylethanolamines in plasma to determine the potential usefulness of these molecules as an early biomarker of chemotherapy-related neuropathic pain or gastrointestinal function. Several studies have shown that plasma/serum levels of these lipidic compounds can be used as biomarkers of chronic pain syndromes such as osteoarthritis pain [57], complex regional pain syndrome [58], fibromyalgia [59,60], migraine [61], and interstitial cystitis/bladder pain syndrome [62]. However, we did not observe any significant alteration in the circulating levels of AEA, 2-AG, PEA, and OEA in our plasma samples. Thus, circulating levels of these compounds seem to not be useful as an early biomarker of neuropathic pain induced by a chemotherapeutic treatment. In any case, at this stage we cannot discard the utility of these biomarkers in models of neuropathic pain induced by other chemotherapeutic drugs, such as oxaliplatin, vincristine, etc., or with cisplatin at different time-points or with a different dose regime. Regarding the possible usefulness of endocannabinoid ligands as markers of chemotherapy-induced gastrointestinal dysfunction, plasma levels of endocannabinoid ligands, such as AEA, 2AG, and PEA, were altered in women suffering from gastroparesis, a chronic gastrointestinal motility disorder where there is delayed gastric emptying without gastric outlet obstruction [63]. The authors also analysed the plasma levels of these endocannabinoids in men suffering from this condition, but in this case, no significant effect was observed, which is in line with our current results. A future study in which female rats are also evaluated is required to check if cisplatin-induced delayed gastric emptying may be able to induce, at least in females, alterations in plasma endocannabinoid levels and, hence, be useful as a biomarker of cisplatin-related gastrointestinal effects.

In summary, our results show that a single acute treatment with a 5 mg/kg dose of cisplatin induces important alterations in the ECS of gastrointestinal and nervous systems that could be linked to changes in gastrointestinal motility and early development of neuropathic pain, severe side effects observed in cisplatin-treated cancer patients.