Results and Discussion

Serotonin is synthesized from essential amino acid tryptophan (Fig. 1B) in the brain and other peripheral tissues (Wurtman et al., 1980; Mbongue et al., 2015; O’Mahony et al., 2015). The central effects of serotonin are thought to include modulation of mood, the sleep-wake cycle, hormone secretion, sexual behavior, and thermoregulation (Bijl, 2004). Serotonin syndrome is believed to be caused by increased serotonin at the intrasynaptic cleft as the result of the pharmacological effect of a drug or a combination of drugs that modulate intracerebral serotonin concentrations (Bijl, 2004; Boyer and Shannon, 2005; Cooper and Sejnowski, 2013). There are various mechanisms by which drugs could potentially produce SS via their functions in serotonin clearance. Examples of those drugs include, but are not limited to, SSRIs and MAOIs. The former inhibit serotonin reuptake into presynaptic neurons, while the latter prevent the breakdown of serotonin (Cooper and Sejnowski, 2013).

EPAC is an inhibitor of the enzyme indoleamine IDO1. IDO1 is one of the key enzymes involved in the kynurenine metabolism pathway (Fig. 1B) (Mbongue et al., 2015). It is known that increasing the synthesis of serotonin by elevations in L-tryptophan can potentially lead to SS (Wurtman et al., 1980; Esteban et al., 2004). Its manifestation may be the result of the elevation of the tryptophan concentrations either through systemic circulation or via local effect at the synaptic cleft since serotonin does not cross the BBB. Therefore, it is possible that inhibition of IDO1 could precipitate a SS through alterations in the concentrations of tryptophan, and in turn serotonin. Animal models of SS can be helpful in identifying drugs and drug combinations that have the potential to induce SS in humans (Haberzettl et al., 2013), in addition to uncovering the mechanism behind the development of SS. The effects of a single dose of SSRI on serotonin concentrations using in vivo microdialysis techniques have been studied in rats. Fluoxetine (either through subcutaneous or intraperitoneal injection) augmented the serotonin concentrations by 110%–300% in the rat frontal cortex (Gobert et al., 1997; Koch et al., 2002; Cooper and Sejnowski, 2013). In this current rat microdialysis study, a moderate increase of up to 2-fold in ECF serotonin concentrations was demonstrated at approximately 120 minutes following treatment with fluoxetine (Fig. 2A). In addition, the 2-fold increase in serotonin concentrations was maintained during the treatment period, which is consistent with reported observation and the long half-life of fluoxetine. Some of the reversible inhibitors of monoamine oxidase A, such as linezolid, cannot produce SS when taken alone, even as an overdose (Lawrence et al., 2006). However, the concomitant administration of linezolid and a SSRI, such as fluoxetine, can boost the serotonin concentrations in the synaptic cleft, and therefore can augment the SS (Patel and Galarneau, 2016). In the current study, coadministration with fluoxetine resulted in a time-dependent increase in serotonin ECF concentrations in rats with a pronounced (9-fold) elevation being observed at approximately 300 minutes post-treatment (Fig. 2A).

The effect of EPAC on the ECF serotonin concentration levels was also determined in this study using microdialysis in freely moving conscious rats. Following a 25 mg/kg oral dose of EPAC (a dose that is equivalent to the 200–300 mg twice daily clinical dose), the average serotonin concentrations in brain ECF remained unchanged throughout the 360-minute treatment time period (Fig. 2A). Similar effects were also observed with coadministration of EPAC and linezolid, in that brain ECF serotonin concentrations were comparable to those in both the EPAC monotherapy and vehicle control groups (Fig. 2A).

To have a more direct comparison on the serotonin concentrations following each treatment option, the area under the curve values of serotonin collected during the −100 to 360 minute pre- and post-treatment periods were calculated (Fig. 2B). The values of the area under the curve of serotonin were comparable among the vehicle, EPAC, and combination of EPAC with linezolid groups. However, these values increased 2.1- and 5.5-fold to that of the vehicle group in the fluoxetine alone and the fluoxetine and linezolid cotreated groups, respectively (Fig. 2B). The present findings are in good agreement with literature published reports indicating that either SSRI alone or the combination with MAOI could cause SS. In addition, these data also suggested that the SS is unlikely following the treatment with EPAC, an IDO1 inhibitor, either as a single agent or combination treatment with MAOIs, such as linezolid.

Adequate brain uptake of the perpetrating agent is an expected requirement for modulation of intracerebral serotonin concentrations. Indeed, linezolid has shown good CNS penetration and is a promising candidate for treatment of CNS infections (Villani et al., 2002). In the supporting study, the brain penetration potential of EPAC was determined in rats following intravenous dosing. At steady state (4 hours postdose), the mean brain homogenate concentration of EPAC was 15% of the average plasma concentration of 0.851 µM (Table 1). Moreover, the CSF concentrations of EPAC were below the quantification limit of 0.004 µM, and therefore well below the corresponding free plasma concentration at steady state. These results support minimal brain penetration of EPAC in rats, which is in good agreement with a recent publication demonstrating the low brain penetration potential of EPAC in C57BL/6 mice following a 50 mg/kg oral dose with a correspondent plasma concentration close to 30 µM (Ladomersky et al., 2018). The restricted BBB penetration of EPAC may be due in part to efflux by both P-glycoprotein and/or breast cancer resistance protein (Zhang et al., 2017), two major efflux transporters expressed in the brain microvessel endothelial cells lining the BBB (Giacomini et al., 2010). The poor brain uptake for EPAC suggests that the potential for IDO1 inhibitory activity in the brain following EPAC oral dosing is limited.

As shown in Fig. 2A, treatment with fluoxetine alone or in combination with linezolid significantly increased the brain ECF serotonin concentrations in rats. The impact of these treatments on the brain ECF concentrations of tryptophan was also determined. Although there were large variations in the baseline concentrations of tryptophan between the two treatment groups, treatment with either fluoxetine alone or the combination of fluoxetine and linezolid did not have any significant effect on the brain ECF concentrations of tryptophan (Fig. 2C) in rats. The minimum impact on the ECF concentrations of tryptophan despite a drastic effect on the serotonin concentrations may be partially explained by the fact that these two compounds do not have a direct effect on the kynurenine pathway, which accounts for approximately 95% of tryptophan catabolism (Prendergast et al., 2017). In addition, this observation may also be due in part to the fact that both fluoxetine and linezolid have the brain penetration potential. Therefore, their impact on the local pool of serotonin in brain ECF is significant via interacting directly with the serotonin reuptake pathway (fluoxetine) or the serotonin metabolism pathway (linezolid) in the presynaptic cleft.

The ECF concentrations of tryptophan were also determined following treatment with either EPAC alone or in combination with linezolid. The tryptophan ECF concentrations did not change significantly following treatment with the combination of EPAC plus linezolid (Fig. 2C). Following administration of EPAC alone, there was a trend toward an increase in tryptophan concentrations in brain ECF; however, the changes were well within the normal range for tryptophan across studies. The lack of indoleamine 2,3-dioxygenase inhibitory effect in rat brain is likely a result of poor brain uptake of EPAC. Approximately 95% of the tryptophan metabolism pathway is through the kynurenine catabolic route (Prendergast et al., 2017); therefore, inhibition of IDO1 may have an impact on the tryptophan concentration systemically. In addition, since circulating tryptophan can cross the BBB (Wurtman et al., 1980), the plasma concentrations of tryptophan were determined in the EPAC alone and vehicle groups. Although the tryptophan concentration at 6 hours postdose with 25 mg/kg EPAC (220 ± 170 µM) was greater than that of the vehicle-treated group (157 ± 30 µM), these values were not statistically significantly different from each other, suggesting a minimal impact of EPAC treatment on systemic concentration of tryptophan.

The effect of EPAC (an IDO1 inhibitor that inhibits the conversion from tryptophan to kynurenine) on brain ECF concentrations of serotonin was minimal in rats using the microdialysis method. In addition, codosing of linezolid in the treatment did not have a significant impact on the serotonin concentration. In contrast, either fluoxetine alone or the combination with linezolid significantly increased the local pool of serotonin in the brain ECF of rats. The lack of effect on the serotonin ECF concentrations by EPAC may be partially explained by the fact that it has minimal penetration of the BBB. Furthermore, the fact that the concentration of circulating tryptophan following EPAC treatment was not statistically significantly different from the vehicle treatment group suggested that the impact of EPAC treatment on the circulating tryptophan, and consequently on the brain concentration of serotonin, is minimal.

Across the EPAC monotherapy and combination therapy clinical studies, subjects were closely monitored for clinical symptoms of SS. Only four cases of suspected SS have been reported among those 2490 subjects (six studies of EPAC monotherapy and 22 studies of EPAC in combination with other agents). The first case was a subject in the ECHO 202 Keynote-037/ECHO-202: Clinicaltrials.gov Identifier NCT02178722. (https://clinicaltrials.gov/ct2/show/NCT02178722?term=ECHO+202&rank=1) study that was testing the combination of EPAC with pembrolizumab in advanced or metastatic cancers. The subject reported chills, increased blood pressure, and increased temperature [all common terminology criteria for adverse events (CTCAE) grade 1] on day 1 of treatment. These events resolved within 1 week while dosing was stopped. The subject was treated with lorazepam and toradol. The subject was taking escitalopram, an SSRI. While he experienced mild symptoms, the full constellation of SS was not observed nor could it be ruled out. The SSRI was discontinued and the subject was able to restart EPAC about 1 week later at the same dose level of 100 mg twice a day without further incidents. The second case was a subject treated in the ECHO 203 (ClinicalTrials.gov Identifier NCT02318277. https://clinicaltrials.gov/show/NCT02318277) study that was testing the combination of EPAC with durvalumab in advanced solid tumors. The subject reported tremors (CTCAE grade 1) and agitation (CTCAE grade 2) on study day 47 and was assessed for SS on study day 56. The subject was not on an SSRI but was on a medication for anxiety (Alprazolam). The full constellation of SS was not observed nor could it be ruled out. The events resolved and the dosing with EPAC was interrupted for 1 week. Retreatment started with a lower oral dose of EPAC of 50 mg twice a day from 75 mg twice a day on study day 58 without further incidents. The third case involved a subject treated in ECHO 306 (Keynote-715-05/ECHO-306-05. ClinicalTrials.gov Identifier: NCT03322566. https://clinicaltrials.gov/show/NCT03322566.), a combination study with pembrolizumab and 100 mg EPAC twice a day in nonsmall cell lung cancer, taking granisetron, an antiemetic. The subject reported chills, fever, tremors, tachycardia, and hyper-reflexia (CTCAE grade 1) on day 16. The recovery was confirmed as the symptoms resolved after treatment with cyproheptadine. The subject stopped epacadostat but continued on pembrolizumab and pemetrexed. The fourth case was a subject on study ONC-DPX-Survivac-06, a combination of EPAC 300 mg twice a day with DPX-Survivac Vaccine and cyclophosphamide in ovarian cancer. The subject was receiving multiple medicines including alprazolam and ondansetron. The subject presented on study day 48 with complaints of tachycardia, tachypnea, palpitations, nausea, dizziness, confusion, shaking for the last hour with tremor and weakness in the legs, and jerky movements. The subject was also extremely thirsty and anxious. Lactated Ringer’s solution, lorazepam, and cyproheptadine were given to the subject and the symptoms were reduced within 30 minutes. The subject resumed treatment with no new symptoms reported. All four cases occurred in different clinical trials, were mild in severity, and the subjects recovered after supportive care interventions took place.

Thus, although four SS-like episodes occurred across multiple clinical studies, these episodes were confounded by other medical conditions that were mild in severity, and all events recovered after supportive care interventions took place. In addition, three of the four subjects were able to resume treatment with EPAC. None of the four reports was clinically substantiated to represent a true case of SS. In summary, the preclinical microdialysis data as well as the CNS penetration data in rats suggest that SS is unlikely following treatment with either EPAC alone or in combination with MAOIs such as linezolid. Moreover, no true cases of SS were demonstrated clinically. Based on the preclinical and clinical observations, EPAC does not appear to cause SS and the exclusion of MAOIs from clinical studies with EPAC has been lifted.