Differential involvement of neurotransmitters through the time course of cisplatin-induced emesis as revealed by therapy with specific receptor antagonists

https://doi.org/10.1016/S0959-8049(02)00674-3Get rights and content

Abstract

Advances in antiemetic therapy for chemotherapy-induced emesis have resulted in improved protection against symptoms occurring within 24 h of chemotherapy. However, the vomiting which tends to occur beyond 24 h after chemotherapy (delayed-phase vomiting) is still relatively poorly controlled by the currently available drugs, suggesting that more than one mechanism may mediate these symptoms. The standard antiemetic regimen currently recommended for prevention of chemotherapy-induced emesis includes a serotonin (5-HT3) antagonist and a corticosteroid. The neurokinin-1 (NK1) antagonist aprepitant represents a new class of antiemetic currently in clinical development. Using data obtained in 2 Phase II clinical trials of aprepitant in patients receiving chemotherapy based on the highly emetogenic chemotherapeutic agent cisplatin, we compared the time course of antiemetic effect of aprepitant, a 5-HT3 antagonist, or a combination of both. Over the entire observation period (up to 7 days post-cisplatin), patients who received the NK1 antagonist had a superior prevention of emesis. However, in the first 24 h after cisplatin, emesis occurred in fewer patients who received the 5-HT3 antagonist than in patients who did not receive this class of drug. Furthermore, the majority of treatment failures in patients who received the NK1 antagonist occurred within the first 8–12 h of chemotherapy, whereas the treatment failures in patients who received a 5-HT3 antagonist were more evenly distributed over time. Patients who received both drugs had superior control of symptoms compared with patients who received one or the other. The difference in the time course of emesis blockade observed with two different classes of receptor antagonists provides substantial evidence for involvement of separate pathophysiological mechanisms in chemotherapy-induced vomiting. Serotonin mediates the early vomiting process that occurs within 8–12 h following cisplatin-based chemotherapy, after which time substance P acting at NK1 receptors becomes the dominant mediator of vomiting

Introduction

Vomiting is a fundamental protective reflex mediated by the central nervous system to prevent the harmful consequences of ingested, potentially toxic substances. The vomiting reflex is mediated by several distinct brainstem nuclei which integrate afferent inputs from diverse sources, including: the area postrema (a medullary site which contains the chemoreceptor trigger zone); the vestibular system; the pharynx and gastrointestinal and cardiovascular systems; and higher brainstem or cortical sites [1]. The peripheral afferent input from the gastrointestinal tract is mediated predominantly by the vagus nerve. Afferent inputs to the vomiting centre are coordinated by the brainstem neuronal network of the dorsal vagal complex which includes the nucleus tractus solitarius (NTS). The NTS is a site for convergence of afferent input into the final common efferent pathway, via the dorsal motor nucleus of the vagus, that produces the various visceral and skeletal muscular contractions necessary to produce oral expulsion of gastrointestinal contents, as well as changes in gut motility [2].

Many neurotransmitters have been implicated in the pathogenesis of vomiting, including dopamine, acetylcholine, histamine, opiates, serotonin and substance 1, 3, 4. A more refined understanding of the relative importance of these neurotransmitters and their interrelationships in the regulation of vomiting is necessary for the development of more effective approaches for the treatment or prevention of vomiting.

In addition to its continued importance as a clinical problem, vomiting induced by cancer chemotherapy may serve as an important model for understanding the physiology of vomiting in general. Cisplatin is the single most emetogenic chemotherapeutic agent currently in use and may be considered the benchmark for evaluation of preventive strategies for CIV. At doses >50 mg/m2 and in the absence of prophylactic therapy, cisplatin causes vomiting in virtually all patients [5]. This vomiting typically follows a biphasic time course. Following initiation of the cisplatin infusion there is a latency period of 1–3 h before the onset of vomiting. The peak frequency of vomiting tends to occur at 6–8 h post-initiation of the cisplatin infusion, and this first phase of vomiting diminishes at approximately 12 h. There follows a tendency for less emesis over approximately 4 h, after which the second phase of vomiting begins (approximately 16 h post-initiation of cisplatin). This second phase peaks between 24 and 72 h, although vomiting frequently occurs for several more days. In the setting of clinical studies of cisplatin-induced emesis, CIV has historically been described as occurring in two arbitrarily defined phases: the acute phase (from 0 to 24 h following the initiation of chemotherapy) and the delayed phase (from 24 h onwards); although the definitions are arbitrary, we present our findings in the context of this convention.

The neurotransmitter serotonin (5-hydroxytryptamine or 5-HT) has been shown both clinically and also in relevant animal assays to be an important mediator of the early (“acute”) phase of CIV. Preclinical studies have shown that cisplatin causes a calcium-dependent exocytic release of serotonin from enterochromaffin cells in the gastrointestinal tract, possibly as a result of free radical generation 8, 9, 10. The released serotonin then activates receptors on vagal afferent fibres, which stimulates the CNS centres that mediate the emetic response 8, 9. These receptors are known to be of the 5-HT3 subtype, as 5-HT3 receptor antagonists (RAs) inhibit the acute emetic response in a ferret CIV model 6, 7, an observation which generated further interest in the role of serotonin.

Cisplatin administration in humans has also shown clear evidence for the involvement of serotonin. After cisplatin administration, there ensues a large increase in the urinary output of the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA) within 24 h [11], indicating the release of intracellular serotonin. The time course of this response in the acute CIV period appears to correlate well with the clinical efficacy profile of 5-HT3 RAs, which are most active against the acute-phase CIV associated with cisplatin-based chemotherapy [11].

Although the predominant therapeutic effect of 5-HT3 RAs is believed to be antagonism of peripherally released serotonin, a central effect cannot be completely excluded. 5-HT3 receptors have been shown to exist in the area postrema, NTS, subnucleus gelatinosus and in lower densities in the dorsal motor nucleus of the vagus and the spinal trigeminal tract of a number of species including man 12, 13, 14, 15, 16, 17, 18, 19. It has been shown in dogs that antagonism at 5-HT3 receptors located within the blood–brain barrier can block cisplatin-induced emesis [20], but the relevance of this finding to the pathophysiology of CIV in humans is unclear, as plasma levels of 5-HIAA are not increased by cisplatin in dogs [21]. Higgins and colleagues reported that in the ferret, emesis induced by cisplatin was attenuated, but not blocked, by infusions of 5-HT3 receptor antagonists into the area postrema [22], although the 5-HT3 selective agonist 2-methyl 5-HT did not reproducibly induce emesis when infused via the same route. It is therefore unclear whether 5-HT3 RAs have any central activity in humans. It is widely assumed that the major antiemetic activity of 5-HT3 RAs occurs through inhibition of afferent vagal stimulation in the periphery. Underscoring the fact that interpretation of the effects of 5-HT3 RAs is not a simple matter, Andrews and colleagues have reported that although ferrets are initially refractory to radiation-induced emesis following vagotomy, the emetic reflex returns in some animals [23]. This finding suggests that emetic reflex pathways may have some plasticity, and could help to explain why the antiemetic effects of 5-HT3 RAs are more pronounced in some patients than in others.

Although important in the acute phase, serotonin is not believed to be a significant mediator of emesis occurring more than 24 h after chemotherapy (historically known as delayed vomiting). Delayed CIV responds poorly to 5-HT3 antagonists in both humans and animal models [11], and it is therefore highly likely that other neurotransmitters are involved in the pathogenesis of delayed-phase symptoms.

Substance P, a member of the tachykinin family of neuropeptides, was first implicated as a potential mediator of vomiting when Amin and colleagues described high levels of this peptide in the area postrema of dogs [24]. Subsequently, studies in ferrets showed that the potent capsaicin analogue resinferatoxin blocked the emetic response to both centrally and peripherally acting emetic agents [23]. It was suggested that this antiemetic effect was mediated by resinferatoxin-induced depletion of sensory neurotransmitters such as substance P in the NTS in the brainstem. In support of this concept, animal studies using both centrally and peripherally active emetogenic stimuli demonstrated that vomiting was prevented by non-peptide antagonists of the neurokinin-1 (NK1) receptor, a site at which substance P is thought to act 4, 25, 26. Substance P is co-localised with serotonin in enterochromaffin cells in the gastrointestinal tract, and substance P levels in the peripheral circulation have been reported to be elevated following cisplatin administration in patients [27]. Substance P has been shown in animals to cross the blood-brain barrier, which raises the possibility that substance P of peripheral origin may act centrally to induce emesis [28]. CNS penetration by the NK1 RAs has been shown to be essential for the prevention of vomiting in the first 4 hours following cisplatin-based chemotherapy, which suggests that the antiemetic effect of NK1 RAs is mediated centrally, probably in region of the NTS [29].

The spectrum of antiemetic activity observed with NK1 RAs in preclinical studies was broader than with other antiemetics such as 5-HT3 RAs. Specifically, NK1 RAs prevented both acute and delayed vomiting induced by cisplatin in the ferret 4, 30, whereas 5-HT3 RAs prevented only acute vomiting in this model. These preclinical data, especially those derived from the ferret model of CIV, were compelling enough to justify clinical evaluation of NK1 RAs.

Section snippets

Clinical data with aprepitant (MK-869) and L-758, 298

Aprepitant (MK-869) is a powerful and selective brain-penetrant NK1 RA that can be administered orally, and L-758,298 is an intravenous pro-drug for aprepitant. Four studies have been published demonstrating the efficacy of these NK1 RAs in the prevention of CIV associated with high-dose cisplatin 31, 32, 33, 34. Their particular efficacy in delayed CIV represents a potentially important medical advance in the treatment of a condition for which current therapy is sub-optimal. Using the

Single-agent comparison of L-758,298 and ondansetron

The time course of emesis was initially analysed in a study comparing a single dose of L-758,298 with a single dose of ondansetron in patients receiving their first course of high-dose cisplatin (⩾50 mg/m2) (efficacy and tolerability results previously published) [31].

A post hoc analysis of data collected up to 7 days post-cisplatin in this study showed that a differential pattern of emesis occurred in the two treatment groups. Over the entire observation period of 168 h, 31.0% of patients had

Discussion

The role of serotonin-related mechanisms in cisplatin-induced emesis is now well established. 5-HT3 RAs are the cornerstone of current preventive regimens for CIV, given their efficacy and excellent tolerability. When used with cisplatin, the 5-HT3 RAs are most effective in the prevention of emesis occurring in the first 24 h (acute emesis) although they appear to be distinctly less effective for emesis developing after 24 h (delayed emesis). Only recently have the mechanisms underlying delayed

Acknowledgments

This research was funded by Merck and Co., Inc., manufacturers of aprepitant.

References (37)

  • F.D. Tattersall et al.

    The novel NK1 receptor antagonist aprepitant (L-754,030) and its water soluble phosphoryl prodrug, L-758,298, inhibit acute and delayed cisplatin-induced emesis in ferrets

    Neuropharmacol.

    (2000)
  • V. Cocquyt et al.

    L-758,298, a prodrug for the selective neurokinin-1 antagonist, L-754,030, compared to ondansetron for the prevention of cisplatin-induced emesis

    Eur. J. Cancer

    (2001)
  • S.M. Grunberg et al.

    Control of chemotherapy-induced emesis

    New Engl. J. Med.

    (1993)
  • C. Veyrat-Follet et al.

    Physiology of chemotherapy-induced emesis and antiemetic therapy

    Drugs

    (1997)
  • J. Herrstedt

    New perspectives in antiemetic treatment

    Support Care Cancer

    (1996)
  • P. Diemunsch et al.

    Potential of substance P antagonists as antiemetics

    Drugs

    (2000)
  • M.G. Kris et al.

    Are more antiemetic trials with a placebo necessary? Report of patient data from randomized trials of placebo antiemetics with cisplatin

    Cancer

    (1996)
  • W.D. Miner et al.

    Inhibition of cisplatin-induced vomiting by selective 5-hydroxytryptamine M-receptor antagonism

    Br. J. Pharmacol.

    (1986)
  • Cited by (227)

    • GDF15 Induces Anorexia through Nausea and Emesis

      2020, Cell Metabolism
      Citation Excerpt :

      Cisplatin induced c-Fos co-localized with GFRAL-positive neurons in the AP/NTS in mice, supporting our hypothesis that GFRAL-positive neurons are stimulated by cisplatin treatment, presumably via induction of peripheral GDF15 secretion. Although we did not directly test whether GFD15 release following cisplatin treatment plays a causal role in cisplatin-induced anorexia, we did detect elevated GDF15 during the same time window of the onset of cisplatin-induced emesis or emetic-like behavior in humans and animal models (Alhadeff et al., 2015; Hesketh et al., 2003; Rudd et al., 2018; Yamamoto et al., 2014). Data in GFRAL KO mice (Hsu et al., 2017) and in tumor models using a GDF15 antibody (Johnen et al., 2007) would suggest that blunting GDF15-GFRAL signaling would lead to reduced chemotherapy-induced malaise.

    View all citing articles on Scopus
    View full text