Although several chemotherapy regimens have been associated with Diarrhea to varying degrees (Table 1), most basic research into the mechanisms underlying CID has focused on irinotecan and its active metabolite SN38 (Gibson and Keefe, 2006). As diarrhea is a well-recognized side-effect of irinotecan treatment, the histological changes that occur throughout the GI tract in response to irinotecan administration have been examined in several animal studies (Araki et al., 1993; Ikuno et al., 1995; Takasuna et al., 1996; Gibson et al., 2003). Pronounced crypt ablation, villus blunting and epithelial atrophy in the small and large intestines have been reported (Logan et al., 2008), resulting in mucosal damage and degeneration being a major theme throughout the literature surrounding CID. Although patients do not routinely have imaging or endoscopy to diagnose the chemotherapy-induced mucosal inflammation (Touchefeu et al., 2014), CID is still largely believed to be a form, or by-product, of GI mucositis. Mucositis is defined as mucosal injury presenting as inflammation and ulceration, resulting in alterations of intestinal microflora and GI secretion (Stringer, 2009; Stringer et al., 2009a,b). The basic pathophysiology of mucositis can be broken into 5 sequential phases: (i) initiation; (ii) up-regulation; (iii) signaling and amplification; (iv) ulceration and inflammation; and (v) healing (Sonis et al., 2004; Lee et al., 2014).

Initiation of mucositis is believed to result from direct or indirect effects of cytotoxic chemotherapeutics on the rapidly dividing epithelial cells in GI tract, triggering apoptosis. This leads to reductions in crypt length and villus area, coupled with activation of nuclear factor-kappa B (NFκB) and subsequent up-regulation of pro-inflammatory cytokines including interleukin 1 (Lawrence, 2009), which contribute to ulceration and inflammation in the mucosal epithelium (Gibson et al., 2003; Stringer et al., 2007, 2008, 2009a,d; Logan et al., 2008). Intestinal microbiota is known to play an integral role in intestinal homeostasis and are now believed to play a key role in the development of mucositis (van Vliet et al., 2010; Touchefeu et al., 2014). Recent studies have revealed that chemotherapeutic administration has effects on intestinal microbial composition (Stringer et al., 2009a,b), and fecal microbiota (Touchefeu et al., 2014).

Much of the research investigating the effects of chemotherapeutic administration on microbiota has focused primarily on topoisomerase I inhibitor, irinotecan, due to the involvement of microbiota in its metabolism (Stringer, 2013). Upon metabolism in the liver, irinotecan is converted to its active metabolite SN-38 by enzyme carboxylesterase, before being deactivated through glucuronidation by uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1) to form SN38 glucuronide (SN38-G). However, SN38G may be reactivated to SN38 in the presence of enzyme β-glucuronidase, which may be produced by the intestinal microbiome. Several studies have shown a shift in commensal bacteria, in particular Bifidobacterium spp. toward Salmonella spp. and Escherichia coli following irinotecan administration (Stringer et al., 2009b). Of the β-glucuronidase-producing bacteria, Bacteroides spp. has been shown to decrease following irinotecan treatment, concurrently Staphylococcus spp., Clostridium spp. and E. coli have been found to be increased, whilst presence of beneficial bacteria, Lactobacillus spp. and Bifidobacterium spp. was decreased following irinotecan treatment (Stringer et al., 2007). When given in combination with antimetabolite 5-fluorouracil, both Clostridium cluster XI and Enterobacteriaceae presence was found to be increased, whilst treatment with 5-fluorouracil alone has also been found to increase the presence of Clostridium spp. and Staphylococcus spp. at 24 h post-treatment (Stringer et al., 2009c).

These changes in microbiota are believed to play an important role not only in maintaining intestinal homeostasis and integrity but in the modulation of inflammatory responses through interaction with Toll-like receptors and the nucleotide oligomerization domain receptors that activate NFκB (van Vliet et al., 2010). In the healing phase, proliferation and differentiation of the GI epithelium return approximately 2 weeks post-chemotherapy (Sonis et al., 2004; Lee et al., 2014), but functional changes persist after recovery of morphological changes (Keefe et al., 2000; Rubenstein et al., 2004). The pathophysiology underlying these persistent changes in GI functions includes several overlapping secretory, osmotic, inflammatory, and neurogenic mechanisms (McQuade et al., 2014).

Disruption to water and electrolyte balance within the GI tract is a key component in the pathophysiology of all types of diarrhea. Direct mucosal damage has been suggested as a major contributor to malabsorption and hypersecretion associated with CID (Richardson and Dobish, 2007; Stringer et al., 2007, 2009b; Stein et al., 2010). Studies using animal models of CID have demonstrated increased apoptosis in the crypts of both the jejunum and colon, resulting in metaplasia of goblet cells and excessive mucous secretion (Ikuno et al., 1995; Gibson et al., 2003). Hyperplasia of the rapidly dividing crypt cells in the epithelium of the gut probably results in heightened proportions of immature secretory cells, leading to increased secretion and decreased absorptive capacity of the villi, thereby contributing to the onset of diarrhea (Castro-Rodríguez et al., 1997). Retention of non-absorbable compounds within the lumen triggers an osmotic shift of water into the lumen (Castro-Rodríguez et al., 1997; Richardson and Dobish, 2007; Stringer et al., 2007). This reduced absorptive capacity and increased secretion in the small intestines results in increased fluid and solutes in the intestinal lumen and overwhelms the absorptive capacity of the colon resulting in diarrhea (Gibson and Keefe, 2006).

Secondary to mucosal damage, CID has been associated with mucosal inflammation throughout the GI tract (Logan et al., 2008). Increased expression of cyclooxygenase (COX)-2, associated with increased release of prostaglandin E2 (PGE2), is seen in rat colon following irinotecan administration (Yang et al., 2005). PGE2 stimulates colonic secretion and hyperperistalsis of the gut, whilst inhibiting sodium, potassium and adenosine triphosphatase, and triggering excessive chloride secretion, all of which further contribute to the onset of diarrhea (Kase et al., 1997a,b; Leahy et al., 2002; Yang et al., 2005). Further, irinotecan stimulates the production of thromboxane A2, a potent physiological stimulant of chloride and water secretion in the colon (Sakai et al., 1997; Suzuki et al., 2000) as well as tumor necrosis factor- α (TNF-α) a pro-inflammatory cytokine and a primary mediator of immune regulation associated with CID (Yang et al., 2005).

Chemotherapy can induce damage to the ENS (Vera et al., 2011; Wafai et al., 2013) which may also underlie GI secretory disturbances involved in pathophysiology of CID. Innervation of the GI tract is primarily from the ENS, sometimes referred to as “the second brain” due to its ability to function autonomously of the central nervous system (Phillips and Powley, 2007). The ENS is comprised of ganglia, primary interganglionic fiber tracts as well as secondary and tertiary fibers which project to many of the effector systems of the gut including muscle cells, glands, and blood vessels (Hansen, 2003). The ENS is divided into two major ganglionated plexi, the myenteric (Auerbach’s), and submucosal (Meissner’s), which are responsible for controlling gut functions including motility, secretion, absorption and vascular tone. Enteric neuropathy has been linked to a variety of GI pathologies, in part due to its regulation of intestinal epithelial function and colonic motility (De Giorgio et al., 2000, 2004; De Giorgio and Camilleri, 2004; Chandrasekharan et al., 2011; Furness, 2012). However, effects of chemotherapeutics on enteric neurons and GI dysfunction have been largely overlooked until recently. It has been shown that chronic treatment with cisplatin results in myenteric neuronal loss, increase in amplitude of the neurally induced contractions of the gastric fundus strips in mice and occasional diarrhea (Pini et al., 2016). Thus enteric neuropathy may be an underlying cause of chemotherapy-induced GI dysmotility.

Movement of fluid between the lumen of the intestine and the body fluid compartments is a complex and tightly regulated process involving neural, endocrine, paracrine, and autocrine systems that act via the enteric neurons within the submucosal plexus (Lundgren et al., 2000; Johnson et al., 2012). Situated superficially to the mucosa, the submucosal plexus lies between the circular muscle and muscularis mucosa layer of the mucosa and derives innervation from neurons in the myenteric plexus as well as direct innervation from branches of the sympathetic and parasympathetic nervous systems. The submucosal plexus innervates the mucosal epithelium and submucosal arterioles to control and maintain water and electrolyte balance, secretion and vascular tone (Furness, 2012). Fluid is absorbed from the lumen containing nutrients via ion-coupled transporters and returned through secretomotor reflexes. Through activation of secretomotor neurons, water and electrolytes are moved from the interstitium of the lamina propria to the lumen, drawn from both the circulation and the absorbed fluids. Neural control of secretion and absorption of water and electrolytes occurs on multiple interacting levels. While there are secretomotor circuits confined to the submucosal plexus, they can be directly controlled by circuitry within the myenteric plexus. Despite the important role of the ENS in controlling secretory function, very little research has been undertaken to elucidate the relationship between the ENS and CID. Enteric neuropathy and/or neuronal dysfunction may be a contributing factor in chemotherapy-induced secretory dysfunction.

Chemotherapy-induced diarrhea may be classified as uncomplicated (grade 1–2 with no complications) or complicated (grade 3–4 with one or more complicating signs or symptoms), early onset (<24 h after administration) or late onset (>24 h after administration) and may be categorized as persistent (present for >4 weeks) or non-persistent (present for <4 weeks) according to the National Cancer Institute’s Common Terminology Criteria for Adverse Effects grading system (Stein et al., 2010). Although uncomplicated CID may be managed by modification of the diet and administration of standard anti-diarrheal drugs such as loperamide, octreotide and tincture of opium, complicated diarrhea requires aggressive high dose anti-diarrheal administration and hospitalization (McQuade et al., 2014). The recommendations on the management of CID were published in 1998 and updated in 2004 (Wadler et al., 1998; Benson et al., 2004a), providing guidelines for evaluation and management of CID. These guidelines have not been updated since 2004. Currently the only drugs recommended in the updated treatment guidelines are opioid derivatives such as loperamide and deodorized tincture of opium (DTO), and octreotide.

Normal bowel function requires the coordination of motility, mucosal transport, and defecation reflexes (Mancini and Bruera, 1998). Broadly constipation can be classified into three categories: normal-transit constipation, defecatory disorders and slow-transit constipation (Lembo and Camilleri, 2003). Normal-transit constipation is the most common form of constipation, where frequency of colonic evacuation is normal, yet patients believe they are constipated due to a perceived difficulty with evacuation or the presence of hard stools. Symptoms of normal-transit constipation include bloating and abdominal pain or discomfort, as well as increased psychosocial distress (Ashraf et al., 1996). Constipation resulting from defecatory disorders is most commonly due to dysfunction of the pelvic floor or anal sphincter. Defecatory disorders may result from prolonged avoidance of the pain associated with the passage of a large, hard stool or painful, anal fissure or hemorrhoid (Loening-Baucke, 1996). Structural abnormalities, such as rectal intussusception, rectocele, obstructing sigmoidocele, and excessive perineal descent, are less common causes of defecatory disorders (Lembo and Camilleri, 2003). Slow-transit constipation is associated with infrequent urge to defecate, bloating, and abdominal pain or discomfort.

Though little clinical research has been undertaken to elucidate the underlying pathology in CIC, it has been hypothesized that CIC may result from effects of chemotherapy on nerve endings in the gut (Ghobrial and Rajkumar, 2003). The GI tract is innervated by the ENS together with fibers from extrinsic sympathetic, parasympathetic (vagus nerve) and sensory afferent neurons (Phillips and Powley, 2007). Both the extrinsic and intrinsic innervation play an important role in the motor activity of the GI tract. The internal circular smooth muscle layer and the external longitudinal smooth muscle are controlled by two main mechanisms: non-neural pacemaker cells, interstitial cells of Cajal (ICCs), which generate myogenic activity and enteric neurons which provide neurogenic supply. Neuronal terminals are closely associated with ICCs which are linked to smooth muscle cells via gap junctions. Within the ENS, three main neuronal classes of myenteric neurons govern the complex motor reflex pathways: sensory neurons, interneurons, and motor neurons. The integration of inputs from these neurons and ICCs to smooth muscle cells in the colon allows expression of various motor patterns including phasic contractile activity and tonic contractile activity which contribute to colonic motor activity and the peristaltic reflex (Gwynne et al., 2004; Dinning et al., 2009; Huizinga and Lammers, 2009; Kuizenga et al., 2015).

Subtle changes to the ENS, not evident in conventional histological examination, have been suggested as a potential underlying mechanism for abnormal colonic motor function leading to constipation (Bassotti and Villanacci, 2011). For instance, alterations in the number of myenteric neurons expressing the excitatory neurotransmitter substance P, as well as abnormalities in the inhibitory neurotransmitters, vasoactive intestinal peptide and nitric oxide, and a reduction in the number of ICCs (Cortesini et al., 1995; Tzavella et al., 1996; He et al., 2000) have been observed in patients with slow-transit constipation. However, the effects of chemotherapeutics on ENS and GI dysfunction have been largely overlooked until recently. A study investigating the effects of 5-fluorouracil-induced dysmotility in mice uncovered myenteric neuronal loss alongside delayed GI transit and inhibition of propagating colonic contractions (McQuade et al., 2016). Similar results have been demonstrated following oxaliplatin administration in mice, and cisplatin administration in rats, where enteric neuronal loss was associated with a reduction in colonic motor activity and reduced GI transit time, respectively (Vera et al., 2011; Wafai et al., 2013). Loss of enteric neurons following administration of cisplatin and oxaliplatin has been correlated with an increase in a population of the myenteric neurons expressing neuronal nitric oxide synthase (Vera et al., 2011; Wafai et al., 2013) and changes in glial cell populations (Robinson et al., 2016). These studies emphasize the importance of enteric neuronal integrity in GI function whilst suggesting neuroprotection as a potential therapeutic pathway for the treatment of chemotherapy-induced GI disorders.

As previously mentioned, a major limitation in the estimation and evaluation of true CIC is the onset of secondary constipation, namely opioid-induced constipation produced by opioid analgesia. Whilst opioid analgesics are the gold standard in pain relief for cancer patients, adverse effects such as opioid-induced bowel dysfunction (OIBD) and opioid-induced constipation (OIC) severely compromise their therapeutic potential (Gonzalez and Halm, 2016). Incidence of OIC ranges from 50 to 87% in terminally ill cancer patients and is positively associated with chronic opioid treatment (Abernethy et al., 2009; Abramowitz et al., 2013). Opioid receptors are located throughout the central and peripheral nervous system and are involved in pain transmission (Camilleri, 2011). In the GI tract, μ-receptors are widely distributed throughout the ileum, stomach and proximal colon where they contribute to the control of fluid and electrolyte transport as well as motility (McKay et al., 1981; Fickel et al., 1997; Garg et al., 2016). Opioid analgesics interfere with GI motility by delaying transit, stimulating non-propulsive motility and altering GI segmentation and tone through their effects on enteric neurons (De Schepper et al., 2004; Wood and Galligan, 2004). These changes coupled with activation of mucosal sensory receptors that trigger a reflex arc facilitate excessive fluid reabsorption, resulting in OIC (Panchal et al., 2007; Camilleri, 2011).

Whilst administration of laxatives remains the first-line treatment option for OIC, this intervention alone is frequently ineffective (Gatti and Sabato, 2012). Selective μ-opioid receptor antagonists are emerging as a promising first line treatment for OIC, in particular treatment with methylnaltrexone bromide (methylnaltrexone) has been found to improve GI transit in chronically ill patients and has been recommended for use in cancer patients (Gatti and Sabato, 2012). Methylnaltrexone has demonstrated efficacy in improving opioid-induced delay in the oral–caecal transit time and inducing laxation in both healthy subjects and advanced illness patients (Culpepper-Morgan et al., 1992; Thomas et al., 2005; Thomas et al., 2008). Similarly, treatment with peripheral μ-opioid receptor antagonist Alvimopan has been found to increase the frequency of spontaneous bowel movements in non-cancer patients with opioid induced bowel dysfunction (Webster et al., 2008).

The management of constipation can be divided into general interventions and therapeutic measures. The general interventions involve increasing physical exercise, fluid intake and fiber consumption, availability of comfort, privacy and convenience during defecation as well as elimination of medical factors that may be contributing to constipation (Mancini and Bruera, 1998). Therapeutic interventions for the management of constipation, including CIC involve the administration of both oral and/or rectal bulk-forming, emollient, osmotic/saline, stimulant, and lubricant laxatives (Connolly and Larkin, 2012). Laxative compounds may fall into one of several categories depending on their mechanism of action.

Chloride is an essential ion in intestinal secretion and absorption. Secretory diarrhea, such as that experienced in irinotecan-treated patients, results from a combination of excessive secretion and reduced absorption in the intestinal lumen (Thiagarajah and Verkman, 2012). Excessive fluid secretion is driven by active chloride secretion, followed by secondary movement of water and sodium into the intestine. Although there is a lack of selective potent inhibitors of voltage gated chloride channels, inhibition of calcium-activated chloride channels throughout the intestines successfully reduced secretion of chloride into the intestinal lumen (Thiagarajah and Verkman, 2013; Thiagarajah et al., 2015). In a mouse model of rotavirus-induced severe secretory diarrhea, inhibition of calcium-activated chloride channels with a red wine extract reduced intestinal fluid secretion, diminishing the symptoms of diarrhea (Ko et al., 2014).

Conversely, chloride channel activation has been used in the management of chronic idiopathic constipation and constipation related to irritable bowel syndrome (IBS-C). Lubiprostone is a bicyclic fatty acid derived from prostaglandin E1 that specifically activates chloride channels in the intestine, whilst having no effect on smooth muscle contraction (Jun, 2013). The underlying mechanism of lubiprostone involves stimulation of electrogenic chloride secretion though activation of chloride channel type-2 (Lacy and Levy, 2007) and cystic fibrosis transmembrane conductance regulator chloride channels (Bijvelds et al., 2009) in the apical membrane of intestinal epithelial cells. Activation of these epithelial channels results in active secretion of chloride into the intestinal lumen followed by a passive secretion of electrolytes and water increasing the liquidity of the luminal contents (June, 2013). Resulting luminal distension from increased intestinal fluid content promotes GI motility and increases intestinal and colonic transit. In healthy volunteers, daily lubiprostone delays gastric emptying, increases fasting gastric volume, reduces maximum tolerated gastric volume, and accelerates small bowel and colon transit (Camilleri et al., 2006). In randomized trials involving patients with IBS-C, lubiprostone twice daily reduced abdominal pain and increased complete spontaneous bowel movement and improved stool consistency, straining, and bloating (Schey and Rao, 2011). Currently, oral lubiprostone is approved for IBS-C at 8 μg twice daily and CIC at doses of 24 μg twice daily, but approval for CIC is limited to only women who have not responded to laxatives (Davis and Gamier, 2015). At present there are no studies investigating the efficacy of lubiprostone for CID.

Cannabinoids mediate their effects via binding to two main G-protein coupled receptors, CB₁ and CB₂, widely expressed in the GI tract (Abalo et al., 2012). Although the activity of the endocannabinoid system varies between species and different regions of the GI tract within the same species, activation of CB₁ receptors coupled to cholinergic motor neurons has been found to inhibit excitatory neuromuscular transmission in human colonic circular muscle (Hinds et al., 2006) and inhibit colonic propulsion in mice and rat (Pinto et al., 2002; Abalo et al., 2015). In recent human trials, dronabinol, a non-selective cannabinoid receptor agonist, was found to inhibit colonic motility in both healthy subjects (Esfandyari et al., 2006, 2007) and patients with IBS-related diarrhea (IBS-D) (Wong et al., 2011). Conversely, a CB₁ receptor inverse agonist, taranabant, has been shown to improve symptoms related to slow GI motility and abdominal pain when administered in vivo in mice (Fichna et al., 2013). Taranabant increased the number of bowel movements after systemic and oral administration and significantly increased fecal pellet output in mice with constipation induced by ipratropium (Fichna et al., 2013). It has been demonstrated that a low dose of a non-selective cannabinoid agonist WIN55,212-2 reduced the severity of 5-fluorouracil-induced diarrhea in rats (Abalo et al., 2016).

Guanylate cyclase C is the principal receptor for heat-stable enterotoxins and plays a major role in E. coli-induced secretory diarrhea (Camilleri, 2010). Enterotoxins and endogenous peptides bind to guanylate cyclase C and stimulate the production of intracellular cyclic guanosine monophosphate (cGMP). Increased levels of cGMP activate the secretion of chloride ions through the cystic fibrosis transmembrane conductance regulator. Linaclotide is a minimally absorbed 14-aminoacid peptide that selectively stimulates intestinal epithelial cell guanylate cyclase C receptors, resulting in increased intracellular and extracellular cGMP leading to accelerated stool transit and laxation (Harris and Crowell, 2007). In phase II and III placebo-controlled studies in chronically constipated and IBS-C patients, linaclotide was found to accelerate colonic transit and improve abdominal pain and symptoms of constipation (Andresen et al., 2007; Johnston et al., 2009, 2010; Lembo et al., 2010). Linaclotide is particularly interesting in that it is both a laxative and analgesic, reducing visceral hypersensitivity with very few drug interactions, it is presently licensed for chronic idiopathic constipation and IBS-C in the USA (Davis and Gamier, 2015), but no trials on CIC have been reported to date.

With the recognition that intestinal microbiota play key roles in the pathophysiology of mucositis and development of CID/CIC, both antibiotics and probiotics have emerged as promising therapeutic options. Administration of probiotics have been shown to prevent CID in both 5-fluorouracil and irinotecan-treated animals (Bültzingslöwen et al., 2003; Bowen et al., 2007). Similarly, a combination of Lactobacillus rhamnosus and fiber has been found to reduce the severity of grade 3/4 5-fluorouracil/leucovorin-induced diarrhea by 15% in a randomized study of patients treated for colorectal cancer (Österlund et al., 2007). Administration of oral antibiotics, such as fluoroquinolone, has also been recommended for aggressive treatment of CID (Benson et al., 2004a; Maroun et al., 2007).

The selective inhibition of bacterial β-glucuronidase has recently been shown to alleviate drug-induced GI toxicity in mice (Wallace et al., 2015). A low-potency β-glucuronidase inhibitor showed promise in reducing the GI toxicity associated with irinotecan in rats (Fittkau et al., 2004). Similarly, oral administration of potent bacterial β-glucuronidase inhibitors has been found to reduce the severity of irinotecan-induced toxicity (Wallace et al., 2010). In clinical trials, Kampo medicine Hangeshashinto (TJ-14) which contains baicalin, a β-glucuronidase inhibitor, has been found to successfully reduce both the incidence and duration of chemotherapy-induced oral mucositis in colorectal cancer patients when compared to placebo patients (Matsuda et al., 2015). In non-small-cell lung cancer patients TJ-14 alleviated irinotecan-induced diarrhea (Mori et al., 2003). Compared with control patients, the TJ-14-treated patients showed a significant improvement in both diarrhea grade, as well as a reduced frequency of grade 3 and 4 diarrhea (Mori et al., 2003).