Chemical and Physical Properties

Senna is a pod (fruit) or leaf of Senna alexandrina P. Mill. (Leguminosae). The plant is a shrub usually 0.7 to 1 m tall, native to Africa, India, and Asia and cultivated in Sudan, China, India, and Pakistan (Blumenthal, 2000; Cupp, 2000). The plants have compound pinnate leaves with three to eight pairs of leaflets per leaf. Lanceolate and pale-green leaflets are 1.5 to 5 cm long and 5 to 15 mm wide. The pale-green pods turn greenish brown during maturity and dark brown after drying. The pods contain five to 10 obovate and green to pale brown seeds (Franz, 1993; WHO, 1999; Blumenthal, 2000; Cupp, 2000; Srivastava et al., 2006).

Fruits and leaves of these plants contain dianthrone O-glycosides (sennoside A, sennoside A1, sennoside B, sennoside C, sennoside D, sennoside D1, sennidin-A-monoglucoside, sennidin-A1-monoglucoside, and sennidin-B-monoglucoside), dianthrones (sennidin A and sennidin B), 1,8-dihydroxy anthraquinones (rhein, aloe-emodin, emodin, and chrysophanol), aloe-emodin diglucoside, rhein glucoside, naphthalene derivatives, flavonoids (kaempferol, isorhamnetin, quercetin diglucoside, kaempferol diglucoside, and isorhamnetin diglucoside), vicenin-2, β-sitosterol, salicylic acid, oxalate, resin, saponins, polyol, sugars, and polysaccharide hydrocolloids (Figure 1; Appendix D) (Lemli et al., 1981; Franz, 1993; WHO, 1999; Blumenthal, 2000; Terreaux et al., 2002).

Figure 1

Primary Structures of Selected Components of Senna. Some are present as glycosides (Franz, 1993; Appendix D). Glu=glucose

Dianthrone O-glycosides are major proactive constituents and make up 1.3% to 3.9% (w/w) of senna pods or leaves (Grimminger and Witthohn, 1993; Shah et al., 2000). Sennosides A and B are stereoisomers. The structure of rhein is similar to danthron, the laxative withdrawn from the market due to possible carcinogenic activity in humans (FDA, 1999; NTP, 2004).

Production, Use, and Human Exposure

Pods and leaves are harvested three times after sowing and dried in shade (Patra et al., 2005). Commercial senna comes from Egypt, Sudan, and India (Blumenthal, 2000). From April 2004 through March 2005, India exported 719,570 kg of senna leaf and pod to the United States (ITC, 2006). From April through September of 2006, India exported 319,900 kg of senna leaf and pod to the United States (ITC, 2007).

Senna is available in over-the-counter laxatives and is classified as a stimulant laxative (Leng-Peschlow, 1992; 21 CFR § 310.545; Rao, 2009). In the United States, senna is available as a tablet (sennosides), fluid extract, leaf, pod, oral solution, or tea (Cupp, 2000). Senna is used clinically for management of constipation resulting from opioid use or when treatment with bulking or osmotic agents has failed (Agra et al., 1998; Foxx-Orenstein et al., 2008; Bouras and Tangalos, 2009). In 1999, due to findings of carcinogenic activity in experimental animal models, the FDA reclassified the stimulant laxatives phenolphthalein and danthron as category II (21 CFR § 310.545) and declared them unsafe for use as over-the-counter laxatives (FDA, 1999). Therefore, it was expected that use of senna as a laxative would increase.

As a laxative, the adult daily dose of senna is 0.5 to 2 g, or 12 to 60 mg of sennosides as a single dose at bedtime. The recommended daily dose is 8.8 to 26.4 mg sennoside in children aged 6 to 12 years and 4.4 to 13.2 mg in children aged 2 to 6 years (Martindale, 1999; PDR for Herbal Medicines, 2007; Novartis, 2010).

In the United States, 2% to 28% of the population suffers from constipation (Higgins and Johanson, 2004). An estimated 7.95 million ambulatory care visits annually (0.18% of the total annual ambulatory care visits) during the period from 2001 to 2004 were for treatment of constipation with 4.6% of patients receiving prescriptions for stimulant laxatives (Shah et al., 2008). The estimated United States population using senna in 2002 was 361,000 (0.9%) (Kennedy, 2005). In the United Kingdom, senna-based products accounted for 23% of laxative use (Heaton and Cripps, 1993). A study of 1,012 patients in Seattle with functional bowel disorders found that 2.4% of those with irritable bowel syndrome and 8.2% of those with functional constipation used senna (van Tilburg et al., 2008). The prevalence of constipation in children is 0.7% to 29.6% (Van den Berg et al., 2006). In interviews with 107 caregivers of African-American children under 2 years of age, seen in a pediatric clinic, 4.7% reported using senna extract to treat colic (Smitherman et al., 2005). Women have approximately a twofold higher prevalence of constipation than men (Higgins and Johanson, 2004). During pregnancy, up to 40% of women have constipation (Cullen and O’Donoghue, 2007). In a study of 3,354 women in the Quebec Pregnancy Registry, 0.4% of pregnant women reported taking senna (Moussally et al., 2009). Adults over age 65 may have a higher risk of constipation (Higgins and Johanson, 2004); this population uses stimulant laxatives more often than bulk laxatives or stool softeners (Ruby et al., 2003). In a survey of 338 geriatric clinic patients aged 65 or older, 1.8% reported using senna tea (Cherniak et al., 2008). A high dose of senna (1 mg/kg, up to 158 mg, of sennosides) is used for cleansing the bowel in preparation for a colonoscopy or surgery (Kleibeuker et al., 1995; Martindale, 1999; Occhipinti and Di Palma, 2009). Senna’s use as a laxative is also listed by the International Association for Hospice and Palliative Care as one of the essential medicines for palliative care (De Lima and Doyle, 2007). Extract of senna leaf is also listed in the European Commission Database (2010) as a skin conditioner.

Regulatory Status

Senna is considered a dietary supplement as specified by the Dietary Supplement Health and Education Act (DSHEA) of 1994 and the DSHEA places dietary supplements in a special category under the general umbrella of “foods” (FDA, 1994). The FDA lists Alexandrian senna (Cassia acutifolia Delile) as a natural flavoring substance for use in food (21 CFR § 172.510). As of November 1993, senna is listed as an over-the-counter category III (safety and effectiveness are not yet established) digestive aid (21 CFR § 310.545; 21 CFR Part 334) Senna is designated pregnancy-category C because no animal reproduction studies and no adequate and well-controlled reproduction studies in humans are available (Tillett et al., 2003; Prather, 2004).

Absorption, Distribution, Metabolism, and Excretion

Experimental Animals

The pharmacological activity of senna is associated with sennosides A and B, the most abundant anthranoids and the precursors of the active metabolite, rhein anthrone (also known as rhein-9-anthrone) (Breimer and Baars, 1976; Sasaki et al., 1979; Lemli and Lemmens, 1980, Hietala et al., 1987; de Witte, 1993; Franz, 1993; Figure 2). The in vivo fate of the sennosides and the other senna anthranoids are described here. Sennosides are not readily absorbed from the mammalian gut; therefore, the activity depends on formation of rhein anthrone following deconjugation and reduction (Figure 2) by microflora in the large intestine (Breimer and Baars, 1976; Sasaki et al., 1979; Lemli and Lemmens, 1980; Dreessen et al., 1981). Lemli and Lemmens (1980) postulated that rhein anthrone arises through formation of a free radical following reduction of sennidin in the gastrointestinal tract. The systemic bioavailability of rhein anthrone is low, putatively due to limited absorption associated with binding to gut contents and rapid oxidation to rhein and sennidin once it is absorbed (Lemli and Lemmens, 1980; Grimminger and Leng-Peschlow, 1988; de Witte, 1993). Lemli and Lemmens (1980) recovered less than 4% of an oral dose Small amounts of free anthraquinones and their glycosides are present in senna (Franz, 1993; Newall et al., 1996; Figure 1; Appendix D), including rhein. Absorption of rhein from the rat gut appears to be greater than absorption of rhein anthrone (de Witte and Lemli, 1988). The cumulative urinary excretion of a single dose of ¹⁴C-labeled rhein or ¹⁴C-labeled rhein anthrone over 5 days was 37.1% and 2.8%, respectively. Rhein was primarily excreted as glucuronide and/or sulfate conjugates. Most of the rhein anthrone-derived ¹⁴C excreted in urine was recovered as rhein following oxidative hydrolysis and extraction of the urine. In a separate study, rhein-derived ¹⁴C was highest in the tissues of the gastrointestinal tract in gavaged rats (Lang, 1988). Other tissues contained low levels of radioactivity 7 days following dosing, probably due to protein binding in the blood. The total absorbed dose was estimated to be 50% and was excreted in urine, primarily as conjugates. In a study conducted by Dahms et al. (1997), specific metabolites, mostly glucuronide and sulfate conjugates, were identified in the urine of rats, rabbits, and dogs receiving oral doses of ¹⁴C-labeled rhein. Qualitative and quantitative differences in rhein metabolism were observed between species. For instance, rabbit urine contained a glucuronide conjugate of a potentially reactive quinoid metabolite, not evident in the urine of the other species.

Figure 2

Metabolism of Sennosides A and B in Mammals. (Adapted from Lemli and Lemmens, 1980 and Dahms et al., 1997. R, R₁, and R₂ are H, glucuronic acid, or sulfate; Glu=glucose)

The kinetics of anthraquinones were investigated following oral administration of senna-containing products to rats (Mengs et al., 2004; Mitchell et al., 2006). Concentrations of rhein and aloe-emodin were determined in the blood of rats at various time points on days 90 and 91 of a 13-week gavage study (100 to 1,500 mg/kg) of powdered Tinnevelly senna fruit (Mengs et al., 2004). The concentrations in plasma were proportional to dose from 100 to 750 mg/kg, and were generally higher in females than in males. Chrysophanol was detected in some plasma samples. Blood was sampled at 6, 12, and 24 months in a 2-year oral carcinogenicity study in rats receiving powdered Tinnevelly senna fruit (25 to 300 mg/kg) (Mitchell et al., 2006). In the study, emodin and chrysophanol were generally not detected and aloe-emodin was only detected in the plasma of the high-dose group. Concentrations of rhein were higher in females at 6 and 12 months, but were similar to males at 24 months. In a radiolabeled study of aloe-emodin in gavaged rats, the ¹⁴C was absorbed, distributed to all assayed tissues, and 30% of the total dose was excreted in urine as rhein, an unidentified metabolite, and their conjugates (Lang, 1993). The remainder of the dose was excreted in feces. The biotransformation of emodin and chrysophanol were investigated in induced liver microsomes from male and female rats (Mueller et al., 1998). Emodin was metabolized to omega-hydroxyemodin and 2-hydroxyemodin; whereas chrysophanol was metabolized to emodin.

Pharmacology

The glucose in sennosides gives characteristics of a prodrug for laxative effect. Glucose makes the molecule hydrophilic. Therefore, absorption of sennosides is prevented before they reach the target organ (the large intestine). Sennosides precipitate at low pH levels in the stomach further inhibiting absorption (de Witte, 1993; van Gorkom et al., 1999). Due to their β-glycosidic bond, sennosides are not hydrolyzed in the stomach or in the small intestine. Normal microflora in the large intestine are important for the laxative effect of senna. For example, oral administration of sennoside A produced a laxative effect in conventional male and female C3H mice but not in germ-free male and female C3H mice (Dreessen et al., 1981). Sennosides A and B are metabolized by bacteria in the large intestine to rhein-9-anthrone, the active form for laxative effect (Dreessen et al., 1981; Grimminger and Leng-Peschlow, 1988). Sennidins A and B, rhein-9-anthrone, and small amounts of rhein were produced from sennosides A and B incubated with cecal extracts from conventional female Fischer rats, whereas incubation with cecal extracts from germ-free female Fischer rats produced no metabolites (Dreessen et al., 1981). Further, incubation of sennidins A and B with cecal extracts from conventional female Fischer rats produced rhein-9-anthrone and small amounts of rhein; however, no rhein-9-anthrone was detected upon incubation of sennidins with cecal extracts from germ-free female Fischer rats. Rhein-9-anthrone was relatively stable in the rat large intestine in the presence of bacteria but oxidized readily in germ-free rats and in buffers (Dreessen et al., 1981; Grimminger and Leng-Peschlow, 1988). In female albino mice, the ED₅₀s of intracecal administration of rhein anthrone, rhein, and aloe-emodin were 11.4, 91.0, and 246.3 µmol/kg, respectively (Yagi and Yamauchi, 1999). Aloe-emodin anthrone has a weaker laxative effect than rhein anthrone. However, intracecal administration of an equimolar mixture of aloe-emodin anthrone and rhein anthrone produced a synergistic laxative effect in female albino mice (Yagi et al., 1997).

A laxative effect is driven by increasing peristalsis and reduced absorption of water and electrolytes (Leng-Peschlow, 1986). In female Wistar rats, oral administration of 50 mg/kg sennosides reduced large intestinal transit time. Intracecal administration of equimolar doses of sennosides A+B, sennidins A+B, and rhein-9-anthrone produced similar responses in reduced large intestine transit time and increased soft feces (Leng-Peschlow, 1988). Application of rhein anthrone on mucosa of isolated guinea pig ileum dose-dependently increased parameters of peristaltic reflex (longitudinal muscle tension, intraluminal pressure, and volume displacement) (Nijs et al., 1993). In isolated large intestine from male Wistar rats, application of rhein (1 mM) in the lumen increased contractility in the colon, the number of migrating contractions, and fluid flow (Rumsey et al., 1993). In humans, intraluminal introduction of senna, which was preincubated with feces or Escherichia coli, produced peristalsis within 1 hour. Intraluminal application of rhein anthrone also produced peristalsis within 1 hour (Hardcastle and Wilkins, 1970). Oral administration of sennosides in rats in vivo, as well as application of rhein on the mucosal side of isolated rat colon in vitro, decreased absorption of water and sodium; enhanced secretion of water, sodium, and potassium; and reduced Na⁺, K⁺-ATPase activity (Leng-Peschlow, 1986, ^{1989, 1993}; Wanitschke and Karbach, 1988). In addition, in humans, perfusion with rhein reversed absorption of water and sodium into secretion in the jejunum and colon, increased chloride secretion in the jejunum, reduced chloride absorption in the colon, and enhanced potassium secretion in the jejunum and colon (Ewe, 1980).

Toxicity

Reproductive and Developmental Toxicity

Carcinogenicity

Genetic Toxicity

The genetic toxicity of senna products was reviewed in detail by Brusick and Mengs (1997). These authors concluded that most studies of the genetic toxicity of senna products gave negative results, but results from some studies indicated that certain components of senna products, particularly emodin and aloe-emodin, are genotoxic. However, these authors suggested that an overall assessment of the genotoxicity profile of senna, its constituents, and its metabolites, in light of other data from animal and human metabolism or kinetic studies, human clinical trials, and rodent carcinogenicity studies, did not support an increased risk for genotoxicity of senna laxatives in humans when these products were consumed under prescribed-use conditions.

Background on the Genetically Modified Mouse (GMM) Model Used in the Senna Study

The p53 tumor suppressor gene suppresses cancer in both humans and mice. The p53 protein is critical to cell cycle control, DNA repair and apoptosis, etc., and is often mutated or lost in human and rodent cancers. The haploinsufficient Trp53 tumor suppressor gene mouse model heterozygous for wildtype and null (+/−) Trp53 alleles (Donehower et al., 1992, 1995) was used in these studies. In this model, a Trp53 null mutation was introduced by homologous recombination in AB1 murine embryonic stem cells which were derived from a black agouti 129Sv inbred mouse. By targeted insertion of a polII neo cassette, an engineered null mutation was induced as a result of the deletion of a 450-base pair gene fragment from the Trp53 gene that included 106 nucleotides of exon 5 and approximately 350 nucleotides of intron 4 that eliminated both mRNA and p53 protein expression from this allele. This Trp53 protein haploinsufficient mouse model has been extensively tested as a short-term cancer bioassay mouse model (Tennant et al., 1995; Dunnick et al., 1997; French et al., 2001a,^b; Pritchard et al., 2003; French, 2004) based upon the observation that mice with only a single wildtype Trp53 allele show a significant decrease in the time required for genotoxic carcinogen-induced tumors to develop. These tumors are often associated with either a mutation and/or a loss of heterozygosity of the remaining wildtype Trp53 allele. Few to no sporadic tumors occur in concurrent or historical study control groups in this GMM model, which allows tests to be conducted with fewer animals and direct analysis of the target wildtype Trp53 allele to test for genotoxicity in vivo as a mode of action.

For these studies, an outcross between C3H/HeNTac (C3) female mice homozygous for the wildtype (+) Trp53 allele and the C57BL/6.129Sv-Trp53^tm1Brd N12 congenic (abbreviated B6.129-Trp53^tm1Brd) N12 backcross generation males homozygous for the Trp53 null (−) allele produced C3B6.129F1/Tac-Trp53^tm1Brd N12 progeny heterozygous for a Trp53 wildtype (+) and null allele (−) inbred mouse progeny [hereafter referred to in the abbreviated form as the heterozygous F1 p53^+/− mouse, Taconic Laboratory Animals and Services (Germantown, NY)]. The heterozygous F1 p53^+/− mouse was selected for the 40-week study of senna because the B6.129-Trp53^tm1Brd (N5) haploinsufficient male and female mice (backcrossed to C57BL/6, subline unspecified, for two generations and then to C57BL/6NTac females for an additional three generations) were not sufficiently inbred. The N5 generation of this line retained both C57BL/6 and 129Sv strain allele heterozygosity at both the Trp53 locus and the flanking region on chromosome 11 and at unknown loci throughout the genome of this line. This residual heterozygosity in the B6.129-Trp53^tm1Brd N5 backcross generation mice was one covariate that may have been responsible for large variations in the p-cresidine induced urinary bladder tumors (0% to 80%, 10 of 11 studies were positive) in males, which was used as a positive control genotoxic carcinogen in the ILSI/HESI Alternatives to Carcinogenicity Testing initiative (Storer et al., 2001). Therefore, additional inbreeding to the N12 generation was anticipated to decrease the variance in tumor incidence and stabilize the penetrance of tumor phenotypes in NTP studies.

The majority of B6.129-Trp53^tm1Brd homozygous null females die in utero and only a few are born alive and most die early. Thus, the B6.129-Trp53^tm1Brd N12 line is maintained by intercross of the B6.129-Trp53^tm1Brd female heterozygote with the B6.129-Trp53^tm1Brd homozygous null male to produce a 1:2 population of homozygous null males and heterozygous null males and females. Therefore it is necessary to select the B6.129-Trp53^tm1Brd homozygous null male as the carrier of the null allele. However, the selection of the C3H/HeNTac female as the wildtype Trp53 allele carrier provides 1) increased fecundity and maternal instincts, 2) increased hybrid vigor of an F1 outcross that increases the number of progeny, 3) the advantage of expanding the pattern of tumor susceptibility associated with this genetic background, and 4) a genetic background similar to the B6C3F1 mouse used in NTP studies (NTP, 2010). Together, these factors provided a rational basis for selection of this GMM test model. In addition, the NTP study to be reported on 3′-azido-3′-deoxythymidine (AZT) and 3′-azido-3′-deoxythymidine/lamivudine (AZT/3TC), also used the C3B6.129F1/Tac-Trp53^tm1Brd N12 haploinsufficient GMM model (NTP, 2012), and the background rate for spontaneous tumors in the control group C3B6.129F1-Trp53^tm1Brd haploinsufficient mice in all three studies (AZT, AZT/3TC, and senna) was not statistically different from the background rates for spontaneous tumors observed in control B6.129-Trp53^tm1Brd (N5) haploinsufficient mice used in previous NTP GMM studies (NTP, 2005a,^{b; 2007a,b,c,d,e;} 2008).

Study Rationale

Senna was nominated for study by the Center for Drug Evaluation and Research, United States Food and Drug Administration (FDA) due to the wide use of laxative preparations, positive genotoxicity in vitro for some senna components or metabolites, and unknown carcinogenic potential. Increased use of senna was expected due to the removal of danthron and phenolphthalein from the market. Because the 2-year rat study was ongoing by the manufacturer (Mitchell et al., 2006), the FDA requested that the NTP conduct a senna study in the p53^+/− mouse.