Alkaloids in food: a review of toxicity, analytical methods, occurrence and risk assessments

Abstract

Alkaloids have been utilized by humans for years. They have diverse applications in pharmaceuticals. They have been proven to be effective in treating a number of diseases. They also form an important part of regular human diets, as they are present in food items, food supplements, diet ingredients and food contaminants. Despite their obvious importance, these alkaloids are toxic to humans. Their toxicity is dependent on a range of factors, such as specific dosage, exposure time and individual properties. Mild toxic effects include nausea, itching and vomiting while chronic effects include paralysis, teratogenicity and death. This review summarizes the published studies on the toxicity, analytical methods, occurrence and risk assessments of six major alkaloid groups that are present in food, namely, ergot, glycoalkaloids, purine, pyrrolizidine, quinolizidine and tropane alkaloids.

Introduction

Alkaloids are naturally occurring nitrogen-containing compounds that are found in plants, fungi, bacteria and animals (Cushnie et al., 2014). The presence of nitrogen atoms in their structures is responsible for their alkaline nature, and Fig. 1 shows the structures of some main alkaloids. Alkaloids are found primarily in plants as secondary metabolites, secreted to assist in their survival and reproduction (Acamovic et al., 2004; Jing et al., 2014). They are found in many higher plants, such as Papaveraceae (poppy), Ranunculaceae (buttercups) and Solanaceae (nightshades) (Aniszewski, 2007; Jing et al., 2014).

Fig. 1.

Structures of some a ergot alkaloids, b glycoalkaloids, c purine alkaloids, d pyrrolizidine alkaloids, e quinolizidine alkaloids, f tropane alkaloid

Humans have made use of alkaloids for various purposes, such as stimulants, narcotics, aphrodisiacs, poisons and medicines. Additionally, plants containing alkaloids form a regular part of human diets, both in food and drinks (Koleva et al., 2012). Common examples of alkaloids that are found in human diets include caffeine from coffee seeds, theobromine and caffeine from cacao seeds, theophylline and caffeine from tea leaves, tomatine from tomatoes, solanine from potatoes and caffeine from Coca-Cola (Kurek, 2019).

Alkaloids possess a broad spectrum of biological activities, such as antiparasitic (Fernandez et al., 2010), antiplasmodial (Frédérich et al., 2002), anticorrosive (Capasso et al., 2002), antioxidative (Czapski et al., 2014), antibacterial (Karou et al., 2006), anti-HIV (Zhang et al., 2015)and insecticidal activities (Ge et al., 2015). They have also been found to have good applications in human and animal health. Morphine, strychnine, quinine, ephedrine, colchicine and nicotine are common alkaloids that are particularly useful in clinical settings (Kurek, 2019). However, a number of alkaloids have toxic effects on humans and animals. Commonly abused drugs include alkaloids, such as morphine, cocaine, caffeine and nicotine (Beyer et al., 2009). Ergot alkaloid contamination results in direct smooth muscle stimulation, central sympatholytic activity and peripheral alpha-adrenergic blockade (Merhoff Portel, 1974). Toxic symptoms from ingesting quinolizidine alkaloids, which results from the consumption of lupin seeds that were not debittered, include blurry vision, facial flushing, dry mouth and confusion (Koleva et al., 2012). A number of pyrrolizidine alkaloids have been found to be hepatotoxic and carcinogenic (Cushnie et al., 2014).

Therefore, this review focuses on the analytical methods and toxicity of some these alkaloid groups, as well as their occurrence and risk assessments in food.

Toxicity

A wide range of toxic effects have been associated with alkaloids being present in food. Toxicity could vary from mild effects, such as vomiting and nausea, to more chronic effects, such as teratogenicity and sudden death. A summary of some of the toxic effects associated with these alkaloids is presented in Table 1.

Table 1.

Toxic effects of alkaloids

Alkaloids	Toxic Effects	References
Ergot Alkaloids	Decreased body weight gain, increased glycogen storage, increased lipogenesis, decreased glucose and thyroxin levels, increased organ weights, enlarged parathymal lymph nodes	Peters-Volleberg et al. (1996); Janssen et al. (2000a)
	Developmental toxicity, Genotoxicity	Kraicer and Strauss (1970); Carpent and Desclin (1969); Witters et al. (1975); Grauwiler and Schon (1973); Sommer and Buchanan (1955); Schaar and Clemens (1972); Roberts and Rand (1977a); Roberts and Rand (1977b); Roberts and Rand (1978); Matter (1976), Matter (1982); Dighe and Vaidya (1988)
Glycoalkaloids	Irritation, congestion in the epididymis and testes, gut bleeding, gastric glandular mucosal necrosis, intestinal mucosal necrosis, gastric distension lesions, severe renal and hepatic congestion, leukocytic infiltration, increased haemoglobin and haemocrit, decreased liver weight, increased white and red blood cells, developmental toxicity, genotoxicity	Al Chami et al. (2003); Phillips et al. (1996); Baker et al. (1989); Sharma et al. (1979); Phillips et al. (1996); Langkilde et al. (2008); Langkilde et al. (2009); Langkilde et al. (2012); Dixit and Gupta (1982); Gupta and Dixit 2002; Dixit et al. (1989); Gaffield and Keeler (1993), Gaffield and Keeler (1996); Renwick et al. (1984); Munari et al. (2012); Freidman and Henika (1992); Almeida et al. (2010).
Purine Alkaloids	Genotoxicity, carcinogenicity	Kulmann Fromme (1968); Choudhury and Palo (2004); Haynes et al. (1996); Aeschbacher et al. (1986); NTP (1998); Yu et al. (2011)
Pyrrolizidine Alkaloids	Vasoconstriction and medial hypertrophy of pulmonary arteries, cytoplasmic and pulmonary edema, chronic lung lesions, liver damage, hemangiosarcoma, alveolar/bronchiolar neoplasms, hepatopathy, hepatocytomegaly, developmental toxicity, genotoxicity, carcinogenicity	Wagenvoort et al. (1974); Miller et al. (1978); Culvenor et al. (1976); Schoental and Magee (1959); Chan et al. (1994); Chan et al. (2003); Peterson and Jago (1980); Green and Christie (1961); Schoental (1959); Silva-Neto et al. (2010); Petry et al. (1986); Hirono et al. (1977), Hirono et al. (1979); Rao and Reddy (1978)
Quinolizidine Alkaloids	Haemocrit, trembling, tonic–clonic convulsions, tonic–clonic spasms, liver problem, squinting, bilateral pupillary dilation, reddening of the eyelids, genotoxicity	Pothier et al. (1998); Yovo et al. (1984); Butler et al. (1996); Stanek et al. (2015); NTP (1978); Silva et al. (2014); Winks et al. (1998).
Tropane Alkaloids	Genotoxicity; carcinogenicity	McCann et al. (1975); Waskell (1978); EFSA Journal (2013); Cabello et al. (2001); Tatsuka et al. (1989).

Acute toxicity

Ergot alkaloids

LD₅₀ values have been determined for several ergot alkaloids, including ergometrine, ergotamine, ergosine and ergostine, via intravenous, subcutaneous or oral administration in rabbits and mice (Griffith et al., 1978). Generally, acute toxicity displayed by these alkaloids are low following oral administration, but the orally administered LD₅₀ values are always higher than intravenous LD₅₀ values for similar animal species, indicating low absorption and high pre-systemic metabolism subsequent to oral administration (Griffith et al., 1978).

Glycoalkaloids

When studied for toxicity in rats, α-solamargine has been found to have an intraperitoneal LD₅₀ of 42 mg/kg bw. At up to 50 mg/kg bw dose, it causes slight irritation and congestion in the epididymis and testes (Chami et al., 2003).

Poisoning from potato glycoalkaloids (PGA), particularly α-solanine and α-chaconine, can be attributed to the consumption of green potatoes (Phillips et al., 1996). In one study, a 25 mg/kg bw intraperitoneal injection of a mixture of α-solanine and α-chaconine (1:1, w/w) that was administered to hamsters had a lethal effect, caused by bleeding in the gut, but a 50 mg/kg bw oral dose of the same mixture had no effect (Phillips et al., 1996).

In another study, the oral administration of ground plant materials from Solanum plant species was shown to have lethal effects on some Syrian hamsters. These deaths were characterized by mucosal necrosis in the stomach and small intestine with little inflammation. Some animals survived but developed lesions that were characteristic of abdominal enlargement (Baker et al., 1989).

Another study reported increased levels of serum glutamic oxaloacetic transaminase (SGOT) and serum glutamic pyruvic transaminase (SGPT) and the decreased activity of serum cholinesterase (ChE) and microsomal enzymes, including cytochrome P-450 following the intraperitoneal administration of a 250 mg/kg dose of solanine to male rats (Dalvi, 1985). The histopathological effects of α-solanine and α-chaconine (which have intraperitoneal LD₅₀ of 32.3 mg/kg and 19.2 mg/kg, respectively) include severe renal and hepatic congestion and leukocytic infiltration (Sharma et al., 1979).

Pyrrolizidine alkaloids

Several pyrrolizidine alkaloids have been reported to cause acute toxicity. Senecionine and adonifoline have an acute oral toxicity corresponding to 57 mg/kg b.w. and 163 mg/kg b.w., respectively (Wang et al., 2011). Oral or intraperitoneal LD₅₀ values range from 34 to 38 mg/kg for retrorsine administration to male rats, while its N-oxide has LD₅₀ values of 48 mg/kg and 250 mg/kg for oral and intraperitoneal administration, respectively (Mattocks, 1971). Single oral doses of heliotrine, lasiocarpine and its N-oxide, retrorsine, riddelliine and seneciphylline produced significant toxicity in rats (Schoental and Magee, 1959).

Symptoms associated with acute toxicity from pyrrolizidine alkaloids include vasoconstriction, hypertrophy of the right ventricles and pulmonary arteries, pulmonary edema, cytoplasmic edema, lung lesions leading to chronic liver lesions and chronic liver damage (Culvenor et al., 1976; Miller et al., 1978; Schoental and Magee, 1959; Wagenvoort et al., 1974).

Quinolizidine alkaloids

LD₅₀ values associated with quinolizidine alkaloids include 100 mg/kg for sparteine, 250 mg/kg for lupanine and lupine extracts (Pothier et al., 1998), a range of 750–4000 mgkg⁻¹ body weight for several fractions of Lupinus angustifolius seeds (Stobiecki et al., 1993), 2279 mg/kg as a result of the oral administration of L. angustifolius, 1464 mg/kg and 177 mg/kg for the oral and intraperitoneal administration of lupanine, respectively, 199 mg/kg for the intraperitoneal injection of 13-hydroxylupanine (Petterson et al., 1987) and 36 mg/kg and 175 mg/kg for the oral administration and 220 mg/kg and 410 mg/kg for the intraperitoneal administration of sparteine and lupanine. (Yovo et al., 1984). At these lethal doses, the symptoms of intoxication include trembling, tonic–clonic convulsions and death by breathing arrest (Pothier et al., 1998; Yovo et al., 1984).

Repeated dose toxicity

Ergot alkaloids

A 28–32-day study on the subacute dietary toxicity of α-ergocryptine in rats showed body weight loss, body weight gain, food intake and food deficiency in U-shaped dose–response curves with a NOAEL of 4 mg/kg food (Janssen et al., 2000b). U-shaped dose–response relationships were also observed for hormonal and metabolic changes, which resulted in increased glycogen storage and lipogenesis. This situation could lead to an increase in protein catabolism (Janssen et al., 2000a).

A 4-week study on the subacute toxicity of ergometrine maleate in rats showed decreased glucose and thyroxin levels and increased organ weights when administered at high doses and an increase in the incidence of enlarged mediastinal and parathymal lymph nodes in male rats, which was proportional to the administered dose (Peters-Volleberg et al., 1996).

Glycoalkaloids

Langkilde et al. (2008; Langkilde et al. 2009; Langkilde et al. 2012) reported a series of experiments using α-solanine mixed with α-chaconine in female hamsters. Decreases in food consumed and relative weight of the liver and increases in haemoglobin and hematocrit all occurred within a 5-day period following the first day of the gavage administration of 75 mg TGA/kg bw per day of a mixture of the two glycoalkaloids in a 1:2.5 ratio (Langkilde et al., 2008). As with the previous dose, lethality was also observed following the gavage administration of 100 mg TGA/kg bw per day in a quarter of the tested animals due to increased adrenals, red blood cells and serum creatinine and decreased body weight (Langkilde et al., 2008).

In another study, a NOAEL of 10 mg TGAs/kg bw per day was gotten for a 1:3.7 ratio of α-solanine and α-chaconine, at a daily dose of 33 mg TGAs/kg b.w. that was administered via gavage to hamsters. The symptoms, which were mostly associated with the blood occurred within a month of the first day of administration (Langkilde et al., 2009). For a ratio of 1:7 at the same dose, death occurred within 2 h after the gavage administration in one of four treated animals (Langkilde et al., 2009). In a study on the compositional similarity and differential toxicity of a genetically modified potato line (SGT 9–2), the compositional differences observed were of no significance to nutritional value or safety. Additionally, differential toxicity data revealed a few significant differences between potato lines with different α-solanine: α-chaconine ratios, but these differences were determined to pose no safety concerns (Langkilde et al., 2012).

The consumption of green potatoes can lead to poisoning due to the presence of PGA, particularly α-solanine and α-chaconine, which increases in concentration when greening is taking place (Phillips et al., 1996). One study found that a single intraperitoneal injection of a 25 mg/kg bw dose or greater of a mixture of the two PGA (1:1, w/w) was lethal to laboratory animals and led to bleeding in the gut. Generally, the effects of the high concentrations of cytotoxic potato glycoalkaloids that are found in some potato tops have been shown to be minimal in laboratory animals (Phillips et al., 1996).

Pyrrolizidine alkaloids

The toxicity, carcinogenicity and the respective dose–response relationships of riddelliine in mice and rats showed lethal symptoms in rats that were administered 1.0 mg/kg doses. Non–neoplastic lesions were also observed on the livers and kidneys of the rats (Chan et al., 2003). As for the mice, males that were fed with 3.0 mg/kg doses experienced marked increases in hemangiosarcoma in their livers while females that were fed with the same dose experienced significant increases in alveolar/bronchiolar neoplasms (Chan et al., 2003).

In a similar experiment, an inverse relationship was observed between increases in body weight and dose among rats and mice. Dose-related hepatopathy and intravascular macrophage accumulation were observed in rats while mice experienced hepatocytomegaly. (Chan et al., 1994).

In another study, male rats showed more sensitivity than female rats, with increased cases of liver lesions, spleen, lungs and pancreas at a dose of 1.0 mg/kg or higher. The effects were less severe in female rats who received doses of 3.3 mg/kg per day or higher. Increased liver weights and hepatic cytomegaly were the only observed effects in male and female mice. Generally, the mice were considered to be less sensitive (Chan, 1993).

Quinolizidine alkaloids

In one study, feeding a group of rats a diet that was based on lupin flour led to reductions in hematocrit (HCT) and blood cell counts, depending on the dose received. The male group also experienced decreased mean cell volumes (MCVs) (Butler et al., 1996). In a similar experiment, a NOAEL of 330 ppm was reported when considering food intake levels and lower body weights at 1000 ppm and 5000 ppm doses while a NOAEL of 1000 ppm was predicted on the basis of hematology, growth rates, clinical chemistry and histopathological results (Robbins et al., 1996).

In another study, reduced food intake, changes in alanine transaminase activity and the parenchymal degeneration of the liver were all observed in animals that were fed with seeds from blue lupin cultivars (Stanek et al., 2015). Debittered Lupinus mutabilis seeds supplemented with DL-methionine could offer a good alternative source of protein for humans and animals as no harmful effects were observed in any of the parameters measured in a study on rats (Schoeneberger et al., 1987). Another study revealed normal weights in vital organs, such as the liver, kidneys, heart, spleen and adrenals, in rats that were dosed with 20% lupin protein, which was supplemented with DL-methionine. The growth rates and histology of the lungs and kidneys were also normal (Ballester et al., 1980).

The National Toxicology Program (NTP) in the USA reported a series of experiments on the toxicity of (-)-scopolamine hydrobromide trihydrate in rats and mice. In a 16-day study, there were no significant differences between the body weight gain and the final mean body weights of the control and dosed groups. However, death occurred in male and female mice that received the highest dose and a female mouse in the lowest dose group (NTP 1997). Squinting and the bilateral dilation of pupils were observed in all of the treated animals. Bilateral pupillary dilation and the reddening of eyelids were observed in a different 14-week study. In a 2-year study, bilateral dilation of the pupil occurred in rats and mice treated with (-)-scopolamine hydrobromide trihydrate but no changes were observed in their ophthalmic examinations (NTP 1997).

Developmental toxicity

Ergot Alkaloids

Ergocornine blocks ovulation in rats, which is an event that peaks at a time called the critical period. Ergocornine can also inhibit ovulation before the critical period but to a lesser extent. It also causes fluid persistence, which distends the proestrus uterus (Kraicer and Strauss, 1970). A study found that a 1 mg administration of ergocornine to a group of animals on the eighth day of pregnancy disturbed the pregnancy and had teratogenic effects on the fetuses (Carpent and Desclin, 1969). The intraperitoneal administration of elymoclavine, which is produced by Claviceps purpurea, at particular doses has been found to have teratogenic effects on mice and can kill embryos (Witters et al., 1975). The teratogenic effects include vertebral defects and the fusion of ribs.

Another study found that injections of ergotamine or ergotoxine into rats during gestation had adverse effects on lactation by not only diminishing the yield of milk but also the fat content of the secretory material (Sommer and Buchanan, 1955). Pregnant mice, rats and rabbits have been shown to suffer fetal retardation and increases in prenatal mortality as a result of the vasoconstrictive action of ergotamine that was orally administered (Grauwiler and Schon, 1973). In a different study, the subcutaneous administration of ergocornine hydrogenmalienate, ergonovine maleate, ergotamine tartrate and ergocryptine mesylate were found to affect the pituitary gland, which led to depressed levels of serum prolactin, thereby depressing lactation (Shaar and Clemens, 1972).

Glycoalkaloids

In one study, epididymal dysfunction occurred when solasodine was orally administered to dogs, demonstrating the antiandrogenic nature of the compound. Other effects included reductions in total protein, sialic acid, glycogen and acid phosphatase activity (Gupta and Dixit, 2002). Administering solasodine to dogs has also been found to render them infertile as sperm were absent in the cauda epididymis and ductus deferens (Dixit and Gupta, 1982). In another study on rhesus monkeys, solasodine interfered with the spermiogenesis process and led to a marked decrease in cauda sperm count (Dixit et al., 1989).

In a different study, α-solanine, α-chaconine as well as their aglycone, solanidine administered orally to Syrian hamsters resulted in craniofacial malformations (Gaffield and Keeler, 1996). Teratogenicity resulting from the administration of steroidal alkaloids, including solanidanes, to hamsters has been shown to have a direct relationship with or without unsaturation in the C-5 and C-6 positions of the alkaloids (Gaffield and Keeler, 1996). Renwick et al. (1984) reported the teratogenic nature of potato sprouts, which was attributed to the presence of solanidine triglycosides, α-chaconine and higher levels of α-solanine.

Pyrrolizidine alkaloids

Developmental toxicity following intraperitoneal administration of heliotrine and dehydroheliotridine to rats has been linked to distorted ribs, retarded ossification, long bones, cleft palates and foot defects (Green and Chrsitie, 1961; Peterson and Jago, 1980). Lasiocarpine and retrorsine have also been shown to produce significant liver lesions in the offspring of lactating rats but with no significant effects on the mothers (Schoental, 1959).

Genotoxicity

Ergot alkaloids

Some derivatives of ergot alkaloids, such as ergotamine, dihydroergotoxine and methysergide, can induce considerable levels of chromosomal damage in cultured human lymphocytes. These derivatives have been found to induce damage in male mice and early fetal deaths in mice at a dose of 100 mg/kg of dihydroergotoxine and methysergide (Roberts and Rand, 1977a). At all concentrations tested in one study, sister chromatid exchange (SCE) was observed in the ovaries of female hamsters following exposure to ergotamine, ergonovine and methylergonovine. Ergocristine has only been shown to induce SCE at high concentrations while α-ergocryptine has not been found to induce SCE at any tested concentrations (Dighe and Vaidya, 1988). However, in another study, three tested derivatives of ergot alkaloids showed no mutagenic effects in mice or Chinese hamsters at subtoxic or therapeutic doses (Matter, 1982; Matter, 1976,1976).

Glycoalklaoids

One study showed that Solanum lycocarpum fruit extract demonstrated protective activity against genomic and chromosomal damage caused by methanesulfonate, a DNA damage-inducing agent, rather than genotoxic effects in the lung fibroblasts (V79 cells) of Chinese hamsters (Friedman and Henika, 1992). In another study, ethanolic extract of Solanum palinacanthum and purified solamargine were tested using Ames test and showed no mutagenic effects in Salmonella typhimurium strains (Almeida et al., 2010).

Purine alkaloids

Caffeine is mutagenic in Drosophila melanogaster, can cause chromatid breakage in human cells and has dominant lethal effects in male mice (Kuhlmann and Fromme, 1968). At higher doses, it can induce sister chromatid exchange, micronuclei and chromosome aberrations in mice and Chinese hamsters (Choudhury and Palo, 2004; Haynes et al., 1996; Aeschbacher et al., 1986). Since caffeine is metabolized in humans and many mammals at a rapid rate, only high doses (100 mg/kg or higher) can induce chromosomal damage; and therefore, caffeine has low genotoxic potential when used as a flavoring agent (EFSA CONTAM Panel, 2017).

Pyrrolizidine alkaloids

Exposure to monocrotaline and dehydromonocrotaline can induce significant DNA damage and apoptosis in glioblastoma of human GL-15 cell line (Silva-Neto et al., 2010). Monocrotaline and jacobine have been shown to induce DNA crosslinks (DNA–DNA or DNA–protein), which has certain effects in adult male rats (Petry et al., 1984; Petry et al.,1986). These crosslinks have also been found to be induced by eight pyrrolizidine alkaloids and follow a particular potency order (Hincks et al., 1991). Further studies on crosslinks associated with these alkaloids have also been reported (Kim et al., 1995; Pereira et al., 1998; Wang et al., 2005).

Quinolizidine alkaloids

Sparteine (0.5 mM) and extracts from Lupinus mexicanus and Lupinus montanus have shown in vitro genotoxic activity when assessed using alkaline comet assays (Silva et al., 2014). In an experiment where ethanolic extracts of Lupinus termis were subjected to three genotoxic studies, the extracts were found to be non–genotoxic in all of the assays (Santiago Quiles et al., 2010). Additionally, in another study, binding or intercalation with deoxyribonucleic acid (DNA) did not occur for several quinolizidine alkaloids (Winks et al., 1998).

Tropane alkaloids

Atropine sulphate (McCann et al., 1975) and (-)-scopolamine (Waskell, 1978) have shown no mutagenic effects on different Salmonella typhimurium strains. EFSA has since confirmed that atropine and (-)-scopolamine produce no genotoxic effects (EFSA CONTAM Panel, 2013).

Carcinogenicity

Purine alkaloids

When a rats and mice were dosed with 75 mg/kg of theophylline in one study, there were no observed carcinogenic effects (NTP, 1998). Coffee has the potential to reduce the overall incidence of cancer and has inverse relationships with some types of cancers (Yu et al., 2011). The IARC (2016) concluded that “overall coffee drinking was evaluated as inclassifiable (group 3) as to its carcinogenicity to humans”.

Pyrrolizidine alkaloids

Riddelliine is classified as “possibly carcinogenic to human” (group 2B), according to the IARC (2002). The NTP classified it as “reasonably anticipated to be a carcinogen” (NTP, 2003). Liver hemangiosarcoma, hepatocellular adenoma and mononuclear cell leukemia all increased according to a carcinogenicity study on the administration riddelliine to rats and mice (Chan et al., 2003). The IARC classifies lasiocarpine as “possibly carcinogenic to humans” (group 2B) (IARC, 1976). In one study, liver angiosarcoma occurred in experimental animals following the administration of lasiocarpine via their diet, leading to the conclusion that lasiocarpine is carcinogenic in animals (NTP, 1978). The dietary administration of lasiocarpine to rats resulted in liver hemangiosarcoma and hepatocellular carcinoma in one study (Rao and Reddy, 1978) while intraperitoneal administration resulted in liver tumors and hepatocellular carcinoma in another study (Svoboda and Reddy, 1972). Other carcinogenic pyrrolizidine alkaloids include clivorine (Kuhara et al., 1980), petasitenine (Hirono et al., 1977), senkirkine and symphytine (Hirono et al., 1979).

Tropane alkaloids

In separate studies on Sprague–Dawley rats, atropine showed no effects on tumor development (Cabello et al., 2001; Schmäahl and Habs, 1976). However, atropine significantly increased gastric tumors per animal in a study that was conducted to investigate the possibility of atropine promoting experimental carcinogenesis in the stomachs of rats caused by N-methyl-N’-nitro-N-nitrosoguanidine (MNNG) (Tatsuta et al., 1989).

Analytical methods

For most alkaloids, instrumental analyses are usually preceded by extraction and sample clean–up steps. A summary of the various analytical methods that are used for alkaloids is presented in Table 2.

Table 2.

Analytical methods for the detection of alkaloids

Alkaloids	Samples	Sample preparation	LOD	LOQ	Analytical method	References
Ergot Alkaloid	Cereals and foods	SPE		0.17–2.78 μg/kg	LC–MS/MS	Krska et al. (2008)
	Rye and wheat	SPE		0.01–1.0 µg/kg (wheat) 0.01–10 µg/kg (rye)	UPLC-MS/MS	Kokkonen and Jestoi (2010)
	Beer				U-HPLC-Orbitrap MS	Zachariasova et al. (2010)
	Rye	SPE	0.25–1.65 μg/kg	1.10–6.65 μg/kg	HPLC–FLD	Franzmann et al. (2010)
	Rye and rye products	Solid-phase filtration	0.02–1.10 μg/kg	0.08v3.30 μg/kg	HPLC–FLD	Muller et al. (2009)
			8.1–22.3 μg/kg	25.5–66.5 μg/kg
	Claviceps purpurea				GC–MS	Jegorov et al. (1997)
	Sorghum ergot		0.01 mg/kg	0.1 mg	ELISA	Molloy et al. (2003)
	Grasses				ELISA	Tunali et al. (2000)
	Straw, Seed Digesta		1.5625 ng/mL		ELISA	Schnitzius et al. (2001
Glycoalkaloids	Tomato		0.39 μg		HPLC–UV	Kozukue et al. (2004)
			0.94 μg
	Potatoes	SPE; C-18 cartridge			RP-HPLC	Houben and Brunt (1994)
	Blood	SPE	1 μg/L	2 μg/L	UHPLC-MS/MS	Nara et al. (2019)
	Milk	LLE		0.28 mg/L	GC-NPD	Bushway et al. (1984)
	Potato and tomato				ELISA	Stanker et al. (1994)
	Potatoes				ELISA	Morgan et al. (1985)
Purine Alkaloids	Tea, Coffee and Urine	Filtration	0.10 μg/mL (caffeine)	0.33 μg/mL	RP-HPLC/UV	Aragao et al.(2005)
			0.07 μg/mL (theobromine)	0.23 μg/mL
			0.06 μg/mL (theophylline)	0.18 μg/mL
	Tea	Infusion	0.07–0.47 mg/L	0.21–1.41 mg/L	HPLC–DAD	Azevedo et al. (2019)
	Tea		0.025 μg/mL (caffeine)	0.08 μg/mL	UHPLC-MS	Aqel et al. (2019)
			0.015 μg/L (theobromine)	0.05 μg/mL
			0.01 μg/mL (theophylline)	0.03 μg/mL
	Tea	Leaching			HPLC	Jankech et al. (2019)
	Beverage and Food	Drop-to-drop solvent microextraction	4.0 ng/mL		GC–MS	Shrivas and Wu (2007)
	Beverage	Ultrasonication	0.01 mg/L		GC–MS	Zou and Li (2006)
	Soft drinks, tea and coffee				FTIR	Paradkar and Irudayaraj (2002)
	Yerba mate	Cellulose acetate membrane	1.6 × 10–3 mg/kg	4.2 × 10–3 mg/kg	NIR	Mazur et al. (2014)
	Coffee and tea	Infusion			UV–Visible spectrophotometry	Lopez-Martinez et al. (2003)
Pyrrolizidine Alkaloids	Herbal Medicines	Mixed mode (MCX) cationic exchange SPE cartridge	0.03–3.37 μg/kg		LC–ESI–MS/MS	Jeong et al. (2021)
	Fodder, hay, silage	SPE, Strata 60 mg, 3 cc cartridges	3–10 μg/kg	10 – 30 μg/kg	UHPLC-MS/MS	Mudder et al. (2009)
	Plant, milk and urine	SPE		0.05–0.2 μg/L (milk)	LC–MS/MS	Hoogenboom et al. (2011)
				0.2v0.5 μg/g (dried plant)
				0.2–0.5 μg/L (urine)
	Honey and herbal tea	SPE	0.5 μg/kg (honey)	1 μg/kg (honey)	UHPLC-MS/MS	Reinwaldt et al. (2017)
			2.5–10 μg/kg (herbal tea)	5–20 μg/kg (herbal tea)
	Senecio inaequidens	Soxhlet			GC–MS	Caniato et al. (1989)
	Senecio alpinus	Soxhlet			GC–MS	Luthy et al. (1981)
	Senecio jacobaea	Extraction with DCM			GC–MS	Pelser et al. (2005)
	Senecio species	Extraction with H2SO4			GC–MS	Witte et al. (1992)
	Herbal remedies	SPE			GC–MS	Conradie et al. (2005)
	Petasites hybridus	Extraction with methanol	0.11–1.53 ng/mL	0.1 ppm	ELISA	Langer et al. (1996)
	Senecionine		0.5–10 ppb		ELISA	Roseman et al. (1996)
			5–100 ppb
			600–1000 ppb
	Antibodies				ELISA	Zundorf et al. (1998)
Quinolizidine Alkaloids	Holstein cattle	Rotary evaporation, Mechanical rotation, DCM			HPLC–MS	Green et al. (2015)
	Lupin flours, Lupin-based ingredients and food	Soxhlet extraction	2 mg/kg		GC–MS	Resta et al. (2008)
	Lupin seeds	Soxhlet	2 mg/kg		GC–MS	Boschin et al. (2008)
	Lupin seeds and Lupin-containing foods	Centrifugation, silica cartridge	1 mg/kg		GC–MS/GC-FID	Reinhard et al. (2006)
			6 mg/kg
			2 mg/kg
	Lupinus Species	Centrifugation			GLC, GC-EIMS	Wink et al. (1995)
Tropane Alkaloids	Plasma	Centrifugation	0.004–0.01 ng/mL	0.05–0.8 ng/mL	LCI-ESI–MS/MS	John et al. (2010)
	Atropine	Water dilution			HPLC–UV/MS	Breton et al. (2005)
	Feedstuff and Biological Samples	SPE, C-18 cartridge	12.05 ng (scopolamine)	37.98 ng (scopolamine)	RP-HPLC/UV	Papadoyannis et al (1993)
			13.25 ng (hyoscyamine)	38.17 ng (hyoscyamine)
	Buckwheat	Soxhlet	0.3 μg/kg (atropine)	1 μg/kg (atropine)	GC–MS	Caligiani et al. (2011)
			1.0 μg/kg (scopolamine)	6 μg/kg (scopolamine)
	Antibodies	Methanol, DCM	0.1 ng		ELISA	Fliniaux and Jacquin-Dubreuil. (1987)
	Datura	28% EtOH/NH4OH (9:1), Centrifugation	0.15 ng		Radioimmunoassay	Savant and Dougall (1990)

Extraction and sample clean-up

Liquid extraction using various solvent mixtures is a common approach that is applied to extract ergot alkaloids from sample sources. Extraction can be achieved using non–polar solvents or the addition of methanol and ammonium hydroxide to some solvents, such as dichloromethane and ethyl acetate, which raises the pH and increases the solubility of ergot alkaloids in the organic layer (Franzmann et al., 2010; Lauber et al., 2005; Müller et al., 2009). In addition, buffers or dilute acid can be added to polar solvents to lower the pH and facilitate the extraction of ergot alkaloids (Krska et al., 2008; Mulder et al., 2012; Storm et al., 2008). Solid–phase extraction (SPE) using alumina cartridges (Franzmann et al., 2010; Müller et al., 2009), dispersive primary and secondary amine (PSA) SPE cartridges (Krska et al., 2008), cation ion exchange cartridges (Storm et al., 2008) and C18 cartridges (Fajardo et al., 1995; Mohamed et al., 2006) have all been used for sample clean up.

Glycoalkaloid extraction needs to be controlled to avoid the enzymatic degradation of glycoside side–chains by glycosidases that are present in fresh potato tubers (Freidman et al., 1997). The extraction can be achieved by using mixtures of aqueous and acidic solvents (Bodart et al., 2000). Further sample clean–up is needed and can be achieved using solid phase extraction cartridges (Freidman et al., 1997).

Purine alkaloids are mostly extracted using hot water (Aqel et al., 2019; Azevedo et al., 2019; Khanchi et al., 2007; Mazur et al., 2014) or water mixed with organic solvents, such as methanol and acetonitrile (Cimpoiu et al., 2010; Uysal et al., 2009). Extraction can also be achieved using supercritical extraction plants, which are designed for working conditions of up to 690 bar in pressure and 250 °C in temperature (Saldaña et al., 2002).

Extraction of pyrrolizidine alkaloids from sample sources can be achieved by using organic solvents (e.g., methanol) (Langer et al., 1996), sulfuric acid (Reinwaldt et al., 2017) or methanol mixed with sulfuric acid (Jeong et al., 2021). Sample clean–up is achieved using SPE cartridges (Jeong et al., 2021; Reinwaldt et al., 2017).

In some cases, the extraction of quinolizidine alkaloids from powdered lupin seeds is usually preceded by a hexane-defatting step (Boschin et al., 2008; Pothier et al., 1998; Yovo et al., 1984). After alkalinizing the defatted powder using aqueous ammonia, extraction can then be achieved using dichloromethane (Pothier et al., 1998) or chloroform (Yovo et al., 1984). Acidic solutions can also be used to extract quinolizidine alkaloids from sample matrices (Boschin et al., 2008; Kamel et al., 2016).

Pressurized solvent extraction, supercritical fluid extraction and microwave-assisted extraction methods can all be applied for the extraction of tropane alkaloids (Christen et al., 2008). Similar to quinolizidine alkaloids, samples need to first be defatted using hexane, alkalinized and then subjected to static extraction using dichloromethane to obtain tropane alkaloids (Caligiani et al., 2011). Different modifications to the solid phase extraction method have also been employed for extraction and sample clean–up (Long et al., 2012; Mroczek et al., 2006; Papadoyannis et al., 1993).

Chromatographic techniques

Chromatography is an important technique for detecting alkaloids in various sample sources. Methods based on high-performance liquid chromatography, gas chromatography and thin-layer chromatography have all been used to confirm the presence of various alkaloids in different samples.

HPLC-based methods

HPLC and LC–MS/(MS) are commonly to detect alkaloids. LC–MS/MS has been applied to analyze ergot alkaloids in various products, including rye and wheat, and it has the advantage of short analysis times. (Kokkonen and Jestoi, 2010; Krska et al., 2008). Liquid chromatography can be used in combination with various detectors to produce hyphenated methods, such as UHPLC-Orbitrap-MS (Zachariasova et al., 2010) and HPLC–FLD (Franzmann et al., 2010; Müller et al., 2009), which have found applications in the analysis of mycotoxins in beer and ergot alkaloids in rye and rye products.

The HPLC–UV method has been used successfully for the analysis of various glycoalkaloids in food (Friedman et al., 1997; Houben and Brunt, 1994; Kozukue et al., 2004). LC–MS/MS, LC-HRMS and LC-Orbitrap-MS have grown increasingly popular for glycoalkaloid analysis as they can be used in the targeted analysis of glycoalkaloids, exploratory research and metabolomic studies (Caprioli et al., 2014; Hossain et al., 2015; Moco et al., 2007; Nara et al., 2019).

Several purine alkaloids have been analyzed using various modifications to the liquid chromatography techniques. HPLC (Jankech et al., 2019), RP-HPLC/UV (de Aragão et al., 2005), HPLC with a diode array detector (Azevedo et al., 2019) and UHPLC-MS (Aqel et al., 2019) have all found applications in methylxanthine analysis.

The non–destructive nature of liquid chromatography techniques allows for the simultaneous determination of N-oxides and free bases in a single analytical run without the initially reducing the oxides (Crews et al., 2010). This technique has been used widely to analyze pyrrolizidine alkaloids in herbal medicines (Jeong et al., 2021), honey and herbal teas (Reinwaldt et al., 2017) and fodder, hay or silage from Senecio plant parts (Mulder et al., 2009), as well as plants, milk, urine and feces (Hoogenboom et al., 2011).

LC-based methods are uncommon for detecting quinolizidine alkaloids. However, there have been a few reports on the application of this method for determining quinolizidine alkaloids, such as in the serum analysis of Holstein cows after exposure to Lupinus leucophyllus (Green et al., 2015), as well as some conference abstract papers (Mulder and de Nijs, 2015; Vanérková et al., 2014).

UV and MS are common detectors that are used with liquid chromatography to analyze tropane alkaloids. The detection of these alkaloids in plasma (John et al., 2010), fodder and biological samples (Papadoyannis et al., 1993) and the chiral separation of atropine (Breton et al., 2005) have all been carried out using liquid chromatography.

GC-based methods

The determination of ergot alkaloids can be achieved using GC-based methods. The detection of ricinoleic acid, which is a marker for ergot alkaloid contamination in rye and products made from rye (Franzmann et al., 2010), and homoisoleucine, an amino acid present in ergogaline (Jegorov et al., 1997) are examples of GC applications in ergot alkaloid determination.

GC can be applied to glycoalkaloid analysis in two ways. The first method is hydrolyzing glycoalkaloids into their aglycones followed by GC analysis. The second method involves the derivatization and subsequent determination of glycosides (Friedman et al., 1997). These principles have been applied to the glycoalkaloid analysis of bovine milk using a nitrogen–phosphorus detector (Bushway et al., 1984) and the separation of solanidine, demissidine and all other C27–steroidal alkaloids in a certain solanum species (van Gelder, 1985).

A few reports have been published on the use of GC–based methods to analyze purine alkaloids. A method for the rapid determination of caffeine in beverages and food was developed by Shrivas and Wu. This method adopts a drop-to-drop solvent microextraction using gas chromatography–mass spectrometry (Shrivas and Wu, 2007). Zou and Li (2006) also determined caffeine concentrations in beverages using GC–MS. The LOD of this method is 0.001 mg/L.

GC–MS has been widely applied in the determination of pyrrolizidine alkaloids in different samples. There are a few reports on the application of this method for determining quinolizidine alkaloids in different Senecio species (Caniato et al., 1989; Luthy et al., 1981; Pelser et al., 2005; Witte et al., 1992). Edgar and Smith (2000) used GC–MS to study the transfer of pyrrolizidine alkaloids in eggs. There has also been a case report about using GC–MS to identify toxic pyrrolizidine alkaloids in traditional medicines that were administered to two sets of twins in Africa (Conradie et al., 2005).

GC–MS is a common method used to detect quinolizidine alkaloids in different variety of lupin-based products. Lupanine, 13α-hydroxylupanine, angustifoline and other pyrrolizidine alkaloids have been found in lupin seeds, lupin flours and different lupin species (Boschin et al., 2008; Reinhard et al., 2006; Resta et al., 2008; Wink et al., 1995). GC–MS and GC–FID can also be applied in parallel for this analysis (Reinhard et at., 2006).

A number of reviews have been conducted on the GC analysis of tropane alkaloids (Aehle and Dräger, 2010; Drager, 2002). GC–MS in single-ion mode has been used to detect toxic tropane alkaloids in commercial buckwheat (Caligiani et al., 2011). Tropane alkaloids are abundant in Datura stramonium seeds and their presence can be confirmed using GC–MS (El Bazaoui et al., 2009; Philipov and Berkov, 2002).

Immunoassays

Most of the quantitative immunoassays for detecting rye ergot alkaloids that have been described in the literature are insensitive to dihydroergosine (Molloy et al., 2003). An enzyme-linked immunosorbent assay (ELISA) method for measuring dihydroergosine was reported by Molloy et al., which involved using dihydroergosine–specific monoclonal and polyclonal antibodies against dihydroergosine conjugated to bovine serum albumin from mouse and rabbits, respectively (Molloy et al., 2003). Tunali et al. (2000) used ELISA and HPLC to measure levels of ergot alkaloids present that were present in native Turkish grasses. The ELISA method is more desirable for analyzing large numbers of forage samples and can detect the quantities of non–specific ergot alkaloids in tall fescue infected by endophyte (Schnitzius et al., 2001).

In the literature, ELISA assays have also been applied to quantify glycoalkaloids. For example, the total glycoalkaloid content in potato tubers has been determined using ELISA assays. The ELISA methods show good sensitivity and no cross–reactivity with tomatidine derivatives (Morgan et al., 1985). Stanker et al. (1994) was able to develop an ELISA method that helps to distinguish between glycoalkaloids present in potato and tomato and their corresponding aglycons.

Langer et al. developed an immunoassay using antibodies against retrorsine to determine the pyrrolizidine alkaloid senecionine, which is found in Asteraceae (Langer et al., 1996). Roseman et al. (1996) used an ELISA method based on polyclonal antibodies to detect the pyrrolizidine alkaloids retrorsine, monocrotaline and retronecine. A polyclonal antibody-based ELISA has also been applied to detect riddelliine and its N-oxide, with an LOD of 25 µg/L for riddelliine-type pyrrolizidine alkaloids in blood (Lee et al., 2001). In contrast to other references, Zündorf et al. (1998) used a monoclonal antibody-based ELISA for the detection retrorsine.

Savary and Dougall, (1990) developed a sensitive radioimmunoassay to investigate the yields of scopolamine from Datura species cultures. The commercially labeled antigen L-(N-methyl [³H])-scopolamine methyl chloride is adopted in the assay, which measures scopolamine in the range of 0.15–3.0 ng (Savary and Dougall, 1990). The wide specificity of antibodies from immunized rabbits using racemic tropic acid conjugated with bovine serum albumin has been shown to enable the simultaneous analysis of the main tropane alkaloids in Solanaceous plants (Fliniaux and Jacquin-Dubreuil, 1987).

Spectroscopic methods

Spectroscopic methods can also be applied to detect and quantify some alkaloids. Near–infrared (NIR) spectroscopy was used to detect and quantify ergot alkaloids in tall fescue (Roberts et al., 1997) and cereals (Vermeulen et al., 2012). Hyperspectral images have also been adopted to measure ergot contamination in cereals. (Vermeulen et al., 2012).

Paradkar and Irudayaraj, (2002) determined the amount of caffeine in a variety of soft drinks and the total methylxanthine content in tea and coffee using a rapid Fourier–transform–infrared (FT-IR) spectroscopic method in a single calibration model. Two spectral regions (1500–1800 cm⁻¹ and 2800–3000 cm⁻¹) were identified in the FT–IR spectrum of pure caffeine and were used for quantitative estimation. NIR spectroscopy has been used in combination with multivariate calibration techniques to quantify methylxanthines in yerba mate (Ilex paraguariensis) (Mazur et al., 2014). López-Martinez et al. used UV–visible spectrophotometry and the partial least squares method to simultaneously determine methylxanthines in coffee and teas. This method is not statistically significantly different from the standard HPLC technique (López-Martínez et al., 2003).

Nuclear magnetic resonance (NMR) spectroscopy is mainly used for the structural elucidation of pyrrolizidine alkaloids. Comprehensive tables of carbon–13 NMR (Roeder, 1990) and proton NMR (Logie et al., 1994) spectroscopic data for pyrrolizidine alkaloids are available in the literature. Birecka et al., (1981) stoichiometrically reacted protonated pyrrolizidine alkaloids with methyl orange and measured the intensity of the dye spectrophotometrically after the methyl orange had been released from the complex. This method is capable of detecting alkaloids at a concentration of 0.5 µg/mL. Alternatively, Azadbakht and Talavaki (2022) used Ehrlich’s reaction and spectrophotometric detection to measure pyrrolizidine alkaloids in wheat and flour. This method allows for the calculation of pyrrolizidine alkaloids and their N-oxides on the basis of senecionine (Azadbakht and Talavaki (2022).

Occurrence in food

Ergot alkaloids

A comparative study of total ergot alkaloids in rye meals and grains harvested using organic and conventional methods in 2003 and 2004 was conducted by Lauber et al. (2005). Using the HPLC-FLD method, they reported higher levels of ergot alkaloids in 2003 compared to 2004. The same method was adopted to analyze 12 ergot alkaloids in rye and rye products in another study. Of the 39 samples analyzed, the total mean concentrations of ergot alkaloids in rye flour, rye, coarse rye and rye flakes were reported as 137.5 µg/kg, 62.2 µg/kg, 157.7 µg/kg and 26.4 µg/kg, respectively. The maximum LOQ has also been reported for ergometrine (3.3 µg/kg) using the blank method and ergocristine (66.5 µg/kg) using the calibration method.

Crews et al. (2009) used LC–MS/MS to determine six ergot alkaloids and their epimers (i.e., the “ine” and “inine” forms) in 28 rye-based cereal products from the United Kingdom. The analyzed samples included rye crispbread, bread mix and crackers, etc. Other than three samples, ergot alkaloids were present in all the other samples. The highest levels of alkaloids were found in the rye crispbread, with concentrations of 131 µg/kg, 183 µg/kg and 340 µg/kg (Table 3). The LOD and LOQ values of the alkaloids ranged from 0.4–2.7 µg/kg to 0.5–2.8 µg/kg respectively.

Table 3.

Occurrence of alkaloids in food

Alkaloids	Food	Alkaloid concentration (mg/kg)	References
		Mean (Minimum – Maximum)
Ergot Alkaloids	Rye meals and rye grains	0.588 (max. 3.28)	Lauber et al., (2005)
	Rye Products	0.119 (max. 0.118)	Muller et al., (2009)
	Rye-based cereal	0.4872 (max. 0.340)	Crews et al., (2009)
	Rye flour	0.213 (max. 1.063)	Reinhold and Reinhardt, (2011)
	Wheat	0.444 (0.0103–0.975)	Topi et al., (2017)
Glycoalkaloids	Potatoes	405.55 (1.3—352.6)	Freidman et al., (2003)
	Potatoes	81.05 (33.69—167.79)	Urban et al., (2018)
	Potatoes Tomatoes	410.35 (94.40—750.0) (42—1498) dehydrotomatine (521—16,285) α-tomatine	Musita et al., (2020) Kozukue et al., (2004)
	Eggplant	(10.00—221.00)	Sanchez-Mata et al., (2010)
Purine Alkaloids	Beverage	(1 × 10–6—350)	Bispo et al., (2002)
	Leaves, branches and fruits	(400—8600)	Borre et al., (2010)
	Teas and coffees	(< 0.06—771.40)	Aragao et al., (2005)
	Teas, coffees, energy drinks, chocolates	80—18,740	Sanchez, (2017)
	Teas	ND—11,064 (ND—17,177.8)	Baek et al., (2022)
Pyrrolizidine Alkaloids	Tea, herbal drugs and honey	0.05—1.86 (< LOD—5.647) tea	Bodi et al., (2014)
		ND—0.45 (< LOD—3.098) herbal drugs
		0.006—0.015 (0.0003—0.234) honey
	Herbal teas and medicines	(< LOQ—7.88)	Chen et al., (2019)
		(0.3—1.12)
	Wheat flour Honey and bee pollen	(0.045—5.4) 0.026 (0.001—0.267)	Kakar et al., (2010) Dubecke et al., (2011)
		1.846 (0.011—37.855)
	Pollen products	5.17 (1.08—16.35)	Kempf et al., (2010)
Quinolizidine Alkaloids	Lupin	(ND—1664)	Gresta et al., (2010)
	Lupin flours, Lupin-based ingredients and food	(8.45—1247)	Resta et al., (2008)
		(< LOQ—86.5)
		(< LOQ—56.3)
	Lupin seed crops	(130 -52,380)	Magalhães et al., (2017)
	Lupin species	(20—21,920)	Cortes  et al., (2005)
		(ND—5670)
		(10—27,580)
	Lupin seeds and foods	(202—19,800) bitter	Reinhard et al., (2006)
		(44—2120) sweet
Tropane Alkaloids	Buckwheat	(ND—1283)	Caligiani et al., (2011)
	Teas and herbal teas	(0.005—4.34)	Romera-Torres et al., (2018)
	Cerealsand Solanaceae seeds	(0.003—1.847)	Marin-Saez et al., (2017)
	Cereals	(ND—0.0096)	Basle et al., (2020)
	Cereal-based food	0.00037—0.00304 (max. 0.0808)	Mulder et al., (2015)

Reinhold and Reinhardt, (2011) determined the levels of six ergot alkaloids and their corresponding epimers in 31 samples of rye flour and whole-meal rye flour from stores that sell retail products in Germany. Analysis was done using HPLC–MS/MS, with the concentrations of the alkaloids and the LOQ given as the sums of the alkaloids and their epimers. Individual alkaloid concentrations ranged from 86 µg/kg for ergocornine and its epimer to 365 µg/kg for ergotamine and its epimer.

In another study, the occurrence of twelve ergot alkaloids was determined in seventy-one samples of wheat harvested between 2014 and 2015 in Albania using LC–MS/MS. The concentrations of the alkaloids ranged from 17.3 µg/kg to 975.5 µg/kg in 2014 and from 10.3 µg/kg to 390.5 µg/kg in 2015. LOD and LOQ values for all the ergot alkaloids were 3 µg/kg and 10 µg/kg, respectively (Topi et al., 2017).

Glycoalkaloids

Friedman et al. (2003) analyzed two major glycoalkaloids (α-solanine and α-chaconine) in eight potato cultivars using HPLC. The analysis was carried out on freeze-dried potato, fresh (wet) potato, potato flesh, potato flesh, potato peel and whole potatoes from eight potato cultivars grown in the USA. The highest value for the sum of the two analyzed glycoalkaloids was found in the potato peel of the cultivar from Snowden (3526 mg/kg) while the lowest value was obtained in the wet potato of the cultivar from Russet Norkota.

Urban et al., (2018) reported the effects of genotype, flesh color and environment on the glycoalkaloid contents of potato tubers. They compared the levels of α-chaconine, α-solanine and total glycoalkaloids in 14 cultivars that had red and purple flesh to those in cultivars that had yellow and white flesh. The three-year study revealed that the total glycoalkaloid content in tuber flesh ranged from 33.69 to 167.77 mg/kg fresh matter (FM) while the ratio of α-chaconine to α-solanine ranged from 1.18 to 3.78. The total glycoalkaloid contents of the cultivars were affected by the cultivar genotype, location and weather conditions but were not affected by flesh color. The cultivar with purple flesh from Bora Valley had the highest average total glycoalkaloid content (165 mg/kg FM) while the cultivar with red flesh from Red Emmalie had the lowest total glycoalkaloid level (43.6 mg/kg FM).

A study on the glycoalkaloid levels of commercial potato varieties in Nairobi, Kenya was conducted by Musita et al. (2020) using HPLC. The potatoes were sampled randomly from the varieties that are sold in supermarkets and open-air markets. The highest glycoalkaloid level was found in the Shangi variety while the Royal variety had the lowest glycoalkaloid level.

Kozukue et al. (2004) determined dehydrotomatine and α-tomatine levels in tomatoes and vegetative plant tissues using the HPLC–UV method. In tomatoes, dehydrotomatine and α-tomatine were observed in the ranges of 42–1498 µg/g and 521–16,285 µg/g, respectively.

In one study, liquid chromatography–mass spectrometry was used for the analysis of two glycoalkaloids (α-solasonine and α-solamargine) in Gboma (Solanum macrocarpon L.) and Scarlet (Solanum aethiopicum L.), which are eggplants that are common in traditional sub-Saharan African culture. The results were compared and showed that S. macrocarpon had α-solasonine and α-solamargine in similar proportions (76–89% α-solamargine) unlike S. aethiopicum, which had a more variable composition (48–89% α-solamargine). The implication of the obtained results is that S. macrocarpon is unsafe for human consumption as its glycoalkaloid levels are too high (5–10 times higher than recommended levels in food), but S. aethiopicum has glycoalkaloid levels that are similar to S. melongena and can be consumed by humans (Sánchez-Mata et al., 2010).

Purine alkaloids

Sanchez (2017) determined the methylxanthine contents of some commonly consumed food products in Spain. Of all the samples investigated, caffeine was the most common methylxanthine that was found. In green teas, the mean concentration of caffeine was 13,085 mg/kg, with a range of 8693–18,740 mg/kg. In black teas, the concentration range of caffeine was 7754–19,803 mg/kg, with a mean concentration of 14,530 mg/kg. The concentration levels of caffeine in energy drinks and espresso coffee were similar, with concentration ranges of 80–160 mg. Theobromine was the methylxanthine that was most detected in chocolate, with concentrations ranging between 4000 and 10,000 mg/kg for dark chocolate.

In another investigation, HPLC was used to identify three methylxanthine derivatives in urine and drinks. The concentrations of methylxanthine, caffeine, theobromine, and theophylline were found in the urine and beverages in the ranges of 0.1 pg/mL–350 g/mL, 0.1 pg/mL–32, 0.1 pg/mL–13.2 g/mL, and 0.1 pg/mL–47, 0.1 pg/mL–66.3 g/mL, respectively (Bispo et al., 2002).

Another study used RP-HPLC/UV to measure caffeine, theobromine and theophylline levels simultaneously in tea, coffee and human urine. LOD values were calculated as 0.10 µg/L for caffeine, 0.07 µg/L for theobromine and 0.06 µg/L for theophylline. The highest concentration of caffeine was found in soluble coffee, with a value of 771.40 µg/mL. Mate tea had the highest concentration of theobromine (13.40 µg/mL) while Sene tea (15.50 µg/mL) had the highest concentration of theophylline. The concentrations of these methylxanthines in human urine was between 0.8 to 8.70 µg/mL for caffeine, 1.40 to 52.4 µg/mL for theobromine and from1.20 to 42.30 µg/mL for theophylline (de Aragão et al., 2005).

Baek et al. (2022) determined the methylxanthine contents of different tea types in Korea using HPLC. The mean contents of caffeine, theobromine and theophylline in the 83 leached extracts of 11 different types of tea were 5561.5, 407.3 and 24.8 mg/kg, respectively. The concentrations ranges were ND–17177.8, ND–2220.5 and ND–223.2 mg /kg for caffeine, theobromine and theophylline, respectively. The limit of quantitation values were 0.0254, 0.0144 and 0.0721 mg/kg while the limit of detection values were 0.0084, 0.0047 and 0.0238 mg/kg for caffeine, theobromine and theophylline, respectively.

In a study to compare methylxanthine, phenolic and saponin contents in the leaves, branches and unripe fruits of Ilex paraguariensis A. St.-Hil (Mate), the lowest concentrations of methylxanthines were found in unripe fruits. The highest concentrations of caffeine were in leaves while the highest concentrations of theobromine were in branches. The concentrations of caffeine and theobromine in commercial mate products were 0.79 and 0.30 g/100 g, respectively (Borré et al., 2010).

Pyrrolizidine alkaloids

Bodi et al. (2014) determined the pyrrolizidine alkaloids (PAs) in 274 teas, 41 herbal drugs and 87 honey samples. In the tea samples, the concentrations of pyrrolizidine alkaloids ranged from < LOD to 5647.2 µg/kg while the mean concentrations of the alkaloids ranged from 51.6 to 1856 µg/kg. The maximum PA concentration among the herbal drugs was found in anise, with a value of 3098.8 µg/kg and a mean concentration of 452.6 µg/kg. The honey samples recorded PA concentrations ranging from 0.3 to 234.5 µg/kg and a mean concentration range of 6.1–14.5 µg/kg.

Chen et al. (2019) conducted a risk assessment of pyrrolizidine alkaloid absorption from herbal teas and medicines and determined the concentration range of PAs in the tested samples to be < LOQ–7883 µg/kg for herbal medicines and 30.7–1120 µg/L for herbal teas.

Since pyrrolizidine alkaloids are known to cause hepatic veno-occlusive disease, plant extracts and wheat flour were tested for contamination by pyrrolizidine alkaloids in one study following an outbreak of the disease in Western Afghanistan. In terms of plant matter, lasiocarpine from the affected area (quantified against senecionine) had the highest concentrations of PAs in its stems, leaves, fresh roots and seeds. The flour samples from the affected area had a median PA concentration of 5.6 mg/kg with heliotrine-N-oxide having the highest median value of 5.4 mg/kg (Kakar et al., 2010).

Dübecke et al. (2011) analyzed 3917 honey samples and 119 bee pollen samples using LC–MS/MS. In total, 94% of retail honeys contained pyrrolizidine alkaloids while 66% of raw honeys tested positive for the presence of PAs. PAs were found in 60% of the bee pollen samples, with concentrations ranging from 11 to 37,855 µg/kg.

Pollen products, which are used as food supplements, were also tested for the presence of pyrrolizidine alkaloids in another study. Of the 55 commercial samples tested, 17 (31%) contained pyrrolizidine alkaloids, with a concentration range of 1080 to 16,350 µg/kg. (Kempf et al., 2010).

Quinolizidine alkaloids

Magalhães et al. (2017) identified and quantified alkaloids from lupin derivatives, which are of commercial interest in Europe. The quinolizidine alkaloids that were detected included lupanine, angustifoline, α-isolupanine and sparteine, with the total concentration of alkaloids in the seeds ranging from 690 to 24,950 mg/kg. The concentration range of each alkaloid expressed as percentage of the total quinolizidine alkaloids was 550–20,750 mg/kg for lupanine, 90–3830 mg/kg for angustifoline, 40–550 mg/kg for α-isolupanine and not detected–10 mg/kg for sparteine.

GC–MS/GC-FID was used in parallel for the analysis of the quinolizidine alkaloids lupanine, 13α-hydroxylupanine, α-isolupanine and angustifoline in lupin flours with high alkaloid contents. The total alkaloid concentrations in the bitter varieties ranged from 10,800 to 19,800 mg/kg and those in the sweet varieties ranged from 44 to 2120 mg/kg (Reinhard et al., 2006).

In another study, a total of 27 lupin-based products, including flours, protein isolates and food products, were analyzed for their quinolizidine alkaloid contents using GC–MS (Resta et al., 2008). In total, five alkaloids were detected in L. albus cv Arés (albine, lupanine, 13α-hydroxylupanine, 13α-angeloyloxylupanine and 13α-tigloyloxylupanine) and L. albus cv Typtop (albine, lupanine, multiflorane, 13α-hydroxylupanine and 13α-angeloyloxylupanine). Only lupanine was detected in the protein isolates. The total alkaloid concentrations in the Arés and Typtop were 146 mg/kg and 1247 mg/kg, respectively (Resta et al., 2008).

Additionally, Cortes et al. (2005) investigated variations in alkaloids during germination in different lupin species. The total quinolizidine alkaloids contents in the raw seeds were 1.51 g/100 g for L. angustifolius, 2.36 g/100 g for L. albus and 2.45 g/100 g for L. campestris. In L. albus, lupanine increased during germination while albine and 13-hydroxylupanine decreased significantly.

Sparteine, lupanine, 13α-hydroxylupanine, angustifoline and α-isolupanine were all detected using GC–MS in a study by Gresta et al. (2010), which evaluated the productive and nutritional characteristics of several sweet varieties of lupin seeds. In the Multitalia variety, the total concentration of alkaloids was 1664 mg/kg. Lupanine had a total composition of 1499 mg/kg while the other quinolizidine alkaloid compositions were 99, 34, 9.1 and 2.1 mg/kg for 13α-hydroxylupanine, angustifoline, α-isolupanine and sparteine, respectively. LOQ for the study was 0.2–0.4 mg/kg.

Tropane alkaloids

In one study, the tropane alkaloids atropine and scopolamine were detected in buckwheat fruits, flours and commercial food products using GC–MS. Highest concentration of alkaloids was found in Datura stramonium seeds, with atropine and scopolamine having concentration values of 1283 and 678 mg/kg, respectively (Caligiani et al., 2011).

Romera-Torres et al. used HPLC/Exactive–Orbitrap analyzer to detect tropane alkaloids in teas and herbal teas. The limit of quantitation values for the study was between 5 and 20 µg/kg. Highest concentration of atropine was present in herbal teas (27 µg/kg). Other tropane alkaloids present in the tested samples included physoperuvine, pseudotropine, tropine, homatropine and apoatropine (Romera-Torres et al., 2018).

A similar method was used in another to analyze tropane alkaloids in cereals and Solanaceae seeds. Concentrations of up to 694 µg/kg were obtained for scopolamine in samples contaminated with Brugmansia seeds and 1847 µg/kg for atropine in samples contaminated with Stramonium seeds. Millet flour samples also tested positive for scopolamine, tropinone, atropine, anisodamine and littorine (Marín-Sáez et al., 2017).

Baslé et al. (2020) used LC–MS/MS to determine tropane alkaloids in cereals from Asia and Africa. Of the 95 cereal and cereal-based products that were analyzed, only one sample contained low concentrations of the targeted tropane alkaloids.

LC–MS/MS was also used in another study to quantify tropane alkaloids in 113 cereal products for infants and young children from 2011, 2012 and 2014 in the Netherlands. The mean total concentrations of atropine and scopolamine for the three years under consideration were 3.04, 2.37 and 0.37 µg/kg while the maximum total concentrations were 80.8, 57.6 and 3.9 µg/kg (Mulder et al., 2015).

Risk assessments

Ergot alkaloids

Rye and rye products contain high amount of ergot alkaloid. The highest average human chronic exposure to ergot alkaloids was found in “toddlers” and “other children”, with upper bound (UB) estimates of 0.47 and 0.46 µg/kg b.w. per day, respectively. The highest P95 exposure to ergot alkaloids was also observed in “toddlers”. In general, the values derived for the lower bound (LB) estimates were four times lower than the UB estimates, on average (EFSA Journal, 2017). The average acute exposure (MB) was estimated to be in the range of 0.02–0.32 µg/kg bw per day for “infants” and “other children”, respectively while the highest estimated value for the 95^th percentile of acute dietary exposure was recorded for “other children” (EFSA, 2017).

Available data indicate that there are no health concerns associated with this alkaloid, based on the determined BMDL₁₀, group ARfD and group TDI values, which are 0.33, 1 and 0.6 µg/kg b.w. per day, respectively (EFSA CONTAM Panel, 2012).

Glycoalkaloids

In one study, the levels of exposure to PGAs, mainly α-solanine and α-chaconine, were determined using occurrence data for raw primary commodities (RPCs) (EFSA CONTAM Panel, 2020). The values obtained for the mean UB and the P95 occurrence were 51.2 mg/kg and 116.8 mg/kg, respectively, with lowest and highest concentrations of 1.1 and 276.9 mg/kg, respectively. Across the conducted surveys, the mean UB potato glycoalkaloid exposure was between 23.3 µg/kg b.w. per day in “adults” to 174.0 µg/kg bw per day in “toddlers” and the P95 exposure ranged from 78.3 µg/kg bw per day in “adults” to 535.1 µg/kg b.w. per day in “toddlers” (EFSA CONTAM Panel, 2020).

The possible health risks resulting from acute exposure to PGAs through food were determined using MOE approach with a LOAEL of 1 mg/kg PGAs adopted as the reference point for acute risk characterization. As shown by the MOE, possible health risks for younger age groups were identified in both food consumption survey with the highest mean exposure and P95 exposure in all surveys. Meanwhile, only the food consumption survey with the highest P95 exposure identified health risks for adult age groups (EFSA CONTAM Panel, 2020).

Purine alkaloids

Using data from the seventh Korean National Health and Nutrition Examination Survey (KNHANES), one study estimated the daily intake values of the purine alkaloids caffeine, theobromine and theophylline to be 30.819, 1.408 and 0.011 µg/kg bw per day, respectively. For caffeine and theobromine, the obtained values were below the recommended daily intake values; therefore, exposure to these alkaloids in teas poses no health concerns (Baek et al., 2022).

The EFSA Panel on Food Contact Materials, Enzymes, Flavorings and Processing Aids estimated the daily exposure through diet to caffeine and theobromine when added as chemically defined flavoring substances. They concluded that caffeine and theobromine do not present any safety concerns based on the estimated intake levels when they are used as flavoring substances (EFSA CONTAM Panel, 2017).

Pyrrolizidine alkaloids

In one study, data regarding the consumption of plant-based food in European populations were used to determine acute dietary exposure to pyrrolizidine alkaloids. Overall, a range of 34.5–48.4 ng/kg b.w. per day (LB–UB) was derived as the mean chronic dietary estimate in toddlers while a range of 154–214 ng/kg b.w. per day (LB–UB) was derived for highly exposed populations (EFSA CONTAM Panel, 2017).

In a more conservative scenario, a mean exposure of up to 311 ng/kg b.w. per day and a P95 exposure of up to 821 ng/kg b.w. per day were the highest estimates, with tea and herbal infusions being the major contributors to exposure to pyrrolizidine alkaloids (EFSA CONTAM Panel, 2017). For honey, the mean chronic exposure range was 0.1–7.4 ng/kg b.w. per day in adult populations and 0.4–18 ng/kg b.w. per day (minimum LB–maximum UB) for high consumers. The average consumption of honey among the younger populations was estimated to be 0.3–27 ng/kg b.w. per day (minimum LB–maximum UB) while it was estimated to be 0.7–31 ng/kg b.w. per day (minimum LB–maximum UB) among high consumers. A chronic exposure level was observed when food with pollen-based supplements was consumed, with exposure ranges of from 0.7 to 12 ng/kg b.w. per day (minimum LB–maximum UB) and 2.8–44 ng/kg b.w. per day (minimum LB–maximum UB) for high consumers only. Exposure to up to 890 ng/kg b.w. per day of pyrrolizidine alkaloids can be reached from the consumption of a 150 mL infusion containing 2 g of selected plant extracts (for example, borage infusion) (EFSA CONTAM Panel, 2017).

On the basis of occurrence data that were limited to honey, the 2011 EFSA CONTAM Panel concluded that the consumption of large amounts of honey could pose health risks to toddlers and children (EFSA CONTAM Panel, 2017). To assess the risks associated with pyrrolizidine alkaloids as possible carcinogens, the Panel used the new reference point of 237 µg/kg b.w. per day and found that pyrrolizidine alkaloid exposure has possible health implications for humans, especially among people that frequently consume teas and herbal infusions (EFSA CONTAM Panel, 2017).

Quinolizidine alkaloids

The EFSA CONTAM Panel limited their risk assessment of quinolizidine alkaloids in food for human and animal health to only the alkaloids contained in Lupinus species. By identifying the lowest single oral effective dose of 0.16 mg sparteine/kg body weight as a reference point, a margin of exposure approach was used to characterize the risks of acute exposure. Dietary exposure could only be estimated for specific scenarios as there were limited data on the occurrence and consumption of this alkaloid. Therefore, it was impossible to fully estimate the risks of exposure to quinolizidine alkaloid for humans. According to the results from the MOE, examples of scenarios that could pose high risks to human health in terms of exposure to quinolizidine alkaloids included the consumption of lupin seeds that had not been debittered, the consumption of lupin seeds that contained high levels of the alkaloid despite debittering and the consumption of lupin-based imitation meat (EFSA CONTAM Panel, 2019).

Tropane alkaloids

Tropane alkaloids are found in plants that have been used in human medicines for centuries, including those used for the treatment of wounds, gout and sleeplessness and as a pre-anesthesia (EFSA CONTAM Panel, 2013). In one study, a mixture of (-)-hyoscyamine and (-)-scopolamine was administered to human volunteers in their food and a NOAEL of 0.16 µg/kg b.w. was determined, which was used to establish a group ARfD. Further observations revealed that a dose of 0.16 µg/kg b.w., resulted in a transient decreased heart rate, which posed no risk to healthy people but could be harmful to vulnerable individuals (EFSA CONTAM Panel, 2013).

Abbreviations

ARfD

Acute reference dose

BMDL₁₀

Benchmark dose at 10% extra risk

DNA

Deoxyribonucleic acid

EFSA

European food safety association

ELISA

Enzyme-linked immunosorbent assay

FAB-MS

Fast-atom bombardment mass spectrometry

FID

Flame ionization detector

GC

Gas chromatography

GC-EI-MS

Gas chromatography-electron impact-mass spectrometry

GC–MS

Gas chromatography–mass spectrometry

HPLC

High-performance liquid chromatography

HRGC

High-resolution gas chromatography

IARC

International agency for research on cancer

LC–ESI–MS/MS

Liquid chromatography-electrospray ionization-tandem mass spectrometry

LC-HRMS

Liquid chromatography-high-resolution mass spectrometry

LC–MS/MS

Liquid chromatography-tandem mass spectrometry

LC-Orbitrap-MS

Liquid chromatography-orbitrap mass spectrometry

LD

Lethal dose

LLE

Liquid–liquid extraction

LOAEL

Lowest-observed-adverse-effect-level

LOD

Limit of detection

LOQ

Limit of quantitation

MS

Mass spectrometry

MS/MS

Tandem mass spectrometry

NOAEL

No-observed-adverse-effect-level

NTP

National toxicology program

RP-HPLC

Reverse-phase high-performance liquid chromatography

TDI

Tolerable daily intake

TGA

Total glycoalkaloid

UHPLC

Ultra-high-performance liquid chromatography

UPLC-MS/MS

Ultra-performance liquid chromatography-tandem mass spectrometry

UPLC-Q-TOF

Ultra-performance liquid chromatography-quadrupole-time of flight

UV

Ultraviolet

WHO

World Health Organization

Declarations

Conflicts of interest

There are no conflicts of interests.