Cyanobacterial Toxins: Anatoxin-A

Home , Cyanotoxin, Cylindrospermopsin, Geosmin, Gonyautoxin, Microcystin, Neurotoxicity

WHO/SDE/WSH/xxxx English only

Cyanobacterial toxins: Anatoxin-a

Background document for development of WHO Guidelines for Drinking-water Quality and Guidelines for Safe Recreational Water Environments

Version for Public Review Nov 2019

Preface

Information on cyanobacterial toxins, including anatoxins, is comprehensively reviewed in a recent volume to be published by the World Health Organization, “Toxic Cyanobacteria in Water” (TCiW; Chorus & Welker, in press). This covers chemical properties of the toxins and information on the cyanobacteria producing them as well as guidance on assessing the risks of their occurrence, monitoring and management. In contrast, this background document focuses on reviewing the toxicological information available for guideline derivation and the considerations for deriving the guideline values for microcystins in water. Sections 1-3 and 8 are largely summaries of respective chapters in TCiW and references to original studies can be found therein.

To be written by WHO Secretariat

Acknowledgements

To be written by WHO Secretariat

Abbreviations used in text

AF drinking-water allocation factor ALP alkaline phosphatase ALT alanine aminotransferase ASP aspartate aminotransferase ATX anatoxin-a bw body weight C daily drinking-water consumption CYN cylindrospermopsin dw dry weight DWI daily water intake ELISA enzyme-linked immunosorbent assay FSH follicle stimulating hormone GD gestational day GGT gamma-glutamyl transferase GSH glutathione GST-P glutathione S-transferase placental form-positive GTX gonyautoxin GV guideline value HBV hepatitis B virus HPLC high performance liquid chromatography HTX homoanatoxin i.p. intraperitoneal i.v. intravenous LC-MS liquid chromatography – mass spectrometry LDH lactate dehydrogenase LH luteinizing hormone LOAEL lowest-observed-adverse-affect level MC microcystin(s) NOAEL no-observed-adverse-affect level P fraction of the TDI allocated to drinking-water PoD point of departure PP1 protein phosphatase-1 PP2A protein phosphatase-2A PSP Paralytic Shellfish Poisoning PST paralytic shellfish toxin STX saxitoxin STXOL saxitoxinol STXs saxitoxins, comprising all analogues TDI tolerable daily intake UF uncertainty factor

Table of Contents 1.0 EXECUTIVE SUMMARY ...... 7 2.0 GENERAL DESCRIPTION ...... 7 2.1 Identity ...... 7 2.2 Physical and Chemical Properties ...... 7 2.3 Organoleptic Properties ...... 7 2.4 Major Uses and Sources...... 8 2 ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE ...... 8 2.1 Air ...... 8 2.2 Food ...... 8 2.3 Water ...... 9 2.4 Estimated total exposure and relative contribution of drinking-water ...... 9 3 KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS ...... 9 Absorption ...... 9 Distribution ...... 10 Metabolism ...... 10 Elimination ...... 10 4 EFFECTS ON HUMANS ...... 10 5 EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO SYSTEMS ...... 10 Acute exposure ...... 10 Short-term exposure ...... 11 Genotoxicity and carcinogenicity ...... 12 In vitro systems ...... 12 Mode of action ...... 12 6 OVERALL DATABASE AND QUALITY OF EVIDENCE ...... 12 6.1 Summary of Health Effects ...... 12 6.2 Quality of Evidence ...... 12 7 PRACTICAL CONSIDERATIONS ...... 13 7.1 Source control ...... 13 7.2 Monitoring ...... 14 7.3 Analytical methods and achievability ...... 14 7.4 Treatment methods and performance ...... 15 8 CONCLUSIONS ...... 15 8.1 Derivation of the guideline-value ...... 15 8.2 Considerations in applying the health-based reference values ...... 16 REFERENCES (still need some tidying up) ...... 17

1.0 EXECUTIVE SUMMARY

[to be completed]

2.0 GENERAL DESCRIPTION

2.1 Identity

Anatoxin-a (ATX; CAS Reference Number 64285-06-9), or 2-acetyl-9-azabicyclo[4:2:1]non- 2-ene, is a tropane-related bicyclic, secondary amine alkaloid. Figure 2.1 shows the presence of an additional methyl group (CH) on carbon atom 11 (C11) which differentiates homoanatoxin- a (HTX; CAS Reference Number 142926-86-1; Fig. 2.1) from its analog ATX. Both molecules share almost identical toxicological properties. Other derivatives of ATX have been identified in cyanobacterial cultures or in field samples, including 2,3-epoxy-anatoxin-a, 4-hydroxy- and 4-oxo-derivatives, dihydroanatoxin-a and dihydrohomoanatoxin-a, that possibly represent degradation products (TCiW; (Testai, in press).

H H a N b N

O O Figure 2.1: Structures of anatoxin-a (a) and homoanatoxin-a (b)

2.2 Physical and Chemical Properties

ATX has a molecular formula of C10H15NO and an average molecular weight of 165.232 Da (monoisotopic MW 165.115 Da). ATX is highly soluble in water with a computed Kow of 0.8 and has a high boiling point of 291°C. It has a density of 1.04 and a low vapor pressure of 0.002. Other physico-chemical properties such as the soil adsorption coefficient (Koc), how it volatizes from water, and its distribution in the atmosphere (Henry’s Law constant) are unknown. Limited information on the chemical breakdown, biodegradation and distribution in the environment is available (TCiW; (Testai, in press).

Table 2.1: Physical and chemical properties of anatoxins. N/A: not applicable.

Property anatoxin-a homoanatoxin CASRN1 64285-06-9 142926-86-1 Chemical Formula C10H15NO C11H17NO Average MW2 (g/mole) 165.237 179.264 Monoisotopic MW (g/mole) 165.115 179.131 Color/Physical State Boiling Point 291°C N/A 3 Kow 0.8 1.3 Solubility in Water High High Solubility in Other Solvents 1 Chemical Abstracts Service Registry Number 2 Molecular Weight 3logP computation with XLogP3 (Cheng et al., 2007)

2.3 Organoleptic Properties

While none of the known cyanobacterial toxins have been shown to affect the taste or odour of water, some cyanobacterial species produce other compounds such as geosmin and methyl- isoborneol that do cause taste and odour of water, thus indicating the presence of cyanobacteria in raw water. However, as this applies only to some species, the absence of these typical tastes

and odours are not a reliable indicator for the absence of cyanotoxins. Taste or odour thresholds in water are 0.004 ppb for geosmin and 0.006 ppb for methyl-isoborneol (TCiW, Kaloudis, in press).

2.4 Major Uses and Sources

ATX occurs naturally (although high concentrations are typical for waterbodies influenced by human activity, i.e. effluents from wastewater or run-off from agricultural land) and there are no known commercial applications. ATX is produced by a variety of cyanobacteria species belonging to Nostocales (Anabaena flos-aquae, A. lemmermannii, Chrysosporum (Aphanizomenon) ovalisporum, Cuspidothrix sp., Cylindrospermopsis sp., Raphidiopsis mediterranea, Cylindrospermum, Dolichospermum), Oscillatoriales (Oscillatoria sp., Planktothrix sp., Phormidium sp., Tychonema sp.), and possibly Chroococcales (Microcystis sp., Woronichinia sp.). Producing and non-producing strains are known for all species for which ATX-production has been observed.

The production of ATX is both species- and strain-specific. ATX has been found to be co- produced with HTX in varying shares. A few cyanobacterial strains have been reported to produce both microcystins and ATX. The ATX content (or cell quota) of individual strains generally varies depending on different growth conditions and environmental factors in a range of 2-4 fold, maximally 7 fold. ATX contents are not consistently related to cell growth phases.

The few data available on ATX/HTX cell quota range from 90 fg/cell in Aphanizomenon sp. to 500 fg/cell in Phormidium sp. Highest toxin contents of maximally 13 mg/g dry weight were reported from Anabaena sp. and Oscillatoria sp. while from other genera contents were 1-2 orders of magnitude lower.

The biosynthesis of the ATXs involves a polyketide synthases (PKS). Complete gene cluster sequences (anaA-G, ca. 25 kbp) are available for strains from several genera (Oscillatoria, Cylindrospermum, Cuspodothrix) and the individual steps of the biosynthesis have been studied. The information on the molecular regulation of biosynthesis is scarce and does not allow generalizations.

For more details on ATX producing organisms and biosynthesis see TCiW (Testai, in press).

2 ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE

2.1 Air

ATX (HTX) is not volatile and hence exposure via inhalation would require their dissolution in aerosols. It could also occur through cyanobacterial cells carried in spray, e.g. during storms or in the wake of a power boat. No information on exposure via this route was found nor were data on concentrations in aerosols.

2.2 Food

ATX has been detected in contents of <50 µg/g in fish harvested from the environment or experimentally exposed to high concentrations of ATXs. Lower contents were observed in mussels, but the low number of studies does not allow any generalization. Analyses of ATXs in dietary supplements also yielded inconsistent results. No evidence for bioaccumulation of ATX is available.

For more details on ATX in food and dietary supplements see TCiW (Dietrich, in press; Ibelings & Chorus, in press).

2.3 Water

In many settings the major water-borne route of human exposure to ATX will be the consumption of drinking-water, most likely where it is produced from surface waters with insufficiently effective or non-existent treatment. A further exposure route – important in some settings – is the recreational use of lakes and rivers. Depending on the seasonal patterns of cyanobacterial blooms and water body use, patterns of exposure may be episodic. Although cyanobacteria potentially producing ATXs occur widely in diverse freshwater environments around the globe, ATXs are generally detected less frequently than microcystins and cylindrospermopsin.

Concentrations rarely exceed some tens of µg/L in open water but in surface blooms they have been reported to exceed 1000 µg/L. Only a few reports on ATX in drinking-water are available, with concentrations generally ranging in the low µg/L range. Among anatoxin-producing species, several do not form water blooms but occur as benthic mats in rivers (i.e., on the sediment surface) or periphyton (attached to higher aquatic plants), resulting in a highly patchy distribution of ATXs in the environment. Reported intoxications of pet dogs or livestock occurred after ATX uptake with cyanobacterial lumps detached from these surfaces and containing high amounts of ATX. Only limited and inconclusive information is available on the release of ATXs from cells and its persistence in surface waters. Available data, however, indicate that ATXs are largely confined to viable cells and released primarily through cell lysis, followed by rapid degradation.

For more details on anatoxin-a and homoanatoxin occurrence in the environment and drinking- water see TCiW (Testai, in press).

2.4 Estimated total exposure and relative contribution of drinking-water

As for the other cyanotoxins, drinking-water is the most likely source of exposure to ATX where surface water sources are used. However, this assumption serves as a starting point and country- or region-specific circumstances should take into account the contribution of foods and of recreational activities in lakes with cyanobacterial blooms. While human exposure through food harvested from aquatic environments has not been evidenced, the above-mentioned contents found in fish show that it to be possible. Significant dermal or inhalational uptake of ATX during recreational exposure is unlikely, leaving the oral route as the main route of concern.

Exposure patterns and durations are strongly influenced by region and lifestyle. Estimating total exposure or the relative contribution of particular exposure routes (e.g., food, drinking-water) requires specific analyses of concentrations in respective media in a given setting.

A specific aspect of exposure to ATX is the potential for high concentrations in beached or floating benthic mat material or macrophytes to which ATX-producing cyanobacteria are associated. While high ATX concentrations appear to be limited to the direct vicinity of such material, animal fatalities due to its ingestion give rise to concern regarding direct contact and possible unintentional ingestion. Contact with such material should therefore be avoided.

3 KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS

Absorption

Acute toxicity studies in animals indicate that ATX is rapidly absorbed from the gut following an oral exposure because signs of neurotoxicity, including loss of coordination, muscular

twitching and death from respiratory paralysis, occur within minutes of oral exposure (Stevens & Krieger, 1991; Fitzgeorge et al., 1994).

Distribution

As with absorption, the rapid appearance of symptoms following oral exposure suggests rapid serum distribution to affected tissues. Studies designed to examine tissue distribution were not identified.

Metabolism

No published studies were identified that have investigated in vivo mammalian metabolism of ATX.

Elimination

It appears that at least some ATX is excreted unchanged in urine and bile since it was detected at high concentrations in these fluids from a dog poisoned by toxic Phormidium (Puschner et al., 2010). However, the analytical method used (ion trap mass spectrometry) was not capable of detecting metabolites (Puschner et al., 2008).

4 EFFECTS ON HUMANS

The only reported case of suspected human poisoning by ATX occurred in a 17-year-old boy who died 2 days after swallowing water while swimming in a golf course pond containing an Anabaena flos-aquae bloom. A peak in the chromatograms with the retention time and mass of ATX from samples of liver, blood and fluids collected post-mortem was, however, later identified as phenylalanine (Carmichael, in prep.). In a review of 11 disease outbreaks reported to be associated with cyanobacterial blooms in the US in 2009-2010, ATX was detected as a water quality indicator in 3 outbreaks (concentration range 0.05-15 µg/L). However, in all 3 cases, microcystins were also detected at often substantially higher concentrations (0.3->2,000 µg/L), and in one case cylindrospermopsin and saxitoxin were also present (Hillborn et al., 2014). Nevertheless, it is noteworthy that neurological symptoms were reported in all 3 ATX associated outbreaks but in none of the other outbreaks.

5 EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO SYSTEMS

Acute exposure

Deaths of dogs, livestock and waterfowl due to ATX poisoning have been reported many times throughout the world, including Scotland, Ireland, New Zealand, France, the Netherlands and the US (Testai et al., 2016; Health Canada, 2017). In many cases, the producing cyanobacteria were benthic species, detached and washed ashore, with dogs consuming lumps of material. Also, planktonic species can stick to an animal’s fur, which are then licked off upon grooming. These circumstances may increase the chance that a lethal dose is ingested and that the incident is reported. Although ATX has been detected in the stomach contents and other tissues of poisoned dogs at necropsy, no estimates of a lethal dose have been made from these case studies.

The acute oral LD50 of a synthetic (+)-anatoxin-a preparation was 13.3 mg/kg bw (95% CI: 12.8-14.1) and the acute intraperitoneal LD50 was 0.21 mg/kg bw (95% CI: 0.20–0.24) (Stevens & Krieger, 1991). However, when ATX was derived from cyanobacterial extracts some preparations were 2-3 times more potent via the oral route than expected from their ATX content or intraperitoneal potency (Stevens & Krieger, 1991; Fitzgeorge et al., 1994). The reason for this discrepancy is not known, but it suggests that ATX content of cyanobacterial extracts may not be an accurate predicter of oral potency. Exposure to a lethal dose via either route causes death by respiratory paralysis within a few minutes.

The acute i.p. toxicity of HTX was reported to be similar to that of ATX (i.p. LD50 = 0.25 mg/kg bw in mice), producing the same symptoms and death within 7–12 min (Skulberg et al., 1992). When given orally by gavage, HTX was 10 times less toxic than via the i.p. route (Lilleheil et al., 1997). The ATX degradation product, dihydroanatoxin-a, is about one tenth as toxic as ATX (Mann et al., 2012).

Short-term exposure

Astrachan et al. (1980) conducted a 54-day drinking-water study in female Sprague-Dawley rats. ATX was partially purified from cultures of Anabaena flos-aquae strain NRC-44-1, with the final concentration determined by molar absorptivity. Groups of 20 rats were exposed to 0, 0.51 or 5.1 ppm ATX in drinking-water and the authors estimated the daily exposures to be 0.051 and 0.51 mg/kg, respectively. Food consumption and body weights were monitored and haematology, serum enzymes, histopathology and liver mixed function oxidase activities analysed at the end of the study. No adverse effects were seen in any animal.

Fawell et al. (1999) conducted a 5-day repeated dosing trial in Crl:CD-1(ICR)BR (VAF plus) mice to determine a maximum tolerated dose for a 28-day study. Doses of (+)-anatoxin-a HCl used in the 5-day trial were 1.5, 3.0, 7.5 or 15 mg/kg. All animals in the top dose group plus one in the 7.5 mg/kg dose group died within 5 min of dosing, so 3.0 mg/kg was chosen as the top dose for the 28-day study. The 28-day study used 4 dose groups of ten mice of each sex dosed daily by gavage with 0, 0.12, 0.6, or 3.0 mg/kg (+)-anatoxin-a HCl (equivalent to 0, 0.098, 0.49 or 2.46 mg/kg pure (+)-anatoxin-a). Bodyweight, food consumption and signs of illness were monitored in all mice through the trial, and detailed histopathology, haematology and serum biochemistry analyses were conducted for control and high dose animals at the end of the study. One mouse in each of the top two dose groups died within 2.5h of dosing and necropsy did not show any signs of a cause, meaning that ATX toxicity could not be excluded. No other treatment-related effects were seen in any animal for any parameter examined. The authors therefore designated 0.098 mg/kg pure (+)-anatoxin-a as the no-observed-adverse- effect level (NOAEL), but noted that the NOAEL could actually be 2.46 mg/kg.

No long-term studies on the systemic effects of ATX were identified.

5.1.1 Neurological effects

ATX is a neurotoxin hence the primary effects caused are on the nerves and the muscles they control. Two daily intraperitoneal injections of 0.21 mg ATX/kg or higher in mice caused decreased motor activity, altered gait, difficulty breathing and convulsions 5–6 minutes post- exposure, with death occurring within 10 minutes. Less severe clinical signs were observed in animals that survived lower doses, with recovery after 15–20 minutes (Rogers et al., 2005). Locomotor activity was reduced in rats by a single subcutaneous injection of 0.06 to 0.225 mg ATX/kg bw, and tolerance did not develop after a regime of 4 x weekly injections (Stolerman et al., 1992; MacPhail et al., 2007). In contrast, tolerance was seen in behavioural responses in trained rats given 4 x weekly subcutaneous injections of 0.05-0.1 mg/kg toxin, but not at 0.2 mg/kg (Jarema et al., 2008).

5.1.2 Reproductive and developmental effects

Non-lethal exposures of pregnant mice did not cause any adverse effects on their off-spring (Fawell et al., 1999; Rogers et al., 2005). However, 7 daily intraperitoneal injections of ATX (50, 100 or 150 μg/kg bw per day) to male mice caused significant reductions in sperm count as well as a range of other adverse effects in the testes (Yavasoglu et al., 2008). More research is needed into the developmental and reproductive effects of orally administered ATX.

5.1.3 Immunological effects

The mammalian immunotoxicity of ATX does not appear to have been studied. 11

Genotoxicity and carcinogenicity

(+)-Anatoxin-a fumarate (0.312, 0.625, 1.25, 2.5, 5 or 10) was not mutagenic in any of 6 strains of S. typhimurium tested (Sieroslawska, 2013). No genotoxicity studies have been conducted using mammalian cell lines and no in vivo carcinogenicity studies on ATX were identified.

In vitro systems

(+)-Anatoxin-a HCl (4 µg/ml) caused significant reductions in viability (LDH leakage, MTT reduction) and increases in apoptosis and DNA fragmentation in cultured Vero cells and primary rat thymocytes. Levels of reactive oxygen species increased in treated thymocytes (Rao et al., 2002).

Mode of action

Anatoxin-a binds with high affinity to the nicotinic acetylcholine receptors of motor neurons, stimulating muscle cell contraction. Unlike the natural neurotransmitter acetylcholine, ATX is not degraded by acetylcholine esterase, so the ATX stimulatory signal to the muscle cells is not switched off as it would be with acetylcholine. The muscle cells become fatigued and eventually paralysed. When this happens in the muscles involved in respiration, the animal dies. ATX also affects other nicotinic cholinergic receptors in the cardiovascular system to increase blood pressure and heart rate, and in the brain (Health Canada, 2017).

Homoanatoxin-a exhibits very similar toxicological properties and mode of action to (+)- anatoxin-a (Lilleheil et al., 1997).

6 OVERALL DATABASE AND QUALITY OF EVIDENCE

6.1 Summary of Health Effects

Anatoxin-a has high acute oral toxicity based on an LD50 of 13 mg/kg in mice (Stevens & Krieger, 1991). ATX chronically stimulates nicotinic cholinergic receptors in peripheral nerves leading to muscular twitching, fatigue and paralysis. Severe overstimulation of respiratory muscles results in respiratory arrest and death within minutes. No chronic dosing studies have been conducted. In a 28-day study, a NOAEL of 0.098 mg/kg bw per day for (+)-anatoxin-a was identified due to the unexplained deaths of single mice in each of the two highest dose groups (n=20 per group). However, the acute LD50 is 13 mg/kg, and if these deaths are ignored the NOAEL would be 2.46 mg/kg bw per day indicating that there is high uncertainty regarding the NOAEL of 0.098 mg/kg bw per day (Fawell et al., 1999). However, lethality is not generally considered an appropriate end-point for deriving a reference value. In a 54-day drinking-water study in rats, the highest dose of partially purified ATX of 0.51 mg/kg bw per day did not cause any adverse effects (Astrachan et al., 1980). In the neurodevelopmental study by Rogers et al. (2005), the NOAEL for maternal toxicity was 0.125 mg/kg bw per day for reduced motor activity seen at 0.2 mg/kg bw per day, but this value is affected by the route of administration (i.p., known to give rise to higher toxicity than the oral route) and the use of a racemic mixture (no relative content of the two stereoisomers was given and purity was only 90%).

6.2 Quality of Evidence

The overall database quality for ATX is considered to be very low. There are a large number of deficiencies in the toxicological information available including: Lack of oral repeat dosing studies into sub-lethal effects including locomotor, behavioural, neurodevelopmental and reproductive outcomes; Lack of kinetic studies including metabolism; Lack of chronic exposure studies; Lack of almost any toxicological information on HTX, or on natural mixtures of ATX with its degradation products. 12

The reliability of the studies that are available is also considered to be low. Very few studies used well-characterized (usually synthetic) (+)-anatoxin-a. Many have used partially purified preparations, some of which have shown discrepancies between the expected potency, based on known ATX content, and the actual potency in the animals. That this appears to apply particularly to oral dosing is especially worrying. None of the sub-chronic dosing studies appear to have checked the stability of ATX in their dosing solutions despite its known instability in light. Very few studies used the oral route of exposure. Most animal studies available for risk assessment analysed endpoints such as histopathology, haematology and serum chemistry, which may not be strongly affected by a neurotoxin.

The only conclusion about the toxicity of ATX that is able to be stated with any confidence is that it is a potent and fast-acting neurotoxin that chronically activates nicotinic cholinergic receptors, leading to an over-stimulation of muscles resulting in muscle fatigue and paralysis. When this occurs in the respiratory muscles, death ensues.

7 PRACTICAL CONSIDERATIONS

ATX occurs less frequently in lakes and reservoirs than other cyanobacterial toxins like microcystins and cylindrospermopsin, and persistent, planktonic blooms of anatoxin-producing cyanobacteria with health-relevant concentrations of ATX do not appear to be common. However, benthic cyanobacteria producing ATXs have been reported in several countries. The often patchy occurrence of benthic or periphytic species hampers systematic monitoring of entire water bodies. Lethal animal intoxications (of wild as well as of domestic animals) have been attributed to ATX, including from benthic species. These can cause considerable public health concern, requiring investigation, risk assessment and possibly protective action.

Chapters 7 – 10 of TCiW give guidance on multiple barriers against cyanotoxins in water including controlling nutrient loads from the catchment, managing water bodies, optimizing sites for drinking-water offtakes or recreation, applying drinking-water treatment to remove cyanobacteria and cyanotoxins and providing information or warnings for recreational use of water bodies with blooms. This includes guidance on planning, managing and documenting the barriers used to mitigate cyanotoxin risks through developing a water safety plan (TCiW; Chorus & McKeown, in press; Bartram et al., 2009).

7.1 Source control

For planktonic toxic cyanobacteria the prevention of blooms in source waters is the key to long- term control of the risks they represent. The most sustainable approach to achieve this is to keep concentrations of plant nutrients low. Most cyanobacteria typically proliferate under eutrophic conditions i.e., at elevated concentrations of nutrients, in particular of phosphorus, and total phosphorus concentrations below 20-50 µg/L will limit the development of cyanobacterial blooms in most situations (TCiW; Chorus & McKeown, in press; Zessner & Chorus, in press). A number of measures within water bodies can mitigate cyanotoxin occurrence, including e.g. artificial water column mixing, nutrient reduction through sediment removal or treatment, or biomanipulation. Their success is highly dependent on the specific conditions in the water body, as discussed in TCiW (Burch et al., in press).

Many reservoir off-take structures (towers) can take water from multiple depths to account for vertical heterogeneity. Variable offtakes enable avoiding water layers containing the highest concentrations of cyanobacteria. If multiple off-takes are not available (e.g. in small systems) it may be possible, as a temporary measure, to siphon water from a specific depth. Where conditions allow, the use of bank filtration between source waters and treatment plant inlets can be very effective both for removing cyanobacteria and for biodegradation of dissolved MC (TCiW; Brookes et al., in press). Where possible, sites for recreational activities are best located upwind of bays where scums tend to accumulate.

7.2 Monitoring

While cyanobacteria can be present in surface waters at low numbers throughout the year the occurrence of blooms producing significant concentrations of ATX tend to be short-lived and often seasonal events. Monitoring of source waters should include surveillance for factors that can support the growth of cyanobacteria including total phosphorus, temperature, water residence time, pH and Secchi disc transparency (for detail see TCiW; Padisák et al., in press). On site visual assessment of turbidity with greenish discolouration or scums and microscopy are effective low cost direct methods that can trigger increased vigilance if CYN-producing cyanobacteria are observed. In many cases monitoring over several seasons can establish the likely occurrence and timing of favourable conditions for cyanobacterial growth as well as the taxonomic composition and magnitude of blooms. For example, a lake with regular seasonal blooms of Aphanizomenon in late summer is unlikely to shift to perennial blooms of Cylindrospermopsis from one year to the next. (TCiW; Ibelings et al., in press).

Monitoring programmes should be adaptive with sampling and testing being increased when there is evidence of increasing cell numbers. Alert Level Frameworks (ALF) have been described both for drinking- water and for recreational water use. These include various criteria to trigger particular analyses and risk mitigation measures (TCiW; Humpage et al., in press; Chorus & Testai, in press). As described in the ALFs monitoring of source waters can start with simple site inspections for appearance of visible blooms, assessing transparency using a Secchi disc. However, not all ATX producers form surface scums or strong discoloration, and these may be overlooked. Therefore, if the presence of cyanobacteria is suspected, microscopic examination for the presence of potentially ATX producing cyanobacteria is important. As blooms develop monitoring can be expanded to include quantitative measures of cyanobacterial biomass indicating potential toxin concentrations such as cyanobacterial biovolumes or chlorophyll-a, or direct analyses of ATX concentrations. While the detection of potentially ATX producing cyanobacteria indicates possible ATX occurrence, this will not include ATX dissolved in water. Therefore, and also because concentrations associated with cyanobacterial blooms can vary substantially, where possible toxin analyses should be performed if ATX is suspected. Toxin data may well allow avoiding or lifting restrictions of site use where these were based on biovolume or chlorophyll-a concentrations.

ATXs can be associated with benthic mats which occur patchily and ephemerally. If benthic mats are considered an issue in individual water bodies testing for ATX will have a greater priority. In a number of reported cases, analysis of ATX followed animal intoxication incidences.

7.3 Analytical methods and achievability Analytical techniques are available for the range of parameters associated with cyanobacterial blooms and associated ATX. The complexity, expertise requirements and costs of monitoring increase from relatively simple visual inspections to testing for phosphorus, pH, Secchi disc transparency, cell numbers, species identification, biovolumes and chlorophyll-a determination. Testing for ATX using liquid-chromatography-mass spectrometry (LC/MS) or high-performance liquid chromatography mass spectrometry (HPLC) is the most complex and time consuming.

For cell-bound ATX, an extraction step is performed prior to analysis. For routine analysis, HPLC and LC-MS are the best methods to assess ATXs. USEPA Method 545 based on Liquid Chromatography Electrospray Ionization Tandem Mass Spectrometry (LC/ESI-MS/MS) (US EPA, 2015) has a lowest minimum reporting concentration of 0.018 µg/L (18 ng/L). Certified reference material for ATX is commercially available. ELISA kits are available which detect both anatoxin-a and homoanatoxin-a with a reported quantification range of 0.15 – 5 µg/mL (for more detail see TCiW; Metcalf et al., in press).

While these methods were developed for the analysis of water samples, applying them to more

complex matrices such as food or stomach contents requires prior clean-up.

7.4 Treatment methods and performance

Treatment processes to reduce ATX in drinking-water are based on two approaches: reducing the cell-bound toxins by physical removal of the cells and reducing dissolved ATX (TCiW; Newcombe et al., in press). Reports of the cell-bound proportion of ATX varies but in healthy blooms the majority of ATX is likely to be cell-bound and therefore effectively removable by physical processes, i.e. coagulation followed by flocculation, clarification and rapid media filtration as well as by slow sand filtration or membrane filtration. Care needs to be taken to avoid or minimise pre-filtration treatment as this causes cell lysis and release of ATX. Further, as cells may lyse in more acidic water the pH should be kept above 6. Care also needs to be taken to ensure that cyanobacterial and ATX concentrates (e.g. filter backwash, sludges and sludge supernatants) are not allowed to return to the head of the filtration plant during a bloom. Dissolved ATX can be removed by adsorption onto powdered or granular activated carbon (PAC or GAC). Efficacy of removal can be influenced by the type of activated carbon, doses and points of application (PAC), contact times (PAC), flow rates (GAC) and water quality. Biological degradation of ATX during slow sand filtration and on GAC filters can be very effective, although it may require a lag phase for the degrading bacteria to establish.

ATX are more resistant to oxidation than the other cyanotoxins. Ozone has been shown to be effective against ATX but chlorine is not reliably effective. Other oxidants such as chloramine and chlorine dioxide are ineffective against ATX at doses and contact times normally used in drinking-water treatment. The treatment methods discussed above are able to reduce ATX concentrations well below the Health-Based Reference Value of 24 μg/L given below. However, validation of efficacy under specific local conditions is important, and this applies in particular to slow sand filtration and oxidation: their efficacy is highly dependent on the specific water quality and further conditions in the treatment system. Validation may include field trials and laboratory investigations such as jar testing. Verification of removal during blooms should be undertaken by monitoring ATX in finished drinking-water. After effective treatment it is important to ensure drinking-water remains safe and free of cyanobacterial regrowth. This can be accomplished by ensuring that any channels and storages are covered and dark, so that cyanobacteria lack light necessary for growth. Maintaining chlorine residuals throughout the distribution system will also suppress cyanobacterial regrowth.

8 CONCLUSIONS

8.1 Derivation of the guideline-value

Acute exposure in animals led to deaths within minutes of gavage administration (Astrachan et al., 1980; Fawell et al., 1999). Since neither of the available repeated toxicity studies identified a non-lethal dose that caused lasting adverse effects, a formal guideline value (provisional or otherwise) for lifetime exposure cannot be derived based on the available information. In the 28-day study of Fawell et al. (1999), one of 20 animals in each of two dose groups died without signs that could attribute the cause to non-treatment effects. If it is conservatively assumed these animals died due to the effects of the toxin, the NOAEL would be 98 μg/kg, but it could otherwise be as high as 2.4 mg/kg if these two animals were excluded (Fawell et al., 1999). A highly conservative assumption has been used in defining the NOAEL for the short-term drinking-water health-based reference value below (the oral NOAEL of 98 µg/kg is lower than 15

the i.p. NOAEL for maternal toxicity identified by Roger et al., 2005): therefore, an uncertainty factor for database deficiencies was not applied. These derivations for ATX are based on repeated dosing and are relevant for short-term or acute exposure.

There is insufficient information to develop a long-term health-based reference value for anatoxin-a.

Calculation of short-term drinking-water health-based reference value for anatoxin-a:

NOAEL ∗ bw ∗ AF 98 ∗ 60 ∗ 1.0 HBRV = = µg L−1 = 29.4 µg L−1 ≈ ퟑퟎ 훍퐠/퐋 short−term UF ∗ C 2 ∗ 100

HBRVshort-term: short-term drinking-water “health-based reference value”; bw: adult body weight (WHO standard is 60 kg for an adult); C: daily water consumption (assumed to be 2 L for an adult); NOAEL: no-observed-adverse-effect level (=98 μg/kg bw per day, based on Fawell et al., 1999); AF: proportion of exposure assumed to come from drinking-water, 1.0 ; UF: uncertainty factor = 10 for interspecies and 10 for intraspecies variation

Calculation of recreational water health-based reference value for anatoxin-a:

NOAEL ∗ bw 98 ∗ 15 HBRV = = µg L−1 = 58.8 μg 퐿−1 ≈ ퟓퟗ 훍퐠 퐋−ퟏ rec UF ∗ C 100 ∗ 0.25

HBRVrec: recreational-water health-based reference value; bw: body weight (assumed to be 15 kg for a child); C: incidental water ingestion assumed to be 250 mL during primary contact (energetic play, swimming, falling out of a boat, etc.); NOAEL: 98 μg/kg per day based on Fawell et al., 1999; UF: uncertainty factor (10 for intraspecies × 10 for interspecies).

8.2 Considerations in applying the health-based reference values

The derivation of the health-based reference values for anatoxin-a follows a highly conservative, precautionary approach. Nonetheless, the considerations for applying guideline values for other cyanotoxins (see WHO, in prep.) are equally relevant to these values. One is that informing the public about cyanobacterial blooms in source waters is important, as cyanobacterial blooms tend to impair the taste and odour of drinking-water. Unless informed that this water is nonetheless safe to use, the public may turn to other, less safe sources of water.

For recreational sites with blooms, information and warnings are particularly important. The most common situation is that monitoring cannot occur at sufficiently short time intervals (i.e. daily rather than weekly) to ensure that it captures situations with heavy scums. Site users therefore need information about avoiding scum contact and ingestion as well as situations with pronounced greenish turbidity, i.e. to the extent that one cannot see one’s feet when knee-deep into the water. Temporary closure of sites is an option if blooms contain high toxin concentrations, exceeding the recreational guideline values (for further detail see TCiW, d’Anglada et al.). It may also be important to provide information about the possibility that detached aquatic plant-like material, either floating or accumulated in lumps on the beach, can contain high ATX concentrations.

The health-based reference values provided above are not intended as WHO guideline values due to inadequate data. Nevertheless, it is recognised that some orientation as to a “bounding value” may be useful to risk assessors. Based on the limited currently available studies of acute and sub-chronic anatoxin-a toxicity, the values provided are unlikely to cause adverse effects

in exposed adults. Since infants and children can ingest a significantly larger volume of water per body weight (e.g. up to 5 times more drinking-water/kg bw for bottle-fed infants compared to an adult), it is recommended, as a precautionary measure, that alternative water sources such as bottled water are provided for bottle-fed infants and small children when ATX concentrations are greater than 6 µg/L for short periods.

The drinking water health-based reference value is based on a 28-day repeated dose study and so is applicable for short-term exposure. However, ATX is acutely toxic and so it is recommended that any exposure above this value be avoided.

The health-based reference values are based on toxicological data for ATX. It is recommended that these values be applied to total HTX as gravimetric or molar equivalents based on a reasonable assumption that HTX has similar toxicity to ATX.

The health-based reference value applied an allocation factor of 100% because drinking-water is usually the most likely long-term source of exposure. However, there is very limited data to suggest that fish and shellfish may become contaminated with ATX. Therefore in some regions, food could be a significant source of exposure, particularly in tropical locations where the duration of blooms is long and there is high consumption of local aquatic food. In such situations, consideration should be given to either reducing the allocation factor based on relative consumption data from the exposed population or analysing ATX concentrations in the local aquatic food.

REFERENCES (still need some tidying up) Astrachan NB, Archer BG, Hilbelink DR (1980). Evaluation of the subacute toxicity and teratogenicity of anatoxin-a. Toxicon, 18(5–6): 684–688. Bartram J, Correales L, Davison A, Deere D, Drury D, Gordon B et al. (2009). Water safety plan manual: step-by-step risk management for drinking-water suppliers. World Health Organization, Geneva, CH. 103 pp. ISBN 978 92 4 15626 8. Brookes JD, Burch MD, Grützmacher G, Klitzke S (in press). Managing cyanotoxin risks at the drinking water offtake. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Burch MD, Brookes JD, Chorus I (in press). Assessing and controlling the risk of cyanobacterial blooms: Water-body conditions. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Carmichael WW, Yuan M, Friday CF. (2004). Human mortality from accidental ingestion of toxic cyanobacteria – a case re-examined [abstract]. Sixth International Conference on Toxic Cyanobacteria, Bergen pp 61–62. Cheng T, Zhao Y, Li X, Lin F, Xu Y, Zhang X et al. (2007). Computation of octanol− water partition coefficients by guiding an additive model with knowledge. J Chem Inform Model. 47:2140-8. Chorus I, McKeown RM (in press). Introduction to understanding and managing water use systems. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Chorus I, Testai E (in press). Understanding Exposure: Recreation and occupational activities. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Dietrich DR (in press). Understanding exposure: Cyanobacteria as dietary supplements. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health

Organisation, Geneva, CH. Fawell JK, Mitchell RE, Hill RE, Everett DJ. (1999). The toxicity of cyanobacterial toxins in the mouse: II. Anatoxin-a. Hum. Exp. Toxicol 18(3): 168–173. Fitzgeorge NLM, Clark SA, Keevil CW.. (1994). Routes of intoxication. In: G.A. Codd et al. (eds.), Detection methods for cyanobacterial (blue-green algal) toxins. First International Symposium on Detection Methods for Cyanobacterial (Blue-Green Algal) Toxins. Royal Society of Chemistry, Cambridge, UK. pp. 69–74. [cited in Chorus and Bartram, 1999]. Funari E, Testai E. (2008) Human health risk assessment related to cyanotoxins exposure. Crit Rev Toxicol 38: 97-125. Health Canada (2017) Humpage AR, Cunliffe DA, Chorus I (in press). Understanding Exposure: Drinking Water. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Ibelings BW, Chorus I (in press). Understanding exposure: Food. In: Chorus I, Welker M, editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Jarema KA, Poling A, MacPhail RC. (2008). Effects of weekly exposure to anatoxin-a and nicotine on operant performance of rats. Neurotoxicol Teratol 30:220–227. Lilleheil G, Andersen RA, Skulberg OM et al. (1997). Effects of a homoanatoxin-a-containing extract from Oscillatoria formosa (cyanophyceae/cyanobacteria) on neuromuscular transmission. Toxicon 35:1275–1289. MacPhail RC, Farmer JD, Jarema KA (2007). Effects of acute and weekly episodic exposures to anatoxin-a on motor activity in rats: Comparison with nicotine. Toxicology. 234: 83-89. Mann S, Cohen M, Chapuis-Hugon F et al. (2012). Synthesis, configuration assignment, and simultaneous quantification by liquid chromatography coupled to tandem mass spectrometry, of dihydroanatoxina and dihydrohomoanatoxin-a together with the parent toxins, in axenic cyanobacterial strains and in environmental samples. Toxicon 60:1404–1414. Metcalf JS, Žegura B, Junek R, Welker M, Törökné AK, Bláha L (in press). Analysing Cyanotoxins. In: Chorus I, Welker M, editor^editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH.

Newcombe G, Ho L, Capelo-Neto J (in press). Controlling cyanotoxin occurrence: Drinking- water treatment. In: Chorus I, Welker M, editor^editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Padisák J, Chorus I, Welker M, Maršálek B, Kurmayer R (in press). Laboratory analyses of cyanobacteria and water chemistry. In: Chorus I, Welker M, editor^editors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Puschner B, Hoff B, Tor ER. (2008). Diagnosis of anatoxin-a poisoning in dogs from North America. J Vet Diagn Invest 20:89–92. Puschner B, Pratt C, Tor ER. (2010). Treatment and diagnosis of a dog with fulminant neurological deterioration due to anatoxin-a intoxication. J Vet Emerg Crit Care 20:518–522. Rao PL, Bhattacharya R, Gupta N et al. (2002). Involvement of caspase and reactive oxygen species in cyanobacterial toxin anatoxin-a-induced cytotoxicity and apoptosis in rat thymocytes and Vero cells. Arch Toxicol 76:227–235. Rogers EH, Hunter ES, Moser VC, Phillips PM, Herkovits J, Muňoz, L. et al. (2005). Potential developmental toxicity of anatoxin-a, a cyanobacterial toxin. J. Appl. Toxicol., 25(6): 527–534. Sieroslawska A. (2013). Assessment of the mutagenic potential of cyanobacterial extracts and pure cyanotoxins. Toxicon 74:76–82. Skulberg OM, Charmichael WW, Anderson RA et al. (1992). Investigation of a neurotoxic 18

Oscillatorialean strain (cyanophiceae) and its toxin. Isolation and characterization of homoanatoxin-a. Env Toxicol Chem 11:321–329. Stevens DK and Krieger RI (1991). Effect of route of exposure and repeated doses on the acute toxicity in mice of the cyanobacterial nicotinic alkaloid anatoxin-a. Toxicon, 29(1): 134–138. Stewart I, Carmichael WW, Backer LC et al. (2011). Recreational Exposure to Cyanobacteria. In Encyclopaedia of Environmental Health. Nriagu, JO (Ed). Elsevier, Amsterdam. P776-788. Stolerman IP, Alburquerque EX, Garcha HS (1992). Behavioural effects of anatoxin, a potent nicotinic agonist, in rats. Neuropharmacol 31(3): 311–314. Testai et al. 2016 Testai E (in press). Anatoxins. In: Chorus I, Welker M, editorêditors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. US-EPA (2015). Method 545: Determination of cylindrospermopsin and anatoxin-a in drinking water by liquid chromatography electrospray ionization tandem mass spectrometry (LC/ESI- MS/MS). United States Environmental Protection Agency, Washington (DC), USA. . Welker M (in press). Fieldwork: Site inspection and sampling. In: Chorus I, Welker M, editorêditors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Welker M, Chorus I, Schaeffer B, Urquhart EA (in press). Planning monitoring of cyanobacterial occurrence. In: Chorus I, Welker M, editorêditors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH. Yavasoglu A, Karaaslan MA, Uyanikgil Y et al. (2008). Toxic effects of anatoxin-a on testes and sperm counts of male mice. Exp Toxicol Pathol 60(4–5): 391–396. Zessner M, Chorus I (in press). Catchment management. In: Chorus I, Welker M, editorêditors: Toxic Cyanobacteria in Water. 2nd ed. World Health Organisation, Geneva, CH.