Cylindrospermopsins

Andrew Humpage and Jutta Fastner

The cyanobacterium Raphidiopsis raciborskii (the renaming from Cylindrospermopsis has been widely accepted; see Chapter 3) first came to notice after the poisoning of 138 children and 10 adults on Palm Island, a tropical island off Townsville in central Queensland, Australia (Byth, 1980). Cultures of the organism were found to produce effects in mice similar to those seen in the human victims (Hawkins et al., 1985). The pure toxin - named cylindrospermopsin - was identified in 1992 (Ohtani et al., 1992).

Chemical Structures

Cylindrospermopsins (CYNs, Figure 2.2) are alkaloids comprising a tricyclic guanidino moiety linked via a hydroxylated bridging carbon (C7) to uracil (Ohtani et al., 1992). Four structural variants have been identified (Table 2.4): 7-epi-cylindrospermopsin (7-epi-CYN), 7-deoxy-cylindrospermopsin (7- deoxy-CYN), 7-deoxy-desulpho-cylindrospermopsin and 7-deoxy- desulpho-12-acetylcylindrospermopsin (Norris et al., 1999; Banker et al., 2000; Wimmer et al., 2014). The assignments of the absolute configurations of CYN and 7-epi-CYN have been exchanged, but this has little practical bearing as they are both equally toxic (Banker et al., 2000; White & Hansen, 2005). Pure CYN is a white powder and is very water soluble. It is stable to boiling and a wide range of pH (Chiswell et al., 1999).

Toxicity: Mode of Action

The toxic effects of cylindrospermopsin, summarised in the following, are described in detail in the WHO Background Document on Cylindrospermopsins (WHO, 2020); see there for further information and references). Based on available studies, the liver, kidneys and erythrocytes may

Figure 2.2 Molecular structure of common cylindrospermopsins.

Table 2.4 Congeners of cylindrospermopsin and their molecular masses

be important targets of CYN toxicity although studies using radiolabelled CYN suggest that it is distributed to all major organs. Skin patch testing produced only mild skin irritation. Since CYNs are hydrophilic molecules, facilitated transport systems mediate their intestinal absorption and uptake into other cell types, including hepatocytes. However, due to the small size of these molecules, a limited passive diffusion through biological membranes is expected. Although not clearly understood, the specific mechanism for toxicity may involve more than one mode of action, depend on the magnitude and frequency of dose, exposure duration, life stage, age or sex of the organism and the duration that an animal is observed post-dosing. At low concentrations, inhibition of protein synthesis (Terao et ah, 1994; Froscio et ah, 2003) appears to be the primary effect, which is mediated by the parent compound, whereas at higher exposures, CYN toxicity appears to involve metabolites and other mechanisms that are cytochrome P450-dependent. Reactive oxygen species and induction of stress responses may also be involved in the mode of action.

Cylindrospermopsins have been shown to be genotoxic in various mammalian cells and tissues using both in vitro and in vivo models. The extent and quality of toxicological data on CYN is quite limited, particularly because many studies have used cell extracts rather than pure toxin.

Derivation of Provisional Guideline Values

The following section is taken directly from the WHO chemicals background document on cylindrospermopsins (WHO, 2020) which discusses the considerations for the derivation of provisional guideline values for exposure to cylindrospermopsins in more detail. The Point of Departure has been identified as the no observed adverse effect level (NOAEL) of 30 pg/kg bw per day from the Humpage and Falconer (2003) study. By applying an uncertainty factor (UF) of 1000 (100 for inter- and intraspecies variability and 10 for the lack of chronic toxicity studies and deficiencies in the overall toxicological database), a provisional tolerable daily intake TDI (NOAEL/UF) of 0.03 pg/kg bw per day can be derived. The value is provisional because of deficiencies in the CYN toxicological database, essentially related to the limited availability of studies with purified toxins, lack of in vivo data on reproductive end-points and the unclear role of metabolites, especially related to potential genotoxicity. The Sukenik et al. (2006) 42-week drinking-water study provides supporting qualitative evidence for CYN toxicity, but the experimental design does not allow derivation of a robust reference value (Funari &c Testai, 2008). The study by Chernoff et al. (2018) observed many of the same effects as seen previously and demonstrates that the NOAEL is below 75 pg/kg bw per day.

The toxicological database is more limited for CYNs than for microcys- tin-LR - for example, data on on reproductive effects following oral dosing are lacking. Critically, there is evidence for potential in vivo genotoxicity of CYN. However, the lack of chronic dosing studies does not affect derivation of the short-term GV. Therefore, an uncertainty factor of 3 was used to allow for these uncertainties in the derivation of the provisional short-term drinking-water GV and recreational water GV.

For deriving the provisional lifetime drinking-water GV, the fraction of exposure allocated to drinking-water was 80% because drinking-water is expected to be the most likely long-term source of exposure. For deriving the provisional short-term drinking-water GV, the default allocation factor for short-term values of 100% was selected, considering that drinking-water is usually the most likely exposure source.

The provisional recreational water GV, which aims to protect from systemic effects, is based on a conservative scenario of a 15-kg child swallowing 250 mL of water (WHO, 2003).

Calculation of provisional lifetime drinking-water GV for CYN:

where

GV_chronk = GV for chronic (lifetime) exposure

NOAEF = no-observed-adverse-effect level (30 pg/kg bw per day, based on Humpage & Falconer, 2003)

bw = body weight (default = 60 kg for an adult)

P = fraction of exposure allocated to drinking-water (80%, because other sources of exposure, such as air, food and soil, are considered minor)

UF = uncertainty factor (1000 = 10 for interspecies variation x 10 for intraspecies variation x 10 for database deficiencies, including use of a subchronic study)

C = daily drinking-water consumption (default = 2 F for an adult).

Calculation of provisional short-term drinking-water GV for CYN:

To develop a short-term GV, the same logic was applied except that a UF of 3 was used for database limitations:

where

GV_short._term = GV for short-term exposure

NOAEL = no-observed-adverse-effect level (30 pg/kg bw per day, based on Humpage & Falconer, 2003)

bw = body weight (default = 60 kg for an adult)

P = fraction of exposure allocated to drinking-water (default for shortterm exposure = 100%, as drinking-water is expected to be the most likely source of exposure)

UF = uncertainty factor (300 = 10 for interspecies variation x 10 for intraspecies variation x 3 for database deficiencies)

C = daily drinking-water consumption (default = 2 L for an adult).

Calculation of provisional recreational water GV for CYN:

where

GV_recreation = GV for recreational water exposure

NOAEL = no-observed-adverse-effect level (30 pg/kg bw per day, based on Humpage & Falconer, 2003) bw = body weight (default = 15 kg for a child)

UF = uncertainty factor (300 = 10 for interspecies variation x 10 for intraspecies variation x 3 for database deficiencies)

C = daily incidental water consumption (default = 250 mL for a child).

Considerations in applying the provisional guideline values

The provisional GVs are based on toxicological data for CYN. The limited evidence on the relative potency of other CYN congeners suggests they are probably similar in potency to CYN. Therefore, for assessing risk, as a conservative approach, it is suggested that the sum of of CYNs (on a molar basis), be evaluated against the GV.

In some regions, others sources of exposure besides drinking-water can be significant (see chapter 5). This includes food from locations where blooms have a long duration and there is high consumption of locally affected food items. In such situations, it may be appropriate to consider reducing the allocation factor for the lifetime and short-term drinking-water GVs based on relative exposure data for the population.

The short-term drinking-water GV is based on exposure of 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 than for adults), it is recommended that alternative water sources such as bottled water are provided for bottle-fed infants and small children when CYN concentrations are greater than 0.7 pg/L even for short periods, as a precautionary measure.