- Email: [email protected]
Relative toxicity of dinophysistoxin-2 (DTX-2) compared with okadaic acid, based on acute intraperitoneal toxicity in mice
ARTICLE IN PRESS
Toxicon 49 (2007) 1–7 www.elsevier.com/locate/toxicon
Relative toxicity of dinophysistoxin-2 (DTX-2) compared with okadaic acid, based on acute intraperitoneal toxicity in mice Tore Aunea,, Stig Larsena, John A.B. Aasena, Nils Rehmannb, Masayuki Satakec, Philipp Hessb a
Norwegian School of Veterinary Science, P. O. Box 8146 Dep., 0033 Oslo, Norway b Marine Institute, Rinville, Oranmore, Co. Galway, Ireland c Tohuku University, Sendai, Japan
Received 26 June 2006; received in revised form 21 July 2006; accepted 25 July 2006 Available online 14 August 2006
Abstract When substituting the mouse bioassay for lipophilic marine algal toxins in shellfish with analytical methods, science based factors of relative toxicity for all analogues that contribute to health risk to consumers are necessary. The aim of this paper is to establish the relative intraperitoneal toxicity of dinophysistoxin-2 (DTX-2) compared with okadaic acid (OA). The study was performed as an open, randomised parallel group trial with a four level response surface design within each of the two parallels. In accordance with the response surface design model, the LD50 for DTX-2 and OA was 338 and 206 mg/kg, respectively. By use of common regression analysis, the LD50 of DTX-2 and OA were estimated to 352mg/kg and 204mg/kg, respectively. The deviations between the LD50 estimates by the two methods was 4% for DTX-2 and less than 1% for OA. Taken together, these results indicate that the relative toxicity of DTX-2 is about 0.6, compared to OA. Results from the PP2A assay correspond very well with the results obtained by the mouse bioassay. The IC50 concentrations for DTX-2 and OA were 5.94 and 2.81 ng/mL, respectively. This indicates that OA is about twice as toxic as DTX-2. Since inhibition of PP2A is acknowledged as the main mechanism of toxicity of the OA group toxins, this supports the establishment of a relative toxicity factor of DTX-2 of 0.6 compared with OA. r 2006 Elsevier Ltd. All rights reserved. Keywords: Dinophysistoxin-2; DTX-2; OA; Toxicity; Relative toxicity factor
1. Introduction The okadaic acid (OA) group of lipophilic toxins, which includes okadaic acid and dinophysistoxins 1,2 and 3 (DTX-1,-2,-3) are commonly found in shellfish in Europe. Consumption of shellfish containing these toxins can lead to human illness whose Corresponding author.
E-mail address: [email protected] (T. Aune). 0041-0101/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2006.07.033
symptoms include nausea, vomiting and diarrhoea. This form of human illness has been termed diarrhetic shellfish poisoning (DSP) and the traditional method, originally established for the analysis and control of these toxins in shellfish is the mouse bioassay (MBA, Yasumoto et al., 1978). This method of analysis is still used in monitoring programmes throughout Europe and is currently the Reference method under EU legislation. Although this bioassay was well suited for control
ARTICLE IN PRESS 2
T. Aune et al. / Toxicon 49 (2007) 1–7
of shellfish to avoid (DSP) when toxins from the okadaic acid (OA) group toxins were present alone, it has become increasingly clear that the MBA is unfit for control of the varying mixture of lipophilic marine biotoxins that may accumulate in shellfish. In addition to the OA group toxins, other toxin groups, including yessotoxins (YTXs), pectenotoxins (PTXs), azaspiracids (AZAs) and cyclic imines will also contribute to the outcome of the MBA. There is growing concern about the continued use of animal assays and a recognised urgent need to replace routine methods that brings pain to many experimental animals, with alternative, non-animal methods (Hess et al., 2006). According to the FAO/ IOC/WHO Expert Consultation in 2004 (Anonymous, 2005), the most promising alternative methods (and candidates for status as future reference methods) for the lipophilic toxins, are based on LC-MS. Furthermore, several international expert groups, among them the FAO/IOC/WHO Expert Consultation (2004) have suggested changes in the maximum tolerance levels (MLs)/guidance levels (GLs) for the different toxin groups. Individual levels for the different toxin groups require analytical methods able to identify and quantify each toxin group. Consequently, the method of analysis of lipophilic toxins in shellfish should no longer be based on the MBA, since this method is lacking both sensitivity and specificity, and cannot distinguish between toxins from the different groups (Hess et al., 2006). However, when applying instrument based analytical methods such as LC-MS in official control programmes, knowledge of the relative toxicities of analogues within each toxin group is required in order to calculate total toxicity. DTX-2 was first described in extracts from Irish mussels by Hu et al. (1992). Their preliminary studies indicated that DTX-2 exerts phosphatase inhibition comparable to OA and DTX-1 and comparable activity in the rat bioassay. DTX-2 has been shown to be the major toxin associated with elevated levels of OA-group toxins in cultivated mussels in Ireland, (Carmody et al., 1995, 1996). While OA is found world wide, and is reported to be the dominating OA-group toxin in Europe (Kumagi et al., 1986), DTX-2 is reported in shellfish from Spain (Gago-Martı´ nez et al., 1996) and Portugal (Vale and Sampayo, 2002) in addition to Ireland. From the beginning of the twenty-first Century, DTX-2 has also constituted a significant
part of OA-group toxins in mussels from Southern Norway (unpublished results). In Ireland DTX-2 is typically detected in shellfish, principally mussel (Mytilus edulis), in late July early August and has been associated with the occurrence of the dinoflagellate Dinophysis acuta in the plankton (McMahon et al, 1996). According to Vale (2004), however, in Portugal increasing levels of DTX-2 compared to OA in mussels in the fall is due to slower elimination of the former, and not higher levels in the toxin producing Dinophysis spp. He made similar observations in clams (Donax spp.). For the OA-group, relative toxicity factors have been established for OA and DTX-1 (both with factor 1). Furthermore, the relative toxicity of DTX-3 (the 7-O-acyl fatty acid esters of OA, DTX-1 and DTX-2) has been suggested at the same level (Professor Yasumoto, contribution to the FAO/IOC/WHO Expert Consultation, 2004). To date, however, the relative toxicity of DTX-2 has not been scientifically established. For several years, the lipophilic toxins have been analysed both by the MBA and LC-MS in parallel by the Norwegian National Reference Laboratory. Results from recent years indicated lower acute toxicity of DTX-2 in the MBA, compared with OA and DTX-1. Consequently, it is important to find the correct relative toxicity factor for DTX-2 when substituting the MBA with LC-MS analyses. In order to study this, elucidation of the relative intraperitoneal (i.p.) toxicities of DTX-2 and OA was undertaken and the results of this work are reported in this paper. 2. Materials and methods 2.1. Study design The study was performed as an open, randomised parallel group designed study. Within each of the two parallel arms, a four level response surface design was applied (Figs. 1 and 2). The mice were equally allocated to the two parallel arms by block randomisation. In the arm randomised to DTX-2, five mice were randomly selected and injected with 500 mg/kg DTX-2 (Table 1). If the prevalence of death among these mice is found 450%, seven new mice would be given 400 mg/kg DTX-2. Otherwise, the seven new mice would receive 600 mg/kg. In the third and the fourth step, additionally seven and nine mice were included, respectively (Fig. 1). In the
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
3
Fig. 1. The four level pathway design used for DTX-2. The possible doses to be used on each design level are given in the boxes. The obtained pathway in the present study is shadowed and the ratios between dead mice and number of mice used are given in brackets.
Fig. 2. The four level pathway design used for OA. The possible doses to be used on each design level are given in the boxes. The obtained pathway in the present study is shadowed and the ratios between dead mice and number of mice used are given in brackets.
arm randomised to OA, five mice were randomly selected and injected with 300 mg/kg OA in the first step. An identical procedure as previously described for DTX-2 was used (Fig. 2).
2.2. Mouse bioassay Mouse strain, CD-1 (Harlan), females, weighing 19-22.5 g were used. The mice were injected body
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
4 Table 1 Number of mice used at each step Toxin injection
Design level
Total
1
2
3
4
DTX-2 Okadaic acid
5 5
7 7
7 7
9 9
28 28
Total
10
14
14
18
56
weight adjusted solutions of the toxins, 1.0 mL/20 g body weight, and they were observed for 24 hours, in accordance with the MBA for DSP toxins. A total of 28 mice in each of two study arms were included in the trial (Table 1). Symptoms and time to death were recorded. 2.3. Preparation of DTX-2 toxin 1 kg of digestive gland (hepatopancreas, HP) was obtained from 6 kg of pre-cooked and shelled mussels (Mytilus edulis) (Castletownbere, Ireland, August 2001). The HP was extracted twice with methanol (3 L). The extract was evaporated to dryness and partitioned twice between hexane (1 L) and 80:20 v/v methanol/water (1 L). Both aqueous phases were combined and evaporated. The extract was then partitioned twice between ethyl acetate (1 L) and water (1 L). The organic phases were combined and evaporated. The sample was suspended in acetone and loaded onto a silica gel column (500 40 mm) (Silica Gel 60 Merck). The column was first eluted with three times the bed volume acetone and then with three times the bed volume methanol. The toxins were found in the methanolic fraction using LC-MS. The sample was again evaporated and dissolved in 700:300:1 v/v/v propanol/water/acetic acid and separated on Toyopearl HW-40 (Tosoh, 20 mm i.d. 25 cm) with 700:300:1 v/v/v propanol/ water/acetic acid as mobile phase at 1 mL/min. 5 mL fraction were collected and the toxins were found to elute around fractions 6–9 (30–45 min). The toxic fractions were combined and separated on Sephadex LH-20 (Amersham Biosciences, 10 mm i.d. 40 cm). Methanol was used as a mobile phase at 1 mL/min. The toxic fractions were combined and the solvent was evaporated. The residue was then further purified on Develosil ODS (Phenomenex) using 850:150:1 v/v/v methanol/water/acetic acid as mobile phase at 1mL/min. The two toxin groups
(AZAs and DTXs) were almost entirely separated. A mixed fraction, containing DTX-2 and AZA, was re-chromatographed on the Develosil to achieve complete separation. The DTX-2 containing fractions were combined and passed through DEAE anion-exchange material (Tosoh). A step gradient was run from 80:20 v/v methanol/water and then 850:150:1 v/v/v methanol/water/acetic acid at flowrates of 0.5 ml/min (60 min) and 1mL/min (30 min) respectively. DTX-2 eluted in the latter, acidic fraction. DTX-2 was purified using isocratic separation on a semi preparative C18-silica column (ZORBAX Eclipse, Agilent, 9.4 mm i.d. 25 cm) with 60:40 v/ v acetonitrile/water containing ammonium formate and formic acid as a buffer. The flow rate was set to 3 ml/min. The chromatogram was monitored using the MS detector and fractions were collected manually. An MS-Scan was carried out on the DTX-2 isolated in this step. The scanned masses ranged from 50–1300 da. No other compound eluted in this trace over 15 min, with DTX-2 eluting around 2.8 min. NMR analysis showed no major contamination of the pure compound and LC-MS analysis showed 800 mg of toxin present. Okadaic acid was bought from CALBIOCHEM, (La Jolla, CA, USA) in 100mg ampoules, (cat. 495624), and quantified against a certified standard of okadaic acid from NRC, Canada. (NRC-CRMOA-b ) before use. Protein phosphatase 2A (PP2A) assay: The assay was performed with the Promega PPase-2A kit, applying enzyme from human red blood cells. The assay was performed according to the procedure suggested by the producer. The assay measures the inhibitory effect of the test compound on PP2Aassociated dephosphorylation of the substrate paranitrophenyl-phosphate (PNPP). The concentration range applied was 1–10 ng/mL for both toxins. Statistical analyses: All categorised variables and factors are given in contingency tables and prevalence given in percent with 95% confidence intervals constructed by use of simple binomial sequences (Agresti, 2002). In order to estimate LD50, both the response surface design and a polynomial regression analysis were performed (Myers and Montgomery 2002, Kleinbaum et al., 1998). 3. Results Before the main study, crude estimates of the acute i.p. toxicity of OA and DTX-2 were compared
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
Table 2 Dose levels and the prevalence of death for DTX-2 and Okadaic acid Toxin
DTX-2
Okadaic acid
Dose levels used (mg/kg bw)
Number of mice used
Number of dead mice
500
5
4
400
7
4
350
7
6
325
9
1
300
5
4
250
7
6
225
9
7
200
7
3
Prevalence of death in % with 95% Confidence interval 80.0 28.4–99.5 57.1 18.4–90.1 85.7 42.1–99.6 11.1 0.3–48.3 80.0 28.4–99.5 85.7 42.1–99.6 77.8 40.0–97.2 42.9 9.9–81.6
Table 3 Estimated LD50 for DTX-2 and Okadaic acid in mg/kg bw, based on regression analysis and the response surface model Toxin
Regression model
Response surface model
DTX-2 Okadaic acid
352 204
338 206
Prevalence
0.80
0.60
0.40
0.20 DTX-2 0.00 300
350
400
450
Dose
(A)
0.8
0.6 Prevalence
by applying a dose-regime between 200 and 400 mg/ kg bw for OA and 200-600 mg/kg bw for DTX-2, each dose in duplicate. The preliminary results indicated that the i.p. toxicity of OA was about twice as high, compared with DTX-2, and, consequently, the initial doses were chosen at 300 mg/kg OA and 500 mg/kg of DTX-2, respectively. In the DTX-2 arm of the study, a total of 15 of the 28 mice died (Table 2). At the first dose level of 500 mg/kg, four of five (80%) of the used mice died which led to a dose reduction to 400 mg/kg (Fig. 1). At this level four of seven mice (57.1%) died and the dose was further reduced to 350 mg/kg. The third level resulted in six of seven (85.7%) dead mice and the dose was further reduced to 325 mg/kg. This gave a death prevalence of 11.1% (one of nine mice). By direct interpolation in the repose surface design, the LD50 was estimated to 338 mg/kg (Table 3). The prevalence of death in this DTX-2 arm was best fitted by a second degree polynome in which the dose explains 45.3% of the variation (Fig. 3A). By use of this regression model, the LD50 of DTX-2 was estimated to be 352 mg/kg (Table 3). The percent deviation between LD50 estimated by the design and the regression model was 4%. Of the 28 mice used in the OA arm, 20 mice died (Table 2). At the first dose level using 300 mg/kg, four
5
0.4
0.2
OA
0.0 200
(B)
220
240
Dose
Fig. 3. The prevalence of death given as a function of A) DTX-2 dose and B) Okadaic acid dose, predicted by use of second-degree polynome.
of five (80%) of the used mice died (Fig. 2). This led to a reduced dose of 200 mg/kg for the second level. At this level three of seven (42.9%) died and the doses was increased to 250 mg/kg. The third level resulted in six of seven (85.7%) dead mice and the dose was again reduced to 225 mg/kg. This gave a death prevalence of 77.8% (seven of nine mice). By direct interpolation in the response surface design, using the results obtained for the two levels closest to LD50, the LD50 was estimated to 206 mg/kg (Table 3). Also in the OA arm, the prevalence of death was best
ARTICLE IN PRESS 6
T. Aune et al. / Toxicon 49 (2007) 1–7
Fig. 4. The PP2A activity rate for (A) DTX-2 and (B) Okadaic acid (OA) expressed as a second-degree polynome of ng/mL toxin with shadowed 95% confidence boundary.
fitted by a second degree polynome in which the dose explains 90.2% of the variation (Fig. 3B). By use of this regression model, the LD50 of OA was estimated to be 204 mg/kg (Table 3). The percent deviation between LD50 estimated by the design and the regression model was found less then 1%. The PP2A assay was performed four times for DTX-2 and 6 times for OA. For both toxins, the protein phosphatase activity rate was best expressed by a second-degree polynome of the dose (Fig. 4A and B). In this study, 97.4% and 96.5% of the variation in the inhibition rate for DTX-2 and OA, respectively was explained by the dose. The IC50 was estimated to 5.94 ng/ml for DTX-2 (Fig. 4A) and 2.81 ng/ml for OA (Fig. 4B). 4. Discussion Results from the mouse bioassays carried out in this study clearly indicate that DTX-2 is less potent
compared with OA when injected intraperitoneally in mice. The relative i.p. toxicity of DTX-2 is about 0.6 that of OA. When calculating total toxicity of OA group toxins based on results from LC-MS, it is suggested that analytical results of DTX-2 should be multiplied by the relative toxicity factor of 0.6. Ideally, toxicity data from experimental animals should be based on results from exposure via the oral route, since the results are used to protect human consumers. However, scarcity of pure toxins has made it necessary to apply data from i.p. injections up till now. Similar practice seems to have functioned satisfactorily for other toxin groups such as the saxitoxins, where toxicity equivalence factors (TEFs) according to Professor Oshima are accepted and used internationally (Anonymous, 2005). The results from the PP2A assay, support the data from the MBA since the IC50 of DTX-2 is about twice the value for OA. This fits well with the internationally accepted consensus that inhibition of protein phosphatase 2A constitutes the main mechanism of toxicity of the OA toxin group. Interestingly, Pere´z-Go´mez et al. (2004) found DTX-2 to be less potent than OA when comparing their neurotoxic effect on cultured cerebellar neurons from rats. In all laboratory animal studies it is of utmost importance to reduce the number of animals to an absolute minimum without reducing the needed information. The only way to obtain this is by optimization of the study design. A common approach in dose finding studies is first to establish a dose window and then divide this into equal distanced doses. An equal number of animals will then be used on these dose levels and the data obtained analysed by a regression model. Usually, a lot of the animals included in such a study approach will give none or limited information. This problem might be overcome by selecting doses closer to the actual level. In most cases, prior knowledge of such an actual dose is limited. However, by using a response surface approach, this can be achieved (Myers and Montgomery, 2002). It has been previously shown that by changing from the common approach as described above to a response surface method, the sample size was reduced from 196 to 64 animals (Ryeng et al., 2001). In the present study, a four level pathway design has been used. By increasing the number of levels in the design, the information will obviously increase rapidly, but so will the number of mice to be included. In order to keep the number of animals to
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
an absolute minimum, only four levels were used. The amount of information will increase considerably by including one dose level more and if possible this is to be recommended. In such a situation at least 9 more mice had to be included in each arm. In the present study this would have resulted in 9 mice at 337.5 mg/kg DTX-2 and 9 mice on 212.5 mg/kg OA. In both cases, this would have been very close to the estimated LD50. Current work at the Marine Institute investigates the possibility of producing a certified standard for DTX-2, as well as a certified mussel tissue reference material containing DTX-2, in collaboration with the National Research Council Canada. Acknowledgements Financial support from the Norwegian Research Council, Project No. 172561/S40, is very much appreciated. Technical assistance from Rolf Harestad on PP2A assay and Brit Heidenreich on the mouse bioassay is also very much appreciated. Financial support was obtained from the Irish National Development Plan and the Marine Institute Ireland, through the ASTOX-project (ST-0202). We gratefully acknowledge critical comments to the manuscript by Dr. Terry McMahon. References Anonymous (2005). Report of the Joint FAO/IOC/WHO ad hoc Expert Consultation on Biotoxins in Molluscan Bivalves (Oslo, Norway, 26-30 September 2004), 31 pp. Rome, Italy: Food and Agriculture Organization. Website http:// www.fao.org/es/esn/food/risk_biotoxin_en.stm. Agresti, A., 2002. Categorical Data Analysis, 2nd edition. Wiley, New Jersey. Carmody, E.P., James, K.J., Kelly, S.S., Thomas, K., 1995. Complex diarrhetic shellfish toxin profiles in Irish mussels. In: Lassus, P., Arzul, G., Erard, E., Gentien, P., Marcaillou, C. (Eds.), Harmful Marine Algal Blooms. Lavoisier Science Publishers, Paris, pp. 273–278.
7
Carmody, E.P., James, K.J., Kelly, S.S., 1996. Dinophysistoxin2: The predominant diarrhoetic shellfish toxin in Ireland. Toxicon 34 (3), 351–359. Gago-Martı´ nez, A., Rodrı´ guez-Va´zquez, J.A., Tibault, P., Quilliam, M.A., 1996. Simultaneous occurrence of diarrhetic and paralytic shellfish poisoning toxins in Spanish mussels in 1993. Nat. Toxins 4 (2), 72–79. Hess, P., Grune, B., Anderson, D.B., et al., 2006. Three Rs approaches in marine biotoxin testing. ATLA 34, 193–224. Hu, T., Doyle, J., Jackson, D., Marr, J., Nixon, E., Pleasance, S., Quilliam, M.A., Walter, J.A., Wright, J.L.C., 1992. Isolation of a new diarrhetic shellfish poison from Irish mussels. J.Chem.Soc.Chem.Commun,, 39–41. Kleinbaum, D.G., Kupper, L.L., Muller, K.E., Nizam, A., 1998. Applied Regression Analysis and other Multivariable Methods. Duxbury Press, Pacific Grove. Kumagi, M., Yanagi, T., Murata, M., Yasumoto, T., Kat, M., Lassus, P., Rodriguez-Vazquez, J.A., 1986. Okadaic acid as the causative toxin of diarrhetic shellfish poisoning in Europe. Agic. Biol. Chem. 50, 2853–2857. McMahon, T., Nixon, E., Silke, J., Taffe, B., Nolan, A., McGovern, E., Doyle, J., 1996. Seasonal variation in diarrhetic shellfish toxins in mussels from the southwest coast of Ireland in 1994. In: Keegan, B.F., O’Connor, R. (Eds.), Irish Marine Science 1995. Galway University Press, Galway, pp. 417–432. Myers, R.H., Montgomery, D.C., 2002. Response Surface Methodology: Process and Product Optimisation Using Designed Experiments, 2nd Edition. Wiley, New Jersey. Pere´z-Go´mez, A., Garcia-Rodrı´ guez, A., James, K.J., ferreroGutierrez, A., Novelli, A., Fernandez-Sa´nchez, M.T., 2004. The marine toxin dinophysistoxin-2 induces differential apoptotic death of rat cerebellar neurons and astrocytes. Toxicological Sciences 80, 74–82. Ryeng, K.A., Arnemo, J.M., Larsen, S., 2001. Determination of optimal immobilizing doses of medetomidine hydrochloride and ketamine hydrochloride in captive reindeer. AJVR 62, 119–126. Vale, P., 2004. Differential dynamics of dinophysistoxins and pectenotoxins between blue mussels and common cockle: a phenomenon originating from the complex toxin profile of Dinophysis acuta. Toxicon 44 (2), 123–134. Vale, P., Sampayo, M.A.deM., 2002. Esterification of DSP toxins by Portuguese bivalves from the Northwestern coast determined by LC-MS—a widespread phenomenon. Toxicon 40, 33–42. Yasumoto, T., Oshima, Y., Yamaguchi, M., 1978. Occurrence of a new type of shellfish poisoning in the Tohuku District. Bull. Jap. Soc. Sci. Fish. 44, 1249–1255.
Toxicon 49 (2007) 1–7 www.elsevier.com/locate/toxicon
Relative toxicity of dinophysistoxin-2 (DTX-2) compared with okadaic acid, based on acute intraperitoneal toxicity in mice Tore Aunea,, Stig Larsena, John A.B. Aasena, Nils Rehmannb, Masayuki Satakec, Philipp Hessb a
Norwegian School of Veterinary Science, P. O. Box 8146 Dep., 0033 Oslo, Norway b Marine Institute, Rinville, Oranmore, Co. Galway, Ireland c Tohuku University, Sendai, Japan
Received 26 June 2006; received in revised form 21 July 2006; accepted 25 July 2006 Available online 14 August 2006
Abstract When substituting the mouse bioassay for lipophilic marine algal toxins in shellfish with analytical methods, science based factors of relative toxicity for all analogues that contribute to health risk to consumers are necessary. The aim of this paper is to establish the relative intraperitoneal toxicity of dinophysistoxin-2 (DTX-2) compared with okadaic acid (OA). The study was performed as an open, randomised parallel group trial with a four level response surface design within each of the two parallels. In accordance with the response surface design model, the LD50 for DTX-2 and OA was 338 and 206 mg/kg, respectively. By use of common regression analysis, the LD50 of DTX-2 and OA were estimated to 352mg/kg and 204mg/kg, respectively. The deviations between the LD50 estimates by the two methods was 4% for DTX-2 and less than 1% for OA. Taken together, these results indicate that the relative toxicity of DTX-2 is about 0.6, compared to OA. Results from the PP2A assay correspond very well with the results obtained by the mouse bioassay. The IC50 concentrations for DTX-2 and OA were 5.94 and 2.81 ng/mL, respectively. This indicates that OA is about twice as toxic as DTX-2. Since inhibition of PP2A is acknowledged as the main mechanism of toxicity of the OA group toxins, this supports the establishment of a relative toxicity factor of DTX-2 of 0.6 compared with OA. r 2006 Elsevier Ltd. All rights reserved. Keywords: Dinophysistoxin-2; DTX-2; OA; Toxicity; Relative toxicity factor
1. Introduction The okadaic acid (OA) group of lipophilic toxins, which includes okadaic acid and dinophysistoxins 1,2 and 3 (DTX-1,-2,-3) are commonly found in shellfish in Europe. Consumption of shellfish containing these toxins can lead to human illness whose Corresponding author.
E-mail address: [email protected] (T. Aune). 0041-0101/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2006.07.033
symptoms include nausea, vomiting and diarrhoea. This form of human illness has been termed diarrhetic shellfish poisoning (DSP) and the traditional method, originally established for the analysis and control of these toxins in shellfish is the mouse bioassay (MBA, Yasumoto et al., 1978). This method of analysis is still used in monitoring programmes throughout Europe and is currently the Reference method under EU legislation. Although this bioassay was well suited for control
ARTICLE IN PRESS 2
T. Aune et al. / Toxicon 49 (2007) 1–7
of shellfish to avoid (DSP) when toxins from the okadaic acid (OA) group toxins were present alone, it has become increasingly clear that the MBA is unfit for control of the varying mixture of lipophilic marine biotoxins that may accumulate in shellfish. In addition to the OA group toxins, other toxin groups, including yessotoxins (YTXs), pectenotoxins (PTXs), azaspiracids (AZAs) and cyclic imines will also contribute to the outcome of the MBA. There is growing concern about the continued use of animal assays and a recognised urgent need to replace routine methods that brings pain to many experimental animals, with alternative, non-animal methods (Hess et al., 2006). According to the FAO/ IOC/WHO Expert Consultation in 2004 (Anonymous, 2005), the most promising alternative methods (and candidates for status as future reference methods) for the lipophilic toxins, are based on LC-MS. Furthermore, several international expert groups, among them the FAO/IOC/WHO Expert Consultation (2004) have suggested changes in the maximum tolerance levels (MLs)/guidance levels (GLs) for the different toxin groups. Individual levels for the different toxin groups require analytical methods able to identify and quantify each toxin group. Consequently, the method of analysis of lipophilic toxins in shellfish should no longer be based on the MBA, since this method is lacking both sensitivity and specificity, and cannot distinguish between toxins from the different groups (Hess et al., 2006). However, when applying instrument based analytical methods such as LC-MS in official control programmes, knowledge of the relative toxicities of analogues within each toxin group is required in order to calculate total toxicity. DTX-2 was first described in extracts from Irish mussels by Hu et al. (1992). Their preliminary studies indicated that DTX-2 exerts phosphatase inhibition comparable to OA and DTX-1 and comparable activity in the rat bioassay. DTX-2 has been shown to be the major toxin associated with elevated levels of OA-group toxins in cultivated mussels in Ireland, (Carmody et al., 1995, 1996). While OA is found world wide, and is reported to be the dominating OA-group toxin in Europe (Kumagi et al., 1986), DTX-2 is reported in shellfish from Spain (Gago-Martı´ nez et al., 1996) and Portugal (Vale and Sampayo, 2002) in addition to Ireland. From the beginning of the twenty-first Century, DTX-2 has also constituted a significant
part of OA-group toxins in mussels from Southern Norway (unpublished results). In Ireland DTX-2 is typically detected in shellfish, principally mussel (Mytilus edulis), in late July early August and has been associated with the occurrence of the dinoflagellate Dinophysis acuta in the plankton (McMahon et al, 1996). According to Vale (2004), however, in Portugal increasing levels of DTX-2 compared to OA in mussels in the fall is due to slower elimination of the former, and not higher levels in the toxin producing Dinophysis spp. He made similar observations in clams (Donax spp.). For the OA-group, relative toxicity factors have been established for OA and DTX-1 (both with factor 1). Furthermore, the relative toxicity of DTX-3 (the 7-O-acyl fatty acid esters of OA, DTX-1 and DTX-2) has been suggested at the same level (Professor Yasumoto, contribution to the FAO/IOC/WHO Expert Consultation, 2004). To date, however, the relative toxicity of DTX-2 has not been scientifically established. For several years, the lipophilic toxins have been analysed both by the MBA and LC-MS in parallel by the Norwegian National Reference Laboratory. Results from recent years indicated lower acute toxicity of DTX-2 in the MBA, compared with OA and DTX-1. Consequently, it is important to find the correct relative toxicity factor for DTX-2 when substituting the MBA with LC-MS analyses. In order to study this, elucidation of the relative intraperitoneal (i.p.) toxicities of DTX-2 and OA was undertaken and the results of this work are reported in this paper. 2. Materials and methods 2.1. Study design The study was performed as an open, randomised parallel group designed study. Within each of the two parallel arms, a four level response surface design was applied (Figs. 1 and 2). The mice were equally allocated to the two parallel arms by block randomisation. In the arm randomised to DTX-2, five mice were randomly selected and injected with 500 mg/kg DTX-2 (Table 1). If the prevalence of death among these mice is found 450%, seven new mice would be given 400 mg/kg DTX-2. Otherwise, the seven new mice would receive 600 mg/kg. In the third and the fourth step, additionally seven and nine mice were included, respectively (Fig. 1). In the
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
3
Fig. 1. The four level pathway design used for DTX-2. The possible doses to be used on each design level are given in the boxes. The obtained pathway in the present study is shadowed and the ratios between dead mice and number of mice used are given in brackets.
Fig. 2. The four level pathway design used for OA. The possible doses to be used on each design level are given in the boxes. The obtained pathway in the present study is shadowed and the ratios between dead mice and number of mice used are given in brackets.
arm randomised to OA, five mice were randomly selected and injected with 300 mg/kg OA in the first step. An identical procedure as previously described for DTX-2 was used (Fig. 2).
2.2. Mouse bioassay Mouse strain, CD-1 (Harlan), females, weighing 19-22.5 g were used. The mice were injected body
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
4 Table 1 Number of mice used at each step Toxin injection
Design level
Total
1
2
3
4
DTX-2 Okadaic acid
5 5
7 7
7 7
9 9
28 28
Total
10
14
14
18
56
weight adjusted solutions of the toxins, 1.0 mL/20 g body weight, and they were observed for 24 hours, in accordance with the MBA for DSP toxins. A total of 28 mice in each of two study arms were included in the trial (Table 1). Symptoms and time to death were recorded. 2.3. Preparation of DTX-2 toxin 1 kg of digestive gland (hepatopancreas, HP) was obtained from 6 kg of pre-cooked and shelled mussels (Mytilus edulis) (Castletownbere, Ireland, August 2001). The HP was extracted twice with methanol (3 L). The extract was evaporated to dryness and partitioned twice between hexane (1 L) and 80:20 v/v methanol/water (1 L). Both aqueous phases were combined and evaporated. The extract was then partitioned twice between ethyl acetate (1 L) and water (1 L). The organic phases were combined and evaporated. The sample was suspended in acetone and loaded onto a silica gel column (500 40 mm) (Silica Gel 60 Merck). The column was first eluted with three times the bed volume acetone and then with three times the bed volume methanol. The toxins were found in the methanolic fraction using LC-MS. The sample was again evaporated and dissolved in 700:300:1 v/v/v propanol/water/acetic acid and separated on Toyopearl HW-40 (Tosoh, 20 mm i.d. 25 cm) with 700:300:1 v/v/v propanol/ water/acetic acid as mobile phase at 1 mL/min. 5 mL fraction were collected and the toxins were found to elute around fractions 6–9 (30–45 min). The toxic fractions were combined and separated on Sephadex LH-20 (Amersham Biosciences, 10 mm i.d. 40 cm). Methanol was used as a mobile phase at 1 mL/min. The toxic fractions were combined and the solvent was evaporated. The residue was then further purified on Develosil ODS (Phenomenex) using 850:150:1 v/v/v methanol/water/acetic acid as mobile phase at 1mL/min. The two toxin groups
(AZAs and DTXs) were almost entirely separated. A mixed fraction, containing DTX-2 and AZA, was re-chromatographed on the Develosil to achieve complete separation. The DTX-2 containing fractions were combined and passed through DEAE anion-exchange material (Tosoh). A step gradient was run from 80:20 v/v methanol/water and then 850:150:1 v/v/v methanol/water/acetic acid at flowrates of 0.5 ml/min (60 min) and 1mL/min (30 min) respectively. DTX-2 eluted in the latter, acidic fraction. DTX-2 was purified using isocratic separation on a semi preparative C18-silica column (ZORBAX Eclipse, Agilent, 9.4 mm i.d. 25 cm) with 60:40 v/ v acetonitrile/water containing ammonium formate and formic acid as a buffer. The flow rate was set to 3 ml/min. The chromatogram was monitored using the MS detector and fractions were collected manually. An MS-Scan was carried out on the DTX-2 isolated in this step. The scanned masses ranged from 50–1300 da. No other compound eluted in this trace over 15 min, with DTX-2 eluting around 2.8 min. NMR analysis showed no major contamination of the pure compound and LC-MS analysis showed 800 mg of toxin present. Okadaic acid was bought from CALBIOCHEM, (La Jolla, CA, USA) in 100mg ampoules, (cat. 495624), and quantified against a certified standard of okadaic acid from NRC, Canada. (NRC-CRMOA-b ) before use. Protein phosphatase 2A (PP2A) assay: The assay was performed with the Promega PPase-2A kit, applying enzyme from human red blood cells. The assay was performed according to the procedure suggested by the producer. The assay measures the inhibitory effect of the test compound on PP2Aassociated dephosphorylation of the substrate paranitrophenyl-phosphate (PNPP). The concentration range applied was 1–10 ng/mL for both toxins. Statistical analyses: All categorised variables and factors are given in contingency tables and prevalence given in percent with 95% confidence intervals constructed by use of simple binomial sequences (Agresti, 2002). In order to estimate LD50, both the response surface design and a polynomial regression analysis were performed (Myers and Montgomery 2002, Kleinbaum et al., 1998). 3. Results Before the main study, crude estimates of the acute i.p. toxicity of OA and DTX-2 were compared
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
Table 2 Dose levels and the prevalence of death for DTX-2 and Okadaic acid Toxin
DTX-2
Okadaic acid
Dose levels used (mg/kg bw)
Number of mice used
Number of dead mice
500
5
4
400
7
4
350
7
6
325
9
1
300
5
4
250
7
6
225
9
7
200
7
3
Prevalence of death in % with 95% Confidence interval 80.0 28.4–99.5 57.1 18.4–90.1 85.7 42.1–99.6 11.1 0.3–48.3 80.0 28.4–99.5 85.7 42.1–99.6 77.8 40.0–97.2 42.9 9.9–81.6
Table 3 Estimated LD50 for DTX-2 and Okadaic acid in mg/kg bw, based on regression analysis and the response surface model Toxin
Regression model
Response surface model
DTX-2 Okadaic acid
352 204
338 206
Prevalence
0.80
0.60
0.40
0.20 DTX-2 0.00 300
350
400
450
Dose
(A)
0.8
0.6 Prevalence
by applying a dose-regime between 200 and 400 mg/ kg bw for OA and 200-600 mg/kg bw for DTX-2, each dose in duplicate. The preliminary results indicated that the i.p. toxicity of OA was about twice as high, compared with DTX-2, and, consequently, the initial doses were chosen at 300 mg/kg OA and 500 mg/kg of DTX-2, respectively. In the DTX-2 arm of the study, a total of 15 of the 28 mice died (Table 2). At the first dose level of 500 mg/kg, four of five (80%) of the used mice died which led to a dose reduction to 400 mg/kg (Fig. 1). At this level four of seven mice (57.1%) died and the dose was further reduced to 350 mg/kg. The third level resulted in six of seven (85.7%) dead mice and the dose was further reduced to 325 mg/kg. This gave a death prevalence of 11.1% (one of nine mice). By direct interpolation in the repose surface design, the LD50 was estimated to 338 mg/kg (Table 3). The prevalence of death in this DTX-2 arm was best fitted by a second degree polynome in which the dose explains 45.3% of the variation (Fig. 3A). By use of this regression model, the LD50 of DTX-2 was estimated to be 352 mg/kg (Table 3). The percent deviation between LD50 estimated by the design and the regression model was 4%. Of the 28 mice used in the OA arm, 20 mice died (Table 2). At the first dose level using 300 mg/kg, four
5
0.4
0.2
OA
0.0 200
(B)
220
240
Dose
Fig. 3. The prevalence of death given as a function of A) DTX-2 dose and B) Okadaic acid dose, predicted by use of second-degree polynome.
of five (80%) of the used mice died (Fig. 2). This led to a reduced dose of 200 mg/kg for the second level. At this level three of seven (42.9%) died and the doses was increased to 250 mg/kg. The third level resulted in six of seven (85.7%) dead mice and the dose was again reduced to 225 mg/kg. This gave a death prevalence of 77.8% (seven of nine mice). By direct interpolation in the response surface design, using the results obtained for the two levels closest to LD50, the LD50 was estimated to 206 mg/kg (Table 3). Also in the OA arm, the prevalence of death was best
ARTICLE IN PRESS 6
T. Aune et al. / Toxicon 49 (2007) 1–7
Fig. 4. The PP2A activity rate for (A) DTX-2 and (B) Okadaic acid (OA) expressed as a second-degree polynome of ng/mL toxin with shadowed 95% confidence boundary.
fitted by a second degree polynome in which the dose explains 90.2% of the variation (Fig. 3B). By use of this regression model, the LD50 of OA was estimated to be 204 mg/kg (Table 3). The percent deviation between LD50 estimated by the design and the regression model was found less then 1%. The PP2A assay was performed four times for DTX-2 and 6 times for OA. For both toxins, the protein phosphatase activity rate was best expressed by a second-degree polynome of the dose (Fig. 4A and B). In this study, 97.4% and 96.5% of the variation in the inhibition rate for DTX-2 and OA, respectively was explained by the dose. The IC50 was estimated to 5.94 ng/ml for DTX-2 (Fig. 4A) and 2.81 ng/ml for OA (Fig. 4B). 4. Discussion Results from the mouse bioassays carried out in this study clearly indicate that DTX-2 is less potent
compared with OA when injected intraperitoneally in mice. The relative i.p. toxicity of DTX-2 is about 0.6 that of OA. When calculating total toxicity of OA group toxins based on results from LC-MS, it is suggested that analytical results of DTX-2 should be multiplied by the relative toxicity factor of 0.6. Ideally, toxicity data from experimental animals should be based on results from exposure via the oral route, since the results are used to protect human consumers. However, scarcity of pure toxins has made it necessary to apply data from i.p. injections up till now. Similar practice seems to have functioned satisfactorily for other toxin groups such as the saxitoxins, where toxicity equivalence factors (TEFs) according to Professor Oshima are accepted and used internationally (Anonymous, 2005). The results from the PP2A assay, support the data from the MBA since the IC50 of DTX-2 is about twice the value for OA. This fits well with the internationally accepted consensus that inhibition of protein phosphatase 2A constitutes the main mechanism of toxicity of the OA toxin group. Interestingly, Pere´z-Go´mez et al. (2004) found DTX-2 to be less potent than OA when comparing their neurotoxic effect on cultured cerebellar neurons from rats. In all laboratory animal studies it is of utmost importance to reduce the number of animals to an absolute minimum without reducing the needed information. The only way to obtain this is by optimization of the study design. A common approach in dose finding studies is first to establish a dose window and then divide this into equal distanced doses. An equal number of animals will then be used on these dose levels and the data obtained analysed by a regression model. Usually, a lot of the animals included in such a study approach will give none or limited information. This problem might be overcome by selecting doses closer to the actual level. In most cases, prior knowledge of such an actual dose is limited. However, by using a response surface approach, this can be achieved (Myers and Montgomery, 2002). It has been previously shown that by changing from the common approach as described above to a response surface method, the sample size was reduced from 196 to 64 animals (Ryeng et al., 2001). In the present study, a four level pathway design has been used. By increasing the number of levels in the design, the information will obviously increase rapidly, but so will the number of mice to be included. In order to keep the number of animals to
ARTICLE IN PRESS T. Aune et al. / Toxicon 49 (2007) 1–7
an absolute minimum, only four levels were used. The amount of information will increase considerably by including one dose level more and if possible this is to be recommended. In such a situation at least 9 more mice had to be included in each arm. In the present study this would have resulted in 9 mice at 337.5 mg/kg DTX-2 and 9 mice on 212.5 mg/kg OA. In both cases, this would have been very close to the estimated LD50. Current work at the Marine Institute investigates the possibility of producing a certified standard for DTX-2, as well as a certified mussel tissue reference material containing DTX-2, in collaboration with the National Research Council Canada. Acknowledgements Financial support from the Norwegian Research Council, Project No. 172561/S40, is very much appreciated. Technical assistance from Rolf Harestad on PP2A assay and Brit Heidenreich on the mouse bioassay is also very much appreciated. Financial support was obtained from the Irish National Development Plan and the Marine Institute Ireland, through the ASTOX-project (ST-0202). We gratefully acknowledge critical comments to the manuscript by Dr. Terry McMahon. References Anonymous (2005). Report of the Joint FAO/IOC/WHO ad hoc Expert Consultation on Biotoxins in Molluscan Bivalves (Oslo, Norway, 26-30 September 2004), 31 pp. Rome, Italy: Food and Agriculture Organization. Website http:// www.fao.org/es/esn/food/risk_biotoxin_en.stm. Agresti, A., 2002. Categorical Data Analysis, 2nd edition. Wiley, New Jersey. Carmody, E.P., James, K.J., Kelly, S.S., Thomas, K., 1995. Complex diarrhetic shellfish toxin profiles in Irish mussels. In: Lassus, P., Arzul, G., Erard, E., Gentien, P., Marcaillou, C. (Eds.), Harmful Marine Algal Blooms. Lavoisier Science Publishers, Paris, pp. 273–278.
7
Carmody, E.P., James, K.J., Kelly, S.S., 1996. Dinophysistoxin2: The predominant diarrhoetic shellfish toxin in Ireland. Toxicon 34 (3), 351–359. Gago-Martı´ nez, A., Rodrı´ guez-Va´zquez, J.A., Tibault, P., Quilliam, M.A., 1996. Simultaneous occurrence of diarrhetic and paralytic shellfish poisoning toxins in Spanish mussels in 1993. Nat. Toxins 4 (2), 72–79. Hess, P., Grune, B., Anderson, D.B., et al., 2006. Three Rs approaches in marine biotoxin testing. ATLA 34, 193–224. Hu, T., Doyle, J., Jackson, D., Marr, J., Nixon, E., Pleasance, S., Quilliam, M.A., Walter, J.A., Wright, J.L.C., 1992. Isolation of a new diarrhetic shellfish poison from Irish mussels. J.Chem.Soc.Chem.Commun,, 39–41. Kleinbaum, D.G., Kupper, L.L., Muller, K.E., Nizam, A., 1998. Applied Regression Analysis and other Multivariable Methods. Duxbury Press, Pacific Grove. Kumagi, M., Yanagi, T., Murata, M., Yasumoto, T., Kat, M., Lassus, P., Rodriguez-Vazquez, J.A., 1986. Okadaic acid as the causative toxin of diarrhetic shellfish poisoning in Europe. Agic. Biol. Chem. 50, 2853–2857. McMahon, T., Nixon, E., Silke, J., Taffe, B., Nolan, A., McGovern, E., Doyle, J., 1996. Seasonal variation in diarrhetic shellfish toxins in mussels from the southwest coast of Ireland in 1994. In: Keegan, B.F., O’Connor, R. (Eds.), Irish Marine Science 1995. Galway University Press, Galway, pp. 417–432. Myers, R.H., Montgomery, D.C., 2002. Response Surface Methodology: Process and Product Optimisation Using Designed Experiments, 2nd Edition. Wiley, New Jersey. Pere´z-Go´mez, A., Garcia-Rodrı´ guez, A., James, K.J., ferreroGutierrez, A., Novelli, A., Fernandez-Sa´nchez, M.T., 2004. The marine toxin dinophysistoxin-2 induces differential apoptotic death of rat cerebellar neurons and astrocytes. Toxicological Sciences 80, 74–82. Ryeng, K.A., Arnemo, J.M., Larsen, S., 2001. Determination of optimal immobilizing doses of medetomidine hydrochloride and ketamine hydrochloride in captive reindeer. AJVR 62, 119–126. Vale, P., 2004. Differential dynamics of dinophysistoxins and pectenotoxins between blue mussels and common cockle: a phenomenon originating from the complex toxin profile of Dinophysis acuta. Toxicon 44 (2), 123–134. Vale, P., Sampayo, M.A.deM., 2002. Esterification of DSP toxins by Portuguese bivalves from the Northwestern coast determined by LC-MS—a widespread phenomenon. Toxicon 40, 33–42. Yasumoto, T., Oshima, Y., Yamaguchi, M., 1978. Occurrence of a new type of shellfish poisoning in the Tohuku District. Bull. Jap. Soc. Sci. Fish. 44, 1249–1255.