Registration Dossier

Ecotoxicological information

Toxicity to aquatic algae and cyanobacteria

Currently viewing:

Administrative data

Link to relevant study record(s)

Reference
Endpoint:
toxicity to aquatic algae and cyanobacteria
Type of information:
experimental study
Adequacy of study:
key study
Study period:
from 1980-08-21 to 1980-09-04
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
test procedure in accordance with national standard methods with acceptable restrictions
Remarks:
The study was conducted according to an appropriate national standard method (US EPA 1979), with acceptable restrictions. The study was not compliant with GLP, no analytical monitoring was carried out and no solvent control was adopted in the study.
Qualifier:
according to guideline
Guideline:
other: Miller et al. 1978, EPA-600/9-78-018
Deviations:
not specified
GLP compliance:
no
Analytical monitoring:
no
Vehicle:
no
Details on test solutions:
PREPARATION AND APPLICATION OF TEST SOLUTION

- Method: exposure concentrations were prepared from an unfiltered stock solution of 101 mg active acid/L

- Controls:

- Chemical name of vehicle (organic solvent, emulsifier or dispersant): AAM

- Concentration of vehicle in test medium (stock solution and final test solution(s) including control(s)): 105 µg/L
Test organisms (species):
Pseudokirchneriella subcapitata (previous names: Raphidocelis subcapitata, Selenastrum capricornutum)
Details on test organisms:
TEST ORGANISM

- Strain: EPA-Corvallis

- Source (laboratory, culture collection): EPA-Corvallis, original source; since maintained at SRI
Test type:
static
Water media type:
freshwater
Limit test:
no
Total exposure duration:
14 d
Nominal and measured concentrations:
Nominal concentrations: 0, 0.57, 1.02, 5.70, 10.19, 57.04 and 101.85 mg active acid/L
Details on test conditions:
TEST SYSTEM

- Test vessel: flasks

- Size, fill volume: 100 ml diluent in 500 ml flasks

- Aeration: none reported

- Initial cells density: 10 000 cells/ml

- Control end cells density:

- No. of vessels per concentration (replicates): 3

- No. of vessels per control (replicates): 4


GROWTH MEDIUM

- Standard medium used: yes, as described by Miller et al. 1978, in EPA-600/9-78-018


TEST MEDIUM / WATER PARAMETERS
- Source/preparation of dilution water: not reported

- Intervals of water quality measurement: not reported



OTHER TEST CONDITIONS

- Sterile test conditions: not reported

- Adjustment of pH: not reported

- Photoperiod: continuous

- Light intensity: 400 ft candles



EFFECT PARAMETERS MEASURED:
- Determination of cell concentrations: cell counts were taken at 96 hours and every other data (except weekends) thereafter with an electronic particle counter.


TEST CONCENTRATIONS

- Spacing factor for test concentrations: 2 and 5

- Range finding study: none
Reference substance (positive control):
no
Duration:
96 h
Dose descriptor:
NOEC
Effect conc.:
10.19 mg/L
Nominal / measured:
meas. (initial)
Conc. based on:
act. ingr.
Remarks:
active acid
Basis for effect:
growth rate
Remarks on result:
other: calculated by study reviewer
Duration:
96 h
Dose descriptor:
LOEC
Effect conc.:
57.04 mg/L
Nominal / measured:
meas. (initial)
Conc. based on:
act. ingr.
Remarks:
active acid
Basis for effect:
growth rate
Remarks on result:
other: calculated by study reviewer
Duration:
96 h
Dose descriptor:
EC50
Effect conc.:
28.13 mg/L
Nominal / measured:
meas. (initial)
Conc. based on:
act. ingr.
Remarks:
active acid
Basis for effect:
cell number
Duration:
14 d
Dose descriptor:
EC50
Effect conc.:
27 mg/L
Nominal / measured:
meas. (initial)
Conc. based on:
act. ingr.
Remarks:
active acid
Basis for effect:
cell number
Details on results:
- Any stimulation of growth found in any treatment: Analysis of variance showed that at 96 hours growth was stimulated at 10 mg/l, and at 14 days growth was stimulated at 0.57-10 mg/l.
Reported statistics and error estimates:
ANOVA for EC50 values reported in study report. Statistical analysis was conducted by the reviewer using the method described in the OECD test guidance 201, based on growth rate over 96 hours to calculate the NOEC and LOEC. These values were calculated in R Studio using a Levene's test for homogeneity, an ANOVA and Tukey's post hoc test.

Table 1. Summary of findings

Nominal concentrations (mg/l)  Mean cells/ml
 96 h  14 d
 Control  3.25 x 10^6 4.65 x 10^6
 0.57  4.20 x 10^6 6.54 x 10^6
 1.02  3.33 x 10^5 6.21 x 10^6
 5.70  3.89 x 10^6 6.20 x 10^6
 10.19  5.0 x 10^6* 7.10 x 10^6*
 57.04  15 x 10^3* 15 x 10^3*
 101.85  21 x 10^3* 76 x 10^3*

* p ≤ 0.05 vs. zero concentration

 
Validity criteria fulfilled:
no
Remarks:
solvent control not present
Conclusions:
A 96 h EC50 value of 28 mg active acid/L and a 14 d EC50 value of 27 mg active acid/L have been determined for the effects of the test substance on cell numbers of the freshwater algae Selenastrum capricornutum (new name: Pseudokirchnerella subcapitata). A 96 hour NOEC value of 10.19 mg active acid/L and a LOEC of 57.04 mg active acid/L were calculated by the reviewer for the effect of the test substance on the growth rate of Selenastrum capricornutum (new name: Pseudokirchnerella subcapitata). The study was not conducted according to GLP but analytical monitoring was carried out. It is possible that the outcome of the test was influenced by the nutrient complexing properties of HMDTMP. In addition, no solvent control was used during the test and the full name of the solvent control is not reported.

Description of key information

96 hour NOEC 10 mg active acid/L, Pseudokirchneriella subcapitata, reliability 2 (Monsanto, 1980, read-across from HMDTMP acid)

Due to the essential nutrients present in the test medium being complexed by the phosphonates, it is likely that the test organisms were exposed to phosphonate-metal complexes. The effects seen in the studies may be a result of nutrient complexation rather than a reflection of the true toxicity of the test substance.

Key value for chemical safety assessment

EC50 for freshwater algae:
28 mg/L
EC10 or NOEC for freshwater algae:
10 mg/L

Additional information

A 96 hour NOEC value of 10 mg active acid/L and a LOEC of 57 mg active acid/L were calculated by the reviewer for the effect of HMDTMP-H on the growth rate of Selenastrum capricornutum (new name: Pseudokirchnerella subcapitata). A 96-h and 14-d EC50value of 28 and 27 mg active acid/L respectively, were also reported in the same study for the effects of HMDTMP-H on the cell number of Selenastrum capricornutum (new name: Pseudokirchneriella subcapitata), based on initial mean measured concentrations.

The acid, sodium and potassium salts in the HMDTMP category are freely soluble in water. The HMDTMP anion can be considered fully dissociated from its sodium or potassium cations when in dilute solution. Under any given conditions, the degree of ionisation of the HMDTMP species is determined by the pH of the solution. At a specific pH, the degree of ionisation is the same regardless of whether the starting material was HMDTMP-H, HMDTMP.4Na, HMDTMP.7K or another salt of HMDTMP.

Therefore, when a salt of HMDTMP is introduced into test media or the environment, the following is present (separately):

1. HMDTMP is present as HMDTMP-H or one of its ionised forms. The degree of ionisation depends upon the pH of the media and not whether HMDTMP (4-7K) salt, HMDTMP (4-7Na) salt, HMDTMP-H (acid form), or another salt was used for dosing. At pH 5.5 - 6, the HMDTMP anions would be present on average as the HMDTMP trivalent anion according to the pH curves.  At neutral pH (7), the HMDTMP anions would be present on average as the HMDTMP pentavalent anion according to the pH curves. At pH 8, the HMDTMP anions would be present on average as the HMDTMP hexavalent anion according to the pH curves.

2. Disassociated potassium or sodium cations. The amount of potassium or sodium present depends on which salt was added.

3. It should also be noted that divalent and trivalent cations would preferentially replace the sodium or potassium ions. These would include calcium (Ca2+), magnesium (Mg2+) and iron (Fe3+). These cations are more strongly bound by HMDTMP than potassium and sodium. This could result in HMDTMP-dication (e.g. HMDTMP-Ca, HMDTMP-Mg) and HMDTMP-trication (e.g. HMDTMP-Fe) complexes being present in solution.

It is a functional property of phosphonate substances that they form stable complexes (ligands) with metal ions. In algal toxicity tests essential nutrients will thus be bound to the phosphonates according to the 'ligand binding model' [1] (Girling et al. 2018). In algal growth medium some metals form strongly-bound complexes and others form weakly-bound ones (Girling et al. 2018). The phosphonates possess multiple metal-binding capacities, and pH will affect the number of binding sites by altering the ionisation state of the substance. However, the phosphonate ionisation is extensive regardless of the presence of metals (Girling et al. 2018).

The phosphonate-metal complexes may be very stable due to the formation of ring structures ("chelation"). This behaviour ensures that the phosphonic acids effectively bind and hold the metals in solution and renders them biologically less available As a result when a trace metal is complexed, its bioavailability is likely to be negligible (Girling et al. 2018, SIAR 2005). However, there is no evidence of severe toxicity from metal complexes of the ligands (Girling et al. 2018).

In algal growth inhibition tests, complexation of essential trace nutrients (including Fe, Cu, Co, and Zn) by phosphonate substances can lead to inhibition of cell reproduction and growth. Guidelines for toxicity tests with algae do not typically describe procedures for mitigating against this behaviour. For example the standard OECD Guideline 201, describing the algal growth inhibition test, only specifies that the “chelator content” should be below 1 mmol/l in order to maintain acceptable micronutrient concentrations in the test medium (SIAR 2005).

OECD guidance on the testing of difficult substances and mixtures (OECD, 2000) does include an annex describing “toxicity mitigation testing with algae for chemicals which form complexes with and/or chelate polyvalent metals”. The procedure is designed to determine whether it is the toxicity of the substance or the secondary effects of complexation that is responsible for any observed inhibition of growth. It involves testing the substance in its standard form and as its calcium salt in both standard algal growth medium and in medium with elevated CaCO3hardness. Calcium is non-toxic to aquatic organisms and does not therefore influence the result of the test other than by competitively inhibiting the complexation of nutrients (SIAR 2005). By increasing the calcium content it may be that the nutrient metals are released from their complexed form although this may not always apply. The outcome of the test however only determines whether nutrient complexation is the cause of apparent toxicity and does not determine the inherent toxicity of the test substance for the reasons explained by the Ligand binding model (Girling et al. 2018).

The magnitude of the stability constants depends on the properties of the metal and also of the ligand, in respect of the type of bonding, the three dimensional shape of the complexing molecule, and the number of complexing groups. The SIAR provides two tables of stability constants (effectively the strength of the complexation), one from Lacour et al. (1999) and one from Gledhill and Feijtel (1992). The Gledhill and Feijtel constants show a range of values for important divalent metal ions, cited as having been obtained from Monsanto internal reports (Owens, 1980). They show that ATMP, HEDP, DTPMP, EDTMP and HMDTMP are strong complexing agents, with stability constant values ranging from 5 to 24 (log10values). ATMP and DTPMP are analogues of HMDTMP, therefore the data available for these substances are used as surrogates for HMDTMP. Please refer to IUCLID Section 13 and Annex 6 of the CSR for further information on the read-across between ATMP and DTPMP to HMDTMP.

The complexation constant for phosphonates with iron (III) has been estimated by TNO (1996a) to be around log K = 25 (Girling et al. 2018).

Calculations based on the known stability constants show that, even where the OECD-recommended approach to add additional calcium to the test media is used, that the key nutrients would still be complexed by the phosphonates in preference to complexation of calcium and magnesium. Therefore the calcium complex (most representative of the environmental species) can never be maintained in the test medium in the presence of other key nutrient ions such as Co, Zn, Mn and Fe (Girling et al. 2018). As a result, the complexed nutrients will almost certainly not be bioavailable to aquatic plants and this can lead to inhibition of algal growth. Growth inhibition via this mechanism is a secondary effect and does not reflect the inherent toxicity of the test substance (Girling et al. 2018).

The available evidence suggests that toxic effects observed in the tests are a consequence of complexation of essential nutrients and not of true toxicity (SIAR 2005). A study designed to ensure adequate levels of bioavailable nutrients with either of the phosphonates would result in the test substance being a phosphonates-Fe complex. Under conditions where iron is readily available to counteract the effects of nutrient complexation it is unlikely that the substance would have a negative effect on algal growth (Girling et al. 2010). The nutrient complexing behaviour of phosphonate substances therefore renders testing to determine their intrinsic toxicity to algae impractical.

A detailed interpretation of the effects of nutrient complexation, and photolytic release of phosphorus from, phosphonic acids on algal growth in toxicity studies is given in Annex V to the phosphonic acid SIARs (2005).The principal and somewhat contrasting conclusions of the review are that:

   

1) Algal growth may be stimulated by the presence of supplementary phosphorous released by the photolytic degradation of phosphonic acids .

2) Algal growth may be inhibited by the complexation of micronutrients (trace metals) by phosphonic acids. This inhibition is an algistatic rather than algicidal effect. Under the standard test conditions used for most studies, the trace metals will be fully and strongly bound to the HMDTMP, with the strong possibility that their bioavailability will have been reduced considerably.

These two phenomena can occur at different stages during the same algal test and at different exposure levels of the substance.

Complexing agents, such as HMDTMP, inhibit algal growth because of their capacity to limit the bioavailability of trace metal micronutrients that are essential for growth. This has been illustrated by the following studies.

Hanstveit and Oldersma (1996) have conclusively demonstrated the importance of taking chelation/complexation into account in tests with another phosphonic acid, DTPMP. They have shown that DTPMP exhibits apparent toxicity (95-h ErC50 = 0.45 mg/l) to Selenastrum capricornutum in growth inhibition tests. These tests were carried out using standard OECD growth medium containing concentrations of Cu, Co and Zn that had been increased above the guideline concentration (Cu: up to 30 times, Co: up to 30x and Zn: up to 300x) in line with their predicted speciation. However, when the test medium was also supplemented with Fe at up to 300x the guideline concentration, no growth inhibition was observed at the highest test concentration of 10 mg/l. The explanation given for the absence of toxicity was that the addition of the Fe ensured that the free ion concentrations of all four of these essential nutrients were now in accordance with those specified in the standard OECD medium as being necessary for healthy algal growth. The experimental concept was for the iron to preferentially bind the DTPMP. The key role of Fe in determining the free concentration of the other elements was based on speciation calculations and an estimated value of the DTPMP-iron stability constant.

Similar findings have been reported by Schowanek et al (1996) from algal toxicity tests carried out with the unrelated chelating substance [S,S]-ethylene diamine disuccinate ([S,S]-EDDS). Chlorella vulgaris was tested according OECD 201, water hardness and trace metal concentrations were varied. In standard media with different water hardness (24-375 mg/l CaCO3), addition of 1 mg/l [S,S]-EDDS reduced the growth rate by 53%, independent on the water hardness. Speciation calculations showed that in the standard medium [S,S]-EDDS is mainly associated with Zn, Cu, and Co. To test the hypothesis that the apparent toxicity was caused by nutrient deficiency, growth experiments in metal-enriched medium were performed. With increasing concentrations of Zn, Co, and Cu, the algal growth increased, reaching a maximum and then falling. The maximum growth was obtained with 1 mg/l (= 3.4 µM) [S,S]-EDDS, 0.62 µM Co, 0.051 µM Cu, and 2.9 µM Zn, where the levels of free Cu, Co and Zn were the same as in standard medium without the chelator. With lower [S,S]-EDDS concentrations, growth is decreased, mainly caused by Zn toxicity.

The interpretation of these data is also consistent with findings presented in the risk assessment being carried out for the chelating agent EDTA (in draft), which is actually a weaker complexing agent than HMDTMP. It has been demonstrated that for EDTA it is not the absolute concentration, but rather the ratio of the EDTA concentration to that of the metal cations that is crucial to determining algal growth under the conditions of a toxicity test (EC, 2003).

The ability of iron to catalyse photodegradation of HMDTMP means that the interpretation of algal growth data can be somewhat uncertain; this applies to the complexing agents discussed above including EDTA. However, limitation of micronutrient availability is considered to be a sufficiently generic phenomenon to explain effects observed in toxicity tests with substances that have the capacity to chelate cationic metals.

Conclusions: Great care has to be exercised in interpreting the results of the algal tests carried out with phosphonic acids. The significant potential for nutrient complexation by HMDTMP and/or release of phosphorous from degradation of HMDTMP to respectively either inhibit or stimulate algal growth makes definitive interpretation difficult. However the available evidence suggests that toxic effects observed in tests with structurally analogous substances are a consequence of complexation of essential nutrients and not of true toxicity. These effects do not obey a classic dose response and as such extrapolation using an assessment factor is inappropriate. In addition, similar effects would not be anticipated in natural environmental waters.

Please see the attached position paper which further discusses algal tests with phosphonate substances and presents arguments against further algal testing.


[1] 'Ligand’ is a general term used to describe a molecule that bonds to a metal; in the present case the phosphonate can form several bonds and the resultant chelated complex can be a very stable entity. It is possible that two molecules could bind to the individual metal, or that one molecule could bind two metals. In dilute solution a 1:1 interaction is the most probable. To simplify discussion, the ligand is considered to be able to form a strongly-bound complex with some metals, and a more weakly-bound complex with others.