The Acute Toxicity of Thulium to Hyalella Azteca and the Inuence of Toxicity Modifying Factors.

The demand for rare earth elements (REEs) is growing and as a result, environmental exposure is a concern. The objective of this research was to evaluate the acute toxicity of Tm to Hyalella azteca and to understand the potential for toxicity modication by dissolved organic matter (DOM) and the cations Ca 2+ , Mg 2+ and Na + . Standard methods were followed for 96 h static exposures in a medium with a hardness of 60 mg CaCO 3 /L, pH of 7.3 at 23°C. H azteca neonates (2-9 d of age) were used and in unmodied media the LC50 concentration was 3.4 µM (95% CI 2.9-3.9 µM; 573 µg/L (482-663)) based on measured dissolved concentrations at the end of the test. Tests done with different concentrations of Ca (0.25, 0.5 and 1.5 mM) did not show consistent trends and there was no clear evidence of a protective effect from Ca. Variations in Na (0.26, 0.5 and 1.6 mM) resulted in no signicant changes in toxicity. Similarly, Mg (0.07, 0.14 and 0.4 mM) did not result in signicant changes in LC50 values, except for a reduction in toxicity for measured total Tm at the lowest Mg concentration. Our results indicate that Tm toxicity is not inuence by cationic competition (Ca, Na and Mg). Dissolved organic matter (sourced from Luther Marsh ON) offered signicant protection against Tm toxicity. Additions over 3 mg DOC/L resulted in signicantly increased LC50 values. This study contributes toward understanding the toxicity of Tm and the importance of considering dissolved organic matter in estimating the potential for environmental risk of REEs.


Introduction
Global demand for REEs has increased dramatically in recent years and Canada is home to signi cant deposits potentially making it a leading global producer (Humphries 2013;Yin et al. 2021). There is limited information about the potential environmental impacts of REEs in aquatic systems (Gwenzi et al. 2018) and this is particularly the case for thulium (Tm). In a comparison of acute toxicities, Borgmann et al. (2005) found Tm to have the lowest LC50 value to Hyalella azteca with a measured dissolved concentration of 0.01 µg/L (0.00059 µM) in very soft water (12 mg CaCO 3 /L). This dissolved concentration was associated with a nominal exposure concentration of 721 µg/L (4.3 µM) indicating that the majority of the Tm in test solutions had precipitated (Borgmann et al. 2005). In spite of the fact that Tm appears to have a higher toxicity compared to the other REEs tested by Borgmann et al. (2005) there have been few studies (if any) on the aquatic toxicity of this heavy REE.
It is well known that the aquatic toxicity of inorganic forms of metals can be in uenced by its geochemical speciation. For many of the well studied metals, acute toxicity results from the uptake of free metal ions into the organism and the resulting disruption of essential ion balance (Niyogi and Wood, 2004; Mebane et al. 2020). Toxicity modifying factors (TMFs) for metals are grouped as either cations that compete for uptake, or negatively charged ligands that complex free ions and thereby reduce bioavailability (Di ; Santore et al. 2001). Toxicity reduction through cationic competition occurs because uptake of the toxic free metal cation occurs via mechanisms for uptake of essential ions, particularly Ca 2+ , Mg 2+ or Na + , and therefore is dependent on concentrations of the latter. This is particularly the case for monovalent and divalent metals such as Cu 2+ , Zn 2+ , Pb 2+ , Ag + and Co 2+ (Niyogi and Wood, 2004). Anions such as HCO 3 and Clalong with negatively charged moieties within dissolved organic matter (DOM) form complexes with free metal ions and thereby reduced the concentration of the most toxic form of the metal without changing the overall total concentration in solution (Mebane et al. 2020). DOM is recognized as having an important role in the mitigation of metal toxicity (Wood et al. 2011). It is ubiquitous in natural aquatic systems where it arises from both autochthonous and terrigenous inputs with the latter representing a key input to food webs (Tanentzap et al. 2014 (2015) used the same species and tested the uptake of Tm in the presences of known complexing ligands, citric acid, malic acid, nitrilotriacetic acid. Internalization of Tm into algae was found to be correlated with Tm 3+ concentrations but there was also uptake of ligand bound Tm (Zhao and Wilkinson 2015). The authors were able to rule out direct uptake of Tm complexes and also uptake via anion channels at the algal surface and concluded that our current understanding metal bioavailability is insu cient in terms of describing the internalization of Tm.

Tm Water Chemistry and Characterization
In the trial to characterize Tm concentrations in RM only (H. azteca not exposed), the Tm -Rec concentrations were generally less than the planned nominal concentrations (  Tm acute toxicity in RM H. azteca mortality increased with increasing concentrations of Tm (Table 2). Based on the results from the bench test, we expected measured Tm concentrations at test beginning and at test end to be relatively similar however, this was not always the case (Table 2). An equilibration time of 24 h may not have been su cient and it is unknown if the addition of organisms in uenced the geochemistry of test solutions.
Calculations of the standard acute toxicity endpoint (96 h LC50) were done with measured concentrations and this was possible for both Tm -T or Tm -D and either at the beginning of the test or at the end ( Table 2). In general, the measured concentrations were lowest at the end of the tests and therefore, as a conservative approach, calculations were based on samples collected at 96 h.

In uence of Cations on Tm toxicity
The results for effects of Ca on Tm toxicity (Fig. 1) were somewhat di cult to interpret. On the basis of Tm -T concentrations the lowest LC50 was at the 0.5 mM Ca treatment and toxicity was signi cantly reduced at both lower and higher concentrations (0.25 and 1.5 respectively). However, for Tm -D there was no trend evident for the low Ca exposure because the LC50 values at 0.25 and 0.5 Ca were not signi cantly different. At the higher Ca treatment Tm -D showed higher toxicity (Fig 1). There was no protective effect with increasing Na concentrations (Fig. 2). Similarly, we did not see a protective effect with Mg as there were no signi cant differences in LC50 values across the range of Mg tested (Fig. 3).
We hypothesized that increases in Ca would have a protective effect to Tm toxicity but there was no consistent trend across the range of Ca tested (Fig 1). In these studies the changes in toxicity are linked to hardness and so cannot be exclusively attributed to Ca. As discussed above, changes in exposure hardness did not result in changes of Tm toxicity (Borgmann et al 2005) and this is consistent with our results but only nominal LC50 values are available. It may be that Tm uptake and toxicity is not in uenced by competitive interaction with Ca and therefore it is unlike other REEs (e.g. Dy, La, Ce and Y) that are. There was no protective effect with increasing Na concentrations (Fig. 2)  where Mg additions did not show a protective effect on Dy toxicity.
We had hypothesized that Tm toxicity would be in uenced by cationic competition, particularly Ca 2+ . Previous studies on the toxicity of inorganic forms of metals attribute the toxicity reduction achieved by cations to direct competition at the site of uptake of essential ions such as Ca 2+ and Na + (Niyogi and and Pb 2+ inhibit Ca 2+ uptake while monovalent metals such as Ag + disrupting Na + uptake (Niyogi and Wood 2004). From this perspective the mechanism by which a trivalent REE free ion would interact (compete) with an essential divalent (Ca or Mg) or monovalent (Na) cation is not clear. However, studies have highlighted the similar properties of trivalent REEs, particularly in relation to Ca 2+ (Evans 1983) and it is well known that La 3+ is an effective analogue. There is some evidence of direct competition between In uence of DOM on Tm toxicity In solutions with added Luther Marsh DOM toxicity was signi cantly reduced above concentrations of 3 mg DOC/L (Fig 4). The addition of DOM also altered the relative concentrations of Tm -T and Tm -D and it appeared that the precipitation threshold may have been increased with elevated DOC content (Table 2). DOM has been shown to reduce the toxicity of numerous metals in a concentration dependent manner (Wood et al. 2011). It is a complex heterogeneous molecule with a variety of negatively charged moieties that are capable of interacting with cationic metals. Complexation of the free ion form of the metal reduces the availability for uptake thereby reducing toxicity. This was evident in our study for Tm (Fig 4) and we assumed that mitigation of toxicity was due to reduced Tm 3+ concentrations. However, this is only an assumption as the bioavailable forms of Tm associated with toxicity and the mechanism of uptake at the biotic surface are unknown. The predicted Tm 3+ concentrations in test solutions with added DOM were at least 140 fold lower (highest Tm -D with lowest DOC) and ranged up to 2.5x10 6 fold lower (lowest Tm -D with highest DOC) than the corresponding Tm 3+ concentrations in solutions with no added DOM. Clearly the predicted Tm 3+ estimates were not linked to the acute toxicity of Tm. One possible conclusion is that Tm 3+ is not associated with toxicity in Hyalella and that other (or additional) geochemical forms are. It is also possible that WHAM is predicting a much higher level of complexation of Tm 3+ than is actually occurring in our test solutions. Either way, DOM signi cantly reduces Tm toxicity and an improved understanding of the geochemical speciation of Tm in relation to acute toxicity is required.
There are relatively few studies on the effects of REEs on aquatic biota and even fewer investigating the potential in uence of DOM on toxicity. DOM has been highlighted as an important factor to include in water quality derivations for La (Hermann et al 2016). Vukov et al (2016) used Suwannee River DOM to show 3-4 fold reductions of Dy toxicity to Hyalella at a DOC concentration of 13 mg/L. The biouptake of Sm 3+ (as measured directly by ion exchange technique) into the unicellular green algae Chlamydomonas reinhardtii was signi cantly reduced by DOM in a DOC concentration dependent manner (Rowell et al 2018). In that study four different sources of DOM were tested, including Luther Marsh DOM, and even very small additions of 0.5 mg DOC/L dramatically reduced uptake by 10 fold (Rowell et al 2018). In tests with the synthetic organic ligands malic acid, diglycolic acid and citric acid, Sm uptake to C. reinhardtii was reduced but the possibility of complexed Sm also being taken up could not be ruled out (Tan et al 2016). Similar reduced uptake results into Chlorella vulgarize were shown for La, Gd and Y using the organic ligands citrate, nitriloacetic acid and ethylenediamine tetraacetic acid (Sun et al 1997). In natural waters the important role that DOM has in complexing REEs is well recognized (

Conclusions
This study provided data on the role of water chemistry in the toxicity of Tm to aquatic invertebrates. As observed in other studies, the formation of insoluble species can predominate at elevated concentrations. While we equilibrated our test solutions for 24 h before beginning tests, this may not have been su cient, and we observed Tm -T and Tm -D changes between the beginning and end of tests. LC50 determinations based on measured concentrations at the end of the test provided a conservative approach to characterizing effects. In tests with different concentrations of Ca or Na or Mg we found no consistent toxicity modi cation and conclude that cationic competition does not in uence Tm toxicity. This would appear to be unlike some of the other REEs where cations, or more commonly, water hardness has been shown to reduce toxicity. Acute toxicity was in uenced by DOM and concentrations above 3 mg DOC/L resulted in signi cantly increased 96 h LC50 values. In estimating the potential of Tm effects in natural water it appears that complexation is an important consideration.

Declarations
Funding -see Acknowledgements Con icts of interest/Competing interests -The authors have no con icting/competing nancial interests nor personal relationships that in uence the work reported in this paper.
Availability of data and material -Data is available from the corresponding author  The 96 h LC50 values (with upper 95% con dence intervals) for Hyalella azteca exposure to Tm at different Na concentrations. The LC50 values are based on measured total (Tm-T black bars) and measured dissolved (Tm-D grey bars) concentrations at the end of the test. There were no signi cant differences for either Tm-T or Tm-D compared to the corresponding acute test in RM at 0.5 mM Na.  The 96 h LC50 values (with upper 95% con dence intervals) for Hyalella azteca exposed to Tm with

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Tables.pdf