Effects of Low, Subchronic Exposure of 2,4-Dichlorophenoxyacetic Acid (2,4-d) and Commercial 2,4-d Formulations on Early Life Stages of Fathead Minnows (Pimephales promelas).

Abstract

The 2005 US Environmental Protection Agency (USEPA; 2005) 2,4-dichlorophenoxyacetic acid (2,4-D) reregistration eligibility decision explicitly stated that there was limited literature on the effects of low-level chronic 2,4-D exposure to fish. Motivated by this deficiency and a dearth in the peer-reviewed literature on impacts of 2,4-D formulations on fish, DeQuattro and Karasov (2016) performed the first 2,4-D studies from our laboratory using a flow-through exposure system built similarly to that used at the USEPA laboratory in Duluth (MN, USA) and described in Ankley et al. (2001). DeQuattro and Karasov (2016) found that fathead minnow larvae raised in waters with the 0.045 ± 0.003 ppm 2,4-D commercial herbicide formulation DMA®4IVM had significantly reduced larval survival compared with controls. This observed response took place when larval fish were exposed to 2,4-D concentrations 3 orders of magnitude lower than the documented no-observed-effect concentrations (NOECs) for 2,4-D acid and amine salt forms (14.2–63.4 ppm), 2 orders of magnitude below permitted levels for lakes (4 ppm for spot treatments; 2 ppm for whole-lake treatments), and within the range of actual 2,4-D concentrations used to treat invasive aquatic plants in Wisconsin (USA; Figure 1; US Environmental Protection Agency 2005; Nault et al. 2014, 2018). Given these findings, it was scientifically reasonable to attempt to replicate this finding and then extend our understanding of decreases in larval survival by identifying a critical window of exposure for fathead minnow larvae with the overarching goal being the identification of a mechanism of action. These were the goals of Dehnert et al. (2018), which became the subject of a recent letter to the editor of this journal by Giddings and Habig (2019). Giddings and Habig (2019) stated, “Dehnert et al. suggest that their study was the first study of this type (exposure of eggs, hatchlings, and juvenile fish) using 2,4-D,” which was not what was written in our peer-reviewed publication. We were aware of an earlier industry-sourced report referenced by Giddings and Habig (Mayes et al. 1990) that was not published in peer-reviewed literature nor available to the public, and we specifically wrote “To our knowledge, no published studies have tested the impacts of environmentally relevant doses (0.00–2.00 ppm) of pure 2,4-D exposure on survival from fertilization to 14 d post hatch (dph) in any fish species.” However, it is most important to note that in a series of individual experiments, Dehnert et al. (2018) found that fathead minnow larval survival (14 dph) was decreased following exposure of eggs and larvae to either of 2 commercial 2,4-D amine salt formulations, or technical grade 2,4-D acid at 2,4-D concentrations well below the putative NOEC (DeQuattro and Karasov 2016; Dehnert et al. 2018). Based on our replicated findings, we made the cautious suggestion that permitting these herbicide formulations should be re-evaluated. In fact, we suggest that the present range of interest for new studies should be between 0 and 2.00 ppm 2,4-D. Giddings and Habig (2019) took issue with several features of water chemistry and modifications to the early life stage exposure assay in our experiments. In fact, total hardness, measured using a titration method of 20 water samples randomly chosen from all experiments, averaged 238 ± 8 mg CaCO3/L (n = 10, 0.00 ppm 2,4-D; n = 10, 2.00 ppm 2,4-D; US Environmental Protection Agency 1982), which is within range of concentrations found in Wisconsin lakes (0–370 ppm; National Water Quality Monitoring Council 2016) and within range of USEPA guidelines (0–250 ppm; US Environmental Protection Agency 2016a, 2016b). Low aeration was used in the exposure tanks to ensure a percentage saturation of oxygen >90% (Ankley et al. 2001; US Environmental Protection Agency 2016b). A separate trial in conjunction with experiment 2, but with no aeration (oxygen saturation still >90%), showed the same negative impacts on larval survival of fathead minnows after exposure to DMA4 from fertilization to 14 dph. We used a flow rate that achieved 24 turnovers/d in 15-L tanks (mistakenly reported as 10 L in Dehnert et al. 2018). The USEPA states that the minimum turnover rate should be 5 turnovers/d, and an increase can be used to increase water quality (US Environmental Protection Agency 2016b). In each of the experiments, we observed >95% survival for control larvae, which exceeded the USEPA's minimum of 70% survival for posthatch fathead minnows (US Environmental Protection Agency 2016b). Contrary to the assertions of Giddings and Habig (2019), all water parameters in Dehnert et al. (2018) are highly appropriate and all modifications made (i.e., aeration, flow rate, concentration groups) to the early life stage exposure assay are scientifically sound and reasonable in allowing us to assess the impacts of ecologically relevant concentrations of 2,4-D and 2,4-D commercial herbicide formulations on fathead minnows. Giddings and Habig (2019) took issue with Dehnert et al.'s (2018) measurement of 2,4-D concentrations in exposure tanks. All 2,4-D concentrations were measured using an enzyme-linked immunosorbent assay (ELISA) and were reported as 2,4-D ppm in Dehnert et al. (2018), which referred to 2,4-D ppm acid equivalence (a.e.; as stated on the product labels of Weedestroy®AM40 and DMA®4IVM). Analyses of random water samples from exposure tanks were carried out independently by the Wisconsin State Laboratory of Hygiene (University of Wisconsin–Madison, Madison, WI, USA). The 2,4-D ELISA is validated in the literature (Hall et al. 1993; Lawruk et al. 1994), is widely used for the detection of 2,4-D (by the Wisconsin State Laboratory of Hygiene and the University of Florida Center for Aquatic and Invasive Plants, Gainseville, FL, USA), and is used to verify permitted applications of 2,4-D in the field (Williams et al. 1997; Chuang et al. 2005; Glomski and Netherland 2010; Vdovenko et al. 2013; Nault et al. 2014, 2018). To use the ELISA, the Wisconsin State Laboratory of Hygiene confirmed that the quality control standard included in every kit was ±5 ppb, and each calibration curve had an R2 value of >0.99. Hall et al. (1993) found that “the correlation between the results from EIA [ELISA] and GC (gas chromatography) analysis yielded a correlation coefficient of 0.92 and an equation of the line y = 0.99x – 0.08” (ppb), and stated that “the regression analysis verifies that EIA is a suitable method for detection of 2,4-D in surface water.” Moreover, Lawruk et al. (1994) found that the correlation between the ELISA and the GC–mass spectrometry methods, using a 2,4-D USEPA standard reference, yielded a correlation coefficient of 0.970 and an equation line of y = 0.938x + 2.21 (ppb; n = 56 samples) and stated that the ELISA is suitable for 2,4-D monitoring in water, soil, food, and solid waste. Giddings and Habig (2019) suggest relatively high variability in measured test concentrations in Dehnert et al. (2018). However, the US Environmental Protection Agency (2016a) states that “the test will not be considered to be unacceptable or invalid solely on the grounds that measured concentrations deviated by more than 20% from nominal concentrations.” Over all 30-d exposure experiments, Dehnert et al. (2018) reported 3 significantly distinct concentration groups that never overlapped and contained appropriate control groups that remained free of 2,4-D (below the detection level of ≤1 ppb 2,4-D; Figure 1). Giddings and Habig (2019) suggest that “2,4-D degrades rapidly in aquatic systems and is not persistent” and therefore could not have putative impacts in ecological situations. Granted, in a laboratory setting and under highly aerobic conditions, 2,4-D undergoes rapid degradation (half-life of 15 d) and is also susceptible to aqueous photolysis (half-life of 12.9 d; US Environmental Protection Agency 2005); however, this is not representative of ecological scenarios in which 2,4-D is used. The USEPA reregistration eligibility decision (US Environmental Protection Agency 2005) states that the measured range of 2,4-D degradation in water is 14 to 333 d. The degradation of 2,4-D in water is highly variable, depending on microbial presence, light, oxygen saturation, nutrient levels, temperature, and pH, and whether the water had been previously contaminated with 2,4-D or other phenoxyacetic acids (Sinton et al. 1986; Howard 1991; US Environmental Protection Agency 2005; Sandoval-Carrasco et al. 2013; Nault et al. 2014, 2018). Even considering the most optimistic degradation of 2,4-D (13-d half-life), native fish species in treated waters could still be exposed to 2,4-D long enough to experience deleterious impacts on larval survival after 14 d of exposure, as indicated in Figures 1a, 3b and c, and 4b of Dehnert et al. (2018). The Wisconsin Department of Natural Resources (Madison, WI, USA) recently measured the half-life of 2,4-D as up to 76 d in permitted 2,4-D whole-lake treatments and spot treatments (Nault et al. 2014; 2018). These studies illustrate that even with a low concentration application and without reapplication, initial whole-lake treatments can result in a prolonged presence of 2,4-D in the water column (Nault et al. 2018). Hence, 2,4-D exposure to lake biota may often exceed 21 d during permitted applications, which encompasses the observed 2,4-D critical window of exposure leading to decreased survival in fathead minnow larvae (Dehnert et al. 2018). In Dehnert et al. (2018), an increase in growth parameters was observed in some 2,4-D treatments, but Giddings and Habig (2019) appear to have missed the point that 2,4-D was not likely to be the cause of the increase in growth. Dehnert et al. (2018) reported that larval fathead minnows exposed to Weedestroy®AM40 and DMA4 (2.00 ppm 2,4-D) had higher total length and mass compared with controls, but that larvae exposed to pure 2,4-D (2.00 ppm) showed no significant difference from controls. Thus, Dehnert et al. (2018) wrote, “the observed results could be due to other compounds in the formulations acting directly or through an interaction between the labeled inert ingredients and 2,4-D.” Giddings and Habig's (2019) commentary disregards the present understanding of the basic principles of hermetic responses to xenobiotics and the role of endocrine disruption in risk assessment. In both Dehnert et al. (2018) and DeQuattro and Karasov (2016), a strict ordered sequence of decrease in dose–response patterns (i.e., monotonic decline) was not apparent in the data; however, nonmonotonic responses (i.e., hermetic) were observed. Nonmonotonic responses have been observed throughout the peer-reviewed literature and explained by others as being consistent with underlying endocrine mechanisms that have been shown to cause deleterious effects (Arcand-Hoy and Benson 1998; Jobling and Tyler 2003; Hotchkiss et al. 2008; Tillitt et al. 2010; Vandenberg et al. 2012; Bencic et al. 2013; Zoeller and Vandenberg 2015; DeQuattro and Karasov 2016). Furthermore, previous evidence has shown that 2,4-D can impact endocrine physiology in fish (Xie et al. 2005; DeQuattro and Karasov 2016). We do not claim, and have not claimed in any of our reports, that the nonmonotonic patterns we observed are conclusive evidence of underlying endocrine mechanisms, only that the patterns we observed are consistent with a nonmonotonic response, and by no means a reason to discount statistically significant impacts of 2,4-D and 2,4-D formulations on larval survival. We agree with Zoeller and Vandenberg (2015) in that large effects at low doses and nonmonotonic dose–response relationships are a challenge to more conventional dose–response relationships in risk assessment, “but an understanding of mechanism should not be required to accept observable, statistically valid phenomena.” Furthermore, it is widely accepted by researchers and regulation entities alike that the presence of nonmonotonic relationships for a given compound, especially those exhibiting deleterious responses, must be accounted for in risk assessment for any compound (Vandenberg et al. 2012, 2013; Zoeller and Vandenberg 2015). In addition to possible endocrine disruption, Dehnert et al. (2018) suggested another possible hypothesis for the impacts of 2,4-D formulations on fathead minnow larval survival in that exposure to 2,4-D may increase oxidative stress; 2,4-D moieties have been shown to increase such stress in previous peer-reviewed literature (Oruc et al. 2004; Atamaniuk et al. 2013; Li et al. 2017). Most recently, we found that 2,4-D at concentrations below putative NOECs alters the development and function of neural circuits underlying vision of larval zebrafish (Danio rerio), and thereby reduces visually guided behaviors essential to survival (Dehnert et al. 2019). Even without an adequate explanation for the Dehnert et al. (2018) results, it would be irresponsible to ignore the results concerning the use of 2,4-D for aquatic weed control. The 2005 USEPA reregistration eligibility decision explicitly stated that there was limited literature on the effects of low-level chronic 2,4-D exposure on fish. In the past 3 yr, DeQuattro and Karasov (2016) and Dehnert et al. (2018, 2019) have all now reported declines in survival, starting as early as 6 d post fertilization in larval fish after exposure to 2,4-D concentrations below the previously presumed NOECs (US Environmental Protection Agency 2005). Moreover, other researchers have shown impacts on growth, acetylcholinesterase activity, and metabolic parameters at 2,4-D concentrations below presumed NOECs (0.10–10 ppm; Xie et al. 2005; da Fonseca et al. 2008; Menezes et al. 2015). Therefore, our message stands that the use of 2,4-D and 2,4-D commercial herbicides could present a risk to larval survival of native fishes and that its use should be reevaluated. We encourage other researchers to look at this issue and share their work in the peer-reviewed literature.