- Email: [email protected]
Does long term low impact stress cause population extinction?
Environmental Pollution xxx (2016) 1e10
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Does long term low impact stress cause population extinction?* M.J.B. Amorim a, *, C. Pereira a, A.M.V.M. Soares a, J.J. Scott-Fordsmand b a b
Department of Biology & CESAM, University of Aveiro, 3810-193 Aveiro, Portugal Department of Bioscience, Aarhus University, Vejlsøvej 25, 8600 Silkeborg, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 April 2016 Received in revised form 10 November 2016 Accepted 14 November 2016 Available online xxx
This study assessed and monitored 40 consecutive reproduction tests - multigenerational (MG) - of continuous exposure to Cd (at 2 reproduction Effect Concentrations (EC): EC10 and EC50) using the standard soil invertebrate Folsomia candida, in total 3.5 years of data were collected. Endpoints included survival, reproduction, size and metallothionein (MTc) gene expression. Further, to investigate adaptation to the toxicant, additional standard toxicity experiments were performed with the MG organisms of F6, F10, F26, F34 and F40 generations of exposure. Exposure to Cd EC10 caused population extinction after one year, whereas populations survived exposure to Cd EC50. Cd induced the up-regulation of the MTc gene, this being higher for the higher Cd concentration, which may have promoted the increased tolerance at the EC50. Moreover, EC10 induced a shift towards organisms of smaller size (positive skew), whereas EC50 induced a shift towards larger size (negative skew). Size distribution shifts could be an effect predictor. Sensitivity increased up to F10, but this was reverted to values similar to F0 in the next generations. The maximum Cd tolerance limits of F. candida increased for Cd EC50 MG. The consequences for risk assessment are discussed. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Multigenerational Tolerance Population size distribution Skewness Kurtosis
1. Introduction Long term exposure to chemicals is a common scenario, with consequent multigenerational effects on organisms. This is particularly the case in the context of persistent chemicals at sub-lethal concentrations and, especially, within a solid compartment (e.g. soil). Current guidelines to assess environmental effects of chemicals are based on toxicity within one generation, even for POPs (Persistent Organic Pollutants). Moreover, even when effects are assessed at long term chronic levels, this constitutes an exposure during a small fraction of the organisms’ life cycle, e.g. from mature adults till offspring reproduction (e.g (OECD, 2004).). Obviously, rather than using only a fragment of a life cycle, a full life cycle test would be preferable, but such tests are seldom available for terrestrial invertebrates (Bicho et al., 2015). When it comes to assessing effects over more life-cycles, i.e. for more than one generation, the tests documenting the possible effects are even scarcer.
*
This paper has been recommended for acceptance by Prof. von Hippel Frank A. * Corresponding author. E-mail addresses: [email protected] (M.J.B. Amorim), CeciliaManuela.Pereira@ UGent.be (C. Pereira), [email protected] (A.M.V.M. Soares), [email protected] (J.J. ScottFordsmand).
Hence, given that the hazard is predicted based on current standard duration tests, the Environmental Risk Assessment (ERA) framework will also, in principle, only predict shorter term risk and will not include potential effects of multigenerational (MG) exposure. This is a gap and a concern, as sustainability is the aim. The ability of organisms to survive and reproduce in contaminated habitats has been widely reported and has caused, among others, increased resistance of pests and pathogens to pesticides and antibiotics (Bickham et al., 2000). For instance, studies by Ward and Robinson (2005) showed an increased resistance in Daphnia magna populations when exposed to Cd along eight generations. Studies from historically polluted sites have shown that the impact can persist years after original source input (Roelofs et al., 2009). Effects can vary, e.g. species selection, development of resistance to higher levels of the source pollutant through adaptation or tolerance, epigenetics, etc. For instance, changes in genetic variability and allele frequencies of populations, which result from induced mutations and population bottlenecks (Bickham et al., 2000), can induce increased survival in contaminated environments (Timmermans et al., 2005). However, the associated changes often carry a decrease in the overall fitness, i.e. less viability to secondary stresses: resistant animals may be less fit and eliminated when exposed to different stressors (Ward and Robinson, 2005). Overall, few multigenerational studies have been conducted.
http://dx.doi.org/10.1016/j.envpol.2016.11.044 0269-7491/© 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
2
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Some of the progress has been driven by endocrine disruptors research (e.g. (OECD, 2005; Segner et al., 2003)), where effects are not observed until a second generation. In terms of soil invertebrates, the examples include mainly few (1e3) generation studies, e.g. Enchytraeus crypticus (Menezes-Oliveira et al., 2013), Enchytraeus albidus (Lock and Janssen, 2002), Folsomia candida (Campiche et al., 2007; Ernst et al., 2016) with one example of 10 generations (Paumen et al., 2008), Eisenia fetida (Spurgeon and Hopkin, 2000), Lumbricus rubellus (Langdon et al., 2009, 2003). Menezes-Oliveira et al. (2013) studied Cu exposed E. crypticus along 2 generations, showing an increase in sensitivity in the 2nd generation. Studies using F. candida showed that pre-exposure to certain endocrine disrupters (not all) caused an increase in sensitivity in the 2nd generation (Campiche et al., 2007). In an other study with F. candida (Ernst et al., 2016), a 2 generation design is proposed to differentiate between substances of potential longterm or low risk. Also with F. candida, exposure to phenantrene EC50 showed complete reproductive failure and subsequent extinction of the population after four generations (Paumen et al., 2008). Spurgeon and Hopkin (Spurgeon and Hopkin, 2000) showed that E. fetida developed resistance to Zn after pre-exposure up to 2 generations (higher LC50, LC90, LC99). Also, studies over 3 generations with another earthworm species, L. rubellus (Langdon et al., 2009) showed resistance to As for pre-exposed animals. On the other hand, no resistance to Cu was demonstrated (Langdon et al., 2003) showing the metal specificity. Lock and Janssen (2001) showed that E. albidus populations pre-exposed to Cd (18 months) had increased metallothionein levels, but had no changes in acute and chronic toxicity. Hence, as outlined above it is envisaged that even in a laboratory setting a multigenerational experiment will provide additional information on the long-term consequences of pollution of natural populations [such phenomena are notoriously difficult to study directly on natural populations, i.e. in natural environments]. The aim of this study was to assess and monitor the impact along multigenerational exposures and to do this along a substantial period of time fully monitored. The standard soil invertebrate Folsomia candida was used. Cadmium (Cd) was selected as a test chemical because it is both relevant and well-studied in this species and is also non-essential and persistent. This was a very long term study, which lasted 41 consecutive reproduction tests (each test was 28 days) of continuous exposure to Cd, i.e. more than 3 years of data was collected. The recovery potential was also assessed in each generation by transfer to control soil tested in parallel. Moreover, to investigate adaptation to the toxicant, additional standard toxicity experiments were performed after 6, 10, 26, 34 and 40 generations of exposure. 2. Materials and methods 2.1. Test organism The standard soil test organism Folsomia candida (Collembola) was used. Organisms were kept in laboratory in petri dishes in a substrate consisting of a mixture of plaster of Paris and activated charcoal (8:1). Cultures were kept at 19 C, photoperiod 16:8 h (light-dark) and were fed twice a week with dried bakers’ yeast. For testing, organisms were of synchronized age (10e12 days old). 2.2. Test soil The natural standard soil LUFA 2.2 was used. The properties can be summarised as follows: pH ¼ 5.5, organic matter ¼ 3.9%, grain size distribution: 6% clay; 17% silt; 77% sand.
2.3. Test chemicals and spiking procedures Cadmium chloride anhydrous (CdCl2, Sigma-Aldrich, 99%) was used. Cadmium was added to the test soil as aqueous solution in deionised water. Test concentrations were 32 and 60 mg Cd/kg dry soil, i.e. the known approximate EC10 and EC50 for reproduction (van Gestel and Mol, 2003). The spiked soil was allowed to equilibrate for three days prior to test start (McLaughlin et al., 2002). Moisture was adjusted to 50% of the maximum water holding capacity (WHC). 2.4. Experimental procedures 2.4.1. Effect concentration response exposure The standard guideline (OECD, 2009) was followed. In short, 10 juvenile organisms were selected randomly from cultures of synchronized age (10e12 days) and introduced in each test vessel containing the moist test soil and food supply. Four replicates per treatment were performed. One additional container was prepared for pH measurements at the beginning and at the end of the experiment. Test concentration range was 0-32-64-128-256 mg Cd/ kg soil DW, following the same described spiking procedures. Exposure lasted 28 days, at 20 ± 2 C, 16:8 h photoperiod. Water content and food were replenished weekly. At test end, the adults (F0) and hatched juveniles (F1) were recovered by flotation and counted. The process consisted of adding water into the test vessel, transferring all into a 500 ml glass beaker and gently stirring with a spatula to maximize flotation and spread the organisms on the water surface. The adults were counted and removed. A picture of the organisms on the water surface was recorded with a digital camera and the number (and size (length) in the MG test) of organisms was obtained using image analysis software (SPSS, 2003). Concentration response experiments, to assess the Effect Concentration (ECx), were assessed at generations F6, F10, F26, F34 and F40, always performing additional F0 in parallel, i.e. using organisms from synchronized cultures. 2.4.2. Multigenerational (MG) exposure The standard guideline (OECD, 2009) was followed with adaptations. The maturity of F. candida is usually reached around 15e16 days (6th instar (Snider, 1973),) in control animals, but this may take longer time under exposure conditions. In short, 10 juvenile organisms were selected randomly from cultures of synchronized age (10e12 days) and introduced in each test vessel containing the moist test soil and food supply. Samples were named as Ct_EC10 and Ct_EC50 (for the controls of the respective Cd treatment) and Cd_EC10 and Cd_EC50 (for the Cd treatments). In terms of replicates, four replicates were performed for control and ten replicates per Cd treatments (32 and 60 mg/kg) to ensure enough organisms to continue sequential generations, i.e. 4 replicates of Ct_EC10, 10 replicates of Cd_EC10, 4 replicates of Ct_EC50 and 10 replicates of Cd_EC50. At each test end, after removing the adults (Fx), Fxþ1 juveniles were transferred from the water surface using a small spoon to a recipient with a layer of mixed plaster of Paris and activated charcoal, adsorbing extra existing water. After this, 10 juvenile organisms were selected within similar size (selecting the larger sized animals) and transferred to a new test vessel with soil spiked under the exact same conditions as the previous generation, including the water control. Extra juvenile organisms were snap frozen in liquid nitrogen and kept at 80 C for further analysis. This procedure was repeated over all generations for EC10 and EC50, respectively. A schematic representation of the experimental design is given in Fig. 1. For an overview, the experimental set up included the following
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
3
Fig. 1. Experimental design used for the multigenerational exposure of Folsomia candida to cadmium (Cd), indicating the generations (Fx) and corresponding time in days and years. The population exposed to Cd_EC10 was extinct after ca. 1 year (F13), whereas to Cd_EC50 it continued until termination (F41).
endpoints and samplings per generation (Table S1). 2.5. Methallothionein (MT) gene expression measurements (qPCR) Total RNA of samples from selected generations: F1, F6, F10, F11 (Ct_EC10, Ct_EC50, Cd_EC10 and Cd_EC50) and F13, F14, F24, F25, F28, F32, F33, F34, F35, F36, F38, F39, F41 (Ct_EC50 and Cd_EC50) was extracted using the SV Total RNA Isolation system (Promega). Each sample consisted of approx. 60 juveniles per respective generation. Organisms from the culture with no history of exposure to Cd were sampled and used as reference (F0) to compare the metallothionein-like motif (MTc) expression levels against all other multigeneration samples. Approximately 0.7 mg of each sample total RNA was then converted into cDNA through a reverse transcription reaction using the SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen). The cDNA was 4x diluted and 2 mL were used in 20 mL PCR reaction volumes containing forward and reverse primers and Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen). Primers used for amplification of the MTc gene like (Nakamori et al., 2010) (MTc; GenBank accession number: AB509260) were: 50 -AGCCAATATTTTCGAGTGGAGA-30 (forward) and 50 -CAAGATGCTCGAATAGCAACAGTA-30 (reverse). The reference gene used for normalization was succinate dehydrogenase (SDHA; Collembase (http://www.collembase.org) accession number: Fcc06005) with the following primer sequences: 50 -ACACTTTCCAGCAATGCAGGAG-30 (forward) and 50 -TTTTCAGCCTCAAATCGGCA-30 (reverse). Efficiency and specificity of each primer were determined by observing the obtained standard and melting curves, respectively, for all primer sets. Quantitative real-time PCR (qPCR) was performed in a 7500 Real-Time PCR System (Applied Biosystems) with three biological replicates per treatment, applied in triplicate on a 96-well optical plate (GeneAmp®, Applied Biosystems). Reaction conditions consisted of one initial cycle at 50 C for 2 min, followed by a denaturation step at 95 C for 2 min, 40 cycles at 95 C for 32 s and 1 cycle at 60 C for 1 min. Finally, a dissociation step was made consisting of 15 s at 95 C, 1 min at 60 C, and 15 s at 95 C. A mean normalized expression value was calculated from the obtained Ct values of the test gene with Relative Expression Software Tool (REST-MSC). 2.6. Data analysis Three population parameters were evaluated: survival, reproductive output and animal size distribution. For survival and reproduction, the Effect Concentration (ECx) and its related confidence intervals were calculated using a logistic 2 parameters regression model (Toxicity Relationship Analysis Program (TRAP)
eversion 1.20a, US EPA). Organisms' size distribution was analysed in terms of changes in the average size (mean of fitted normal distribution), the broadness in size distribution (standard deviation), possible shift between small and larger animals’ distribution (skewness), and the shape of the distribution (kurtosis), using the data description option in SigmaPlot software for Windows Version 11.0 (Systat software Inc., San Jose, CA, USA). The populations were checked for clutch effects, i.e. multiple size classes, by bin distribution and using Freedman-Diaconis rule for bins number definition, although such classes could not be found e.g. as in (Marks et al., 2015). The standard deviations as shown in Fig. 3 by the horizontal dashed lines are the standard deviations calculated for all the values plotted in the figure (i.e. for all the generations). The purpose is to show the deviating population. Significance of differences in gene expression between control and treatments were calculated using t-test (p < 0.05). 3. Results 3.1. Multigenerational (MG) exposure 3.1.1. Survival, reproduction, size Results in terms of survival and reproduction along the MG exposure to Cd can be seen for the Cd_EC10 and for the Cd_EC50 in Fig. 2. For both MG exposures, survival and reproduction increased until generation 6e7, after which a decrease was observed for the subsequent 6 generations. As can be observed, the MG exposure to Cd EC10 (32 mg/kg) lead to population extinction after approx. 1 year (F13), whereas the exposure to the EC50 (60 mg/kg) continued until F41, being terminated without being extinct. In the latter case, the population numbers varied along generations 13e41, shifting from minor increase to decrease. Results for MG regarding size data analysis are summarised in Fig. 3. For Cd_EC10, the mean of the size distribution showed a fluctuation over time similar to that of the controls; the size distribution had a tendency to symmetric (y0) to moderate positive skewness and a positive kurtosis, similar to the control, till the time the population collapsed. For Cd_EC50, the mean of the size distribution showed a fluctuation over time, as for the controls, but with the mean being smaller than for both control and EC10 exposed organisms. For the Cd_EC50 animals, there was an initial moderate to high skewness, which moved towards symmetric to moderate negative skewness over time. The kurtosis fluctuated around the mean, but with a higher frequency of positive kurtosis in the beginning and negative
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
4
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Fig. 2. Results in terms of survival and reproduction for the multigenerational exposure of Folsomia candida to cadmium (Cd_EC10: 32 mg Cd/kg and Cd_EC50: 60 mg Cd/kg) and control (Ct_EC10 and Ct_EC50) along 13 and 41 generations (corresponding to generation 12 and 40 in terms of adults) for the EC10 and EC50, respectively. Results are expressed as average ± standard error (Av±SE). Grey e white background delimit the yearly rotation.
towards the end (F29-F41). There was no clear correlation between skewness and kurtosis. The standard deviation shows a fluctuation in variation that is more or less similar along the generations, with
few exceptions of higher and smaller at a particular generation. For guidance (Fig. 4), please note that negative skewness means a higher number of large animals (compared to mean) than small,
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
5
Fig. 3. Size distribution plots including data regarding mean (normalized to start (F0) size as 0) and standard deviation (±average SD represented by the horizontal dashed lines, which refer to all generations data), skewness and kurtosis values for the multigenerational exposure of Folsomia candida to cadmium in the various treatments (Ct_EC10: control; Cd_EC10: 32 mg Cd/kg; Ct_EC50: control; Cd_EC50: 60 mg Cd/kg).
Fig. 4. Schematic illustration of the population distribution for size variation in terms of skewness (measure of symmetry, or lack of it): negative/left (more organisms of large size) and positive/right (more organisms of small size), [normal (no skew): larger amount of organisms of average size]. Kurtosis is a measure of heavy or light tails, also defined as the degree of peakedness, negative indicates broad variation and tends to signal 2 populations, positive indicates narrow variation of 1 population.
positive skewness means higher number of small animals (compared to mean) than large. The right side graph shows an example of the individual size distributions for 3 (out of 728 in total) individual replicates (F0_EC10, F19_EC50, F39_EC50), where the bins and Gaussian distributions are plotted. Hence, the observations were that for EC10 MG exposures there was a tendency towards relatively more small animals with time (positive skew), after which the population collapsed. For EC50 MG exposures, the shift was towards large sized animals (negative skew), without termination.
3.1.2. Effect concentration response tests Results from effect concentration tests performed with the MG exposed animals are summarised in Table 1. Additionally, results from the standard tests (F0) are also shown for reference. For details of the fit to model, please see Fig. 5. For exposures to the EC10 (32 mg/kg), F. candida populations were equally sensitive after 6 MG as the F0 and more sensitive after 10 MG, as can be depicted by the EC50 (significant decrease to approximately half at F10). For exposures to the EC50 (60 mg/kg), a similar increase in sensitivity was measured with the F10 MG organisms. For F26, F34 and F40, the population sensitivity did not vary
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
6
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Table 1 Summary of the Effect Concentrations (ECx, in bold) estimations (logistic 2 parameters) obtained for range finding tests with Folsomia candida exposed to Cd, using organisms directly from the cultures (F0) and previously exposed to cadmium (32 and 60 mg/kg) for 6, 10, 26, 34 and 40 generations (F6, F10, F26, F34 and F40, respectively). n.d.: not determined. NE: No Effect. CI: Confidence Interval. Generation F0 (reference)
EC10
23 6 < CI < 42 MG exposed animals [EC10] ¼ 32 mg Cd/kg F6 28.5 7.4 < CI < 50 F10 n.d. 95 < CI < 35 MG exposed animals [EC50] ¼ 60 mg Cd/kg F6 23 13 < CI < 60 F10 n.d. 43 < CI < 14 F26 n.d. 230 < CI < 85 F34 n.d. 64 < CI < 36 2 F40 31 < CI < 35
EC50
EC90
LC50
68 59 < CI < 78
113 87 < CI < 139
239 150 < CI < 327
79.2 66 < CI < 93 31 10 < CI < 53
130 97 < CI < 162 92 28 < CI < 156
164 105 < CI < 222 240 3442 < CI < 3922
83 61 36 25 59 14 45 27 54 40
143 90 < 87 60 < 191 27 < 104 54 < 106 69 <
261 239 < CI < 283 297 86 < CI < 680 120 108 < CI < 131 NE
< CI < 106 < CI < 47 < CI < 105 < CI < 64 < CI < 68
CI < 196 CI < 113 CI < 355 CI < 155
NE CI < 144
Fig. 5. Results of the range finding tests in terms of survival (left) and reproduction (right) for Folsomia candida exposed to a range of concentrations of CdCl2 (mg Cd/Kg soil DW). Cd_EC10: organisms exposed to 32 mg Cd/kg for x (Fx) generations. Cd_EC50: organisms exposed to 60 mg Cd/kg for x (Fx) generations. Lines represent the fit to model (logistic 2 parameters), dotted line is the simple link between means for tracing.
significantly compared to F0 (similar EC50), but there was a clear increase in the variability (larger confidence intervals). The LC50 showed a tendency to increase (except for F26), which is clear after generation 34 and 40, where no mortality occurs: organisms seem to tolerate higher maximum concentrations after MG exposures to Cd. 3.1.3. Metallothionein (MT) like gene expression The MT like gene qPCR ratios can be seen in Fig. 6, including the direction (up- or down-regulation). MT expression was increased after one Cd exposure cycle (for
both EC10 and EC50). The MT fold change was positively concentration dependent, being nearly twice as high in the animals exposed to the EC50 (average ca. 1200, versus ca. 600 for the EC10 in F6). This upregulation of MT was turned off when organisms were transferred back to control soil. Moreover, the controls show a clear decrease e.g. in the exposure to the EC10_Cd, from F1 to F6. In the exposure to the EC50_Cd, from F24 to F41 there is an upregulation of the MT gene. There is a strong downregulation at F25 control.
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
7
Fig. 6. Results in terms of metallothionein like (MTc) gene expression (qPCR) for the multigenerational cadmium exposed Folsomia candida in the various treatments (Cd_EC10: 32 mg Cd/kg; Ct_EC10: control; Cd_EC50: 60 mg Cd/kg; Ct_EC50: control) at selected generations. Break in the Y scale is between 100 and 250. Dashed line at the 0 level for reference to the control; *: p < 0.05.
4. Discussion Despite the high importance and current increased interest for transgenerational effects, there are still few studies available with invertebrates, and those that do exist mostly report two to eight generations (e.g. Vandegehuchte et al., 2009, 2010; Ward and Robinson, 2005). To our knowledge, no other study has performed such an extensive long term and multigenerational exposure, except studies on the moth beet armyworm (Spodoptera exigua) exposed to Cd or Zn contaminated food over many generations (Kafel et al., 2014, 2012). In the present study, the MG effects for the exposure to the Cd EC10 caused population extinction after one year, whereas the exposure to the Cd EC50 allowed the population to continue to survive and reproduce. As showed by Ward and Robinson (2005), Cd-acclimated daphnids exposed for eight successive generations had a decrease in genetic diversity compared to unexposed animals. Hence, since the organisms are parthenogenetic, the issue here may be a lack of genetic diversity/adaptation to cope with the environmental variable. This could be one of the factors explaining the extinction observed for Cd EC10 MG, although it is also possible that some stochasticity was involved, as the EC50 organisms also reached a minimum for adults around the same generation, but a few generations later for the juveniles. Finally, the experimental setup had a time frame (28 days), hence, a long delay in adulthood may have caused the extinction. This is partly supported by the observation that there was a continuous low number of adults, but high numbers of juveniles in the following generations, indicating a change in developmental speed, growth and reproductive performance and timing. Under field conditions it is likely that sometimes timing (i.e. a delay) is extremely important and will cause
extinction (e.g. changing weather conditions), whereas in other cases a delay may not matter (e.g. due to stable weather conditions). As noted by (Kozlowski and Gawelczyk, 2002) optimum size is the one that ensures maximum reproduction, and this depends on the adults’ potential to survive and reproduce. That the EC10, but not the EC50 exposed animals, became extinct could be due to enhanced adaptive mechanisms at higher exposure concentrations, e.g. through the observed difference in size distribution between EC10 and EC50 exposed animals. Lower Cd levels (EC10) appear to have induced a shift to select organisms of smaller size, whereas comparatively higher Cd levels (EC50) induced a shift to select organisms of larger size (relatively, i.e. within populations of one generation). This trait may have offered an advantage to the EC50 exposed population in terms of survival/ reproduction in the long term. An ant population study reported by (Grzes et al., 2015) found that ants along a pollution gradient (combination of Zn, Cd and Pb) had a skewness in head size distribution significantly correlated with the pollution level of the site, indicating that the frequency of small ants grew as the pollution level increased. The authors discuss that the bias towards the higher frequency of smaller animals could be due to 1) metal toxicity, 2) detoxification costs, 3) shifts in energy allocation, but may also have an adaptive function. The fact that collembolans may come in clutches (although these could not be separated within this study) suggests that the changes in population size distribution may possibly be composed of two phenomenon's e (1) a change in the relative size of the individual organisms due to developmental changes of the individual and (2) a change in the relative numbers between clutches, e.g. due to different sensitivities of the juveniles, but also the parent fitness (Scott-Fordsmand et al., 2000). Regarding metal toxicity (1), several studies have shown the
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
8
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
negative effect of metals on growth, including Cd (see e.g. Fountain and Hopkin, 2001; Ward and Robinson, 2005). The slower growth rate of springtails (Fountain and Hopkin, 2001) occurred at the highest exposure concentrations and was partly the result of reduced food consumption (contaminated food avoidance): EC50 values were approx. 4 times higher for exposure via food compared to soil (Fountain and Hopkin, 2001). Besides epidermal moulting, collembolans also have the ability to excrete metals from the body by intestinal exfoliation, which adds on to their mechanisms. Increased moulting frequency, however, consumes more energy and may result in reduced size as well (Posthuma et al., 1993). Regarding costs (2), detoxification to maintain body metal concentration on a certain level can be costly, e.g. due to the production of metallothioneins (discussed in more detail ahead). As a result, in more polluted sites less energy can be invested in growth. Shifts in energy allocation (3) can take place to maintain the number of organisms under stress, hence, developmental stages might become shorter, resulting in smaller body size. Pollutioninduced reduction of life span has been documented before (e.g. Grzes, 2010). Kozlowski and Gawelczyk (2002) conclude, among others, that size changes are the result of a compromise, where the optimal size for best reproductive success during a life time is prioritized. This means that for example the largest adult animals are not necessarily the most successful. The selected size will be the one that makes the best compromise between optimum reproduction and the optimum survival. Overall, it seems reasonable to assume that the observed size changes of F. candida along MG EC50 exposure is caused by a specific tradeoff with the imposed stress. Cadmium pre-exposed organisms can respond using different mechanisms, e.g. fish have been reported to have decreased uptake of Cd (Xie and Klerks, 2004). It is also known that such mechanisms can be concentration dependent, e.g. for Daphnia magna Hg uptake was reduced, but only at lethal concentrations (Tsui and Wang, 2006). It is hypothesized that animals develop defense mechanisms to reduce the metal uptake when this becomes lethal. This would support the differences observed between EC10 and EC50 found here, where possible defense mechanisms are induced more in the EC50 than in the EC10 exposed organisms. As observed before (Hensbergen et al., 2000; Nakamori et al., 2010), Cd induces the up-regulation of the Metallothionein (MTc) gene, with levels being exposure concentration related: higher Cd concentration results in higher MTc (until a plateau). It is suggested that MT Like Proteins (MTLP) are more important in detoxification at a higher metal pre-exposure or in a more contaminated environment (Tsui and Wang, 2007). Hence, at the tested sub-lethal ECx (EC10 and EC50) there are differences in terms of energy detoxifying costs, e.g. the cost associated with MT gene expression and the related translation onto protein. Such differences in energy allocation may play a role in the follow up mechanisms for the organisms. Contrary to the mechanism described above for fish (Xie and Klerks, 2004), daphnids have been observed to increase the (aqueous) uptake rate of metals after metal pre-exposure (e.g. Cd), while tolerance to metal toxicity increases (Tsui and Wang, 2007). This could seem contradictory, but although more metals are taken up in the pre-exposed animals, this can be partitioned to the already induced MTLP pool, hence, even more metals can be taken up and organisms suffer less than non-exposed organisms. It is not possible to clearly identify whether F. candida developed an adaptation level to Cd in the present experimental time (EC50 exposed); MT measurements seem to indicate acclimation because MT returns to basal levels when organisms are offered non-spiked soil. Nevertheless, the population survival and increased fitness are
due to a given trait. We could be observing slightly deleterious and slightly advantageous mutations, which have small effects on fitness, or deleterious and advantageous mutations, which have larger fitness effects (Lanfear et al., 2014), this will vary depending on the Ne (effective population number). Some suggest that the rate of adaptation is independent of Ne, and that it is determined by the rate of environmental change and number of traits upon which selection acts (Lourenço et al., 2013). Studies with Daphnia longispina from metal-contaminated waters suggest that organisms with adaptation for many generations are better at keeping their tolerances compared to the metalacclimated individuals (Lopes et al., 2006). Historically, contaminated field collembolan populations are also known to develop tolerance, e.g. the well-studied Cd tolerant Orchesella cincta (Roelofs et al., 2007). The mechanisms developed by this species include heritable increase of excretion efficiency, decrease in Cdinduced growth reduction, and over-expression of the MT gene. Moreover, results suggest that besides MT up-regulation, constitutive MT mRNA expression plays an important role in protection against Cd toxicity (Timmermans et al., 2005). Hence, the Cd detoxification mechanism is constitutively overexpressed in tolerant animals and cellular homeostasis is not disturbed upon exposure to Cd (Roelofs et al., 2009). Studies in F. candida (Nota et al., 2008) showed that exposure to Cd activates response mechanisms at various levels, e.g. immune system, hypoxia or detoxification. As quantified in a selection of samples, when exposed F. candida was transferred from Cd spiked soil to control, the MT levels dropped. This was not surprising, but what was intriguing was the larger decrease (negative values) in specific generations (EC10_Ct_F11 and EC50_Ct_F25). This larger MT reduction preceded a high reduction in the reproduction (EC10_Cd_F12/13, EC50_Cd_F26), hence, the mechanisms could be associated with the measured effect. Why this happened and in these particular generations it is unknown. MG exposed populations showed small changes in the reproduction ECx (Fx range finding tests) compared to the F0 standard reference. F10 had increased sensitivity, but there was no continuous increase with the number of generations. In a study with fish, it was demonstrated that the mRNA of MTLP in a Cd-tolerant fish was maternally transferred to the next generation (Lin et al., 2000); vertebrates and invertebrates can vary substantially in such mechanisms. For instance, in a study with D. magna continuously exposed to Hg, the MTLP was not transferred to the next generation (Tsui and Wang, 2006), but there was increased tolerance transference. Although the MG exposed populations showed minor changes in the ECx compared to the F0, there was an increase in the LC50 (e.g. F34, F40). This could indicate an increase in the maximum tolerance limits of Cd for F. candida exposed to Cd EC50 in terms of survival. The study by Kafel et al. (2012) also suggests lower metal tolerance among one-generation exposed insects in comparison with larvae exposed for many generations: the MG organisms had lower accumulation of Cd and increased anti-oxidant response, which could contribute to a greater efficiency in diminishing the toxic effects of the metal. The possibility of transgenerational inheritance of environmentinduced epigenetic changes to non-exposed subsequent generations has been shown for several organisms. For instance, in D. magna this was observed for MG exposure to 5-azacytidine (active pharmaceutical compound known to inhibit DNA methyltransferases) (Vandegehuchte et al., 2010). Interestingly, a concurrent reduction in body length was observed. This could be an indication that size/growth may be a good phenotypical endpoint for transgenerational effects. An experiment with D. magna and exposure to Cd over two generations suggests that Cd does not
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
affect their DNA methylation status (Vandegehuchte et al., 2009), although epigenetics by other mechanisms cannot be excluded. So far, it is unknown whether Cd (or other compounds) causes epigenetics in F. candida, but this would be interesting to investigate. 4.1. Consequences for risk assessment Currently, too little is known in terms of multigenerational effects to be able to include them reliably in hazard and, subsequently, risk predictions. As observed in the present study, up to F6 (6 months) the results could indicate an overall improvement in terms of survival and reproduction for MG exposed organisms, although no changes in the ECx. From F6 to F12, the shift to a decrease in overall performance occurred, this with highest consequences for the EC10 exposure where extinction occurred. The decrease in the EC50 values to about half, as measured with the F10 MG organisms, could be an indication of increased sensitivity, but still the effects after were unpredictable (EC10 crash and EC50 survived). Multigenerational studies with F. candida exposed to phenantrene EC50 showed a significant increase in the effect: complete reproductive failure and subsequent extinction of the population after four generations (Paumen et al., 2008). On the other hand, exposure to lower concentrations caused no effects on survival and reproduction, probably due to biotransformation. Hence, these F. candida studies (Paumen et al., 2008 and our own) show two opposite effects and consequences of MG for RA, 1. phenantrene with lack of adaptation and the all-or-nothing effect (below the threshold, no adaptation, and above this concentration the population went extinct) and 2. Cd with tolerance or adaptation potential (at EC10 no adaptation and population went extinct, at the EC50 adaptation or tolerance and population was maintained). Hence, the studies point to the fact that the effect levels do change, and even when effect levels do not change, the population structure does, which has consequences not only for the population, but also for the community, e.g. the predatory-prey relationship. This clearly supports the notion that the standard toxicity studies, where exposure is performed within one (or less than one) full generation cycle, are not adequate to assess the risk of chemicals. The methodology implemented here for multigenerational exposure worked well and can be used or replicated to test other substances. We recommend its use; the number of generations can be adjusted according to the purpose. Moreover, we recommend the recording of size as an additional endpoint to the standard test for F. candida [not excluding other endpoints in other invertebrate groups], which will not only add in terms of effect level, but also on the potential predictive value. 5. Conclusions Multigenerational exposure of F. candida to Cd showed that long term effects varied with concentration: EC10 exposure caused extinction after 1 year and EC50 continued until >3 years without becoming extinct. The reasons for the differences could be associated with underlying stress response mechanisms. We hypothesize the following, based on the measured endpoints (survival, reproduction, size, metallothionein (MTc) level and ECs monitored along MG): the EC50 exposed organisms responded with higher MTc expression and detoxifying mechanisms (compared to the EC10 exposed), which required more energy and change in strategy. This higher energy cost may be among the reasons for the smaller size of the EC50 exposed animals. The maximum tolerance levels of Cd increased for survival, but not for reproduction (e.g. F40_LC50 > F10_LC50). These changes seem to have favoured the survival of the population.
9
Acknowledgements Thanks are due for the financial support to CESAM (UID/AMB/ 50017), to FCT/MEC through national funds, and the co-funding by FEDER, within the PT2020 Partnership Agreement and Compete 2020. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2016.11.044. References Bicho, R.C., Santos, F.C.F., Gonçalves, M.F.M., Soares, A.M.V.M., Amorim, M.J.B., 2015. Enchytraeid reproduction testplus: hatching, growth and full life cycle testdan optional multi-endpoint test with Enchytraeus crypticus. Ecotoxicology 1053e1063. http://dx.doi.org/10.1007/s10646-015-1445-5. Bickham, J.W., Sandhu, S., Hebert, P.D., Chikhi, L., Athwal, R., 2000. Effects of chemical contaminants on genetic diversity in natural populations: implications for biomonitoring and ecotoxicology. Mutat. Res. 463, 33e51. Campiche, S., L'Ambert, G., Tarradellas, J., Becker-van Slooten, K., 2007. Multigeneration effects of insect growth regulators on the springtail folsomia candida. Ecotoxicol. Environ. Saf. 67, 180e189. http://dx.doi.org/10.1016/ j.ecoenv.2006.11.009. Ernst, G., Kabouw, P., Barth, M., Marx, M.T., Frommholz, U., Royer, S., Friedrich, S., 2016. Assessing the potential for intrinsic recovery in a Collembola twogeneration study: possible implementation in a tiered soil risk assessment approach for plant protection products. Ecotoxicology 25, 1e14. http:// dx.doi.org/10.1007/s10646-015-1560-3. Fountain, M.T., Hopkin, S.P., 2001. Continuous monitoring of Folsomia candida (Insecta: collembola) in a metal exposure test. Ecotoxicol. Environ. Saf. 48, 275e286. http://dx.doi.org/10.1006/eesa.2000.2007. Grzes, I.M., 2010. Zinc tolerance in the ant species myrmica rubra originating from a metal pollution gradient. Eur. J. Soil Biol. 46, 87e90. http://dx.doi.org/10.1016/ j.ejsobi.2009.11.005. Grzes, I.M., Okrutniak, M., Woch, M.W., 2015. Monomorphic ants undergo withincolony morphological changes along the metal-pollution gradient. Environ. Sci. Pollut. Res. 22, 6126e6134. http://dx.doi.org/10.1007/s11356-014-3808-5. Hensbergen, P.J., van Velzen, M.J., Nugroho, R. a, Donker, M.H., van Straalen, N.M., 2000. Metallothionein-bound cadmium in the gut of the insect Orchesella cincta (Collembola) in relation to dietary cadmium exposure. Comp. Biochem. Physiol. C. Toxicol. Pharmacol. 125, 17e24. ska, E., Zawisza-Raszka, A., Migula, P., 2014. The effects Kafel, A., Rozpedek, K., Szulin of cadmium or zinc multigenerational exposure on metal tolerance of Spodoptera exigua (Lepidoptera: noctuidae). Environ. Sci. Pollut. Res. 21, 4705e4715. http://dx.doi.org/10.1007/s11356-013-2409-z. ska, E., 2012. Effects of multigenerational cadKafel, A., Zawisza-Raszka, A., Szulin mium exposure of insects (Spodoptera exigua larvae) on anti-oxidant response in haemolymph and developmental parameters. Environ. Pollut. 162, 8e14. http://dx.doi.org/10.1016/j.envpol.2011.09.034. Kozlowski, J., Gawelczyk, A.T., 2002. Why are species ’ body size distributions usually skewed to the right;? Funct. Ecol. 16, 419e432. Lanfear, R., Kokko, H., Eyre-Walker, A., 2014. Population size and the rate of evolution. Trends Ecol. Evol. 29, 33e41. http://dx.doi.org/10.1016/ j.tree.2013.09.009. Langdon, C.J., Morgan, a J., Charnock, J.M., Semple, K.T., Lowe, C.N., 2009. As-resistance in laboratory-reared F1, F2 and F3 generation offspring of the earthworm Lumbricus rubellus inhabiting an As-contaminated mine soil. Environ. Pollut. 157, 3114e3119. http://dx.doi.org/10.1016/j.envpol.2009.05.027. Langdon, C.J., Piearce, T.G., Meharg, A. a, Semple, K.T., 2003. Inherited resistance to arsenate toxicity in two populations of Lumbricus rubellus. Environ. Toxicol. Chem. 22, 2344e2348. Lin, H., Hsu, S., Hwang, P., 2000. Maternal transfer of cadmium tolerance in larval oreochromis mossambicus. J. Fish. Biol. 57, 239e249. Lock, K., Janssen, C.R., 2001. Tolerance changes of the potworm Enchytraeus albidus after long-term exposure to cadmium. Sci. Total Environ. 280, 79e84. Lock, K., Janssen, C.R., 2002. Multi-generation toxicity of zinc, cadmium, copper and lead to the potworm Enchytraeus albidus. Environ. Pollut. 117, 89e92. http:// dx.doi.org/10.1006/eesa.2002.2155. Lopes, I., Baird, D., Ribeiro, R., 2006. Genetic adaptation to metal stress by natural populations of Daphnia longispina. Ecotoxicol. Env. Saf 63, 275e285. min, S., Galtier, N., 2013. The rate of molecular adaptation in a Lourenço, J.M., Gle changing environment. Mol. Biol. Evol. 30, 1292e1301. http://dx.doi.org/ 10.1093/molbev/mst026. Marks, E.A.N., Molowny-Horas, R., Domene, X., 2015. Investigating the use of multigaussian nonlinear regressions to separate collembolan size classes for soil quality assessment. Water. Air. Soil Pollut. 226 http://dx.doi.org/10.1007/ s11270-015-2503-6. McLaughlin, M., Hamon, R., Parker, D., Pierzynski, G., Smolders, E., Thornton, I.,
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
10
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Welp, G., 2002. Test methods to determine hazards of sparingly soluble metal compounds in soils. In: Fairbrother, A., Glazebrook, P.W., Tarazona, J.V., S.N. (Eds.), Soil Chemistry. Society of Environmental Toxicology and Chemistry (SETAC) Press, Pensacola, Florida, pp. 5e85. Menezes-Oliveira, V.B., Damgaard, C., Scott-Fordsmand, J.J., Amorim, M.J.B., 2013. Interaction between density and Cu toxicity for Enchytraeus crypticus e comparing first and second generation effects. Sci. Total Environ. 458e460, 361e366. Nakamori, T., Fujimori, A., Kinoshita, K., Ban-nai, T., Kubota, Y., Yoshida, S., 2010. mRNA expression of a cadmium-responsive gene is a sensitive biomarker of cadmium exposure in the soil collembolan Folsomia candida. Environ. Pollut. 158, 1689e1695. http://dx.doi.org/10.1016/j.envpol.2009.11.022. €n, J., De Nota, B., Timmermans, M.J.T.N., Franken, O., Montagne-Wajer, K., Marie Boer, M.E., De Boer, T.E., Ylstra, B., Van Straalen, N.M., Roelofs, D., 2008. Gene expression analysis of collembola in cadmium containing soil. Environ. Sci. Technol. 42, 8152e8157. http://dx.doi.org/10.1021/es801472r. OECD, 2004. OECD Guidelines for the Testing of Chemicals 220 (Enchytraeid Reproduction Test). OECD, 2005. OECD Environment Health and Safety Publications Series on Testing and Assessment No. 50: Detailed Review Paper on Aquatic Arthropods in Life Cycle and Two-generation Toxicity Tests. OECD, 2009. OECD Guidelines for Testing of Chemicals - Collembolan Reproduction Test in Soil. Guideline No.232. Paumen, M.L., Steenbergen, E., Kraak, M.H.S., van Straalen, N.M., Van Gestel, C.A.M., 2008. Multigeneration exposure of the springtail Folsomia candida to phenanthrene: from dose - response relationships to threshold concentrations. Environ. Sci. Technol. 42, 6985e6990. Posthuma, L., Hogervorst, R.F., Joosse, E.N.G., Van Straalen, N.M., 1993. Genetic variation and covariation for characteristics associated with cadmium tolerance in natural populations of the springtail orchesella cincta (L.). Evol. (N. Y) 47, 619e631. €n, J., Roelofs, D., Janssens, T.K.S., Timmermans, M.J.T.N., Nota, B., Marie Bochdanovits, Z., Ylstra, B., Van Straalen, N.M., 2009. Adaptive differences in gene expression associated with heavy metal tolerance in the soil arthropod Orchesella cincta. Mol. Ecol. 18, 3227e3239. http://dx.doi.org/10.1111/j.1365294X.2009.04261.x. €n, J., Van Straalen, N.M., 2007. Differential gene expression profiles Roelofs, D., Marie associated with heavy metal tolerance in the soil insect Orchesella cincta. Insect biochem. Mol. Biol. 37, 287e295. http://dx.doi.org/10.1016/j.ibmb.2006.11.013. Scott-Fordsmand, J.J., Krogh, P.H., Weeks, J.M., 2000. Responses of Folsomia
fimetaria (Collembola: isotomidae) to copper under different soil copper contamination histories in relation to risk assessment. Environ. Toxicol. Chem. 19, 1297e1303. €fers, C., Segner, H., Caroll, K., Fenske, M., Janssen, C.R., Maack, G., Pascoe, D., Scha Vandenbergh, G.F., Watts, M., Wenzel, A., 2003. Identification of endocrinedisrupting effects in aquatic vertebrates and invertebrates: report from the European IDEA project. Ecotoxicol. Environ. Saf. 54, 302e314. http://dx.doi.org/ 10.1016/S0147-6513(02)00039-8. Snider, R., 1973. Laboratory observations on the biology of Folsomia candida (Wil lem) (collembola: isotomidae. Rev. D’Ecologie Biol. du Sol. 10, 103e124. SPSS, 2003. SigmaScan Pro 5dImage Analysis. SPSS for Windows 12.0. Spurgeon, D.J., Hopkin, S.P., 2000. The development of genetically inherited resistance to zinc in laboratory-selected generations of the earthworm Eisenia fetida. Environ. Pollut. 109, 193e201. Timmermans, M.J.T.N., Ellers, J., Roelofs, D., van Straalen, N.M., 2005. Metallothionein mRNA expression and cadmium tolerance in metal-stressed and reference populations of the springtail orchesella cincta. Ecotoxicology 14, 727e739. http://dx.doi.org/10.1007/s10646-005-0020-x. Tsui, M., Wang, W.-X., 2006. Acute toxicity of mercury to daphnia magna under different conditions. Env. Sci. Technol. 40, 4025e4030. Tsui, M.T.-K., Wang, W.-X., 2007. Biokinetics and tolerance development of toxic metals in Daphnia magna. Environ. Toxicol. Chem. 26, 1023e1032. http:// dx.doi.org/10.1897/06-430r.1. van Gestel, C. a. M., Mol, S., 2003. The influence of soil characteristics on cadmium toxicity for Folsomia candida (Collembola: isotomidae). Pedobiol. (Jena) 47, 387e395. http://dx.doi.org/10.1078/0031-4056-00202. Vandegehuchte, M., Kyndt, T., Vanholme, B., Haegeman, A., Gheysen, G., Janssen, C., 2009. Occurrence of DNA methylation in daphnia magna and influence of multigeneration Cd exposure. Environ. Int. 35, 700e706. http://dx.doi.org/ 10.1016/j.envint.2009.01.002. re, F., Vanhaecke, L., Vanden Berghe, W., Janssen, C.R., Vandegehuchte, M.B., Lemie 2010. Direct and transgenerational impact on daphnia magna of chemicals with a known effect on DNA methylation. Comp. Biochem. Physiol. C. Toxicol. Pharmacol. 151, 278e285. http://dx.doi.org/10.1016/j.cbpc.2009.11.007. Ward, T.J., Robinson, W.E., 2005. Evolution of cadmium resistance in Daphnia magna. Environ. Toxicol. Chem. 24, 2341e2349. Xie, L., Klerks, P., 2004. Changes in cadmium accumulation as a mechanism for cadmium resistance in the least killifish Heterandria fomosa. Aquat. Toxicol. 66, 73e81.
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Does long term low impact stress cause population extinction?* M.J.B. Amorim a, *, C. Pereira a, A.M.V.M. Soares a, J.J. Scott-Fordsmand b a b
Department of Biology & CESAM, University of Aveiro, 3810-193 Aveiro, Portugal Department of Bioscience, Aarhus University, Vejlsøvej 25, 8600 Silkeborg, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 April 2016 Received in revised form 10 November 2016 Accepted 14 November 2016 Available online xxx
This study assessed and monitored 40 consecutive reproduction tests - multigenerational (MG) - of continuous exposure to Cd (at 2 reproduction Effect Concentrations (EC): EC10 and EC50) using the standard soil invertebrate Folsomia candida, in total 3.5 years of data were collected. Endpoints included survival, reproduction, size and metallothionein (MTc) gene expression. Further, to investigate adaptation to the toxicant, additional standard toxicity experiments were performed with the MG organisms of F6, F10, F26, F34 and F40 generations of exposure. Exposure to Cd EC10 caused population extinction after one year, whereas populations survived exposure to Cd EC50. Cd induced the up-regulation of the MTc gene, this being higher for the higher Cd concentration, which may have promoted the increased tolerance at the EC50. Moreover, EC10 induced a shift towards organisms of smaller size (positive skew), whereas EC50 induced a shift towards larger size (negative skew). Size distribution shifts could be an effect predictor. Sensitivity increased up to F10, but this was reverted to values similar to F0 in the next generations. The maximum Cd tolerance limits of F. candida increased for Cd EC50 MG. The consequences for risk assessment are discussed. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Multigenerational Tolerance Population size distribution Skewness Kurtosis
1. Introduction Long term exposure to chemicals is a common scenario, with consequent multigenerational effects on organisms. This is particularly the case in the context of persistent chemicals at sub-lethal concentrations and, especially, within a solid compartment (e.g. soil). Current guidelines to assess environmental effects of chemicals are based on toxicity within one generation, even for POPs (Persistent Organic Pollutants). Moreover, even when effects are assessed at long term chronic levels, this constitutes an exposure during a small fraction of the organisms’ life cycle, e.g. from mature adults till offspring reproduction (e.g (OECD, 2004).). Obviously, rather than using only a fragment of a life cycle, a full life cycle test would be preferable, but such tests are seldom available for terrestrial invertebrates (Bicho et al., 2015). When it comes to assessing effects over more life-cycles, i.e. for more than one generation, the tests documenting the possible effects are even scarcer.
*
This paper has been recommended for acceptance by Prof. von Hippel Frank A. * Corresponding author. E-mail addresses: [email protected] (M.J.B. Amorim), CeciliaManuela.Pereira@ UGent.be (C. Pereira), [email protected] (A.M.V.M. Soares), [email protected] (J.J. ScottFordsmand).
Hence, given that the hazard is predicted based on current standard duration tests, the Environmental Risk Assessment (ERA) framework will also, in principle, only predict shorter term risk and will not include potential effects of multigenerational (MG) exposure. This is a gap and a concern, as sustainability is the aim. The ability of organisms to survive and reproduce in contaminated habitats has been widely reported and has caused, among others, increased resistance of pests and pathogens to pesticides and antibiotics (Bickham et al., 2000). For instance, studies by Ward and Robinson (2005) showed an increased resistance in Daphnia magna populations when exposed to Cd along eight generations. Studies from historically polluted sites have shown that the impact can persist years after original source input (Roelofs et al., 2009). Effects can vary, e.g. species selection, development of resistance to higher levels of the source pollutant through adaptation or tolerance, epigenetics, etc. For instance, changes in genetic variability and allele frequencies of populations, which result from induced mutations and population bottlenecks (Bickham et al., 2000), can induce increased survival in contaminated environments (Timmermans et al., 2005). However, the associated changes often carry a decrease in the overall fitness, i.e. less viability to secondary stresses: resistant animals may be less fit and eliminated when exposed to different stressors (Ward and Robinson, 2005). Overall, few multigenerational studies have been conducted.
http://dx.doi.org/10.1016/j.envpol.2016.11.044 0269-7491/© 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
2
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Some of the progress has been driven by endocrine disruptors research (e.g. (OECD, 2005; Segner et al., 2003)), where effects are not observed until a second generation. In terms of soil invertebrates, the examples include mainly few (1e3) generation studies, e.g. Enchytraeus crypticus (Menezes-Oliveira et al., 2013), Enchytraeus albidus (Lock and Janssen, 2002), Folsomia candida (Campiche et al., 2007; Ernst et al., 2016) with one example of 10 generations (Paumen et al., 2008), Eisenia fetida (Spurgeon and Hopkin, 2000), Lumbricus rubellus (Langdon et al., 2009, 2003). Menezes-Oliveira et al. (2013) studied Cu exposed E. crypticus along 2 generations, showing an increase in sensitivity in the 2nd generation. Studies using F. candida showed that pre-exposure to certain endocrine disrupters (not all) caused an increase in sensitivity in the 2nd generation (Campiche et al., 2007). In an other study with F. candida (Ernst et al., 2016), a 2 generation design is proposed to differentiate between substances of potential longterm or low risk. Also with F. candida, exposure to phenantrene EC50 showed complete reproductive failure and subsequent extinction of the population after four generations (Paumen et al., 2008). Spurgeon and Hopkin (Spurgeon and Hopkin, 2000) showed that E. fetida developed resistance to Zn after pre-exposure up to 2 generations (higher LC50, LC90, LC99). Also, studies over 3 generations with another earthworm species, L. rubellus (Langdon et al., 2009) showed resistance to As for pre-exposed animals. On the other hand, no resistance to Cu was demonstrated (Langdon et al., 2003) showing the metal specificity. Lock and Janssen (2001) showed that E. albidus populations pre-exposed to Cd (18 months) had increased metallothionein levels, but had no changes in acute and chronic toxicity. Hence, as outlined above it is envisaged that even in a laboratory setting a multigenerational experiment will provide additional information on the long-term consequences of pollution of natural populations [such phenomena are notoriously difficult to study directly on natural populations, i.e. in natural environments]. The aim of this study was to assess and monitor the impact along multigenerational exposures and to do this along a substantial period of time fully monitored. The standard soil invertebrate Folsomia candida was used. Cadmium (Cd) was selected as a test chemical because it is both relevant and well-studied in this species and is also non-essential and persistent. This was a very long term study, which lasted 41 consecutive reproduction tests (each test was 28 days) of continuous exposure to Cd, i.e. more than 3 years of data was collected. The recovery potential was also assessed in each generation by transfer to control soil tested in parallel. Moreover, to investigate adaptation to the toxicant, additional standard toxicity experiments were performed after 6, 10, 26, 34 and 40 generations of exposure. 2. Materials and methods 2.1. Test organism The standard soil test organism Folsomia candida (Collembola) was used. Organisms were kept in laboratory in petri dishes in a substrate consisting of a mixture of plaster of Paris and activated charcoal (8:1). Cultures were kept at 19 C, photoperiod 16:8 h (light-dark) and were fed twice a week with dried bakers’ yeast. For testing, organisms were of synchronized age (10e12 days old). 2.2. Test soil The natural standard soil LUFA 2.2 was used. The properties can be summarised as follows: pH ¼ 5.5, organic matter ¼ 3.9%, grain size distribution: 6% clay; 17% silt; 77% sand.
2.3. Test chemicals and spiking procedures Cadmium chloride anhydrous (CdCl2, Sigma-Aldrich, 99%) was used. Cadmium was added to the test soil as aqueous solution in deionised water. Test concentrations were 32 and 60 mg Cd/kg dry soil, i.e. the known approximate EC10 and EC50 for reproduction (van Gestel and Mol, 2003). The spiked soil was allowed to equilibrate for three days prior to test start (McLaughlin et al., 2002). Moisture was adjusted to 50% of the maximum water holding capacity (WHC). 2.4. Experimental procedures 2.4.1. Effect concentration response exposure The standard guideline (OECD, 2009) was followed. In short, 10 juvenile organisms were selected randomly from cultures of synchronized age (10e12 days) and introduced in each test vessel containing the moist test soil and food supply. Four replicates per treatment were performed. One additional container was prepared for pH measurements at the beginning and at the end of the experiment. Test concentration range was 0-32-64-128-256 mg Cd/ kg soil DW, following the same described spiking procedures. Exposure lasted 28 days, at 20 ± 2 C, 16:8 h photoperiod. Water content and food were replenished weekly. At test end, the adults (F0) and hatched juveniles (F1) were recovered by flotation and counted. The process consisted of adding water into the test vessel, transferring all into a 500 ml glass beaker and gently stirring with a spatula to maximize flotation and spread the organisms on the water surface. The adults were counted and removed. A picture of the organisms on the water surface was recorded with a digital camera and the number (and size (length) in the MG test) of organisms was obtained using image analysis software (SPSS, 2003). Concentration response experiments, to assess the Effect Concentration (ECx), were assessed at generations F6, F10, F26, F34 and F40, always performing additional F0 in parallel, i.e. using organisms from synchronized cultures. 2.4.2. Multigenerational (MG) exposure The standard guideline (OECD, 2009) was followed with adaptations. The maturity of F. candida is usually reached around 15e16 days (6th instar (Snider, 1973),) in control animals, but this may take longer time under exposure conditions. In short, 10 juvenile organisms were selected randomly from cultures of synchronized age (10e12 days) and introduced in each test vessel containing the moist test soil and food supply. Samples were named as Ct_EC10 and Ct_EC50 (for the controls of the respective Cd treatment) and Cd_EC10 and Cd_EC50 (for the Cd treatments). In terms of replicates, four replicates were performed for control and ten replicates per Cd treatments (32 and 60 mg/kg) to ensure enough organisms to continue sequential generations, i.e. 4 replicates of Ct_EC10, 10 replicates of Cd_EC10, 4 replicates of Ct_EC50 and 10 replicates of Cd_EC50. At each test end, after removing the adults (Fx), Fxþ1 juveniles were transferred from the water surface using a small spoon to a recipient with a layer of mixed plaster of Paris and activated charcoal, adsorbing extra existing water. After this, 10 juvenile organisms were selected within similar size (selecting the larger sized animals) and transferred to a new test vessel with soil spiked under the exact same conditions as the previous generation, including the water control. Extra juvenile organisms were snap frozen in liquid nitrogen and kept at 80 C for further analysis. This procedure was repeated over all generations for EC10 and EC50, respectively. A schematic representation of the experimental design is given in Fig. 1. For an overview, the experimental set up included the following
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
3
Fig. 1. Experimental design used for the multigenerational exposure of Folsomia candida to cadmium (Cd), indicating the generations (Fx) and corresponding time in days and years. The population exposed to Cd_EC10 was extinct after ca. 1 year (F13), whereas to Cd_EC50 it continued until termination (F41).
endpoints and samplings per generation (Table S1). 2.5. Methallothionein (MT) gene expression measurements (qPCR) Total RNA of samples from selected generations: F1, F6, F10, F11 (Ct_EC10, Ct_EC50, Cd_EC10 and Cd_EC50) and F13, F14, F24, F25, F28, F32, F33, F34, F35, F36, F38, F39, F41 (Ct_EC50 and Cd_EC50) was extracted using the SV Total RNA Isolation system (Promega). Each sample consisted of approx. 60 juveniles per respective generation. Organisms from the culture with no history of exposure to Cd were sampled and used as reference (F0) to compare the metallothionein-like motif (MTc) expression levels against all other multigeneration samples. Approximately 0.7 mg of each sample total RNA was then converted into cDNA through a reverse transcription reaction using the SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen). The cDNA was 4x diluted and 2 mL were used in 20 mL PCR reaction volumes containing forward and reverse primers and Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen). Primers used for amplification of the MTc gene like (Nakamori et al., 2010) (MTc; GenBank accession number: AB509260) were: 50 -AGCCAATATTTTCGAGTGGAGA-30 (forward) and 50 -CAAGATGCTCGAATAGCAACAGTA-30 (reverse). The reference gene used for normalization was succinate dehydrogenase (SDHA; Collembase (http://www.collembase.org) accession number: Fcc06005) with the following primer sequences: 50 -ACACTTTCCAGCAATGCAGGAG-30 (forward) and 50 -TTTTCAGCCTCAAATCGGCA-30 (reverse). Efficiency and specificity of each primer were determined by observing the obtained standard and melting curves, respectively, for all primer sets. Quantitative real-time PCR (qPCR) was performed in a 7500 Real-Time PCR System (Applied Biosystems) with three biological replicates per treatment, applied in triplicate on a 96-well optical plate (GeneAmp®, Applied Biosystems). Reaction conditions consisted of one initial cycle at 50 C for 2 min, followed by a denaturation step at 95 C for 2 min, 40 cycles at 95 C for 32 s and 1 cycle at 60 C for 1 min. Finally, a dissociation step was made consisting of 15 s at 95 C, 1 min at 60 C, and 15 s at 95 C. A mean normalized expression value was calculated from the obtained Ct values of the test gene with Relative Expression Software Tool (REST-MSC). 2.6. Data analysis Three population parameters were evaluated: survival, reproductive output and animal size distribution. For survival and reproduction, the Effect Concentration (ECx) and its related confidence intervals were calculated using a logistic 2 parameters regression model (Toxicity Relationship Analysis Program (TRAP)
eversion 1.20a, US EPA). Organisms' size distribution was analysed in terms of changes in the average size (mean of fitted normal distribution), the broadness in size distribution (standard deviation), possible shift between small and larger animals’ distribution (skewness), and the shape of the distribution (kurtosis), using the data description option in SigmaPlot software for Windows Version 11.0 (Systat software Inc., San Jose, CA, USA). The populations were checked for clutch effects, i.e. multiple size classes, by bin distribution and using Freedman-Diaconis rule for bins number definition, although such classes could not be found e.g. as in (Marks et al., 2015). The standard deviations as shown in Fig. 3 by the horizontal dashed lines are the standard deviations calculated for all the values plotted in the figure (i.e. for all the generations). The purpose is to show the deviating population. Significance of differences in gene expression between control and treatments were calculated using t-test (p < 0.05). 3. Results 3.1. Multigenerational (MG) exposure 3.1.1. Survival, reproduction, size Results in terms of survival and reproduction along the MG exposure to Cd can be seen for the Cd_EC10 and for the Cd_EC50 in Fig. 2. For both MG exposures, survival and reproduction increased until generation 6e7, after which a decrease was observed for the subsequent 6 generations. As can be observed, the MG exposure to Cd EC10 (32 mg/kg) lead to population extinction after approx. 1 year (F13), whereas the exposure to the EC50 (60 mg/kg) continued until F41, being terminated without being extinct. In the latter case, the population numbers varied along generations 13e41, shifting from minor increase to decrease. Results for MG regarding size data analysis are summarised in Fig. 3. For Cd_EC10, the mean of the size distribution showed a fluctuation over time similar to that of the controls; the size distribution had a tendency to symmetric (y0) to moderate positive skewness and a positive kurtosis, similar to the control, till the time the population collapsed. For Cd_EC50, the mean of the size distribution showed a fluctuation over time, as for the controls, but with the mean being smaller than for both control and EC10 exposed organisms. For the Cd_EC50 animals, there was an initial moderate to high skewness, which moved towards symmetric to moderate negative skewness over time. The kurtosis fluctuated around the mean, but with a higher frequency of positive kurtosis in the beginning and negative
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
4
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Fig. 2. Results in terms of survival and reproduction for the multigenerational exposure of Folsomia candida to cadmium (Cd_EC10: 32 mg Cd/kg and Cd_EC50: 60 mg Cd/kg) and control (Ct_EC10 and Ct_EC50) along 13 and 41 generations (corresponding to generation 12 and 40 in terms of adults) for the EC10 and EC50, respectively. Results are expressed as average ± standard error (Av±SE). Grey e white background delimit the yearly rotation.
towards the end (F29-F41). There was no clear correlation between skewness and kurtosis. The standard deviation shows a fluctuation in variation that is more or less similar along the generations, with
few exceptions of higher and smaller at a particular generation. For guidance (Fig. 4), please note that negative skewness means a higher number of large animals (compared to mean) than small,
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
5
Fig. 3. Size distribution plots including data regarding mean (normalized to start (F0) size as 0) and standard deviation (±average SD represented by the horizontal dashed lines, which refer to all generations data), skewness and kurtosis values for the multigenerational exposure of Folsomia candida to cadmium in the various treatments (Ct_EC10: control; Cd_EC10: 32 mg Cd/kg; Ct_EC50: control; Cd_EC50: 60 mg Cd/kg).
Fig. 4. Schematic illustration of the population distribution for size variation in terms of skewness (measure of symmetry, or lack of it): negative/left (more organisms of large size) and positive/right (more organisms of small size), [normal (no skew): larger amount of organisms of average size]. Kurtosis is a measure of heavy or light tails, also defined as the degree of peakedness, negative indicates broad variation and tends to signal 2 populations, positive indicates narrow variation of 1 population.
positive skewness means higher number of small animals (compared to mean) than large. The right side graph shows an example of the individual size distributions for 3 (out of 728 in total) individual replicates (F0_EC10, F19_EC50, F39_EC50), where the bins and Gaussian distributions are plotted. Hence, the observations were that for EC10 MG exposures there was a tendency towards relatively more small animals with time (positive skew), after which the population collapsed. For EC50 MG exposures, the shift was towards large sized animals (negative skew), without termination.
3.1.2. Effect concentration response tests Results from effect concentration tests performed with the MG exposed animals are summarised in Table 1. Additionally, results from the standard tests (F0) are also shown for reference. For details of the fit to model, please see Fig. 5. For exposures to the EC10 (32 mg/kg), F. candida populations were equally sensitive after 6 MG as the F0 and more sensitive after 10 MG, as can be depicted by the EC50 (significant decrease to approximately half at F10). For exposures to the EC50 (60 mg/kg), a similar increase in sensitivity was measured with the F10 MG organisms. For F26, F34 and F40, the population sensitivity did not vary
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
6
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Table 1 Summary of the Effect Concentrations (ECx, in bold) estimations (logistic 2 parameters) obtained for range finding tests with Folsomia candida exposed to Cd, using organisms directly from the cultures (F0) and previously exposed to cadmium (32 and 60 mg/kg) for 6, 10, 26, 34 and 40 generations (F6, F10, F26, F34 and F40, respectively). n.d.: not determined. NE: No Effect. CI: Confidence Interval. Generation F0 (reference)
EC10
23 6 < CI < 42 MG exposed animals [EC10] ¼ 32 mg Cd/kg F6 28.5 7.4 < CI < 50 F10 n.d. 95 < CI < 35 MG exposed animals [EC50] ¼ 60 mg Cd/kg F6 23 13 < CI < 60 F10 n.d. 43 < CI < 14 F26 n.d. 230 < CI < 85 F34 n.d. 64 < CI < 36 2 F40 31 < CI < 35
EC50
EC90
LC50
68 59 < CI < 78
113 87 < CI < 139
239 150 < CI < 327
79.2 66 < CI < 93 31 10 < CI < 53
130 97 < CI < 162 92 28 < CI < 156
164 105 < CI < 222 240 3442 < CI < 3922
83 61 36 25 59 14 45 27 54 40
143 90 < 87 60 < 191 27 < 104 54 < 106 69 <
261 239 < CI < 283 297 86 < CI < 680 120 108 < CI < 131 NE
< CI < 106 < CI < 47 < CI < 105 < CI < 64 < CI < 68
CI < 196 CI < 113 CI < 355 CI < 155
NE CI < 144
Fig. 5. Results of the range finding tests in terms of survival (left) and reproduction (right) for Folsomia candida exposed to a range of concentrations of CdCl2 (mg Cd/Kg soil DW). Cd_EC10: organisms exposed to 32 mg Cd/kg for x (Fx) generations. Cd_EC50: organisms exposed to 60 mg Cd/kg for x (Fx) generations. Lines represent the fit to model (logistic 2 parameters), dotted line is the simple link between means for tracing.
significantly compared to F0 (similar EC50), but there was a clear increase in the variability (larger confidence intervals). The LC50 showed a tendency to increase (except for F26), which is clear after generation 34 and 40, where no mortality occurs: organisms seem to tolerate higher maximum concentrations after MG exposures to Cd. 3.1.3. Metallothionein (MT) like gene expression The MT like gene qPCR ratios can be seen in Fig. 6, including the direction (up- or down-regulation). MT expression was increased after one Cd exposure cycle (for
both EC10 and EC50). The MT fold change was positively concentration dependent, being nearly twice as high in the animals exposed to the EC50 (average ca. 1200, versus ca. 600 for the EC10 in F6). This upregulation of MT was turned off when organisms were transferred back to control soil. Moreover, the controls show a clear decrease e.g. in the exposure to the EC10_Cd, from F1 to F6. In the exposure to the EC50_Cd, from F24 to F41 there is an upregulation of the MT gene. There is a strong downregulation at F25 control.
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
7
Fig. 6. Results in terms of metallothionein like (MTc) gene expression (qPCR) for the multigenerational cadmium exposed Folsomia candida in the various treatments (Cd_EC10: 32 mg Cd/kg; Ct_EC10: control; Cd_EC50: 60 mg Cd/kg; Ct_EC50: control) at selected generations. Break in the Y scale is between 100 and 250. Dashed line at the 0 level for reference to the control; *: p < 0.05.
4. Discussion Despite the high importance and current increased interest for transgenerational effects, there are still few studies available with invertebrates, and those that do exist mostly report two to eight generations (e.g. Vandegehuchte et al., 2009, 2010; Ward and Robinson, 2005). To our knowledge, no other study has performed such an extensive long term and multigenerational exposure, except studies on the moth beet armyworm (Spodoptera exigua) exposed to Cd or Zn contaminated food over many generations (Kafel et al., 2014, 2012). In the present study, the MG effects for the exposure to the Cd EC10 caused population extinction after one year, whereas the exposure to the Cd EC50 allowed the population to continue to survive and reproduce. As showed by Ward and Robinson (2005), Cd-acclimated daphnids exposed for eight successive generations had a decrease in genetic diversity compared to unexposed animals. Hence, since the organisms are parthenogenetic, the issue here may be a lack of genetic diversity/adaptation to cope with the environmental variable. This could be one of the factors explaining the extinction observed for Cd EC10 MG, although it is also possible that some stochasticity was involved, as the EC50 organisms also reached a minimum for adults around the same generation, but a few generations later for the juveniles. Finally, the experimental setup had a time frame (28 days), hence, a long delay in adulthood may have caused the extinction. This is partly supported by the observation that there was a continuous low number of adults, but high numbers of juveniles in the following generations, indicating a change in developmental speed, growth and reproductive performance and timing. Under field conditions it is likely that sometimes timing (i.e. a delay) is extremely important and will cause
extinction (e.g. changing weather conditions), whereas in other cases a delay may not matter (e.g. due to stable weather conditions). As noted by (Kozlowski and Gawelczyk, 2002) optimum size is the one that ensures maximum reproduction, and this depends on the adults’ potential to survive and reproduce. That the EC10, but not the EC50 exposed animals, became extinct could be due to enhanced adaptive mechanisms at higher exposure concentrations, e.g. through the observed difference in size distribution between EC10 and EC50 exposed animals. Lower Cd levels (EC10) appear to have induced a shift to select organisms of smaller size, whereas comparatively higher Cd levels (EC50) induced a shift to select organisms of larger size (relatively, i.e. within populations of one generation). This trait may have offered an advantage to the EC50 exposed population in terms of survival/ reproduction in the long term. An ant population study reported by (Grzes et al., 2015) found that ants along a pollution gradient (combination of Zn, Cd and Pb) had a skewness in head size distribution significantly correlated with the pollution level of the site, indicating that the frequency of small ants grew as the pollution level increased. The authors discuss that the bias towards the higher frequency of smaller animals could be due to 1) metal toxicity, 2) detoxification costs, 3) shifts in energy allocation, but may also have an adaptive function. The fact that collembolans may come in clutches (although these could not be separated within this study) suggests that the changes in population size distribution may possibly be composed of two phenomenon's e (1) a change in the relative size of the individual organisms due to developmental changes of the individual and (2) a change in the relative numbers between clutches, e.g. due to different sensitivities of the juveniles, but also the parent fitness (Scott-Fordsmand et al., 2000). Regarding metal toxicity (1), several studies have shown the
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
8
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
negative effect of metals on growth, including Cd (see e.g. Fountain and Hopkin, 2001; Ward and Robinson, 2005). The slower growth rate of springtails (Fountain and Hopkin, 2001) occurred at the highest exposure concentrations and was partly the result of reduced food consumption (contaminated food avoidance): EC50 values were approx. 4 times higher for exposure via food compared to soil (Fountain and Hopkin, 2001). Besides epidermal moulting, collembolans also have the ability to excrete metals from the body by intestinal exfoliation, which adds on to their mechanisms. Increased moulting frequency, however, consumes more energy and may result in reduced size as well (Posthuma et al., 1993). Regarding costs (2), detoxification to maintain body metal concentration on a certain level can be costly, e.g. due to the production of metallothioneins (discussed in more detail ahead). As a result, in more polluted sites less energy can be invested in growth. Shifts in energy allocation (3) can take place to maintain the number of organisms under stress, hence, developmental stages might become shorter, resulting in smaller body size. Pollutioninduced reduction of life span has been documented before (e.g. Grzes, 2010). Kozlowski and Gawelczyk (2002) conclude, among others, that size changes are the result of a compromise, where the optimal size for best reproductive success during a life time is prioritized. This means that for example the largest adult animals are not necessarily the most successful. The selected size will be the one that makes the best compromise between optimum reproduction and the optimum survival. Overall, it seems reasonable to assume that the observed size changes of F. candida along MG EC50 exposure is caused by a specific tradeoff with the imposed stress. Cadmium pre-exposed organisms can respond using different mechanisms, e.g. fish have been reported to have decreased uptake of Cd (Xie and Klerks, 2004). It is also known that such mechanisms can be concentration dependent, e.g. for Daphnia magna Hg uptake was reduced, but only at lethal concentrations (Tsui and Wang, 2006). It is hypothesized that animals develop defense mechanisms to reduce the metal uptake when this becomes lethal. This would support the differences observed between EC10 and EC50 found here, where possible defense mechanisms are induced more in the EC50 than in the EC10 exposed organisms. As observed before (Hensbergen et al., 2000; Nakamori et al., 2010), Cd induces the up-regulation of the Metallothionein (MTc) gene, with levels being exposure concentration related: higher Cd concentration results in higher MTc (until a plateau). It is suggested that MT Like Proteins (MTLP) are more important in detoxification at a higher metal pre-exposure or in a more contaminated environment (Tsui and Wang, 2007). Hence, at the tested sub-lethal ECx (EC10 and EC50) there are differences in terms of energy detoxifying costs, e.g. the cost associated with MT gene expression and the related translation onto protein. Such differences in energy allocation may play a role in the follow up mechanisms for the organisms. Contrary to the mechanism described above for fish (Xie and Klerks, 2004), daphnids have been observed to increase the (aqueous) uptake rate of metals after metal pre-exposure (e.g. Cd), while tolerance to metal toxicity increases (Tsui and Wang, 2007). This could seem contradictory, but although more metals are taken up in the pre-exposed animals, this can be partitioned to the already induced MTLP pool, hence, even more metals can be taken up and organisms suffer less than non-exposed organisms. It is not possible to clearly identify whether F. candida developed an adaptation level to Cd in the present experimental time (EC50 exposed); MT measurements seem to indicate acclimation because MT returns to basal levels when organisms are offered non-spiked soil. Nevertheless, the population survival and increased fitness are
due to a given trait. We could be observing slightly deleterious and slightly advantageous mutations, which have small effects on fitness, or deleterious and advantageous mutations, which have larger fitness effects (Lanfear et al., 2014), this will vary depending on the Ne (effective population number). Some suggest that the rate of adaptation is independent of Ne, and that it is determined by the rate of environmental change and number of traits upon which selection acts (Lourenço et al., 2013). Studies with Daphnia longispina from metal-contaminated waters suggest that organisms with adaptation for many generations are better at keeping their tolerances compared to the metalacclimated individuals (Lopes et al., 2006). Historically, contaminated field collembolan populations are also known to develop tolerance, e.g. the well-studied Cd tolerant Orchesella cincta (Roelofs et al., 2007). The mechanisms developed by this species include heritable increase of excretion efficiency, decrease in Cdinduced growth reduction, and over-expression of the MT gene. Moreover, results suggest that besides MT up-regulation, constitutive MT mRNA expression plays an important role in protection against Cd toxicity (Timmermans et al., 2005). Hence, the Cd detoxification mechanism is constitutively overexpressed in tolerant animals and cellular homeostasis is not disturbed upon exposure to Cd (Roelofs et al., 2009). Studies in F. candida (Nota et al., 2008) showed that exposure to Cd activates response mechanisms at various levels, e.g. immune system, hypoxia or detoxification. As quantified in a selection of samples, when exposed F. candida was transferred from Cd spiked soil to control, the MT levels dropped. This was not surprising, but what was intriguing was the larger decrease (negative values) in specific generations (EC10_Ct_F11 and EC50_Ct_F25). This larger MT reduction preceded a high reduction in the reproduction (EC10_Cd_F12/13, EC50_Cd_F26), hence, the mechanisms could be associated with the measured effect. Why this happened and in these particular generations it is unknown. MG exposed populations showed small changes in the reproduction ECx (Fx range finding tests) compared to the F0 standard reference. F10 had increased sensitivity, but there was no continuous increase with the number of generations. In a study with fish, it was demonstrated that the mRNA of MTLP in a Cd-tolerant fish was maternally transferred to the next generation (Lin et al., 2000); vertebrates and invertebrates can vary substantially in such mechanisms. For instance, in a study with D. magna continuously exposed to Hg, the MTLP was not transferred to the next generation (Tsui and Wang, 2006), but there was increased tolerance transference. Although the MG exposed populations showed minor changes in the ECx compared to the F0, there was an increase in the LC50 (e.g. F34, F40). This could indicate an increase in the maximum tolerance limits of Cd for F. candida exposed to Cd EC50 in terms of survival. The study by Kafel et al. (2012) also suggests lower metal tolerance among one-generation exposed insects in comparison with larvae exposed for many generations: the MG organisms had lower accumulation of Cd and increased anti-oxidant response, which could contribute to a greater efficiency in diminishing the toxic effects of the metal. The possibility of transgenerational inheritance of environmentinduced epigenetic changes to non-exposed subsequent generations has been shown for several organisms. For instance, in D. magna this was observed for MG exposure to 5-azacytidine (active pharmaceutical compound known to inhibit DNA methyltransferases) (Vandegehuchte et al., 2010). Interestingly, a concurrent reduction in body length was observed. This could be an indication that size/growth may be a good phenotypical endpoint for transgenerational effects. An experiment with D. magna and exposure to Cd over two generations suggests that Cd does not
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
affect their DNA methylation status (Vandegehuchte et al., 2009), although epigenetics by other mechanisms cannot be excluded. So far, it is unknown whether Cd (or other compounds) causes epigenetics in F. candida, but this would be interesting to investigate. 4.1. Consequences for risk assessment Currently, too little is known in terms of multigenerational effects to be able to include them reliably in hazard and, subsequently, risk predictions. As observed in the present study, up to F6 (6 months) the results could indicate an overall improvement in terms of survival and reproduction for MG exposed organisms, although no changes in the ECx. From F6 to F12, the shift to a decrease in overall performance occurred, this with highest consequences for the EC10 exposure where extinction occurred. The decrease in the EC50 values to about half, as measured with the F10 MG organisms, could be an indication of increased sensitivity, but still the effects after were unpredictable (EC10 crash and EC50 survived). Multigenerational studies with F. candida exposed to phenantrene EC50 showed a significant increase in the effect: complete reproductive failure and subsequent extinction of the population after four generations (Paumen et al., 2008). On the other hand, exposure to lower concentrations caused no effects on survival and reproduction, probably due to biotransformation. Hence, these F. candida studies (Paumen et al., 2008 and our own) show two opposite effects and consequences of MG for RA, 1. phenantrene with lack of adaptation and the all-or-nothing effect (below the threshold, no adaptation, and above this concentration the population went extinct) and 2. Cd with tolerance or adaptation potential (at EC10 no adaptation and population went extinct, at the EC50 adaptation or tolerance and population was maintained). Hence, the studies point to the fact that the effect levels do change, and even when effect levels do not change, the population structure does, which has consequences not only for the population, but also for the community, e.g. the predatory-prey relationship. This clearly supports the notion that the standard toxicity studies, where exposure is performed within one (or less than one) full generation cycle, are not adequate to assess the risk of chemicals. The methodology implemented here for multigenerational exposure worked well and can be used or replicated to test other substances. We recommend its use; the number of generations can be adjusted according to the purpose. Moreover, we recommend the recording of size as an additional endpoint to the standard test for F. candida [not excluding other endpoints in other invertebrate groups], which will not only add in terms of effect level, but also on the potential predictive value. 5. Conclusions Multigenerational exposure of F. candida to Cd showed that long term effects varied with concentration: EC10 exposure caused extinction after 1 year and EC50 continued until >3 years without becoming extinct. The reasons for the differences could be associated with underlying stress response mechanisms. We hypothesize the following, based on the measured endpoints (survival, reproduction, size, metallothionein (MTc) level and ECs monitored along MG): the EC50 exposed organisms responded with higher MTc expression and detoxifying mechanisms (compared to the EC10 exposed), which required more energy and change in strategy. This higher energy cost may be among the reasons for the smaller size of the EC50 exposed animals. The maximum tolerance levels of Cd increased for survival, but not for reproduction (e.g. F40_LC50 > F10_LC50). These changes seem to have favoured the survival of the population.
9
Acknowledgements Thanks are due for the financial support to CESAM (UID/AMB/ 50017), to FCT/MEC through national funds, and the co-funding by FEDER, within the PT2020 Partnership Agreement and Compete 2020. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2016.11.044. References Bicho, R.C., Santos, F.C.F., Gonçalves, M.F.M., Soares, A.M.V.M., Amorim, M.J.B., 2015. Enchytraeid reproduction testplus: hatching, growth and full life cycle testdan optional multi-endpoint test with Enchytraeus crypticus. Ecotoxicology 1053e1063. http://dx.doi.org/10.1007/s10646-015-1445-5. Bickham, J.W., Sandhu, S., Hebert, P.D., Chikhi, L., Athwal, R., 2000. Effects of chemical contaminants on genetic diversity in natural populations: implications for biomonitoring and ecotoxicology. Mutat. Res. 463, 33e51. Campiche, S., L'Ambert, G., Tarradellas, J., Becker-van Slooten, K., 2007. Multigeneration effects of insect growth regulators on the springtail folsomia candida. Ecotoxicol. Environ. Saf. 67, 180e189. http://dx.doi.org/10.1016/ j.ecoenv.2006.11.009. Ernst, G., Kabouw, P., Barth, M., Marx, M.T., Frommholz, U., Royer, S., Friedrich, S., 2016. Assessing the potential for intrinsic recovery in a Collembola twogeneration study: possible implementation in a tiered soil risk assessment approach for plant protection products. Ecotoxicology 25, 1e14. http:// dx.doi.org/10.1007/s10646-015-1560-3. Fountain, M.T., Hopkin, S.P., 2001. Continuous monitoring of Folsomia candida (Insecta: collembola) in a metal exposure test. Ecotoxicol. Environ. Saf. 48, 275e286. http://dx.doi.org/10.1006/eesa.2000.2007. Grzes, I.M., 2010. Zinc tolerance in the ant species myrmica rubra originating from a metal pollution gradient. Eur. J. Soil Biol. 46, 87e90. http://dx.doi.org/10.1016/ j.ejsobi.2009.11.005. Grzes, I.M., Okrutniak, M., Woch, M.W., 2015. Monomorphic ants undergo withincolony morphological changes along the metal-pollution gradient. Environ. Sci. Pollut. Res. 22, 6126e6134. http://dx.doi.org/10.1007/s11356-014-3808-5. Hensbergen, P.J., van Velzen, M.J., Nugroho, R. a, Donker, M.H., van Straalen, N.M., 2000. Metallothionein-bound cadmium in the gut of the insect Orchesella cincta (Collembola) in relation to dietary cadmium exposure. Comp. Biochem. Physiol. C. Toxicol. Pharmacol. 125, 17e24. ska, E., Zawisza-Raszka, A., Migula, P., 2014. The effects Kafel, A., Rozpedek, K., Szulin of cadmium or zinc multigenerational exposure on metal tolerance of Spodoptera exigua (Lepidoptera: noctuidae). Environ. Sci. Pollut. Res. 21, 4705e4715. http://dx.doi.org/10.1007/s11356-013-2409-z. ska, E., 2012. Effects of multigenerational cadKafel, A., Zawisza-Raszka, A., Szulin mium exposure of insects (Spodoptera exigua larvae) on anti-oxidant response in haemolymph and developmental parameters. Environ. Pollut. 162, 8e14. http://dx.doi.org/10.1016/j.envpol.2011.09.034. Kozlowski, J., Gawelczyk, A.T., 2002. Why are species ’ body size distributions usually skewed to the right;? Funct. Ecol. 16, 419e432. Lanfear, R., Kokko, H., Eyre-Walker, A., 2014. Population size and the rate of evolution. Trends Ecol. Evol. 29, 33e41. http://dx.doi.org/10.1016/ j.tree.2013.09.009. Langdon, C.J., Morgan, a J., Charnock, J.M., Semple, K.T., Lowe, C.N., 2009. As-resistance in laboratory-reared F1, F2 and F3 generation offspring of the earthworm Lumbricus rubellus inhabiting an As-contaminated mine soil. Environ. Pollut. 157, 3114e3119. http://dx.doi.org/10.1016/j.envpol.2009.05.027. Langdon, C.J., Piearce, T.G., Meharg, A. a, Semple, K.T., 2003. Inherited resistance to arsenate toxicity in two populations of Lumbricus rubellus. Environ. Toxicol. Chem. 22, 2344e2348. Lin, H., Hsu, S., Hwang, P., 2000. Maternal transfer of cadmium tolerance in larval oreochromis mossambicus. J. Fish. Biol. 57, 239e249. Lock, K., Janssen, C.R., 2001. Tolerance changes of the potworm Enchytraeus albidus after long-term exposure to cadmium. Sci. Total Environ. 280, 79e84. Lock, K., Janssen, C.R., 2002. Multi-generation toxicity of zinc, cadmium, copper and lead to the potworm Enchytraeus albidus. Environ. Pollut. 117, 89e92. http:// dx.doi.org/10.1006/eesa.2002.2155. Lopes, I., Baird, D., Ribeiro, R., 2006. Genetic adaptation to metal stress by natural populations of Daphnia longispina. Ecotoxicol. Env. Saf 63, 275e285. min, S., Galtier, N., 2013. The rate of molecular adaptation in a Lourenço, J.M., Gle changing environment. Mol. Biol. Evol. 30, 1292e1301. http://dx.doi.org/ 10.1093/molbev/mst026. Marks, E.A.N., Molowny-Horas, R., Domene, X., 2015. Investigating the use of multigaussian nonlinear regressions to separate collembolan size classes for soil quality assessment. Water. Air. Soil Pollut. 226 http://dx.doi.org/10.1007/ s11270-015-2503-6. McLaughlin, M., Hamon, R., Parker, D., Pierzynski, G., Smolders, E., Thornton, I.,
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044
10
M.J.B. Amorim et al. / Environmental Pollution xxx (2016) 1e10
Welp, G., 2002. Test methods to determine hazards of sparingly soluble metal compounds in soils. In: Fairbrother, A., Glazebrook, P.W., Tarazona, J.V., S.N. (Eds.), Soil Chemistry. Society of Environmental Toxicology and Chemistry (SETAC) Press, Pensacola, Florida, pp. 5e85. Menezes-Oliveira, V.B., Damgaard, C., Scott-Fordsmand, J.J., Amorim, M.J.B., 2013. Interaction between density and Cu toxicity for Enchytraeus crypticus e comparing first and second generation effects. Sci. Total Environ. 458e460, 361e366. Nakamori, T., Fujimori, A., Kinoshita, K., Ban-nai, T., Kubota, Y., Yoshida, S., 2010. mRNA expression of a cadmium-responsive gene is a sensitive biomarker of cadmium exposure in the soil collembolan Folsomia candida. Environ. Pollut. 158, 1689e1695. http://dx.doi.org/10.1016/j.envpol.2009.11.022. €n, J., De Nota, B., Timmermans, M.J.T.N., Franken, O., Montagne-Wajer, K., Marie Boer, M.E., De Boer, T.E., Ylstra, B., Van Straalen, N.M., Roelofs, D., 2008. Gene expression analysis of collembola in cadmium containing soil. Environ. Sci. Technol. 42, 8152e8157. http://dx.doi.org/10.1021/es801472r. OECD, 2004. OECD Guidelines for the Testing of Chemicals 220 (Enchytraeid Reproduction Test). OECD, 2005. OECD Environment Health and Safety Publications Series on Testing and Assessment No. 50: Detailed Review Paper on Aquatic Arthropods in Life Cycle and Two-generation Toxicity Tests. OECD, 2009. OECD Guidelines for Testing of Chemicals - Collembolan Reproduction Test in Soil. Guideline No.232. Paumen, M.L., Steenbergen, E., Kraak, M.H.S., van Straalen, N.M., Van Gestel, C.A.M., 2008. Multigeneration exposure of the springtail Folsomia candida to phenanthrene: from dose - response relationships to threshold concentrations. Environ. Sci. Technol. 42, 6985e6990. Posthuma, L., Hogervorst, R.F., Joosse, E.N.G., Van Straalen, N.M., 1993. Genetic variation and covariation for characteristics associated with cadmium tolerance in natural populations of the springtail orchesella cincta (L.). Evol. (N. Y) 47, 619e631. €n, J., Roelofs, D., Janssens, T.K.S., Timmermans, M.J.T.N., Nota, B., Marie Bochdanovits, Z., Ylstra, B., Van Straalen, N.M., 2009. Adaptive differences in gene expression associated with heavy metal tolerance in the soil arthropod Orchesella cincta. Mol. Ecol. 18, 3227e3239. http://dx.doi.org/10.1111/j.1365294X.2009.04261.x. €n, J., Van Straalen, N.M., 2007. Differential gene expression profiles Roelofs, D., Marie associated with heavy metal tolerance in the soil insect Orchesella cincta. Insect biochem. Mol. Biol. 37, 287e295. http://dx.doi.org/10.1016/j.ibmb.2006.11.013. Scott-Fordsmand, J.J., Krogh, P.H., Weeks, J.M., 2000. Responses of Folsomia
fimetaria (Collembola: isotomidae) to copper under different soil copper contamination histories in relation to risk assessment. Environ. Toxicol. Chem. 19, 1297e1303. €fers, C., Segner, H., Caroll, K., Fenske, M., Janssen, C.R., Maack, G., Pascoe, D., Scha Vandenbergh, G.F., Watts, M., Wenzel, A., 2003. Identification of endocrinedisrupting effects in aquatic vertebrates and invertebrates: report from the European IDEA project. Ecotoxicol. Environ. Saf. 54, 302e314. http://dx.doi.org/ 10.1016/S0147-6513(02)00039-8. Snider, R., 1973. Laboratory observations on the biology of Folsomia candida (Wil lem) (collembola: isotomidae. Rev. D’Ecologie Biol. du Sol. 10, 103e124. SPSS, 2003. SigmaScan Pro 5dImage Analysis. SPSS for Windows 12.0. Spurgeon, D.J., Hopkin, S.P., 2000. The development of genetically inherited resistance to zinc in laboratory-selected generations of the earthworm Eisenia fetida. Environ. Pollut. 109, 193e201. Timmermans, M.J.T.N., Ellers, J., Roelofs, D., van Straalen, N.M., 2005. Metallothionein mRNA expression and cadmium tolerance in metal-stressed and reference populations of the springtail orchesella cincta. Ecotoxicology 14, 727e739. http://dx.doi.org/10.1007/s10646-005-0020-x. Tsui, M., Wang, W.-X., 2006. Acute toxicity of mercury to daphnia magna under different conditions. Env. Sci. Technol. 40, 4025e4030. Tsui, M.T.-K., Wang, W.-X., 2007. Biokinetics and tolerance development of toxic metals in Daphnia magna. Environ. Toxicol. Chem. 26, 1023e1032. http:// dx.doi.org/10.1897/06-430r.1. van Gestel, C. a. M., Mol, S., 2003. The influence of soil characteristics on cadmium toxicity for Folsomia candida (Collembola: isotomidae). Pedobiol. (Jena) 47, 387e395. http://dx.doi.org/10.1078/0031-4056-00202. Vandegehuchte, M., Kyndt, T., Vanholme, B., Haegeman, A., Gheysen, G., Janssen, C., 2009. Occurrence of DNA methylation in daphnia magna and influence of multigeneration Cd exposure. Environ. Int. 35, 700e706. http://dx.doi.org/ 10.1016/j.envint.2009.01.002. re, F., Vanhaecke, L., Vanden Berghe, W., Janssen, C.R., Vandegehuchte, M.B., Lemie 2010. Direct and transgenerational impact on daphnia magna of chemicals with a known effect on DNA methylation. Comp. Biochem. Physiol. C. Toxicol. Pharmacol. 151, 278e285. http://dx.doi.org/10.1016/j.cbpc.2009.11.007. Ward, T.J., Robinson, W.E., 2005. Evolution of cadmium resistance in Daphnia magna. Environ. Toxicol. Chem. 24, 2341e2349. Xie, L., Klerks, P., 2004. Changes in cadmium accumulation as a mechanism for cadmium resistance in the least killifish Heterandria fomosa. Aquat. Toxicol. 66, 73e81.
Please cite this article in press as: Amorim, M.J.B., et al., Does long term low impact stress cause population extinction?, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.11.044