Environ. Sci. Technol. 2005, 39, 9001-9008
Development of Estrogen-Responsive Transgenic Medaka for Environmental Monitoring of Endocrine Disrupters ZHIQIANG ZENG, TAO SHAN, YAN TONG, SIEW HONG LAM, AND ZHIYUAN GONG* Department of Biological Sciences, National University of Singapore, Singapore
To develop a transgenic fish system to monitor environmental pollution, we generated a mvtg1:gfp transgenic medaka line, in which the gfp reporter gene was under the control of medaka vitellogenin1 (mvtg1) gene promoter. In this transgenic line, GFP was exclusively expressed in the liver of the mature adult female. Male and juvenile transgenic fish did not express GFP but could be induced to express GFP in the liver after exposure to 17-β-estradiol (E2). Concurrent accumulation of mvtg1 and gfp mRNAs was observed during both development and estrogen treatment, indicating that the gfp transgene was faithfully expressed under the mvtg1 promoter. Dose- and time-dependent induction of GFP expression by E2 was investigated in male transgenic fish. The lowest-observed-effect concentration (LOEC) of E2 to induce GFP expression was 0.5 µg/L by observation of live fish and 0.05-0.1 µg/L by observation of dissectionexposed liver in a 30 day exposure experiment. GFP expression was observed within 36 h after treatment in high concentrations of E2 (5 µg/L), and it took longer to detect GFP expression under lower concentrations of E2. By removal and readdition of E2, we demonstrated that GFP expression was repeatedly induced. Finally, we also demonstrated that GFP expression could be induced by other estrogenic compounds, including 17-R-ethynylestradiol (EE2, 0.05 µg/L), diethylstibestrol (DES, 5 µg/L), estriol (10 µg/ L), and bisphenol A (BPA, 1 mg/L), but not by weak estrogenic chemicals such as nonylphenol (NP, up to 1 mg/ L) and methoxychlor (MXC, up to 20 µg/L). Our experiments indicated the broad application of the transgenic line to monitor a wide range of estrogenic chemicals.
Introduction Natural and synthetic estrogenic compounds derived from industrial wastes, pesticides, and sewage pollute aquatic environments and cause endocrine and reproductive impairments, termed endocrine disruption, in humans and wildlife (1, 2). The possibility of adverse health effects of endocrine disruption has received considerable public attention (3, 4). As many pollutants released into the environment eventually reach bodies of water, they are concentrated in the tissues of aquatic organisms including fish (5). The use of fish is a popular approach to monitor and assess risks of exposure to chemicals in aquatic environments (6-8), as * Corresponding author phone: 65-68742860; fax: 65-67792486; e-mail:
[email protected]. 10.1021/es050728l CCC: $30.25 Published on Web 10/06/2005
2005 American Chemical Society
the fish endocrine system shares many similarities with those of higher vertebrates including humans (9). It has been known that endocrine disrupters, which can cause intersexuality and reproduction decline in fish, can also induce synthesis of female-specific proteins such as vitellogenin (Vtg) and zona radiata proteins (10). Vtg is a precursor of egg yolk protein and is normally synthesized only in the liver of adult females in oviparous vertebrates under the control of the female hormone. Vtg and vtg mRNA expression can also be induced in juvenile and male fish by estrogen or xenoestrogenic compounds. Thus, induction of Vtg in vivo has been widely accepted as an indicator for monitoring environmental estrogenic compounds (11-14). In the past few years, our laboratory has generated several gfp transgenic zebrafish (Danio rerio) lines using tissuespecific promoters. Expression of the gfp reporter gene can be easily monitored in live fish either under a fluorescent microscope or by direct observation with the naked eye (15, 16). The inducibility of Vtg expression by estrogenic compounds, as well as wide acceptance of using fish for endocrine disruption studies, prompted us to develop a gfp transgenic fish under an estrogen-inducible promoter from a vtg gene. Although the zebrafish has been widely used for transgenic research (17, 18), we chose medaka for this work because the medaka has excellent background in fish toxicological studies. The medaka is a small, well-characterized laboratory model fish sharing essentially all the advantageous features attributed to the zebrafish (19). In addition, there are several other features that make the medaka a more ideal species for monitoring endocrine disrupters. First, its sex is clearly determined by genetics or XY chromosomes and can be readily distinguished from secondary sex characteristics. Second, we previously observed that the medaka is more sensitive to respond to estrogen exposure than zebrafish; the LOEC of E2 for vtg mRNA induction is 0.1 µg/L in medaka compared with 1 µg/L in zebrafish (20). Third, the medaka is more transparent, and GFP expression in internal organs can be more sensitively detected in the medaka than in the zebrafish. Finally, the medaka is a hardier fish and can adapt to wide range of salinity and temperature. Thus, in the present study, we have developed a gfp transgenic medaka line under an estrogen-inducible promoter derived from the medaka vitellogenin 1 gene, mvtg1 (20). We demonstrated that transgenic GFP is faithfully expressed in female livers and in estrogen-induced male livers, perfectly mimicking the expression pattern of the endogenous mvtg1. More interestingly, GFP expression can be induced by several other estrogenic compounds; thus, there is a potential to use the transgenic line to develop a biomonitoring system for control of environmental pollution.
Materials and Methods Fish. The medaka fish (orange-red strain) stock was initially obtained from the medaka stock center of Nagoya University, Nagoya, Japan. Fish were maintained under an artificial photoperiod of 14 h of light and 10 h of darkness at ambient temperature (∼27 °C). Staging of medaka embryos and fry was based on Iwamatsu (21). Isolation of Medaka mvtg1 Promoter and Construction of pMVTG1-EGFP Plasmid. An mvtg1 genomic clone with an insert of ∼14 kb was isolated from a medaka genomic DNA library purchased from Stratagene (U.S.A.). The 5′ flanking region of about 4500 bp was completely sequenced, and a 3005 bp fragment of the proximal promoter region was amplified by PCR and cloned into the pEGFP-1 vector (BD VOL. 39, NO. 22, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9001
Biosciences Clontech, U.S.A.) to construct the chimeric plasmid, pMVTG1-EGFP. Microinjection. Eggs were collected within 30 min of fertilization. To prevent the egg chorion from hardening after fertilization, the eggs were incubated in distilled water containing 2 mM glutathione at 6 °C for up to 2 h. As the mvtg1 promoter is not functional in early embryos, plasmid pCK-RFP, which contains a zebrafish skin epithelium-specific promoter with a red fluorescent protein gene (22), was coinjected with pMVTG1-EGFP in order to monitor the success of microinjection and subsequent transgenic screening. Both pMVTG1-EGFP and pCK-RFP plasmids were linearized and injected at final concentrations of 200 and 50 ng/µl, respectively. Microinjection was performed at the 1-2 cell stage, and each egg was injected with ∼2 nl of mixed DNA plasmids. Injected eggs were incubated at 26 °C in the embryo-rearing medium according to the medaka handbook (23). Under a fluorescence microscope, injected embryos were examined for RFP expression, and these expressing embryos were retained and raised to adult. Transgenic Fish Screening. Sexually mature founder fish were individually mated with wild-type fish. PCR was adopted to examine the existence of the gfp transgene from pooled embryos. Of the 21 founders screened, one germlinetransmitted transgenic founder was identified. F1 fish fry obtained from the positive founder were raised and further screened by PCR using DNA from cut tail fins. Finally, the mvtg1:gfp transgenic line was confirmed by 50% inheritance in the F2 generation. Observation of GFP Expression by Fluorescent Microscopes. Two types of observation of GFP expression were performed in the present study: live fish observation and exposed liver observation. For live fish observation, fish were anaesthetized by 0.1% 2-phenoxyethanol (Sigma) and observed on a Petri dish under a fluorescence stereomicroscope (LEICA MZ12) or an inverted fluorescence microscope (ZEISS Axiovert 25) with the Zeiss filter set 09 (excitation: 450 to ∼490 nm). Fish were returned to tanks after observation to continue exposure. For exposed liver observation, fish were slit open in the abdomen and the liver observed under the same inverted fluorescent microscope. For estimation of GFP expression level, GFP expression was recorded by photography at a fixed exposure (1.532 s) using a MagnaFire SP imaging system (OPTRONICS, CA). Expression levels were estimated and graded according to the intensity of green fluorescence (see Figures 3 and 4). For observation of GFP expression in cryosections, E2-treated transgenic juvenile fish were fixed in 4% PFA/PBS for 1 h and mounted in tissuefreezing medium. The whole juvenile fish were sectioned at 10 µm thickness using a cryostat (Leica CM1900). Cryosections were mounted with 50% glycerol/PBS and coverslides. Observation of GFP on glass slides was performed under a fluorescent microscope (ZEISS, Axioskop2 mot plus) equipped with the AxioVision 4.0 (Carl Zeiss Vision GmbH) digital imaging system. Chemical Exposure Tests. Chemicals used in this study were 17-β-estradiol (E2), 17-R-ethynylestradiol (EE2), diethylstilbestrol (DES), estriol, bisphenol A (BPA), methoxychlor (MXC), 11-ketotestosterone (11-KT), and trenbolone (TB) from Sigma-Aldrich Company; and 4-n-nonylphenol (NP) from Lancaster Synthesis. All stock solutions were prepared in 100% ethanol and stored at -20 °C. The appropriate volume of the stock preparation was added to dechlorinated tap water; the volume of ethanol (solvent vehicle) did not exceed 1/20 000 of the total volume. The control group received vehicle only. The chemical-containing water was changed every 2 days throughout the experiment. Real-Time RT-PCR for Quantification of mvtg1 and gfp mRNAs. Quantification of mvtg1 and gfp mRNAs was conducted by real-time RT-PCR using a LightCycler RNA 9002
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005
amplification kit SYBR Green I (Roche Applied Science, Germany) as previously described by Tong et al. (20). Total RNAs were extracted from whole fish by using TRIzol reagent (Gibco-BRL). Each real-time PCR run generally consisted of a set of 10-fold serial dilution of mvtg1 or gfp positive control RNA samples for construction of a standard curve, test samples, and negative control (template-free). The positive RNA samples were isolated from spawning female fish and were also used to normalize the variation among different runs. The PCR products were also analyzed by agarose gel electrophoresis, and a single band with the predicted size of 127 bp (mvtg1 amplicon) or 208 bp (gfp amplicon) were observed (data not shown).
Results Generation of mvtg1:gfp Transgenic Medaka. Linearized pMVTG1-EGFP was injected into medaka eggs, and the injected embryos were raised to adult stage. Sexually mature founder fish were crossed with wild-type fish to obtain F1 transgenic embryos. Genomic DNA was extracted from pools of F1 embryos and screened by PCR. Of 21 founder fish screened, one fish was identified to be capable of germline transmission of the transgenes. The transmission rate was ∼25% (n ) 84) from founder to F1. From F1 to subsequent generations, standard Mendelian ratios of transgene inheritance were always observed, suggesting a single integration of the transgene. Expression of GFP in mvtg1:gfp Transgenic Medaka. In mvtg1:gfp transgenic fish, GFP expression was detected only in the liver of spawning females but not of males (Figure 1A). Dissection of male and female fish confirmed that GFP expression was specifically in female liver but not in male liver (Figure 1B). Although there was no liver expression of GFP in immature fish, ectopic GFP expression was found in two small symmetrical tissues dorsal-posterior to the gills on both sides (Figure 1C). There was no indication of endogenous expression of mvtg1 as investigated by both RTPCR and in situ hybridization assays (data not shown). The reason for the ectopic expression is unknown and may be related to the transgene integration site. However, the ectopic expression is beneficial for early selection of transgenic individuals as the onset of the ectopic expression occurred only 10 days after hatching and in 100% of transgenic fry irrespective of their sex. In comparison, visible GFP expression in female liver only occurred about 1 week prior to spawning, or about 10-11 weeks after hatching. In male transgenic fish, there was no detectable GFP expression, but like endogenous mvtg1 mRNAs (20), GFP expression was induced in male liver by immersion treatment of 17-β-estradiol (E2) (Figure 1D). Thus, transgenic gfp expression behaved like the endogenous mvtg1 and can be induced by estrogens. To determine the earliest induction of GFP expression, transgenic embryos were treated with 5 µg/L of E2 before liver formation (stage 28, 64 hpf at 25 °C). Induced GFP expression was first observed in embryonic liver 1 day after hatching when liver morphogenesis was completed (∼10 days postfertilization). In the second day after hatching, GFP expression in the liver was significantly increased in most embryos (Figure 1E). Tissue sections of E2-induced juvenile fish further indicated that GFP expression was limited to the liver only (Figure 1, parts F and G). Correlation of mvtg1 and gfp mRNA Expression in mvtg1: gfp Transgenic Medaka during Development and E2 Induction. To further characterize transgenic GFP expression in comparison with the endogenous mvtg1 mRNA expression, real-time RT-PCR was performed to quantify the relative levels of mvtg1 and gfp mRNAs during development. As shown in Figure 2A, five fish were sampled each week from week 7 to week 10 after hatching. A concurrent accumulation of gfp and mvtg1 mRNAs was observed. Both mvtg1 and gfp mRNAs
FIGURE 1. GFP expression in mvtg1:gfp transgenic medaka. (A) Live fish observation of GFP expression in mvtg1:gfp transgenic male (top) and female (bottom) adults. (B) Liver observation of GFP expression in abdomen-opened mvtg1:gfp transgenic male (top) and female (bottom) adults. (C) Comparison of transgenic (top) and nontransgenic (bottom) juvenile fish (1 month old). Ectopic GFP expression dorsal-posterior to the gills is boxed and enlarged as an inset. (D) Induction of GFP expression by 1 µg/L E2 in a transgenic male (bottom) compared to that of an untreated control male (top). (E) Precocious induction of GFP expression in transgenic fry (bottom, 2 days after hatching) by 5 µg/L E2 in comparison with that of an untreated transgenic control fry (top). Positions of livers are indicated with asterisks. (F and G) Liver-specific induction of GFP expression by E2. A cryosection of a 1 month old juvenile fish after treatment with 1 µg/L of E2 for 2 days was viewed under a bright light (F) and a blue light for GFP observation (G). Scale bars are indicated. Abbreviations: G, gonad; I, intestine; K, kidney; L, liver. began to accumulate in some fish at week 9, when most of the fish had yet to show secondary sexual characteristics. By week 10, the sex of the fish was clearly identifiable and all females showed increased expression of mvtg1 and gfp mRNAs although there was no observable GFP fluorescence until week 11. To compare the time course of induced gfp mRNA with that of endogenous mvtg1 mRNA, real-time RT-PCR was also performed with 3 month old transgenic adult males treated with 1 µg/L E2. As shown in Figure 2B, endogenous mvtg1 mRNA was induced by E2 as early as 12 h after exposure and reached a peak at day 7 with a level as high as 1.9-fold that of untreated transgenic females. Interestingly, a very similar induction curve was also observed for gfp mRNA. Thus, these experiments demonstrated that transgenic gfp was faithfully expressed under the mvtg1 promoter. Induction of GFP Expression in Male Transgenic Fish by Exogenous E2. As GFP expression can be induced from
nonexpression background of male transgenic fish, it is interesting to use the male fish to develop a biomonitoring system for estrogenic compounds. To explore this potential, E2 induction of GFP expression in male fish was characterized by examining the dose effect, time course, and repeated induction. First, the dose effect was examined. In this experiment, 3 month old F2 transgenic male fish were exposed to E2 for 30 days at each of the following concentrations: 0, 0.05, 0.1, 0.5, and 1 µg/L. GFP expression was examined every 1 or 2 days by observation of live and anaesthetized fish under a fluorescence microscope. To achieve a semiquantitative assessment, GFP expression in individual fish was arbitrarily graded into five levels from 0 (no detectable GFP expression) to 4 as shown in Figure 3A-D and described in the legend of Figure 3. The averaged GFP expression levels of each concentration group (10 fish) are presented in Figure 3E. There was no visible induction of GFP expression throughout VOL. 39, NO. 22, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9003
FIGURE 2. Concurrent expression of endogenous mvtg1 mRNA and transgenic gfp mRNA. (A) Developmental expression of endogenous mvtg1 and transgenic gfp mRNAs. Five fish were sampled each week from 7 to 10 weeks posthatching, and total RNAs were isolated from the whole fish. Both mvtg1 and gfp mRNAs were quantified by real-time RT-PCR. In weeks 7-9, fish have yet to show clear secondary sexual characteristics and fish were sampled randomly. In week 10, the sex of all fish was obvious and only female fish were sampled. (B) Comparison of the induced gfp and mvtg1 mRNAs expression. Male fish were treated with 1 µg/L E2, and four fish were sampled at each time point. Induced gfp mRNA and endogenous mvtg1 mRNAs were quantified by real-time RT-PCR, with the RNA from untreated transgenic female (a pool of 10 fish) as control. The value for each time point is the mean value of four individual fish. the 30 day test with 0.05 and 0.1 µg/L of E2. In the 0.5 µg/L group, initial GFP expression was observed on day 6. GFP expression was gradually increased in the next few days with a peak on day 18, followed by a slight decline toward to the end of 30 day exposure. A similar trend of GFP expression
was also observed in the 1 µg/L group, but GFP expression first appeared 1 day earlier and was also stronger in the first few days. Treatment with a higher concentration of E2 caused even earlier induction of GFP expression (see Figure 5). An essentially identical observation was also made in the F3 transgenic male after treatment with different concentrations of E2 (data not shown). Thus, there was a dose-dependent induction of GFP expression and a sensitivity to E2 induction at 0.5 µg/L by observation of GFP expression in live fish. As thick tissue may compromise the detection limit of GFP expression in live fish, we also carried out another set of experiments by examining GFP expression in the liver after opening the abdomen at the end of the 30 day E2 exposure to determine if the detection sensitivity could be improved. The level of GFP expression was estimated based on GFP fluorescence intensity in the exposed livers (Figure 4A-E). As summarized in Figure 4F, while 0.05 µg/L E2 did not induce clearly detectable GFP expression in the exposed liver, GFP expression could be seen in most fish (four out of five) in the 0.1 µg/L group. Treatment with 0.2 µg/L or higher E2 caused 100% of the male fish to express GFP at an even higher level. Thus, the sensitivity of induction of GFP expression (0.05-0.1 µg/L E2) is the same as induction of mvtg1 mRNAs as we previously determined (20). The rate of GFP expression after removal of E2 and the feasibility of reinduction of GFP expression were also investigated (Figure 5). In this experiment, male transgenic fish were first treated with 5 µg/L E2. Initial GFP expression could be detected 36 h after exposure. GFP expression was induced to level 3 (see Figure 3C) in all fish by day 6, and these treated fish were then transferred to estrogen-free water. The brightness of the GFP fluorescence remained unchanged for the next 3 days, then slightly decreased, and completely disappeared by day 32 after removal of E2. After another week, these fish were then again treated with 5 µg/L E2. The second induced GFP expression was first detected as soon as 29 h after treatment, and a level 3 GFP expression was achieved in all fish by day 5. Thus, it seems that these fish become more sensitive in response to the second E2 treatment. This experiment also indicated the potential that
FIGURE 3. E2 sensitivity and the time course of induction of GFP expression in male transgenic fish as observed in live fish. Male transgenic fish were treated with E2 at different concentrations (0.05, 0.1, 0.5, and 1.0 µg/L) for 30 days. Ten male fish were used in each concentration. GFP expression in each fish was examined under a fluorescent microscope, and its expression level was recorded by photographing at a fixed exposure time setting. GFP expression levels of the fish were estimated according to the following arbitrary standards: level 1 (+), initial appearance of GFP fluorescence just behind the gill cover and below the pectoral fin (A); level 2 (+ +), GFP fluorescence expanded to a larger area (B); level 3 (+ + +), GFP fluorescence spread over the whole liver region, fluorescence intensity has also increased (C); level 4 (+ + + +), further increase of GFP fluorescent intensity leading to saturated exposure under our standard photograph setting (D). The time course of GFP induction was computed by averaging the levels from the four fish at each time point (E). Different E2 concentrations are represented by different symbols as indicated in the diagram. Dosages of 0.05 and 0.1 µg/L did not induce detectable GFP by observation of live fish, and the values with 0.05 µg/L (closed diamonds) are overshadowed by the values with the 0.1 µg/L (open circles). Scale bars, 1 mm. Abbreviations in Panes A-D: GC, gill cover; PF, pectoral fin. 9004
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005
FIGURE 4. Sensitivity of GFP induction in response to E2 treatments as observed in dissection-exposed liver of transgenic male fish. Five transgenic male fish were in each treatment group and were exposed to E2 at a particular concentration (0.05, 0.1, 0.2, 0.3, 0.5, or 1.0 µg/l) for 30 days. Fish were then dissected to expose the liver, and GFP expression in the liver was examined under a fluorescence microscope and recorded by photographing at a fixed exposure time. Expression levels were estimated by the following arbitrary standards: level 1 (+), GFP fluorescence in less than 50% of liver cells (A); level 2 (+ +), GFP fluorescence in more than 50% of liver cells (B); level 3 (+ + +), GFP fluorescence in the whole liver (near 100% of cells) (C); level 4 (+ + + +), GFP fluorescence showing saturated exposure (white color) at the fixed exposure time setting (1.53 s) (D); level 5 (+ + + + +), GFP fluorescence showing overwhelmingly saturated exposure in nearly the whole liver (E). (F) Summary of GFP expression levels in various concentrations of E2. Each bar represents an individual fish. All of the five fish in the 0.05 µg/L group and one of the fish in the 0.1 µg/L group did not show clear GFP expression. Only three fish survived the 30 day experiment in the 0.2 µg/L group.
FIGURE 5. Reinduction of GFP expression by a second E2 treatment. Ten transgenic male fish were treated with 5 µg/L E2, and GFP expressions were induced to level 3 in all of the fish by day 6. Thereafter, all the fish were transferred into E2-free water on day 8 (or day 0 for E2-free water). GFP expression was gradually decreased and was completely invisible by day 32. To ensure that GFP extinguished completely, these fish were kept in estrogenfree water for another week and then treated with second dose of E2 at 5 µg/L. GFP expression was continuously monitored daily. The level of GFP expression was estimated as described in Figure 3. The first E2 exposure and removal is represented by triangles and the second E2 exposure by squares. the transgenic fish can be used repeatedly in monitoring E2 compounds. Induction of GFP Eexpression by Other Estrogenic Chemicals. To test the applicability of mvtg1:gfp transgenic medaka to monitor environmental estrogens, several related estrogenic chemicals were also used to treat transgenic male fish, including 17-R-ethynylestradiol (EE2), estriol, diethylstilbestrol (DES), bisphenol A (BPA), 4-n-nonylphenol (NP), and methoxychlor (MXC). These chemicals have been implicated as environmental endocrine disrupters because of their strong or weak estrogenic effects in the environment (2). For each chemical, preliminary tests were performed to determine sublethal concentrations. Then several different
concentrations below sublethality were used to treat transgenic male fish. As summarized in Table 1, GFP expression could be strongly induced in transgenic males treated with both natural (E2 and estriol) and synthetic estrogens (EE2 and DES). In general, there was a dose-dependent increase of GFP signal and the detection of GFP expression in exposed liver was more sensitive than in live fish. Furthermore, visible GFP induction was observed by treatment with a weak estrogenic compound, BPA. However, no visible GFP induction was observed for the other two weak estrogenic compounds, NP and MXC. Endogenous mvtg1 mRNAs were also determined by real-time RT-PCR. Consistent with the observation of GFP induction, in all cases where GFP was induced after treatment with E2, EE2, estriol, DES, and BPA, mvtg1 mRNA were also induced. While GFP was not induced by treatment with NP and MXC, no endogenous mvtg1 mRNA was detected (Table 1). Thus, the lack of GFP induction was not due to the decrease of induction sensitivity in the transgenic system but to the limitation of induction of the Vtg biomarker. Further increase of the concentration of these chemicals generally caused acute lethality. Over 80% of the fish died within 7 days. The relative estrogenic potency of these compounds was also estimated based on their initial appearance of GFP expression at the lowest concentrations used. As shown in Table 1, the estrogenic potency is basically about the same for E2, EE2, and DES, while estriol is about 10 times lower and BPA is about 5000 times lower. These relative estrogenic potencies are generally consistent with those reported in previous literature (e.g., 24-27).
Discussion Estrogenic pollutants, which cause endocrine disruption and reproduction impairment, are now increasingly becoming a VOL. 39, NO. 22, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9005
TABLE 1. Induction of GFP Expression in mvtg1:gfp Transgenic Medaka by Estrogenic Compoundsa GFP fluorescence chemical concn (µg/L)
live fish observation
exposed liver observation
mvtg1 mRNAb
E2 0.05 0.1 1.0
(0.18 nM) (0.37 nM) (3.67 nM)
++
+/+ +++++
2.0 × 10-3 2.2 × 10-1 9.7 × 10-1
EE2 0.05 0.1 1.0
(0.17 nN) (0.34 nM) (3.38 nM)
++++
+ +++ +++++
2.0 × 10-3 6.0 × 10-3 1.8 × 10-1
Estriol 0.5 1 10
(1.74 nM) (3.47 nM) (34.7 nM)
+++
+++++
0 1.4 × 10-2 4.5 × 10-2
DES 5
(18.7 nM)
+++
+++++
3.7 × 10-1
BPA 50 200 1000 5000
(0.22 µM) (0.88 µM) (4.39 µM) (21.9 µM))
++
++ +++++
0 7.0 × 10-3 8.4 × 10-3 4.2 × 10-1
NP 50 100 500 1000
(0.23 µM) (0.45 µM) (2.27 µM) (4.55 µM)
-
-
0 0 0 0
MXC 5 10 20
(14.5 nM) (28.9 nM) (57.8 nM)
-
-
0 0 0
E2 potencyc 1
∼2
∼0.1
∼0.3 ∼0.0002
a Method: For each concentration of chemicals, five homozygous F3 transgenic male fish were used. The duration of treatment was 21 days. For live fish GFP observation, fish were examined daily. At the end of day 21, fish were sacrificed for liver GFP observation. GFP fluorescence levels were recorded as described in Figures 3 and 4. After liver GFP observation, whole fish were used for RNA extraction to determine mvtg1 mRNA by real-time RT-PCR. b The level of mvtg1 mRNA is relative to that in pooled female control samples; “0” indicates undetectable. c E2 potency is relative to E2 (0.1 µg/mL) as 1 and was estimated based on the initial appearance of GFP expression at the lowest concentration from exposed liver observation.
public concern. A variety of in vitro and in vivo test methods have been developed for assessing the estrogenic potency of chemicals (28, 29). Most of the in vitro methods are cellbased, such as receptor binding assays, cell proliferation assays, and reporter gene assays (30-33). Although being relatively simple, high-throughput, sensitive, and quantifiable, they are generally designed to evaluate the estrogenic potency of unknown chemicals in laboratories and are not suitable for environmental monitoring. In contrast, in vivo assays are based on physiological effects of chemicals on target organs, such as the classical uterotrophic assay and vaginal cornification assay. Other in vivo assays, such as intersexuality of roach fish and induction of Vtg in male rainbow trout, have been used to assess the adverse effects of estrogenic compounds in the field (8). However, these in vivo assays are generally labor-intensive, expensive, and lowthroughput. Thus, in the present study, we developed an estrogen-inducible gfp transgenic medaka line, and this transgenic fish should combine some of the advantages of both in vitro and in vivo assays, such as simplicity, highthroughput, in vivo response, etc. More importantly, the transgenic line has a potential to develop an on-line and on-site monitoring system for endocrine disrupters in netted aquatic environments such as rivers and lakes. Previously it has been reported that in medaka the LOEC of E2 varies from 0.01 to 0.1 µg/L by measuring Vtg protein or mRNA (13, 20, 34). In our current study, by observing induced GFP expression in our transgenic line, we found that the LOEC of E2 for induction of visible GFP expression was essentially the same as the assay for Vtg protein or mRNA; 9006
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005
thus, our transgenic line retains the same level of sensitivity. The LOEC of 0.1 µg/L determined by observation of GFP expression in the present study is actually within the range of environmental E2 concentrations of 0.027-0.24 µg/L in sewage effluent, which was extrapolated from Vtg induction studies in fish (34). It is worth noting that in all experiments described in this report, we used only hemizygous transgenic fish for the tests. By using homozygous transgenic fish by the E2 test, we found that the LOEC of E2 can be improved to 0.05 µg/L by observation of exposed liver (Table 1). The detection of GFP expression in live fish observation may be further improved by using transparent medaka strains. The most promising example is the see-through medaka that is completely transparent, and its internal organs including the liver, spleen, gut and gonad can be easily viewed externally (35). In addition, the responsive sensitivity may also be improved by adding more estrogen responsive elements (EREs) to the mvtg1 promoter. For example, Catherino and Jordan (36) reported that increasing the number of EREs to three tandem copies in the promoter resulted in up to a 15-fold increase in reporter gene transcription. To further demonstrate the application of the transgenic line in environmental monitoring, we also tested the transgenic fish with several compounds proven or implicated with estrogenic activities, including EE2, DES, estriol, BPA, NP, and MXC. EE2 is a synthetic estrogenic steroid widely used in oral contraceptive pills (37), and DES is a pharmaceutical estrogen that was heavily prescribed to prevent miscarriage and other pregnancy complications from the late1940s to the 1970s before it was banned due to its carcinogenicity
(38). Both chemicals have been proven in fish, including medaka, to induce Vtg or vtg mRNAs (24, 39, 40). Consistent with these reports, we found that both transgenic GFP and endogenous mvtg1 mRNA can be readily induced by the two chemicals with potencies similar to that of E2 in our mvtg1: gfp transgenic medaka (Table 1). Estriol is a weak natural estrogen that is found in the urine during pregnancy (41) and has been reported to induce medaka Vtg, albeit 7 times lower than E2 (27). Similarly in this study, estriol was able to induce GFP expression in our transgenic line at about 10% of the E2 potency (Table 1). BPA and NP are industrial intermediates, and their estrogenic activities have been proven in the laboratory (8), while MXC is a pesticide and has been implicated with estrogenic activity (42, 43). All of these three weak estrogenic compounds have been tested in medaka and/or other fish to induce Vtg by long-term immersion exposure (g5 weeks) (44-46). However, with our transgenic line, we could observe the induction of transgenic GFP and endogenous mvtg1 mRNA only with BPA but not NP and MXC, even if we applied a higher concentration of NP than those reported. One possibility is that longer exposure time is required for the weaker estrogenic compounds to induce visible GFP fluorescence. Consistent with this, we have observed that the appearance of induced GFP fluorescence is usually delayed by exposure to lower concentrations of E2 (Figure 3). The current mvtg1:gfp transgenic line was designed to monitor female estrogenic compounds by using non-GFPexpressing male fish. We also explored the possibility of monitoring male androgenic compounds by observation of GFP expression in transgenic females treated with 5 µg/L 11-ketotestosterone (11-KT, natural androgen) or 5 µg/L trenbolone (TB, synthetic androgen and known environmental pollutant) (47, 48). After a 30 day exposure, the intensity of GFP fluorescence did not show an obvious decrease although endogenous mvtg1 and transgenic gfp mRNAs were suppressed to
