Comparative aquatic toxicity of propranolol and its photodegraded mixtures: Algae and rotifer screening

Authors


  • Published on the Web 7/29/2009.

Abstract

Transformation products of pharmaceuticals formed by human metabolism within sewage treatment plant or receiving waters are predicted, in most cases, to be less toxic than the parent compound to common aquatic species. However, there is little available data to demonstrate whether this is generally the case. In the present study, a framework was developed to guide testing of transformation products using phototransformation of the β-blocker propranolol to test the hypothesis for this particular transformation route. Phototransformation is an important depletion mechanism of some pharmaceuticals in surface waters with fast reaction rate constants at environmentally relevant conditions. Samples of propranolol in deionized water (DIW) and river water (RW) were exposed to a solar simulator (λ: 295–800 nm) and comparative toxicity of propranolol and its degraded mixtures measured using algal (Pseudokirchneriella subcapitata) and rotifer (Brachionus calyciflorus) screening tests. Results suggested a reduction of toxicity in photodegraded mixtures compared to the parent active pharmaceutical ingredient in all samples tested. Chemical analysis of effect test solutions supported the hypothesis that propranolol was transformed into compounds that appear to be less toxic to the organisms tested under the study conditions. Although the reactions were much faster in RW than in DIW, profiles of transformation products were similar in both matrices at two starting concentrations (1 and 10 mg/L). Results for propranolol implied that the reduction of toxicity using algal and rotifer screening tests was probably due to the production of more hydrophilic and more polar transformation products. Such results will provide useful insights into the environmental risk assessment of pharmaceuticals by taking into account their transformation products.

INTRODUCTION

Pharmaceuticals in the environment have been extensively investigated because of the concern over their continuous release into the environment through sewage treatment plants (STPs) after patient use [1–7]. Active pharmaceutical ingredients (APIs) belong to a diverse group of bioactive chemicals that are metabolized in human bodies. They are excreted intact or as metabolites, both of which may be in conjugated form, before entering into STPs. Conventional STPs are designed for removing the majority of dissolved organic carbon, heavy metals, and nutrients and may not always be effective in removing APIs [8–13]. Therefore, some APIs have been repeatedly found in sewage effluents and receiving surface waters at ng/L to low μL levels [3–6,11,12]. This has led to public scrutiny over the potential long-term effects of APIs, such as β-blockers and their transformation products, on aquatic species [14,15] and human health issues over releases of APIs to groundwater or drinking water supplies [16,17; http://www.peer.org/news/news_id.php?row_id=1011].

Concerns over the potential environmental fate and effects of metabolites and transformation products are reflected in environmental regulations. For example, the European Medicines Agency regulatory guidance for the environmental risk assessment (ERA) of pharmaceuticals makes specific reference to the use of such data for refinement of the ERA [18]. In general, it is considered sufficient to conduct the ERA on the parent compound or the active moiety in the case of pro-drugs. This is a pragmatic approach that effectively assumes that any metabolites or transformation products are likely to be less toxic than the parent and hence is considered to represent a reasonable worst-case scenario for ERA. However, if the ratio of predicted environmental concentration to predicted no-effect concentration is above one based on the ERA of the parent compound, further evaluation may be needed, including fate and effects testing of metabolites and/or transformation products. However, no guidance is provided on how this should be done or how such information will be used.

This article forms part of a project that aims to develop a test strategy to fill in the gap of undertaking ERA of pharmaceutical transformation products and metabolites and to go beyond standard ERA guidance by linking fate and effect of pharmaceuticals and assessing transformation products in degraded mixtures. It also aims to provide a better understanding of the comparative ecotoxicity of different types of metabolites or transformation products compared with propranolol and to test the hypothesis that transformation products are likely to be less toxic to aquatic species than the parent API [15]. The selection of suitable chronic ecotoxicological screening tests was considered to be an important part of the work after reviewing available mode of action data [19], properties of test substance, and the numbers of organisms needed. The latter is especially important from an ethical point of view considering reduction, refinement, and replacement of vertebrate animals in scientific procedures.

To support the refinement of ERA, a tiered testing framework was developed in order to investigate potential effects of human metabolites and environmental transformation products (Fig. 1). One of the ideas with this framework is to be able to, in a tailored way, decide which approach to take for the testing of human metabolites and/or transformation products when ERA refinement is deemed necessary. For example, it may be more useful to test human metabolites than the parent of a compound that is totally metabolized when excreted from the human body. Alternatively, if it is known that an API is excreted largely unchanged, it may be appropriate to test the toxicity of biodegradation products resulting from sewage treatment. Similarly, if it is known that an API is susceptible to light, testing of phototransformation products in the environment may be appropriate. Thus, there are three key elements to the present study that address comparative ecotoxicity of APIs and their human metabolites, STP degradation products, and environmental transformation products.

Figure Fig. 1..

Framework showing the work breakdown structure for the research into comparative toxicity of transformation products/human metabolites versus the parent active pharmaceutical ingredient. MOA = mode of action; QSAR = quantitative structure-activity relationship; STP = sewage treatment plant.

Data from STP biodegradation and human metabolism studies are reported separately. In this article, transformation products formed via phototransformation are investigated. It has been demonstrated that phototransformation is an important environmental depletion mechanism for removing some APIs in surface waters with fast reaction rate constants under environmentally relevant conditions [20–29]. Environmental relevance has been considered by exposure to natural sunlight or solar simulator at low starting concentrations without causing a concentration effect for phototransformation in surface waters and measurement in natural matrices, such as STP effluents and river waters as well as pure water [20,23,24,27,28]. Phototransformation is particularly relevant to hydrophilic APIs with multiple chromophores, such as propranolol [24–28]. However, there are relatively few measured data [29] to determine whether transformation products are less toxic than the parent substances.

The majority of publications on ecotoxicity testing of phototransformation products describe procedures based on the identification, isolation, and toxicity testing of individual components (parent vs phototransformation products) [30–36]. In the approach here, we have attempted to quantify the overall ecotoxicity of the mixture of transformation products derived from the phototransformation experiments, using both deionized water (DIW) and river water (RW) test media. The profiles of transformation products against fate exposure time were investigated in DIW and RW samples. Chemical analysis was also included in the algal and rotifer screening tests to monitor whether there was significant reduction of parent compound during the effect exposure. This approach provided environmental relevance by considering that transformation products usually coexist in the environment and may undergo temporal variations. Furthermore, it is pragmatic and can reduce the need for developing isolation techniques and separate tests for individual substances. Although technically challenging, this enables a more robust interpretation of the data in the context of hazard and risk assessments of APIs and their transformation products in the aquatic environment.

Using propranolol as a case study, a testing program was designed to achieve the objectives of developing robust and cost-effective methods to provide comparative ecotoxicity data for APIs and their photodegraded mixtures, testing the hypothesis that phototransformation products from propranolol are less toxic than the parent API to aquatic species, and developing a test strategy and weight-of-evidence approach to show the relationship between primary degradation and reduction of toxicity. Propranolol is a nonselective β-blocker used as a cardiovascular API for treatment of angina pectoris, hypertension, and cardiac arrhythmia and was selected as a water-soluble API with established toxicity to some aquatic species and phototransformation data [24–28].

MATERIALS AND METHODS

Test substance

The test substance (Chemical Abstracts Service: 318–98–9) was a racemic mixture supplied by AstraZeneca UK Limited, that is, propranolol hydrochloride, (±)-1-isopropyl-amino-3-(1-naphthyloxy)propan-2-ol hydrochloride (99.8% w/w, subsequently referred to as propranolol).

Chemicals and reagents

River water samples were taken from the Tamar River (a rural site in southwestern England) in August 2007. River water samples were taken just below the water surface, downstream of an STP, for degradation studies and nutrient analysis. Nutrients analysis included dissolved organic carbon, nitrate, nitrite, ammonia nitrogen, total oxidized nitrogen, orthophosphate, total suspended solids, and particle size distribution [28].

High-purity methanol, acetonitrile, water, and formic acid were purchased from Fisher Scientific and were all high-performance liquid-chromatographic gradient grades.

Stock standard solutions of 10 mg/L were prepared in deionized water and stored in a refrigerator for less than 14 d until use. Appropriate working solutions (1–100 μL) were prepared by diluting the stock solution with a mixture of water:acetonitrile:formic acid (50:50:1, v/v/v). Unspiked RW samples were also analyzed to determine background concentration of propranolol.

Photolysis of propranolol in DIW and RW samples

Studies were performed with a DIW and a RW sample. The photolysis method used was based on Organization for Economic Cooperation and Development (OECD) guideline [37] with some modifications made [24,28]. Exposure concentrations of 1 and 10 mg/L parent compound propranolol were prepared in DIW or RW. Prior to use in the photolysis study, DIW and RW were sterilized at 121°C for 15 min to remove any microbial activity (biodegradation) and checked for dissolved oxygen (DO) concentration. If the DO concentration was below 80 to 100% saturation at test temperature, the batch of medium was gently aerated, avoiding introduction of microbial contaminants, to achieve the requisite DO concentration prior to the start of the photolysis experiment.

Samples (30 ml × 2) were transferred into borosilicate glass reaction vessels with a quartz glass lid. The depth of the samples in solutions was measured to give the light path length (l, cm) of the samples. The vessels containing samples were placed under the solar irradiator Hanau Suntest CPS (Heraeus Equipment) and irradiated for 48 h with a xenon (Xe) arc lamp equipped with a system of filters (290–800 nm). The light intensity and spectra of the Xe lamp were measured at the beginning and end of the experiment with a spectroradiometer (SpectRad, Jobin Yvon). Samples were compared with dark-control solutions with the same concentration but kept in darkness. Samples were taken at 0-, 6-, 24-, and 48-h intervals. Dark controls were included for both chemical analysis and ecotoxicity testing at each sampling point.

Algal microplate screening methods

The algal microplate screening method was based on the inhibition of algal growth, employing a procedure to monitor fluorescence as an indirect measure of algal biomass [38]. The test organism was the unicellular freshwater green alga Pseudokirchneriella subcapitata (strain ATCC 22662), maintained routinely and precultured according to OECD 201 [39].

Algal microplate tests were performed in white 96-well polystyrene microplates (Whatman) with 200 μl of test solution per well. The experimental design consisted of 10 replicates of the control and five of each concentration of the test sample in individual wells of the microplate. Corresponding blanks in the test media without alga were also included.

The growth medium for algal culture and testing was prepared according to OECD test guideline 201 [39]. At the start of the test, each well was inoculated with an initial cell density of 0.25 × 104 algal cells/ml. Cell density was determined by Coulter counter, counting at a lower threshold that was equivalent to a diameter of approximately 2.3 μm. After inoculation, the microplates were closed with the microplate cover and sealed additionally with Parafilm M (Laboratory Film, American National Can TM) to minimize evaporative losses and cross contamination but to ensure an exchange of carbon dioxide and oxygen. Microplates were then incubated at 24 ± 2°C under continuous illumination (maximum 13,000 lux) for a period of 72 h. After incubation, the density of algal cells in each well, including the blanks, was determined by fluorescence measurement using a SPECTRAFluor Plus microplate fluorimeter (Tecon AG). The excitation filter was set at 440 nm with a bandwidth of 40 nm, and the emission filter was at 690 nm with a bandwidth of 8.4 nm. Fluorescence may be used as an indirect measurement of algal biomass in the microplate method, and the expression of results based on algal biomass is preferred to measurements based on growth rate in order to obtain a high sensitivity of the test procedure. Physical parameters (pH, dissolved oxygen, and temperature) were measured daily; samples for chemical analysis of test solutions were taken at 0 h (excess test solution) and after 72 h (direct from microplate wells).

Rotifer microplate screening methods

The rotifer test was based on the inhibition of reproduction in the freshwater rotifer Brachionus calyciflorus as described in a draft International Organization for Standardization (ISO) standard (ISO/DIS 20666) [40]. In summary, young female rotifers less than 3 h old at the beginning of the test were exposed individually to a range of concentrations of the test sample for a period of 48 h. Test organisms (i.e., rotifers at 0 to 3 h old) were hatched from cysts incubated at 25°C for approximately 18 h in synthetic freshwater (U.S. Environmental Protection Agency moderately hard water of 80–100 mg/L as CaCO3) under low light intensity. Rotifer tests were performed in 24-well polystyrene microplates, 10 replicates per concentration (minimum n = 5) and dilution water control, with each well containing 2 ml of test sample and a suspension of 2 × 106 cells/ml of the green alga Chlorella vulgaris. The test was initiated by transferring one rotifer into each well of the microplates and incubating it at 25°C for 48 h in the dark. After 48 h, the total numbers of living female rotifers per well were counted and results were expressed as the corresponding median effective concentration (EC50) values, no-observed-effect concentration (NOEC), and lowest-observed-effect concentration (LOEC) for reproduction.

Chemical analysis for propranolol and its transformation products

For the profiles of transformation products, the method was described previously [24,28]. Exposed samples in RW and DIW (1 and 10 mg/L) were taken at 0, 6, 24, and 48 h. Analysis was by positive ion liquid chromatography/mass spectrometry (Thermo LCQ Advantage) operated in full scan mode. Transformation products with masses m/z at 292, 276, 308, and 264 were profiled, and the percentage of these products was calculated based on their peak areas relative to the peak area of parent API (propranolol) from standard solutions. Response factors were not determined for this purpose; therefore, the results here provided only comparison of transformation products in RW and in DIW.

Quantification of propranolol was also undertaken to confirm whether biological effects were due to the loss of propranolol from phototransformation. Both dark controls and samples from the algal and rotifer studies were taken at the beginning and end of the study. Samples from the beginning of the studies were taken from the excess exposure solutions and diluted into a range of 0.01, 0.05, 0.1, 0.5, 1, 5, and 10 mg/L with DIW. Samples from the end of the studies were taken from the microplate wells after centrifugation at 800 g and then diluted into a range with DIW. Analysis was by positive ion liquid chromatography/mass spectrometry (Thermo LCQ Advantage) using selected reaction monitoring [24].

Statistical analysis

For both algal and rotifer tests, EC50 values and the associated 95% confidence intervals, based on algae biomass and rotifer numbers and growth rate, were calculated using Weibull or Probit analysis. In addition, NOEC and LOEC values were calculated using appropriate parametric statistical procedures (one-way analysis of variance followed by Dunnett's t test, p = 0.01 and/or p = 0.05).

Selection of ecotoxicity test species and analytical methods

Selection of ecotoxicity test species and methods was based on a number of criteria. From a scientific perspective, ERA of pharmaceuticals tends to focus on chronic ecotoxicity testing in order to evaluate potential sublethal effects on organisms from different taxonomic groups, such as algae, invertebrates, and fish [18]. Second, laboratory photolysis used in the present study generated only relatively small sample volumes, which limited the use of fish for sublethal studies. Third, use of vertebrates, such as fish, also needs to be justified to address animal welfare principles. Finally, and most important, algal and invertebrate (rotifer) microplate screening methods provided robust approaches for the measurement of growth and reproductive end points, respectively, within relatively short exposure times. Therefore, the final choice of algal and rotifer screening methods met the selection criteria with a focus on measurement of sublethal toxicity, providing data in a cost-effective and ethical way.

Chemical analysis of test solutions was undertaken at the start and finish of the algal and rotifer studies. Measured concentrations of dark controls did not show any significant change of propranolol during exposure to algal and rotifer (Supporting Information, Tables S1–S4; http://dx.doi.org/10.1897/09–071.S1); therefore, results were calculated using nominal concentrations of propranolol. In optimizing reproductive output of rotifers in the required time period, a relatively high concentration of algal cells was added to the prepared test solutions. The algal particles showed the potential to interfere with chemical analyses; hence, the experimental design used in the microplate test included the addition of blank wells containing test solution but no algae.

The number of replicates for the rotifer tests was increased from 8 (as recommended in the draft guideline [40]) to 10 in order to increase statistical robustness with minimum effect on the time taken to perform the test. Test volume per cell well was increased from 1 to 2 ml to help reduce evaporative losses and provide more test volume for rotifers to reproduce.

Figure Fig. 2..

Phototransformation kinetics of propranolol measured in deionized (DIW) and in rotifer test media. The starting concentrations for rotifer tests were 1.25, 2.5, and 5 mg/L.

RESULTS AND DISCUSSION

Phototransformation in DIW and RW

After photolysis in DIW and prior to ecotoxicity testing in rotifer test media, measured first-order phototransformation kinetics were almost identical (Fig. 2) with a calculated half-life of 16.5 h. The similarity of propranolol profile in photolysis and rotifer test media suggested that possible media effects in the ecotoxicity samples were minimal. The kinetics data measured here were consistent with previously measured summer phototransformation kinetics of propranolol [20,24].

Reaction in RW was much faster than in DIW, with a rapid decrease of propranolol and major transformation products over time in the RW (Fig. 3), confirming a photosensitization effect in RW for propranolol as reported previously [27,28]. The phototransformation products from propranolol in RW were similar to those identified in DIW (Fig. 3a–d). Propranolol profile after photolysis is given in Figure 3e for the comparison to the transformation products. The profile is in agreement with our previous results of indirect photolysis of propranolol [28]. Although approximately 20 products were observed, only five were major products, and all of them appeared to be less than 10% of propranolol after 70-h exposure under the Xe lamp (Fig. 3). No concentration effect on the transformation products was seen in RW samples; that is, the profiles at 1 and 10 mg/L in RW were almost identical (Fig. 3c and d). Products m/z 292 showed structural isomers in two clusters, m/z 292 (1) and m/z 292 (2) [28], possibly because of light-induced ring opening and rearrangement [24]. Addition of three oxygen atoms to propranolol (m/z 308) also generated four main structural isomers [28]. Product m/z 276 had only one peak and showed transient existence, particularly in RW. However, as m/z 292 (1) decreased and m/z 292 (2) stayed stable over exposure time in RW, m/z 264 and m/z 310 showed a trend of increase over the time.

Figure Fig. 3..

Profile of transformation products from propranolol in deionized water (DIW) and river waters. (a) DIW, 1 mg/L, (b) DIW, 10 mg/L, (c) river water, 1 mg/L, and (d) river water, 10 mg/L. (e) Phototransformation profile of propranolol in DIW and river waters.

Physical-chemical properties and published aquatic toxicity of propranolol

Structures, properties, and literature ecotoxicity values of propranolol and its main phototransformation products are given in Table 1. Measured log DOW for propranolol at pH 7 was 0.78 [41]. The published propranolol ecotoxicity data are included here for guiding our selection of aquatic species to test. Huggett et al. measured the 48-h fish (Japanese medaka) median lethal concentration value of propranolol at ≥100 mg/L [14]. The measured 10-d fish (medaka) EC50 value of propranolol was 24.3 mg/L [42]. Growth rate of rainbow trout was affected only at >1 mg/L for 10-d growth and at >10 mg/L for 40-d growth [43]. However, it was difficult to compare available ecotoxicity data because they were generated from different studies using either measured or nominal exposure concentrations. According to published aquatic toxicological values so far, propranolol demonstrated higher acute toxicity than other β-blockers [14]. This could be partly due to its relatively higher log DOW value and the fact that propranolol is a strong membrane stabilizer [14,44]. However, it was reported that propranolol was more sensitive to phytoplankton and zooplankton than fish [14,42,45]. Ceriodaphnia dubia showed higher sensitivity than Daphnia magna or other zooplankton organisms [45]. The blue-green alga Synechococcus leopolensis reacted most sensitively within phytoplankton [42]. Based on the literature search, the consideration of available sublethal ecotoxicity screening methods and the requirement for small sample volumes, green algae (P. subcapitata) and invertebrate rotifer (B. calyciflorus) were selected for ecotoxicity tests in this study.

Table Table 1.. Structures, properties, and literature values of ecotoxicity (mg/L) for propranolol and its major phototransformation products (NA = not available)
NamePropranolol inline imageProduct 292 inline imageProduct 264 inline image
  1. a These are from KOW Win estimations (Syracuse Research Corporation).

  2. b log DOW: pH-dependent octanol-water partition coefficient; measured value from Avdeef et al. [41].

  3. c Ferrari et al. [42].

  4. d Fent et al. [45].

  5. e Huggett et al. [14].

  6. f Owen et al. [43]. EC50 = median effective concentration, which is the concentration causing a response in 50% of the test organisms in the time period specified; NOEC = no-observed-effect concentration. LOEC = lowest-observed-effect concentration.

log KOWa2.60.962.04
log Dow (pH 7)0.78bNANA
Water solubilitya281.81.4 × 1041,066
pKa9.5NANA
Synechococcus leopolensis 96-h EC500.668eNANA
Ceriodaphnia dubia 48-h EC500.8dNANA
Daphnia magna 48-h EC501.6eNANA
C. dubia NOEC/LOEC0.125/0.25eNANA
Hyallela azteca 27-d reproduction0.1eNANA
Oncorhynchus mykiss (rainbow trout) 10-d growth1.0 (NOEC)f 10 (LOEC)fNANA
O. mykiss (rainbow trout) 40-d growth10 (NOEC)fNANA
Table Table 2.. Measured 1/EC50, NOEC, and LOEC values of propranolol and its photodegraded mixtures exposed to freshwater species in deionized water (DIW) and river water (RW) (NA = not available)a
SpeciesTime (h)Fate test matrixbSample1/EC50° (1/mg/L)(95% CI) l/EbC50b (1/mg/L)NOEC (mg/L)LOEC (mg/L)
  1. a EC50 = the median effective concentration, which is the concentration causing a response in 50% of the test organisms in the time period specified. The results for algae refer to biomass. NOEC = no-observed-effect concentration; LOEC = lowest-observed-effect concentration.

  2. b Fate test matrices were DIW (deionized water) and RW (river water). River water samples were taken from the Tamar River (UK) in the summer of 2007.

  3. c Algal results were calculated using Weibull analysis based on biomass after 72 h of exposure to Pseudokirchneriella subcapitata. Rotifer results were based on rotifer numbers after 48 h of exposure to Brachionus calyciflorus. All results were based on nominal concentrations of propranolol.

AlgaePropranolol1.3(0.94–2.0)<0.780.78
 0DIWSample0.80(0.77–0.83)0.156>0.156
   Dark control0.51(0.33–1.1)<0.1560.156
  RWDark control0.79(0.68–0.93)0.156>0.156
 6DIWSample0.45(0.44–0.46)0.3120.625
   Dark control0.67(0.51–0.96)0.3120.625
 24DIWSample0.33(0.30–0.37)0.6251.25
   Dark control0.46(0.43–0.51)1.252.5
  RWSample0.21(0.15–0.31)0.3130.625
   Dark control1.1(0.98–1.3)0.1560.313
 48DIWSample0.26(0.22–0.32)1.252.5
   Dark control0.97(0.86–1.1)0.3120.625
  RWSample<0.20NA2.55.0
   Dark control1.2(0.98–1.6)<0.1560.156
Rotifer Propranolol0.45(0.40–0.50)1.03.2
 0DIWDark control and sample0.26(0.22–0.29)2.55.0
 6DIWSample0.20(0.16–0.28)2.55.0
   Dark control0.20(0.16–0.23)2.55.0
 24DIWSample<0.20NA>5.0
   Dark control0.27(0.21–0.32)2.55.0
 48DIWSample<0.20NA>5.0
   Dark control0.16(0.13–0.19)>5.0

Comparative toxicity using algal and rotifer screening methods

Concurrent ecotoxicity tests of photodegraded samples and their corresponding dark controls provided a direct comparison between photodegraded samples at 6, 24, and 48 h and nonphotodegraded samples at 0 h (Table 2). It also provided comparison between photodegraded samples and photolysis dark controls. For photodegraded samples taken at 0 h, dark controls and samples were identical; therefore, only one of these samples was subjected to ecotoxicity testing (Table 2). Results from both algal and rotifer studies on photodegraded solutions of propranolol indicated a reduction of 1.8 to 3 times in toxicity compared to the parent compound, that is, DIW-0 and RW-0 (Figs. 4 and 5). The relative toxicity to both species was also generally lower in photodegraded samples than in dark controls at each photolysis sampling point (6, 24, and 48 h exposure to the Xe lamp) with higher EC50 values, hence lower toxicity, in samples than in dark controls (Table 2).

Figure Fig. 4..

Algal screening results for propranolol and photodegraded mixtures: (a) 0- and 24-h photolysis samples in deionized water (DIW), (b) 0- and 48-h photolysis samples in deionized water, (c) 0- and 24-h photolysis samples in river water, and (d) 0- and 48-h photolysis samples in river water.

Algal screening results

Algal results, expressed in toxic units (1/EC50) for samples taken at 0 h were similar in DIW and RW despite the use of different media, that is, RW and deionized water for photolysis (Table 2). These results were also in agreement with the toxicity measured for the parent compound (Table 2). Compared with the 0-h DIW sample, there was a reduction of 2.4 to 3 times in toxicity for 24- and 48-h photolysis samples (Fig. 4). Furthermore, after 24 and 48 h of photolysis, photodegraded samples in DIW showed a reduction in toxicity (by a factor of 1.4–3.1) compared with corresponding photolysis dark controls (Table 2). A larger reduction in toxicity (by a factor of 5.5 to >6) was observed in corresponding samples in RW (Table 2).

Prior to ecotoxicity tests, the measured concentrations of propranolol in test dilutions of sample DIW-0 (i.e., not subjected to photolysis) were similar to nominal (96–104%). Similar results were obtained for samples incubated in the dark for periods of 6, 24, and 48 h (96–104% of nominal). Mean measured concentrations of propranolol in photodegraded solutions, prior to toxicity testing, declined to approximately 85, 48, and 27% of nominal values after 6, 24, and 48 h, respectively (Supporting Information, Tables S1–S2). This followed first-order reaction kinetics and was in agreement with previous results for direct photolysis of propranolol [24]. Subsequently, measured concentrations of propranolol changed little over the 72-h exposure duration of the algal test (Supporting Information, Table S2). Further changes in measured concentrations of propranolol were most evident in sample DIW-48 (48 h of photolysis, followed by 72 h of exposure in algal microplate test), and mean measured concentrations ranged from approximately 9 to 15% of nominal concentration. These results suggest that the observed reduction in toxicity of propranolol samples after photolysis was due to loss of parent compound after phototransformation in DIW or RW.

The algal microplate procedure has been shown to provide data that were comparable with the regulatory algal shake flask method [39], provided that physical-chemical properties of the tested substances are taken into consideration [46,47]. The physical-chemical properties include the variation of partitioning behavior (log DOW) and ionization of propranolol (pKa) as a function of pH values of environmental waters (pH ∼7). For example, comparative testing of the algal microplate and shake flask method yielded results that were within a factor of <5 (usually less toxic) for organic chemicals [38].

Rotifer screening results

For the parent compound, all rotifers were dead after 48 h at a nominal propranolol concentration of 10 mg/L. Compared to the dilution water control, there was no effect on rotifer numbers at a nominal 1.0 mg/L; however, there was a 72% reduction in rotifer numbers at 3.2 mg/L. Consequently, based on these results, a dilution series starting from a nominal 5.0 mg/L downward should have captured any changes in toxicity following photolysis of the parent compound. For the photodegraded mixtures, samples showed a reduction in toxicity compared with corresponding samples incubated in the dark after 24 and 48 h of photolysis (Table 2). After 24 and 48 h of photolysis (i.e., DIW-24 and DIW-48 samples), toxicity was observed only in the highest exposure concentration (nominal 5.0 mg/L propranolol) (Fig. 5). The inhibitory effect of photodegraded mixtures was less than 50% of the dilution water controls for the concentration range tested and was not sufficient to produce valid EC50 values that were needed to determine the precise magnitude of reduction in toxicity. However, the approach here was designed to provide comparative ecotoxicity of transformation products to the parent substance (i.e., propranolol). For this purpose, the results were sufficient to suggest that photodegraded mixtures showed less inhibitory effect to rotifer than propranolol itself at 5 mg/L nominal concentration (Table 2 and Fig. 5).

Figure Fig. 5..

Rotifer screening results for propranolol and photodegraded mixtures in deionized water (DIW): (a) 0- and 24-h photolysis samples and (b) 0- and 48-h photolysis samples.

DISCUSSION AND CONCLUSION

The test strategy and results described here seek to answer a fundamental question on how transformation products should be taken into account for overall environmental risk assessment and how the assessments should be performed. The current work is included in one of three key elements included in the tiered testing framework outlined in the Introduction and describes phototransformation of propranolol and the possible impact of the phototransformation products on aquatic species.

Propranolol is a nonselective β-blocker. Because of the naphthalene skeleton, propranolol is liable to undergo fast phototransformation either in solid phase [48] or in water solutions [24–28]. It was hypothesized that phototransformation products from propranolol would be less toxic because of their presumed higher polarity and hydrophilicity than the parent compound, propranolol. In this study, a literature search, practical considerations, and animal welfare principles were applied for selecting suitable aquatic species for testing using a weight-of-evidence approach. Algal and rotifer ecotoxicity screening tests were then used for measurement of comparative toxicity of propranolol and its photodegraded mixtures. The results from these tests support the hypothesis that phototransformation products from propranolol are less toxic to green algae and rotifer under the test conditions.

To assess potential risks posed by transformation products compared to parent APIs, it needs to be recognized that transformation products are the results of reaction mixtures in various environmental matrices. However, there is little information on how the reaction mixtures interact with each other to vary their behaviors in different environmental compartments and how competitive physical-chemical processes (e.g., abiotic and biotic transformation, partitioning, volatilization, and precipitation) occur within mixtures. Profiles and concentrations of transformation products undergo spatial and temporal changes [49]. Adsorption/desorption can also be dynamic and takes time to reach equilibrium [26]. In addition, assessing absolute mixture toxicity is very difficult because of the limited understanding of mechanisms of action for mixtures [50]. Experimental approaches have been used to assess ecotoxicity of testing isolated photoproducts from some pharmaceuticals [30–34]. However, it is considered that the comparative toxicity of parent APIs and their degraded mixtures is a pragmatic approach that may be used as an alternative to those used for testing isolated transformation products.

Future work

The algal and rotifer microplate screening tests employed in this study provided rapid (48–72 h), cost-effective, and practical tools for measuring sublethal ecotoxicity of transformation products versus parent APIs. The use of fish in the proposed testing scheme required careful consideration; for example, practical limitations arose from the large volumes of test solution required to conduct sublethal studies. Although methods for fish life cycle assessment of pharmaceutical effects have been developed [51], it remains difficult to expand the approach to a wide range of APIs because of the costs and animal welfare issues. To address these issues, intelligent testing schemes should take into consideration the pharmaceutical mode of action to identify which taxonomic groups require ecotoxicity assessment for risk assessment purposes [19]. Fish should be used only if considered absolutely necessary, and practical obstacles need to be overcome. Fish possess β-adrenergic receptors, and, in the case of β-adrenergic receptor blockers such as propranolol, acute-to-chronic ratios for fish were reported to be very high [52], indicating that this taxonomic group may be sensitive to such substances. The in vitro method (e.g., fish cell lines) may be an interim step between nontesting strategies, such as quantitative structure-activity relationship, mode of action, and full fish studies, and work is under way to investigate what methods can be used to provide sublethal fish data when needed.

SUPPORTING INFORMATION

Table S1. Chemical confirmation: measured concentrations of propranolol in test solutions after 0 h exposure to algae.

Table S2. Chemical confirmation: measured concentrations of propranolol after 72 h exposure to algae.

Table S3. Chemical confirmation: measured concentrations of propranolol in test solutions after 0 h exposure to rotifer.

Table S4. Chemical confirmation: measured concentrations of propranolol in test solutions after 48 h exposure to rotifer.

All found at DOI: 10.1897/09–071.S1 (33 KB PDF).

Acknowledgements

The authors are grateful to Sarah MacLean, Doug Morgan, and Pete Johnson, who assisted in the laboratory measurements, and Andrew Riddle, who provided useful comments of the manuscript. Qintao Liu is very grateful to Stewart Owen for a series of discussions on fish toxicity of β-blockers.

Ancillary