In Silico Models for Ecotoxicity of Pharmaceuticals

Abstract

Pharmaceuticals and their active metabolites are one of the significantly emerging environmental toxicants. The major routes of entry of pharmaceuticals into the environment are industries, hospitals, or direct disposal of unwanted or expired drugs made by the patient. The most important and distinct features of pharmaceuticals are that they are deliberately designed to have an explicit mode of action and designed to exert an effect on humans and other living systems. This distinctive feature makes pharmaceuticals and their metabolites different from other chemicals, and this necessitates the evaluation of the direct effects of pharmaceuticals in various environmental compartments as well as to living systems. In this background, the alarming situation of ecotoxicity of diverse pharmaceuticals have forced government and nongovernment regulatory authorities to recommend the application of in silico methods to provide quick information about the risk assessment and fate properties of pharmaceuticals as well as their ecological and indirect human health effects. This chapter aims to offer information regarding occurrence of pharmaceuticals in the environment, their persistence, environmental fate, and toxicity as well as application of in silico methods to provide information about the basic risk management and fate prediction of pharmaceuticals in the environment. Brief ideas about toxicity endpoints, available ecotoxicity databases, and expert systems employed for rapid toxicity predictions of ecotoxicity of pharmaceuticals are also discussed.

1 Introduction

A significant amount of pharmaceuticals and their metabolites have been found in the various environmental compartments causing damage to the environment and hazard to the living systems. Due to an increase in application of human and veterinary medicines manyfold, pharmaceuticals and their metabolite residues have been found in rivers, sewage effluents, streams and in surface, ground, and potable water, creating a big concern for the ecologists [1]. The primary routes of entrance of pharmaceuticals into the environment are domestic, hospital, and industrial wastes [2]. Pharmaceuticals are excreted in urine or feces as a mixture of unchanged chemicals and metabolites and enter into the environment through septic tank and sewage systems [1]. On the other hand, ecotoxicity data of pharmaceuticals are available in the literature for less than 1 % of the drugs, and only a small number of pharmaceuticals and their residues have been subjected to risk assessment employing ecotoxicological tests.

Pharmaceuticals are intentionally designed to have a specific mode of action and exert an effect on specific organs, tissues, cells, or biomolecules in humans, mammals, or other vertebrates, and many of them are persistent in the body [3]. As a consequence, when pharmaceuticals and their unaltered metabolites enter into the environment by different means, they can affect humans as well as other living species. There are many drugs whose specific effects or modes of action are not well known, and they often produce effects through several modes of action. These distinguished features make pharmaceuticals dissimilar from others and this is the sole reason to assess the potential acute and chronic effects of pharmaceuticals in diverse environmental compartments. It is quite apparent that the toxic effects of pharmaceuticals on diverse organisms in aquatic as well as nonaquatic environment are due to their long persistent and bio-accumulative nature [4]. In view of the serious issue of pharmaceutical toxicity to the environment, it is vital to categorize the proper source, occurrence, effects, and fate of each individual pharmaceutical product as well as to perform the risk assessment and risk management of ecotoxicological effects of the pharmaceutical chemicals and their metabolites [1, 2].

Antibiotics are one of the majorly used pharmaceuticals in human and veterinary medicines. The world consumption of antibiotics has risen radically in the last decade, also increasing the elimination of their metabolites in their original form. Most antibiotics are poorly metabolized after ingestion, probably resulting in a fraction of antibiotics from 25 to 75 % leaving the bodies in an unaltered form after consumption [5]. Additionally, a high percentage of the antibiotics added to the animal feed are excreted in urine or manure. In some cases, as much as 90 % of the antibiotic administered orally may pass through the animal unchanged and excreted in urine and manure. Thereafter, these antibiotics can enter surface and groundwater and be strongly adsorbed in soils and are not readily degradable [6]. Vidaver [7] estimates that 53,000 ha of fruit and vegetable plants are sprayed annually with antibiotics. For example, streptomycin and oxytetracycline are registered by the US Environment Protection Agency (USEPA) for use in plant agriculture. Utilization of transgenic plants to produce inexpensive antibiotics may also be a cause of environmental hazards due to the existence of crop residues, roots, and root exudates in the soil which can act as a continuous source of residual antibiotics to soil fauna and flora [8].

While pharmaceuticals and their metabolite residues are detected in rivers, streams, sewage influents and effluents, surface, ground, and potable waters [9], it may be noted that the drinking water treatment methods reduce residues, but they are incapable of removing the contaminant pharmaceuticals absolutely. According to a nationwide study of “emerging pollutants” in waters, the US Geological Survey (USGS) tested for pharmaceuticals in 139 rivers in 30 states of the USA, detecting diverse therapeutic classes of biologically active compounds [10]. The cardiovascular drug propranolol has been reported downstream from the sewage treatment plant [11]. The antiepileptic drugs carbamazepine and clofibrate are two most persistent pharmaceuticals which have been detected in the environment [2]. Major detected drugs in rivers were beta blockers (e.g., metoprolol up to 1.54 μg/l) and beta-sympathomimetics, estrogens (e.g., 17β-estradiol up to 0.013 μg/l) [12], analgesic and anti-inflammatory drugs (e.g., Diclofenac up to 1.2 μg/l) [13], and also antibiotics (e.g., erythromycin up to 1.7 μg/l) [12], as well as lipid-lowering agents (e.g., clofibrinic acid up to 0.2 μg/l) [14] and antiepileptic drugs (e.g., carbamazepine up to 2.1 μg/l) [13]. Presence of clofibric acid, propylphenazone, and diclofenac has been reported in the drinking water of Berlin in the concentration range of several hundreds of nanograms per liter [15]. Paracetamol, diclofenac, and carbamazepine were monitored in drinking water in Southern France [16], and clofibric acid and diazepams were detected in treated drinking water in Milan, Italy [17]. Psychoactive and illicit drugs amphetamine, cocaine and its metabolite benzoylecgonine, morphine, 6-acetylmorphine, 11-nor-9-carboxy-delta-9-tetrahydrocannabinol, methadone and its main metabolite 2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine have been detected in surface and waste waters [18]. Schultz and Furlong found highest concentrations of antidepressant drugs venlafaxine, citalopram, and bupropion 1000 ± 400 ng/l, 90 ± 20 ng/l, and 60 ± 40 ng/l, respectively, in samples collected downstream from a water reclamation plant [19]. The maximum determined concentration of fluoxetine was 0.099 ng/l in wastewater treatment plant (WWTP) effluents in Canada [20].

Nonprescription drugs like caffeine, cotinine, and acetaminophenone are found in samples of potable water collected near Atlanta, Georgia [21]. Tauber detected carbamazepine and gemfibrozil in drinking waters in ten cities in Canada that were examined for a 44-drug subset consisting pharmaceuticals including sulfonamides, quinolones, tetracyclines, and macrolide antibiotics [22]. Oraine and Pettigrove identified and quantified ibuprofen (0.93 μg/l) and ibuprofen methyl ester (4.95 μg/l) in finished water in alarming quantity [23]. Median concentrations of 0.02 μg/l and 0.12 μg/l were reported for ciprofloxacin and norfloxacin, respectively, for samples from 139 surface streams across the USA. Ciprofloxacin in the range 0.7–124.5 μg/l was found in wastewater of a Swiss hospital [24]. Hellweger et al. [25] claimed that environmental concentrations of tetracycline in surface waters are usually less than 0.11 mg/l, although higher values of up to 6.8 mg/l have been observed. Estrogens, a sex hormone, have been detected in plasticizers and preservatives, while 17α-ethinylestradiol (EE2) used as a component of contraceptive pills has been identified in ground and tap water samples [26].

The presence of human and veterinary pharmaceuticals and their residues into the environment has impelled the introduction of different risk assessment guidelines in the European Union by the European Medicines Evaluation Agency (EMEA) and in the USA by the Food and Drug Administration (FDA). According to the European Commission guideline [27], a medicinal product for human use must be accompanied by environmental risk assessment data. The EMEA has released a guideline for the assessment of potential environmental risks in 2006 [28]. According to the US FDA guidelines for the risk assessments of human drugs, applicants have to provide an environmental assessment report when the expected concentration of the active pharmaceuticals in the aquatic environment is ≥1 μg/l [29]. Additionally, the FDA Center for Drug Evaluation and Research (CDER) issued a guidance document “Guidance for Industry for the Submission of an Environmental Assessment in Human Drug Application and Supplements” in 1995 [30]. In case of veterinary medicines, environmental risk assessments have been required in the USA since about 1980 and Europe since 1997 [31].

The need for a practical approach in gathering data on the environmental toxic effects of pharmaceuticals has been identified by the European Union Commission’s scientific committee on toxicity, ecotoxicity, and environment (CSTEE). The four classes of special environmental feature-specific concerns, which are stereotypically not evaluated in traditional ecotoxicity testing under EU directive 1488/94 [28] are antibiotics [resistance issue], antineoplastics [mutagenicity], sex hormones [endocrine disruption], and cardiovascular high potential hazard. Therefore, it is acknowledged that a prioritization technique needs to be developed for environmental risk assessment of pharmaceuticals, and this should follow the general scheme for chemicals according to the REACH guidelines [27], where the implication of in silico methods specifically the quantitative structure–activity relationship (QSAR) method is stressed.

In this perspective, to make the information regarding ecotoxicity of diverse pharmaceuticals available, different government and nongovernment regulatory authorities are recommending the application of fast and economical in silico methods for prediction of the elementary physicochemical and fate properties of pharmaceuticals as well as their ecological and direct human health effects before they reach into market for usage. Computer-aided toxicity models allow for the effects of pharmaceuticals (physicochemical properties, toxicological activity, distribution, fate, etc.) to be easily predicted. These predictions may be obtained from the knowledge of chemical structure alone, provided that the structure can be described in two or three dimensions. Employing these methods, ecotoxicity information on pharmaceuticals may be obtained without toxicity testing, and/or even before synthesis of the compound. Therefore, use of QSAR as one of the non-experimental methods is significant in order to lessen time, animal usage and cost involvement in design, development, and discovery process of drugs and/or pharmaceuticals.

There is a significant lack of knowledge about the environmental fate of a huge number of pharmaceuticals and their metabolites. On the contrary, only a limited number of in silico models have been developed so far to predict the risk of pharmaceuticals to the environment. This chapter aims to provide information regarding occurrence of pharmaceuticals and their residues in the environment, their persistence, environmental fate, and toxicity as well as application of in silico methods to predict risk and fate properties of pharmaceuticals to the environment. Concise ideas about ecotoxicity endpoints, available ecotoxicity databases and expert systems employed for rapid ecotoxicity predictions of pharmaceuticals are discussed in this chapter.

2 Ecotoxicity of Pharmaceuticals: A General Overview

2.1 Source and Entry Routes

Identification of proper sources and routes of entry of pharmaceuticals into diverse environmental compartments is the first step to get a proper view of the ecotoxicity problem due to pharmaceuticals. The most obvious and common pathways for environmental contamination of pharmaceuticals are discussed below.

• (a) Urine and feces: Major and most common entry routes for pharmaceuticals into the environment are via urine and feces of the patients. Not only active ingredients, but also the metabolites are excreted through the urine and feces as many drugs are metabolized into hydrophilic compounds for excretion. The risk of these metabolites is completely different from the parent drugs in majority of cases which make the risk assessment study more critical one.

• (b) Direct exposure of diagnostic compounds: Contrast media like diatrizoate, iohexol, iomeprol, and iopromide are used as diagnostic tools for capturing detailed X-ray images of soft tissues. Iodinated X-ray contrast media are highly hydrophilic substances which are extensively applied and eliminated without proper treatment; as a result they persist for a long time in the environment [32].

• (c) Household disposal: Either out-of-date or unwanted medicines are discarded through the sink/toilet or via waste collection, before being taken to landfill sites where they appear as terrestrial ecosystem contaminants. Less than 20 % users had ever been given instructions about medication dumping by a health care provider. In a study, causes for possessing unused medication were found to be due to an alteration of medication by the doctor (48.9 %), or self-discontinuation (25.8 %) [33]. The most common method of disposal was to throw unused medicines in the trash (76.5 %) or flush them down the drain (11.2 %) [33].

• (d) Manufacturers: According to the regulation of the Good Manufacturing Practices (GMP), the active pharmaceutical emissions during manufacturing have been thought to be insignificant. But recently it has been found that in Asian countries concentrations up to several milligrams per liter can be found in effluents for single compounds [34].

• (e) Hospital influent and effluent: Point sources such as hospital effluents are likely to be another significant source. There are up to 16 pharmaceuticals including antiepileptics and anti-inflammatories which were found in the hospital waste water according to a study [35]. Several studies suggested the existence of the pharmaceuticals in the effluent and influent of the sewage treatment plants and it was proved that the elimination of the pharmaceuticals is partial [35].

• (f) Animal husbandry and veterinary medicine: Veterinary medicines and their metabolites are also excreted through urine and feces. Apart from the potential for direct soil contamination, there is also the risk of run-off with heavy rain, thus potentially contaminating both the surrounding surface and groundwater. Other sources include direct application in aqua farming, manure run-off, run-off from the application of sewage sludge and manure on farmland as fertilizers, or, finally, via landfill leaching [36].

• (g) Aquaculture: Sewage Treatment Plant (STP) sludge is habitually employed as fertilizer on agricultural land which is a rich source of non-suspected drugs [37]. According to the Food and Agriculture Organization (FAO), antibiotics have been utilized in aquaculture primarily for therapeutic purposes and as prophylactic agents. Antibiotics authorized for use in aquaculture are florfenicol, oxytetracycline, sarafloxacin, premix, erythromycin sulfonamides potentiated with ormethoprim, or trimethoprim [38].

• (h) Plant agriculture: Antibiotics are comprehensively employed to control bacterial diseases of plants. Streptomycin with oxytetracycline to a minor extent is very commonly used antibiotic in plant agriculture in controlling bacterial diseases of tree fruits. Primary uses are on apple, pear, and related fruit trees for the control of fire blight caused by Erwinia amylovora. According to a report, antibiotics applied to plants account for less than 0.5 % of total antibiotic use in the USA [39]. In Fig. 1, we have represented different sources, routes, fate of pharmaceuticals.

  Fig. 1

  

  Common sources, routes and fate of pharmaceuticals

2.2 Occurrence

Pharmaceuticals are among the most common personal care products in day to day life. Medicines are regularly used in human and veterinary health care, farming, and aquaculture in the modern era. Country specific consumption for groups of drugs in defined daily doses (DDDs) can be found for Europe on the European Surveillance of Antimicrobial Consumption (ESAC) homepage [40]. In the last decade, a large number of studies covering occurrence of pharmaceuticals in water bodies, sewage treatment plants, manure, soil, and air dust have been published. The most concerning issue is that under the environmental conditions, these molecules can be neutral, cationic, anionic, or zwitterionic which make the risk assessment study of pharmaceuticals more difficult. In Table 1 we have presented the reported concentrations of diverse pharmaceuticals from various therapeutic classes in different samples of different countries and probable ecotoxicity data to particular toxicological endpoints [3, 41–68].

Table 1 Various therapeutic classes of pharmaceuticals and their reported concentration in different samples of different countries as well as toxicological endpoints and probable ecotoxicity data

2.2.1 Waterbodies

The presence of pharmaceuticals in the various waterbodies in the environment has been quite extensively studied by different research groups. Quinolones (predominantly ciprofloxacin) and other pharmaceuticals have been detected in the effluent of hospitals up to a low μg/l range. Another study reveals that β-lactams (including penicillins, cephalosporins, carbapenems, monobactams, β-lactamase inhibitors) were detected in the lower μg/l range in hospital effluent and in the influent of a municipal STP [69]. NSAIDs have the higher concentrations recorded in surface water, ranging between 0.4 ng/l and 15 μg/l, diclofenac, paracetamol, and ibuprofen being the most quantitatively found [70]. Drugs like caffeine with a maximum concentration of 6 μg/l and sulfamethoxazole with 1.9 μg/l in the USA, carbamazepine up to 1.3 μg/l in Germany and in Canada, gemfibrozil up to 790 ng/l, ranitidine up to 580 ng/l, atenolol with 241 ng/l in Italy, and metformin up to 150 ng/l are detected in surface water [71]. In the effluent of WWTP and STP, the concentrations of estrogenic compounds usually are below 50 ng/l, but there are unexpected high concentrations of estriol and 17α-estradiol (about 590 ng/l and 180 ng/l respectively) found in the USA [72].

2.2.2 Manure and Soil

Antibiotics have been detected in soil in concentrations in the mg/kg range [73]. Generally, the concentrations of pharmaceuticals detected in the soils are quite low when compared with that of pharmaceuticals in water resource. According to the literature, the six most common pharmaceuticals found in soil are the antibacterials (trimethoprim, sulfadiazine, and triclosan), analgesics (ibuprofen and diclofenac) and antiepileptic (Carbamazepine). Extensive studies have detected tetracyclines and sulfonamides in liquid manure at concentrations of up to 20 and 40 mg/l, respectively. Antibiotics like virginiamycin, sarafloxacin, tetracycline, oxytetracycline, chlortetracycline, and cyclosporine A have quite slow biodegradability in soil. Tylosin disappeared soon after the application of manure. Hamscher et al. [74] detected tetracycline and chlortetracycline in 10 out of 12 soil samples. The highest average concentration of 86.2 μg/kg (0–10 cm), 198.7 μg/kg (10–20 cm), 171.7 μg/kg (20–30 cm) tetracycline, and 4.6–7.3 μg/kg (in all three sublayers) chlortetracycline were found. Carbamazepine is the most frequent compound detected in soil among five studies [75].

2.2.3 Air Dust

Several comprehensive reports have been published on environmental concentrations of antibiotics in dust originating from a pig-fattening house [76]. In a large-scale pig production, veterinary antibiotics are hugely used. This production system is represented as a considerable source of dust.

2.3 Effects

2.3.1 Antibiotics

Pharmaceuticals may have potential adverse effects on aquatic and terrestrial organisms by directly reaching into the environment. Organisms like bacteria, fungi, and microalgae are primarily affected as antibiotics are designed to inhibit the microorganisms. Antibiotics have the potential to affect the microorganisms in sewage systems and waste water treatment plant too. The inhibition of wastewater bacteria may seriously affect organic matter degradation and nitrification process which is a vital step in wastewater purification and elimination of toxic ammonia [77]. Lincomycin showed significant inhibition of the nitrification activity [78]. Ciprofloxacin was found to be active against Vibrio fischeri at a concentration of 5 mg/l [79]. Thomulka and McGee [80] have performed two bioassays to evaluate the toxicity of antibiotics like novobiocin, chloramphenicol, tetracycline, ampicillin, and streptomycin to Vibrio harveyi, and approximately no toxic effects were identified after short incubation times where the employed endpoint was luminescence. Common receptors have been identified in plants for a number of antibiotics affecting transcription and translation (tetracyclines, macrolides, lincosamides, aminoglycosides, and pleuromutilins), metabolic pathways such as folate biosynthesis (sulfonamides), chloroplast replication (fluoroquinolones), and fatty acid biosynthesis (triclosan) [81].

Antimicrobials can affect the degradation of organic matter in large extent as well as have effects upon sediment’s microbial community [82]. Strong inhibitory effects on several bacteria and diminution in the length of the hyphae of lively molds in forest soil have been observed when antibiotics are added in concentrations of 10 mg/kg soil. A transitory effect on sulfate reduction was detected when antibiotics were mixed to sediment [83]. Allergic risks may arise from the high exposure of antibiotics dust particle in the air. Tylosin and sulfamethazine, which occurred in 80 % and 65 % of the samples respectively, are drugs with known allergic potential. Therefore, the high incidence of the asthma disease occurred among children living on farms. A survey on dust in pig fattening buildings in Europe exposed an average concentration of inhalable airborne dust of 2.2 mg/m³ [84]. Chloramphenicol is extensively employed in farming resulting in severe hazardous effects including myelosuppression to farmers; that is why it was totally banned for food-producing animals within the EU and the USA in 1994 [85].

Another important aspect is the emergence of resistance due to enormous application of antibiotics in human medicine, veterinary medicine, and animal husbandry. Resistance is one of the most concerning issue in medical field due to its accumulating and accelerating nature. On the contrary, the techniques combating resistance are diminishing in power and number. Antibiotics in sub-inhibitory concentrations can have an influence on cell functions and modify the genetic expression of virulence factors or the transfer of antibiotic resistance. The most prominent medical examples are vancomycin-resistant enterococci (VRE), methicillin-resistant Staphylococcus aureus (MRSA), and multiresistant pseudomonads [86].

2.3.2 Analgesics and Nonsteroidal Anti-inflammatory Drugs (NSAIDs)

Cleuvers [45] evaluated that acute toxicities of NSAIDs were relatively low, with half-maximal effective concentration (EC₅₀) values obtained using Daphnia in the range from 68 to 166 mg/l and from 72 to 626 mg/l in the algal test. With EC₅₀ values of 23.6 mg/l (ibuprofen), 23.8 mg/l (diclofenac), and 38.2 mg/l (naproxen), chronic ecotoxicity was somewhat higher, but still the values are far above the concentrations detected in surface water. A prominent confirmation of diclofenac residues in dead cattle has been observed in Pakistan [87]. Only in Germany, in 2002, 93.5 million prescriptions for NSAIDs were made with a transaction volume of about 1562 million Euros [88]. Due to higher usage and pharmacokinetic and pharmacodynamic properties, analgesics and anti-inflammatory drugs can reach considerable (up to >1 μg/l) concentrations in the environment. Few NSAIDs are detected in very low doses even in drinking water. Reports suggested the presence a concerning amount of diclofenac and ibuprofen in Swiss lakes and rivers, as well as in water bodies from the UK, Spain, Brazil, Greece, and the USA [15].

Diclofenac seems to be the compound having the highest acute toxicity with the effective concentrations below 100 mg/l within the class of NSAIDs. Short-term acute toxicity was analyzed in algae and invertebrates, phytoplankton was found to react more sensitively [lowest EC₅₀ (96 h) = 14.5 mg/l] than zooplankton [lowest EC₅₀ (96 h) = 22.43 mg/l] [89]. Diclofenac is commonly found in wastewater at median concentration of 0.81 μg/l whereas the maximal concentration in wastewater and surface water is up to 2 μg/l [90]. Acetylsalicylic acid affected reproduction in D. magna and D. longispina at concentrations of 1.8 mg/l [90]. Water flea Daphnia magna population growth rate was considerably reduced for concentrations ranging from 0 to 80 mg/l due to chronic toxicity of ibuprofen. Acute toxicity tests showed that naproxen had LC₅₀ and EC₅₀ values within the 1–100 mg/l range for the water flea Ceriodaphnia dubia, the rotifer Brachionus calyciflorus, and the fairy shrimp Thamnocephalus platyurus. But the most sensitive reported species was D. magna for which EC₅₀ values were 30.1 or 50 mg/l. Another most commonly prescribed NSAID is paracetamol which is present in concentration below to 20 ng/l to 4.3 μg/l in STP effluents; in surface waters, the values can reach 78.17 μg/l, which are values higher than the predicted no-effect concentration (PNEC) of 9.2 μg/l [3]. Hence, paracetamol might represent a threat for nontarget organisms.

2.3.3 Blood Lipid-Lowering Agents

Statins have the capability to subdue synthesis of the juvenile hormone in insects and may also produce detrimental effect to protozoan parasites, inhibiting growth and development. Reports suggested that a proliferation of peroxisomes in rodent livers is caused by fibrates. Embryonic development of nontarget organisms that share these receptors can be stopped by simply inhibiting cellular differentiation. Fibrates present in the micromolar concentration range are sufficient to cause it in zebrafish (Danio rerio) and amphibians [3]. Quinn et al. [91] classified gemfibrozil as toxic (EC₅₀ between 1 and 10 mg/l) and bezafibrate as harmful for nontarget organisms (EC₅₀ between 10 and 100 mg/l).

Clofibrate is classified as harmful to aquatic organisms as it showed LC₅₀ values in the range of 7.7–39.7 mg/l. The fish Gambusia holbrooki [LC₅₀ (96 h) = 7.7 mg/l] seems to be the most sensitive organism to acute clofibrate concentrations [92]. Clofibrate has an immunosuppressive action in mammalian hosts, suppressing the production of IgM but not IgE antibodies, allowing an amplified number of encysted larvae of the nematodes T. spiralis and Trichinella nelsoni to occur and a decrease in the rate of exclusion of adult worms from the intestines, although the effects differed between parasite species and host strain [93]. Fibrates have been assessed by conventional toxicity tests and the following no-observed-effect-concentration (NOEC) were found for clofibric acid in C. dubia [NOEC (7 days) = 640 μg/l], the rotifer B. calyciflorus [NOEC (2 days) = 246 μg/l], and in early life stages of zebrafish [NOEC (10 days) = 70 mg/l] [94]. Clofibrate was observed to produce no effect on in vitro growth of T. bruceii but did reduce the incidence of P. berghei and the invasiveness and development of Acanthomoeba culbertsoni in exposed mammalian hosts [95]. Lovastatin hinders the egg production of the trematode S. mansoni and subsequently there is a decline in pathogenic granulomas typically associated with the eggs in the mammalian liver [96].

2.3.4 Beta-blockers

Beta-blockers act by competitive inhibition of beta-adrenergic receptors which is critical for normal functioning in the sympathetic branch of the vertebrate autonomic nervous system. Among beta-blockers, propranolol shows the highest acute toxicity and highest log K _ow which proves the fact that it is a strong membrane stabilizer than other examined beta-blockers [97]. Undefined antagonists such as propranolol may be active in fish as they contain β₂-receptors in heart and liver as well as in reproductive tissues [98]. There is a prominent evidence that propranolol not only has chronic cardiovascular toxicity, but also has toxic effect on reproduction system. The NOEC and lowest-observed-effect-concentration (LOEC) of propranolol affecting reproduction in C. dubia were 125 and 250 μg/l, and reproduction was affected after 27 days of exposure in H. azteca at 100 μg/l [97].

Beta-blockers may also affect parasite functional biology. Aqueous exposure of propranolol may negatively affect swimming behavior, survival, and phototaxis of free living aquatic stages of trematodes. Propranolol may also considerably decrease the number of Dirofilaria immitis nematode larvae capable of finishing third-stage molt, and in vitro prevent the growth of the malaria parasite Plasmodium falciparum [99]. Fathead minnows exposed to atenolol throughout embryo-larval growth showed NOEC and LOEC values for growth rate of 3.2 mg/l and 10 mg/l, respectively [3]. At 48-h exposure to propranolol, LC₅₀ values of 29.8, 1.6, and 0.8 mg/l were obtained for H. azteca, D. magna, and C. dubia, respectively, while acute exposure to nadolol did not affect the survival of the invertebrates [3]. Encystment of the protozoan Entamoeba invadens was inhibited in the presence of metoprolol [100].

2.3.5 Antineoplastic Drugs

Antineoplastic drugs are designed to kill the proliferating cells in cancer. As a consequence, a parallel effect can be expected on normally growing eukaryotic organisms. It is expected that antineoplastic drugs possess mutagenic, genotoxic, teratogenic, carcinogenic, and fetotoxic properties, and 14–53 % of the administered drugs can be excreted in unchanged form through urine [101]. Methotrexate revealed teratogenicity for fish embryos with an EC₅₀ of 85 mg/l after 48 h of exposure and acute effects in the ciliate Tetrahymena pyriformis with an EC₅₀ for 48 h of 45 mg/l [102]. Due to immunosuppressant property, methotrexate and cyclophosphamide are reported to cause a proliferation in disease incidence and intensity in host–parasite systems [103]. Acute toxicity of methotrexate is reported on highly proliferative species like the ciliate Tetrahymena pyriformis [EC₅₀ (48 h) = 45 mg/l] [104]. On the contrary, cyclophosphamide appears to have a little effect on them. Methotrexate has been shown to have no or little effect on certain protozoans including Toxoplasma gondii, Babesia bovis, and Leishmania tropica, perhaps as they have different mechanisms of drug metabolism [105]. Development and growth of helminths in both mammalian and bird hosts were detrimentally effected by methotrexate and cyclophosphamide. Abnormal teratogenicity was noticed in fish embryos at higher concentrations [EC₅₀ (48 h) = 85 mg/l]. Biomphalaria glabrata, a freshwater snail is largely affected with the long-term exposure to methotrexate [106]. Doxorubicin, tamoxifen, and methotrexate have all been reported as effective parasiticide agents against many protozoan species [107].

2.3.6 Neuroactive Compounds (Antiepileptics, Antidepressants)

A very limited number of studies on the effects of neurological agents on host–parasite dynamics have been studied, despite phenothiazine has been used as a parasiticide for long time [108]. The serotonin re-uptake inhibitor (SSRI) fluoxetine is deceptively the most acute toxic human pharmaceutical with toxicity ranging from EC₅₀ (48 h, alga) = 0.024 mg/l to LC₅₀ (48 h) = 2 mg/l so far [2]. Sertraline exhibits highly toxic properties to rainbow trout (LC₅₀ of 0.38 mg/l) at a 96-h exposure [109]. SSRIs were also tested on algae by evaluating the growth inhibition induced. Chronic toxicity tests proved that the organisms were sensitive with NOEC values below 1 mg/l [110]. C. vulgaris was shown to be the least sensitive species for all SSRIs tested. Fluvoxamine provided escalation to the highest EC₅₀ values for all algae species tested (3563–10,208 μg/l).

Under the category of benzodiazepines, diazepam and nitrazepam were identified to increase the number of microfilariae of Setavia cervi liberated from the lungs into the peripheral blood circulation in rats [111]. Caffeine was found to stimulate the growth of Plasmodium gallinaceum and P. falciparum, while the antipsychotic haloperidol and the mood stabilizer valproic acid effectively inhibited the in vitro growth of T. gondii [112]. Diazepam and carbamazepine (antiepileptics) are classified as potentially detrimental to aquatic organisms as most of the acute toxicity data are below 100 mg/l. Conventional toxicity studies showed chronic toxicity of carbamazepine in C. dubia [NOEC (7 days) = 25 μg/l], in the rotifer B. calyciflorus [NOEC (2 days) = 377 μg/l], and in early life stages of zebrafish [NOEC (10 days) = 25 mg/l] [94]. Carbamazepine is carcinogenic to rats but does not have mutagenic properties in mammals [113]. It is also lethal to zebrafish at the 43 μg/l level and produces sublethal changes in Daphnia sp. at 92 μg/l [113]. Growth of D. magna was inhibited for concentrations of carbamazepine above 12.7 mg/l, showing acute toxicity at 17.2 mg/l [113].

2.3.7 Sex Hormones

Sex hormones are one of the extremely important biologically active compounds emerged as most serious aquatic environmental toxicants due to extensive use of human contraceptives. Exposure of mammalian hosts infected with the blood trematode S. mansoni to contraceptive pills resulted in a noteworthy modification in a range of liver cell’s ultrastructure and function. Ethinylestradiol (EE2) is a synthetic estrogen found in oral contraceptive pills with noticeable estrogenic effects in fish. The life-cycle exposure of fathead minnows to EE2 concentrations below 1 ng/l produced a noteworthy decline in fertilization success, an increased egg production and decreased expression of secondary male sex characteristics. Life-long exposure of zebrafish to 5 ng/l to EE2 has led to reproductive failure due to the nonexistence of secondary male sex characteristics [63]. Exposure to 17β-estradiol caused an increased susceptibility to the protozoan T. gondii in mice, while increased pathology occurred in mammals infected with Leishmania mexicana amazonensis and exposed to either estradiol or testosterone [114]. Estradiol increased the susceptibility of cyprinids to hemoflagellates by the suppression of lymphocyte proliferation [115]. At relatively high concentrations, hydrocortisone can cause an increase in the intensity of ectoparasitic infections in fish.

2.3.8 Antiparasitic Compounds

A study was performed on farms in the UK and the report suggested that concentrations of antiparasitic compounds of 0.112 mg/kg (doramectin) to 1.85 (ivermectin) mg/kg in dung were found. On the contrary, in same place, concentrations of these drugs in soil were considerably lower up to 0.046 mg/kg [116]. In a study performed in Slovenia, it was found that high concentrations of abamectin and doramectin were found in feces (0.2–0.8 mg/kg and 0.4–1.2 mg/kg, respectively) during the first 20 days after treatment, reaching concentrations of about 0.2 mg/kg after 70 and 50 days, respectively [117]. Grønvold et al. [118] found that ivermectin and fenbendazole affect the survival of the nematode Pristionchus maupasi at concentrations higher than 3 mg dung/kg (w/w) and 10–20 mg dung/kg, respectively. Svendsen et al. [119] showed that ivermectin and the fenbendazole did not affect earthworms. However, the disappearance of dung was affected by the avermectin but not by the fenbendazole. Avermectin B_1A with LC₅₀ value of 17.1 mg/kg in soil was found with the compost worm Eisenia fetida [120]. Eprinomectin did not affect survival or biomass of the earthworm species Lumbricus terrestris in laboratory tests at concentrations up to 0.43 mg/kg dung (w/w) or 3.3 mg/kg dung [121].

2.3.9 Antivirals

Tamiflu [oseltamivir ethylester-phosphate (OP)] and Relenzas (zanamivir) belong to a novel class of antiviral drugs under the neuraminidase inhibitors category. National storing of neuraminidase inhibitors in the USA began with the emergence of the 2009 influenza pandemic (H1N1) [122]. Tamiflu tablet largely dominated Relenza (disk inhaler) due to its relative ease of administration. Tamiflu is a prodrug, which is converted to the active drug oseltamivir carboxylate (OC) in the liver. About 80 % of an oral dose of Tamiflu is excreted as OC in the urine and the remaining portions are excreted as OP in the feces. Therefore, both the parent chemical and its bioactive metabolite eventually are projected to reach a mean of 2–12 mg/l in WWTPs during a moderate and severe pandemic [122]. Current evidences suggested that rivers receiving WWTP effluent would also be exposed to OC throughout a pandemic. The OC concentrations between 293 and 480 ng/l have been recorded in rivers receiving WWTP effluent during the 2009 pandemic [123].

2.3.10 Pharmaceuticals Mixtures

Pharmaceuticals are identified as multicomponent mixtures rather than isolated pure substance in diverse environmental compartments. Majority of pharmaceuticals will either be transformed by physical and chemical means and/or subsequently biotransformed by some organisms. Multicomponent mixtures are the foremost concerning issue for the ecotoxicity. The following characteristics also make their joint toxic effects a major issue for hazard and risk assessment:

• 1. The toxicity of a mixture has always a synergistic effect than the effects produced by a single component.

• 2. A mixture can have a substantial ecotoxicity, even if all components exist only in low concentrations that do not aggravate noteworthy toxic effects if acting separately on the exposed systems.

A combination of fluoxetine and clofibric acid is lethal for more than 50 % of a water-flea (Daphnia) population after an exposure of 6 days, although the individual drugs did not show any significant effect when present separately at same concentrations [124]. A substantial swing in sex ratio was perceived after an exposure to a three-component mixture of erythromycin, triclosan, and trimethoprim. Again, individual components did not elicit significant individual effects. These studies are very important to show that mixture effects have to be taken into consideration to identify the effects of pharmaceuticals.

2.4 The Environmental Risk Assessment

Exposure assessment is the procedure of determining or assessing the intensity, frequency, and extent of environment and human exposure to an existing pharmaceutical product, or of estimating theoretical exposure that might rise from the discharge of new pharmaceuticals into the environment. The concept of “exposomics,” which integrates a top-down and bottom-up approach to identification of relevant exposure biomarkers, will be an important component of future exposure science [125]. The major aims of environmental risk assessment (ERA) should be risk mitigation and risk management. In order to alleviate or accept risks, a risk assessment has to be performed both for products and for activities followed by generation of report based on the characteristics of the product, its possible environmental exposure, fate and effects, and risk extenuation strategies. The inference of the report should be based on sound scientific reasoning supported by adequate studies. If other applicable data are accessible, they should also be submitted.

The outline of the registration process and the ERA consist of European Commission and Council directives and regulations on registration, European policy, case law, and global (trade) agreements. The decision-making process and the risk models should elevate the expenses to society in terms of ecotoxicity and financial loss. Also the assessment method itself should obstruct neither product development nor timely action to eradicate hazards.

2.4.1 Risk Assessment Approaches

The most commonly employed approaches for risk assessment are hazard identification, dose-response assessment, exposure assessment, and risk characterization of pharmaceuticals and its metabolites in various environment compartment [126].

Hazard Identification

The first step for risk assessment is hazard identification which supports the intensity of risk for a particular product. Although in vitro test studies provide useful data on the toxicity of environmental hazards, the majority of scientists rely heavily on the outcome of animal toxicity tests for hazard identification. As a consequence, a greater stress should be provided on the implication of in vitro assays in human cells and QSAR analysis, as well as the use of computational techniques in systems biology [127].

Dose-Response Assessment

Identification of the threshold dose of the toxic effect of any product is very much essential for scientific risk assessment. Dose-response information over a wide range of test concentrations should be assessed employing Quantitative high throughput screening (q-HTS). There should be availability of sensitive assays capable of detecting toxicity at very low doses or below environmental levels experienced by human populations. Statistical approaches can be used to estimate yardstick concentrations for adaptive and adversarial responses and to assess critical concentrations [128]. As discussed in subheading “Hazard Identification”, the extrapolation techniques will be required to interpret in vitro test results to in vivo utilizing an appropriate internal tissue dose metric [129].

Dose and Species Extrapolation

The major problems of risk assessment are low-dose and interspecies extrapolation. In silico models and expert systems have supported such extrapolations, including linear and threshold models for low-dose extrapolation and body weight or surface area alterations for interspecies extrapolation. New extrapolation complications are dose extrapolation of molecular and cellular pathway responses, and extrapolation from the short-term in vitro to longer term in vivo exposure. In vitro to in vivo extrapolation and physiologically based pharmacokinetic (PBPK) models are amenable to sensitivity, variability, and uncertainty analysis employing conventional tools [130]. Computational biology systems will back the application of tools for determining variability and uncertainty from the pharmacologically based pharmacokinetics (PBPK) information as the pathway components imitate more targeted molecular elements and their interactions [131].

Exposure Assessment

In present scenario, human exposure assessment is made principally on the measured levels of environmental agents in the human environment [132]. In few cases, internal dose measures may also be calculated using biomonitoring [133] or pharmacokinetic modeling [134]. For superior exposure assessment, the focus should be more on direct measures of critical toxicity pathway agitations in humans by employing innovative biomonitoring techniques coupled with advanced new high throughput approaches [135].

2.4.2 Environmental Risk Assessment Modeling of Pharmaceuticals

The risk assessment model consists of the risk assessment process, including their harmonization and communication with the risk management process. The risk model interprets the safety issues in quantities like probabilities, concentrations, dosages, and risk quotients of each pharmaceutical product. The simplest approaches to estimating concentrations of a pharmaceutical in diverse compartments are provided in the guidance for environmental assessments for regulatory drug approvals by the US FDA [30] or the EU EMA [28]. In Fig. 2, the risk assessment is harmonized with risk management process.

Fig. 2

Possible steps for risk assessment and risk management

Before designing or modeling a toxicological study, it is very beneficial to assess exposure of any pharmaceutical by the following way [136]:

• The exposure is measured in form of the environmental concentration (occurrence) to which the biological system is exposed, the duration and frequency being not on the concentrations to which each individual is actually exposed. The actual exposure is subjected to many other factors such as, the fate, sorption effects, metabolism and transformation processes.

• The life stage and behavioral patterns should also be taken into account for any organism or living system.

• The bioavailability and toxicokinetics of the drug are studied.

• The pathways and target sites in the biological system are explored.

• The mode of action which depicts steps and processes to molecular and functional effects is determined.

• As pharmacokinetics and pharmacodynamics can influence the dose of pharmaceuticals, one has to consider these aspects to assess the dose which ultimately reaches the environment taking into account possible absorption, distribution, and elimination mechanisms.

• The hazard due to the inherent toxicity of the pharmaceutical according to its chemical properties is also studied.

2.5 Risk Management

Risk management is “the process of identifying, evaluating, selecting, and implementing actions to reduce risk to human health and to ecosystems. The goal of risk management is scientifically sound, cost-effective, integrated actions that reduce or prevent risks while taking into account social, cultural, ethical, political, and legal considerations” [137]. For eco-friendly risk management, one may select a combination of apposite tactics to balance risks, costs and benefits, taking into account social values and economic considerations.

2.5.1 Implementation of Precautionary Measures

The application of pharmaceuticals and their after use toxic effects cannot be stopped but the probable risk of pharmaceutical products related to environmental can be controlled by implementing proper precaution and safety measures. The EMEA 2006 guideline demonstrates following steps as safety measures for risk management:

• 1. Calculation of product risks initially

• 2. Proper product labeling and summary product characteristics (SPC)

• 3. Package leaflet (PL) for each pharmaceutical for patient use to inform the probable toxic effects

• 4. Appropriate and safe storage of pharmaceutical product

• 5. Safe and proper scientific disposal of pharmaceuticals

2.5.2 Lessening the Input of Pharmaceuticals into the Environment

To diminish the occurrence of pharmaceuticals into the different compartments of the environment, one has to follow the principle of sustainability where the entire life cycle of a pharmaceutical has to be taken into consideration to categorize the opportunities for risk management. For diminishing the input of pharmaceuticals into the environment, following steps can employed effectively [138].

Training and Awareness

The most important step to reduce the occurrence of pharmaceuticals in the environment is proper training and awareness of users who are the major source points. A proper usage and disposal of pharmaceutical is the responsibility of the shareholders and people using the compounds, including patients, doctors and nurses, and pharmacists. Industrial sectors should have the major role to treat the failed active pharmaceutical under quality control category properly before it reach to the environment. Additionally, each pharmaceutical product should consist of materials safety data sheet (MSDS) intended to provide workers and emergency personnel with procedures for handling or working with that substance in a safe manner, information such as physical data, toxicity, health effects, first aid, reactivity, storage, disposal, protective equipment, and spill-handling procedures. Appropriate and effective risk management strategies need basic knowledge of entry routes of pharmaceuticals. Therefore, one has to identify the bulk of drug flows connected with the diverse sources of pharmaceuticals such as households, industries, hospitals and pharmacy.

Improved Sewage Treatment

The most technical and extensively considered approach for risk management is improvement of sewage treatment. Analyzing Table 1, one can easily identify the presence of threatening amount of pharmaceutical wastes after sewage and waste water treatment also. The purpose of advanced and improved sewage and waste water treatment is to further reduce the ecotoxicity, hormonal effects and pathogenic effects of the effluent. In recent years, advanced effluent treatment has been studied extensively. The advanced treatment of sewage influents and effluents as well as waste water treatment can be done employing photochemical oxidation processes, filtration, and application of powdered charcoal and constructed wetlands [139].

Green and Sustainable Pharmacy

The third approach is evolving from the knowledge of green and sustainable pharmacy which states that substitution of the compound with a more environmentally benign compound [138]. Though this approach is less practiced, in terms of sustainability, it appears to be the most encouraging one in the long run. The prime principle of green chemistry is easy and fast degradability of pharmaceuticals after their application. Understanding of full life cycle of drugs will lead to a different understanding of the functionality necessary for a pharmaceutical.

Additionally, other crucial issues like (a) development of improved drug delivery systems so that lower doses are required; (b) upgradation of packaging and package sizes to prolong shelf life and lessen the amount of the product that expires and rejection of unused products; and (c) changes in prescription and animal farming practices are substantial options for minimizing or eliminating emissions to the environment. Potential processes and measures to decrease the environmental toxicity by carious stakeholders are addressed in Fig. 3 for a better understanding.

Fig. 3

Probable actions to be taken for reduction of the occurrence of pharmaceuticals in the environment by different stakeholders

3 Regulatory Agencies for the Risk Assessment and Management of Ecotoxicity Pharmaceuticals

Immense exposure of pharmaceuticals and their metabolites to the environment is a matter of concern and a burning global issue at recent times. The risk effects are not only related with the environment, it is also directly related to human health to a large extent. As a consequence, release of these pharmaceutical products, their risk assessment as well as risk management are controlled and regulated at local, national and international levels by different governments and regulatory agencies worldwide. As experimental data of environmental fate and toxicity of pharmaceuticals are absent or some time not sufficient, there is a strong urge to predict physical and chemical properties, environmental fate, ecological effects and health effects of pharmaceuticals and their metabolites. Several government organizations have been applying the approaches of structure–activity relationship (SAR) and QSAR to develop the predictions for untested existing as well as newly introduced pharmaceuticals. To establish proper identification of environmental hazards, their risk assessment and fate modeling, SAR and QSAR approaches along with other predictive in silico tools are employed by Australian, Canadian, Danish, European, German, Japanese, Dutch, and US Government organizations [28, 30, 140–144].

QSAR models can be generated for prediction of the following ecotoxicity related properties or effects:

• 1. Physicochemical properties

• 2. Toxic potential and potency

• 3. Environmental distribution and fate in different compartments (air, water and soil) of environment

• 4. Biokinetic processes (absorption, distribution, metabolism, and excretion) of pharmaceuticals and their metabolites

Areas where QSARs can be applied by governmental regulatory agencies are as follows:

• 1. Prioritization of existing pharmaceuticals for toxicity testing to environment.

• 2. Classification and labeling of new pharmaceuticals according to their safe use.

• 3. Risk assessment of new and existing pharmaceuticals.

• 4. Guiding experimental design of regulatory tests or testing strategies.

• 5. Providing mechanistic information

• 6. Filling up the large data gaps.

• 7. Building a proper database of each pharmaceutical to different species regarding environmental toxicity.

• 8. Development of expert systems for each therapeutic classes for different compartments of the environment.

• 9. Construction of efficient interspecies models to extrapolate data from one species to another species when data of a particular species is absent.

Global regulatory authorities and agencies [28, 30, 140–144] for the risk identification, risk assessment and finally risk management of ecotoxicity pharmaceuticals are listed in Table 2.

Table 2 Global regulatory bodies and agencies for the hazard and risk assessment of ecotoxicity pharmaceuticals

The most common endpoints associated with various test methods proposed under Organization for Economic Co-operation and Development (OECD) are the following ones:

• Physical-chemical properties: Most commonly evaluated properties are melting point, boiling point, vapor pressure, octanol–water partition coefficient, organic carbon–water partition coefficient, and water solubility.

• Ecological effects on endpoints: Acute fish-toxicity, long-term toxicity, acute Daphnia toxicity, algal toxicity, terrestrial toxicity, marine organism toxicity, microorganism toxicity in sewage treatment plant.

• Environmental fate: Biodegradation, hydrolysis in water, atmospheric oxidation, and bioaccumulation;

• Human health effects: Acute oral, acute dermal, acute inhalation, eye irritation, skin irritation, skin sensitization, repeated dose toxicity, genotoxicity, reproductive toxicity, developmental toxicity, systemic toxicity, mutagenicity, carcinogenicity, etc.

OECD’s database on risk assessment models:

In silico models that are employed by the OECD countries to predict health or environmental hazards, exposure potential, and probable effects were organized into a searchable database. This database is intended as an information resource only. The models are listed by countries and by the property or effect included. The models can be useful as a screening tool, when there is a lacking of chemical-specific data, for establishing priorities for chemical assessment and for identifying issues of potential concern [140]. Areas of assessment and category of information for predicting human health and environment according to OECD’s guidelines are represented in Figs. 4 and 5, respectively.

Fig. 4

Areas of assessment and risk models for predicting human health and environment according to the OECD database

Fig. 5

Category of information included in predicting health and environmental effects according to the OECD guidelines

4 In Silico Modeling of Ecotoxicity Using SAR and QSAR Approaches

The toxic potential of large quantities of industrial chemicals including pharmaceuticals, cosmetics, pesticides and other synthetic or semisynthetic chemicals is often required to be assessed by using standard animal models, comprising the basic test protocol for risk assessments for their approval as a registered product to launch into the market. With increasing concern about the environmental pollution and human health, the manufacture, storage, distribution, and release of these hazardous substances after their application to the environment are controlled and regulated at various levels by different governments and regulatory agencies worldwide. Applications of analogues, SAR and QSAR of different pharmaceuticals are also providing useful information in a regulatory decision making context in the absence of experimental data [140]. Most commonly employed predictive in silico tools are depicted in Fig. 6.

Fig. 6

Predictive in silico tools for the prediction of ecotoxicity of pharmaceuticals

Among the available in silico predictive models for ecotoxicity, majority of the models are constructed employing QSAR techniques. Therefore, in this book chapter, a special importance is given to the discussion of QSAR models. The QSAR approach attempts to correlate structural/molecular properties with biological activities/toxicities, for a set of compounds by means of statistical methods. As a result, a simple mathematical relationship is established:

$$ \mathrm{Biological}\kern0.28em \mathrm{activity}\kern0.28em \mathrm{or}\kern0.28em \mathrm{toxicity}\kern0.28em =\kern0.28em f\left(\mathrm{chemical}\kern0.28em \mathrm{structure}\kern0.28em \mathrm{or}\kern0.28em \mathrm{property}\right). $$

Applications of QSAR can be extended to any molecular design purpose, prediction of different kinds of biological activities and toxicities, lead compound optimization, classification, diagnosis, and elucidation of mechanisms of drug action, toxicity prediction of environmental toxicants (pollutant pharmaceuticals, chemicals, gas, etc.), and prediction of drug-induced toxicity [145]. The major objective of structure–activity/toxicity relationship modeling is to investigate and identify the decisive factors for the measured activity/toxicity for a particular system, in order to have an insight of the mechanism and behavior of the studied system. For such a purpose, the employed strategy is to generate a mathematical model that connects experimental measures with a set of chemical descriptors determined from the molecular structure for a set of compounds. The derived model should have as good predictive capabilities as possible to predict the studied biological/toxicological or physicochemical behavior for new compounds. The factors governing the events in a biological system are represented by a multitude of physicochemical descriptors, which can include parameters to account for hydrophobicity, electronic properties, steric effects, and topology, among others [145].

With the constant progress of QSAR techniques, many methods, algorithms, and techniques have been discovered and applied in QSAR studies. The development of a QSAR model follows five major steps:

• 1. Selection of a dataset with series of known response data

• 2. Calculation of descriptors

• 3. Splitting of the dataset into training and test sets for model development and its subsequent validation

• 4. Construction of models using different chemometric tools, and

• 5. Validation of the developed model based on internal and external validation statistics

Additionally, the development of 3D-QSAR models includes two more steps for their successful execution: conformation analysis of the molecules and their alignment status with respect to the most active compound. The most important feature for an acceptable and reliable QSAR model is predictive capability for new set of compounds. The predictive quality of the developed model is determined based on different validation statistics. Thus, validation of QSAR models plays the most crucial role in defining the applicability of the QSAR model for the prediction of untested compounds. Initially, verification of the correlation between chemical features of the molecules and the biological activity/toxicity was of prime interest during the development of a QSAR model. Later, the focus gradually shifted toward the predictive power of the model than simply unveiling the quantitative relationships [146].

To validate a QSAR model, one has to follow OECD principles for acceptable predictions in order to make the model as a reliable screening tool for future toxicity prediction of untested pharmaceuticals. A meeting of QSAR experts held in Setúbal, Portugal in March 2002 reported guidelines for the validation of QSAR models for regulatory purposes. The OECD principles were agreed by OECD member countries, QSAR and regulatory communities at the 37th Joint Meeting of the Chemicals Committee and Working Party on Chemicals, Pesticides and Biotechnology in November 2004. These principles are listed here: Principle 1: a defined endpoint; Principle 2: an unambiguous algorithm; Principle 3: a defined domain of applicability; Principle 4: appropriate measures of goodness-of fit, robustness, and predictivity; Principle 5: a mechanistic interpretation, if possible [147]. Different quality metrics for QSAR models can be categorized into two classes: one determining the fitting ability of the model while the other analyzing the predictive potential of the developed model [146].

4.1 Why In Silico Models Should Be Developed for Ecotoxicity Predictions of Pharmaceuticals?

4.1.1 The 3R Concept

The 3R concept represents three words “Reduction,” “Replacement,” and “Refinement”. The concept brought about an imperative notification about animal experimentation in the scientific communities. The word ‘Reduction’ refers to the diminution in number of animals used to get results of a defined precision. Next, ‘Replacement’ corresponds to the use of nonliving resources to replace conscious living higher animals, and ‘Refinement’ means decline in the severity or cruelty of inhuman methodologies applied to the experimental animals [148]. As a consequence, to establish the 3R concept, in silico techniques are one of the front runners.

4.1.2 Ban of Animal Experimentation

There are different social as well as governmental organizations that consider reduction or complete ban of animal experimentations [149]. Here, we have listed a few of them:

• 1. The European Centre for the Validation of Alternative Methods (ECVAM) was established in the year 1991 that agrees the principle of 3Rs.

• 2. The European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Procedures.

• 3. Council Directive 86/609/EEC on the Approximation of Laws, Regulations and Administrative Provisions of the Member States Regarding the Protection of Animals Used for Experimental and Other Scientific Purposes.

• 4. Johns Hopkins Center for Alternatives to Animal Testing (CAAT), a US based organization focussing on the reduction of animal experimentations.

• 5. The testing ban on the finished cosmetic products applies since 11 September 2004; the testing ban on ingredients or combination of ingredients applies since 11 March 2009. The marketing ban applies since 11 March 2009 for all human health effects with the exception of repeated-dose toxicity, reproductive toxicity, and toxicokinetics. For these specific health effects, the marketing ban applies since 11 March 2013, irrespective of the availability of alternative non-animal tests [150].

• 6. India and Israel have also banned animal testing for cosmetic products, while the USA has no such ban in place [151].

• 7. China is the only major market where testing all cosmetics on animals is required by law, and foreign companies distributing their products to China must also have them tested on animals. [152] China has announced that its animal testing requirement will be waived for shampoo, perfume, and other so-called “non-special use cosmetics” manufactured by Chinese companies after June 2014. “Special use cosmetics,” including hair regrowth, hair removal, dye and permanent wave products, antiperspirant, and sunscreen, will continue to warrant mandatory animal testing.

4.1.3 Regulatory Decision Making

SARs and QSARs are employed to predict aquatic toxicity, physical or chemical properties, and environmental fate parameters as well as to predict specific health effects of organic chemicals by Australian, Canadian, Danish, European, German, Japanese, Dutch, and US Government organizations.

4.1.4 Filling Data Gaps

Acceptable toxicity data of pharmaceuticals to environment and human health is considerably less than 5 % [153]. Computer-aided prediction has the competency to assist in the prioritization of pharmaceuticals for testing, and for predicting specific toxicities to allow for classification. As the number of reliable models for toxicity predictions is increasing, they can be employed as one of the major sources for filling the missing data of pharmaceutical toxicity to ecosystem.

4.1.5 Development of Understanding of Biology and Chemistry

In the modeling of acute toxicological endpoints, much has been gained regarding mechanisms of action. For many modeling approaches, it may be assumed that compounds fitting the same QSAR are acting by the same mechanism of action. This has allowed workers to define the chemical domain of certain mechanisms. There are countless examples where knowledge of biology and chemistry has been advanced by modeling in the field of toxicological and fate effects [154].

4.1.6 Cost and Time Reduction

Toxicity study is very costly in terms of the animals employed for testing and time taken. Even a simple ecotoxicological assay may cost several thousand dollars, and a 2-year carcinogenicity assay may cost several million dollars. Cost is a clear issue to fill the data gaps for the many new compounds that have not been tested. On the other hand, prediction of various toxicity endpoints for pharmaceuticals at an early stage of design can save a large amount of expenses for such compounds which may be found toxic at a later stage of drug development program [155].

4.1.7 Identification of New Toxicological Problems

The development of computational techniques not only allows for the prediction of the potential risk of pharmaceuticals but also allows for rational direction to be given to the testing programs.

5 Review of Literature on In Silico Ecotoxicity Modeling of Pharmaceuticals

Kar and Roy [156] have constructed robust quantitative interspecies toxicity correlation models for Daphnia magna and fish evaluating the ecotoxicity of structurally diverse 77 pharmaceuticals. They have demonstrated that the keto group and the

(aasC) fragment are principally responsible for higher toxicity of pharmaceuticals to D. magna. On the other hand, for fish toxicity, along with the keto group, structural fragments like X=C=X, R–C(=X)–X, and R–C≡X are largely responsible for the toxicity. The interspecies models were also used to predict fish toxicities of 59 pharmaceuticals (for which Daphnia toxicities were present) and Daphnia toxicities of 30 pharmaceuticals (for which fish toxicities were present). They established that the interspecies correlation study would permit an improved and inclusive risk assessment of pharmaceuticals for which toxicity data was missing for a particular endpoint.

Das et al. [157] attempted to develop interspecies correlation models taking rodent toxicity as dependent variable so that any drug without reported rodent toxicity can be predicted using fish, daphnia, or algae toxicity data which can be further extrapolated to human toxicity. Interspecies extrapolation QSAR models were developed employing multiple validation strategies. Analyzing the models, the authors concluded that heteroatom atom count and charge distribution were significant determinants of the rodent toxicity, and that the atom level log P contributions of various structural fragments and various extended topochemical atom (ETA) indices reflecting electronic information and branching pattern of molecules were important determinants for the rodent toxicity. In addition, from interspecies aquatic toxicity modeling, it was established that apart from the algae toxicity, atom level log P contributions of different fragments, charge distribution, shape, and ETA parameters were important in describing the daphnia and fish toxicities in the interspecies correlation models with algae toxicity. The toxicity of chemicals to rodents bears minimum interspecies correlation with other mentioned nonvertebrate and vertebrate toxicity endpoints.

The acute toxicity was predicted (>92 %) using a generic quantitative structure–toxicity relationship (QSTR) model developed by Sanderson and Thomsen [158] suggesting a narcotic mechanism of action (MOA) of 275 pharmaceuticals. An analysis of model prediction error suggests that 68 % of the pharmaceuticals have a nonspecific MOA. Authors have compared the measured effect data to the predicted effect concentrations using ECOSAR regarding the predictability of ecotoxicity of pharmaceuticals and accurate hazard categorization relative to Global Harmonized System (GHS). Molecules were predicted using the model resulting in 71 % algae, 74 % daphnia, 83 % fish datasets that could be compared.

Escher et al. [159] constructed QSAR models with the total toxic potential of mixtures of the β-blockers and related human metabolites for the phytotoxicity endpoint. They have assumed two scenarios for this study. In the first scenario, the metabolites lose their explicit activity and act as baseline toxicants. In the second scenario, the metabolites reveal the identical specific mode of action like their parent drug. β-Blockers are secondary amines and are, therefore, fully protonated at environmental pH. The authors accounted for their positive charge in the QSAR analysis and have experimentally determined the liposome–water partition ratios at pH 7 to make QSAR analysis more robust.

Berninger and Brooks [160] considered the mammalian Acute to Therapeutic Ratio (ATR) to predict pharmaceuticals which may result in comparatively high Acute to Chronic Ration (ACR) in fish models. The authors identified a statistically significant relationship between mammalian ATRs and fish ACRs (p < 0.001, r ² = 0.846). In this model, they only included chronic responses of fish to pharmaceuticals which appear to have been elicited through a therapeutic MOA for calculating ACRs and for statistical analysis of the relationship with mammalian ATRs. Utilizing this approach, mammalian ATR values can be used for predicting pharmaceuticals with higher fish ACRs if the chronic response used in ACR calculation is reasonably linked to the therapeutic MOA of a pharmaceutical.

Sanderson et al. [161] employed the US EPA generic aquatic (Q)SAR model ECOSAR to screen more than 2800 pharmaceuticals and provided a baseline to fill the screening data regarding parent pharmaceuticals environmental toxicity. The model can be used to predict both acute and chronic aquatic toxicity.

Sanderson and Thomsen [162] overestimated the toxicity for 70 % of the 59 pharmaceuticals by ECOSAR v3.20 which contains both measured and modeled data. For the remaining 30 % pharmaceuticals, more than 94 % of the predictions underestimated toxicity by less than a factor of 10. This is an indication that a narcosis based model is conservative relative to experimental values around 70 % of the time, thus implying that for at least 70 % of the Active Pharmaceutical Ingredients (APIs), the acute mode of action (MOA) can be elucidated by baseline toxicity. The authors have observed determination coefficients (r ²) ranging from 0.73 to 0.76 between all the modeled Log EC₅₀ and Log K _ow. The slopes of the Log EC₅₀–Log K _ow regressions based on measured data from the USA National Oceanic and Atmospheric Administration (NOAA) database for both fish and daphnia equal −0.86 which suggest a narcotic MOA.

Lienert et al. [163] assessed the ecotoxicological risk potential of 42 pharmaceuticals from 22 therapeutic classes, including metabolites formed in humans. They considered each parent drug and its metabolites as a mixture of equally acting compounds, and in case when effect data were missing, they estimated these with QSAR models. They have collected data on the identity and excretion pathways of human metabolites and, where available, experimental ecotoxicity data (EC/LC₅₀) from pharmaceutical compilations and from diverse literature sources. They have compiled physicochemical data like structure, molecular weight, octanol–water partition coefficient K_ow, acidity constant pK _a mainly from the Physical Properties Database (http://www.syrres.com/esc/physprop.htm). Moreover, they have generated a risk quotient (RQ_mixture) using simple predictions of drug concentrations in wastewater which can be useful for risk assessment of pharmaceuticals.

Christen et al. [164] developed VirtualTox Lab [165] to predict the effects of pharmaceuticals in the aquatic system. The study leads to the inference that the mode of action perception is most appropriate for the identification of highly active compounds (HC). As suggested by the authors, modification can be done by balancing this concept by the QSAR model (VirtualTox Lab), whereas the fish plasma model seemed to be less apposite due to the requirement of environmental concentration above 10 ng/l for the identification of a risk. The practice of the VirtualTox Lab will support the mode of action concept and may be beneficial to recognize surplus targets of the pharmaceutical to assess the ecotoxicity.

Escher et al. [166] predicted baseline toxicity of the 100 molecules using established QSARs for algae, daphnia, and fish. The QSARs were selected from the Technical Guidance Document of the EU. The logarithm of D _lipw (liposome water distribution coefficient) was employed in the model development for baseline toxicity to calculate the toxicity of the compound towards the stated species.

The environmental risk assessment of 26 pharmaceuticals and personal care products have been performed by De García et al. [167] based on the ecotoxicity values generated by bioluminescence and respirometry assays. Then the compounds were classified following the Globally Harmonized System of Classification and Labelling of Chemicals by predictions using the US EPA ecological structure–activity relationship (ECOSAR™). The real risk of impact of these pharmaceuticals in wastewater treatment plants (WWTPs) and in the aquatic environment was predicted according to the criteria of the European Medicines Agency. According to their studies, in at least two ecotoxicity tests, 65.4 % of the PPCPs showed prominent toxicity to aquatic organisms. There study showed some type of risk for the aquatic environments and/or for the activated sludge of WWTPs for pharmaceuticals like acetaminophen, ciprofloxacin, clarithromycin, clofibrate, ibuprofen, omeprazole, triclosan, parabens, and 1,4-benzoquinone.

Here we have discussed available in silico models on ecotoxicity of pharmaceuticals. Due to the limited availability of the in silico models on ecotoxicity of pharmaceuticals, there is a need to develop more in silico models in order to reduce time and cost involvement as well as reduction of animal usage in getting relevant data and for better and fast risk assessment of pharmaceuticals. It is not possible to experimentally study toxic effects of each pharmaceutical in different species. Most active pharmaceutical ingredients have available rodent toxicity information. As a result, if this data could be extrapolated or modeled to different other species, this would be a noteworthy resource for prioritization of pharmaceuticals with regards to diverse environment hazards. However, very limited papers have been published on interspecies models to predict environmental toxicity for pharmaceuticals, and there are relatively few statistical models available to bridge the chronic toxicity data information gap [168, 169].

6 Endpoints

Toxicity of a molecule should be assayed on specific toxicity endpoints for the generation of data which are employed commonly to develop in silico models. This why a clear concept is required about the endpoints or test batteries as they are employed for the experimental toxicity studies and for understanding the mode of toxicity with respect to that particular endpoint [110]. We list the most commonly employed endpoints for this purpose in Table 3.

Table 3 Most commonly employed test batteries (endpoints) for the modeling of ecotoxicity

7 Databases

A good quality of ecotoxicological data of pharmaceuticals and their metabolites is required for the development of accurate and reproducible in silico models. A significant number of chemical/drug/agrochemical/pesticide toxicity databases towards environment are publicly accessible, and such numbers are growing. But one cannot deny that the existing databases are very few compared to drug discovery compound libraries. Recent initiatives requiring superior use of in silico technologies have called for transparency and expansion of toxicity database information that is available to the public at no cost. Table 4 represents publicly available toxicity databases describing environmental as well as human health effects of pharmaceuticals useful in risk assessment, risk management, safety evaluation, and hazard characterization.

Table 4 List of available databases comprising information of the ecotoxicity due to pharmaceuticals^a

8 Expert Systems

Expert systems allow for the direct entry of a structure into software followed by the calculation or prediction without the requirement to compute descriptors and re-perform the modeling process. This makes expert system a more convenient option for toxicity prediction over traditional QSARs. Expert systems have been frequently employed by regulatory agencies, academia and industries worldwide for more efficient and fast prediction. The foremost criterion of toxicity prediction is to differentiate between toxicologically active and inactive molecules. Multiple mechanisms can lead to the identical toxic effect and this intricacy requires the accessibility of predictive tools that are able to discriminate manifold regions in the activity space. This necessitates the development of so-called expert systems, which try to cover broader structural and activity regions in comparison to the local models. Table 5 summarizes different freely available and commercial expert systems to predict endpoints related toxicity predictions.

Table 5 A complete list of exert systems to predict various endpoints of ecotoxicity

9 Green and Ecological Pharmacy

The role of green chemistry and principles are very important for risk management of pharmaceuticals. The principles of green chemistry state that the functionality of a chemical should not only comprise the properties of a chemical essential for its application, but also quick and trouble free degradability after its usage. Improvement of synthesis and renewable feedstock are very important issues for preparation of environment friendly pharmaceuticals. Employing these principles and the awareness of green chemistry to pharmaceuticals are necessary [138]. In this perspective, a system called “benign by design” can be considered which means easy degradability after application is considered even before a pharmaceutical’s synthesis. This approach is not completely new. For instance, it is a general practice during the development of pharmaceuticals that adverse side effects are to be taken into consideration. This can also result in economic rewards in the long run and will fit into green pharmacy [170]. But one has to note that a pharmaceutical may also lose its specific therapeutic action due to the structural modification while introducing green chemistry. However, this approach can be employed at least for the optimized and new synthesis routes [170]. Again, it is true that finding good lead compounds is a major task even without considering the environment toxicity issue. However, there is no requirement to find a new lead compound at first. The modification of known lead structures can be the best option to do. Responding to the green and justifiable pharmacy challenge may also result in new marketing opportunities with help of appropriate and scientific research within industry and academia.

10 Overview and Conclusion

A huge number of reports and publications have been published in the last decade about the ecotoxicity due to human and veterinary pharmaceuticals, but it is still too meager to permit us to execute a systematic and precise risk assessment and appropriate risk management. There is still a huge need of filling the missing data gaps in our knowledge. Due to biologically active and persistence nature of pharmaceuticals, they are one of the most serious threats to human health and environment stability. Additionally, their specific modes of action and specific effects on living systems make pharmaceuticals distinctly different from other chemicals. This sole feature is sufficient reason to assess the potential effects of pharmaceuticals in diverse environmental compartments. The problem is more horrifying as the occurrence level of pharmaceuticals in different environmental compartments is largely varied. The variations in drug occurrences from country to country and also within the different regions of a country make the assessment of pharmaceuticals a troublesome job for the environmental scientist. The interactions between pharmaceuticals and natural stressors of aquatic and terrestrial communities remain to be unexplained. Along with that, the proper risk assessment of mixtures of pharmaceutical products is another area where more introspection is required in present times.

In this book chapter, the hazardous effects of the most common therapeutic classes of pharmaceutical to the living ecosystems and environment are discussed. Furthermore, specific information on the sources, fate, and effects of pharmaceuticals in the environment and their possible negative impact on different ecosystems are explored. There is a lack of sufficient information and scientific data on effects of long-term exposure to nontarget organisms. It is also important to assess the presence of pharmaceuticals and their metabolites and transformation products in several environmental compartments. One can find only a few reports on the quantitative effects of pharmaceuticals, but the effects of metabolites are not sufficiently explored by the scientific community. One has to accept that the identification of risk assessment and management are not sufficient if they are not properly implemented in right way. In these perspectives, the major role should be played by government authorities and agencies by implementing various guidelines and rules for the reduction of toxicity of pharmaceuticals to the environment.

Scarcity of adequate ecotoxicity data related to the diverse classes of pharmaceuticals and their metabolites has stalled appropriate computational modeling and development of expert systems. As a consequence, there are only a very limited number of models developed so far for the risk assessment of pharmaceuticals and their metabolites as well as for the pharmaceutical mixtures. Hence, a sufficient number of models should be developed to address the risk assessment and risk management in an efficient way by minimizing the requirement of time, animal testing and cost. This will also help in gathering the ecotoxicity data as soon as a new pharmaceutical product comes to the market. In this perspective, expert systems are more reliable and results may be easily available in no time. There is a need of more expert systems for prediction of toxicity of pharmaceuticals from diverse classes of therapeutic actions and their metabolites against different endpoints. It is true that in silico techniques cannot substitute “wet” experiments but both of them can be utilized together for a better risk management of pharmaceuticals in near future.