Académique Documents
Professionnel Documents
Culture Documents
To cite this article: D. B. Huggett, J. C. Cook, J. F. Ericson & R. T. Williams (2003): A Theoretical Model for Utilizing
Mammalian Pharmacology and Safety Data to Prioritize Potential Impacts of Human Pharmaceuticals to Fish, Human and
Ecological Risk Assessment: An International Journal, 9:7, 1789-1799
To link to this article: http://dx.doi.org/10.1080/714044797
11:33
TJ883-18
INTRODUCTION
The presence and potential hazards of pharmaceuticals in the aquatic environment have received increased attention recently (Daughton and Ternes 1999;
Kummerer 2001). This attention is largely the result of a growing number of peerreviewed papers reporting trace levels of pharmaceuticals in municipal effluent, surface water, groundwater and to a lesser extent, drinking water. For instance, a national
reconnaissance by the United States Geological Survey (USGS) detected 82 pharmaceutical and personal care products in surface waters of the United States (Kolpin
et al. 2002). While pharmaceuticals are being detected in the aquatic environment,
Received 11 November 2002; revised manuscript accepted 11 April 2003.
Address correspondence to Duane B. Huggett, Pfizer Global Research and Development,
Eastern Point Road, Mailstop 8118D-4050, Groton, CT 06340, USA. E-mail: duane huggett@
groton.pfizer.com
1789
11:33
TJ883-18
D. B. Huggett et al.
the levels are low and the science is generally not yet developed enough to determine
whether these low-levels pose a risk to environmental species.
Environmental hazard assessment frameworks based on an initial evaluation of
acute ecotoxicity endpoints have been developed to evaluate the potential impact
of pharmaceuticals on aquatic species (FDA 1998; EMA 2001). The science of these
evaluations centers on: 1) the development of a predicted environmental concentration (PEC) to estimate potential aquatic organism exposure, 2) determining acute
toxicity concentrations using algae, Daphnia sp. and fish, and 3) application of multipliers (safety or assessment factors) to account for uncertainties such as inter- or
intra-species variability or extrapolation from acute to chronic toxicity. Hazard assessment frameworks may trigger investigation of increasingly complex and difficult
to measure chronic toxicity endpoints. Depending upon the properties of the drug
under investigation, these tiered evaluation frameworks may result in a conclusion
that no further action is required or the development of a detailed risk characterization with recommendations for potential precautionary measures as needed. In
addition, properties of the compound such as Log P and bioaccumulation potential
may trigger chronic toxicity evaluation outside of the standard acute toxicity/safety
factor framework.
Acute toxicity tests assess toxicity following short exposures (96 hr) using the
primary endpoint of death. Hence, these tests generally do not assess the potential
of a test substance to produce effects following chronic, low level exposures. Acute
toxicity of pharmaceuticals at environmentally relevant concentrations has been
suggested to not be a significant issue because the acute effects concentrations historically determined are generally well above predicted environmental levels (Webb
2001). The biologically active nature of pharmaceuticals, however, has led some
to suggest that aquatic organisms may be impacted in ways that are not being detected using current test methodologies and data evaluation strategies (Daughton
and Ternes 1999).
The mammalian nonclinical safety profile developed to support registration of
a pharmaceutical and use in a patient population is based on data from a series
of in vitro and in vivo studies (Table 1). In vitro studies may be used to determine
key characteristics of a pharmaceutical such as metabolism and stability, bioavailability, mutagenicity and other selected endpoints (e.g., toxic effects to specific tissues).
In vivo safety studies are typically performed in rodent and non-rodent animal species
using standard, non-clinical toxicology study designs and the intended clinical dosing route. A battery of genotoxicity tests are also conducted. Safety pharmacology
studies are conducted to evaluate potential adverse effects at high dosages of the
drug on critical organs such as the heart, lungs, brain, and other organ systems based
on the pharmacology of the drug. The safety assessment studies typically increase
in duration as development of a drug continues. Testing includes acute through
sub-chronic (up to three month) studies in two species prior to Phase I clinical investigation, where initial human exposure in small populations of healthy subjects
occurs.
In conjunction with the progressive phases of clinical development, non-clinical
studies of increasing duration are performed. During Phase II clinical trials, fertility,
developmental toxicity, and pre/postnatal studies (commonly referred to as Segments or Studies I-III) are conducted to assess the safety of the drug in women of
1790
11:33
TJ883-18
Table 1.
Test
Rat, dog/primate
Rat, dog/primate
6-month rat, 9-12 month dog/primate
Mouse, rat
Rat
Rat, rabbit
Rat
Rat, dog
child bearing potential and chronic studies (6-month duration in rodent, one-year
duration in non-rodent) are conducted to assess safety during long-term use. During Phase III, carcinogenicity studies are conducted in mice and rats if the drug is
intended for chronic usage in patients.
In many ways, fish are not that biochemically different from mammals. Aquatic
vertebrates appear to have very similar enzyme and receptor systems as humans
(Table 2) (Evans 1993). The National Institutes of Health recognize the zebrafish
(D. reno) as a biomedical model to elucidate an understanding of vertebrate development and disease (NIH 2002). A large number of receptors in fish have been
identified through cloning and sequencing techniques in both the central and peripheral organs of fish (Table 2). Sequence homologies for receptors and enzymes
in fish range from 31 to 88 %. For example, -adrenergic receptors, peroxisome proliferation activated receptors (PPAR ), and serotonin receptors (5-HT) have been
identified in representative fish species with sequence homologies of 63, 47 and
72%, respectively (Nickerson et al. 2001; Yamaguchi and Brenner 1997; Andersen
et al. 2000). Since fish have many similar enzyme and receptor systems, this potentially makes them susceptible to similar biochemical and physiological mechanisms
of activation/inactivation. Because of the conservation of enzyme and receptor systems between mammals and these fish, chronic and target organ toxicity identified
in mammalian safety assessments is likely to be useful in predicting the need for
additional toxicity evaluation in teleosts.
While reports indicate that vertebrates may have receptors similar to those in humans (Evans 1993), the data set for invertebrates is much more limited. In addition,
receptor and enzyme systems in invertebrates may be structurally similar to those in
mammals, but be physiologically distinct with regard to functionality. For instance,
17-ethinyl estradiol regulates ovarian activity in mammals, but it does not appear
to elicit aquatic invertebrate reproductive responses through arthropod hormonal
Hum. Ecol. Risk Assess. Vol. 9, No. 7, 2003
1791
11:33
TJ883-18
D. B. Huggett et al.
Table 2.
Species
Homology1
Reference
Receptor
PPAR
ER/ER
AR
5-HT1A /5-HT1D
2 -adrenoceptor
A1A -adrenoceptor
NMDA (NR1 subunit)
GlyR ( subunit)
GluR (R3 subunit)
IGF
IR
AH
OR1
NPY
BK2
IL1
GnRH
FSH
LH
TSH
GH
RAR
SST
S. salar
D. rerio
P. major
F. rubripes
O. mykiss
O. latipes
A.leptorhychus
D. rerio
Oreochromis
D. rerio
Salmon
D. rerio
D. rerio
Cod
D. rerio
S. salar
O. mykiss
I. punctatus
I. punctatus
M. saxatilis
C. auratus
F. rubripes
C. auratus
47
47/47
45
72/71
63
61
88
77
87
63
83
43
66
50
35
31
45
53
47
57
42
58
62
Enzyme
Lipoprotein Lipase
P450 1beta-hydroxylase
COX
P450 Aromatase
Creatine Kinase
Caspase-3
Stearoyl-CoA desaturase
ADH
iNOS
AChE
O. mykiss
O. mykiss
D. rerio
D. labrax
D. rerio
D. rerio
C. chanos
D. rerio
O. mykiss
D. rerio
56
33
67
50
86
60
63
81
62
62
System
1792
11:33
TJ883-18
systems (Hutchinson 2002). As another example, HMG-COA reductase is an important enzyme in invertebrate juvenile hormone production, compared to being a
mediator in cholesterol formation in mammals (Debernard et al. 1994). The most
extensive comparative dataset with the best homology exists between mammals and
fish (Table 2). Therefore, use of this model to target areas worthy of additional study
initially should be limited to fish.
The purpose of this paper is to present a theoretical model which utilizes the extensive scientific and mechanistic understanding of pharmaceuticals in mammalian
systems to aid in prioritizing which pharmaceuticals may require additional chronic
ecotoxicity testing in teleosts. The premise for the model described in this paper is
that the drugs with the highest potential for additional chronic testing in fish are
those where the anticipated plasma concentration in fish approaches the plasma concentration in humans at which therapeutic effects are observed. It is assumed that
a pharmacological response would occur prior to a toxicological response, due to
the specificity of a compound for its respective target. Much like in human pharmacology, the dose determines the response, whether it be pharmacological or toxicological. Therefore, a receptor/enzyme mediated therapeutic response in non-target
organisms may be viewed as a precursor or surrogate response when compared to
traditional toxicological effects (survival, growth, or reproduction).
The European Agency for the Evaluation of Medicinal Products guidance on environmental risk assessment for medicinal products (EMA 2001) specifically requires
that the assessment utilize toxicology, mechanism of action, and other information
provided in the marketing application to determine if there is any potential for adverse ecotoxicological effects. This model is not intended to change the conditions
that trigger an environmental assessment for pharmaceuticals either in the United
States or Europe. However, with validation and further development, the model has
the potential to be a valuable component of the environmental impact assessment
process.
MODEL DESCRIPTION
The key assumption for the model is that many receptors and enzyme systems
are conserved across mammalian and non-mammalian species, thus making mechanism of action extrapolations possible (Table 2). The model requires comparison
of human and aquatic vertebrate plasma concentrations for a given pharmaceutical. Ideally, the model would be based on the highest human plasma steady state
concentration measured that corresponds to a No Observable Effect Concentration
(NOEC). Such a human plasma concentration at the NOEC (HSS PC NOEC) for
a given pharmaceutical could then be compared to a measured fish steady state
plasma concentration (FSS PC ) derived from an environmental exposure. If these
data were available, the ratio of these two plasma concentrations could be expressed
as a plasma based margin of safety:
Margin of SafetyPB = HSS PC NOEC/FSS PC
Ideally, anthropogenic compounds would have a margin of safety that is as large as
possible. Unfortunately, drug levels in humans and/or mammalian test species are
Hum. Ecol. Risk Assess. Vol. 9, No. 7, 2003
1793
Figure 1.
Conceptual derivation of a safety ratio based on mammalian pharmacological and environmental concentration data.
1794
11:33
Human and Ecological Risk Assessment
TJ883-18
11:33
TJ883-18
traditionally reported at the recommended human dose rather than at the NOEC.
As would be expected, there is a very limited amount of published steady state fish
plasma data for pharmaceuticals linked to an actual aqueous exposure. These two
situations of limited data availability mean that this margin of safety approach for
comparing human and environmental species plasma levels is currently not feasible.
As an alternative to the margin of safety approach, an effect ratio (ER) can be
calculated (Figure 1). The human therapeutic plasma concentration (HT PC) at the
maximum dose of a drug is commonly available and can be utilized instead of a
HSS PC NOEC. For the fish plasma level, a predicted or measured environmental
concentration can be used in conjunction with its respective Log Kow to calculate a
FSS PC. These values can then be used to derive an ER:
ER = HT PC/FSS PC
As in the case with a margin of safety, the ER ratio would ideally be as large as possible
under environmental exposure conditions. An ER 1 indicates that the predicted
drug concentration in the fish plasma is equal to or greater than the drug concentration in human plasma that elicits a therapeutic effect. This plasma concentration
relationship would indicate that there is a potential for receptor mediated responses
in fish, especially if the target enzyme or receptor is expressed in the fish. Therefore,
the lower the ER, the greater the potential need for additional chronic investigation
in fish. Conversely, if the ER > 1 then the predicted fish plasma drug concentration
is less than the human therapeutic plasma drug concentration, indicating a lower
likelihood of receptor mediated responses in fish.
HT PC Derivation
Human therapeutic plasma concentrations (HT PC) are determined as a standard
part of the drug development process and are expressed as Cmax (maximum concentration) or AUC (area under the curve) values. HT PC describes the presence of
a drug in the systemic circulation system at either a single point in time (Cmax) or
as a function of time (AUC) (Hardman et al. 1996). For the ER model, the Cmax
and AUC values can be obtained from the Online Physicians Desk Reference (2002)
or DrugDex Drug Evaluations (2002).
FSS PC Derivation
Fish steady state plasma concentrations (FSS PC) can be predicted utilizing a
Log Kow value and a measured or predicted environmental concentration (EC).
In general, Log Kow values and measured ECs for drugs can be obtained from the
literature or calculated (ACD/LogD software, ACD Labs, Ontario Canada; Huang
and Sedlak 2001; Ternes 1998; Synder et al. 1999; Koplin et al. 2002).
For the purpose of this model, fish steady state blood concentrations were based
solely on calculations using hydrophobicity . Accumulation via the food chain was
not considered (i.e ., trophic transfer). Partitioning between the aqueous phase and
arterial blood in trout has been described with the equation (Fitzsimmons et al.
2001):
Log PBlood:Water = 0.73 Log Kow 0.88.
Hum. Ecol. Risk Assess. Vol. 9, No. 7, 2003
1795
11:33
TJ883-18
D. B. Huggett et al.
A very similar relationship was described by Veith et al. (1979), indicating that fish
bioconcentration was related to the following relationship:
LogBCF = 0.85(logKow) 0.70.
Utilizing these relationships, drug partitioning between blood and water for a
given predicted/measured EC can then be equated to a fish steady state plasma
concentration:
FSS PC = EC (PBlood:Water ).
The major assumption in FSS PC derivation is that the driving force behind a
compound crossing from the aquatic media into the blood stream of a fish is its
hydrophobicity (i.e ., Log Kow ). The reasonableness of this assumption for estimation
purposes has been demonstrated (Veith et al. 1979; Fitzsimmons et al. 2001). The
proposed model further assumes that no metabolism, excretion or protein binding
occurs in the fish (this is a worst case scenario for producing and maintaining a
maximum blood level). Once a compound crosses the gill into the arterial blood,
we assume that it stays in the blood plasma fully dissolved and bioavailable.
ER Derivation
The outcome of the model will be an effect ratio (ER). The ER will be defined as:
ER = HT PC/FSS PC.
An ER 1 indicates that the predicted drug concentration in the fish plasma is equal
to or greater than the drug concentration in human plasma that elicits a therapeutic
effect, while an ER > 1 occurs when the fish drug plasma concentration is lower
than the plasma drug concentration in humans that elicits a therapeutic effect. The
lower the ER, the greater the potential need for additional chronic fish testing to
determine if there are concerns.
Safety Factor Analysis
Safety factors should be used to refine the degree of uncertainty expressed in the
model. Using standard U.S. Environmental Protection Agency (USEPA) guidelines,
a safety factor of 1000 could be used as an initial approach (USEPA 1989). This
factor is derived by applying a 10-fold factor for extrapolation of humans to animals,
a 10-fold factor for sensitivity differences, and 10-fold factor for extrapolating from
mammalian to non-mammalian species. Based on these assumptions, compounds
with an ER < 1000 might warrant additional assessment in fish. The SF of 1000 should
be viewed as an initial assessment, since no fish uptake and receptor activation data
has been developed with human pharmaceutical compounds. As data are developed
in support of this model, the SF can be further refined.
CONCLUSIONS
This model approach has begun to leverage the extensive mammalian pharmacological and toxicological data that are available on pharmaceutical products. The
1796
11:33
TJ883-18
ACKNOWLEDGMENTS
The authors thank Dr. Robert Chapin, Pfizer Inc, for review and constructive
comments made regarding this manuscript.
REFERENCES
Andersen O, Eijsink VGH, and Thomassen M. 2000. Multiple variants of the peroxisome
proliferator-activated receptor (PPAR) y are expressed in the liver of the Atlantic salmon
(Salmo salar). Gene 255:4118
Andreasen EA, Hahn ME, Heideman W, et al. 2002. The zebrafish (Danio rerio) aryl hydrocarbon receptor type 1 is a novel vertebrate receptor. Molecular Pharmacology 62:23449
Bertrand C, Chatonnet A, Takke C, et al. 2001. Zebrafish acetylcholinesterase is encoded by a
single gene localized on linkage group 7. J Biological Chem 276:46474
Chan SJ, Plisetskaya EM, Urbinati E, et al. 1997. Expression of multiple insulin and insulin
like growth factor receptor genes in salmon gill cartilage. Evolution 94:1244651
Chang HM, Wu YM, Chang YC, et al. 1998. Molecular and electrophysiological characterizations of fGlu3, an ionotropic glutamate receptor subunit of a teleost fish. Molecular Brain
Res 57:21120
Dasmahapatra AK, Doucet HL, Bhattacharyya C, et al. 2001. Developmental expression of
alcohol dehydrogenase in zebrafish (Danio rerio). Biochemica Biophysica Research Communications 286:10826
Daughton CG and Ternes TA. 1999. Pharmaceuticals and personal care products in the environment: Agents of subtle change. Environ Health Perspect 107:90738
David-Watine B, Goblet C, Jan D, et al. 1999. Cloning, expression and electrophysiological
characterization of a glycine receptor alpha subunit from zebrafish. Neuroscience 90:303
17
Debernard S, Rossignol F, and Couillaud F. 1994. The HMG-CoA reductase inhibitor fluvastatin inhibits insect juvenile hormone biosynthesis. General and Comparative Endocrinology 95:928
Dickmeis T, Rastegar S, Aanstad R, et al. 2001. Expression of brain subtype creatine kinase in
the zebrafish embryo. Mechanisms of Develop 109:40912
Dunn RJ, Bottai D, and Maler L. 1999. Molecular biology of the Apteronotus NMDA receptor
NR1 subunit. J Experimental Biol 202:131926
Duner T, Conlon JM, Kukkonen JP, et al. 2001. Cloning, structural characterization and functional expression of a zebrafish bradykinin B2-related receptor. Biochemistry J 364:81724
European Agency for the Evaluation of Medicinal Products (EMA). 2001. Draft CPMP discussion paper on environmental risk assessment of non-genetically modified organism
(NON-GMO) containing medicinal products for human use. CPMP/SWP/4447/00
Evans DH. 1993. The Physiology of Fishes. CRC Press, Ann Arbor, MI, USA
FDA (Food and Drug Administration). 1998. Guidance for Industry: Environmental Assessment of Human Drug and Biologics Applications. Center for Drug Evaluation and Research
and Center for Biologics Evaluation and Research www.fda.gov/cder/guidance/index.htm)
Hum. Ecol. Risk Assess. Vol. 9, No. 7, 2003
1797
11:33
TJ883-18
D. B. Huggett et al.
Fitzsimmons PN, Fernadez JD, Hoffman AD, et al. 2001. Branchial eliminationof superhydrophobic organic compounds by rainbow trout (Oncorhynchus mykiss). Aquatic Toxicol
55:2334
Grosser T, Yusuff S, Cheskis E, et al. 2002. Developmental expression of functional cyclooxygenases in zebrafish. Proceed Natural Acad Sci 99:841823
Hsieh SL, Liao WL, and Kuo CM. 2001. Molecular cloning and sequence analysis of stearoylCoA desaturase in milkfish, Chanos chanos. Comparative Biochem Physiology B 130:467
77
Huang CH and Sedlak DL. 2001. Analysis of estrogenic hormones in municipal wastewater
effluent and surface water using enzyme-linked immunosorbent assay and gas hromatography/tandem Mass spectrophotometry. Environ Toxicol Chem 20:1339
Hutchinson TH. 2002. Reproductive and developmental effects of endocrine disrupters in
invertebrates: In vitro and in vivo approaches. Toxicology Letters 131:7581
Koplin DW, Furlong ET, Thurman EM, et al. 2002. Pharmaceuticals, hormones and other
organic wastewater contaminants in U.S. streams, 19992000. A national reconnaissance.
Environ Sci Technol 36:120211
Kumar RS, Ijiri S, and Trant JM. 2001. Molecular biology of channel catfish gonadotropin
receptors: 2. Complementary DNA cloning, functional expression, and seasonal gene expression of the follicle-stimulating hormone receptor. Biol Reproduct 65:71017
Kumar RS, Ijiri S, and Trant JM. 2001b. Molecular biology of channel catfish gonadotropin
receptors: 1. Cloning of a functional luteinizing hormone receptor and preovulatory induction of gene expression. Biol Reproduct 64:10108
Kumar RS, Ijiri S, Kight K, et al. 2000. Cloning and functional expression of a thyrotropin
receptor from the gonads of a vertebrate (bony fish): potential thyroid-independent role
for thyrotropin in reproduction. Molecular Cellular Endocrinology 167:19
Kummerer K. 2001. Pharmaceuticals in the Environment. Springer-Verlag, Berlin, Gemany
Kusakabe M, Kobayashi T, Todo T, et al. 2002. Molecular cloning and expression during spermatogenesis of a cDNA encoding testicular 11beta-hydroxylase (P45011beta) in rainbow
trout (Oncorhynchus mykiss). Molecular Reproduct Develop 62:45669
Lee L, Nong G, Chan YH, et al. 2001. Molecular cloning of a teleost growth hormone receptor
and its functional interaction with human growth hormone. Gene 270:1219
Lin X, Janovick JA, Cardenas R, et al. 2000. Molecular cloning and expression of a type-two
somatostatin receptor in goldfish brain and pituitary. Molecular Cellular Endocrinology
166:7587
Lindberg A and Olivercona G. 2002. Lipoprotein lipase from rainbow trout differs in several
respects from the enzyme in mammals. Gene 292:21323
Madigou T, Mananos-Sanchez E, Hulshof S, et al. 2000. Cloning, tissue distribution and central expression of the gonadotropin-releasing hormone receptor in the rainbow trout
(Oncorhynchus mykiss). Biol Reproduct 63:185766
Maures T, Chan SJ, Xu B, Sun H, et al. 2002. Structural, Biochemical and Expression nalysis of two distinct insulin like growth factor 1 receptors and their ligands in zebrafish.
Endocrinology 143:185871
Menuet A, Pellegrini E, Anglade I, et al. 2002. Molecular characterization of three estrogen
receptor forms in zebrafish:Binding characteristics, transactivation properties and tissue
distributions. Biol Reproduct 66:188192
Nickerson JG, Dugan SG, Drouin G, et al. 2001. A putative 2-adrenoceptor from the rainbow
trout (Oncorhynuchus mykiss). European J Biochemistry 268:646572
NIH (National Institutes of Health). 2002. http://www.nih.gov/science/models/zebrafish
Rodriguez RE, Barrallo A, Garcia-Malvar F, et al. Characterization of ZFOR1, a putative deltaopioid receptor from the teleost zebrafish (Danio rerio). Neuroscience Letters 288:20710
Rxlist. 2002. http://www.rxlist.com
1798
11:33
TJ883-18
1799