Reviews Ofenvironmentalcontaminationand Toxicology

Reviews ofEnvironmentalContaminationand Toxicology

Pharmaceuticalsmay disruptnaturalchemical informationflowsand species interactionsin aquaticsystems:ideas and perspectiveson a hidden global change

--ManuscriptDraft--

Manuscript

Click here todownload Manuscript:VanDonk etal_RVTC15sept.docx

Click here toview linked References

1Pharmaceuticalsmaydisrupt naturalchemicalinformation flowsand

2speciesinteractionsin aquatic systems: ideasand perspectivesonahidden

3globalchange

4

5Ellen Van Donk*, Scott Peacor,KatharinaGrosser,Lisette NDeSenerpont Domis, Miquel

6Lürling

7E. Van Donk

8NetherlandsInstituteofEcology(NIOO-KNAW), Department ofAquatic Ecology,

9Wageningen,Netherlands and Department of EcologyandBiodiversity, Universityof

10Utrecht, Utrecht, Netherlands. e-mail:

11Scott Peacor

12NetherlandsInstituteofEcology(NIOO-KNAW), Department ofAquatic Ecology,

13Wageningen,Netherlands and Department of Ecologyand Department ofFisheriesand

14Wildlife, Michigan State University, EastLansing,Michigan,USA. e-mail:

15Katharina Grosser

16NetherlandsInstituteofEcology(NIOO-KNAW), Department ofAquatic Ecology,

17Wageningen,Netherlands andGerman CentreforIntegrativeBiodiversityResearch(iDiv)

18Halle-Jena-Leipzig,Deutscher Platz5e, 04103Leipzig, Germany.e-mail:

20Lisette NDeSenerpont Domis

21NetherlandsInstituteofEcology(NIOO-KNAW), Department ofAquatic Ecology,

22Wageningen,Netherlands andAquaticEcologyWaterQualityManagement Group, Dept.

23Environmental Sciences,WageningenUniversity,Wageningen,Netherlands. e-mail

25MiquelLürling

26NetherlandsInstituteofEcology(NIOO-KNAW), Department ofAquatic Ecology,

27Wageningen,Netherlands and AquaticEcologyWaterQualityManagement Group, Dept.

28Environmental Sciences,WageningenUniversity,Wageningen,Netherlands. e-mail:

30

31*Correspondingauthor

32

33Contents

341 Introduction

352 Natural information transfer via infochemicals in aquatic ecosystems

363 Current found levels of pharmaceuticals affectorganisms

373.1 Direct and indirecteffects

383.2Infodisruption andmimicry

394 Conclusions and futuredirections

405 Summary

41References

42

43

441 Introduction

45Over thelast decades, anthropogenic activities havedischarged into the environment many

46manmade chemicals.Thereis a risingconcern regardingpharmaceutical products and their

47spread into the environment (e.g.Kümmerer 2008). Dueto the enormous quantities

48consumed, anti-inflammatories, antibiotics, anti-depressives, hormones andblood lipid

49regulators arefound in almostallaquatic environments(Kolpin et al. 2002;Loos et al. 2009).

50Mostpharmaceuticals tend to enter the aquatic environment continuously(but seeSacher et

51al. 2008 forseasonal exception) in contrast to otherpollutants such as herbicides and

52insecticides which areapplied onlyat specifictimes related to thelifecycleof thetarget

53organism, or in responseto observed pest outbreaks(Rosi-Marshalland Royer 2012).

54Pharmaceuticals aredesigned to bebiologicallyactive at verylowconcentrations and end up

55in surfacewaters either unchanged, oras activemetabolites/polar conjugates, mostlyvia

56municipal wastewater and agricultural discharges(Boxallet al. 2012).

57Inreceivingsurfacewaters, organisms livein a seaof natural chemical substances,

58released byotherorganisms, to which manyreact. Combinations of such chemicals, referred

59to as infochemicals, constitutea “smellscape”important in shaping and functioningof aquatic

60ecosystems (seesection2). Pollutantscan disrupt these chemically-mediated information

61flows at several levels inthe chemical signalingpathways/networks (reviewed inLürlingand

62Scheffer 2007; Boyd 2010;Olsén 2011;Lürling2012).However, one classof emerging

63pollutants has receivedfar less attentionthan others: pharmaceuticals.Evenat verylow

64concentrations pharmaceuticals maymimicinfochemicals or interfere with their operation,

65dueto their structuraland functional similaritytothe originalcompounds(Klaschka2008).

66Although direct effects of pharmaceuticals on manyorganisms havebeenwidely

67studied, verylimited research is devoted toahidden aspect of pharmaceuticals:do

68pharmaceuticals affect interactions between species bydisruptinginfochemical pathways?

69This review aims to explore ideasand perspectives for the potential impactof pharmaceuticals

70on the structure andfunctioningof aquatic ecosystems viadisruptingnatural chemical

71information networks between aquaticorganisms.

72

73

742Naturalinformation transfer viainfochemicalsinaquatic ecosystems

75Organisms use chemicalcues in their surroundings, so-called infochemicals, as an important

76sourceof information on their environment (Brönmark and Hansson 2012;Vos et al. 2006).

77Such infochemicals arecompoundsreleased byorganisms and playa critical role across

78different organismal functions and interactions, includingcompetition, predation, navigation

79to and choiceof mates, location of resources,andnavigation to breeding grounds. For

80example, infochemical facilitated predatoreffectson preyphenotypes havebeen shown to be

81widespreadacross diversetaxa, from phytoplankton responding with morphologicalchanges

82to zooplankton herbivores(Van Donk etal.2011),to vertebratepredator-preyinteractions,

83across almost everyimaginable aquaticecosystemfrom streams, ponds, lakes, and marine

84habitats (reviewed inBrönmark and Hansson 2012; Dodson et al. 1994; Tollrian and Harvell

851999)(Fig. 1).

86Lakemesocosm experiments provideevidence ofthe potential profound role of

87infochemicals in food web interactions between fish, zooplankton, and phytoplankton(Boeing

88and Ramcharan 2010). Some ofthe zooplankton (Daphnia) clones responded to fish

89infochemicals bymigratinglower in thewater column to avoid predators.The effects of

90infochemicals cascaded through thefoodweb: without an infochemical induced behavioral

91response, Daphnia were driven to nearextinction with associated algal blooms and reduction

92in fish biomass. On theotherhand, with thebehavioral response,Daphniapopulation

93

94

95

96

97

98

99

100

101

102

dynamicswerestable and fish biomass increased.The abilityto assess predation risk via infochemicalsgoesfar beyond simplysensingasingle predator's infochemicals. Species have been shown to differentiate thescent of different predators and modulatetheirresponse accordingly(Dodson et al.1994). Furthermore, the abilityofapreyspeciesto perceiverisk changes with thediet ofits predator (Dodsonet al.1994), and somespeciescan even balance risk based on the presenceof conspecifics or otherpreyof thepredators(e.g. higher competitor densities represent a weaker predation risk at thesame infochemical level, Van Buskirk et al. 2011).In essence, ecologists arediscoveringthat species navigatea complex chemical smell-scapeof infochemicals togaugepredation risk, avoid competitors and find

foodor mates (Fig. 2).

103

104

105

106

107

108

109

110

111

112

Thenatureof the chemicals that serve as infochemicals and transfer information between organisms isdiverse.It ranges from chemicals that could be considered metabolic products that leak to theenvironment and fortuitouslyconveyinformation,to chemicals created byorganisms to serveparticularpurposessuch as alarm signals(Dodsonet al. 1994). Chemistshaveidentified alargenumber of such substances involved in interactions among organisms in terrestrial and aquaticsystems (e.g.Pohnert et al. 2007). Manyof the infochemicalsin aquatic ecosystems occur at verylow concentrations. Over evolutionarytime the chemosensorysystems have evolved to befinelytuned to detect and react to these compounds.

113

3Current foundlevelsof pharmaceuticalsaffectorganisms

114

3.1 Direct and indirect effects

115

116

117

118

119

120

121

122

Recent studies on the effects of pharmaceuticals show that theycan haveprofound effects on organisms at levels found in natural systems.At themostextreme, pharmaceutical compoundshavebeen found tolead directlyto thedeath ofindividuals. Threespecies of vulturein Asiahavebeenbrought to the edgeof extinction dueto renal failurefollowing consumption of carcassesof diclofenac-treated livestock(Oakset al. 2004).If animals are preyed upon such directeffects maynot onlyaffect theprey, buttravel further through the foodweb via indirect effects on species that interact with the prey(Relyeaand Hoverman

2006).

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

Inaquatic environments, pharmaceutical concentrations areingeneral orders of magnitudelower than theconcentrations exerting direct lethal effects in laboratoryassays (Santoset al. 2010). Nonetheless,agrowingnumber of laboratorystudies report density- mediated directeffects atrealistic concentrations.For instance,tetracyclineconcentrationsas low as 0.5 μg/Lled to lower bacteriaandcyanobacteria biomass in periphyton of artificial streams (Quinlan etal. 2011);amixtureof low concentrations17α-ethynylestradiol(10 ng/L) and fluoxetine(10 ng/L)significantlyreduced population growth rates forPhysa pomilia snails (Luna et al. 2013);a 21 dayexposureofadultmalefathead minnows(Pimephales promelas)to environmentallyrealistic concentrations ofsertraline(5.2 ng/L)and venlafaxine (305ng/L)resulted in mortality(Schultzet al. 2011). Other reports showdirect effects on endpoints that likelycorrelatewithgrowth rate ormortality:clotrimazole isfoundin low concentrations (ng/L)in natural systems and asimilar concentration(17 ng/L)caused inhibition of algal14α-demethylase in labexperiments (OSPAR Commission, 2013); exposureto diclofenac(1 μg/L) causedstructuraldisruptions in the kidney and intestine of

rainbow trout (Mehinto et al. 2010).

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

Importantly, direct effects mayonlybecome apparent afterprolonged exposure, potentiallythrough multigenerationaleffects. Whilestandard acute andchronic assays with the zooplankton grazer Ceriodaphniaindicated that toxic effects ofsertraline at

environmentallyrelevant concentrations wereunlikely,asimple extension of the experimental duration showed that in thethird generation effects on growthand reproduction occurredat a concentration of4.8 μg/L, which isonlyafew times higher thanlevelsthat havebeen encountered in nature(Lamichhaneet al. 2014).Growth in fathead minnows was reduced

after 58 daysexposureto 4 ng/L17α-ethynylestradiolin the F0parent population, but growth reduction occurredalreadyafter 28 days at 0.2 ng/Lin theoffspringF1population (Länge et al. 2011). Exposureof fathead minnowduringthreeyearstoenvironmentalrealistic concentrationsof17α-ethynylestradiol(4.8– 6.1 ng/L), ledtoa completecollapseof the population in thetreatedlake (Kidd et al. 2007). These examples illustrate thatcontinuous exposureto a pharmaceuticalduringmultiplegenerations, maynot onlylead to increased sensitivityover time(Längeet al. 2011;Lamichhane et al. 2014), butalso that longterm

exposuremayhavean impact on the wholepopulation(Kidd et al. 2007).

153

Higher mortalityand reduced populationgrowthof moresensitive species(S1 in Fig.

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

(Fleegeret al. 2003).In fact, thesedensity-mediated indirect effects caused bypollutants might be quite common (Relyeaand Diecks 2008). However, the effect ofonespecies on anothermightnot just travel through densities, but species interactions canalso be influenced bychanges in activities, behaviour orphenotypictraits(Fig. 3B).Low concentrations of pollutants and pharmaceuticalsencountered in nature, rather thanprimarilyleadingtoeffects on mortality, mayalsoshow changes in species’traits (Fig. 3B). Thereis now convincing evidencethat pharmaceuticals maycausefeminization ofmalefishat estrogenconcentrations in thelower ng/Lrange(Gross-Sorokin et al. 2006). Behavioral changes of fish at environmental concentrationsof 1.8μg/Loxazepam(Brodin et al. 2013), 0.12 μg/Lsertraline (Hedgespethet al. 2014)and1.1 μg/Loxazepam(Klaminder etal. 2014)havebeen reported. In theamphipod Gammarus pulex exposureto environmentallyrealistic concentrationsof ibuprofen (10ng/L) or fluoxetine(100ng/L) led to decreasedactivity(DeLange et al. 2006), whereas ibuprofen(1, 10and 100 ng/L), fluoxetine(10 and 100 ng/L)andcarbamazepine (1 and 10 ng/L)caused elevated ventilation (De Langeet al. 2009).As with direct effects on density, direct effects ofpharmaceuticals on aspecies’traits couldalsoindirectlyaffect other species (Fig. 3B, S1S2) through modification of the interaction strengths between the affected species with theotherspecies. This is analogous to the case, forexample, when predators cause(throughinduction) a changein preytraits that leads to indirect effects of the

predator on resources, competitors and otherpredators ofthe prey(Wernerand Peacor2003).

174

3.2 Infodisruptionandmimicry

175

176

177

178

179

180

181

182

183

184

Thefact that pharmaceutical concentrationsoccurat levels in the natural environmenthigh enough toaffect species traits isaforeboder that theselevels may affect infochemical pathways as well(Fig3C-D). Wedividethe potential effects on species interactions into two broad categories.First, (Fig3C) pharmaceuticalsmaydisrupts the transferof information by the infochemicaleither by(a) affectingtheproduction ofthe infochemicalbythe sender(Fig.

3C, arrow i), or(b)affectingthe reception oftheinfochemical bythereceiver (Fig. 3C,arrow ii). Second, pharmaceuticals maymimicinfochemicals (Fig. 3D). Both disruption (Fig. 3C) and mimicry(Fig3D)could affect thereceiverspecies and indirectlyaffect other species by affectingthereceiverspecies traits.

An analysisof publishedliteratureshows astrong increasein the amountof

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

pharmaceuticals. Although this subject is little investigated,wereviewbelow the studies that pointto thepotential role pharmaceuticals could play.

Themajorityof studies ofpharmaceutical effectson infochemical-mediated interactions haveexamined apathwayin which pharmaceuticals affect thereception (i.e. perception and processing) of infochemicals bythe receiver(Fig. 3C-ii). For example,an antidepressant, fluoxetine, has been shown in laboratorystudies to interferewith reception in anumberof fish species by(a) disruptingthe integration ofpheromonecues to control sexual behaviors in malegold fish(Mennigen et al. 2010– 54 μg/L), (b)causingelevated alarm responses in Arabian killifish(Barry2013 – 0.03 to 3 μg/L) and(c) slowing predator avoidanceresponse in larval fathead minnows(Painter et al. 2009 – 25 ng/L). Other effects havebeen found in vastlydifferent pharmaceutical compounds, including (a) propranolol loweringtheresponseofamphipods to predator cuesalbeit at rather high concentrationsof

100 to 5000 μg/L(Wiklundet al. 2011), (b)the painkiller acetylsalicylicacid(1 mM) impairingthe larval metamorphosis ofthe largeseasnailqueen conch (Strombus gigas) that is triggered byred algal (Laurencia poitei) chemical cues (Boettcher and Target 1998), and c)

the veterinarypharmaceutical ivermectin(at 10 mg/L) -abroad-spectrum antiparasitic agent- blockingthechemoreception of allelochemicals and pheromones in nematodes(Rolfe et al.

2001). Thereisalso circumstantial evidencehow pharmaceuticals could interferewith the reception ofinfochemicals (Fig. 3C-ii).Forexample, the antidepressant fluoxetine affected amphipod photo- and geotaxisat 0.1 μg/Land higher levels, whichcouldimpair anti-predator

behavior (Guler andFord2010).

208

209

210

211

212

213

214

215

216

217

218

219

Thereis much less evidenceforthepotential influenceof pharmaceuticalson the production ofinfochemicals bythe sender(Fig. 3C-i). We areonlyaware of studiesusingfish as amodel organism, wherehormonal steroids that arereleased into the watercanact as

potent pheromones (Stacey et al. 2011).Infemalegoldfish duringvitellogenesis (yolk incorporation in theoocytes), the hormone 17β-estradiolstimulates urinaryreleaseof an unidentified pheromonethat attracts males, whiletheoocyte maturation-inducingsteroid

17,20β-dihydroxy-4-pregnen-3-one(17,20β-P) acts as apheromone affectingmalehormone levels and sexual behaviors (Staceyetal.2011).The antidepressant fluoxetinereduces the level of 17β-estradiolin femalegoldfish(Mennigen et al. 2008), whereas theoral contraceptivelevonorgestrel reduces 17,20β-Pin femalefathead minnows (Overturfet al.

2012), and thus effects on pheromone communication arelikely.A simplified workingmodel on how theneuroendocrinedisruption in fish byfluoxetinemight affect pheromone

220

221

222

223

communication via sexsteroidsisvisualized in Figure5 in Mennigen etal. (2011).In addition, gestagens (natural progestogensand synthetic progestins) havebeen identified as class of pharmaceuticals that need to be studied inrelation to potential effects on pheromonal

communication (Orlandoand Ellestad 2014).

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

A number ofstudies suggest that pharmaceuticalsmaymimicinfochemicals, representing anothermechanismbywhich infochemicals disrupt interactions (Fig. 3D).For instance, several androgens and progestins-that weredetected in effluent atconcentrationsup to 14.9 ng/L, oftenexceedingolfactorydetection thresholds forpheromones in fish- are expected to disrupt pheromone communication in fish eitherthrough elicitingresponses at inappropriate times or through competitivebindingto olfactoryreceptors(Kolodziejet al.

2003).In asecond example, the antidepressants venlafaxine and citalopram caused foot detachment in freshwatersnailsat environmentallyrealistic concentrationsas low as 313 pg/L and 405 pg/L,respectively (FongandHoy2012),which in marinesnails isaknown chemical stimuli-mediated escaperesponse to predatorystarfish (Lemmnitzet al. 1989). Examples of fluoxetine-induced spawningof freshwatermussels(Bringolf et al. 2010) -even at low fluoxetine concentrationsof 50 nM (Fong1998) and 20 ng/L(Lazarraet al.2012)might point at mimicryasduringmass spawningevents, mussels usesexpheromones for attractingthe sexual partner and coordinated releaseofgametes byboth partners(Paul etal.2011). Pharmaceuticals mediated spawningunder unfavorable environmental conditions can potentiallyreducemusselreproduction and ultimatelylead to a changein thetrophic

structure.

241

242

243

244

245

246

247

248

249

250

251

These examples underscorethat pharmaceuticals can impact infochemical-mediated interactions and mighttherebyaffectthefitness ofthe organisms involved and potentially food-web structure. There arepresentlyfew studies in this area.Indeed, weareawareof no studies that haveinvestigated potential ensuingindirect effects on other species(i.e. S2 in Figs. 3Cand 3D) and in turn thelarger ecologicalcommunity.To ourknowledge, all published effects of pharmaceuticals on species interactions viainfochemical pathways are examples in which the sender and/or receiverspecies are affected. Presumablysuch direct effects will propagate to indirect effects on other species viathe pathways outlined in Fig. 3C

and 3D. For instance, theselectiveserotonin reuptakeinhibitor sertraline –impaired feedingof perch on the zooplanktongrazerDaphnia in a concentration rangebetween0.12 and 89 μg/L (Hedgespeth et al. 2014).Incontrast,Brodin et al.(2013) found thatlow concentrations (1.8

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

μg/L)of theanxiolyticdrugoxazepamincreased feedingactivityof perch onDaphnia. Likewise, Klaminderet al. (2014)foundincreasedactivityand lower mortalityratesin perch (Perca fluviatilis) exposed to1.1/1.2 μg/Loxazepam.In latter twostudies,perch showed besides increased feeding activityonDaphnia, also less socialityand morebold behaviour (Brodin etal. 2013;Klaminder et al. 2014). Although boldindividuals tend to growfaster, such behavioural changes maylead to higher riskof predation(Hellström and Magnhagen

2011) and thus the outcome ofbehavioural changeswilldepend on the environmental context.

Either way, astronger orreduced predation pressureof perch on thegrazerDaphnia will influencethe information flowbetweentheseorganisms that will travel furtherto phytoplankton (Ringelberg2009) which mayhaveconsequences forother trophicgroups as well (Lürling and Van Donk1997). Webelievefurther research in this areawilluncover such indirect effectsbecausethe perch-Daphnia-algaetritrophicfood chainrepresents wellknown infochemical pathways and as illustrated byRingelberg (2009) theinformation network can

beviewedas superimposed on and tightlyconnected to theflowof matter (see Fig13.2 in Ringelberg2009). Therefore, itis highlylikelythat effects on perchwillnot onlyinfluence the flow ofmatter, but also thesuperimposed information network. Such model systems will makegood candidates to explorethe role of pharmaceuticals, as effects on the information flow (infochemical network) can beseparatedfromeffects onthe energyflow usingexisting, well-developed bioassays.

272

4Conclusions and future directions

273

274

275

276

277

278

279

280

281

282

283

Manyaquaticorganisms useinfochemicalsnot onlyto findpartners and food, but also to sensethe presenceof natural enemiesand to avoid predation.Although thereis agrowing bodyofevidencethat awidevarietyofanthropogenicpollutants commonlyfound in surface waters -at environmentallyrealistic concentrations- can impairchemical communication betweenaquaticorganisms, the impact of pharmaceuticalshas received farless attentionthan otherpollutants. Ourreview indicates that at verylow concentrations pharmaceuticals may mimicinfochemicals or interferewith their operation, due to their structural and functional similarityto theoriginalcompounds. So thesebiologicallyactivepharmaceuticalsmayposea risk ofdisruption ofthe ubiquitousnatural chemical information transfer between organisms. Combined with a plethoraof potential otherstressors influencingtheirmodeof action, this makes pharmaceuticalsatruehiddenglobal change.

284

285

286

287

288

289

290

291

292

293

294

295

296

297

In thefutureseveral major challenges need to beaddressedto further substantiatethe incidenceand scaleat which infodisruption takes place. Virtuallyallstudies refer to laboratoryexperiments with single pollutants, while in theirnatural environmentorganisms arepotentiallyconfronted withmultiple infodisruptors actinginconcertunder varying conditions.Importantly, tounderstand the impactof such infodisruption on natural populations and ecosystems, multi-species and multi-trophic experimentsinmesocosms, combined with multi-compound exposures and model studies areneededto advancethefield. Our review underlines that effects of pharmaceuticals go beyond commonpracticeendpoints, wethereforewould liketo promotethe initiative to extendcurrent ecotoxicological testingof pharmaceuticals(Klaminder et al. 2014;Brodin etal. 2014). Useof standard, well-known model systems, such as thefish-Daphnia (Ringelberg2009) andDaphnia-Scenedesmus systems (LürlingandVan Donk 1997), as wellasbenthicsystems using Gammarussp. (De Lange et al. 2005; 2006;2009) would beparticularlyuseful inthis new generation of

exotoxicologicalexperiments.

298

299

300

301

302

303

304

305

306

Anotherstep forwardwould bethe prolonged exposureof above-mentionedmodel systems as wellas morecomplexcommunities toblends oflow concentrations of pharmaceuticals or evenmixed with other pollutantsor stressors.A recent studyshowed that the chemosensoryperception ofpredators bythegraytreefrogwas reduced by50%when tadpoles werehoused inpolluted stream waterand wastewater effluent compared to clean tap water(Troyerand Turner2015). Thesubstances identified to have an info-disruptingeffect haveoften been hitupon bychance. A systematicscan of selected chemicals and natural- pollutingmixes as discharged from wastewater treatments should provideabroader imageof

the problem.