Environmental Pollution 233 (2018) 356e363

Contents lists available at ScienceDirect

Environmental Pollution
journal homepage: www.elsevier.com/locate/envpol

Evaluation of acute and chronic ecotoxicity of cyclophosphamide,

ifosfamide, their metabolites/transformation products and UV treated
samples*

Chiara Russo a, Margherita Lavorgna a, Marjeta Cesen b, c
, Tina Kosjek b, c,
b, c, ** a, *
Ester Heath , Marina Isidori
a  della Campania L. Vanvitelli, Via Vivaldi 43, I-81100 Caserta, Italy
Dipartimento di Scienze e Tecnologie Ambientali, Biologiche e Farmaceutiche, Universita
b
Jozef Stefan International Postgraduate School, Jamova 39, 1000 Ljubljana, Slovenia
c
Jozef Stefan Institute, Department of Environmental Sciences, Jamova 39, 1000 Ljubljana, Slovenia

a r t i c l e i n f o a b s t r a c t

Article history: Cyclophosphamide (CP) and Ifosfamide (IF) are two nitrogen mustard drugs widely prescribed in cancer
Received 28 July 2017 therapy. They are continuously released via excreta into hospital and urban wastewaters reaching
Received in revised form wastewater treatment plants. Although CP and IF, their metabolites and transformation products (TPs)
12 October 2017
residues have been found in the aquatic environment from few ng L
1 to tens of mg L
1, their environ-
Accepted 15 October 2017
Available online 5 November 2017
mental toxic effects are still not well known. The present study aimed to investigate the acute and
chronic ecotoxicity of CP and IF and their commercially available human metabolites/TPs, i.e. carboxy-CP,
Keto-CP and N-dechloroethyl-CP on different organisms of the aquatic trophic chain. The experiments
Keywords:
Cyclophosphamide
were performed using the green alga Pseudokirchneriella subcapitata, the rotifer Brachionus calyciflorus
Ifosfamide and the crustaceans Thamnocephalus platyurus and Ceriodaphnia dubia. Moreover, to assess the treatment
Acute toxicity conditions in regards to parent compound removal and formation of new TPs, CP and IF were UV-
Chronic toxicity irradiated for 6 h, 12 h, 24 h, 36 h and 48 h, followed by toxicity evaluation of treated samples by algae,
Metabolites/TPs rotifers and crustaceans. Between the parent compounds, IF resulted as more toxic drug under tested
UV-treatment conditions, exerting both acute and chronic effects especially on C. dubia (LC50:196.4 mg L
1,
EC50:15.84 mg L
1). Among the tested metabolites/TPs, only carboxy-CP inhibited the reproduction in
the rotifer. However, LOEC and NOEC values were calculated for CP and IF for all organisms. In addition,
despite a low degradation of CP (28%) and IF (36%) after 48 h UV-irradiation, statistically significant effect
differences (p < 0.05) from not-irradiated and irradiated samples were observed in both acute and
chronic assays, starting from 6 h UV-irradiation. Our results suggest that the toxic effects found in the
aquatic organisms may be attributable to interactions between the parent compounds and their me-
tabolites/TPs.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction crosslinking and strand breaking of DNA and inhibition of cell di-
vision with consequent cell death (Ralhan and Kaur, 2007). Ac-
Alkylating agents have been used in cancer therapy since the cording to ATC classification, anticancer drugs are divided into the
1940s. They are known to directly interact with DNA, what causes following groups: L01AA Nitrogen mustard analogues, L01AB Alkyl
sulfonates, L01AC Ethylene imines, L01AD Nitrosoureas, L01AG
Epoxides and L01AX Other alkylating agents. In this broad spec-
*
This paper has been recommended for acceptance by Klaus Kummerer. trum, nitrogen mustards, represent the most widely prescribed
* Corresponding author. Dipartimento di Scienze e Tecnologie Ambientali, Bio- chemotherapeutics used for the treatment of lymphoma, leukemia
 della Campania L. Vanvitelli, Via Vivaldi 43, I-
logiche e Farmaceutiche, Universita
and several solid tumors, where they nonspecifically alkylate the N7
81100 Caserta, Italy.
** Corresponding author. Jozef Stefan International Postgraduate School, Jamova
of the guanine bases by the aziridinium group in tumor tissue (Xie,
39, 1000 Ljubljana, Slovenia. 2012). However, cytotoxic effects were observed towards healthy
E-mail addresses: [email protected] (E. Heath), [email protected] cells with adverse effects such as carcinogenicity, teratogenicity,
(M. Isidori).

https://doi.org/10.1016/j.envpol.2017.10.066
0269-7491/© 2017 Elsevier Ltd. All rights reserved.
 C. Russo et al. / Environmental Pollution 233 (2018) 356e363 357

fetotoxicity both in patients and in health care workers in reference structural analogue 3- dechloroethyl-IF. This paper follows the
to the first applied nitrogen mustards Cyclophoshamide (CP, 
research conducted by Cesen and collaborators (2016b) where the
L01AA01) and Ifosfamide (IF, L01AA06) (Lekskulchai, 2016). They toxic effects of these compounds have been evaluated on pro-
are polar compounds, highly soluble in water, with low octanol- ducers, the alga Pseudokirchneriella subcapitata and the cyanobac-
water partition coefficient (log Kow of 0.6 and 0.9 for CP and IF, teria Synecococcus leopoliensis. Since the evaluation of the possible
respectively, Table S1) and low adsorption potential to sewage ecotoxicological effects on organisms from different trophic levels
sludge or sediments, which suggest they can be mainly found in the is necessary for the environmental risk assessment of such com-
water phase (Kosjek and Heath, 2011; Xie, 2012). pounds, in the present study, the acute and the chronic toxicity of
Based on pharmacokinetics, CP and IF are pro-drugs, which are CP, IF and the metabolites/TPs was tested in primary consumers.
metabolised and excreted together with their metabolites from The following organisms were utilized: the environmentally
human body via urine and/or faeces. About 62% of the initial CP abundant rotifer Brachionus calyciflorus, the anostracan crustacean
dosage is excreted via urine unaltered together with its metabolites Thamnocephalus platyurus, which was used because of its high
when administration lasts over 48 h (Eitel et al., 2000). The total sensitivity in acute toxicity testing and the cladoceran crustacean
amount of IF and its metabolites excreted over a 24 h period via Ceriodaphnia dubia, which is distributed in freshwaters worldwide
urine is ~ 50% of the injected IF dose. The excretion % of e.g. and can be employed in both, acute and chronic tests. Moreover,
carboxy-CP, keto-CP and N-dechloroethyl-CP relative to the dose growth inhibition of CP, IF and their metabolites/TPs was also
applied is 11e23%, 0.4e0.6% and 2e3% (as CP metabolite) and 11% evaluated in the green alga Pseudokirchneriella subcapitata to obtain
(as IF metabolite), respectively (Gilard et al., 1993; Joqueviel et al., their unknown LOEC and NOEC. Finally, to evaluate the treatment
1998; Zhang et al., 2006). These studies report that the excretion efficiency of CP and IF, the effects of UV irradiated CP and IF was
depend on age, gender, physical condition of patient, the admin- evaluated by alga, rotifer and cladoceran crustacean toxicity testing
istered dosage, the duration of the intravenous administration and to assess the risk associated with unknown TP formation.
the accompanying therapy with other pharmaceuticals.
As these residues mainly end up in sewage system, their main 2. Materials and methods
contamination source is represented by wastewater treatment
plant effluents (Kosjek and Heath, 2011; Isidori et al., 2016). 2.1. Standards, reagents and chemicals
The low removal efficiencies for both drugs in wastewater
treatment plants have raised concern for environmental health and CP (99%, CAS 50-18e0 P) and IF (99%, CAS 3778-73-2) were
safety implications despite existence of various removal techniques purchased from Sigma Aldrich (Hong Kong, China). Deuterated
reported in literature such as biotransformation, ozonation, air cyclophosphamide (d6-CP), used as internal standard for chemical
stripping, sorption and phototransformation. In fact, Buerge et al. analysis, carboxy-cyclophosphamide (CAS 22788-18-7), keto-CP
(2006) found that no degradation was observed in activated (CAS 27046-19-1) and N-decl-CP (a structural analogue of CP and
sludge incubation experiments within 24 h at concentrations of IF metabolites/TPs; CAS 36761-83-8) were obtained from Niomech
~100 ng L
1. Lutterbeck et al. (2016) also found that no degradation - IIT GmbH (Bielefeld, Germany). Derivatizing agent trifluoroacetic
was observed for CP using UV and simulated sunlight treatments. anhydride (TFAA, 99%, CAS 407-25-0) was purchased from Fluka

Contrarily, Cesen et al. (2015) showed higher removal efficiencies (Buchs, Switzerland). Acetonitrile and ethyl acetate were purchased
for these two compounds (59% and 35% for CP and IF, respectively) from J. T. Baker (Deventer, Netherlands) and were of analytical
when Moving Bed Biofilm Reactor (MBBR) technique was applied grade purity. To prevent scavenging during UV irradiation experi-
(attached growth biomass) and even higher removals (99% for CP ments, stock solutions of CP and IF were prepared in MilliQ water.
and 94% for IF) when UV treatment and ozonation with addition of Calibration standards were prepared by dilution of stock solutions
H2O2 were sequentially included after biological treatment. In 2015, in MilliQ water.
Lin and colleagues investigated the removal efficiency of ozonation
for different anticancer drugs including CP and IF and reported the 2.2. UV irradiation experiment
resistance of process by-products to further ozonation. This resis-
tance could justify the increase in Microtox acute toxicity results Duplicate UV experiments were performed in a cylindrical glass
after only 30 min of ozonation despite the rapid degradation of CP reactor containing 760 mL of aqueous solution with initial con-
and IF which was caused by the attack of hydroxyl radicals. centration of CP and IF at 200 mg L
1 in MilliQ water. The drugs
Accordingly, CP and IF residues including residues of parent com- selected were tested at high concentrations, certainly far from
pounds, metabolites and TPs have been found in wastewaters at environmental concern, considering CP and IF occurrence in surface
concentrations from pg L
1 to tens of mg L
1 (Buerge et al., 2006; waters, but necessary to obtain ecotoxicological data useful for

Cesen et al., 2015; Isidori et al., 2016). further risk assessment. The high concentrations were also selected
Nonetheless, toxic effects of CP and IF metabolites/TPs are not 
based on our previous study (Cesen et al., 2016b) and the study of

well understood. Cesen et al. (2016a) identified four CP and four IF Sanderson et al. (2003) addressing the toxicity of CP and IF. Tested
TPs formed upon UV treatment and five CP and four IF TPs during solutions were continuously stirred by magnetic stirrer (200 rpm)
UV/H2O2 treatment. In addition, the authors measured CP and IF throughout each experiment. Monochromatic low pressure (LP)
selected metabolites/TPs, i.e. as carboxy-CP, keto-CP and N-dech- mercury lamp was introduced into the reactor and separated from
loroethyl-CP (or structural analogue 3-dechloroethyl-IF) in hospital testing solution by quartz cooling well, which maintained the so-
wastewaters at concentrations from ng to tens of mg L
1. However, lution temperature at approximately 20  C. The LP lamp (6 W) as
to the author's knowledge, there is no evident toxic effects of both irradiation source emits 90% of its radiation at 254 nm. The UV
parent compounds and metabolites/TPs reported in concentrations intensity, which was measured by ferrioxalate actinometry, was
below mg L
1 (Lin et al., 2015). Data on the long-term toxic effects 0.78  10
6 Einstein/sec (Nicole et al., 1990). Fresh MilliQ water was
of CP, IF and their metabolites/TPs in non-target organisms are also used as medium because it has only negligible amounts of organic
lacking. matter (<5 ppb) so that the formation of various reactive oxygen
The aim of this study was to evaluate the ecotoxicity of the species (ROS) induced by UV irradiation could be excluded. The
parent drugs CP and IF and their commercially available metabo- MilliQ water with resistivity of 18.2 MU cm at 25  C which ensures
lites/TPs: carboxy-CP, keto-CP and N-dechloroethyl-CP and its pH of 7.0 was applied. The pH of the media was not measured
358 C. Russo et al. / Environmental Pollution 233 (2018) 356e363

within the experiments in order to avoid microbial contamination. (US EPA, 1993) using less than 24 h old organisms. Neonates of at
In addition, to avoid any significant changes in the pH of the media least third generation came from the mass culture (Aquatic
during UV experiments, experiments were performed in a closed Research Organisms, Inc., Hampton, NH, USA) and were maintained
atmosphere to prevent the drop of pH, which could derive from at 25 ± 1  C in synthetic ISO medium (hardness 250 mg L
1
dissolved ambient CO2. The photodegradation efficiency of each expressed as CaCO3) with a 16:8 h light:dark cycle (600 lux). Tests
analyte was determined by the ratio between concentrations were performed in 24-well plates with 10 crustaceans/well and
determined in treated samples and in non-treated controls. One with either 1.0 mL of the four to seven dilutions of single com-
hundred microliters and 20 mL of samples were collected after 6 h, pounds (starting from the highest concentrations of 2000 mg L
1
12 h, 24 h, 36 h and 48 h of UV irradiation for chemical analysis and and 100 mg L
1 for parent compounds and metabolites respec-
bioassays, respectively. tively, geometric progression 2) or with UV-irradiated samples at
different exposure times. The tests were performed in three
2.3. Sample preparation for chemical analysis replicates.
For all tests, a test-medium control series (negative control) was
As the initial concentration within the UV experiments was used to the test series. Subsequently, all plates were incubated in
high, i.e. 200 mg L
1, preconcentration of samples was not needed. darkness at 25  C for 24 h for each test.
Instead, 100 mL of samples including controls and UV-irradiated Mortality was the end-point considered in the acute tests and
samples were dried under nitrogen at 40  C. The dried samples the concentration resulting in a 50% effect after 24 h-exposure was
were derivatized for 1 h at 60  C with 100 mL of TFAA and 100 mL of indicated as Median Lethal Concentration (LC50).
ethyl acetate. The samples were then dried again under nitrogen to
remove any trifluoroacetic acid formed during derivatization and
re-dissolved in 250 mL of ethyl acetate prior to chemical analysis.
2.5.2. Chronic toxicity tests
2.4. Chemical analysis B. calyciflorus chronic test was based on the offspring reduction
over 48 h exposure (ISO, 20666, 2008) and was performed on less
The samples were analysed using a HP 6890 series (Hewlett- than 2 h old organisms. Cysts were hatched as described for the
Packard, Waldbron, Germany) gas chromatograph with a single acute test. Tests were performed in 48-well plates with one rotifer/
quadrupole mass selective detector. The capillary column was a DB- well and with either 0.9 mL/well of single compounds (in six di-
5MS 30 m  0.25 mm  0.25 mm (Agilent J&W, CA, US). The carrier lutions in moderately hard dilution water, starting from the highest
gas was He (flow rate of 1 mL/min). One microlitre of sample concentrations of 200 mg L
1 and 100 mg L
1 for parent com-
extract was injected in splitless mode with the injector tempera- pounds and metabolites, respectively, with a geometric progression
ture set at 270  C. The temperature program was as follows: an of 2) or with different UV-irradiated samples. All tests were per-
initial temperature of 65  C was ramped at 30  C/min to 300  C, formed in six replicates. The organisms were fed with 0.1 mL of
where it was held for 2 min. Total runtime was 9.83 min. The MS fresh suspension of 107 cells/mL of the unicellular alga
was operated in electron impact (EI) mode with an ionization P. subcapitata. Plates were incubated in darkness at 25  C for 48 h.
voltage of 70 eV. Selected ion monitoring (SIM) mode was used for The test on C. dubia 24 h old neonates was performed over 7
qualitative and quantitative analysis by monitoring the following days (ISO, 20665, 2008). Organisms from at least the third gener-
ions: m/z 309, 307 and 150 for CP, m/z 307, 181 and 150 for IF and ation were individually exposed to 25 mL of test solution in bea-
m/z 313 and 315 for CP-d6. Instrumental control and data pro- kers. All test media were exchanged five times per week in semi-
cessing were performed using ChemStation software. static conditions. From the fourth day-exposure onward, the
offspring produced by each parent animal were counted and
2.5. Toxicity tests removed daily. Crustaceans were fed with a combination of yeast
Saccharomyces cerevisiae, alfalfa and flake food (5 g L
1 of each) in
2.5.1. Acute toxicity tests addition to the unicellular green alga P. subcapitata (108 cells/mL).
B. calyciflorus acute assay was performed following the ASTM E Single compounds (in five to seven dilutions starting from the
1440e91 guidelines. Organisms less than 2 h old were hatched highest concentrations of 200 mg L
1 and 10 mg L
1 for parent
from cysts (MicroBioTest Inc., Nazareth, Belgium) in synthetic compounds and metabolites/TPs, respectively, with a geometric
moderately hard freshwater (80e100 mg L
1 CaCO3, pH 7.5 ± 0.3) at progression of 2) as well as UV-irradiated samples were tested in
25 ± 1  C under continuous illumination (3000e4000 lux) for ten replicates and incubated at 25 ± 1  C, with a 16:8 h light:dark
16e18 h prior to test initiation. 0.3 mL of test solution was used in cycle (600 lux).
each test well (36-well plates, MicroBioTest Inc., Nazareth, P. subcapitata growth inhibition test (Paixao et al., 2008) was
Belgium). Single compounds (in four to five dilutions starting from performed in 96-well microplates. The single compounds (six-
the highest concentrations of 2000 mg L
1 and 100 mg L
1 for seven dilutions starting from the highest concentrations of
parent compounds and metabolites, respectively, with a geometric 200 mg L
1 and 10 mg L
1 for parent compounds and metabolites/
progression of 2) as well as UV-irradiated samples at different TPs, respectively, with a geometric progression 2) and the UV-
exposure times (6 h, 12 h, 24 h, 36 h and 48 h) were tested in six irradiated samples were incubated with 104 cells/mL of algal sus-
replicates with five animals/well. pension in six replicates under continuous illumination (6000 lux)
T. platyurus test (ISO 14380, 2011) was conducted on larvae at 25 ± 1  C on a microplate shaker (450 rpm). The plates were read
hatched from cysts 20e22 h before the assay in the same synthetic at 450 nm (SpectraFluor, Tecan, Switzerland) immediately before
freshwater as the rotifers and under the same experimental con- the test and after 24 h, 48 h and 72 h.
ditions. Tests were performed in 24-well plates using 10 crusta- For all chronic tests, a test-medium control series (negative
ceans/well, 1.0 mL in five dilutions starting from the highest control) was used to the test series. The number of the offspring or
concentrations of 2000 mg L
1 and 100 mg L
1 for parent com- the algal growth outputs were compared to the values obtained for
pounds and metabolites respectively (geometric progression 2) and the negative control in order to determine the chronic effective
in three replicates. percentages and to evaluate the chronic Effective Concentrations
C. dubia test was performed following EPA-600-4-90 procedures (ECx).
 C. Russo et al. / Environmental Pollution 233 (2018) 356e363 359

2.5.3. Data analysis

For each test, two (for UV-irradiated samples) or three (for cðtÞ
ln ¼
kt  t (1)
parent compounds and metabolites/TPs) independent experiments cð0Þ
were performed.
C. dubia independent assays were analysed using ToxRat Pro- where kt is the first-order rate constant including both direct and
fessional software, Ver 2.10.05 (Alsdorf, Germany) for the calcula- indirect oxidation of the investigated compounds (Rosenfeldt and
tion of effective percentages. Linden, 2007). The R2 values of degradation curves were 0.983
For all organisms the effective percentages obtained from in- and 0.963 for CP and IF, respectively. The kt values for CP and IF
dependent experiments were pooled using Prism5 software were 7.00 ± 1.43  10
3 h
1 (or 1.97  10
6 sec
1) and
(Graphpad Inc., CA, USA) to estimate the concentrations giving x% 1.04 ± 0.202  10
2 h
1 (or 2.9  10
6 sec
1), respectively. In
effect (L(E)Cx) by non-linear regression (log agonist vs. normalized addition, poor removal was expected since chemical structures of
response-variable slope). The LC50 value, corresponding to 50% CP and IF do not contain double bonds that could absorb photons

under UV irradiation (Cesen et al., 2015). The kt value for CP
mortality, was the test parameter for organisms in the acute tests,
whereas EC50, EC20 and EC10 were the concentrations giving 50%, determined within this study is of similar order of magnitude as
20% or 10% of the effect used in chronic tests to evaluate the inhi- previously reported by Kim and Tanaka (2009) in ultrapure water
bition reproduction effect or the inhibition of the algal growth. (kt for CP ¼ 8.3  10
5 sec
1). Despite the similarity in obtained kt
For long-term toxicity, the Lowest Observed Effect Concentra- values, the potential formation of radical species during photolysis
tion (LOEC) and the No Observed Effect Concentration (NOEC) were most probably contributed towards the observed CP and IF removal.
estimated by ANOVA and Dunnett's multiple comparison test. Another very important parameter for photodegradation to be
Thus, for the inhibition of the reproduction tests, the compari- considered is the quantum yield of CP and IF in the UV 254 nm, as is
son was evaluated between the mean number of live offspring molar extinction coefficient (ε). A recent study by Zhang et al.
produced per parent organism, which was exposed to chemicals (2017) reports quantum yields for CP and 5-fluorouracil (both cy-
and the offspring mean of the controls. Regarding the algal growth tostatics), which are 54.83  103 mol E
1 and 4.06  103 mol E
1,
inhibition tests, the comparison was between the mean number of respectively. The authors correlate low removal during sole UV
the absorbance values related to different compounds (directly irradiation using LP lamp (5 W, UV irradiation at 254 nm) to low ε
proportional to the number of algal cells) and the control absor- for CP (7 M
1 cm
1), if compared to 5-fluorouracil
bance mean values (after 72 h). (3506 M
1 cm
1), which was degraded approximately 35-times
As regards to UV-irradiated samples, the statistically significant faster. They relate higher removal by high absorbance of 5-
differences between the non-irradiated and irradiated samples (at fluorouracil at UV-254 nm and not to lower quantum yield. Based
different UV exposure times) as well as the significant difference on our results, we assume that the observed removal of CP and
between different UV exposure times and between different or- structurally very similar IF might be attributed to low ε and to
ganism species were evaluated by 2-WAY ANOVA, Bonferroni potential formation of radical species formed upon treatment.
posttests. Furthermore, also the mineralization during photolysis is negli-
gible, what was confirmed in our previous study (Cesen  et al.,
2016b). In there, DOC was monitored in UV irradiated samples
containing CP and IF (at 10 mg L
1) and revealed no significant
3. Results and discussion
decrease in DOC over 48 h of UV irradiation. This suggested that UV
degradation failed to mineralize both compounds.
3.1. Results of chemical analysis

Chemical analysis of UV-treated samples revealed poor degra- 3.2. Results of toxicological tests
dation of both investigated compounds, what was expected ac-
cording to their UV-Vis spectra in the range of 250e370 nm (Lin 3.2.1. Acute and chronic toxicity of parent compounds and their
et al., 2014). After 48 h, CP and IF were degraded only up to 28.3% metabolites/TPs
and 36.5%, respectively, (Table S2, Fig. 1). Degradation of both Table 1 shows acute toxicity of CP and IF expressed as LC50. For
compounds followed pseudo first order kinetics using the following both drugs, the median mortality of all invertebrates was found at
Equation (1): concentrations orders of magnitude higher (196.4e1924 mg L
1)

Fig. 1. Degradation of CP and IF during UV-irradiation of aqueous samples (n ¼ 2).

360 C. Russo et al. / Environmental Pollution 233 (2018) 356e363

Table 1 concentrations are too far from those of environmental concern

LC50 values in mg L
1 for acute toxicity tests with 95% confidence range. (mg-ng L
1) and are dependent on the dilution factor chosen in the
Compounds C. dubia 24 h B. calyciflorus 24 h T. platyurus 24 h experimental plan, from a toxicological point of view the study of
CP 986.6 1924 1396
the lowest effective and non-effective levels is a very useful tool for
(765.3e1272) (1210e3060) (1304e1494) further studies regarding the environmental risk assessment and
IF 196.4 996.3 771.5 the calculation of the Risk Quotient (R.Q.) of the compounds here
(149.1e258.7) (767.3e1294) (627.9e947.8) studied. Furthermore, as observed only in case of CP (Fig. 2), both,
For metabolites/TPs no effect was observed up to concentration of 100 mg L
1. P. subcapitata and C. dubia, increased their growth/reproduction
rate at the lowest tested concentrations confirming adaptation
potential known as hormesis. This useful biological response to
than concentrations found in the aquatic environment (Cesen  et al., toxins and stressors generally allows organisms to react to toxic
2015; Isidori et al., 2016). Between the two parent drugs, IF was compounds. The hormetic doseeresponse curves, observed from
more toxic for all non-target organisms, whose susceptibility was the cellular level to broad categories of organisms, compensate any
found in the following order: B. calyciflorus < T. platyurus < C. dubia imbalance in homeostasis caused by mild stresses. For example,
highlighting that IF acute toxic effects increased not only as the reactive oxygen species are one of the major reasons for most of the
aquatic trophic chain level increased, but also with the increased diseases, but exposure to low quantities reverses its negative effects
stages of phylogenetic tree. This is in accordance with a study by (Ristow and Schmeisser, 2014). Interestingly, in 2015 Gaya et al.
Sanderson et al. (2003), where effects of the same antineoplastic observed that the administration of CP at high doses exerted anti-
drugs were evaluated towards algae, daphnids and fish. The highest tumoral effect with bone marrow suppression as an adverse ef-
toxic effects were reported for fish Pimephales promelas, i.e. the fect while, at low doses, it contributed to anti-tumor immunity.
organism with the highest evolutionary stage among the tested Despite similarity in chemical structure, only carboxy-CP
organisms. The absence of acute toxicity for CP is also confirmed by showed chronic effects among the tested metabolites/TPs. In fact,
Lutterbeck et al. (2014), who did not find any luminescence inhi- carboxy-CP caused inhibition of the reproduction in rotifers in a
bition after 30 min in Vibrio fischeri. In the present study, CP and IF concentration dependent manner (Fig. 3) with a median offspring
metabolites/TPs did not reveal acute toxic effects up to 100 mg L
1, reduction equal to 23.66 mg L
1 (Table 2). The toxicity of carboxy-
which was the highest tested concentration. CP occurring in the range of mg L
1 has already been reported by
In chronic toxicity tests (Table 2, Fig. 2), IF was confirmed to be 
Cesen et al. (2016b), who observed a median toxic effect at
the most toxic compound causing a median offspring reduction in 17.1 mg L
1 with the cyanobacterium Synecococcus leopoliensis. In
C. dubia at 15.84 mg L
1 after 7d-exposure and in B. calyciflorus at line with the EU Directive 93/67/EEC (1996), which classifies sub-
76.05 mg L
1 after 48 h-exposure. CP EC50 was 58.03 mg L
1 in C. stances according to their EC50 values towards aquatic organisms,
dubia with an EC10 of 19.50 mg L
1. This result is partially in those authors defined carboxy-CP as harmful to aquatic organisms
agreement with the EC10 found by Lutterbeck et al. (2014). An having an EC50 in the range from 10 to 100 mg L
1. Since both
increasing growth inhibition effect was observed for algae at con- parent compounds showed an EC50 in the range from 10 to
centrations up to the highest concentration tested of 200 mg L
1 for 100 mg L
1, when tested in C. dubia and B. calyciflorus (Table 2), also
both parent compounds. These findings are in agreement with a CP and IF can be considered harmful to aquatic organisms. Besides,
study by Zounkova et al. (2007), who found a median growth in- these alkylating agents are known as mechlorethamine com-
hibition in P. subcapitata at much higher concentrations pounds, having a bischloroethyl group that reacts, through an
(930 mg L
1) for CP. An evident increase of the effect caused by IF aziridinium intermediate, with electron-rich atoms in macromol-
was observed both in acute and chronic toxicity moving up the ecules to form highly reacting covalent bonds that cause toxic ef-
trophic chain. This suggests an increase of its toxic effect with the fects (Siddick, 2002).
increased complexity of organisms. According to Zhu et al. (2015),
different metabolic pathways could be linked to alkylating agents
metabolism such as GSH metabolism, fatty acid metabolism, 3.2.2. Acute and chronic toxicity of CP and IF after UV-irradiation
biosynthesis of unsaturated fatty acids, fatty acid biosynthesis and The risk associated with unknown TP formation after UV-
steroid hormone biosynthesis pathways that could be involved in irradiation was assessed by the effect-driven approach. In this
bioaccumulation and bioconcentration processes also of human approach, the irradiated mixture of CP and IF underwent toxicity
concern. testing to investigate their ecotoxicity at different UV exposure
Although for P. subcapitata the median toxic effect was not times.
experimentally reached, both, LOEC and NOEC values, were deter- Acute toxicity results of UV-irradiated samples containing CP
mined in the mg L
1 range (Table S3). Even though these observed and IF are shown in Fig. 4, where the mortality percentage obtained

Table 2
EC50, EC20, EC10 values (mg L
1) for chronic toxicity tests with 95% confidence range.

Compounds C. dubia 7d B. calyciflorus 48 h

EC50 EC20 EC10 EC50 EC20 EC10

CP 58.03 28.85 19.50 89.84 14.19 4.82

(37.43e89.98) (15.60e48.08) (9.57e38.64) (67.62e119.4) (7.87e22.18) (0e10.19)
IF 15.84 5.58 3.03 76.05 14.93 5.75
(13.03e19.27) (3.78e7.79) (0e5.00) (48.04e120.4) (6.59e27.10) (0e14.93)
Carboxy-CP 23.66 3.99 1.41
(15.45e36.23) (0e7.57) (0e3.93)
N.D. up to 10
Keto-CP
N.E. up to 100
N-dechloroethyl-CP

N.D.: Not determinable.

N.E.: No effect.
 C. Russo et al. / Environmental Pollution 233 (2018) 356e363 361

Fig. 2. Concentration/effect curves (non-linear regression: log agonist vs. normalized response-variable slope) of CP and IF in C. dubia, B. calyciflorus and P. subcapitata. Bars indicate
standard errors.

(p < 0.001) and 74.8% (p < 0.001), respectively, after only 6 h UV-
treatment, reaching 95%, 96.7% and 98% of mortality after 48 h of
UV-irradiation. The significant mortality increase following the
exposure time is shown in Fig. 5. A clear time exposure-mortality
trend was observed in the rotifer exposed to irradiated CP,
differing from the concentration effect found when exposing the
crustacean to irradiated samples containing CP with no significant
increasing mortality. IF irradiated samples showed the highest
acute toxicity to both exposed organisms. Probably, the acute toxic
potential could be due to a mixture of different TPs formed during
treatment such as the human metabolites and TPs IF-TP199 (N-
dechloroethyl-ifosfamide), IF-TP275 (keto-ifosfamide), IF-TP138
and IF-TP259 (imino-ifosfamide) already tested by Cesen  et al.
(2016a) on Synecococcus leopoliensis.
Chronic toxicity results are shown in Fig. 6, where the offspring
Fig. 3. Concentration/effect curves (non-linear regression: log agonist vs. normalized
reduction and the growth inhibition percentages observed for
response-variable slope) of Carboxy-CP, Keto-CP and N-dechloroethyl-CP in
B. calyciflorus. Bars indicate standard errors. B. calyciflorus/C. dubia and for P. subcapitata, respectively, are
correlated to the CP and IF UV-exposure time. The least toxic effect
was observed in B. calyciflorus exposed to UV-irradiated sample
in B. calyciflorus and in C. dubia after 24 h exposure is given. In containing CP, with initial significant inhibition of the reproduction
addition, at each study time (6 h, 12 h, 24 h, 36 h and 48 h), a (81.05%) obtained after 24 h of UV-irradiation. The highest initial
statistically significant difference among different samples was toxicity level (p < 0.01) was found in B. calyciflorus exposed to 6 h-
calculated. CP UV-irradiated sample showed the lowest toxicity UV irradiated sample containing IF (toxic effect of 86.35%). Chronic
after irradiation in C. dubia reaching a significant effect (p < 0.05, effects in C. dubia of UV-treated samples containing IF are shown in
compared to non-irradiated samples) equal to 23.3% of mortality Fig. 7. The highest toxic result (77.95% offspring reduction) was
after 48 h of UV-irradiation. CP with B. calyciflorus as well as IF with obtained with 1/10 dilution of IF irradiated by UV for 48 h, while an
B. calyciflorus and with C. dubia caused significant mortality rates initial significant toxic effect (p < 0.05) was found when testing 1/
(compared to non-irradiated samples) of 26.7% (p < 0.01), 41.7% 1000 dilution after 6 h of UV irradiation. Figure S1 shows the

Fig. 4. Mortality percentage obtained testing UV-irradiated samples in B. calyciflorus and C. dubia. Significant differences were noted on comparing at each time non-irradiated and
irradiated samples for *p < 0.05; **p < 0.01; ***p < 0.001(2WAY ANOVA, Bonferroni posttests). Different letters were used for each significantly different sample (p < 0.05).
362 C. Russo et al. / Environmental Pollution 233 (2018) 356e363

Fig. 5. Mortality percentage/UV-irradiation time trend obtained in B. calyciflorus and C. dubia. Different letters for p < 0.05(2WAY ANOVA, Bonferroni posttests). Different numbers
were used to identify different samples.

Fig. 6. Offspring reduction and growth inhibition percentages observed respectively for B. calyciflorus/C. dubia and for P. subcapitata exposed to different times of samples UV-
irradiated. Significant differences were noted on comparing at each time non-irradiated and irradiated sample for *p < 0.05; **p < 0.01; ***p < 0.001(2WAY ANOVA, Bonferroni
posttests).

growth inhibition) is related to 36 h of UV exposure, while for CP

(85.16%) it is related to 24 h-exposure. The highest CP toxic effect in
rotifers and in crustaceans was related to 36 h of UV exposure
(96.30 and 88.85% inhibition of the reproduction, respectively),
whereas the first most significant toxic effect of IF was 94.55% in the
rotifer for a sample that was UV-irradiated for 24 h (Figure S1b).
The acute toxicity results were mostly related to the highest
toxic level in IF UV-irradiated samples, while the chronic outcomes
revealed also the ability of CP UV-irradiated samples to restrain
reproduction and exert growth inhibition. CP and IF were degraded
only up to 28% and 36%, respectively, but the findings of this study
suggest that a mixture of parent compounds and metabolites/TPs
could represent a harmful environmental combination exhibiting
synergistic and/or potentiation effects. Since the UV irradiated
mixtures showed an increase in toxicity at the same time as the
decrease in parent compound concentrations, further studies
where TPs will be separated, identified and quantified before
Fig. 7. Offspring reduction percentage testing IF UV-irradiated samples in C. dubia.
toxicity testing, as recommended by exposure-driven approach,
Different letters for p < 0.05(2WAY ANOVA, Bonferroni posttests). Different numbers
were used to identify different dilutions.
what is needed to confirm these assumptions.

chronic effect percentage/UV-irradiation time trend obtained in 4. Conclusions

both P. subcapitata (a) and in B. calyciflorus and C. dubia (b). In
Figure S1a, the highest toxic effect of IF in P. subcapitata (95.67% The findings of the present study reveal that the toxic effects of
 C. Russo et al. / Environmental Pollution 233 (2018) 356e363 363

the tested parent compounds occurred at concentrations several ISO 14380, 2011. Water Quality-Determination of the Acute Toxicity to Thamnoce-
orders of magnitude higher (mg L
1) than their environmental phalus Platyurus (Crustacea, Anostraca). International Organization for Stan-
dardization, Geneva, Switzerland.
relevant concentrations (ng-mg L
1). However, wastewater treat- ISO 20665, 2008. Water Quality-determination of Chronic Toxicity to Ceriodaphnia
ment, metabolic and environmental transformations could lead to Dubia in 7 Days-population Growth Inhibition Test. International Organization
the formation of a mixture of parent compound residues and me- for Standardization, Geneva, Switzerland.
ISO 20666, 2008. Water Quality-determination of Chronic Toxicity to Brachionus
tabolites/TPs representing a harmful combination for the environ- Calyciflorus in 48 H-population Growth Inhibition Test. International Organi-
ment and then for human health. The UV-irradiation treatments, zation for Standardization, Geneva, Switzerland.
here performed, highlighted the increase of toxicity due to the Joqueviel, C., Martino, R., Gilard, V., Malet-Martino, M., Canal, P., Niemeyer, U., 1998.
Urinary excretion of cyclophosphamide in humans, determined by Phosphorus-
presence of these TPs. In this study, the only metabolite/TP showing 31 nuclear magnetic resonance spectroscopy. Drug Metabolism Dispos. 26 (5),
a toxic potential was carboxy-CP although the synergies among the 418e428.
other metabolites/TPs are still unknown and further studies are Kim, I., Tanaka, H., 2009. Photodegradation characteristics of PPCPs in water with
UV treatment. Environ. Int. 35 (5), 793e802.
needed to contribute to the environmental risk assessment. Kosjek, T., Heath, E., 2011. Occurrence, fate and determination of cytostatic phar-
maceuticals in the environment. Trends Anal. Chem. 30 (7), 1065e1087.
Acknowledgements Lekskulchai, V., 2016. Quantitation of anticancer drugs-Cyclophosphamide and
Ifosfamide in urine and water sewage samples by gas chromatography-mass
spectrometry. Int. J. Occup. Med. Environ. Health 29 (5), 815e822.
This work was financially supported by the EU through the EU Lin, A.Y., Lin, Y.C., Lee, W.N., 2014. Prevalence and sunlight photolysis of controlled
FP7 project CytoThreat (Fate and effects of cytostatic pharmaceu- and chemotherapeutic drugs in aqueous environments. Environ. Pollut. 187,
170e181.
ticals in the environment and the identification of biomarkers for
Lin, A.Y., Hsueh, J.H., Hong, P.K., 2015. Removal of antineoplastic drugs cyclophos-
an improved risk assessment on environmental exposure. Grant phamide, ifosfamide, and 5-fluorouracil and a vasodilator drug pentoxifylline
agreement no. 265264). Also, financial support from Slovene from wastewaters by ozonation. Environ. Sci. Pollut. Res. 22 (1), 508e515.
Research Agency is kindly acknowledged (P1-0143, L1-5457, J1- Lutterbeck, C.A., Kern, D.I., Machado, E.L., Kümmerer, K., 2014. Ecotoxicological
evaluation of selected anticancer drugs. Toxicol. Lett. 229, S71eS72.
6752, L1-7544). Lutterbeck, C.A., Wilde, M.L., Baginska, E., Leder, C., Machado, E.L., Kümmerer, K.,
2016. Degradation of cyclophosphamide and 5-fluorouracil by UV and simu-
Appendix A. Supplementary data lated sunlight treatments: assessment of the enhancement of the biodegrad-
ability and toxicity. Environ. Pollut. 208, 467e476.
Nicole, I., De Laat, J., Dore, M., Duguet, J.P., Bonnel, C., 1990. Utilisation du rayon-
Supplementary data related to this article can be found at nementultraviolet dansle traitement des eaux: mesure duflux photonique par
https://doi.org/10.1016/j.envpol.2017.10.066. actinometrie chimique au peroxyde d’hydrogene. Water Res. 24, 157e168.
Paixao, S.M., Silva, L., Fernandes, A., O'Rourke, K., Mendonca, E., Picado, A., 2008.
Performance of a miniaturized algal bioassay in phytotoxicity screening. Eco-
References toxicology 17, 165e171.
Ralhan, R., Kaur, J., 2007. Alkylating agents and cancer therapy. Expert Opin. Ther.
ASTM E 1440e1491, 2004. Standard Guide for Acute Toxicity with the Rotifer Bra- Pat. 17 (9), 1061e1075.
chionus. American Society for Testing and Materials, Philadelphia PA, USA Ristow, M., Schmeisser, K., 2014. Mitohormesis: promoting health and lifespan by
reapproved. increased levels of reactive oxygen species (ROS). Dose-Response Publ. Int.
Buerge, I.J., Buser, H.R., Poiger, T., Müller, M.D., 2006. Occurrence and fate of the Hormesis Soc. 12, 288e341.
cytostatic drugs cyclophosphamide and Ifosfamide in wastewater and surface Rosenfeldt, E.J., Linden, K.G., 2007. The ROH,UV concept to characterize and the
waters. Environ. Sci. Technol. 40 (23), 7242e7250. model UV/H2O2 process in natural waters. Environ. Sci. Technol. 41,

Cesen, M., Kosjek, T., Laimou-Geraniou, M., Kompare, B., Sirok,  B., Lambropolou, D., 2548e2553.
Heath, E., 2015. Occurrence of cyclophosphamide and ifosfamide in aqueous Sanderson, H., Johnson, D.J., Wilson, C.J., Brain, R.A., Solomon, K.R., 2003. Probabi-
environment and their removal by biological and abiotic wastewater treatment listic hazard assessment of environmentally occurring pharmaceuticals toxicity
processes. Sci. Total Environ. 15 (527e528), 465e473. to fish, daphnids and algae by ECOSAR screening. Toxicol. Lett. 144, 383e395.

Cesen, M., Kosjek, T., Busetti, F., Kompare, B., Heath, E., 2016a. Human metabolites Siddick, Z.H., 2002. In: Malcolm, R. (Ed.), The Cancer Handbook, first ed. Alison John
and transformation products of cyclophosphamide and ifosfamide: analysis, Wiley & Sons, Ltd.
occurrence and formation during abiotic treatments. Environ. Sci. Pollut. Res. 23 US EPA, 1993. Methods for Measuring the Acute Toxicity of Effluents and Receiving
(11), 11209e11223. Waters to Freshwater and Marine Organisms. EPA-600-4-90, fourth ed. US

Cesen, 
M., Elersek, T., Novak, M., Zegura, B., Kosjek, T., Filipi
c, M., Heath, E., 2016b. Environmental Protection Agency, Washington DC.
Ecotoxicity and genotoxicity of cyclophosphamide, ifosfamide, their metabo- Xie, H., 2012. Occurrence, ecotoxicology, and treatment of anticancer agents as
lites/transformation products and their mixtures. Environ. Pollut. 210, 192e201. water contaminants. J. Environ. Anal. Toxicol. S 2.
Eitel, A., Scherrer, M., Kümmerer, K., 2000. Handling Cytostatic Drugs: a Practical Zhang, J., Tian, Q., Zhou, S., 2006. Clinical pharmacology of cyclophosphamide and
Guide. Bristol-Myers-Squibb (2000). ifosfamide. Curr. Drug Ther. 1 (1), 55e84.
Gaya, A., Akle, C.A., Mudan, S., Grange, J., 2015. The concept of hormesis in cancer Zhang, Y., Xiao, Y., Zhang, J., Chang, V.W.C., Lim, T., 2017. Degradation of cyclo-
therapye is less more? Cureus 7 (4), 261. phosphamide and 5-fluorouracil in water using UV and UV/H2O2: kinetics
Gilard, V., Malet-Martino, M.C., de Forni, M., Niemeyer, U., Ader, J.C., Martino, R., investigation, pathways and energetic analysis. J. Environ. Chem. Eng. 5 (1),
1993. Determination of the urinary excretion of ifosfamide and its phospho- 1133e1139.
rated metabolites by phosphorus-31 nuclear magnetic resonance spectroscopy. Zhu, H., Long, M., Wu, J., Wang, M., Li, X., Shen, H., Xu, J., Zhou, L., Fang, Z., Luo, Y.,
Cancer Chemother. Pharmacol. 31 (5), 387e394. Li, S., 2015. Ginseng alleviates cyclophosphamide-induced hepatotoxicity via

Isidori, M., Lavorgna, M., Russo, C., Kundi, M., Zegura, B., Novak, M., Filipi c, M., reversing disordered homeostasis of glutathione and bile acid. Sci. Rep. 17536,
Misík, M., Knasmueller, S., Lopez de Alda, M., Barcelo 
 , D., Zonja, 
B., Cesen, M., 1e14.

S cancar, J., Kosjek, T., Heath, Ester, 2016. Chemical and toxicological character- Zounkova, R., Odr aska, P., Dole , L., Hilscherov
zalova a, K., Mars ha, L., 2007.
alek, B., Bla
isation of anticancer drugs in hospital and municipal wastewaters from Ecotoxicity and genotoxicity assessment of cytostatic pharmaceuticals. Environ.
Slovenia and Spain. Environ. Pollut. 219, 275e287. Toxicol. Chem. 26 (10), 2208e2214.