Reproductive Toxicology 50 (2014) 171–179

Contents lists available at ScienceDirect

Reproductive Toxicology journal homepage: www.elsevier.com/locate/reprotox

Safety profiles of gadolinium chelates in juvenile rats differ according to the risk of dissociation Nathalie Fretellier a,∗,1 , Meryam Maazouz a,1 , Alexandrine Luseau b , Fannie Baudimont b , Gaëlle Jestin-Mayer a , Simon Bourgery a , Marlène Rasschaert a , Patrick Bruneval c,d , Cécile Factor a , Fatiha Mecieb a , Jean-Marc Idée a , Claire Corot a a

Guerbet, Research & Innovation Division, 93600 Aulnay-sous-Bois, France Atlantic Bone Screen, 44800 Saint Herblain, France c Department of Pathology, Hôpital Européen Georges Pompidou, 75015 Paris, France d INSERM U970, Department of Pathology, Hôpital Européen Georges Pompidou, 75015 Paris, France b

a r t i c l e

i n f o

Article history: Received 8 July 2014 Received in revised form 8 October 2014 Accepted 28 October 2014 Available online 5 November 2014 Keywords: Gadolinium chelates Gadolinium Contrast agents Safety Juvenile toxicity Immaturity of renal function.

a b s t r a c t This study was designed to compare the safety of two gadolinium chelates (GCs), used as contrast agents for magnetic resonance imaging, in juvenile rats. Juvenile rats received five intravenous administrations (between postnatal day [PND] 4 and 18) of gadoteric acid (macrocyclic ionic GC), gadodiamide (linear nonionic GC) or saline, and sacrificed at PND 25. Gadodiamide induced mortality, alopecia and hyperpigmentation of dorsal skin. Two gadodiamide-treated rats presented severe epidermal and dermal lesions. No abnormal signs were detected following administration of gadoteric acid. Higher tissue gadolinium concentrations were found in the gadodiamide group compared to the gadoteric acid group. Dissociation of gadodiamide was observed in skin and liver, with the presence of dissociated and soluble gadolinium. In conclusion, repeated administration of gadoteric acid was well tolerated by juvenile rats. In contrast, gadodiamide induced significant toxicity and more marked tissue gadolinium retention (at least partly in the dissociated and soluble form). © 2014 Elsevier Inc. All rights reserved.

1. Introduction Gadolinium is a highly toxic lanthanide [1]. Its acute toxicity appears to be related to its ionic radius close to that of Ca2+

Abbreviations: GC, gadolinium chelate; PND, postnatal day; MRI, magnetic resonance imaging; CT, computed tomography; NSF, nephrogenic systemic fibrosis; EMA, European Medicines Agency; FDA, Food and Drug Administration; ELISA, enzyme-linked immunosorbent assay; HES, hematoxylin-eosin-saffron; LOD, limit of detection; LOQ, limit of quantification; HPLC, high pressure liquid chromatography; CNS, central nervous system. ∗ Correspondence to: Guerbet, Research Division, BP 57400, Roissy Charles de Gaulle, France. Tel.: +33 1 45 91 69 64; fax: +33 1 45 91 51 23. E-mail addresses: [email protected] (N. Fretellier), [email protected] (M. Maazouz), [email protected] (A. Luseau), [email protected] (F. Baudimont), [email protected] (G. Jestin-Mayer), [email protected] (S. Bourgery), [email protected] (M. Rasschaert), [email protected] (P. Bruneval), [email protected] (C. Factor), [email protected] (F. Mecieb), [email protected] (J.-M. Idée), [email protected] (C. Corot). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.reprotox.2014.10.024 0890-6238/© 2014 Elsevier Inc. All rights reserved.

[1]. The Gd3+ ion must therefore be chelated by an appropriate ligand to allow intravenous administration to patients as contrast agents for magnetic resonance imaging (MRI). Gadolinium chelates (GCs) are classified according to their chemical structure (linear or cyclic) and the ionic or nonionic nature of the ligand used to chelate Gd3+ . So-called “macrocyclic” GCs consist of a ligand forming a “cage” imprisoning Gd3+ (such as gadoteric acid). “Linear” chelates are molecules in which Gd is associated with a ligand with an “open” structure (such as gadodiamide). The thermodynamic and kinetic stabilities of the various GCs differ according to these structural characteristics [1,2]. Thermodynamic stability describes the strength of the bond between Gd and its ligand. Kinetic stability defines the rate at which equilibrium between the Gd chelate and its dissociated component is reached. The kinetic stability of macrocyclic GCs is much higher than that of linear GCs and the thermodynamic stability of ionic GCs is generally higher than that of nonionic GCs [1–3]. GCs are not metabolized and are almost exclusively excreted by the kidneys by glomerular filtration [1]. MRI has become a major tool for pediatric patients for three main reasons: superior tissue characterization compared to computed tomography (CT), smaller injected volumes of contrast agent are required compared to CT, and no exposure to ionizing radiation,

172

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

which is crucial in children [4]. Contrast agent administration is necessary to allow better visualization and characterization of micro-vessels or lymph, for example, or to distinguish certain tumors. Contrast-enhanced MRI examinations are performed in neonates and young children whenever the following clinical indications [5–7] are present: neonatal tumors, evaluation of lesions detected on prenatal ultrasound, hemodynamically significant vascular malformations, congenitally acquired infections and congenital heart disease. Repeated MRI monitoring may also be necessary in some cases, as the frequency of follow-up may vary depending on the grade of the lesions, biological activity and treatment [8]. For example, in children with brain tumors, serial MRI is commonly performed every 3–6 months after surgery, as long as clinically necessary. Although generally considered to be safe in the majority of patients, GCs can be associated with the development of a severe delayed adverse reaction, nephrogenic systemic fibrosis (NSF). NSF has only been diagnosed in patients with severe or end-stage renal failure [9]. Almost all cases of NSF induced by a single GC (so-called “unconfounded” cases), appear to be related to a specific category of GCs, linear GCs, and more particularly gadodiamide and gadopentetic acid [10], while almost no cases of confirmed NSF have been linked to the administration of other categories of GCs. In particular, no unconfounded case of NSF has been reported for gadoteric acid, evaluated in the present study. The youngest NSF pediatric published case was reported in a six-year-old child [11]. Anatomical development of the kidneys (nephrogenesis) in humans is complete at the 35th week of gestation [12] and is followed by a process of morphological and functional maturation of the kidneys to achieve an adult level of renal function around the age of two years [12]. Immaturity of renal function in neonates and infants may therefore result in increased systemic exposure, tissue retention of Gd3+ and significant adverse effects. The safety of GCs in pediatric patients is a matter of concern for radiologists and health authorities. Relatively few clinical safety data are available for GCs in the pediatric population. Moreover, nonclinical juvenile toxicity studies are rarely carried out to predict safety in the pediatric population [13,14]. A better understanding of the safety profile of GCs in this specific population is essential. The European Medicines Agency (EMA) has contraindicated the use of GCs considered to be associated with a high risk of NSF (gadodiamide, gadopentetic acid and gadoversetamide) in neonates under the age of 4 weeks. However, for GCs considered to be associated with a low risk (gadoteric acid, gadobutrol and gadoteridol) or intermediate risk of NSF (gadobenic acid), only a precaution for use has been issued in newborns up to the age of 4 weeks [15]. In the United States, the Food and Drug Administration (FDA) has not yet approved any GC for use in children under the age of two years. GCs are therefore administered off-label in children under the age of two years. In this study, we evaluated the systemic safety of gadolinium chelates with two different molecular structures and stabilities in juvenile rats. Firstly, we validated changes in renal function in juvenile rats over time. The toxicity profiles of two types of GCs with different thermodynamic and kinetic stabilities (gadodiamide and gadoteric acid) were then evaluated in juvenile rats and tissue Gd3+ concentrations were characterized. 2. Materials and methods The study protocol was approved by the in-house Animal Welfare Ethics Committee. Studies were conducted in accordance with European Directive 2010/63/EU and French legislation on animal welfare. All studies were carried out on Sprague-Dawley rats obtained from Charles River (Charles River Laboratories, L’Arbresle, France).

Animals were born at Charles River and were delivered in litters of 10 rats (five males and five females) at postnatal day (PND) 2. Animals were not weaned during the studies. 2.1. Characterization of renal function in juvenile rats Animals were sacrificed at nine different timepoints: PND 4, 8, 11, 14, 16, 18, 22, 28 and 30. At each timepoint, 3 males and 3 females were sacrificed in order to detect a possible sex-linked effect. At PND 4 and 8, animals were anesthetized by intraperitoneal administration of a mixture of ketamine (95 mg/kg) and xylazine (10 mg/kg). The volume of mixture injected was calculated with a ratio of 0.2 mL/100 g of body weight. From PND 11, animals were anesthetized by isoflurane (5%) supplemented with 1 L/min O2 and were then sacrificed by exsanguination by intracardiac puncture using a 1 mL heparinized syringe (up to PND 22). From PND 28, exsanguination was performed from the abdominal aorta, after laparotomy. 2.1.1. Biochemistry Plasma creatinine levels were assayed by an enzymatic method (Vitros Fusion 5.1, Ortho-Clinical Diagnostics, Inc., Issyles-Moulineaux, France) and plasma cystatin C was assayed by an immunoenzymatic method (enzyme-linked immunosorbent assay, ELISA) (Quantikine® ELISA – mouse/rat cystatin C immunoassay – R&D Systems Europe, Lille, France). 2.1.2. Histology At sacrifice, the kidneys were removed and fixed in 4% neutral buffered formalin. After dehydration, samples were paraffin-embedded, sectioned (5 ␮m thickness) and stained with hematoxylin-eosin-saffron (HES). Kidney samples from rats sacrificed at PND 4, PND 11 and PND 30 were examined under blinded conditions by a certified pathologist. 2.2. Comparative study of saline, gadoteric acid and gadodiamide in juvenile rats Two studies were conducted: one product per litter was administered in Study 1 and three products were administered to the same litter in Study 2. In both studies, two GCs were tested versus a control group (saline). Rats were therefore randomized to receive intravenous injections of 2.5 mmol Gd/kg (5.0 mL/kg) of gadoteric acid (Dotarem® , Guerbet, Villepinte, France), gadodiamide (Omniscan® , GE Healthcare, Chalfont St Giles, United Kingdom) or 5.0 mL/kg of isotonic saline (Lavoisier, Paris, France) on PND 4, 8, 11, 14 and 18 (cumulative dose: 12.5 mmol Gd/kg). Animals were placed in a ventilated chamber heated to 38 ◦ C to maintain a temperature similar to that of the dam before administration. Products were injected into the jugular vein for the first two injections (on PND 4 and 8) and the caudal vein for subsequent administrations (from PND 11) in conscious animals. Both studies were blinded (administrations, clinical examinations and assays). Animals were identified by an ear-notch system and were monitored daily. Clinical followup was evaluated on a previously developed rating scale including behavioral signs (rejection by the dam); not eating or drinking: dehydration with persistent skinfold; motor activity: prostration, attention, immobilization, little or no locomotor activity; behavioral responses to external stimuli: reaction after stimulation, curiosity); physiologic signs (body weight, tibial length, breathing, skin lesions, digestive problems, opening of eyes, presence of chromodacryorrhea and any other abnormal sign). In both studies, the animals were euthanized at PND 25 under isoflurane

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

anesthesia (i.e. one week after the last administration), so that renal function was still not fully mature at sacrifice. 2.2.1. Study 1 The study was conducted on three different litters and one product was injected in each litter. Each litter consisted of 10 rats (five males and five females). Animals were not weaned during the study. All rats from the same litter received the same treatment to avoid cross-contamination. 2.2.2. Study 2 A second study was conducted to evaluate the possibility of a litter-related effect according to exactly the same protocol as for the comparative study (Study 1) except for distribution of the products and the number of animals. One litter was treated as follows: three rats received gadoteric acid, four rats received gadodiamide and three rats received isotonic saline. 2.2.3. Histology In both studies, skin and liver samples were examined by a certified pathologist under blinded conditions. Three types of examinations were performed on skin sections to detect signs of fibrosis: (a) histopathological description of the tissues with semiquantitative evaluation (scoring system from 1 minimal to 5 severe) of elementary findings observed on hematoxylin-eosin (HE) stained sections, (b) histomorphometric analysis of fibrosis identified by picrosirius red staining and (c) histomorphometric analysis of fibroblast cell density on HE stained sections. 2.2.4. Biochemistry As for the characterization Study, plasma creatinine was assayed by an enzymatic method. Urine creatinine, plasma urea, and plasma levels of total calcium, phosphorus and transferrin-bound iron were assayed by a colorimetric method (Vitros DT60II, Ortho-Clinical Diagnostics, Inc., Issy-les-Moulineaux, France). Plasma aspartate transaminase, alanine aminotransferase, alkaline phosphatase, total bilirubin, albumin, globulin, total proteins, cholesterol, uric acid, urea, creatine kinase and amylase were assayed at sacrifice (MScan II, Melet Schloesing, Osny, France). 2.2.5. Determination of tissue total gadolinium Total Gd was assayed on previously mineralized samples by addition of 65% nitric acid at 80 ◦ C for 8 h. Total Gd was assayed in plasma, heart, kidney, skin (dorsal skin biopsy), liver (one sample from the left lobe), and left femur (both epiphyses) by ICP-MS (inductively coupled plasma – mass spectrometry). This technique is unable to distinguish free gadolinium, Gd3+ , from chelated gadolinium (GdL). The concentration obtained is corrected by the mineralization mass over the sample mass and by the dilution factor to obtain a concentration expressed in nmol/g. The limit of detection (LOD) was 0.015 ␮g/L and the limit of quantification (LOQ) was 0.05 ␮g/L. 2.2.6. Determination of dissociated gadolinium in plasma In addition to total gadolinium assays, the presence of dissociated Gd3+ was determined in plasma by high pressure liquid chromatography (HPLC) connected to an ICP-MS system according to Frenzel et al. [16]. The HPLC system (1260 Infinity Bioinerte, Agilent Technologies, Santa Clara, USA) was equipped with a 1mL Chelating Sepharose column (Hi-Trap, GE-Healthcare, Uppsala, Sweden) and was connected to an ICP-MS system (7700x, Agilent Technologies, Santa Clara, USA). The 158 Gd isotope signal intensity was recorded and displayed as a chromatogram. Peak area analysis was performed to determine the fractions of intact Gd complex and dissociated Gd3+ in each

173

sample. Dissociated Gd3+ was quantified against a calibration curve with Gd nitrate ranging from 0.1 to 20 ␮M in rat plasma.

2.2.7. Assessment of dissociation in solid tissues (by relaxometry) The presence of dissociated Gd3+ in femur, skin and liver was assessed by a relaxometry technique, as described elsewhere [17]. Matrices were diluted six-fold (liver) or 11-fold (skin, femur) in a mixture of D2 O/MilliQ water (90/10 v/v). Skin was ground with a Precellys® homogenizer (Bertin Technologies, Montigny-leBretonneux, France). Bone was first immersed in a liquid nitrogen bath (−196 ◦ C) and a first grinding step was then performed in an Ultra-Turrax® homogenizer (IMLAB, Lille, France), and a second grinding step was performed with a Precellys® homogenizer. Liver samples were ground with a GentleMacs Dissociator® (Miltenyi Biotec, Paris, France). Relaxation times T1 were measured with the Minispec® relaxometer (Bruker, Karlsruhe, Germany) at 60 MHz (i.e., a magnetic field of 1.42 Tesla), at 37 ◦ C with inversion/recovery sequences (T1 ). Samples were then stored at −20 ◦ C until total gadolinium assay by ICP-MS. When the 1/T1 − 1/T1 diamagnetic value was less than 20% of 1/Tdiamagnetic in the absence of Gd precipitation (i.e. due to a low total Gd concentration in the sample), it was considered that r1 relaxivity could not be determined. Total Gd was then determined in samples by ICP-MS. Relaxivities (in vivo r1 value) were calculated according to the formula: r1 = (1/T1 sample −1/T1 diamagnetic )/[Gd]sample , with relaxation rate (1/T1 ) expressed in s−1 , Gd concentration in mM and relaxivity r1 in mM−1 s−1 . Relaxometry studies (in vitro spiking studies) on tissue matrices were performed by spiking with GCs (range: 0, 0.005, 0.01, 0.02, 0.04, 0.05, 0.1, 0.5 and 1 mM) on D2 O/H2 O mix or rat femur, skin and liver from non-treated rats.

2.3. Statistical analysis Data are presented as mean ± standard deviation (when n > 2). The effect induced on weight changes, biochemical parameters and total gadolinium concentration was compared by one-way (product) analysis of variance (ANOVA). When an interaction was detected by ANOVA, a pairwise comparison was performed using the Bonferroni test. The 95% confidence interval for the relaxivity constant value r1 obtained in vivo (mean 1.96 ± SD) was calculated and then compared to the reference relaxivity constant value r1 obtained in vitro ± 23% (r1 interval in vitro). When the confidence interval of the mean for the relaxivity constant value r1 obtained in vivo was included in the in vitro r1 interval, the in vitro state of Gd in vivo was considered to be equal to the in vitro state of Gd (i.e., chelated form), otherwise the two states were considered to be different (i.e., dissociated and soluble or dissociated and precipitated forms). Differences were considered to be significant for an ˛ risk ≤ 0.05. Graph Pad Prism® software (GraphPad Software Inc. San Diego, CA, USA) was used for statistical analysis.

3. RESULTS 3.1. Characterization of renal function in juvenile rats 3.1.1. Histology At timepoint PND 4, glomeruli in the superficial cortex were immature, as they still corresponded to the capillary loop stage (Fig. 1). However, deeper glomeruli were mature. From PND 11 onwards, all of the renal parenchyma was mature (Fig. 1).

174

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

Fig. 1. Renal superficial cortex at different timepoints: At PND 4 immature capillary loop stage glomeruli (white arrow) are still present; At PND 11 all the glomeruli are mature showing patent urinary spaces (black arrows) lined by mature podocytes and Bowman’s capsule cells; At PND 30 all the glomeruli are mature and grown up in size (black arrow: urinary space). All the panels are stained with HE; original magnification 40X.

Fig. 2. Changes in plasma cystatin C (A) and creatinine (B) concentrations (individual data) as a function of age (PND corresponds to postnatal day); N = 6/age.

3.1.2. Biochemistry A significant reduction of plasma cystatin C concentrations was observed with age (Fig. 2A). No significant difference in cystatin C concentrations was observed between males and females. Cystatin C concentration decreased with age and became stable from PND 22 (1.73 ± 0.30 mg/L at PND 22 vs. 1.67 ± 0.26 mg/L at PND 30, NS). Plasma creatinine concentration relative to body weight decreased with age until PND 18, when plasma creatinine levels remained statistically stable until PND 30 (Fig. 2B).

No significant sex-linked difference was observed for plasma creatinine levels. 3.2. Comparative study of gadoteric acid and gadodiamide in juvenile rats 3.2.1. Clinical Signs The observed clinical signs (Studies 1 and 2) are summarized in Table 1.

Table 1 Observed clinical signs (Studies 1 and 2) in juvenile rats treated with gadoteric acid, gadodiamide (5 × 2.5 mmol Gd/kg) or saline (5.0 mL/kg) via the intravenous route at PND 4, 8, 11, 14 and 18. Clinical signs

Mortality

Product Study 1

Study 2

Study 1

Study 2

Gadoteric acid Gadodiamide

0/9a 2/10 (from PND 10)

0/3 2/4 (from PND 11)

Saline

0/10

0/3

No abnormal signs • Abnormal hair growth (10 rats; from PND 8) • Piloerection (8 rats; from PND 14) • Brown hyperpigmentation of the skin (8 rats; at PND 25) • Less reactive to stimuli (5 rats; from PND 16) No abnormal signs

No abnormal signs • Abnormal hair growth (4 rats; from PND 8) • Piloerection (2 rats; from PND 14) • Brown hyperpigmentation of the skin (2 rats; at PND 25) • Severe skin lesions and hair loss (2 rats; from PND 18) No abnormal signs

a

1 rat withdrawn from the study in the gadoteric acid group because of incomplete IV injection (at PND 14).

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

175

Fig. 3. Histological skin lesions observed in rats treated with gadodiamide: deep ulceration (black arrows) with dermal inflammation and fibroplasia (A) and diffuse acanthosis with dermal inflammation and fibroplasia (B), compared to normal skin (C). All the panels are stained with HE; original magnification 20X.

No statistically significant difference in plasma cystatin C concentrations, measured at PND 25, was observed in juvenile rats from Study 1 and Study 2. However, plasma cystatin C concentrations in juvenile rats were significantly higher than those of the dams (1.67 ± 0.36 mg/L at PND25 vs. 0.83 ± 0.19 mg/L, p< 0.001) (data not shown). 3.2.1.1. Study 1. One male from the litter that received gadoteric acid was withdrawn from the study and was sacrificed on PND 14 due to injection difficulties. Abnormal hair growth was observed in the group that received gadodiamide from PND 8 and became worse with time in all injected animals. The presence of piloerection was also observed during the study in all rats of the group. One female was found dead at PND 10 and one female was found dead at PND 14. At the time of sacrifice (PND 25), brown hyperpigmentation of the back was observed in all animals who received gadodiamide, while no abnormality was detected in the groups that received saline or gadoteric acid. No significant differences in the body weight of the rats were observed according to treatment. However, the inter-individual variability within the group that received gadodiamide was considerable. A possible effect on growth was evaluated by tibial length, but no significant difference according to treatment was observed. Eye opening was observed in all rats between PND 14 and PND 17. 3.2.1.2. Study 2. The three treatments were administered to a single litter in order to evaluate a possible litter-related effect. Deterioration of the general clinical condition was observed in the animals that received gadodiamide. One rat in this group died on PND 10 and two other rats presented severe skin lesions on the skull from PND 21. One of these animals was sacrificed for ethical reasons on PND 22. One rat that received gadodiamide also presented substantial alopecia of the back starting on PND 18. The three rats that received gadoteric acid or saline did not present any abnormal clinical signs. The body weight gain of the animals that received gadodiamide was lower than that of animals that received saline or gadoteric acid. 3.2.2. Histology 3.2.2.1. Study 1. Initiation of the hair growth phase, corresponding to hypertrophy of hair follicles in the deep hypodermis and marked skin thickening were observed in a skin sample from a rat treated with gadodiamide that was found dead on PND 10. No fibrotic process was observed in the dermis and the morphological differences (minimal to slight changes in fibroblast density, skin thickness and fibrous composition of the hypodermis) were attributed to individual variability and remained within physiologic limits.

Minimal or moderate portal inflammation (mononuclear cells with a few eosinophils), without fibrosis, was observed in liver samples from all rats, including the control group, except for one rat treated with gadoteric acid. These findings were considered to be unrelated to treatment, as they were also observed in control rats. 3.2.2.2. Study 2. More severe inflammation with a diffuse pattern was observed in the dermis of two rats that received gadodiamide compared to the other animals (Fig. 3). One gadodiamide-treated rat presented a widespread dermal granulomatous inflammation with multiple hair shaft debris boils (interpreted as furunculosis) and diffuse dermal fibroplasia. A full-thickness epidermal and superficial dermal necrosis (deep ulceration) was observed for the wide majority of the biopsy. This ulcer was covered by large amount of aggregates of serous exudate, keratin, cell debris including degenerate granulocytes neutrophils. A very superficial bacterial proliferation, considered as secondary opportunistic colonization, was also reported. Moderate dermal perivascular inflammation, composed of mononuclear cells with fewer granulocytes neutrophils, with edema of the superficial dermis were observed in the other rat. The rat also presented mild diffuse fibroplasia and diffuse regular acanthosis of the overlying epidermis (hyperplastic proliferation of epidermal cells) with orthokeratotic hyperkeratosis. No ulceration was present on the biopsy. Both rats treated with gadodiamide exhibited significant skin inflammation with epidermal necrosis and dermal fibrosis for one of them and marked acanthosis and mild fibrosis for the other one. Although these skin findings can be morphologically unspecific, a toxic effect of gadodiamide is highly suspected. 3.2.3. Biochemistry No significant difference was observed between groups, apart from plasma iron, which was significantly higher in rats that received gadoteric acid (74.3 ± 13.0 ␮mol/L) or gadodiamide (87.1 ± 11.7 ␮mol/L) compared to the control group (43.6 ± 6.2 ␮mol/L) (p< 0.001) at PND 25. Minor changes of plasma biochemistry parameters were observed after repeated administrations of gadoteric acid or gadodiamide: – higher mean alanine aminotransferase concentration in the gadoteric acid group (38.1 ± 6.1 U/L vs. 28.8 ± 10.6 U/L in the saline group, p< 0.05) – higher mean creatine kinase concentration in the gadoteric acid group (252.9 ± 74.8 U/L vs. 175.4 ± 54.3 U/L in the saline group p< 0.05) – higher mean bilirubin concentration in the gadodiamide group (3.1 ± 0.3 ␮mol/L vs. 2.5 ± 0.5 ␮mol/L in the saline group, p< 0.05).

176

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

Fig. 4. Total gadolinium concentration measured in plasma (A), kidney (B), heart (C), femur (D), skin (E) and liver (F) of juvenile rats at PND 25 and treated with gadoteric acid, gadodiamide (5 × 2.5 mmol Gd/kg) or saline (5.0 mL/kg) via the intravenous route at PND 4, 8, 11, 14 and 18. ICP-MS measurement, nmol Gd/g). Mean and individual data are shown. * p < 0.05; *** p < 0.001.

3.2.4. Total gadolinium Plasma total Gd concentrations were not significantly different between the gadoteric acid and gadodiamide groups (Fig. 4A); 0.20 ± 0.10 ␮mol/L vs. 0.21 ± 0.07 ␮mol/L, respectively). In contrast, total Gd concentrations in the kidney (Fig. 4B), heart (Fig. 4C), femur (Fig. 4D), skin (Fig. 4E) and liver (Fig. 4F) were significantly higher in animals that received gadodiamide compared to those that received gadoteric acid (p < 0.001 except for kidney samples, p < 0.05). A similar trend was observed in Study 2 (Fig. 4).

3.2.5. Plasma dissociated gadolinium Plasma dissociated Gd concentration was above the LOQ of 0.10 ␮mol/L in only one (from Study 1) of the 12 rats of the gadoteric acid group (0.11 ␮mol/L), while it was above the LOQ in the plasma of 9 of the 11 survivors on PND 25 in the gadodiamide group (mean dissociated Gd concentration: 0.12 ± 0.07 ␮mol/L, n = 11). The dissociated/total Gd ratio was 61 ± 32% in the gadodiamide group.

3.2.6. Relaxometry The relaxivity constant r1 for both studies was calculated according to the relaxometry study (i.e. dividing the relaxation rate (1/T1 ) by the total Gd concentration) in the femur, skin (Fig. 5A) and liver (Fig. 5B). The relaxivity constant r1 could not be determined in the femur due to the low total Gd concentrations (i.e. 0.018 mM) in the samples of the two studies. Total tissue Gd concentrations in the liver and skin in rats treated with gadoteric acid were also too low (i.e. 0.019 mM in the liver and 0.012 mM in the skin) to determine the constant r1 in both studies. On the other hand, the relaxivity constant r1 was determined in the liver of all rats that received gadodiamide and in the skin of 5/8 rats from Study 1 and all rats from Study 2. In these animals, the relaxivity constant r1 was higher than the reference range of r1 values (calculated in vitro), in both the liver and the skin. Total tissue Gd concentrations in the skin of the three rats from Study 1 that received gadodiamide were less than 0.008 mM and the relaxivity constant r1 could therefore not be calculated.

Fig. 5. In vivo relaxivity r1 values (60 MHz, 37 ◦ C) in skin (A) and liver (B) samples of juvenile rats at PND 25 treated with gadoteric acid or gadodiamide (5 × 2.5 mmol Gd/kg) via the intravenous route at PND 4, 8, 11, 14 and 18. Mean and individual data are shown. Colored bars correspond to in vitro r1 range of each GC in tissues (r1 ± 23%).

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

4. Discussion Insufficient data are currently available concerning the safety of GC administration in neonates and infants. Safety data for many drugs in children are often extrapolated from data obtained in adult animals [13,14]. According to health authority recommendations, studies in juvenile animals should only be considered when the available safety data in adult animals and humans are insufficient [18,19]. It can be assumed that children under the age of 2 years, with immature renal function, would be more susceptible to the toxicity of renally-excreted drugs than adults [13,20]. In this context, we considered it important to compare the safety of two marketed GCs with distinct kinetic and thermodynamic stabilities in a juvenile animal model. We therefore decided to compare gadoteric acid, a macrocyclic ionic GC, and gadodiamide, a linear nonionic GC. The choice of animal species must be relevant to human development [13,20]. Rats are often chosen for these studies [20–22]. In our study, histologic examination of the kidneys demonstrated structural maturity of the kidneys at PND 11. These data are consistent with completion of nephrogenesis in rats at the age of 11 days [12]. In humans, nephrogenesis is completed around 35 weeks of pregnancy and therefore does not continue after birth [12]. In our study, two different biochemical markers of renal function in plasma of juvenile rats were used: plasma cystatin C and creatinine levels. Plasma creatinine levels is the most commonly used biological marker to estimate glomerular filtration rate (GFR). However, this marker is of low accuracy and depends on sex and body mass. In contrast, the protein cystatin C is independent of sex, age, height and body mass. Moreover, its plasma concentration only depends on GFR (since it is not secreted in the tubule). Since a rapid decrease in serum cystatin C levels has been observed after birth, it has been proposed that cystatin C may be a better renal function marker than plasma creatinine levels in children below two years old [23]. To our knowledge, the value of this marker has not yet been validated in juvenile rats, although it is already widely used in healthy or renally-impaired adult rats [24]. Both biochemical markers were stable from PND 18 to 22. Our results are consistent with the FDA guidance document indicating that juvenile rats reach adult GFR values between PND 15 and PND 21 [18]. We subsequently compared the safety of two GCs with clearly different stabilities and structure in juvenile rats. Based on data of the characterization study, we administered the products between PND 4 and PND 18, i.e. prior to renal maturity. In man, the GC dose administered in children is usually the same as that used in adults (0.1 mmol/kg) [5,25]. We used the maximum feasible dose (2.5 mmol/kg) based on the concentration of commercial solutions (500 mM) and the dose volume (5.0 mL/kg, maximum volume load) via the intravenous route in neonatal rats. A dose of 2.5 mmol Gd/kg/day in rats is approximately equivalent to a dose of 0.4 mmol Gd/kg in humans, after correction for body surface area, as recommended by the FDA [26]. Five injections were administered to investigate accumulation of Gd in various tissues and to mimic repeated contrast-enhanced MRI studies. Two different methodologies were used. In the first study, one product was injected per litter to avoid cross-contamination. However, this methodology can induce a “litter”-related effect (stress for the dam, rejection of certain animals) or a genetic bias [20]. In the second study, the three test products were therefore administered to the same litter. Similar clinical manifestations (systemic and cutaneous) were observed in both of these studies in gadodiamide-treated rats, but were more severe in the second study. These results confirm the absence of a litter effect in Study 1. A genetic bias in Study 1 was also ruled out, as the litters were composed of animals from different litters adopted by a lactating dam.

177

Systemic safety clearly differed according to the GC administered, as gadodiamide induced significant mortality (4 out of 14 rats), substantial hair growth abnormalities, alopecia, the appearance of severe skin lesions on the head (2 out of 14 rats) and brown pigmentation of the dorsal skin. Two animals (Study 2) with skin lesions presented epidermal necrosis, probably due to a toxic effect. Significant dermal inflammation of the dermis was also observed in these animals with dermal fibrosis. Similarly, alopecia, due to atrophy of the hair bulb and/or necrosis of the sheath of hair follicles could be related to a toxic effect of the GC. The cause of hyperpigmentation of the dorsal skin of the rats treated with gadodiamide remains unclear. This sign is classically observed in NSF patients [27], but the clinical features observed in our study appear to suggest systemic toxicity rather than NSF-like skin lesions. No cutaneous or systemic toxicity was observed after repeated administration of gadoteric acid to juvenile rats in these two studies, confirming the excellent safety of this GC, as reported elsewhere [28–30]. A higher iron concentration was observed in the plasma following administration of both GCs. We have no explanation or hypothesis on this finding. To our knowledge, this effect has never been reported so far in rats. A longer treatment-free period may allow to investigate whether this effect is reversible or not. At the time of sacrifice, dissociated Gd3+ was detected in the plasma of almost all rats receiving gadodiamide (9 out of 11 rats), but in only one gadoteric acid-treated rat, which is consistent with studies performed in adult, renally-impaired rats [17,28], reflecting the higher thermodynamic and kinetic stabilities of the macrocyclic ionic GC gadoteric acid compared to the linear nonionic gadodiamide [1]. However, when the dissociated Gd3+ concentration was quantifiable in the plasma, it was low and close to the limit of quantification, regardless the GC administered. Total tissue Gd concentrations were significantly higher in rats treated with gadodiamide compared to rats treated with gadoteric acid, which is consistent with all studies comparing macrocyclic GCs, especially gadoteric acid, with gadodiamide, in adult rats with either normal renal function [30] or renal failure [28,29,31,32]. Plasma total Gd concentrations at sacrifice were similar for the two GCs studied. The differences in clinical signs observed between the two GCs were probably due to different degrees of tissue Gd retention, related to the thermodynamic stability of the chelate. However, the chelated or dissociated form of Gd measured in tissues could not be distinguished by the mass spectrometry technique used in this study. The relaxometry technique (which measures proton relaxation time T1 ), combined with total Gd assay qualitatively addresses this issue. The calculated relaxivity value r1 is characteristic of a given form of Gd3+ [17]. This technique therefore constitutes a useful tool to investigate in vivo dissociation of GC [17]. Our results suggest the presence of soluble dissociated Gd3+ in the liver parenchyma and skin of animals receiving gadodiamide and therefore in vivo dissociation of this GC. The presence of dissociated Gd in the tissues of juvenile rats, long after repeated administrations of gadodiamide, could be responsible for toxic effects [33,34]. The relaxivity constant r1 could not be determined in the femur due to the low total tissue Gd concentrations, in contrast with previous studies in adult rats in which Gd accumulation (partially dissociated) in the femur was observed with gadodiamide [17]. In juvenile rats as in human neonates, the femur is not totally formed and ossification continues after birth [35,36]. Immature bone also primarily consists of cartilage (collagen and glycosaminoglycans), while adult bone primarily consists of cortical bone (mostly composed of hydroxyapatite) [35,37]. The high affinity of lanthanides, especially Gd, for hydroxyapatite crystals [37,38] could account for these findings. In contrast with gadodiamide-treated rats, the relaxivity constant r1 could not be determined in gadoteric acid-treated rats due

178

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179

to the low total Gd concentrations observed in all organs of these rats, confirming the results of studies in adult renally impaired rats [17,28]. In rats, GFR reaches adult values around PND 20 [12,21,22], corresponding to an age of about 2 years in humans [18,23,39]. The time window selected for GC administrations in our study was therefore roughly equivalent to the first two years of life in humans. Under our conditions, unlike gadoteric acid, repeated intravenous administrations of gadodiamide from PND 4 to PND 18 induced systemic toxicity in juvenile rats. According to the currently approved summary of product characteristics, gadodiamide is approved for central nervous system (CNS) examinations in children from 4 weeks of age in all European countries [40], while the age limit for whole-body MRI varies from one country to another [40]. Wholebody MRI is approved for children over the age of 2 years in the United Kingdom and Romania, but for children over the age of 6 months in many European countries [40]. EMA recommends contraindicating the use of the high-NSF risk category of GCs, including gadodiamide, in neonates up to 4 weeks of age [15]. According to a web-based survey of pediatric radiology departments in the United States and Canada in 2011, gadodiamide had been used in pediatric patients by 10% (n = 4/42) of responding institutions [41]. 5. Conclusions In conclusion, repeated administrations of gadoteric acid were well tolerated in rats despite their immature renal function. In contrast, gadodiamide induced significant morbidity and mortality, alopecia and higher levels of tissue Gd retention (at least partly in the soluble, dissociated form). Studies with a longer follow-up before sacrifice would be useful to evaluate long-term toxicity or reversibility of the observed effects. Conflict of interest N Fretellier, M Maazouz, G Jestin-Mayer, S Bourgery, M Rasschaert, C Factor, F Mecieb, J-M Idée and C Corot are or were employees of Guerbet. Histopathological studies were performed by A Luseau, F Baudimont (Atlantic Bone Screen) and P Bruneval (Department of Pathology, Hôpital Européen Georges Pompidou) under subcontracting arrangements. Transparency document The Transparency document associated with this article can be found in the online version. Acknowledgments The authors thank Dr Anthony Saul for reviewing the English version of the manuscript. The study was sponsored by Guerbet. References [1] Idée JM, Port M, Raynal I, Schaefer M, Le Greneur S, Corot C. Clinical and biological consequences of transmetallation induced by contrast agents for magnetic resonance imaging: a review. Fundam Clin Pharmacol 2006;20:563–76. [2] Port M, Idée JM, Medina C, Robic C, Sabatou M, Corot C. Efficiency, thermodynamic and kinetic stability of marketed gadolinium chelates and their possible clinical consequences: a critical review. Biometals 2008;21:469–90. [3] Laurent S, Elst LV, Copoix F, Muller RN. Stability of MRI paramagnetic contrast media: a proton relaxometric protocol for transmetallation assessment. Invest Radiol 2001;36:115–22. [4] Pearce MS, Salotti JA, Little MP, McHugh K, Lee C, Kim KP, et al. Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 2012;380:499–505.

[5] Emond S, Brunelle F. Gd-DOTA administration at MRI in children younger than 18 months of age: immediate adverse reactions. Pediatr Radiol 2011;41:1401–6. [6] Saunders DE, Thompson C, Gunny R, Jones R, Cox T, Chong WK. Magnetic resonance imaging protocols for paediatric neuroradiology. Pediatr Radiol 2007;37:789–97. [7] Meng H, Grosse-Wortmann L. Gadolinium in pediatric cardiovascular magnetic resonance: what we know and how we practice. J Cardiovasc Magn Reson 2012;14:56. [8] Mentzel HJ, Seidel J, Fitzek C, Eichhorn A, Vogt S, Reichenbach JR, et al. Pediatric brain MRI in neurofibromatosis type I. Eur Radiol 2005;15:814–22. [9] Girardi M, Kay J, Elston DM, Leboit PE, Abu-Alfa A, Cowper SE. Nephrogenic systemic fibrosis: clinicopathological definition and workup recommendations. J Am Acad Dermatol 2011;65:1095–106, e7. [10] Thomsen HS, Morcos SK, Almén T, Bellin M-F, Bertolotto M, Bongartz G, et al. Nephrogenic systemic fibrosis and gadolinium-based contrast media: updated ESUR Contrast Medium Safety Committee guidelines. Eur Radiol 2013;23:307–18. [11] Nardone B, Saddleton E, Laumann AE, Edwards BJ, Raisch DW, McKoy JM, et al. Pediatric nephrogenic systemic fibrosis is rarely reported: a RADAR report. Pediatr Radiol 2014;44:173–80. [12] Zoetis T, Hurtt ME. Species comparison of anatomical and functional renal development. Birth Defects Res B Dev Reprod Toxicol 2003;68:111–20. [13] Beck MJ, Padgett EL, Bowman CJ, Wilson DT, Kaufman LE, Varsho BJ, et al. Nonclinical juvenile toxicity testing. Dev Reprod Toxicol: Pract Approach 2006:263–328. [14] Piersma AH, Tonk EC, Makris SL, Crofton KM, Dietert RR, van Loveren H. Juvenile toxicity testing protocols for chemicals. Reprod Toxicol 2012;34:482–6. [15] EMA. Assessment report for gadolinium-containing contrast agents. London: European Medicines Agency; 2010 (EMA/740640/2010). [16] Frenzel T, Lengsfeld P, Schirmer H, Hutter J, Weinmann HJ. Stability of gadolinium-based magnetic resonance imaging contrast agents in human serum at 37 degrees C. Invest Radiol 2008;43:817–28. [17] Fretellier N, Idée JM, Dencausse A, Karroum O, Guerret S, Poveda N, et al. Comparative in vivo dissociation of gadolinium chelates in renally impaired rats: a relaxometry study. Invest Radiol 2011;46:292–300. [18] FDA. Guidance for industry: nonclinical safety evaluation of pediatric drug products. Rockville, MD: Center for Drug Evaluation and Research; 2006. [19] EMA. Guideline on the need for non-clinical testing in juvenile animals of pharmaceuticals for paediatric indications. London: European Medicines Agency; 2008 (EMEA/CHMP/SWP/169215/2005). [20] De Schaepdrijver LM, Bailey GP, Coogan TP, Ingram-Ross JL. Juvenile animal toxicity assessments: decision strategies and study design. Pediatric drug development: concepts and applications. 2013, 201–21. [21] Fleck C. Determination of the glomerular filtration rate (GFR): methodological problems, age-dependence, consequences of various surgical interventions, and the influence of different drugs and toxic substances. Physiol Res 1999;48:267–79. [22] Falk G. Maturation of renal function in infant rats. Am J Physiol 1955;181:157–70. [23] Fischbach M, Graff V, Terzic J, Bergere V, Oudet M, Hamel G. Impact of age on reference values for serum concentration of cystatin C in children. Pediatr Nephrol 2002;17:104–6. [24] Bokenkamp A, Ciarimboli G, Dieterich C. Cystatin C in a rat model of end-stage renal failure. Ren Fail 2001;23:431–8. [25] Bird CR, Drayer BP, Medina M, Rekate HL, Flom RA, Hodak JA. Gd-DTPA-enhanced MR imaging in pediatric patients after brain tumor resection. Radiology 1988;169:123–6. [26] FDA. Guidance for industry: estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers. Rockville, MD: Center for Drug Evaluation and Research (CDER); 2005. [27] Cowper SE, Rabach M, Girardi M. Clinical and histological findings in nephrogenic systemic fibrosis. Eur J Radiol 2008;66:191–9. [28] Fretellier N, Bouzian N, Parmentier N, Bruneval P, Jestin G, Factor C, et al. Nephrogenic systemic fibrosis-like effects of magnetic resonance imaging contrast agents in rats with adenine-induced renal failure. Toxicol Sci 2013;131:259–70. [29] Haylor J, Schroeder J, Wagner B, Nutter F, Jestin G, Idée JM, et al. Skin gadolinium following use of MR contrast agents in a rat model of nephrogenic systemic fibrosis. Radiology 2012;263:107–16. [30] Sieber MA, Lengsfeld P, Frenzel T, Golfier S, Schmitt-Willich H, Siegmund F, et al. Preclinical investigation to compare different gadolinium-based contrast agents regarding their propensity to release gadolinium in vivo and to trigger nephrogenic systemic fibrosis-like lesions. Eur Radiol 2008;18:2164–73. [31] Fretellier N, Idée J, Bruneval P, Guerret S, Daubine F, Jestin G, et al. Hyperphosphataemia sensitizes renally impaired rats to the profibrotic effects of gadodiamide. Br J Pharmacol 2012;165:1151–62. [32] Pietsch H, Lengsfeld P, Steger-Hartmann T, Lowe A, Frenzel T, Hutter J, et al. Impact of renal impairment on long-term retention of gadolinium in the rodent skin following the administration of gadolinium-based contrast agents. Invest Radiol 2009;44:226–33. [33] Manjunath V, Perazella M. Imaging patients with kidney disease in the era of NSF: can it be done safely. Clin Nephrol 2011;75:279–85. [34] Morcos SK. Experimental studies investigating the pathophysiology of nephrogenic systemic fibrosis: what did we learn so far. Eur Radiol 2011;21: 496–500.

N. Fretellier et al. / Reproductive Toxicology 50 (2014) 171–179 [35] Zoetis T, Tassinari MS, Bagi C, Walthall K, Hurtt ME. Species comparison of postnatal bone growth and development. Birth Defects Res B Dev Reprod Toxicol 2003;68:86–110. [36] Horton JA, Bariteau JT, Loomis RM, Strauss JA, Damron TA. Ontogeny of skeletal maturation in the juvenile rat. Anat Rec (Hoboken) 2008;291:283–92. [37] Vidaud C, Bourgeois D, Meyer D. Bone as target organ for metals: the case of f-elements. Chem Res Toxicol 2012;25:1161–75. [38] Abraham JL, Thakral C, Skov L, Rossen K, Marckmann P. Dermal inorganic gadolinium concentrations: evidence for in vivo transmetallation and long-term persistence in nephrogenic systemic fibrosis. Br J Dermatol 2008;158:273–80.

179

[39] Hogg RJ, Furth S, Lemley KV, Portman R, Schwartz GJ, Coresh J, et al. National Kidney Foundation’s Kidney Disease Outcomes Quality Initiative clinical practice guidelines for chronic kidney disease in children and adolescents: evaluation, classification, and stratification. Pediatrics 2003;111:1416–21. [40] HMA. Public assessment report for paediatric studies submitted in accordance with Article 45 of Regulation (EC) No1901/2006, as amended – Gadodiamide. UK: Heads of Medicines Agencies; 2013 (UK/W/063/pdWS/001). [41] Trout AT, Dillman JR, Ellis JH, Cohan RH, Strouse PJ. Patterns of intravenous contrast material use and corticosteroid premedication in children—a survey of Society of Chairs of Radiology in Children’s Hospitals (SCORCH) member institutions. Pediatr Radiol 2011;41:1272–83.