Objective
The objective of this study was to determine the fetal drug compartment concentrations when various concentrations of carboplatin cross the placental-trophoblastic barrier and the effect on the fetal kidneys.
Study Design
An ex vivo human placenta perfusion model was utilized. Term human placentae (n = 9) were collected immediately after delivery and then reperfused with plasma concentrations achieved with carboplatin an area under the curve of 5 (1000 ng/mL), 7.5 (5000 ng/mL), or 11 (11,000 ng/mL). Antipyrine was used as a reference compound. Samples were collected over 2 hours. Placental transfer was evaluated by computation of transport fraction and clearance index. Primary cells isolated by explant culture of 16-18 week old fetal organ tissues were incubated with carboplatin for up to 48 hours with untreated cell as controls. Immunohistochemical, flow cytometry analysis, and immunoblotting were applied for the expression of apoptosis-related proteins.
Results
Mean transport fractions for carboplatin at low, middle, and high concentrations were 0.05 ± 0.02, 0.04 ± 0.01, and 0.10 ± 0.01, respectively, with clearance indexes of 0.22 ± 0.01, 0.14 ± 0.08, and 0.50 ± 0.07, respectively. The fetal peak concentrations of carboplatin achieved were 61 ± 39 ng/mL (low), 375 ± 248 ng/mL (middle), and 2081 ± 529 ng/mL (high). Fetal kidney cells exposed to carboplatin showed a concentration-dependent increased expression of apoptosis-inducing factor and p53 apoptosis proteins and a time-dependent increase in expression Bax apoptosis protein expression. Apoptosis was confirmed at the high concentration by flow cytometry.
Conclusion
Doses of carboplatin up to an area under the curve of 7.5 were not associated with significant placental transfer, fetal exposure, or fetal toxic effects. This suggests it might not be necessary to empirically reduce carboplatin doses in pregnant women.
Gynecological cancers are among the most common malignancies that frequently occur in reproductive-age women. It is estimated that gynecological malignancies complicate approximately 1 in 1000 pregnancies annually. This number is expected to rise as more women delay childbearing into their later reproductive years as maternal age is known to be the most powerful predictor of cancer risk. Although cancer during pregnancy is uncommon, it is not rare. Smith et al reported an incidence of primary malignant neoplasm of 0.71 per 1000 live births in a population of more than 3 million deliveries. According to the National Center for Health Statistics, cancer is the second leading cause of death in women 25-44 years of age, and the most frequent cases involve the hemopoietic and lymphatic systems followed by thyroid, breast, cervix, ovary, colon, and melanoma.
Although chemotherapy may improve outcomes, clinicians are often reluctant to use chemotherapy during pregnancy because of the potential risk of fetal growth restriction, the carcinogenic effects on the fetus, and the fear of inducing fetal malformations. The risk of adverse fetal effects from chemotherapy is primarily influenced by the amount of the chemotherapeutic agent that crosses the placental/trophoblastic barrier (PTB) and the gestational age of the fetus at the time of exposure to these antineoplastic agents. Data are lacking for most antineoplastic drugs, but many clinicians still consider these agents as potentially harmful to the developing fetus, especially if given during the critical period of organogenesis.
Although there are no apparent adverse fetal effects when these agents are used in the second or third trimester, there is a paucity of information regarding the long-term effects on the fetus and child development. Because there is limited knowledge on the safety of chemotherapy on the fetus, oncologists have difficulty weighing the risks and benefits of particular anticancer agents when recommending a course of therapy for the pregnant patient. Furthermore, there are concerns for the loss of efficacy in chemotherapy agents because of the alteration in pharmacokinetics based on increased volume of distribution and/or increased clearance as result of the additional compartments as well as alteration plasma protein binding or changes in cytochrome P450 (CYP450) activity and renal clearance.
The effects of exposure to the common anticancer agents used for the treatment of cancer in pregnant women have not been studied in great depth, specifically with respect to fetal physiology and development. The important factor responsible for the adverse effects of chemotherapy drugs on fetal exposure is the amount of drug that crosses the PTB. Although numerous animal models such as sheep or guinea pig placentas have been utilized to evaluate the permeability and placental transfer of hydrophilic drugs, there is significant interspecies variability with up to 30-fold differences between placenta models. Thus, human tissue based models are required.
The human placental cotyledon model is an ex vivo model that can be used to evaluate maternal drug transfer across the PTB to determine drug exposure to the human fetus. For example, Bawdon and colleagues using an ex vivo human placental perfusion model have analyzed multiple drugs such as bisheteroypiperazine, lamivudine, zidovudine, azidothymidine, 2′, 3′-dideoxyinosine, 2′,3′-dideoxycytidine, and rosiglitazone and determined the amount that crossed the placental/trophoblastic barrier.
Carboplatin is among the most commonly used anticancer agents during pregnancy, especially for ovarian cancers. The extent and duration of carboplatin’s adverse reactions depend on the dose and regimen schedule. Potential toxicities include myelosuppression, neurotoxicity, nephrotoxicity, emesis, alopecia, ototoxicity, and hypomagnesemia. Therefore, the objective of this study was conducted to determine the amount of carboplatin that crosses the placenta barrier using an ex vivo human placental perfusion model and to elucidate its toxic effect on fetal kidney cells, particularly after repeated exposure.
Materials and Methods
Carboplatin placenta perfusion study
Placenta collection
Human term placentas (37-40 weeks’ gestation) were obtained at Memorial Hermann Hospital (Houston, TX) from uncomplicated, HIV-negative, hepatitis B-negative pregnancies immediately after delivery from either cesarean or vaginal deliveries in accordance with the University of Texas Health Science Center at Houston Medical School Institutional Review Board for Human Studies between August 2009 and May 2011. The Institutional Review Board Committee granted a waiver for obtaining an informed consent for the use of discarded tissues in these studies. Placentas were transported to the laboratory in warmed 0.9% saline solution immediately after delivery and reperfused within 30 minutes.
Placenta perfusion
The single cotyledon placenta perfusion system was used as described by Challier. Briefly, a fetal artery and vein on the chorionic plate were cannulated with a 3.0F and a 5.0F catheter, respectively, to complete a vascular circuit and then was gradually reperfused with Eagle’s minimum essential medium containing 1 U/mL heparin from a porcine source (Sigma, St. Louis, MO) to prevent clotting, and 3 g/dL bovine serum albumin, a plasma protein to account for potential drug plasma protein binding (Sigma). Adjustment of pH and buffering was undertaken to achieve a value of 7.40 using sodium hydroxide and 2.2 g/L sodium bicarbonate.
After confirmation of circulatory integrity, the cotyledon and portion of its surrounding placental tissue were transferred to a temperature-controlled (37 °C) chamber in which the fetal circulation was perfused as a closed system at a rate between 4.5 and 4.8 mL/min for approximately 20 minutes to remove any residual blood. The placental chamber pressure was monitored by a BP-1 pressure monitor (World Precision Instruments, Sarasota, FL) and maintained below 60 mm Hg for the entire experiment. To establish a maternal circulation, 3 18-guage needle probes were then inserted into intervillous space of the cotyledon, and then maternal flow rate was initiated at 17 mL/min. The placental cotyledon circulation circuits that leaked or that did not achieve a stable pressure secondary to lack of vascular integrity were discarded.
The media used for circulation of the maternal and fetal compartments consisted of 150 mL of Eagle’s minimal essential medium (pH 7.2-7.4), were aerated with 95% oxygen and 5% carbon dioxide, and continually mixed by a magnetic stirring bar. Placenta transfer studies were performed with carboplatin purchased from The University of Texas M. D. Anderson Division of Pharmacy.
Carboplatin was added to the maternal media at 3 concentrations (low: 1000 ng/mL; medium: 5000 ng/mL; high: 11,000 ng/mL), which are similar to the peak plasma concentrations achieved with carboplatin doses of areas under the curve 5, 7.5, and 11, respectively, that are commonly used alone or in combination regimens for the treatment of solid and hematological malignancies. Antipyrine at a concentration of 100mg/L was added to the media to the maternal side of the model as the positive control to evaluate placenta transfer as described in the reference method by Challier.
The experiments were first conducted in an open-open (no recirculation) setup to determine the transport fraction (portion of drug that crossed placenta from maternal to fetal side). Experiments were then repeated in a closed-closed recirculation system to determine clearance index and accumulation of carboplatin that may occur based on a balance of the transport fraction and the rate of clearance from fetal side to maternal side. The drug will accumulate on the fetal side when the drug crosses the PTB at a higher rate (transport fraction) from the maternal side to the fetal side compared with the rate the drug is eliminated (clearance index) or crosses back from the fetal side to the maternal side. The sample aliquots from both the fetal and maternal compartments were collected every 10 minutes during both the open and closed experiments for further analysis to determine the drug concentrations. All experiments containing the low, medium, and high concentrations of carboplatin were performed in triplicate.
The transport factions, clearance, and clearance index of carboplatin and antipyrine were determined based on the Challier formula for placental transfer. The following equations were used for the determination of placental transport: TF = (CFV-CFA)/(CMA-CFA), where TF is the transport fraction, C is the concentration, M is the maternal perfusate, F is the fetal perfusate, A is the artery, and V is the vein; and Cl = TF∗Q, where Cl is the clearance of the compound, TF is the transport fraction, and Q is the flow rate of the fetal circulation. Ci is the clearance of the carboplatin/clearance of antipyrine, where Ci is clearance index.
Carboplatin assay
Carboplatin concentrations were quantified by a validated atomic absorption assay. After collection, the maternal and fetal samples were centrifuged in Centrifree cartridges purchased from Millipore (Billerica, MA) to remove the protein from sample, and this ultrafiltrate was frozen at –20°C until ready to be analyzed. A flameless Varian SpectrAA-300 atomic absorption spectrometer (Palo Alto, CA) was used to quantify platinum concentrations with a Varian GTA-96 used for detection. A furnace operation using a graphite tube atomizer was used in place of a flame for sample atomization. Platinum concentrations in the samples were determined by the platinum standard curve and were converted to their carboplatin equivalent concentrations. The concentration range of carboplatin was 25-17000 ng/mL with a correlation coefficient (r) of 0.992.
Antipyrine assay
Antipyrine concentrations were quantified by validated high pressure liquid chromatography assay. Placental perfusion samples were kept in the subfreezer (–80°C) until time of analysis and then were thawed at room temperature for analysis. Antipyrine was isolated from fluid aliquots by liquid/liquid extraction with dichloromethane. Liquid chromatographic separation was achieved by isocratic mobile phase consisting of 67 mM phosphate buffer (pH 3.5)-methanol (91:10 to 40:60) with a flow rate of 0.5 mL/min to achieve antipyrine elution at 8.45 minutes from a Waters Nova-Pak C18, 3.9 × 150 mm, 4 μm particle size packing analytical column (Waters Corp, Milford, MA) for the antipyrine assay identified from other peaks using a photodiode array detector at a wavelength of 243 nm. The antipyrine assay was found to be linear over a concentration range of 6.25-100 μg/mL with a correlation coefficient (r) of 0.9981.
Sample size and statistical considerations for placenta model
In the literature placental perfusion studies have included 3 placentas evaluated at each concentration under both conditions of open/open and closed/closed models to determine descriptive statistical parameters such as mean values with SD and coefficient of variance. The low, middle, and high concentration transfer studies for carboplatin were each performed in 3 placentas for a total of 9 placentas that were included to complete the study. Studies were completed first in an open system for 1 hour and then a closed system for 1 hour. All sample analyses were performed in quadruplicate to decrease sample-to-sample data variation. Descriptive statistics including the mean, SD, and the coefficient of variance were used for each concentration of carboplatin and antipyrine.
Correlative fetal kidney toxicology studies
Because carboplatin is exclusively cleared by the kidneys to be eliminated, the focus of toxicity experiments was on the impact on fetal kidney development. These studies were not designed to define the no observed effect level (NOEL) for carboplatin, so organ sites were not evaluated.
Carboplatin (10 mg/mL) was purchased from The University of Texas M. D. Anderson Cancer Center, Division of Pharmacy. Hanks’ balanced salt solution, fetal bovine serum, chicken serum, trypsin-ethylenediaminetetraacetic acid, N-2-hydroxyethylpipera-zine-N′-2-ethane sulfonic acid buffer, collagenase 1A, collagenase 1, and deoxyribonuclease 1 were purchased from GIBCO Invitrogen Co (Carlsbad, CA). Dibutyryl cyclic adenosine monophosphate and antibiotic antimycotic solution were purchased from Sigma-Aldrich Co (St. Louis, MO).
The BCA protein estimation kit was purchased from Pierce (Rockford, IL). All primary antibodies including p53, apoptosis-inducing factor (AIF), Bax, caspase-3, B-actin, and goat antirabbit and goat antimouse IgG purified antibody were purchased from Santa Cruz Biotechnology, Inc (Santa Cruz, CA).
The buffer solutions including 40% acrylamide/Bis solution, 19:1, N,N,N′,N′-tetra-methylethylenediamine, 10% sodium dodecyl sulfate (SDS), 10× Tris/glycine/SDS buffer, 10× Tris/glycine buffer, 10× Tris-buffered saline, Tween 20, blotting grade blocker–nonfat dry milk, Immun-Blot polyvinylidene difluoride membrane for protein blotting (0.2 μm) were all obtained from Bio-Rad Laboratories (Hercules, CA). Finally, the ECL plus Western blotting detection reagents were obtained from Amersham Biosciences Co (Piscataway, NJ). Four percent paraformaldehyde and propidium iodide solutions were purchased from Sigma-Aldrich. Sixteen to 19 week old fetal kidneys were purchased from Advanced Bioscience Resources, Inc (Alameda, CA).
Isolation of primary cells from fetal kidneys
For the isolation of primary cells from fetal kidneys, an explant fetal tissue culture model described by Snyder et al was used. Briefly, fetal kidneys were rinsed once with medium, finely chopped into small pieces with a sterile razor blade, and mixed with 3-5 mL of medium by pipetting up and down several times. This minced tissue was then placed into 2 100 mm dishes on sterile lens paper supported by 4 sterile stainless steel grids and maintained in organ culture for 3 days by changing the medium daily. After 3 days, all tissue was pooled aseptically and suspended in prewarmed (37°C) 15-20 mL digestion buffer that consisted of 2.5 mL chicken serum, 1.25 mL of 500 mM N-2-hydroxyethylpipera-zine-N′-2-ethane sulfonic acid buffer, 50 mg collagenase 1A, 50 mg collagenase 1, 2-3 mg deoxyribonuclease 1 diluted to 25 mL with Dulbeccos’s modified Eagle’s medium (DMEM).
Tissue was then triturated continuously with a 10 mL pipette for 15-20 minutes until all clumps were broken. The digestion buffer tissue suspension was immediately centrifuged at 1400 rpm for 5 minutes and the supernatant was removed. The pellet was resuspended into 25 mL DMEM and after mixing plated 2 mL each on 60 mm extracellular matrix–coated dishes. Cells were then incubated overnight at 37°C with 5% CO 2 . After overnight incubation, cells were cleaned up by washing vigorously 3 times with Hanks’ balanced salt solution to remove all cell debris. Fresh medium supplemented with 1 mM dibutyryl cyclic adenosine monophosphate and 5% antibiotic antimycotic solution was added daily for 5 consecutive days to accelerate the differentiation of primary cells.
In vitro cell culture
To confirm findings from the primary cells that were isolated from explant fetal kidney tissue, a commercially available established fetal kidney cell line HEK-293 was obtained from the American Type Culture Collection (Manassas, VA). Exploratory repeated exposure studies could also be carried out with this cell line. The HEK-293 cells were propagated in a DMEM supplied with 2 mM L-glutamine and Earle’s balanced salt solution adjusted to contain 10% fetal bovine serum and 3% antibiotic/antimycotic solution. The primary cells were isolated from the 18 week old fetal kidneys purchased from Advanced Bioscience Resources, Inc. The same medium (DMEM) was used to isolate and grow the primary cells from fetal kidneys.
Drug treatment for both kidney toxicity studies
One-time exposure
First, to evaluate the toxic effect of carboplatin on the fetal kidney cell line HEK-293 and primary cells isolated from human fetal kidneys, 5 million HEK-293 cells were treated with the carboplatin concentrations achieved in the fetal compartment of the placenta perfusion model for 12 hour, 16 hour, 24 hour, and 48 hour time points. Untreated cells without any drug served as the experimental control (baseline/time zero). The same procedure was repeated for primary cells isolated from human fetal kidneys for baseline, 16 hour, 24 hour, and 48 hour time points. These experiments were repeated in duplicate for each time point.
Repeated exposure
To explore the potential toxic effect of repeated exposure of carboplatin on the fetal kidney cell line HEK-293, 1 million HEK-293 cells were treated with 6 different selected concentrations of carboplatin that were within range (375-2250 ng/mL) of concentrations observed in the fetal compartment in the placenta perfusion studies with samples obtained at 12 hour, 24 hour, and 48 hour time points in 3 different groups. The first group was treated only once; the second group was treated twice, once every 3 weeks; and the third group was treated 3 times, once every 3 weeks. In the time intervals in between the drug treatment, the drug was removed from the cells and the cells were allowed to grow by changing the medium 2 times a week. Untreated cells without any drug served as the experimental control for each concentration and each time point. These experiments were repeated in duplicate for each drug concentration and each time point. The experiment was designed to mimic the use of carboplatin cyclic dosing in the clinical setting. These experiments were repeated in duplicate for each time point.
Immunoblotting
At each of the time points, the respective experiments carried out, HEK-293 cells and/or primary cells isolated from fetal kidneys had samples of cells harvested and then protein extracts prepared by lysing cells on ice in 100-200 μL of Nonidet P-40 (NP40) lysis buffer. NP40 lysis buffer was used to extract the total proteins from cells using standard procedure. Pierce micro-BCA protein assay kit (Pierce) was used to determine the protein concentration. For each series of protein determinations, a standard curve was constructed with known concentrations of bovine serum albumin. For direct immunoblotting, 50 μg of protein were run on 10% SDS-polyacrylamide gel electrophoresis (PAGE) gels, transferred to polyvinylidene difluoride membranes, and probed with the appropriate antibodies using the manufacturer’s protocol (Santa Cruz Biotechnology). β-Actin was used as the gel loading control for all immunoblotting experiments.
Flow cytometry
Following the respective time point treatment of HEK-293 cells and primary cells isolated from fetal kidneys with various concentrations of carboplatin, both floating and adherent cells were pooled and washed in ice-cold phosphate-buffered saline (PBS). Cells were then fixed in ice-cold 1% paraformaldehyde in PBS for 30 minutes on ice. After washing the cells with ice-cold PBS, they were fixed in 70% ethanol overnight. After fixing in ethanol, cells were again washed in ice-cold PBS and stained with propidium iodide in the presence of ribonuclease at 37°C for 15 minutes.
Cell cycle distribution was analyzed on 10,000 cells for each experimental condition. Data analysis was performed using flow cytometry and cellular imaging at the Core Facility of M. D. Anderson Cancer Center. Quantitative data from cell cycle analysis were graphed to illustrate the accumulation of the percentage of cells in sub G 0 phase of cell cycle. The same procedure was followed for the flow analysis of cell samples collected from the second repeated exposure exploratory experiment.
Sample size and statistical considerations for fetal toxicology studies
All tissue drug exposure experiments were completed in duplicate. Analysis for immunoblotting for changes in expression of markers of apoptosis and flow cytometry was completed in triplicate. Hence, there were 6 samples per dose level per time point in each experiment. A paired Student t test was used to evaluate for statistical differences, with a value of P < .05 defined as significant difference.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
