Safety and pharmacokinetic analysis of methotrexate administered directly into the fourth ventricle in a piglet model
We have developed a piglet model to assess chemotherapy administration directly into the fourth ventricle as a potential treatment for medulloblastoma and other malignant posterior fossa tumors. The objective of this study was to assess safety and pharmacokinetics after methotrexate infusions into the fourth ventricle. Catheters were inserted into the fourth ventricle and lumbar cistern in five piglets. Two milligrams of Methotrexate (MTX) was infused into the fourth ventricle on five consecutive days. Safety was assessed by neurological examination, 4.7 T MRI, and post-mortem pathological analysis. MTX levels in serum and cerebrospinal fluid (CSF) were measured, and area under the concentration–time curve (AUC) was calculated for CSF samples. No neurological deficits were caused by MTX infusions. One piglet died from complications of anesthesia induction for MRI scanning. MRI scans showed accurate catheter placement without signal changes in the brainstem or cerebellum. One piglet had asymptomatic ventriculomegaly. Pathological analysis demonstrated meningitis and choroid plexitis consisting predominantly of CD-3 positive T-lymphocytes in all piglets and a small focal area of subependymal necrosis in one. In all piglets, mean peak MTX level in fourth ventricular CSF exceeded that in lumbar CSF by greater than five-fold. Serum MTX levels were undetectable or negligible. Statistically significant differences between fourth ventricle and lumbar AUC were detected at peaks (P = 0.01) and at all collection time points (P = 0.01) but not at troughs (P = 0.36). MTX can be infused into the fourth ventricle without clinical or radiographic evidence of damage. An inflammatory response without clinical correlate is observed. Significantly higher peak MTX levels are observed in the fourth ventricle than in the lumbar cistern.
KeywordsChemotherapy Fourth ventricle Intraventricular Local delivery Piglet
We hypothesize that infusion of chemotherapeutic and other agents directly into the fourth ventricle of the brain may offer distinct advantages over current regimens used to treat medulloblastoma and other malignant tumors situated in this location. Based upon this hypothesis, our laboratory has developed a piglet model to test the safety of chemotherapy infusion into the fourth ventricle and to assess potential pharmacokinetic advantages over other means of local drug delivery.
Medulloblastoma typically fills the fourth ventricle, and gross total surgical resection is often not achieved due to adherence of tumor to the floor of the fourth ventricle . Postoperative adjuvant therapy includes radiation therapy, except in infants, and a variety of systemic chemotherapy regimens. Both radiation therapy and systemic chemotherapy are often associated with significant morbidity. When medulloblastoma recurs, tumor within the posterior fossa is often accompanied by leptomeningeal spread via cerebrospinal fluid (CSF) pathways [2, 3, 4]. Thus, adjuvant therapy approaches must address both local disease within the posterior fossa as well as central nervous system (CNS) metastasis.
In an effort to maximize drug concentration within the CSF while minimizing systemic exposure, previous studies for various malignancies, including medulloblastoma, have included intrathecal or intraventricular chemotherapy [5, 6, 7, 8]. Agents in these studies were administered either into the lumbar cistern by repeated lumbar punctures or into the lateral ventricle via a ventricular access device (Ommaya reservoir) . Both of these approaches have distinct disadvantages when compared to direct infusion into the fourth ventricle. Repeated lumbar punctures are painful, often require sedation in children, and are technically challenging in some patients. Catheter placement into the lateral ventricle, while often regarded as a relatively straightforward surgical procedure, can be associated with significant complications. Catheter placement can be technically challenging in patients with small ventricles, and catheter malposition is common [9, 10, 11, 12]. Because the catheter must pass through normal brain parenchyma in order to reach the lateral ventricle, intraparenchymal hemorrhage may occur [9, 11]. Moreover, even when the catheter tip is within the ventricle, some of its holes may be in brain parenchyma, and drug administration can thereby result in leukoencephalopathy [9, 13, 14, 15, 16].
We propose that surgical placement of a catheter into the fourth ventricle attached to a ventricular access device offers several distinct advantages over lumbar or lateral ventricular administration. At the time of surgery for a fourth ventricular malignant tumor, maximal surgical resection would be performed as per usual practice. If a frozen section diagnosis of medulloblastoma or another malignancy is rendered, a catheter would be placed in the fourth ventricle and connected to a ventricular access device. Because the catheter would be placed at the time of surgery for primary or recurrent tumor resection, an additional operation would be avoided. In contrast, lateral ventricle catheter placement requires a second surgery at a separate site. The risk of intraparenchymal brain hemorrhage would be eliminated, because the catheter would not pass through brain parenchyma. Catheter placement under direct vision at the time of surgery would ensure that all catheter holes were within the ventricle, thereby decreasing the risk of treatment-related leukoencephalopathy. Moreover, direct chemotherapy infusion into the fourth ventricle would potentially enable the highest possible drug concentration at the site of the tumor bed as well as regional distribution throughout CSF spaces. As with other local drug delivery approaches, systemic drug exposure would be minimized.
Previous experiments in our laboratory established that etoposide can be administered into the fourth ventricle in piglets without causing recognized neurological deficits or MRI signal changes in the adjacent brainstem or cerebellum [17, 18]. Autopsy examination showed ventriculitis and meningitis which did not have a clinical correlate. In these studies, drug did not distribute evenly throughout CSF spaces, and significantly higher peak CSF levels were observed in the fourth ventricle than in the lumbar cistern.
The objective of the current study was to assess safety and drug distribution after repeated infusion of methotrexate (MTX) into the fourth ventricle using the piglet model developed in our laboratory. MTX was chosen because of extensive previous experience with intraventricular administration of this agent. Moreover, intrathecal MTX is part of a highly successful published regimen for medulloblastoma in infants in which chemotherapy alone, without radiation, achieved survival equivalent to previous regimens which included both chemotherapy and radiation .
Materials and Methods
Experiments were performed on five Yorkshire piglets after approval was obtained from the University of Miami Miller School of Medicine Institutional Animal Care and Use Committee. Piglets weighed between 14.3 and 15.9 kilograms (kg). General anesthesia was achieved using a previously published regimen, and central and arterial lines were placed for infusion of medications, blood pressure monitoring, and blood gas sampling as previously described . Dexamethasone, 0.5 milligrams (mg) per kg IV, was administered prior to surgical incision and twice daily (0.25 mg per kg) for the remainder of the experiment. Closed-tip silicone lumbar drain catheters (Medtronic, product reference number 46419) were surgically placed into the fourth ventricle and lumbar cistern using surgical approaches previously published by our laboratory [17, 18]. Catheters were pre-cut to a length of 23 cm. At this length, dead space in the catheter was approximately 0.15 ml. Fourth ventricle catheters were placed under direct vision to ensure that all catheter holes were within the fourth ventricle. After water-tight closure of both the posterior fossa and lumbar dura as previously described [17, 18], normal saline was infused into both catheters to ensure that no fluid leaked from the dural closure. Catheters were then tunneled through the skin and secured with sutures, and a luerlock connector was placed for subsequent access. Muscle, fascia, and skin were then closed in routine fashion, and piglets were extubated after being allowed to emerge from anesthesia. After extubation, femoral arterial lines were removed, and central venous catheters were maintained for subsequent access.
Postoperative neurological examinations were performed at least once per day until sacrifice. Piglets were monitored for general level of alertness, pain, and signs of infection such as meningismus or wound erythema. Motor function and coordination were assessed by observing gait and movement of each forelimb and hindlimb. Sensory examination was limited to assessing withdrawal of each limb to non-painful stimulus. A limited cranial nerve examination included assessment of response to visual and auditory threat, corneal reflex testing (by dripping water into the eye), and observation of mouth movement and feeding. Eating and drinking patterns were monitored.
MTX Infusions and CSF and Serum collection
Preservative-free MTX solution (Intas Pharmaceuticals Ltd.) was diluted in sterile preservative-free normal saline so that each infusion contained 2.0 mg of MTX in 2.0 ml of total volume. Immediately after each infusion, 0.25 ml of normal saline was infused into the catheter to ensure that all drug was flushed out of the tubing and into the fourth ventricle.
MTX infusions into the fourth ventricle were performed once per day for five consecutive days beginning at least 2 days after surgery. Piglets were briefly sedated with intravenous propofol for chemotherapy infusions and during collection of serum and CSF samples. CSF samples were obtained simultaneously from the fourth ventricle and lumbar catheters at 15 min and then 1, 2, 4, 8, 12, and 24 h after the first MTX infusion. For the next 4 days, CSF samples were then obtained just before and 15 min after each subsequent MTX infusion to monitor trough and peak levels. When sampling CSF from a catheter, 0.2 ml was first aspirated and discarded to ensure that fluid analyzed was from within the ventricle rather than the dead space of the tubing. 0.5 ml of CSF was then aspirated and saved in a collection tube. 0.7 ml of sterile, preservative-free normal saline was then infused to flush the catheter and replace the aspirated volume.
Serum samples (3 ml) were obtained 2 and 4 h after intraventricular MTX infusion. CSF and serum samples were collected in heparinized collection tubes, centrifuged at 2000 revolutions per minute (rpm) for 5 min, and then stored at a temperature below −20° Celsius until pharmacokinetic analysis was performed.
CSF samples were also collected for gram stain, cultures, and cell count both at the time of surgery and at the conclusion of the experiment.
Pharmacokinetic Analysis of CSF and Serum Samples
MTX concentrations in CSF and serum samples were measured using high performance liquid chromatography (HPLC). Investigators performing this analysis were blinded regarding the time and site of sample collection.
HPLC methods to examine MTX were a modification of those reported previously [19, 20]. CSF samples were filtered through a 0.22 μm filter, centrifuged at 10,000 rpm for 5 min and injected (50 μl) to HPLC. For plasma, samples (200 μl) were mixed with equal volume (200 μl) 10% perchloric acid. The mixture was vortex-mixed for 2 min and then centrifuged at 10,000 rpm for 10 min to precipitate the proteins. Supernatants were filtered through 0.22 μm centrifuge tube filters, and 50 μl of plasma supernatants were injected. The HPLC system consisted of a solvent delivery system (System Gold in the system control center with gold 126 Solvent Module ternary pump) and UV detector (System Gold Det 168 diode array detector with optical and UV scanning), an Autosampler (System Gold 508, Beckman Coulter Inc. Fullerton, CA, USA). A C18 column (Alltima HP, 150 mm × 4.6 mm, 5 μm, USA) was used for the fractionation. A mobile phase consisted of a mixture of 20% Acetonitrile in 50 mM sodium phosphate buffer (pH 5.4), and the solvent was delivered at a flow rate of 1 ml/min at room temperature. MTX was detected at 303 nm wavelength and the retention time was 5.5 min.
For CSF samples, area under the concentration–time curves (AUC) by trapezoidal rule were calculated for each piglet for peak, trough, and all time points of methotrexate level measurement, respectively. The differences in mean AUCs were calculated by comparing mean AUCs between fourth ventricle and lumbar samples. Statistical significance of differences in mean AUCs was tested by paired t-tests at 5% significance level.
Magnetic Resonance Imaging (MRI) Scans
Each piglet underwent a 4.7 T MRI scan after the completion of all MTX infusions in order to assess exact catheter position and to detect any signal changes in the adjacent brainstem or cerebellum. MRI scans were performed after intubation, general anesthesia administration, and new arterial line placement using a previously described protocol [17, 18]. Sagittal and coronal T2-weighted RARE (Rapid Acquisition Relaxation Enhanced) sequences and coronal T1-weighted Fluid Attenuated Inversion Recovery (FLAIR) sequences were obtained in each piglet.
Tissue Preparation and Histological Analysis
Immediately after MRI scans, piglets were anesthetized with intravenous propofol and then sacrificed with intravenous potassium chloride. Trans-cardiac perfusion/fixation of the brain was performed in situ using 10% buffered formalin as previously described . Brains were extracted and placed in fixative for at least 1 week prior to cutting. Tranverse sections of medulla and pons with cerebellum and the dorsal hippocampi were embedded in paraffin, cut, and stained with hematoxylin and eosin. Sections were analyzed by a neuropathologist (CKP) for disruption of cytoarchitecture, necrosis, inflammation, and any other abnormalities. Immunohistochemistry studies were performed with anti-CD3 + monoclonal antibody to further characterize inflammation.
There were no neurological deficits caused by MTX administration into the fourth ventricle. All piglets had a normal level of alertness, and none exhibited nuchal rigidity, head tilt, lethargy, or any other signs of meningitis. All piglets had normal gait and symmetrical forelimb and hindlimb movement both spontaneously and in response to non-painful stimulus. All piglets responded appropriately to auditory and visual threats by moving rapidly away from the stimulus. All piglets demonstrated normal corneal reflexes, spontaneous mouth movement, and appropriate feeding and drinking patterns throughout the experiment. One piglet (Piglet 4) had normal examinations throughout the experiment, as described above, but died suddenly during induction of anesthesia for MRI scanning at the conclusion of the experiment. In this piglet, intubation for the MRI scan was difficult, and cardiac arrest was noted after multiple unsuccessful attempts at intubation. The dead animal was taken immediately to MRI scanning, but the animal’s brain was not harvested for analysis because trans-cardiac perfusion/fixation could not be performed immediately after animal sacrifice as per our protocol.
Brains of all four piglets which were harvested after trans-cardiac perfusion-fixation (Piglets 1, 2, 3, and 5) appeared normal on gross inspection. In Piglet 1, the animal in which ventriculomegaly was noted on MRI scan, the outlet of the fourth ventricle appeared to be occluded both by the catheter and surrounding surgical debris.
Culture and cell count results from CSF obtained at the time of surgery and after completion of 5 days of intraventricular MTX administration
CSF culture (baseline)
CSF culture (conclusion of experiment)
CSF cell count (baseline)
CSF cell count (conclusion of experiment)
1250 wbc (37% L, 9% M, 54% P)
218 wbc (58% L, 7% M, 34% P)
4039 wbc (40% L, 4% M, 54% P)
157 wbc (62% L, 3% M, 35% P)
4605 wbc (49% L, 9% M, 42% P)
939 wbc (19% L, 33% M, 49% P)
2593 wbc (21% L, 26% M, 53% P)
227 wbc (11% L, 14% M, 75% P)
827 wbc (60% L, 6% M, 34% P)
64 wbc (84% L, 4% M, 12% P)
Serum MTX level two and four hours after infusion into the fourth ventricle
MTX level (2 h)a
MTX level (4 h)a
Sample not obtained
AUC analysis of peak MTX levels demonstrated statistically significant differences between fourth ventricle and lumbar CSF samples (ΔAUC = 37638 micromole h/l with 95% CI: 14216, 61061, P = 0.01). Statistically significant differences between fourth ventricle and lumbar MTX AUC were also noted when all collection time points were assessed (ΔAUC = 18951 micromole h/l with 95%CI: 7031, 30871, P = 0.01). No statistically significant difference between fourth ventricle and lumbar CSF MTX levels was detected when trough AUC was compared (ΔAUC = 176 micromole h/l with 95%CI: −293, 644, P = 0.36).
These experiments are based upon the hypothesis that local delivery of chemotherapeutic and other agents may play a role in future treatment strategies for medulloblastoma and other malignant brain tumors which occur in the fourth ventricle. Despite the fact that survival has improved dramatically over the past few decades, novel approaches are warranted because current treatment regimens are associated with significant morbidity and survival rates for patients with recurrent disease are low. Administration of anti-neoplastic agents directly into the fourth ventricle is a potential means of bypassing the blood–brain barrier to achieve high drug levels at the site of disease as well as in the communicating CSF-filled spaces of the remainder of the brain and spine. Catheter placement within the fourth ventricle at the time of tumor resection would avoid anesthetic and procedural risks, pain and inconvenience associated with repeated lumbar punctures or an additional operation required for lateral ventricle catheter and Ommaya reservoir placement. As with other local delivery strategies, this approach would minimize systemic exposure and associated side effects.
Using a piglet model developed in our laboratory, previous studies using this approach with etoposide were recently published [17, 18]. In these studies, etoposide infusions into the fourth ventricle did not cause any recognized clinical neurological deficits or signal changes in the brainstem or cerebellum on MRI scans. Pathological examination revealed an inflammatory response consisting predominantly of T-lymphocytes which did not have a clinical correlate. Pharmacokinetic analysis showed that etoposide did not distribute evenly throughout CSF spaces, as mean fourth ventricular CSF peak etoposide level exceeded mean peak lumbar etoposide levels by greater than 10-fold. Etoposide was not detected in serum samples obtained 2 and 4 h after infusion.
The current experiments used the same piglet model to test the safety and pharmacokinetics associated with MTX administration into the fourth ventricle. MTX was used in this study because it has been administered for decades via lumbar puncture and lateral ventricle infusion for a variety of malignancies, including those involving the central nervous system. Furthermore, intrathecal MTX was part of a highly successful published regimen for medulloblastoma in infants in which chemotherapy alone achieved survival equivalent to previous regimens which included both chemotherapy and radiation . These experiments were performed in normal piglets in order to demonstrate that MTX can be administered into the fourth ventricle without causing injury to a normal adjacent brainstem and cerebellum.
Repeated MTX infusions directly into the fourth ventricle performed daily over a period of five consecutive days were not associated with any recognized clinical neurological deficits. In the one piglet which died, the death was clearly due to complications of anesthesia induction and intubation at the conclusion of the experiment for MRI scanning, as the animal was perfectly normal throughout the study until this point. In all piglets, MRI scans showed the catheter within the fourth ventricle without any evidence of damage to the brainstem or cerebellum. One piglet had ventriculomegaly on MRI which was likely related to occlusion of the fourth ventricular outlet by the implanted catheter and surrounding surgical debris. The animal was asymptomatic and had normal measured intracranial pressure. Hydrocephalus caused by catheter obstruction is much more likely to occur in piglets than in humans because the outlet of the fourth ventricle is much smaller in piglets than in humans and therefore more easily occluded by an implanted catheter.
The main pathological finding after catheter placement and MTX infusions into the fourth ventricle was an inflammatory response consisting of meningitis and choroid plexitis. Subjectively, this inflammation was less severe than the response observed in previous experiments using etoposide. While the decreased severity of the inflammatory response may be due to a less profound reaction to MTX than etoposide, it is also possible that the administration of dexamethasone during the current experiments dampened the inflammatory response. Dexamethasone was not administered during the previous experiments with etoposide.
It was notable that, along with the inflammatory response pathologically, pleocytosis was observed in CSF samples collected at the conclusion of the experiment in each piglet. Pleocytosis was also observed in CSF collected at the time of initial surgery in all animals, but a high red blood cell count in each animal suggests that the initial cell counts are likely a reflection of blood products from the surgery. Only one piglet (Piglet 2) had a positive CSF culture, and this piglet had fewer white cells and an improved red blood cell to white blood cell ratio at the conclusion of the experiment than at the time of initial CSF collection. The positive culture in Piglet 2, as well as the pleocytosis in all piglets, may possibly be attributed to contamination from the many time points at which CSF was accessed from externalized ports despite attempts to maintain sterile technique. Because the number of white blood cells was decreased in every piglet at the conclusion of the experiment compared to before MTX infusions, an inflammatory reaction to the MTX is unlikely. Most importantly, because none of the piglets demonstrated any clinical signs of meningitis, neurological deficits, or signal changes on MRI scan, and the inflammatory reaction was not severe, the significance of both the CSF pleocytosis and the inflammatory changes observed on pathological analysis is uncertain.
Other pathological findings included focal necrosis observed in a limited area at the ependymal surface in one piglet and axonal retraction balls close to the ependymal surface in three piglets. The area of focal necrosis was relatively small and may be due to direct trauma from surgery or catheter placement rather than from MTX infusion, particularly because this was only observed in one small region, the remainder of the ependymal surface anatomy was preserved, and necrosis was not observed in the other piglets. Axonal retraction balls are a non-specific finding typically observed after an axon has been severed, as is commonly seen after traumatic brain injury . Because they were observed close to the ependymal surface and in areas without any other evidence of injury, the significance of the axonal retraction balls in these experiments is uncertain.
MTX levels were determined both from serum as well as CSF samples obtained from the fourth ventricle and lumbar cistern. Ideally, CSF would have been sampled from the lateral ventricle as well to evaluate distribution throughout the ventricles, but the small size of the lateral ventricle in piglets renders collection of adequate CSF for analysis not feasible with this model. We found that serum MTX levels obtained 2 and 4 h after infusion into the fourth ventricle were either so low that they were not measurable using standard HPLC techniques or were extremely low compared to CSF levels. This suggests that systemic toxicity associated with MTX infusions into the fourth ventricle would likely be absent or minimal.
This study is the first publication investigating pharmacokinetics after infusion of MTX directly into the fourth ventricle in either animals or humans. In CSF samples from the fourth ventricle, MTX levels peaked immediately after infusion and then gradually declined. Lumbar CSF MTX levels started lower than fourth ventricle levels and then gradually increased. By between 1 and 4 h after infusion, MTX levels sampled from CSF in the fourth ventricle and lumbar cistern were similar, suggesting that this treatment approach would provide both local and regional therapy if utilized prior to the presence of bulky leptomeningeal disease obstructing CSF flow. Mean peak fourth ventricle CSF MTX levels exceeded lumbar levels by greater than five-fold, and AUC analysis confirmed statistically significant higher drug levels at peak measurements in the fourth ventricle. At all time points, mean CSF MTX level in both fourth ventricle and lumbar CSF samples far exceeded the “therapeutic concentration” required for cell killing in previous studies using leukemia cells, 10−6 M . To date, despite clinical use of intrathecal MTX in medulloblastoma regimens , the therapeutic concentration required for medulloblastoma cell killing in vitro has not been reported in the medical literature.
Intuitively, it may be assumed that regardless of whether an agent is administered in the lateral ventricle, fourth ventricle, or lumbar cistern, the agent will rapidly equilibrate in all of these spaces as CSF circulates. However, previous studies by Shapiro et al. demonstrate that MTX distribution in different CSF-filled spaces may vary significantly depending upon the site of administration . In these studies, administration into the lateral ventricle in humans yielded higher and more consistent drug levels in the lateral ventricle than administration in the lumbar cistern. Similarly, Balis et al. infused MTX into the lateral ventricles in nonhuman primates and measured MTX levels in the fourth ventricle and lumbar cistern . They noted that lumbar MTX levels were significantly lower and more variable than fourth ventricle CSF levels after such infusions. While the benefit of providing a substantially higher drug concentration than needed for cytotoxicity is unknown, we hypothesize that aggressive treatment directly at the site of residual tumor immediately after surgery may offer a theoretical advantage as long as there is no associated toxicity.
In conclusion, MTX can be infused directly into the fourth ventricle in piglets without causing any recognized clinical toxicity or MRI evidence of damage to the brainstem or cerebellum. The most significant pathological finding is an inflammatory response which does not have associated clinical findings. Peak CSF MTX levels are significantly higher in the fourth ventricle than in the lumbar cistern, and serum MTX levels are absent or minimal compared to CSF levels. These findings suggest that direct infusion of MTX into the fourth ventricle will enable cytotoxic concentrations at the site of tumor origin without causing neurological deficits or significant systemic toxicity. Because these experiments study fourth ventricular MTX infusions over a short time period with no long-term follow-up, future studies will be required to evaluate long-term toxicity after such infusions. Future experiments will also be performed in primates to examine the effects of this treatment modality in a species with a neurological examination that can be performed in greater detail and which more closely resembles that of humans.
This study was supported by grants from the Miami Children’s Hospital Foundation and the Women’s Cancer Association of the University of Miami. Catheters used in these experiments were obtained through a grant from Medtronic. The authors thank Ms. Mariana Nunez for her assistance with preparation of pathological specimens.
- 12.Shapiro WR et al (1977) Treatment of meningeal neoplasms. Cancer Treat Rep 61(4):733–743Google Scholar
- 17.Sandberg DI et al (2009) Pharmacokinetic analysis of etoposide distribution after administration directly into the fourth ventricle in a piglet model. J Neurooncol 38(2–3):213–218Google Scholar