Plasma pharmacokinetics and cerebrospinal fluid penetration of hypericin in nonhuman primates.

UNLABELLED: Hypericin, a polycyclic aromatic dianthroquinone, is a natural pigment derived from the plant Hypericum perforatum (St John's Wort). The compound has been synthesized and shown to inhibit the growth of malignant glioma cell lines in vitro via inhibition of protein kinase C. Oral hypericin has entered clinical trials in adults with recurrent malignant glioma. PURPOSE: The present study was performed to characterize the plasma pharmacokinetics (PK) and cerebrospinal fluid (CSF) penetration of hypericin in nonhuman primates. METHODS: Hypericin was administered as an intravenous bolus dose of 2 mg/kg (n = 3) or 5 mg/kg (n = 1). Plasma and CSF (ventricular or lumbar) were sampled prior to administration and at frequent intervals for up to 50 h after administration of the drug. Hypericin concentrations in plasma and CSF were determined using a specific reverse-phase HPLC assay. RESULTS: Mean peak plasma concentration of hypericin following the 2 mg/kg dose was 142 +/- 45 microM. Elimination of hypericin from plasma was biexponential, with an average alpha half-life of 2.8 +/- 0.3 h and average terminal half-life of 26 +/- 14 h. CONCLUSIONS: The 2 mg/kg dose in the nonhuman primate was sufficient to maintain plasma concentrations above 10 microM (the in vitro concentration required for growth inhibition of human glioma cell lines) for up to 12 h. No hypericin was detected in the CSF of any animal (lower limit of detection 0.1 microM); the CSF penetration is therefore less than 1%. A severe dose-limiting photosensitivity skin rash was seen at the 5 mg/kg dose level.

Plasma and cerebrospinal fluid pharmacokinetics of O6-benzylguanine and analogues in nonhuman primates.

O6-Benzylguanine (BG) is a potent, specific inactivator of the DNA repair protein, O6-alkylguanine-DNA alkyltransferase, that enhances the sensitivity of tumor cell lines and tumor xenografts to chloroethylnitrosoureas. To search for BG analogues with greater penetration into the cerebrospinal fluid (CSF), we evaluated plasma and CSF pharmacokinetics of BG, 8-aza-O6-benzylguanine (8-azaBG), O6-benzyl-8-bromoguanine (8-BrBG), O6-benzyl-8-oxoguanine (8-oxoBG), O6-benzyl-8-trifluoromethylguanine (8-tfmBG), and O6-benzyl-2'-deoxyguanosine (B2dG) after i.v. administration of 200 mg/m2 of drug through an indwelling Ommaya reservoir in a nonhuman primate model. BG and its analogues were quantified in plasma and CSF using reverse-phase high-performance liquid chromatography assays. The plasma clearances of the four 8-substituted BG analogues were similar (0.04-0.06 l/h/kg), but half-lives ranged from <2 to >24 h. BG was converted to 8-oxoBG, an equally potent O6-alkylguanine-DNA alkyltransferase inactivator, and the elimination of 8-oxoBG was much slower than that of BG. As a result, the plasma area under the curve of 8-oxoBG was 3.5-fold greater than that of BG. B2dG was metabolized to BG and 8-oxoBG, but this pathway accounted for only 20% of B2dG elimination. The CSF penetration percentages (based on the ratio of AUC(CSF): AUCplasma) for BG, 8-azaBG, 8-oxoBG, 8-tfmBG, 8-BrBG, and B2dG were 3.2, 0.18, 4.1, 1.4, <0.3, and 2.0%, respectively. The CSF penetration of BG and its active metabolite 8-oxoBG is greater than the penetration of 8-azaBG, 8-BrBG, 8-tfmBG, and B2dG.

Methotrexate distribution within the subarachnoid space after intraventricular and intravenous administration.

PURPOSE: Intrathecal methotrexate achieves high concentrations in cerebrospinal fluid (CSF), but drug distribution throughout the subarachnoid space after an intralumbar dose is limited. The objective of this study was to quantify methotrexate distribution in CSF after intraventricular and intravenous administration and to identify factors that influence CSF distribution. METHODS: Nonhuman primates (Macaca mulatta) with permanently implanted catheters in the lateral and fourth ventricles received methotrexate by bolus injection (0.5 mg) and infusion (0.05 to 0.5 mg/day over 24 to 168 h) into the lateral ventricle, as well as intravenous infusions. CSF was sampled from the lumbar space, fourth ventricle and the subarachnoid space at the vertex. Methotrexate in CSF and plasma was measured with the dihydrofolate reductase inhibition assay. RESULTS: After bolus intraventricular injection, methotrexate exposure in lumbar CSF ranged from 11% to 69% of that achieved in the fourth ventricle. During continuous intraventricular infusions, methotrexate steady-state concentrations (C(ss)) in lumbar CSF and CSF from the vertex were only 20% to 25% of the ventricular CSF C(ss). The dose, duration of infusion, and infusate volume did not influence drug distribution to the lumbar CSF, but probenicid increased the lumbar to ventricular C(ss) ratio, suggesting the involvement of a probenicid-sensitive transport pump in the efflux of MTX from the CSF. During the intravenous infusions, the ventricular methotrexate C(ss) was lower than the lumbar C(ss) and the C(ss) in CSF from the vertex. CONCLUSION: Methotrexate CSF distribution after intraventricular injection was uneven, and at steady-state CSF methotrexate concentrations were lower at sites that were more distant from the injection site.

Pharmacokinetics of O6-benzylguanine and its active metabolite 8-oxo-O6-benzylguanine in plasma and cerebrospinal fluid after intrathecal administration of O6-benzylguanine in the nonhuman primate.

O6-Benzylguanine (O6BG) irreversibly inactivates the single-turnover DNA repair protein alkylguanine-alkyltransferase. Thus, O6BG increases tumor-cell sensitivity to alkylating agents such as carmustine, lomustine, procarbazine, and temozolomide. We investigated the pharmacokinetic behavior of O6BG and O6-benzyl-8-oxoguanine (8-oxo-O6BG) in cerebrospinal fluid (CSF) and plasma after intraventricular administration of O6BG in a nonhuman primate model. In our study, three animals received a single 1-mg dose of O6BG into the lateral ventricle. CSF from the 4th ventricle and plasma samples were collected after administration, and O6BG and 8-oxo-O6BG concentrations were measured by high-performance liquid chromatography. Four additional animals received 1 mg of O6BG via the intralumbar route weekly for 6 weeks to assess the feasibility and toxicity of this route of administration. The peak O6BG CSF concentration was 412+/-86 microM, the t1/2 was 0.52+/-0.02 h, the clearance was 0.22+/-0.01 ml/min, and the area under the concentration-time curve was 319+/-15 microM x h in 4th ventricular CSF. The peak CSF concentration of 8-oxo-O6BG in CSF was 1.9+/-0.4 microM, the t1/2 was 0.76+/-0.03 h, and the area under the concentration-time curve was 5.0+/-1.1 microM x h. Both O6BG and 8-oxo-O6BG were detected in the plasma 0.5-3 h after intraventricular O6BG administration. The plasma peak concentration of O6BG was 0.4 microM at 30 min, and the concentration was <0.1 microM by 3 h. The plasma concentration of 8-oxo-O6BG was 0.2 microM at 30 min and 0.6 microM at 3 h. The animals tolerated the single intraventricular dose and 6 weekly intralumbar doses of O6BG without toxicity. We concluded that intrathecal administration of O6BG is well tolerated in the nonhuman primate and seems to have a substantial pharmacokinetic advantage over systemic administration for meningeal tumors.

Pharmacokinetics of the protease inhibitor KNI-272 in plasma and cerebrospinal fluid in nonhuman primates after intravenous dosing and in human immunodeficiency virus-infected children after intravenous and oral dosing.

KNI-272 is a human immunodeficiency virus (HIV) protease inhibitor with potent activity in vitro. We studied the pharmacokinetics of KNI-272 in the plasma and cerebrospinal fluid (CSF) of a nonhuman primate model and after intravenous and oral administration to children with HIV infection. Plasma and CSF were sampled over 24 h after the administration of an intravenous dose of 50 mg of KNI-272 per kg of body weight (approximately 1,000 mg/m2) to three nonhuman primates. The pharmacokinetics of KNI-272 were also studied in 18 children (9 males and 9 females; median age, 9.4 years) enrolled in a phase I trial of four dose levels of KNI-272 (100, 200, 330, and 500 mg/m2 per dose given four times daily). The plasma concentration-time profile of KNI-272 in the nonhuman primate model was characterized by considerable interanimal variability and rapid elimination (clearance, 2.5 liters/h/kg; terminal half-life, 0.54 h). The level of drug exposure achieved in CSF, as measured by the area under the KNI-272 concentration-time curve, was only 1% of that achieved in plasma. The pharmacokinetics of KNI-272 in children were characterized by rapid elimination (clearance, 276 ml/min/m2; terminal half-life, 0.44 h), limited (12%) and apparently saturable bioavailability, and limited distribution (volume of distribution at steady state, 0.11 liter/kg). The concentrations in plasma were maintained above a concentration that is active in vitro for less than half of the 6-h dosing interval. There was no significant increase in CD4 cell counts or decrease in p24 antigen or HIV RNA levels. The pharmacokinetic profile of KNI-272 may limit the drug's efficacy in vivo. It appears that KNI-272 will play a limited role in the treatment of HIV-infected children.

Effect of body position on ventricular CSF methotrexate concentration following intralumbar administration.

PURPOSE: Intralumbar methotrexate is one of the primary therapeutic modalities for the prevention and treatment of meningeal leukemia. However, methotrexate distribution to the ventricles is limited and highly variable following intralumbar dosing, and cytotoxic concentrations of methotrexate are not always achieved or sustained in the ventricular CSF. We used a nonhuman primate model to determine the effect of body position on the caudal distribution of an intralumbar dose of methotrexate. METHODS: Methotrexate (1.0 mg) was administered by intralumbar injection to four animals, which were then immediately placed either in an upright sitting position or in a prone position for 1 hour, then upright. Each animal served as its own control and was studied in each position on at least one occasion. RESULTS: The mean peak ventricular methotrexate concentration was 0.12 mumol/L (range, 0.091 to 0.20) in animals that were immediately placed upright, compared with 2.81 mumol/L (range, 0.21 to 8.9) in animals that remained prone for 1 hour. The mean area under the concentration-versus-time curves (AUC) was 0.51 mumol/L.h (range, 0.26 to 1.1) in the upright animals and 12.0 mumol/L.h (range, 0.9 to 35.4) in the prone animals. CONCLUSION: Maintaining a prone position for 1 hour after an intralumbar dose increased the peak methotrexate concentration and drug exposure in ventricular CSF. CSF drug distribution following intralumbar therapy can be influenced by body position after the injection.

Cerebral subarachnoid sampling of cerebrospinal fluid in the rhesus monkey.

A previously described rhesus monkey model with two intraventricular catheter systems was expanded to provide a means of studying drug concentrations of chemotherapeutic agents such as methotrexate (MTX) in the cerebral subarachnoid cerebrospinal fluid (CSF) following intrathecal drug injection. A continuous intraventricular infusion of MTX was started through the lateral ventricular catheter 44 h before surgical placement of the cerebral subarachnoid catheter to allow for steady-state concentrations to be attained throughout the subarachnoid space. Monkeys were anesthetized intramuscularly with ketamine hydrochloride (7 mg/kg) and xylazine (6 mg/kg) to allow placement of a temporary lumbar catheter for sampling of lumbar CSF Sodium thiopental (2.5%) was given intravenously if needed for intubation and anesthesia was maintained with isoflurane (0.5 to 1.5%) and oxygen during the surgical placement and sampling of the cerebral subarachnoid catheter. A midline incision was made over the frontal bone and a 3/8-inch trephine was used to expose the dura adjacent to the lateral midline. Following puncture of the dura with an 18-gauge spinal needle, the cerebral subarachnoid catheter (0.025-inch [i.d.] x 0.047-inch [o.d.] Silastic tube) was passed into the subarachnoid space for approximately 10 to 17 mm. The CSF from the cerebral subarachnoid catheter was collected by gravity flow Ventricular, lumbar subarachnoid, and cerebral subarachnoid CSF were collected concurrently at 44, 46, and 48 h after the start of MTX infusion. Mean ventricular, lumbar subarachnoid, and cerebral subarachnoid CSF methotrexate concentrations at steady state were 5.8, 1.0, and 1.5 mumol/L, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

Pharmacokinetics of PEG-L-asparaginase and plasma and cerebrospinal fluid L-asparagine concentrations in the rhesus monkey.

The pharmacokinetics of the polyethylene glycol-conjugated form of the enzyme L-asparaginase and the depletion of L-asparagine from the plasma and cerebrospinal fluid (CSF) following an i.m. dose of 2500 IU/m2 PEG-L-asparaginase was studied in rhesus monkeys. PEG-L-asparaginase activity in plasma was detectable by 1 h after injection and maintained a plateau of approximately 4 IU/ml for more than 5 days. Subsequent elimination from plasma was monoexponential with a half-life of 6 +/- 1 days. Plasma L-asparagine concentrations fell from pretreatment levels of 14-47 microM to < 2 microM by 24 h after injection in all animals and remained undetectable for the duration of the 25-day observation period in four of six animals. In two animals, plasma L-asparagine became detectable when the PEG-L-asparaginase plasma concentration dropped below 0.1 IU/ml. Pretreatment CSF L-asparagine levels ranged from 4.7 to 13.6 microM and fell to < 0.25 microM by 48 h in five of six animals. CSF L-asparagine concentrations remained below 0.25 microM for 10-14 days in four animals. One animal had detectable CSF L-asparagine concentrations within 24 h and another had detectable concentrations within 1 week of drug administration despite a plasma PEG-L-asparaginase activity profile that did not differ from that of the other animals. These observations may be useful in the design of clinical trials with PEG-L-asparaginase in which correlations among PEG-L-asparaginase pharmacokinetics, depletion of L-asparagine, and clinical outcome should be sought.

Methotrexate pharmacokinetics following administration of recombinant carboxypeptidase-G2 in rhesus monkeys.

PURPOSE: Carboxypeptidase-G2 (CPDG2) is a bacterial enzyme that rapidly hydrolyzes methotrexate (MTX) into inactive metabolites. As an alternative form of rescue after high-dose MTX (HDMTX), CPDG2 has more potential advantages than standard leucovorin (LV) rescue. In this study, the plasma pharmacokinetics of MTX with and without CPDG2 were evaluated in adult rhesus monkeys. MATERIALS AND METHODS: The plasma pharmacokinetics of MTX were determined in groups of animals that had received a 300-mg/m2 loading dose of MTX followed by a 60-mg/m2/h infusion during an 18-hour period. One group received CPDG2 at the end of the infusion, and the other group served as a control. Two additional animals with high titers of anti-CPDG2 antibody also were studied. RESULTS: During infusion, the steady-state MTX plasma concentration was 11.3 +/- 4.8 mumol/L. Without CPDG2, the postinfusion plasma MTX concentration remained above 0.1 mumol/L for more than 6 hours. After the administration of 50 U/kg of CPDG2, plasma MTX concentrations decreased to nontoxic levels (less than 0.05 mumol/L) within 30 minutes. The initial half-life (t1/2 alpha) of MTX decreased from 5.8 +/- 2.1 minutes to 0.7 +/- 0.02 minutes after enzyme administration. The postinfusion area under the plasma concentration time curve of MTX was 301 +/- 171 mumol/L/min without CPDG2 compared with 19.6 +/- 6.1 mumol/L/min with CPDG2. The immunogenicity studies performed indicated that although animals developed anti-CPDG2 antibodies, none of them manifested allergic symptoms. The effectiveness of CPDG2 was diminished but not eliminated in animals with high titers of anti-CPDG2 antibody. CONCLUSIONS: CPDG2 is capable of rapidly decreasing plasma MTX concentrations to nontoxic levels. The administration of CPDG2 seems safe, well tolerated, and it may be useful as an alternative to LV rescue.

Intrathecal 5-fluorouracil in the rhesus monkey.

Because meningeal spread of both leukemia and solid tumors remains a difficult therapeutic problem, there is a compelling need to develop new agents for intrathecal administration. 5-Fluorouracil (5FU), an active anticancer agent, penetrates into the central nervous system to some degree following intravenous dosing. Significant systemic toxicity, however, is associated with this route of administration. Therefore, the pharmacokinetic behavior of 5FU following its intrathecal administration was studied in a rhesus monkey model. After a 10-mg intraventricular dose, the disappearance of the drug from ventricular cerebrospinal fluid was monoexponential, the half-life being 51 min; the area under the concentration-time curve (AUC) being greater than 18 mM h-1; and the peak ventricular 5FU concentrations ranging between 10 and 15 mM. After a 1-mg intralumbar dose, the AUC was 1235 microM h-1. No toxicity was observed following intraventricular administration of 5FU. After intralumbar administration of either a 10-mg or a 1-mg dose, however, local toxicity was observed in the lumbar spinal cord. These findings suggest that intrathecal administration of 5FU is not presently a feasible means of achieving cytotoxic cerebrospinal fluid concentrations.

Pharmacokinetics of pyrazoloacridine in the rhesus monkey.

Pyrazoloacridine is a rationally synthesized acridine derivative with in vitro activity against solid tumor cell lines, noncycling and hypoxic cells, and tumor cell lines that exhibit the multidrug resistance phenotype. The pharmacokinetic behavior of pyrazoloacridine after a 1- or 24-h i.v. infusion was studied in 5 rhesus monkeys that received a total of 10 courses of pyrazoloacridine at 300 or 600 mg/m2. Pyrazoloacridine levels in plasma and cerebrospinal fluid were measured by high-pressure liquid chromatography. For 1-h infusions, the plasma disappearance was biexponential with a t 1/2 alpha of 31 min and t 1/2 beta of 11 h. The mean volume of distribution at steady state was 1380 liters/m2. The clearance was 1660 ml/min/m2. For the 300 mg/m2 dose, the mean area under the concentration-time curve was 759 microM.min, and the mean peak concentration was 1.3 microM. For the 600 mg/m2 dose, the area under the concentration-time curve was 1330 microM.min, and the peak concentration was 2.5 microM. The steady-state plasma concentrations during the 24-h continuous infusions were 0.27 microM for the 300 mg/m2 dose and 0.45 microM for the 600 mg/m2 dose. The mean clearance calculated from these steady-state concentrations was 2420 ml/min/m2. Cerebrospinal fluid levels were less than 0.1 microM for all doses and schedules. There was no evidence of toxicity at any dose or schedule. These results contrast strikingly with those obtained in mice and dogs in which, despite a more rapid clearance of pyrazoloacridine, significant toxicities were observed at doses that were nontoxic in the monkey. These interspecies differences in the pharmacokinetic and pharmacodynamic behavior of pyrazoloacridine have important implications for the design of Phase I trials in humans.

Rescue of experimental intrathecal methotrexate overdose with carboxypeptidase-G2.

The carboxypeptidase G class of enzymes rapidly hydrolyze methotrexate (MTX) into the inactive metabolites 4-deoxy-4-amino-N10-methylpteroic acid (DAMPA) and glutamate. This study evaluated the use of carboxypeptidase-G2 (CPDG2) as a potential intrathecal (IT) rescue agent for massive IT MTX overdose. The CSF pharmacokinetics of MTX with and without CPDG2 rescue was studied in adult rhesus monkeys (Macaca mulatta) using a nontoxic IT 5 mg dose (equivalent to 50 mg in humans). Without CPDG2 rescue, peak CSF MTX concentration was 2,904 +/- 340 mumol/L. Within 5 minutes of administration of 30 U IT CPDG2, CSF MTX concentrations decreased greater than 400-fold to 6.55 +/- 6.7 microM. Subsequently, groups of three monkeys received either 25 mg IT MTX (equivalent to 250 mg in humans) followed by 150 U IT CPDG2 or 50 mg IT MTX (equivalent to 500 mg in humans) followed by 300 U IT CPDG2. All animals survived without neurotoxicity. Our studies suggest that CPDG2 may prove to be an important addition to the currently recommended strategy for the management of IT MTX overdose.

A rhesus monkey model for continuous infusion of drugs into cerebrospinal fluid.

A new rhesus monkey model with two intraventricular catheter systems was developed to examine the pharmacokinetics and neurotoxicity of chemotherapeutic agents administered by continuous intraventricular infusion. A lateral ventricular catheter system implanted in the lateral ventricle and attached to a subcutaneous access port on the animal's back is used for infusion of drugs into the ventricle. A Pudenz catheter implanted in the fourth ventricle and connected to a subcutaneous Ommaya reservoir permits repetitive CSF sampling in unanesthetized animals. The model was evaluated in five animals for over 12 months for catheter patency, surgical complications, and utility in studying the pharmacokinetics of continuous intraventricular infusion of methotrexate. There were no perioperative complications. Three of the five monkeys maintained both systems successfully. The other two animals developed staphylococcal ventriculitis, one at 7 days as a result of manipulation of the incision by the animal leading to cellulitis around the catheter site and subsequent ventriculitis, the other at 5 months. Both animals were treated successfully with antibiotics and catheter removal. An infusion of 0.05 mg of methotrexate over 24 hours maintained ventricular drug concentrations of 1 mol/L without evidence of neurotoxicity. This new model has applications both for the development of continuous intraventricular infusion as a therapeutic approach for the treatment of meningeal cancers in humans and as a research tool to study the distribution and elimination of drugs from the CSF.

Bioavailability of low-dose vs high-dose 6-mercaptopurine.

The bioavailability of oral 6-mercaptopurine (6MP) at standard doses is very low, largely as a result of extensive first-pass metabolism by xanthine oxidase. Fewer than one third of patients achieve 6MP plasma concentrations known to be cytocidal in vitro (greater than 1 mumol/L). Studies in vitro have suggested that first-pass metabolism can be saturated at higher doses of 6MP. To determine whether saturation occurs in vivo at clinically used doses and whether bioavailability can be enhanced by increasing the dose, the bioavailability of different doses of 6MP was studied first in rhesus monkeys and then in children with acute lymphoblastic leukemia in remission. In monkeys a higher dose of 6MP resulted in enhanced bioavailability, whereas in patients the mean relative bioavailability at the higher dose was significantly less. However, all patients achieved cytocidal (greater than 1 to 10 mumol/L) plasma concentrations at the higher dose without manifesting significant clinical toxicity. Therefore cytocidal levels of 6MP can be achieved in patients with oral 6MP without the risk of unexpectedly high levels caused by saturation of first-pass metabolism.

Pyrimidine dideoxyribonucleosides: selectivity of penetration into cerebrospinal fluid.

Cerebrospinal fluid (CSF)/plasma ratios of 1 to 30% were obtained in rhesus monkeys for the 3'-azido- and 2',3'-dideoxy-analogs of thymidine, deoxycytidine and deoxyuridine. Penetration of thymidine and deoxyuridine analogs was much greater than for deoxycytidine analogs. Octanol/buffer partition coefficients varied more than 30-fold, but did not correlate with CSF entry. Plasma protein binding was insignificant for all compounds. The presence or absence of the azido group at position 3' did not appear to influence the extent of CSF penetration. Although we do not fully understand the mechanistic basis for the penetration of these nucleosides into the CSF, it is apparent that the structural specificity is related more closely to the nucleobase than the sugar. Based upon elimination rates from the CSF after direct intrathecal injection, the differences in net penetration are determined by influx rather than efflux processes.

Cerebrospinal fluid penetration of active metabolites of cyclophosphamide and ifosfamide in rhesus monkeys.

The penetration of the active metabolites of cyclophosphamide (CP) and ifosfamide (IF) into cerebrospinal fluid (CSF) was determined in rhesus monkeys following an i.v. infusion of 1 gm/m2 of CP and IF. Active metabolites were measured using a high-performance liquid chromatography assay with fluorometric detection following derivatization with m-aminophenol. CSF to blood ratios of the active metabolites of CP and IF were found to be 0.17 and 0.13 following systemic dosing of CP and IF, respectively. The levels achieved in the CSF, however, were equivalent to levels known to be cytocidal to malignant cell lines derived from tumors which metastasize to the central nervous system. Only one animal demonstrated neurotoxicity with IF. CSF levels of active metabolite in this animal were similar to those observed in the other animals.