Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Amifostine (an organic thiophosphate ester prodrug) is used clinically for its ability to reduce the renal toxicity associated with cisplatin therapy in patients with ovarian cancer or nonsmall cell lung cancer, and to reduce the incidence of xerostomia in patients undergoing radiation treatment for head and neck cancer (Physician's Desk Reference, 2002). To be active, amifostine must first be dephosphorylated at the tissue site by membrane-bound alkaline phosphatase to its free thiol metabolite, WR-1065¹ (Spencer and Goa, 1995; Capizzi, 1996; Foster-Nora and Siden, 1997; Culy and Spencer, 2001). WR-1065 is then further metabolized through oxidation to a symmetrical disulfide or to mixed disulfides (protein and nonprotein). The active form, WR-1065, is taken up into cells and provides cytoprotection by scavenging oxygen free radicals, donating hydrogen ions to free radicals, depleting oxygen, and binding to and inactivating cytotoxic drugs. Amifostine selectively protects benign tissues without decreasing the response of tumor tissues toward radiation and/or cytotoxic drugs. This selectivity is attributed to the higher capillary alkaline phosphatase activity, higher pH, and better vascularity of normal tissues relative to the tumor tissue. As a result, there is a preferential conversion and uptake of WR-1065 in normal cells.

Given the role of tissue activation in the effective and selective protection of normal tissue, it is important to understand the regional pharmacokinetics of amifostine in specific organs. Unfortunately, the sites and extent of this organ-specific activation are not known. Moreover, studies on the kinetics of amifostine (and WR-1065) are few and incomplete (Spencer and Goa, 1995; Shaw et al., 1996; Culy and Spencer, 2001). In particular, some clinical and preclinical data suggest that amifostine exhibits nonlinear kinetics consistent with saturable metabolism (Mangold et al., 1989; Shaw et al., 1994a, 1999). However, the precise mechanism responsible for a concentration-dependent metabolism has not been elucidated. Thus, a more detailed and precise assessment of the processing of amifostine by major organ systems could provide a better pharmacokinetic model on which to base new applications of amifostine, and a fuller exploration of the potential therapeutic value of this cytoprotector.

With this in mind, the primary objective of the study was to examine the in vivo extraction and clearance of amifostine in tissues that are abundant in alkaline phosphatase (e.g., liver, kidney, intestine and lung). A secondary objective of the study was to probe the potential for saturable kinetics in these organs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Amifostine and WR-1065 were obtained from MedImmune, Inc. (Gaithersburg, MD). The aminothiol analogs (WR-80855 and WR-251833), used as internal standards, were a generous gift from the Walter Reed Army Institute for Research (Washington, DC). Amifostine aqueous solutions were prepared immediately before use by dissolving the powder in saline solution. All other chemical and solvents were reagent grade or better.

Experimental Methods. A total of 10 mixed breed male and female dogs (20.9 ± 3.3 kg) were used in these acute pharmacokinetic studies. The surgical design was adapted from studies performed previously by our group in dogs (Kuan et al., 1996, 1998), with minor modifications. Briefly, after an overnight fast, each dog was administered a preanesthetic mixture consisting of acepromazine maleate (1.1 mg/kg i.m.) and atropine sulfate (0.04 mg/kg i.m.). Anesthesia was then induced with sodium pentobarbital (35 mg/kg i.v.). Supplemental doses of the anesthetic agent were provided on an as-needed basis, and respiration was maintained on a volume-controlled ventilator (Harvard Apparatus, South Natick, MA). In five dogs, catheters for serial blood sampling were placed by laparotomy into the portal and hepatic veins, by thoracotomy into the pulmonary artery, and by cutdown into the carotid artery. Perivascular ultrasonic transit time flow probes (Transonic Systems, Inc., Ithaca, NY) were placed for blood flow measurements around the common hepatic artery, portal vein, and ascending aorta. This allowed regional pharmacokinetic parameters to be determined for the liver, gastrointestinal tract, splanchnic area (i.e., combined liver, gastrointestinal tract, spleen, and pancreas), lungs, and total body. In another five dogs, catheters for serial blood sampling were placed by laparotomy into the hepatic and renal veins, by thoracotomy into the pulmonary artery, and by cutdown into the carotid artery. Perivascular ultrasonic transit time flow probes were placed for blood flow measurements around the common hepatic artery, portal vein, ascending aorta, and renal artery. This allowed regional pharmacokinetic parameters to be determined specifically for the kidneys, along with the splanchnic area, lungs, and total body. For all dogs, the gastrointestinal artery was ligated to eliminate any extrahepatic blood flow from contributing to the measurement of blood flow in the hepatic artery. Catheters and flow probe wires were tunneled subcutaneously to exit the skin, and the abdomen and chest were closed (Fig. 1). In our studies, flow probes were factory tested and precalibrated upon arrival. The probes were also tested/validated before and after experimentation, according to the manufacturer's directions (Transonic Systems, Inc.).

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Schematic of circulation depicting the position of sampling catheters, flow probes, and the drug infusion catheter.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Schematic depicting the experimental design of three sequentially escalated intravenous infusions of amifostine.

Analytical Methods. Amifostine and WR-1065 concentrations were determined using the method of Shaw et al. (1984, 1986a, 1994b) with minor modifications. Analyses were performed using high-performance liquid chromatography coupled to an electrochemical detector, with chemical analogs as the internal standards. Amifostine, WR-1065, and internal standard were detected using a BAS (West Lafayette, IN) LC-4C amperometric detector equipped with a thin film mercury-gold amalgam electrode. The Hg/Au electrode working potential was set at +0.15 V with respect to the Ag/AgCl reference electrode. The column (100 × 3 mm, 3-µm particle size, ODS; BAS) was operated at room temperature. The amifostine chromatography used an isocratic mobile phase consisting of 0.1 M monochloric acetic acid and 3 mM sodium octyl sulfate, pH 3.0, at a flow rate of 1.0 ml/min. The WR-1065 chromatography used an isocratic mobile phase consisting of 0.1 M monochloric acetic acid, 3 mM sodium octyl sulfate, pH 3.0, and 30% methanol at a flow rate of 0.6 ml/min. Peak identification was confirmed by comparing retention times in samples with authentic standards. Quantification was based on the peak area ratio of the compound and the appropriate internal standard (i.e., WR-80855 for amifostine and WR-251833 for the active free thiol). Validation assays were performed for amifostine (0.5, 1, 10, and 40 µM) and WR-1065 (1, 4, 10, and 50 µM), and the interday variability (precision) and accuracy (bias) were less than 12% for all methods.

Pharmacokinetic Calculations. The regional kinetics of amifostine, after extended drug infusions, were based on clearance concepts across an eliminating organ (Gibaldi and Perrier, 1982; Wilkinson, 1987). Thus, the extraction ratios across the liver (E_H), gastrointestinal tract (E_GI), splanchnic region (E_Spl), lungs (E_Lu), and kidneys (E_R) were determined for each infusion as follows:




where C_ave was the average plasma concentration of drug entering the liver. C_HV, C_CA, C_PV, C_PA, and C_RV were the respective plasma concentrations in the hepatic vein, carotid artery, portal vein, pulmonary artery, and renal vein. Because the liver receives drug from two sources, namely, the hepatic artery and portal vein, C_ave was estimated as follows:
in which Q_HA is the hepatic arterial blood flow, Q_PV the portal venous blood flow, and Q_H the total blood flow to liver. The fraction of drug escaping extraction on each pass through the liver (F_H), gastrointestinal tract (F_GI), splanchnic region (F_Spl), lungs (F_Lu), and kidneys (F_R) were calculated as follows:




Blood clearances by the liver (CL_b,H), gastrointestinal tract (CL_b,GI), splanchnic region (CL_b,Spl), lungs (CL_b,Lu), and kidneys (CL_b,R) were then determined as follows:




where Q_CO was the cardiac output and Q_RA was the renal arterial blood flow (one kidney). Finally, the total body plasma clearance of drug (CL_p,TB) was calculated as follows:
in which R_o is the infusion rate of drug in the cephalic vein.

Statistics. Data were reported as mean ± S.D., unless otherwise indicated. To test for parameter differences among treatment groups, an ANOVA was performed with repeated measures. When the F ratio showed that there were significant differences among groups, Tukey's test was used to determine which groups differed ( = 0.05). Means that were significantly different have the same capital letters in Tables 1 to 3. All statistical computations were preformed using SYSTAT version 8.0 (SPSS Inc., Chicago, IL).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The extraction ratios of amifostine across specific organs are shown in Table 1. As observed, hepatic extraction of amifostine remained unchanged at about 90%, whereas gastrointestinal extraction decreased significantly from 43 to 12 to 15% as the dose rates increased. Taken as a whole region, the splanchnic area had a constant extraction ratio for amifostine of about 88%. Pulmonary extraction of amifostine was low at 7%, whereas renal extraction was intermediate at 57%.

Table 2 shows the fraction of amifostine that escaped regional extraction by each organ. In agreement with the extraction data, hepatic and splanchnic availabilities were low but constant, averaging about 10 to 12%. In contrast, the fraction of amifostine surviving elimination by the gastrointestinal tract increased from 57 to 85 to 88% as the dose rate increased. The pulmonary availability remained high at 93% and the fraction of drug surviving renal elimination was 43%.

Blood clearances by the liver, gastrointestinal tract, splanchnic region, lungs, and kidneys, as well as the total body plasma clearance of amifostine are summarized in Table 3. As the dose rate increased, hepatic and splanchnic clearances did not change significantly, maintaining values on the order of 21.6 to 25.6 ml/min/kg. Clearance by the gastrointestinal tract, on the other hand, showed a significant decrease from 9.8 ml/min/kg to 2.8 to 3.3 ml/min/kg at the higher doses. Values for the pulmonary and renal clearance were steady at about 6.1 to 6.2 ml/min/kg. Finally, the total body plasma clearance of amifostine decreased, from 52.6 to about 37.3 ml/min/kg, as the doses increased.

Plasma concentrations of amifostine were measured in those regions that perfuse organs of interest (Table 4). Of particular importance, it seems that carotid arterial concentrations of 15 µM are sufficient to minimize extraction of amifostine by the gastrointestinal tract (15%). As a result, the gastrointestinal tract becomes a less important organ of drug elimination, compared with the other tissues studied, when these concentrations are achieved or surpassed.

Carotid arterial blood samples were analyzed for free WR-1065 but concentrations were below the limit of quantitation (1.0 µM) for all dose rates. However, due to the reversibility of a disulfide bond, the (mixed) disulfides may be regarded as an exchangeable pool of the active free thiol. With this in mind, systemic plasma concentrations of total WR-1065 (i.e., after conversion of all disulfide species), as well as the dose-corrected concentrations, are shown in Table 5. Although speculative, the lack of significant differences among the dose-corrected plasma levels of total WR-1065 suggests that the clearance of WR-1065 and its metabolites may be linear. In addition, based on our inability to detect free WR-1065 in blood and assuming a complete conversion of amifostine to WR-1065, it seems that the total body clearance of free thiol is very high. Unfortunately, the clearance of WR-1065 could not be determined directly because only amifostine was available for administration.

To determine the stability of our experimental system, the hepatic arterial, portal venous, total hepatic, and renal arterial blood flows were measured, as well as the cardiac output, during amifostine infusions (Fig. 3). Blood flow was 6.2 ± 3.2 ml/min/kg in the hepatic artery, 20.2 ± 6.6 ml/min/kg in the portal vein, 26.4 ± 7.7 ml/min/kg in the total liver, 5.5 ± 1.7 ml/min/kg in the renal artery (one kidney), and 112 ± 19 ml/min/kg for cardiac output. These blood flows compared well with values found in the literature (Altman and Dittmer, 1974) except for that of the kidney, which was about 40% lower. The reason for a lower renal blood flow is unclear, although it might represent normal variability among animals or, perhaps, vasoconstriction of the renal artery as a result of contact with the flow probe. Importantly, there were no time-dependent changes (i.e., from 0 to 2, 2 to 4, and 4 to 6 h) in flow for the hepatic artery (P = 0.880), portal vein (P = 0.841), total liver (P = 0.935), renal artery (P = 0.963), or cardiac output (P = 0.371). The systolic blood pressure was also measured during the amifostine infusion studies, as displayed in Fig. 4. The mean blood pressure during the infusion was 122 ± 18 mm Hg. As observed, amifostine infusion did not cause a dose-dependent hypotension under conditions used in these experiments. The volume of normal saline administered during the 10 h of surgical and drug experiments was 4.0 ± 0.5 liters.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Blood flow versus time profiles for hepatic artery (QHA; solid circles), portal vein (QPV; solid squares), total hepatic stream (QH; open diamonds), cardiac output (QCO; solid triangles), and renal artery (QRA; open triangles, one kidney) during three sequentially escalated intravenous infusions of amifostine. Data are reported as mean ± S.E. of 10 dogs (except for QRA, 5 dogs).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Systolic blood pressure versus time profile during three sequentially escalated intravenous infusions of amifostine. Data are reported as mean ± S.E. of 10 dogs.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Amifostine (previously known as WR-2721) has been shown to protect almost all normal tissues from the cytotoxic effects of radiation and some chemotherapeutic agents (Spencer and Goa, 1995; Capizzi, 1996; Foster-Nora and Siden, 1997). To accomplish this, amifostine must be dephosphorylated to the active metabolite WR-1065 at its site of action in tissue and be preferentially taken up by normal as opposed to tumor cells. At present, amifostine is used to protect the kidney against cisplatin toxicity and the parotid gland against radiation treatment for head and neck cancer. However, depending on the outcome of numerous ongoing clinical trials, amifostine may find broader clinical applications, both as a chemo- and radioprotectant (Culy and Spencer, 2001). In this regard, we believe that a systematic examination of the sites and extent of organ-specific activation will facilitate the development of other cancer treatments and novel routes of drug delivery for amifostine.

In the present study, we have demonstrated that amifostine is eliminated more extensively by the liver than by the gastrointestinal tract, lungs, or kidneys. Approximately 90% of the drug was extracted by the liver (i.e., converted to WR-1065), suggesting that the liver may preferentially take up the active free thiol. Amifostine was also shown to have a minimal extraction by the gastrointestinal tract (i.e., less than 15%) at concentrations likely to be achieved during clinical dosing (15 µM). Finally, amifostine exhibited a saturable metabolism in the gastrointestinal tract as well as in the whole body.

Extended infusion studies were specifically designed to probe the extraction, clearance and concentration dependence of amifostine in organs having high levels of alkaline phosphatase (McComb et al., 1979) or thought to be important in drug disposition and exposure. Under these experimental conditions, regional and systemic plasma concentrations of amifostine ranged between 2 and 30 µM, depending on the dose rate. Similar plasma concentrations were observed for total WR-1065, whereas free WR-1065 in blood was undetectable. Thus, despite the application of nonclinical doses (i.e., usually infused as 910 mg/m² over 15 min or 200 mg/m² over 3 min; equivalent to about 8 µmol/min/kg for both regimens), drug levels of amifostine and total WR-1065 in dogs were within the range of plasma concentrations observed for these drug species in cancer patients (Shaw et al., 1986b; Korst et al., 1997). In contrast, after a 740 or 910 mg/m² i.v. dose of amifostine to patients (infused over 15 min), the free WR-1065 levels were reported as being similar to that of amifostine, with area under the plasma concentration-time profiles (AUCs) of both chemical species being about 1/10 that of the disulfides. Although the reason for undetectable blood levels of WR-1065 in our study is not clear, it probably reflects the use of much lower amifostine infusion rates (8- to 64-fold). Still, the effect of animal species and disease differences cannot be ruled out. It is also possible that organ blood flow and organ uptake of drug in an awake, free-moving dog may be markedly different compared with an anesthetized animal. Notwithstanding this uncertainty, maximal plasma concentrations of 16 to 17 µM were reported for amifostine and protein-bound WR-1065 after a 500-mg s.c. dose of drug to healthy subjects (Bonner and Shaw, 2002). Moreover, the AUC of free WR-1065 was less than 12% of the AUC for total WR-1065. These results after s.c. dosing in subjects are strikingly similar to our findings in dogs and may reflect the use of congruous input rates of drug into the systemic circulation (assuming the s.c dose of amifostine is released over 30 min or longer).

Amifostine may exhibit dose-dependent kinetics as suggested by a study (Shaw et al., 1994a) in which cancer patients were administered 15-min i.v. infusions of drug at 910 or 740 mg/m². In the high- versus low-dose groups, mean values were reported for volume of distribution (7.4 versus 8.7 liters), half-life (2.7 versus 1.5 min), and plasma clearance (2.1 versus 4.3 l/min). Thus, a significant increase in clearance was observed when less amifostine was administered intravenously and reduced peak plasma concentrations were obtained (235 versus 100 µM). As also noted, in vitro K_m values for amifostine in human kidney (100 µM) and placenta (80 µM) alkaline phosphatase preparations were well within the range of plasma concentrations achieved during infusions of drug to cancer patients. Preclinical studies also support a concentration- and route-dependent disposition of amifostine (Mangold et al., 1989), in which 150 mg/kg amifostine was administered to monkeys by intravenous and portal venous infusions of 10- and 120-min duration. In comparing equivalent intravenous doses, the AUC of amifostine was 3 times greater during the 10- versus 120-min infusion. After equivalent portal venous doses, the AUC of amifostine was 7 times greater during the 10-min infusion compared with the 120-min infusion. This finding lends strong support to the argument that amifostine exhibits saturable kinetics, particularly after regional dosing to the liver.

In the present study, the gastrointestinal tract displayed a concentration-dependent extraction, first-pass availability and clearance at amifostine plasma concentrations of 15 µM and greater. In contrast, the liver, splanchnic region, lungs, and kidneys did not demonstrate saturability. However, this does not preclude the liver (or other tissues) from exhibiting capacity-limited metabolism when higher concentrations of amifostine are present in the circulating blood. In fact, this is likely to occur during the first hour after clinical doses (or dose rates) of amifostine or, more dramatically, if drug were infused regionally. Given the known clearances of amifostine in the splanchnic region, lungs, and kidneys, one can also estimate the extent of drug elimination by other organs and/or tissues. Because the blood-to-plasma partitioning of amifostine is 0.5 (Souid et al., 1998), the total body clearance of drug, based on blood levels, would be twice that of its total body plasma clearance (i.e., CL_b,TB = CL_p,TB · 2). Therefore, drug elimination by other processes would be determined as CL_b,_other = CL_b,TB  (CL_b,Spl + CL_b,Lu + CL_b,R). Thus, the contribution (percentage) of amifostine elimination by CL_b,_other was estimated at 70% for the low-dose rate and at 50% for the middle- and high-dose rates. Taken as a whole, a substantial portion of amifostine is eliminated by other clearance mechanisms. It should be appreciated that alkaline phosphatase is ubiquitous within the body (McComb et al., 1979). Although speculative, amifostine could be metabolized by many other tissues such as the muscle, brain, bone, adrenal glands, and vascular endothelium. It is also possible that these tissues may be contributing to the nonlinear total body clearance that is observed for amifostine.

A high extraction of amifostine by the liver suggests that it might serve as a protective agent to normal hepatocytes during hepatic radiotherapy or radiochemotherapy. The role of radiotherapy in the management of patients with diffuse intrahepatic cancer has been limited by radiation-induced liver disease. An agent that would protect the normal liver from the effects of ionizing radiation, without compromising tumor cell kill, would be very valuable. Previous studies in rats have demonstrated that systemic administration of amifostine protects irradiated hepatocytes from reproductive cell death with a radiation dose modification factor of 2 (Jirtle et al., 1985) and that the liver is protected from radiation fibrosis with a dose modification factor that is greater than two (Jirtle et al., 1990). In a subsequent study (Symon et al., 2001), either systemic or portal venous administration of amifostine effectively protected hepatocytes from ionizing radiation without compromising tumor cell kill in a rat liver tumor model. In addition, a higher liver/tumor ratio of free WR-1065 was achieved after portal venous administration of amifostine, compared with systemic dosing. These findings suggest that amifostine may be a selective normal tissue radioprotectant in liver cancer and that regional/portal infusions may be preferable to systemic dosing. Our dog data, exhibiting a high activation/extraction of amifostine by the liver, support this contention.

It has been documented that the major source of blood flow to macroscopic hepatic cancers is by way of the hepatic artery (Sigurdson et al., 1987). In contrast, the delivery of nutrients to normal liver tissue. is primarily a function of the portal circulation (Ridge et al., 1987). Thus, amifostine selectivity in liver may not only be enhanced by differences between normal tissue and tumor in alkaline phosphatase activity, but also by differences in the drug's regional route of delivery (i.e., portal vein is favored). We are currently investigating the use of systemic and regional amifostine, as a radiation protector of normal liver, in patients with cancer.