Introduction

Top Abstract Introduction Methods Results Discussion Conclusion References

ILP is a form of regional chemotherapy for cancer that involves the exposure of a tumour-bearing limb to a high concentration of anticancer agent, although sparing the patient from systemic toxicity (Ghussen et al., 1989; Krementz et al., 1985; Krementz et al., 1987; Scott et al., 1992a, Thompson and Gianoutsos, 1992). Creech et al. (1958) first used the isolated cytotoxic limb perfusion technique to treat a patient with multiple superficial recurrences of melanoma localised to the leg. The efficacy and safety of the procedure has improved considerably and ILP is now widely accepted as the treatment of choice for locally recurrent melanoma restricted to the limbs, with response rates of 50 to 80% having been reported (Hartley and Fletcher, 1987; Baas et al., 1988; Krementz et al., 1987; Scott et al., 1992a; Hafstrom et al., 1991; Thompson et al., 1994). ILP has also been successfully used in treatment of sarcomas of the extremities (Englund et al., 1971; Lejeune et al., 1988; Stehlin et al., 1984; Di Filippo et al., 1988; Kettelhack et al., 1990). The drug of choice in ILP for melanoma is melphalan. Although this drug has been in use for many years, an understanding of the pharmacokinetic behaviour of the melphalan in the perfused tissues, on which an optimal dosage and treatment schedule could be based, has not been available (Hafstrom et al., 1984; Briele et al., 1985; Benckhuijsen et al., 1986; Scott et al., 1992b).

Benckhuijsen et al. (1988) estimated the time course of tissue levels of melphalan during ILP using a biexponential model for the perfusate melphalan concentration versus time curves. Three previous studies (Scotter et al., 1987; Scott et al., 1992a; Klaase et al., 1994) have reported melphalan concentrations in the limb perfusate and tissues during ILP. However, the tissue concentrations of melphalan in these studies were generally lower than those calculated by Benckhuijsen et al. (1988) who also suggested that the uptake of melphalan into the normal tissues during ILP greatly exceeded uptake by the tumor tissue. Quantitative data on the amount of drug taken up by limb tissues during perfusion for different perfusate conditions appears undefined but relevant in determining the optimal composition of perfusate to be used in ILP.

Leakage of drug-containing medium into the systemic circulation after ILP is unavoidable, even if limited in extent (Hafstrom et al., 1984; Lejeune and Ghanem, 1987). Systemic melphalan concentrations are about 1 to 4% of peak concentrations in perfusate (Hafstrom et al., 1984; Minor et al., 1985; Scott et al., 1992b). Rauschecker et al. (1991) suggested that between 1.4 and 18% of the total melphalan dose reaches the systemic circulation after reconnection. They suggested that a careful washing-out procedure was needed to ensure that the unwanted escape of melphalan into the systemic circulation was minimised (Rauschecker et al., 1991; Scott et al., 1992b). This washout procedure is undertaken before disconnection of the limb from the perfusion circuit, with removal of the melphalan-containing perfusate by a single-pass perfusion using drug-free perfusate. The washout solutions that have been used are of variable compositions and include plasma expanders mixed with electrolyte solutions (Rauschecker et al., 1991), Ringer's lactate (Scott et al., 1992a&b) and dextran (Krementz et al., 1987; Egerton, 1982). Currently, there is no pharmacokinetic data describing the kinetics of melphalan removal during the washout period.

In our study, we have investigated the tissue and perfusate concentrations of melphalan during and after melphalan perfusion in the IPRH model (Wu et al., 1993). Given that melphalan can exhibit variable protein binding in the perfusate (Wu et al., 1995a), we examined whether melphalan perfusate binding was a determinant of melphalan tissue concentrations during ILP and washout. An overall aim of this study was to examine whether melphalan would be best administered with a perfusate containing either albumin or dextran during ILP and similarly for melphalan removal during the washout period. We therefore quantified tissue and perfusate concentrations of melphalan using BSA and dextran perfusates during both perfusion and washout period. This data enabled us to design a physiologically based pharmacokinetic model which may assist in understanding the processes of melphalan distribution and removal during ILP.

	Methods

Top Abstract Introduction Methods Results Discussion Conclusion References

Rat Hindlimb Perfusion Preparation

The details of the single-pass rat hindlimb perfusion system have been described previously (Wu et al., 1993, 1995b). Briefly, rats (male Wistar, 320 ± 20 g) were anaesthetized, the abdomen opened and the right femoral artery cannulated (PE50) via the dorsal aorta. A second cannula (PE205) was placed in the dorsal vena cava, and the hindlimb perfused in a humidicrib with oxygenated (95% O₂/5% CO₂) Krebs-Heinseleit buffer (pH 7.4, 37°C), containing 4.7% BSA (Fraction V, Sigma Chemical Co., Sydney, Australia) or 2.8% dextran 40 (Sigma Chemical Co.). The oncotic pressures of perfusate containing 2.8% dextran 40 corresponded to the oncotic pressure produced by 4.7% BSA (Cross et al., 1996). In an attempt to modify the allergic response of rats to dextran (hindlimb edema during perfusion), animals were given an i.p. injection of 2.5 ml of 2.8% dextran 40 in perfusate 24 hr before the perfusion studies (Perez-Trepichio et al., 1991). We have validated the stability of the rat hindlimb perfusion for up to 2 hr by testing inflowing and outflowing perfusate samples for pH, dissolved oxygen concentrations, K⁺, lactate dehydrogenase and creatinine kinase as markers of cell damage (Wu et al., 1993). Perfusion flow rates of 3.80 ± 0.21 ml/min, were controlled by a graduated peristaltic pump and were measured at the beginning and end of each perfusion.

Experimental Procedures

Before each experiment, melphalan powder (donated by Welcome Australia) was dissolved in HPLC-grade methanol to give a stock solution of 1 mg/ml. Melphalan solution (15.26 µg/ml) was prepared in the perfusate containing 4.7% BSA or 2.8% dextran 40 immediately prior to melphalan perfusion. Rat hindlimbs were perfused with melphalan perfusate for 60 min, followed by a drug-free either BSA or dextran perfusate washout for 30 min. Perfusate inflow and outflow samples (0.5 ml) were taken at 0, 4, 8, 16, 20, 30, 40, 50, 60, 61, 62, 65, 70 and 90 min. Additional groups of rats were also sacrificed at various times (2, 5, 10, 15, 20, 30, 60, 70 and 90 min) during melphalan perfusion and hindlimb dissected. The perfused rat hindlimb was depilated with commercial Nair hair-removal cream and tissue samples (skin, fat and muscle) (500 mg) were taken in duplicate from selected points in the hindlimb for melphalan analysis.

Determination of Melphalan Concentrations

Melphalan concentrations in perfusate and tissue samples were analyzed by a HPLC assay previously developed in our laboratory (Wu et al., 1995a). Briefly, perfusate samples were analysed following methanol precipitation (100 µl sample with 200 µl methanol containing internal standard, dansyl-arginine), using a phenyl column and fluorescence detection. The detector was programmed to 265 nm excitation and 360 nm emission for melphalan and 265 nm excitation and 575 nm emission for the internal standard. The mobile phase consisted of methanol-water-glacial acetic acid (25:75:2, v/v), pH = 2.7, with 1-octanesulphonic acid added at a concentration of 50 mg/100 ml. The flow rate was 2 ml/min and the injection volume 20 µl.

Tissue samples (100 mg) were minced using scissors and suspended in 200 µl of methanol. The mixture was sonicated, on ice, for 1 min using an ultrasonic microtip and centrifuged at 10,000 × g for 15 min. The supernatant was removed and 20 µl injected into the HPLC system. The limit of quantitation of this assay for melphalan was 7.2 ng for tissue sample on column (Wu et al., 1995a).

Determination of Tissue Vascular Space

To calculate the vascular space in each tissue of the perfused rat hindlimb, ¹²⁵I-labeled albumin was mixed with melphalan in the perfusate of the IPRH. After 60 min perfusion, assuming the albumin had reached steady-state, an inflow sample was taken to measure the dpm/ml of albumin in the perfusate (C_perfusate). The tissue samples (skin, muscle and fat) were taken to determine the concentration of albumin (dpm) per g of tissue (C_tissue). The total weight (g) of each tissue type (Wt) present in the perfused limb was then determined by complete dissection. The skin hair was removed by the commercial Nair hair-removal cream. The quantity of ¹²⁵I-labeled albumin was determined in a Cobra II gamma-counter (Packard, Meriden, CT). The vascular space of each tissue (V_p) (ml) in the perfused rat hindlimb was then determined from the following formula:

(1)

Determination of Tissue Blood Flow

The blood flow in skin, muscle and fat during perfusion at flow rate of 3.8 ml/min in the IPRH were determined using the radiolabeled microsphere (⁵¹Cr, 10 µ) method described in an earlier study (Wu et al., 1995b). The total blood flow to each tissue type in the perfused hindlimb was assumed to equal the blood flow of tissue (ml/min/g of tissue) multiplied by the weight of each tissue type in the perfused hindlimb.

Theoretical Section

Model development. The physiological pharmacokinetic model proposed to describe melphalan distribution into the perfused tissues (skin, fat and muscle) after isolated perfused rat hindlimb is presented in figure 1. Each tissue can be described as a single compartments, connected to a single central plasma compartment (fig. 1). The differential mass balance equations can then be written for four compartments assuming: 1) a pseudo equilibrium exists between a given tissue and the perfusate in the hindlimb; 2) each tissue acts as a well-stirred compartment; 3) lateral spread of melphalan from one tissue to another tissue is negligible; 4) each tissue is a noneliminating organ, except for hydrolysis, where the rate constant in each tissue is the same.

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Diagrammatic representation of the pharmacokinetic model used to describe the disposition of melphalan in the isolated perfused rat hindlimb. Refer to text for meaning of various terms.

(2)

View larger version (15K): [in this window] [in a new window]   Fig. 2.   The concentration-time profile of melphalan in perfusates containing 4.7% BSA () or 2.8% dextran 40 () at 37°C. Mean ± S.D., n = 3.

(6)

(7)

(8)

(9)

Data analysis. The concentration of melphalan in the inflow perfusate at time zero is defined as C₀. The rate constant of melphalan hydrolysis in inflow perfusate was calculated from a semilogarithmic plot of concentration versus time (equation 2). The total amount of melphalan in the perfused rat hindlimb was calculated from the concentration of melphalan in each tissue (µg/g of tissue) multiplied by the weight of each tissue. The total vascular space was taken as the sum of the estimated vascular spaces for skin, muscle and fat.

	Results

Top Abstract Introduction Methods Results Discussion Conclusion References

Tissue vascular volume and blood flow. Individual perfusate flow rates in skin, muscle and fat determined in the rat hindlimb by perfusion with microspheres and perfused tissue vascular volumes determined by ¹²⁵I albumin perfusion are given in table 1. The volume and blood flow of each tissue were in the order of muscle > skin > fat, the vascular volume ratio of skin to fat and muscle to fat were 6.7 and 21.4, respectively, in correspondence with perfusate flow ratios of skin/fat and muscle/fat of 7.8 and 23.0, respectively.

Perfusate kinetics of melphalan. Melphalan hydrolysis in the perfusate followed an apparent monoexponential decline (fig. 2) with a hydrolysis rate constant of 0.011 ± 0.001 (4.7% BSA perfusate) or 0.011 ± 0.0004 min¹ (2.8% dextran 40 perfusate).

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Concentration or amount-time profiles for melphalan () in the outflow perfusate (A), skin (B), muscle (C) and fat (D) during 60 min melphalan perfusion followed by 30-min drug-free washout with 4.7% BSA containing perfusate. The solid line is the prediction from the model shown in figure 1, according to equations 4, 5, 7, 8 and 9. Values are reported as mean ± S.D. (n = 3).

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Concentration or amount-time profiles for melphalan () in the outflow perfusate (A), skin (B), muscle (C) and fat (D) during 60 min melphalan perfusion followed by 30-min drug-free washout with 2.8% dextran 40 containing perfusate. The solid line is the nonlinear regression using the model shown in figure 1, according to equations 4, 5, 7, 8 and 9. The dotted line is that obtained by simulation using parameters derived for a perfusate containing 4.7% BSA (table 2) and fu be altered from 0.521 (4.7% BSA) to 0.873 (2.8% dextran-40). Values are reported as mean ± S.D. (n = 3).

Distribution and pharmacokinetics of melphalan in perfused tissues. The total amount of melphalan in skin, muscle and fat in the IPRH increased with the time during melphalan (15.26 µg/ml) perfusion. The relative amounts of melphalan in each tissue was in the order: muscle > skin > fat for both BSA and dextran perfusions (figs. 3 and 4). A comparison of figures 3 and 4 shows that the amount of melphalan in each tissue was significantly higher using dextran perfusate than using BSA perfusate (P < .05). The amount of melphalan washed out from skin and fat using drug-free perfusate was characterized by an estimated apparent t_1/2 of 59 ± 2 min for both BSA and dextran perfusate (fig. 5A). However, the amount of melphalan removal from muscle was much slower and characterized by an estimated t_1/2 of 79 and 96 min for both BSA and dextran washout perfusates respectively (fig. 5A). The melphalan concentration remaining in the tissues after 30 min washout with dextran drug-free perfusate was slightly higher but not significantly different from a washout with BSA perfusate after 60 min melphalan perfusion with 4.7% BSA perfusate (P > .05) (fig. 5B). A similar result was obtained following perfusion of melphalan for 60 min using 2.8% dextran 40 perfusate (P > .05) (fig. 5C).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   A, Simulation of percentage melphalan remaining in tissues using 4.7% BSA (key of lines: 1. skin, 2. muscle, 3. fat) or 2.8% dextran 40 (key of lines: 4. skin, 5. muscle, 6. fat) washout solutions after melphalan (15.26 µg/ml) perfusions with 4.7% BSA for 60 min. B, The amount of melphalan in skin, muscle and fat with 4.7% BSA perfusate (60 min) and 30-min washout with drug-free perfusate containing 4.7% BSA (open symbol) or 2.8% dextran (filled symbol). C, The amount of melphalan in skin, muscle and fat with 2.8% dextran 40 perfusate (60 min) and 30 min wash with drug-free perfusate containing 4.7% BSA (open symbol) or 2.8% dextran (filled symbol). Values are reported as mean ± S.D. (n = 3).

	Discussion

Top Abstract Introduction Methods Results Discussion Conclusion References

This work has shown that melphalan concentrations differ significantly in muscle, skin and fat after perfusion in the IPRH. Scott et al. (1992a), in measuring the concentration of melphalan in skin and fat during in ILP, had previously shown a significantly lower melphalan concentration in fat (1.48 µg/g) than in skin (3.83 µg/g). Klaase et al. (1994) reported that the uptake of melphalan into muscle (6.5 µg/g) was higher than into skin (3.2 µg/g) after ILP. Our results are consistent with these reports in that the melphalan concentration in each tissue after melphalan perfusion with 4.7% BSA perfusate were of the order: muscle (18.38 µg/g) > skin (12.94 µg/g) > fat (9.76 µg/g). The differences in melphalan tissue concentration in this study compared to earlier work in humans arises in part from the use of the constant infusion input in this study as distinct from the bolus input used in the previous studies (Scott et al., 1992a; Klaase et al., 1994).

In the IPRH, the amount melphalan in each tissue increased during the time of perfusion (60 min) (figs. 3 and 4). The physiological pharmacokinetic model proposed appears to adequately describe the amount of melphalan in each tissue during this perfusion period. The only other data on melphalan tissue concentration-time profiles in a single subject have been reported by Benckhuijsen et al. (1988). However, these authors deduced, rather than measured, the tissue concentrations using a biexponential representation of the melphalan perfusate concentration-time data obtained during 60 min ILP after an initial bolus input of melphalan. They suggested a peak tissue concentration of melphalan at 15 min and the model assumed a single homogeneous tissue compartment in the ILP. In present work, melphalan concentrations had not reached steady-state during the 60 min infusion. This result is consistent with a report by Thompson et al. (1995), who used a noncompartmental moment analysis to show that the mean residence time of melphalan is 43 to 51 min during ILP.

A limitation in the analysis of data shown in figures 3 and 4 is the use of a compartmental model, in which each plasma or tissue compartment is regarded as being well-stirred (Wagner, 1988). The model is consistent with the physiologically based pharmacokinetic model developed by Bischoff and Dedrick (1968) which depicts the body as a number of well-stirred vascular compartments representing individual organs. The model assumes that intercompartmental transport occurs by blood flow only and that instantaneous equilibrium is achieved between tissue and the perfusing blood. The model may be extended to include well-stirred tissue compartments (Gerlowski and Jain, 1983). In reality, the representation of the body as a number of well stirred compartments is not consistent with the actual anatomy and physiology of the organs (Roberts and Rowland, 1985, 1996). In this study, a four-compartment model was used to describe the amount of melphalan in each tissue and in the perfusate. A simpler model, in which each tissue is considered separately as a two-compartment model with distribution of melphalan from the vascular space to the tissue defined by the measured blood flow, was also applied to the tissue data. The summed predicted perfusate concentration time course arising from the three tissues, with corrections for the flow rate in each tissue, appeared to correspond closely to the observed total perfusate outflow concentration profile for melphalan with time. Although the two compartment model of each tissue allows the description of the melphalan disposition in the IPRH, a more complicated model is needed to fit the observed perfusate and tissue data simultaneously.

In this study, the PS of melphalan was estimated to be 1.7 ml/min, the molecular weight (MW) is 305, pKa at pH 7.4 are 1.83 and 9.13 (Fasman, 1976). Sexton and Laughlin (1994) have reported that ⁵¹Cr-EDTA (MW 341) has a PS of 1.18 ml/min. The main source of the difference is likely to be differences of the chemical properties of melphalan and EDTA. The PS products for melphalan in each tissue is of the order: muscle > skin > fat, consistent with the ⁵¹Cr-EDTA results published by Sexton and Laughlin (1994) and Paaske (1976). Rowland and Tozer (1989) reported that the protein concentration in various tissues are also in the order: muscle > skin > fat. As an approximation, the fraction unbound of drug in a given tissue is given by: f_ut = 1/(1 + K_a * P_t), where K_a is the association constant of drug for the protein in the tissue and P_t is the total amount of protein in the tissue. Accordingly, the fraction unbound of drug in each tissue is of the order: muscle < skin < fat. Noting that the distribution volume (V_t) in each tissue is in the order of muscle > skin > fat, the f_ut/V_t in each tissue is predicted to be in the order: f_utf/V_tf > F_uts/V_ts > f_utm/V_tm. Our results for f_ut/V_t in various tissues (table 2) are consistent with this finding. The overall effect of these relationships is an outcome of the amount of melphalan uptake into muscle being significantly higher than that of into skin and fat.

As an alternative to PS and f_ut/V_t, the rate constant of influx in each tissue could also be expressed as K_12i = (f_u/V_p) * PS_i, and the rate constant of efflux from each tissue by K_21i = (f_uti/V_ti) * PS_i, where i represents each tissue. Whereas PS_i and f_uti/V_ti are not significantly different between 4.7% BSA and 2.8% dextran 40 perfusate, for the K_12i, the dextran 40 perfusate is significantly higher than the BSA perfusate. The differences in K_12i arise from the presence of f_u in this parameter. The fraction of melphalan unbound in the dextran perfusate (0.87 ± 0.10) is significantly higher than in BSA perfusate (0.52 ± 0.04) perfusate. In contrast, f_u is not a component of K_21i and similar values for K_21i were obtained for perfusates containing BSA and dextran.

By definition (equations 7, 8 and 9), the amount of melphalan in each tissue is related to the melphalan of f_u in perfusate, PS product and f_ut/V_t. The change of f_u in perfusate will therefore affect melphalan uptake into perfused tissues. The outcome is that a perfusate with low binding capacity for melphalan will facilitate melphalan uptake into the perfused tissues.

Rauschecker et al. (1991) reported that the majority of melphalan in the systemic circulation appeared immediately after reconnection of the vasculature when ILP finished, and that it is of major importance in the consideration of melphalan induced systemic side effects. Melphalan from the perfused limb tissues and vasculature will wash into the systemic circulation after reconnection of the ILP. At the end of the 60-min melphalan perfusion in the IPRH, the vascular space contained only 1.1% of the melphalan remaining in the tissues (figs. 3 and 4). In addition, whereas the perfusate is rapidly cleared of melphalan during first few min of the washout (figs. 3A and 4A), the release of melphalan from tissue is much slower. The slow efflux is consistent with previous studies (Scott et al., 1992a; Rauschecker et al., 1991), reporting that a major fraction of the melphalan was retained in the perfused tissue and that redistribution from the tissue compartment into the systemic circuit was slow. Rauschecker et al. (1991) suggested that 1.4 to 18% of the total melphalan dose reaches the systemic circulation after reconnection in ILP. In present study, approximately 14.5% of total melphalan dose would reach the systemic circulation if a washout procedure was not used.

A number of melphalan-free perfusates have been used in the washout procedure during ILP. They include plasma expander mixed with electrolyte solution (Rauschecker et al., 1991), Ringer's lactate (Scott et al., 1992a) and dextran (Krementz et al., 1987; Egerton, 1982). To minimize the later distribution of melphalan from the perfused limb into the systemic circulation, techniques for the washing-out phase need to be optimised (Rauschecker et al., 1991). We had hypothesized that a drug free washout solution containing BSA and longer washout periods may better facilitate melphalan removal from the perfused limb than a perfusate without any binding capacity for melphalan. In present study, no difference in melphalan removal from skin and fat were found between BSA and dextran perfusates (fig. 5A). This result suggests removal from tissue is effectively into perfusate under sink conditions, i.e., there is no substantial redistribution back to these tissues from the perfusate. Redistribution may be important for muscles as the melphalan removal from muscle was faster using 4.7% BSA than with using 2.8% dextran perfusate. Given that 72 to 74% of the melphalan is in perfused muscle for both BSA and dextran perfusate (figs. 3 and 4), the use of washout perfusate containing BSA to reduce the amount of melphalan reaching the systemic circulation will be that primarily from the perfused muscle.

In this work, a comparison of drug-free BSA or dextran perfusates during the washout period was conducted following identical melphalan perfusion conditions. With either 4.7% BSA or 2.8% dextran perfusate with melphalan perfusion, the amount of melphalan in each tissue after 30 min washout was less for the 4.7% BSA drug free perfusate than for the 2.8% dextran perfusate (fig. 5). This result suggests that using washout solutions that bind melphalan highly may facilitate melphalan removal from the perfused limb. However, the removal of 50% of the melphalan from each tissue by 4.7% BSA perfusate would require more than 1 hr washout time, especially for muscle. At present, the washout time in clinical practice is between 3 min (Scotter et al., 1987) and 4 min (Rauschecker et al., 1991).

A variety of media has been used that are likely to influence oxygen supply, arteriolar recruitment, drug binding in blood and perfusion pressure (Wu et al., 1993). What the desirable composition of the perfusate should be was raised by Kroon (1988) in his review of ILP procedures. Our work has suggested that protein binding influences both the input and output of melphalan in muscle during perfusion and washout. Melphalan has been reported (Greig et al., 1987) to bind to red blood cells (36.7%), plasma (84.1%) and human albumin (46.1%). Wu et al. (1995a) reported that melphalan binding to 4.7% BSA was 47.9%. In figure 6, the effect of changing f_u in the perfusate to the melphalan concentration-time profiles in outflow perfusate, skin, muscle and fat after the application of a constant input source to the IPRH is examined using the pharmacokinetic model (fig. 1) derived in this work. The conditions used were an initial applied concentration of 15.26 µg/ml and a perfusion flow rate of 3.8 ml/min with perfusate containing BSA, dextran, RBC, plasma or other solutions. The melphalan concentrations in the outflow perfusate with varying f_u of melphalan in the perfusate are in the order: f_u = 0.01 > f_u = 0.16 (plasma) > f_u = 0.52 (4.7% BSA) > f_u = 0.63 (RBC) > f_u = 0.87 (2.8% dextran 40) > f_u = 1. Consistent with model predictions, the amount of melphalan uptake into each tissue was highest for the highest fu with an order: f_u = 1 > f_u = 0.87 (2.8% dextran 40) > f_u = 0.63 (RBC) > f_u = 0.52 (4.7% BSA) > f_u = 0.16 (plasma) > f_u = 0.01. It is apparent that a perfusate with a high protein content perfusate impairs melphalan uptake into perfused tissues.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Simulated concentration or amount-time profiles of melphalan in perfusate (A), skin (B), muscle (C) and fat (D) after melphalan (15.26 µg/ml) perfusion for 60 min and washout with drug free perfusate for 30 min according to the model in figure 1 using perfusates in which the melphalan fraction unbound (fu) varies. Key: (1) fu = 0.01, (2) fu = 0.16 (plasma), (3) fu = 0.521 (4.7% BSA), (4) fu = 0.633 (red blood cell), (5) fu = 0.873 (2.8% dextran 40), (6) fu = 1.

Because the commonly applied process of hyperthermia may enhance melphalan hydrolysis, the effect of potential hydrolysis in tissues was examined in figure 7. The simulated melphalan concentration-time profile in the outflow perfusate was constant for K_m ranging from 0 to 0.1. However, the amount of melphalan in each tissue decreased proportionately as K_m increased.

View larger version (35K): [in this window] [in a new window]   Fig. 7.   Simulated concentration or amount-time profiles of melphalan in perfusate (A), skin (B), muscle (C) and fat (D) after melphalan (15.26 µg/ml) perfusion for 60 min and washout by drug free perfusate (4.7% BSA) for 30 min for various melphalan hydrolysis rates constants in tissues. Key: (1) Km = 0, (2) Km = 1014, (3) Km = 104, (4) Km = 0.01 and (5) Km = 0.1.

	Conclusion

Top Abstract Introduction Methods Results Discussion Conclusion References

A physiological pharmacokinetic model has been developed to describe the concentration-time profile of melphalan in both perfusate outflow and tissues in the isolated perfused rat hindlimb during perfusion with melphalan and washout periods. The predicted concentration-time profiles obtained appear to correspond closely to the observed data. Further simulations enabled the prediction of melphalan perfusate and tissue profiles with a dextran perfusate, after correction for difference in melphalan binding to perfusate components. The work also shown that higher melphalan protein binding in the perfusate impairs its uptake into tissue. However, the effect of protein binding on the washout kinetics was small, a comparable rate of elimination being observed for dextran and albumin perfusates. It is implied throughout that this physiological pharmacokinetic model will have use in the understanding of limb exposure to a chemotherapeutic agent such a melphalan during a local limb perfusion. The description of washout of drug from the perfused limb assists in understanding systemic drug exposure.