Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Hepatocyte growth factor (HGF) is a heterodimer with a 69-kDa -chain and a 34-kDa -chain, linked by a single disulfide bridge. HGF is translated from a single mRNA, as a single chain preproHGF (Gak et al., 1992; Naka et al., 1992). Extracellular processing by specific serine protease, HGF-activator or HGF-converting enzyme (Miyazawa et al., 1993), results in conversion from a biologically inactive form to active two-chain mature HGF. The active form of HGF is mainly distributed in the liver, followed by metabolism after endocytosis (Liu et al., 1992).

HGF was initially identified in its partially purified form to be a potent mitogen for mature hepatocytes in primary culture; this was followed by its complete purification and molecular cloning (Matsumoto and Nakamura, 1996). HGF potentially stimulates the proliferation of epithelial cells in the liver, kidney, lung, and other tissues (Higuchi and Nakamura, 1991; Ishiki et al., 1992; Matsumoto and Nakamura, 1996). HGF also promotes their motility and/or morphological change. HGF is a mediator of angiogenesis (Silvagno et al., 1995). DNA synthesis in both osteoclasts and osteoblasts is also stimulated by HGF (Grano et al., 1996). Thus, HGF exerts a variety of biological activities initiated by its binding to the HGF receptor, c-met protooncogene product. Exogenous administration of HGF exhibits antihepatic, antirenal, and antipulmonary failure effects in vivo (Ishiki et al., 1992; Kawaida et al., 1994; Ishii et al., 1995; Roos et al., 1995; Ohmichi et al., 1996). In addition, recent investigations have revealed that HGF exhibits proliferative activity not only in acute hepatitis but also in chronic hepatitis in experimental animals (Matsuda et al., 1995, 1997; Yasuda et al., 1996). Therefore, HGF is expected to act as a therapeutic agent for many types of diseases. However, its wide spectrum of biological activity may hinder its clinical use since the systemic administration of HGF may result in biological activities other than those expected. Therefore, its precise pharmacokinetic profile after systemic administration needs to be investigated to understand its biological activity in vivo.

Gene expression of HGF is enhanced, and its endogenous plasma level, assessed by enzyme-immunoassay (EIA), is increased in liver failure in both humans and rats (Lindroos et al., 1991; Tomiya et al., 1992; Arakaki et al., 1995; Shiota et al., 1995). Thus, HGF may act as a hepatotropic factor to repair the hepatic injury. Miyazawa et al. (1994) reported that most of the endogenous HGF associated with normal tissues is in the inactive form, whereas the active form is found selectively in injured tissues. Thus, it is believed that the conversion is critical before HGF can exert its tissue-specific activities. The inactive precursor of HGF has also been found in circulating plasma of patients with liver failure (Arakaki et al., 1995). Nevertheless, whether or not its active form is present in circulating plasma is still controversial because of the difficulty in detecting the active form in plasma, as it is crucial to avoid any artificial conversion at the time of blood sampling. Therefore, to understand the pathophysiological role of HGF, the disposition of its active form needs to be clarified.

We have been studying the clearance mechanism of HGF in rats using an in vivo, perfused liver system and cultured hepatocytes (Liu et al., 1992, 1995, 1997, 1998a). Based on our previous findings, we believe that its elimination mechanisms involve both receptor-mediated endocytosis and another low-affinity component, probably mediated by cell-surface heparan-sulfate proteoglycans (Liu et al., 1992, 1997, 1998b). To identify the clearance organ involved, we injected ¹²⁵I-labeled HGF into rats to determine the tissue uptake clearance (CL_uptake). The CL_uptake per kilogram of body weight of ¹²⁵I-labeled HGF was much higher in the liver than in other organs (Liu et al., 1992). This suggests that the liver plays a predominant role as the clearance organ. Nevertheless, more detailed information concerning the precise contribution of the liver to the overall elimination is still not available. In addition, it is possible that radiolabeling of HGF may result in a change in its biological activity and pharmacokinetic properties. We have reported that ¹²⁵I-labeled HGF can specifically bind to HGF receptors, and its mitogenic activity on rat hepatocytes in primary culture was almost comparable with that of unlabeled HGF (Higuchi et al., 1991). However, this finding cannot fully ignore the above possibility. For this reason, in the present study, we performed the pharmacokinetic analysis during i.v. infusion of unlabeled HGF to obtain pharmacokinetic information. The purpose of the present study is to clarify the efficiency of HGF extraction by each organ and estimate the individual contribution of each organ to the overall elimination. To this end, we determined both the plasma clearance and organ clearance for several organs (liver, kidney, and lung), which have been already identified as the target organs of HGF where it exhibits proliferative activity in vivo.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Male Wistar rats weighing 250 g (Nisseizai, Tokyo, Japan) were used. All animals were treated humanely. The studies reported in this manuscript have been carried out in accordance with the Declaration of Helsinki and the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health.

Infusion Study of HGF. Rats were divided into four groups with respect to blood sampling sites: group 1 underwent blood sampling only from the femoral artery, group 2 had samples removed from the femoral artery and hepatic vein, group 3 had samples removed from the femoral artery and hepatic and renal veins, whereas group 4 had samples removed from the femoral artery and right atrium. For all of the groups, rats were maintained under light ether anesthesia from the start of the experimental surgery to the final blood sampling point and underwent cannulation of the femoral vein and artery with PE50 polyethylene tubing (i.d. 0.58 mm, A 0.965 mm; Becton Dickinson, San Jose, CA). The human recombinant active form of HGF was purified from a culture medium of C-127 cells transfected with plasmid containing human HGF cDNA (Nakamura et al., 1989) and dissolved in saline containing 0.2% (w/v) rat serum albumin. For the infusion of HGF, stock solutions for all of the animals involved a series of different concentrations. In all rat groups, the infusion of the stock solutions was performed through the femoral vein, and the rate of infusion of such stock solutions was set at 0.02 ml/min/body. For rat group 1, the HGF concentration in the stock solution was set at approximately 0.5, 1.25, 2.5, 10, 50, and 250 µg/ml. For groups 2, 3, and 4, the HGF concentration in the stock solution was set at approximately 1 and 45 µg/ml, 2.5 and 5 µg/ml, and 1 µg/ml, respectively. Thus, the number of infusion rates in groups 1, 2, 3, and 4 were 6, 2, 2, and 1, respectively. To minimize the experimental errors, the actual infusion rate was directly determined by sampling the infused saline at the end of the experiment. The infusion rate thus determined is given in the legends of the corresponding figures and tables. The sampling time was 10, 20, 30, 45, 60, 75, and 90 min for group 1 and 20, 40, and 60 min for groups 2, 3, and 4. The sampled plasma volume was approximately 100 µl. The number of rats was 4, 4 to 5, 3 to 4, and 4 in groups 1, 2, 3, and 4, respectively. Hepatic and renal veins and right atrium were cannulated with SP-31 polyethylene tubing (i.d. 0.5 mm, A. 0.8 mm; Natsume Seisakusyo Co., Ltd., Tokyo, Japan). Blood was collected at specified times and transferred to Eppendorf tubes containing volume of 10% (w/v) citrate solution allowed to stand on ice and then centrifuged at 4°C to obtain plasma. In the case of group 1, the bile duct was also cannulated with PE10 polyethylene tubing (i.d. 0.28 mm, A. 0.61 mm; Becton Dickinson). HGF concentrations in plasma and bile were determined by an EIA kit (code 1EH1, Institute of Immunology, Tochigi, Japan) (Liu et al., 1997). The detection limit of this assay system was 1.2 to 37 pM with a coefficient of variation of 15%. In the present study, we routinely determined the background value using an EIA of blank plasma obtained after surgery but before starting the infusion. These background values were always below the limit of detection. In addition, the EIA kit for human HGF used in the present study was specific for human HGF with less than 1% cross-reactivity with rat HGF. This EIA system recognizes both active and inactive forms of human HGF. The detection limit of this assay system was 1.2 to 37 pM with a coefficient of variation of 15%. Since the extent of HGF distribution to red blood cells was found to be minor in our preliminary study, we monitored the HGF concentration only in the plasma fraction.

Pharmacokinetic Parameters. The mean plasma concentration (C_p) of group 1 after start of the HGF infusion was fitted to the following equation:

(1)

where R₀, CL_p, V_d, and T represent the infusion rate, plasma clearance, volume of distribution, and time, respectively. The steady-state HGF concentration in circulating plasma (C_pss) and extraction ratios in the liver (E_h), kidney (E_r), and lung (E_l) were obtained as follows:

(2)

(3)

(4)

(5)

where C_hvss, C_rvss, and C_rass represent the steady-state plasma HGF concentrations in hepatic vein, renal vein, and right atrium, respectively. Because of the limited number of sampling points, no fitting of plasma concentration-time profiles was performed in groups 2, 3, and 4, and the steady-state plasma concentration was determined as the mean of the plasma concentrations at 40 and 60 min for these groups. The half-life (T_1/2) was obtained as follows:

(6)

The organ clearance in the liver (CL_h) and kidney (CL_r) was obtained as follows:

(7)

(8)

where Q_h and Q_r are the hepatic and renal plasma flow rates. These were determined in the present study and obtained from the literature (I. Kino, Y. Kato, J. H. Lin, and Y. Sugiyama, submitted for publication), respectively.

(9)

Determination of Hepatic Plasma Flow Rate. [³H]Taurocholate (128.4 GBq/mmol, New England Nuclear, Boston, MA) was infused i.v. at a rate of 3.7 kBq/0.02 ml/min. This infusion study of taurocholate was performed in different groups of rats from those used for the HGF infusion study. This infusion study was also performed under ether anesthesia for the duration of the experiment. Taurocholate concentrations in both circulating plasma and hepatic vein were determined at 15, 20, 25, and 30 min by measuring ³H radioactivity. Both C_pss and C_hvss for taurocholate were taken as the mean values of the plasma concentrations at 25 and 30 min. CL_p and E_h for taurocholate were calculated based on eqs. 2 and 3, respectively. Q_h was calculated as follows:

(10)

as this compound is almost completely recovered in bile in its unchanged form. Since the distribution of taurocholate to red blood cells was also minor, we only monitored its plasma concentration.

Statistical Analysis. For the analysis of the difference between two data sets, the test for equal variance (F test) and a subsequent Student's t test were performed on the two means of the unpaired data. A p value of less than .05 was considered to be statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Nonlinear Elimination of HGF in Plasma at Steady State. HGF was administered i.v. in group 1 at various infusion rates, and its plasma concentration was determined (Fig. 1). The plasma HGF concentration reached steady state after the start of infusion at all the infusion rates studied (Fig. 1). The CL_p decreased as the C_pss increased (Fig. 2). This reduction was not complete, and the CL_p reached an almost constant value at a C_pss higher than 100 pM (Fig. 2). Based on a one-compartment analysis taking into account both linear and nonlinear elimination pathways, the K_m, V_max, and CL_ns values were found to be 17.2 ± 8.5 pM, 391 ± 62 fmol/min/kg, and 19.7 ± 4.1 ml/min/kg, respectively (mean ± calculated S.E.). The Akaike's information criterion obtained by this fitting was 3.37. The V_d was 351 ± 84, 582 ± 84, 301 ± 71, 402 ± 56, 323 ± 19, and 211 ± 23 ml/kg at an R₀ of 0.162, 0.378, 0.857, 2.81, 13.0, and 70.4 pmol/min/kg, respectively (mean ± calculated S.E.). The corresponding T_1/2 and Akaike's information criterion were 7.53 and 11.9, 9.57 and 16.3, 8.68 and 10.7, 14.0 and 13.7, 10.8 and 26.9, and 6.69 and 23.1, respectively.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Time profiles of HGF concentration in circulating plasma during i.v. infusion. HGF was infused i.v. at rates of 0.162 ± 0.013 (), 0.378 ± 0.033 (), 0.857 ± 0.038 (), 2.81 ± 0.15 (), 13.0 ± 1.0 (), and 70.4 ± 5.8 () pmol/min/kg. Plasma HGF concentration was determined by EIA. Data represent mean ± S.E. of four rats. The indicated line is the fitted line based on eq. 1.

View larger version (9K): [in this window] [in a new window]   Fig. 2.   Existence of saturable and nonsaturable components of plasma clearance. Both Cpss and CLp were obtained from the data shown in Fig. 1. The horizontal and vertical bars represent the calculated S.E. values for Cpss and CLp, respectively. The indicated line is the fitted line based on eq. 9.

Extraction of HGF by Target Organs. Extraction of HGF by liver, kidney, and lung was examined in rat groups 2, 3, and 4 (Table 1). During the i.v. infusion of HGF at 0.59 pmol/min/kg, the C_hvss was lower than C_pss, whereas C_rvss was not significantly different from C_pss (Fig. 3). The E_h was determined using both C_pss and C_hvss at various infusion rates (Table 1). The E_h thus obtained was 0.60 at a C_pss of 7.5 pM, and this fell to 0.47 at a C_pss of 470 pM (Table 1). Since the C_rvss was almost comparable with the C_pss at the two infusion rates examined, E_r should be much lower than E_h and, at most, 0.1 (Table 1). The CL_h calculated based on eq. 7 was 62 to 73% of the CL_p at any infusion rate (Table 1). To investigate pulmonary extraction, the HGF concentrations in right atrium (inflow to the lung) and femoral artery (outflow from the lung) were determined simultaneously (Fig. 4). These two concentrations were very similar and not significantly different (Fig. 4).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Extraction of HGF by the liver and kidney. During the i.v. infusion of HGF at 0.585 ± 0.041 pmol/min/kg, HGF concentrations in circulating plasma (), hepatic vein (), and renal vein () were determined by EIA. Data represent mean ± S.E. of four rats. *, significantly different from HGF concentration in circulating plasma (p < .05)

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Extraction of HGF by the lung. During the i.v. infusion of HGF at 0.286 ± 0.025 pmol/min/kg, HGF concentrations in circulating plasma () and right atrium () were determined by EIA. Data represent mean ± S.E. of four rats. No significant difference was observed between the concentrations at any time examined (p > .05).

Biliary Excretion of HGF. During the i.v. infusion shown in Fig. 1, the biliary excretion of HGF was also measured (Fig. 5). Biliary excretion of HGF was detected at the higher three infusion rates but undetectable at the lower three infusion rates. The biliary excretion rate exhibited a lag time of 15 to 30 min, followed by an increase, as the infusion time increased, up to, at most, 0.1 to 0.2% of the corresponding infusion rate (Fig. 5).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Biliary excretion of HGF during i.v. infusion. During the i.v. infusion of HGF at 2.81 ± 0.15 (), 13.0 ± 1.0 (), and 70.4 ± 5.8 () pmol/min/kg, biliary excretion of HGF was determined by EIA. Data represent biliary excretion rate, which was normalized by infusion rate (%), during each period and are shown as mean ± S.E. of four rats.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Because HGF is a potent mitogen for a variety of types of epithelial cells including hepatocytes (Matsumoto and Nakamura, 1996), it was expected that it would be developed as a therapeutic agent to treat hepatic failure. Our study of the plasma elimination and tissue uptake of ¹²⁵I-labeled HGF suggests that the liver is the most important organ as far as its clearance is concerned (Liu et al., 1992). To more precisely determine the contribution of the liver to its overall elimination, a steady-state infusion study such as that in the present study is the most appropriate method. The CL_h found at any C_pss examined was approximately 70% of CL_p (Table 1). This result demonstrates that the liver is the major clearance organ, accounting for 70% of the whole body elimination, over a wide range of C_pss (5-3000 pM) covering the endogenous HGF concentration found in human plasma under normal conditions and in liver diseases (3-1000 pM) (Tomiya et al., 1992; Arakaki et al., 1995; Shiota et al., 1995). Approximately 30% of CL_p is governed by extrahepatic clearance. We previously found that the tissue distribution of ¹²⁵I-labeled HGF in liver, kidney, lung, adrenal, and spleen could not be accounted for by its distribution to the extracellular space, and the CL_uptake in these organs is much higher than that in other organs (Liu et al., 1992). Therefore, these five organs may include the extrahepatic clearance organ, which accounts for the remaining 30% of CL_p. In the present study, the renal extraction of HGF was not significant, and the CL_r was much less than the CL_p (Table 1). Therefore, the contribution of the kidney to HGF elimination is quite minor. We also did not find any significant extraction of HGF by lung (Fig. 4). However, this does not necessarily refute the importance of the lung as a clearance organ. Considering the high plasma flow rate of the lung (125 ml/min/kg; Tsuji et al., 1983), the extrahepatic clearance (CL_p  CL_h) observed in the present study (~10 ml/min/kg) might be almost completely explained by pulmonary elimination, even if the pulmonary extraction ratio is only 8%. Because of experimental variability, such a low extraction could go undetected in the present analysis. Further studies have to be performed to identify clearance organs other than the liver. It should be noted that HGF seems to have a high extraction ratio in the liver (Table 1). This indicates that the hepatic clearance of HGF should be affected by any change in hepatic blood flow rate. The present study was performed under ether anesthesia, and it is known that the hepatic blood flow can be influenced by such anesthesia (Tsuji et al., 1983); the hepatic extraction and clearance of HGF may also have been changed by such treatment.

We have reported that receptor-mediated endocytosis, especially in the liver, plays an important role as a clearance site of HGF (Liu et al., 1992). After the induction of down-regulation of HGF receptors either by hepatic malfunction or i.v. administration of excess HGF, the HGF receptor density on the liver cell surface is reduced, resulting in a reduction in the hepatic elimination of HGF (Liu et al., 1995, 1998b). In addition, endocytosis of ¹²⁵I-labeled HGF in perfused rat liver is inhibited by an inhibitor of receptor-mediated endocytosis, phenylarsine oxide (Liu et al., 1992). Specific binding of HGF to its receptor has an equilibrium dissociation constant (K_d) of 20 to 40 pM (Higuchi and Nakamura, 1991). Therefore, under conditions where there is an excess of HGF, much greater than the K_d, then almost all of the receptors should be occupied, resulting in the saturation of receptor-mediated endocytosis. Actually, in the present study, the CL_p exhibited saturation at a K_m of 17.2 pM, which is comparable with the K_d for the HGF receptor. This saturation in CL_p can be explained by saturation of HGF receptors. Nevertheless, the CL_p at a C_pss (~3000 pM) much higher than the K_d was at least half the CL_p at a C_pss under linear conditions (<10 pM) less than the K_d (Fig. 2). This result clearly shows the existence of a low-affinity and high-capacity clearance mechanism for HGF, which cannot be saturated at a C_pss of ~3000 pM. HGF is known to bind to heparan-sulfate proteoglycans expressed on the surface of ubiquitous cells (Lyon et al., 1994a,b), and more than half the surface-bound ¹²⁵I-labeled HGF in perfused rat liver can be washed out by heparin (Liu et al., 1992). Therefore, such proteoglycans may act as a high-capacity clearance system that removes any excess HGF molecules from the circulating plasma.

Although HGF was first identified as a potent mitogen for mature hepatocytes, recent studies have revealed that HGF exhibits biological activity in various types of epithelial cells. Exogenously administered HGF exhibits mitogenic activity not only in the liver but also in the kidney and lung (Ishiki et al., 1992; Kawaida et al., 1994; Ishii et al., 1995; Roos et al., 1995; Ohmichi et al., 1996). Therefore, the pharmacological activity exhibited in each organ depends upon the efficiency of extraction of HGF by each organ. The present findings demonstrate that the hepatic extraction ratio is 40 to 60% at any C_pss examined and much higher than the kidney and lung. Thus, HGF present in circulating plasma is efficiently captured by the liver. Gene expression of HGF is accelerated in liver, kidney, lung, and spleen when various types of tissue injuries are present, including acute hepatitis, partial hepatectomy, acute renal failure, partial nephrectomy, and acute pulmonary injury (Nagaike et al., 1991; Hamanoue et al., 1992; Yanagita et al., 1992; Igawa et al., 1993; Yanagita et al., 1993; Joannidis et al., 1994; Yamaguchi et al., 1996). Miyazawa et al. (1994) have demonstrated that endogenous HGF is synthesized as an inactive form, which is then activated only in the injured organs. Therefore, the active form of HGF can be synthesized in liver, kidney, and lung when these organs are damaged. Based on the present finding (Figs. 3 and 4, Table 1), it can be speculated that most (>90%) of the active form of HGF synthesized in the kidney and lung cannot be extracted by these organs through the first single-pass and should be released into the circulating plasma. This might lead to the biological activity of HGF being exhibited in organs other than those damaged. Also, in the case of liver injury, a portion of the active form of HGF synthesized in the liver should be extracted by the liver through the first single-pass. Nevertheless, such hepatic extraction is at most 60% (Table 1), and the remaining HGF avoiding the hepatic first-pass trap should enter the circulating plasma. Considering that the ubiquitous cells in the body respond to HGF (Matsumoto et al., 1996), such low extraction in each organ seems to result in biological responses to endogenous HGF in many organs other than those injured. This may be unfortunate in view of the aim of endogenous HGF to exert its biological activity only in the damaged organ. Other mechanisms apart from such tissue extraction of HGF may be present for the specific targeting of the active form of HGF to the damaged organ. Therefore, to understand the pathophysiological role of HGF, further studies need to be designed to clarify the fate of this active form of HGF once it is synthesized in damaged organs.

The present findings show that a small portion of the injected dose of HGF is recovered in the bile in unchanged form as determined by EIA. Liu et al. (1994) found that after the injection of ¹²⁵I-labeled HGF into rats, a portion of the radioactivity excreted into bile had a molecular weight similar to authentic ¹²⁵I-labeled HGF. Therefore, it is likely that a minor portion of endocytosed HGF is transported into the bile without lysosomal degradation. Concerning the other ligands that are endocytosed via cell-surface receptors, Burwen et al. (1984) produced evidence for a nonlysosomal pathway, which is independent of the lysosomal degradation system, for epidermal growth factor in rat liver. Such a direct biliary transport pathway in the liver is also known for IgA (Schiff et al., 1984). The physiological meaning of such biliary excretion is still unknown. HGF exhibits the proliferative activity of bile duct epithelial cells (Joplin et al., 1992; Matsumoto et al., 1994), which may respond to and/or endocytose HGF excreted in bile.

It should be noted that in the present study human HGF was used in rats in vivo. The homology of the amino acid sequences between human and rat HGF has been reported to be more than 90% (Tashiro et al., 1990). The homology in amino acid sequence of full-length c-met was 89% between mouse and human (Chan et al., 1988), and the partial sequence of rat c-met was reported to be 95% similar to that of the corresponding sequence in mouse c-met (Tsuji et al., 1994). Thus, the homology in the primary structure of HGF and its receptor is high for the two species. Nevertheless, it is still possible that a species difference exists in the pharmacokinetic properties of HGF. The biological activity of human and rat HGF, assessed as the mitogenic response of primary cultured rat hepatocytes, was similar (Matsumoto and Nakamura, 1996). Therefore, the receptor binding of these two forms of HGF to rat HGF receptors should be similar. In fact, we found there was receptor-mediated endocytosis of ¹²⁵I-labeled human HGF in a perfused rat liver system (Liu et al., 1992). Thus, human HGF can bind to rat HGF receptors, and the present rat model may still be useful to characterize the pharmacokinetics of HGF. It is possible to use rat HGF in rat models to completely avoid any species difference. Further studies are needed to clarify such species differences in the pharmacokinetic properties of HGF.

In conclusion, HGF in circulating plasma is efficiently extracted by the liver, compared with other target organs of HGF, through both receptor-mediated endocytosis and another high-capacity elimination mechanism. The liver is the major clearance organ, and hepatic elimination accounts for 70% of the overall elimination under both linear and nonlinear conditions.