Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Chemokines are potent chemotactic cytokines classified structurally by the invariant position of cysteine residues near the amino terminus of the protein (CC, CXC, C, and CX₃C, where C is the standard letter designation for cysteine and X represents an unspecific amino acid residue). Chemokines attract host defense effector cells along concentration gradients and are involved in numerous physiological processes, including leukocyte trafficking, inflammation, angiogenesis, tumor growth, and human immunodeficiency virus suppression (Rollins, 1997; Baggiolini, 1998; Schwarz and Wells, 1999).

SB-251353 is a novel truncated form of the human CXC chemokine growth-related gene product beta (GRO, residues 5-73), with potent anti-infective and hematopoietic activities (King et al., 2000). SB-251353 is a basic, heparin-binding protein with a molecular mass of ~7500 Da. Studies in mice have shown that SB-251353 increases blood neutrophil counts with a single administration of 0.1 mg/kg. With higher single doses (2.5 mg/kg), SB-251353 mobilizes hematopoietic stem cells into the peripheral blood. Although the details of the molecular transduction pathway initiated by SB-251353 have not been elucidated, it has been shown that stem cell mobilization by GRO-T is mediated by matrix metalloproteinase-9 (King et al., 2001).

Although there exists significant potential for the therapeutic use of this class of protein agents, extremely limited information exists regarding the pharmacokinetics of chemokines in animals (Laterveer et al., 1996; Vetillard et al., 1999), and no studies of the tissue distribution of a chemokine have been reported. To begin to develop an understanding of the mechanism and time course of pharmacologic action of SB-251353, we studied its pharmacokinetics following single and repeated i.v. administration in the mouse (0.1-250 mg/kg). Tissue distribution following i.v. administration of radioiodinated chemokine was studied using both macroscopic and light microscopic autoradiographic techniques (0.1 and 2.5 mg/kg).

These data provide information regarding the likely mechanisms of elimination of the chemokine, including the dose range over which different mechanisms may be influential. The potential influence of receptors, including CXCR2, the Duffy antigen/receptor for chemokines (DARC), as well as tissue glycosaminoglycan, on the fate of SB-251353, is considered. Comparisons are made between the disposition of this chemokine and previously studied growth and hematopoietic factors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Recombinant SB-251353 and RANTES (regulated on activation, normal T cell expressed and secreted) were expressed in Escherichia coli at GlaxoSmithKline (King of Prussia, PA). Goat anti-mouse IgG (Fc-specific polyclonal antibody) and horseradish peroxidase-conjugated streptavidin were purchased from Pierce Chemical Co. (Rockford, IL). Mouse anti-human GRO monoclonal antibody and goat anti-human GRO polyclonal antibody were from R & D Systems (Minneapolis, MN). Na¹²⁵I was from Amersham Pharmacia Biotech (Arlington Heights, IL). Iodobeads were from Bio-Rad (Hercules, CA). All other chemicals were of reagent grade or better.

Preparation of [¹²⁵I]SB-251353. SB-251353 was radio-iodinated using a modified solid-phase chloramine T method (Tsomides et al., 1991). SB-251353 has no tyrosine residues, and therefore the method was designed to incorporate the radiolabel into histidine. The reaction mixture contained 250 µg of SB-251353 and 1 mCi of Na¹²⁵I in 100 µl of 0.2 M sodium borate, 0.05% (v/v) Tween 20, pH 8.7, and a single IODO-BEAD (Pierce Chemical Co.). After a 10-min incubation at ambient temperature, protein-associated radioactivity was separated from free radiolabel by size exclusion chromatography in 10 mM HEPES buffer, 0.05% Tween 20, 150 mM sodium chloride, pH 7.4. Approximately 20% of the radioactivity in the final preparation was not protein-associated and could be separated from the product by diafiltration. For the in vitro blood cell binding studies (below), the preparation was dialyzed before use so that >95% of the total radioactivity was associated with the protein. SDS-polyacrylamide gel electrophoresis (PAGE) was used to assess radiochemical purity of the product (>90%). Specific radioactivity was 0.4 µCi/µg.

Blood Cell Association and Stability in Vitro. Fifty male BALB/c mice were sacrificed by carbon dioxide asphyxiation, and whole blood was collected from the vena cava. [¹²⁵I]SB-251353 was added to ice-cold EDTA whole blood at nominal concentrations of 50 ng/ml and 25 µg/ml. Aliquots were removed for analysis, and the remaining blood was placed in a shaking water bath at approximately 37°C for 10 or 60 min. Following incubation, plasma was separated by centrifugation, and blood cells were washed sequentially (three times) with phosphate-buffered saline. The percentage of trichloroacetic acid soluble radioactivity in plasma was determined as described previously (Davis et al., 1992). Radioactivity in blood, plasma, washes, and blood cells was determined by scintillation counting. In a separate experiment, the effect of RANTES (25 µg/ml) on the blood cell association of [¹²⁵I]SB-251353 (50 ng/ml, on wet ice) was assessed.

Pharmacokinetics. Mice were housed individually in stainless steel cages in a controlled environment (68-76°F; 40-70% relative humidity) with a 12-h light/dark cycle. The mice were offered Certified Rodent Diet (PMI Feeds, Inc., St. Louis, MO) or similar food ad libitum; filtered tap water was available from an automatic watering system. Thirty-six male (30-40 g) and 36 female (20-30 g) CD-1 mice (Charles River Laboratories, Raleigh, NC) were given 14 daily i.v. bolus doses of 2.5, 25, or 250 mg/kg SB-251353 by tail vein injection. The dosing volume was 10 ml/kg in 20 mM succinic acid, 6% (w/v) sucrose, 0.02% (w/v) Tween 80, pH 4. A terminal blood sample was collected from the vena cava of each mouse (following carbon dioxide asphyxiation) such that three mice/sex/time point were sampled for each dose group at nominal times of 5 and 30 min and 1, 3, 6, and 24 h after dosing on days 1 and 14. Single dose pharmacokinetic studies were also performed in CD-1 and BALB/c mice (Charles River Laboratories) at an i.v. bolus dose of 0.1 mg/kg (n = 20-30 per group). In these studies, the dose volume was 4 ml/kg in sterile saline. Terminal blood samples were obtained from three to five mice/time point at nominal times of 2, 10, 20, 40, 60, 80, 120, and 160 min post dose.

Tissue Distribution. Four groups of male BALB/c mice (n = 7 per group) received a single i.v. dose of [¹²⁵I]SB-251353 by bolus injection via a tail vein. Animal husbandry was as described above except that 20 mM sodium iodide was added to the drinking water beginning 72 h before dose administration. Doses were 0.1 mg/kg to groups 1 and 2 and 2.5 mg/kg to groups 3 and 4 (~1 µCi/mouse). The vehicle was phosphate-buffered saline, and the dose volume was 10 ml/kg. Each group was divided into three subgroups [A (n = 3), B (n = 2), and C (n = 2)]. At either 10 min (groups 1 and 3) or 120 min (groups 2 and 4), the mice were sacrificed and samples collected.

Urinary Excretion in the Mouse. Eight BALB/c mice (Charles River Laboratories) were divided into two groups of four animals and acclimated to metabolism cages (Nalgene, Rochester NY). Following i.v. bolus administration of unlabeled chemokine (0.1 or 25 mg/kg; 10 ml/kg phosphate-buffered saline), urine was collected over a 0- to 24-h interval into polycarbonate containers surrounded by dry ice. Urine collection volume was determined gravimetrically. The amount of SB-251353 excreted in urine was calculated from the urine drug concentration (by immunoassay) and the urine collection volume.

Immunoassays for SB-251353. Plasma samples were analyzed for SB-251353 using an enzyme-linked immunosorbent assay. In the assay, microtiter plates were coated with orienting goat anti-mouse IgG (Fc-specific polyclonal antibody), and then the plates were coated with mouse anti-human GRO monoclonal antibody. Neat plasma samples were diluted to within the assay calibration range of 25 to 1600 pg/ml in 50% (v/v) EDTA mouse plasma, 50% (v/v) PBS-Tween-bovine serum albumin buffer (10 mM sodium phosphate, 150 mM sodium chloride, 0.05% Tween 20, 0.1% bovine serum albumin, pH 7.4). SB-251353 was subsequently captured on the plate, and bound SB-251353 was probed with biotin-conjugated goat anti-human GRO polyclonal antibody. The sandwich was detected with horseradish peroxidase-conjugated streptavidin. The urine assay was performed similarly except that the assay calibration range was 125 to 8000 pg/ml in 10% (v/v) urine in PBS-Tween-BSA buffer.

Pharmacokinetic Analysis. For the pharmacokinetic studies, plasma clearance (CL), steady-state volume of distribution (V_ss), and mean residence time (MRT) were estimated by noncompartmental methods (Gibaldi and Perrier, 1982) using WinNonlin Professional version 1.1 (Apex, NC). Mean concentrations and blood collection times for each group were used in the analysis. The initial distribution volume was estimated by dividing the dose by the maximum observed plasma concentration after bolus i.v. administration.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Pharmacokinetics. The single and multiple dose pharmacokinetics of SB-251353 were studied over a broad dose range in the mouse. Plasma concentration time profiles for these studies are depicted in Fig. 1. Results of noncompartmental analyses of these data are summarized in Table 1. Single and multiple dose pharmacokinetic data were similar, indicating no accumulation or time dependence of the chemokine disposition. Similarly, there was no difference in the pharmacokinetics between male and female CD-1 mice over the dose range of 2.5 to 250 mg/kg. Therefore, summary statistics of pharmacokinetic parameters were calculated using pooled data from both males and females from days 1 and 14. 

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Plasma concentration-time profiles after single or multiple daily i.v. administrations of SB-251353 to mice. Diamonds (day 1) and squares (day 14) are CD-1 mouse data (2.5-250 mg/kg). Open diamonds and squares correspond to males; closed diamonds and squares correspond to females. Open circles represent male BALB/c mouse data (0.1 mg/kg); closed circles represent male CD-1 mouse data (0.1 mg/kg).

Tissue Distribution. The mean tissue to plasma radioactivity concentration ratios as determined by direct radioanalysis, at 10 min and 120 min after the intravenous administration of [¹²⁵I]SB-251353 (0.1 and 2.5 mg/kg), are summarized in Fig. 2. Following intravenous administration of radiolabeled chemokine, radioactivity was rapidly and widely distributed into the tissues with the highest concentration found in the kidney. The overall pattern of distribution was similar for both dose levels, and as assessed by the tissue concentration ratios, most of the tissue radioactivity concentrations were dose-proportional at 10 and 120 min post dose. Although the bone marrow appeared to be an exception, variability between animals was high for this tissue (70-100% coefficient of variation). Furthermore, whole body autoradiographic analysis showed no difference in dose-normalized marrow concentration, and these were comparable with blood concentrations. For both dose levels, the concentrations of radioactivity in most tissues were higher at 10 min relative to 120 min; the exceptions that occurred in both dose groups were the cerebellum, cerebrum, eyes, pancreas, skeletal muscle, stomach, submandibular gland, testes, and thyroid/parathyroid.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Mean tissue to plasma radioactivity concentration ratios after a single i.v. administration of [125I]SB-251353 to the mouse. Data based on direct radioanalysis are the mean (n = 3), except the pituitary gland (n = 2). Average coefficient of variation of the mean was ~20% (both time-points and dose groups, all tissues).

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Whole body autoradioluminograph of radioactive distribution after a single i.v. administration of [125I]SB-251353 to the mouse. A, 0.1 mg/kg, 10 min; B, 0.1 mg/kg, 120 min; C, 2.5 mg/kg, 10 min; and D, 2.5 mg/kg, 120 min. Radioactivity was widely and rapidly distributed. Tissue distribution was similar for both dose levels. Radioactivity in the renal cortex was particularly concentrated.

View larger version (142K): [in this window] [in a new window]   Fig. 4.   Light microscopic autoradiographic photomicrographs of mouse tissues after i.v. administration of [125I]SB-251353. A and B, cortex of kidney, 2.5 mg/kg, at 10 (A) and 120 (B) min. At 10 min, moderate numbers of silver grains were localized over the epithelial cells at the urinary pole of the glomerulus and extending down into the proximal tubule (lumen and apical portion of epithelial cells), which decreased to minimal numbers at 120 min. A smaller (mild) number of silver grains were localized over proximal tubules cut in cross-section that where farther away from the glomerulus at both 10 and 120 min. C and D, dermis of skin, 2.5 mg/kg, at 10 (C) and 120 (D) min. Silver grains (mild) were present along the margins of elements of the dense irregular connective tissue of the dermis at both time points. There was an apparent decrease in the number of silver grains present at 120 min. E and F, liver, 2.5 mg/kg, 10 (E) and 120 (F) min. Minimal numbers of silver grains were localized over the cells (endothelial and/or Kupffer cells) lining the sinusoids. G and H, bone marrow, 2.5 mg/kg, 10 (G) and 120 (H) min. Minimal numbers of silver grains were localized over the sinusoids traversing the bone marrow.

Urinary Excretion. To quantify the extent of urinary excretion of SB-251353, groups of male BALB/c mice (n = 4/group) received single i.v. doses of 0.1 or 25 mg/kg unlabeled chemokine, urine was collected over 24 h, and drug in urine was quantified by sensitive immunoassay. Concentrations of SB-251353 were not quantifiable in urine for all four animals in the 0.1-mg/kg group. Using the assay LLQ and the urine collection volume, the extent of excretion of peptide was estimated to be <0.03% of the administered dose for the 0.1-mg/kg group. In the 25-mg/kg group, SB-251353 was quantifiable in the urine for all four animals. Mean urinary excretion was 0.4% (range of 0.005-1.63%) of the administered dose of 25 mg/kg.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The disposition of SB-251353, a novel truncated form of the human CXC chemokine GRO, has been characterized in the mouse following intravenous administration. The integration of pharmacokinetic and tissue distribution data provides insight regarding physiological factors that may potentially influence the pharmacokinetics, and thereby the pharmacodynamics of this chemokine, in animals and humans. Furthermore, these studies provide insight into the disposition of endogenous chemokines generally and potentially their mechanism of action. Extremely limited information currently exists regarding the disposition of any chemokine in vivo.

SB-251353 has nonlinear pharmacokinetics in the mouse. At the highest doses studied, clearance was comparable with the glomerular filtration rate in the mouse (Davies and Morris, 1993). At the lowest dose, clearance was at least 4-fold greater and exceeded the rate of blood flow to the kidney (Davies and Morris, 1993). It is plausible that at the lower dose, clearance of this chemokine occurs through its pharmacologic receptor, CXCR2, via endocytosis with subsequent lysosomal degradation. Granulocyte colony-stimulating factor (GCSF) receptor has similarly been hypothesized to be responsible for the saturable clearance of GCSF (Kuwabara et al., 1994; Houston et al., 1999), and similar phenomena have been reported for hepatocyte growth factor (Liu et al., 1992; Liu et al., 1995).

In the present investigation, saturable binding of SB-251353 to blood cells was observed in vitro. The fact that this was blocked by excess RANTES (a CC chemokine) suggests that binding of SB-251353 was due to DARC on red blood cells (Darbonne et al., 1991; Chaudhuri et al., 1993; Neote et al., 1994). It has been postulated that erythrocyte DARC acts as a sink for circulating chemokines (Darbonne et al., 1991). Erythrocyte DARC is not expected to internalize or degrade chemokine; however, the possibility exists that chemokine clearance may be mediated in part through DARC in tissues (Darbonne et al., 1991; Chaudhuri et al., 1997; Luo et al., 1997).

We did not directly observe binding of [¹²⁵I]SB-251353 to tissues expected to express the CXCR2 receptor or DARC (other than erythrocytes). This may be because the specific radioactivity of the radiolabel was low (<0.5 µCi/µg), as incorporation of ¹²⁵I into histidine was inefficient. The diffuse tissue distribution of CXCR2 may also limit our ability to detect binding to the receptor (Sozzani et al., 1997; Asagoe et al., 1998; Tani et al., 1998). On the other hand, it is plausible that the uptake in the bone marrow detected by direct scintillation counting at 0.1 mg/kg (10 min), which was diminished at the higher dose (Fig. 3), may be an indication of specific, saturable binding to the CXCR2 receptor. Saturable binding of GCSF to the GCSF receptor in bone marrow has been reported previously (Kuwabara et al., 1995).

Macroscopic and microscopic distribution of radioactivity after i.v. administration of SB-251353 to the mouse supports the hypothesis that this chemokine undergoes glomerular filtration. Microscopic distribution suggested uptake into renal proximal tubule epithelial cells. Limited excretion of SB-251353 in the urine at doses up to 25 mg/kg is consistent with catabolism of the chemokine in the tubules. Similar phenomena have been observed for other filtered proteins (for example, Hepburn et al., 1995).

Uptake clearance (CL_uptake) by the kidney can be estimated using the chemokine concentration in the kidney at 10 min (Fig. 2) from the distribution data, and from the area under the plasma concentration time curve from time 0 to 10 min from the pharmacokinetic study (as in Kuwabara et al., 1995). With this approach, mean kidney CL_uptake was 21 ml/min/kg for the 2.5-mg/kg group. CL_uptake was very similar to total CL estimated from the pharmacokinetic studies at the higher doses (2.5-250 mg/kg), and this value is similar to the glomerular filtration rate in the mouse (as discussed above).

As the dose increased from 2.5 to 250 mg/kg, the apparent steady-state distribution volume estimated by noncompartmental analyses increased dramatically (~3.6-fold). Also, initial plasma concentrations increased only 5-fold between 25 and 250 mg/kg. Estimates for the initial distribution volume from the observed C_max data (~1 l/kg) are unusually large for a protein and approach total body water in the mouse (Davies and Morris, 1993). These observations cannot be explained by the limited binding of SB-251353 to erythrocyte DARC.

Increasing volume without a change in clearance may be an indication of increased binding to tissues as the dose was increased. Microscopic autoradiographic data are consistent with this hypothesis, as binding to connective tissue in the liver and dermis was evident at both 0.1 and 2.5 mg/kg (independent of dose over this lower dose range). Because the connective tissue elements contain large amounts of heparin sulfate proteoglycans, it is possible that distribution of the highly basic SB-251353 (pI ~ 10) to these sites reflects a charge interaction. As there is a large amount of heparin sulfate proteoglycan distributed throughout the extracellular matrix, the capacity for binding SB-251353 would be large. This may explain the large, relatively constant distribution volume that was observed from 0.1 to 25 mg/kg. The V_ss increases over the dose range of 2.5 to 250 mg/kg may reflect more complete titration of the lower-affinity binding sites at the highest concentrations studied.

Localization of exogenously administered heparin-binding proteins, particularly growth factors, to heparin sulfate proteoglycans has been reported previously (Liu et al., 1995, 1998). The biological importance of chemokine-glycosaminoglycan complexation has not been fully elucidated, although it may be important for presentation to the pharmacologic receptor or for sequestration (Ruoslahti and Yamaguchi, 1991; Kuschert et al., 1998). Proteolytic release of heparin-bound chemokines may regulate their activity in vivo as suggested for growth factors (Saksela and Rifkin, 1990).

In summary, the pharmacokinetics and tissue distribution of a novel human CXC chemokine have been characterized in the mouse over a broad intravenous dose range. It is proposed that saturable binding and clearance may occur through the pharmacologic receptor, and this leads to nonlinearity in the pharmacokinetics. At low concentrations, the saturable component(s) of the clearance are more significant than clearance due to renal filtration, while at higher concentrations, renal filtration (with subsequent uptake and degradation in the proximal tubule) is the predominant route of elimination. Glycosaminoglycan complexation may be responsible for the large distribution volume of this chemokine and the apparent increase in distribution volume with increasing dose.