264
Bioconjugate Chem. 2001, 12, 264−270
Metabolism and Renal Clearance of Antibody Fragments
111In-Labeled
DOTA-Conjugated
S. W. Tsai,† L. Li,† L. E. Williams,‡ A.-L. Anderson,§ A. A. Raubitschek,§ and J. E. Shively*,† Division of Immunology, Beckman Research Institute of the City of Hope, 1450 East Duarte Road, Duarte, California 91010, Division of Diagnostic Radiology, City of Hope National Medical Center, Duarte, California 91010, and Department of Radioimmunotherapy, City of Hope National Medical Center, Duarte, California 91010. Received August 10, 2000; Revised Manuscript Received December 21, 2000
Radiometal-labeled antibody fragments are promising reagents for radioimmunotherapy due to their high tumor uptake and rapid pharmacokinetics, but their therapeutic potentials are limited by high uptake and retention in the kidney. Identification of metabolic products is a first step in designing rationale approaches to lower kidney uptake. Previous studies in rats have shown that 111In-labeled DTPA-conjugated antibody fragments (via lysine residues) were degraded to an DTPA--amino-lysine derivative and retained in the lysosomal compartments of the liver and kidney [Rogers et al. (1995) Cancer Res. 55, 5714s-5720s]. To determine the metabolic profile of another widely used metal-chelate, [111In]DOTA conjugated to lysines in antibody fragments via active ester chemistry, we analyzed kidney homogenates from nude mice injected with an [111In]DOTA-Fab generated enzymatically from the anti-lymphoma intact antibody Rituxan. The major kidney metabolite was identified as [111In]DOTA-amino-lysine by comparison to an authentic synthetic standard. This end product was also identified in the urine, along with relatively small amounts of [111In]DOTA-Fab. Since injection of [111In]DOTA-amino-lysine into nude mice resulted in rapid clearance into the urine without kidney retention, it is likely that the renal retention observed was due to kidney uptake of [111In]DOTA-Fab, followed by lysosomal degradation to [111In]DOTA--amino-lysine, which is only slowly cleared from this compartment. This observation is supported by autoradiographs of the kidney showing rapid localization of radioactivity into the distal regions of the kidney cortex. To extend this analysis to clinical trials, we have also analyzed urine taken from a patient injected with the intact antibody [111In]DOTA-cT84.66. In that example, we found that the major radioactive species was also [111In]DOTA--amino-lysine.
INTRODUCTION
Antibodies can be used to deliver radionuclides (Klein et al., 1989), toxins (Spitler, 1987), or cytotoxic drugs (Starling et al., 1988) to tumors due to their high specificity for tumor surface antigens. Radiometals that are conjugated to antibody via chelates are a good choice for radioimmunotherpay (RIT),1 as exemplified by clinical studies with radiometal labeled anti-lymphoma (Press, 1999), anti-colon cancer (Wong et al., 1995), and antibreast cancer antibodies (DeNardo et al., 1997). In RIT, the use of intact immunoglobulin conjugates lead to a long circulation time in blood, thus exposing normal tissues to unwanted radiation (LoBuglio et al., 1989). Due to its large size (150 kDa), the IgG molecule also shows a relatively slow diffusion from the vasculature into tumors (Yokota et al., 1992; Blumenthal et al., 1995). While IgG fragments exhibit a faster biodistribution and better tumor penetration than intact IgG (Yokota et al., * To whom correspondence should be addressed. Phone: (626) 359-8111. E-mail:
[email protected]. † Division of Immunology. ‡ Division of Diagnostic Radiology. § Department of Radioimmunotherapy. 1 Abbreviations: DOTA, 1,4,7,10-tetraazacyclododecane-N,N′, N′′,N′′′-tetraacetic acid; DTPA, diethylenetriaminepentaacetic acid; RIT, radioimmunotherapy; IgG, immunoglobulin; EDC, 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide; sulfo-NHS, Nhydroxysulfosuccinimide; CEA, carcinoembryonic antigen; BFC, bifunctional chelate; TFA, trifluroacetic acid; BFC, bifunctional chelate.
1992), this advantage is offset by nonspecific organ uptake, especially in the kidney (Andrew et al., 1988). Although this problem may be partially corrected by using metabolically labile isotopes such as radioiodine (Andrew et al., 1988), this is attained at the price of reduced tumor retention, thus requiring higher amounts of radioactivity to achieve the same radiation absorbed dose achieved with lower amounts of radiometal. Thus, although radiometal labeled antibody fragments have faster blood clearance and better tumor penetration than their intact IgG counterparts, their uptake and retention in the radiation sensitive kidney has limited their therapeutic application. The first step to lowering kidney uptake is to identify the radiolabeled metabolic products. A series of studies with [111In]DTPA-conjugated (via lysine residues) glycoproteins (Duncan and Welch, 1993; Arano et al., 1994) or antibody fragments (Anderson et al., 1995; Rogers et al., 1995) in animal models revealed that these conjugates were metabolized in the lysosomal compartment at both tumor and nontumor sites. For example, Duncan and Welch (Duncan and Welch, 1993) found that [111In]DTPAlabeled glycoproteins targeted to cell surface receptors were degraded to [111In]DTPA--amino-lysine, which was then slowly released from cells and recovered intact in both urine and feces. Rogers et al. (1995) identified [111In]DTPA--amino-lysine as the major metabolites of DTPAconjugated (via lysines) antibodies and antibody fragments. These findings suggest that lysosomal enzymes are unable to cleave the amide bond of [111In]DTPA--
10.1021/bc0000987 CCC: $20.00 © 2001 American Chemical Society Published on Web 03/06/2001
Metabolism of [111In]DOTA-mAb Fragments
amino-lysine and the end product is only slowly cleared from this compartment. Although DTPA is an excellent chelate for radiometalbased immunoconjugates, its radiometal stability constants, especially for the imaging and therapeutic radiometals 111In3+ and 90Y3+, respectively, are such that significant amounts of radionuclide escape into nontarget organs such as the bone (Williams et al., 1996; Govindan et al., 1998). For this reason we have begun to utilize the more stable macrocycle chelate DOTA (Li et al., 1994) which can be conveniently linked to lysines in proteins via simple active ester chemistry (Lewis et al., 1994). Using the metabolic studies on [111In]DTPA-lysine based conjugates as a guide, we speculated that DOTA-lysinebased conjugates would be degraded to [111In]DOTA-amino-lysine and retained in the lysosomal compartment. Therefore, we synthesized this derivative, studied its biodistribution in nude mice, the usual preclinical model for RIT, and showed that it is rapidly cleared into the urine with no net organ retention. However, when [111In]DOTA-Fab fragments are injected into nude mice, high kidney retention of [111In]DOTA--amino-lysine is observed, suggesting that kidney filtration of the fragment results in cellular uptake, followed by metabolism and retention in the lysosomal compartment. EXPERIMENTAL PROCEDURES
Preparation and Radiolabeling of Antibody Fab Fragment Conjugate. Rituxan (75 mg, IDEC) was digested with immobilized papain (Pierce) and purified on a protein A column (Bio-Rad) according to manufacturer’s protocols. Purified Rituxan-Fab (70% yield) gave a single band at 50 kDa on a 10-15% nonreducing SDSpolyacrylamide gel. The Fab was conjugated to DOTA using the water-soluble EDC/sulfo-NHS method according to Lewis et al. (1994) with a DOTA-ester-to-antibody ratio of 100:1. Following conjugation, the protein was dialyzed against 0.2 M NH4OAc, pH 7.2, and concentrated to 3 mg/mL. The antibody was radiolabeled with In-111 at a ratio of 2.5-3 mCi/mg. Indium-111 chloride in 0.05 N HCl was buffered with an equal volume of 0.2 M NH4OAc at pH 7 prior to addition of antibody. The reaction mixture was incubated at 43 °C for 1 h, then 10 mM DTPA solution was added to give a final concentration of 1 mM DTPA. This was followed by a 15 min incubation at room temperature, after which purification was performed by HPLC using a Superdex 75 HR 10/30 gel filtration column (Pharmacia) eluted with normal saline at a flow rate of 0.5 mL/min. Fractions of 0.5 mL each were collected into tubes containing 1 drop 25% human serum albumin. Labeling efficiency, as determined during HPLC purification, ranged from 84 to 96%. Synthesis of DOTA-E-amino-lysine and Radiolabeling. DOTA active ester was prepared as follows: DOTA (trisodium salt, 100 mg, 213 µmol), EDC (81 mg, 425 µmol), and sulfo-NHS (92 mg, 425 µmol) in were reacted in 150 uL of water at RT at pH 7.0 for 1 h. R-BocLys (104 mg, 425 µmol) dissolved in 1 mL of water was reacted with 0.5 equiv of the DOTA active ester for 22 h at RT. The solution was cooled to 4 °C and acidified with 1 mL of HCl and passed over a Bio-Rad Ag50W-X8 column (6 mL in the ammonium form). After extensive washing with water, the product was eluted with stepwise 10 mL aliquots of 0.5, 1.0, 2.0, and 4.0 M NH4OH. The fractions (1 mL) were dried and the product was identified by ESI/MS (Finnigan LCQ ion trap). The product was loaded onto a Dowex AG1-X4 (6 mL in the formate form), washed with water, and eluted stepwise with 10 mL each of 0.1, 0.2, 0.5, 1.0, and 2.0 M formic acid. The fractions (1 mL) were dried, and the product
Bioconjugate Chem., Vol. 12, No. 2, 2001 265
identified as before. The product (29 mg of white powder) had the correct mass by ESI/MS (MH+ of 532.6 expected, 532.9 observed) and the following proton NMR: (D2O, 500 MHz) δ 1.32-1.46 (m, 2H, γ-CH2 of lysine), 1.521.57 (m, 2H, β-CH2 of lysine), 1.80-1.94 (m, 2H, δ-CH2 of lysine), 2.92-3.04 (m, 4H, ring protons of DOTA), 3.06-3.20 (m, 6H, -CH2 of lysine plus ring protons of DOTA), 3.18-3.45 (m, 8H, ring protons of DOTA), 3.46 (s, 2H, acetate protons of DOTA), 3.52 (s, 2H, acetate protons of DOTA), 3.74-3.78 (m, 2H, R-CH2 of lysine), 3,80 (s, 2H ring protons of DOTA), 3.83 (s, 2H, ring protons of DOTA). Unreacted DOTA (60 mg) was recovered and recycled for further reactions. The product (15 µL of a 37.5 mM solution in water) was radiolabeled at 43 °C with 531 µCi of 111InCl3 (an equal volume of 0.2 M 7.0 NH4OAc buffer of pH 7.0 was added to 35 µL of 111 InCl3 in 0.04 M HCl). Biodistributions of [111In]DOTA-Rituxan-Fab and [111In]DOTA-E-amino-lysine. Groups of five nontumorbearing nude mice (Charles River) per time point were injected with 5.0 µCi (2.0 µg) of 111In-labeled DOTARituxan Fab via tail vein. Animals were sacrificed at 0, 1, 2, 4, 6, and 12 h after injection. Major organs and blood were weighed and radioactivity measured and expressed as percentage injected dose per gram of tissue. In a separate experiment, groups of two nontumor-bearing nude mice per time point were injected with 10 µCi (4.0 µg) of [111In]DOTA-Rituxan-Fab or 50 µCi (30 µg) of [111In]DOTA--amino-lysine. Kidneys, blood, and urine samples were collected at 1, 5, and 9 h after injection and the radioactivity measured. Kidney Cryostat-Sections. Kidney tissues were dissected from the mice injected with 5.0 µCi (2.0 µg) Rituxan Fab at 0, 4, and 12 h after injection. Renal tissues were embedded in O. C. T. (Miles Inc.), quick frozen on dry ice, cut into 4 µm thick slices, and were then mounted on slides. Frozen section slides were fixed in cold acetone for 10 min, and then exposed to a phosphor imager (Molecular Dynamics) for 2 weeks. Kidney Metabolites of [111In]DOTA-Rituxan Fab. Kidneys were removed from mice injected with 5 µCi (2.0 µg) or 10 µCi (4.0 µg) of [111In]DOTA-Rituxan Fab. The organs were rinsed with PBS, frozen and thawed, and homogenized in an equal volume of PBS using a Bounce homogenizer. The PBS insoluble fractions were boiled and solubilized in 1% SDS. Samples were then centrifuged in an Eppendorf centrifuge at 12 000 rpm for 60 min, passed through a 0.22 µm filter, and were analyzed by SDS-PAGE 10 to 15% gradient gels (Pharmacia, Phast gel), gel filtration chromatography (Pharmacia, tandem Superose 12 columns, each 1 × 30 cm, flow rate 0.5 mL/min in PBS) and anion-exchange chromatography (Perseptive Biosystems, Poros HQ, 4.6 × 100 mm, flow rate 1 mL/min in 20 mM Tris, pH 8.5. Gradient started 5 min after injection and increased to 0.02 M Tris, pH 7.0, 0.5 M NaCl over 35 min) Analysis of Patient Urine Sample. Urine was collected from a patient enrolled in a phase I therapy trial involving pretherapy imaging with the anti-CEA intact antibody (Williams et al., 1996) [111In]DOTA-chT84.66. The urine sample was collected between 24 and 48 h after administration of the pretherapy [111In]DOTA-chT84.66 antibody (5 mCi, 5 mg of antibody). Aliquots (200 µL) were applied to the anion-exchange column as described above. RESULTS
[111In]DOTA-Rituxan Fab Biodistribution. DOTA was conjugated to the lysine residues of purified Fab fragments of Rituxan using active ester chemistry and
266 Bioconjugate Chem., Vol. 12, No. 2, 2001 Table 1. Biodistribution of
111In-labeled
Tsai et al. DOTA-Rituxan Fab in Non-Tumor-Bearing Nude Micea
organ
0 (h)b
1 (h)
2 (h)
4 (h)
6 (h)
12 (h)
blood liver spleen kidney lung bone carcass
45.21(3.59) 7.11(0.52) 7.40(0.51) 58.82(7.91) 13.23(1.33) 2.90(0.38) 2.10(0.14)
1.44(0.28) 1.65(0.38) 1.31(0.43) 45.36(13.70) 1.57(0.20) 0.96(0.41) 1.46(0.56)
1.00(0.34) 1.70(0.36) 1.03(0.20) 43.20(13.88) 1.18(0.26) 0.77(0.20) 1.05(0.28)
0.76(0.10) 1.79(0.47) 1.10(0.26) 70.28(68.69) 1.01(0.08) 0.71(0.07) 1.08(0.07)
0.53(0.04) 1.43(0.23) 0.98(0.14) 79.14(60.15) 0.82(0.16) 0.65(0.03) 0.89(0.12)
0.30(0.10) 1.29(0.10) 1.01(0.20) 95.87(65.42) 0.62(0.10) 0.66(0.23) 0.82(0.17)
a Groups of five nontumor-bearing mice were injected with 5.0 µCi (2.0 µg) of 111In-labeled DOTA-Rituxan Fab. Mean values ((SD) are in units of percent injected dose per gram (% ID/g) for each organ at each time point. b Approximately 20 min after injection.
Figure 1. Kidney cryostat-sections. Kidneys excised from four mice injected with 5.0 µCi (2.0 µg) [111In]DOTA-Rituxan Fab at t ) 0, 4, and 12 h were cut into 4 µm thickness and then exposed onto a phosphor imager for 2 weeks. Little movement of the activity was observed in these specimens over the 12 h interval.
radioabeled with 111In (5.0 µCi, 2.0 µg). Biodistribution studies in tumor-free nude mice showed rapid blood clearance (1030 M-1 and negligible off rates for trivalent metals such as In3+ and Y3+ (Craig et al., 1989; Deshpande et al., 1990; Moi et al., 1990; Peterson et al., 1999). As with previous generation BFCs, lysines have been the major target residue for coupling to proteins (Moi et al., 1990). In our own work, we have shown that DOTA can be coupled to lysines in proteins using a simple activated ester chemistry (Lewis et al., 1994). In the case of intact antibodies labeled with [111In]DOTA, liver retention in the nude mouse model is moderate (=10% ID/g at 1-48 h) and kidney retention low (=12%
ID/g at 1-48 h) (Williams et al., 1998). However, the same BFC coupled to a Fab antibody fragment and labeled with 111In resulted in 45-95% ID/g in the kidneys (Table 1). This exceptionally high uptake was similar to that observed for [111In]DTPA labeled F(ab′)2 fragments (Kuhlmann and Steinstrasser, 1988; Massuger et al., 1991; Yemul et al., 1993). On the basis of the results for EDTA and DTPA based BFCs, we examined that possibility that the major metabolic end product would be [111In]DOTA--amino-lysine. Our analysis of the kidney homogenates by three methods (gel filtration, SDS-PAGE, and ion-exchange chromatography) revealed [111In]DOTA--amino-lysine as the ultimate metabolic end product with only trace amounts of intact [111In]DOTA-Fab found at the earliest time points. Analysis of the urine shows appreciable amounts of [111In]DOTA-Fab is filtered without degradation at early time points, but later on only [111In]DOTA-amino-lysine is found. Additionally, analysis of urine samples from a patient injected with [111In]DOTAchT84.66 showed the identical end product, thus confirming that this conclusion can be extended to other species and intact antibodies. In earlier studies, Schott et al. (1992) conjugated a benzyliosthiocyanto derivative of DOTA to a recombinant scFv version of the antitumor antibody CC49, radiolabeled it with 177Lu and compared its biodistribution to an 125I-labeled version. At 1 h the kidney uptake was
Metabolism of [111In]DOTA-mAb Fragments
Bioconjugate Chem., Vol. 12, No. 2, 2001 269
Figure 5. Anion-exchange chromatography analysis of patient urine sample. Urine sample from a patient being imaged with [111In]DOTA-chT84.66 was analyzed on a Poros-HQ column. The urine sample corresponded to 24-48 h after administration of the pre-therapy [111In]DOTA-chT84.66 antibody. An aliquot (200 µL) was applied to the column. The elution positions of [111In]DOTA and [111In]DOTA--amino-lysine are indicated by arrows.
(Lamki et al., 1990) showed that chemically stabilized F(ab′)2 fragments have low kidney uptake. An alternative approach to lower renal uptake has been to attach chelates to antibody fragments via a metabolically labile linker such as cleavable ester linkers (Meares et al., 1988; Paik et al., 1989; Weber et al., 1990; Arano et al., 1996). However, the results have been uniformly disappointing, giving only a modest reduction in kidney retention, and sometimes leading to cleavage within the blood. While our study showed that [111In]DOTA-Fab has high kidney uptake, injection of the metabolite [111In]DOTA-amino-lysine, results in excretion directly into the urine. These data suggest that [111In]DOTA--amino-lysine produced by metabolism of radiolabeled DOTA-antibody in other tissues would clear rapidly from the blood and not be retained in the kidney. Thus, the accumulation in kidney occurs at the level of the antibody fragment and failure to excrete the majority of the metabolic product produced there (e.g., see Figure 1). On the basis of these results, we are now designing DOTA-based chelates that are expected to have enhanced kidney clearance. Our strategy is based on the peptide approach of Arano and co-workers (Wakisaka et al. 1997; Arano 1998; Arano, Fujioka et al. 1999) and Meares and co-workers (Li and Meares 1993; Peterson et al. 1999) who have appended their radiolabeled moiety to peptides. This approach promises to decrease renal uptake and/or increase the rate of metabolism in the kidney. 111In-labeled
Figure 4. Anion-exchange chromatography analysis of kidney homogenates and urine samples. Kidney homogenates and urine samples taken at 5 h from mice injected with 10 µCi of [111In]DOTA-Rituxan Fab were analyzed on a Poros-HQ column and compared with markers. (A) [111In]DOTA--amino-lysine. (B) [111In]DOTA. (C) Kidney homogenate, 5 h. (D) Urine, 5 h.
>200% ID/g for the [177Lu]DOTA-scFv and slowly dropped to >150%ID/g by 48 h. As expected, the 125I-labeled version had low kidney uptake (10% ID/g at 1 h and
