In Vivo Biodistribution of a Humanized Anti-Lewis Y Monoclonal Antibody (hu3S193) in MCF-7 Xenografted BALB/c Nude Mice Free

Despite advances in the treatment of breast cancer in terms of new chemotherapeutic and hormonal agents, the median survival of patients with advanced breast cancer has remained virtually constant over the last 30 years, and recurrent breast cancer remains incurable(1). New treatment strategies are required if there is to be any impact on the current plateau in survival. Conventional radiotherapy and chemotherapy are delivered to both normal and neoplastic cells, relying upon the enhanced sensitivity of rapidly dividing cancer cells to achieve preferential killing (2). In contrast,mAbs²can be directed or “targeted” against tumor associated antigens which may be expressed on the surface of the tumor cells(3).

A number of tumor antigens have been shown to have high expression on malignant breast cancer cells, including HER2/neu, MUC,carcinoembryonic antigen, and Le^y(3). The Le^y antigen is a family member of the blood group-related difucosylated oligosaccharides with chemical structure Fucα1→2Galβ1→4[Fucα1→3]GlcNAcβ1→R(4). They are associated with 60–90% of human carcinomas of epithelial cell origin including breast, colon,gastric, and lung cancer (5, 6, 7). The high frequency of Le^y-expressing tumors, high density of Le^y on tumor cell surface, and relatively homogeneous expression in primary and metastatic lesions have led to its selection as an antigenic target.

3S193 is a mAb produced using standard hybridoma techniques following immunization of BALB/c nude mice with Le^y-expressing MCF-7 breast cancer cells. The murine antibody has been shown to have high specificity for Le^y and reacted strongly in rosetting assays and cytotoxic tests with Le^y-expressing cells(8). The use of murine mAbs, however, is limited by their immunogenicity. Murine antibodies are recognized as foreign, and the immune system mounts a classical humoral immune response producing HAMAs. Although clinical HAMA requires rechallenge to the same antibody, serum sickness reactions can occur after the first or repeated infusions and are usually unrelated to dose or rate of administration and are manifested principally by myalgias and arthralgias (9). A further limitation of many murine antibodies is their inability to harness the effector and complement response of the human immune system as effectively as humanized antibodies, depending upon the isotype of the antibody(10).

CDR grafting is one method of humanizing antibodies, combining murine antigen binding regions (CDR) with human V-region framework determinants. CDR grafting aims to make the surfaces of the antibody appear as fully human as possible, but retain the murine antigen-binding packaging and interface reactions determining high-affinity binding (11). 3S193 has been humanized by CDR grafting, and the reactivity for Le^y has been confirmed to be similar to the murine version, and with retained potent effector function of complement-dependent cytotoxicity and antibody-dependent cellular cytotoxicity (8, 12). hu3S193 has only 3–5% murine residues in the antibody variable domain,markedly reducing the potential for an immune response in planned human trials.

The role of conventional radiotherapy in metastatic breast cancer has historically been limited to local palliation. Given that breast cancer is a radiation-sensitive disease (2), targeted delivery of radiation of effective dose to all metastatic sites using mAbs would be a novel and potentially useful therapeutic modality. Radionuclides are attractive agents for selective targeting to tumors by chemical conjugation to antitumor antibodies. The emitted radiation targeted to tumors can be used to kill tumor cells (therapy) or to detect primary or metastatic tumor (imaging), depending on the isotope selected. Considerations in selecting a radioisotope include type of emission,its range, tumor dose rate and total tumor dose, biological half-life,labeling efficiency, and stability. To be effective, the link between the antibody and the radionuclide must be stable in vivo,and the labeling procedure must not alter the biodistribution or binding characteristics of the antibody. This study explores such properties of hu3S193 bound to ¹²⁵I, ¹¹¹In, and ⁹⁰Y.

The generation of murine 3S193 using standard hybridoma technique after immunization of BALB/c mice with Le^y-expressing MCF-7 breast cancer cells and its subsequent humanization has been described previously (8, 12). hu3S193, a CDR grafted version of 3S193, was produced by Scotgen (Aberdeen, Scotland) in conjunction with the Ludwig Institute for Cancer Research. hu3S193 used for the experiments outlined was obtained from both the New York Branch, and the Biological Production Facility of the Melbourne Branch of the Ludwig Institute for Cancer Research. huA33, a humanized antibody directed against a novel antigen found in >95% of colorectal cancers (13), was supplied by the New York Branch of the Ludwig Institute and used as a subclass-specific control.

MCF-7, a Le^y-expressing human breast adenocarcinoma cell line originally derived from the pleural effusion of a 69-year-old woman with estrogen receptor-positive metastatic breast cancer (14), was obtained from the American Type Culture Collection (Rockville, MD). SW1222, a Le^y-negative human colonic cancer cell line, was a gift from the tumor cell bank of the New York Branch of the Ludwig Institute and was used as a control cell line.

Cells were grown in 175-cm² plastic flasks(Nunclon; Nunc, Roskilde, Denmark) and maintained in log-phase growth in RPMI 1640 (Trace Biosciences, Sydney, Australia) supplemented with 10% (MCF-7) or 5% (SW1222) FCS (MultiSer; Trace Biosciences, Sydney,Australia), 100 units/ml penicillin, 100μg/ml streptomycin, 0.25 ml/l insulin, 2 mm glutamine, and essential amino acids. Cells were cultured at 37°C in a 5% CO₂ incubator(Forma Scientific, Inc., Marietta, Ohio) and passaged with 0.05%EDTA/PBS (BDH Chemicals, Merck, Melbourne, Australia). Cell viability in all experiments, as determined by trypan blue exclusion, exceeded 90%.

hu3S193 and control antibody (huA33) were labeled with three isotopes, ¹²⁵I, ¹¹¹In, and ⁹⁰Y. Isotopes were obtained from DuPont (Life Science Products, Boston, MA). Radioiodination was performed using a modification of a previously published chloramine-T reaction(15) using a 2-fold molar excess of chloramine-T (Merck,Darmstadt, Germany) over antibody, dissolved in 0.5 mpotassium phosphate buffer (pH 7). After a brief 2-min incubation period, the reaction was stopped by adding a 5-fold excess of sodium metabisulfite, again dissolved in a 0.5 m phosphate buffer,and then purified through a desalting column (P6DG; Bio-Rad, Sydney,Australia) equilibrated with PBS.

Antibody labeling with ⁹⁰Y and ¹¹¹In was achieved via a bifunctional metal ion chelating agent, CHX-A″-DTPA (16) using a modification of a method published previously (17). In brief, a dialysis bag with a M_r cutoff of 10,000 (Novex,Sydney, Australia) was treated with 10 mm sodium bicarbonate and soaked in deionized water and 1 mm EDTA (pH 7.0) to remove heavy metals. hu3S193 was dialyzed against a 50 mm sodium bicarbonate buffer (pH 8.6) containing 0.15 m NaCl for 6 h. CHX-A″-DTPA was added in molar excess of 5:1 and incubated at room temperature overnight in the dark. Excess unbound antibody was removed by 8 h dialysis with 20 mm sodium acetate buffer containing 0.15 m NaCl (pH 6.3). ¹¹¹In or ⁹⁰Y were bound to CHX-A″-DTPA antibody conjugate under mildly acidic conditions (pH 5.5)for 20 min, and then the pH was raised to 7 by the addition of small aliquots of 2.0 m sodium acetate, followed by 10 mm EDTA. The radiolabeled mixture was purified by centrifugal desalting on a Sephadex G50 column (Pharmacia, Uppsala,Sweden) equilibrated with PBS.

Radiolabeling was performed on the day of injection into mice. Prior to injection, the percentage of unbound radionuclide content was determined by ITLC, and binding ability of the final radiolabeled product was tested by a Le^y-positive cell binding(Lindmo) assay as detailed below. Scatchard analysis was used to determine the binding constant (K_a)and number of antibody molecules bound per cell for ¹²⁵I- and ¹¹¹In-labeled antibody.

The amount of free versus bound antibody following radiolabeling was determined by ITLC as described previously (17, 18). Aliquots of radiolabeled hu3S193 were placed on the origin of standard silica gel impregnated glass fiber ITLC strips (Gelman Scientific, Ann Arbor, MI). To determine the percentage of bound ¹¹¹In, ITLC strips were developed using 10 mm EDTA (pH 4.5) and 0.9% saline/10 mm NaOH solution as solvents. In EDTA, hu3S193 remains at the origin, and the CHX-A″-DTPA-¹¹¹In or ⁹⁰Y, and free ¹¹¹In or ⁹⁰Y, respectively, migrate with the solvent front. In 10 mm NaOH solution, hu3S193 and free radionuclide remain at the origin, and CHX-A″-DTPA-radionuclide migrates with the solvent front. Strips were developed until the solvent migrated to the end, and then they were cut in half and counted on a gamma camera. Radioactivity in the bottom half of the strip was calculated as a percentage of the entire strip (top plus bottom). The percentage of free radionuclide was determined by subtraction(NaOH − EDTA). Results were performed in duplicate. Ten% trichloroacetic acid was used as the solvent to determine the percentage of unbound ¹²⁵I, with free iodine migrating with the solvent front and bound remaining at the origin. The percentage of isotope bound to hu3S193 was >90% in all experiments detailed.

IR of radiolabeled hu3S193 to MCF-7 target cells was determined by linear extrapolation to binding at infinite antigen excess using a“Lindmo” assay (19). Fifteen to 20 ng of radiolabeled hu3S193 were added to a range of MCF-7 cell concentrations and incubated for 45 min at room temperature with continuous mixing throughout to keep the cells in suspension. Cells were washed three times to remove unbound antibody, and pellets were measured in a gamma counter (Cobra II, Model 5002, Auto-gamma; Packard Instruments,Canberra, Australia). Three samples of radiolabeled hu3S193, at the same concentration as that initially added to the cells, were measured at the same time as cell pellets and the percentage of binding of hu3S193 to MCF-7 cells was calculated by the formula: (cpm cell pellet/mean cpm radioactive antibody standards) × 100. The percentage of binding was graphed against MCF-7 cell concentration using Sigmaplot for Windows (Jandel Scientific, San Rafael, CA), and IR was calculated as the Y-intercept of the inverse plot of both values. Cell pellets to which >100-fold unlabeled hu3S193 was added, prior to radiolabeled hu3S193, were used as a control to determine background cpm.

Scatchard analysis was used to calculate the apparent association constant (K_a) and number of antibody molecules bound per cell (19). Unlabeled hu3S193 at concentrations ranging from 5 to 200 μg/ml were added to 12 × 10⁶/ml MCF-7 cells and mixed, and subsequently 30 ng of labeled hu3S193 were added. After 45 min incubation, the cells were washed and counted as described above for Lindmo assay. The IR fraction was taken into account in calculating the amount of free, reactive antibody [(100 − %bound)/100 × total antibody × IR fraction], and specific binding (nm; total antibody × % bound) was graphed against specific binding/reactive free. The association constant was determined from the negative slope of the line. The number of hu3S193 molecules bound per cell was derived by: [(X-intercept of Scatchard plot(nm)/1000) × (6.023 × 10²³)]/number of cells used in the assay(12 × 10⁶).

In vivo tissue biodistribution studies were performed in female athymic BALB/c nude mice, 5–6 weeks of age, homozygous for the nu/nu allele, bred by the SPF Facility, University of South Australia (Adelaide, Australia). Mice were maintained in autoclaved microisolator cages housed in a positive pressure containment rack (Thoren Caging Systems, Inc., Hazelton, PA).

To establish MCF-7 human breast xenografts, mice were supplemented with exogenous estrogen (20, 21). After light Ethrane anesthesia, a 60-day slow release estrogen pellet (0.72 mg estradiol/pellet; Innovative Research of America, Sarasota, FL) was inserted using aseptic techniques into a small s.c. pocket fashioned between the shoulder blades. The surgical incision was closed with a single stitch (6-0 absorbable Dacron or silk). MCF-7 cells(25 × 10⁶) in 100–150 μl of medium were subsequently injected s.c. into the left inguinal mammary line. SW1222, non-Le^y-expressing, colon cancer xenografts were established by injection of 10 × 10⁶ cells in the right inguinal area of each mouse as a control tumor. The growth rates of both tumors differed. To obtain comparable tumor sizes, the more rapidly growing SW1222 colon cancer cells were injected 4 weeks after inoculation of MCF-7 tumors. Both tumors were of a size appropriate for use 14 days after this time.

Tumor volume was calculated by the formula [(length × width²/2] (21, 22), where length was the longest axis and width the measurement at right angles to the length.

Two biodistribution studies were performed. In the first study, 41 mice were injected i.v. via the retro-orbital plexus with hu3S193 labeled with 5 μCi of both ¹²⁵I and ¹¹¹In CHX-A″-DTPA-hu3S193 (total, 10 μg of antibody and 10 μCi of radioactivity) suspended in 150 μl of sterile media. Groups of four to five mice, with a mean (± SD) tumor volume of 647 mg ± 252, were sacrificed by cervical dislocation after light ethrane anesthetic at 4, 24, 48, 72, 96, 120,168, 240, and 360 h after injection of radiolabeled antibody. Mice were bled via cardiac puncture, and blood was collected into heparinized tubes. Tumors and organs [skin, liver, spleen, intestine,stomach, kidneys, brain, bone (femur), lungs, and heart] were immediately removed, blotted dry, and weighed (Sartorius Basic Balance,Ratingen Germany). All samples were counted in a dual chamber gamma scintillation counter (Cobra II,Auto-gamma; Packard Instruments) using a dual tracer program with standard windows set for each isotope, 15–75 Kev for ¹²⁵I and 140–430 Kev for ¹¹¹In. Standards prepared from the injected material were counted each time, with tissues and tumors enabling calculations to be corrected for the physical decay of the isotopes. Results of labeled antibody distribution over time were expressed as the % ID/g [(cpm tissue sample/cpm cpm standard) × 100/weight in grams] and as tumor:blood ratios.

A second biodistribution study was also performed to assess the distribution and localization of ⁹⁰Y-labeled antibody. Twenty μCi of ⁹⁰Y-CHX-A″-DTPA-hu3S193 were injected into the retro-orbital plexus of 25 mice bearing MCF-7 xenografts. One to four mice were sacrificed at time points 4, 24, 48,72, 120, 168, 240, 336, 408, and 504 h after injection and blood and tumors removed. At the 72-h time point, spleen, bone, liver, and kidney were also excised. % ID/g was calculated as described above.

Paired Student t tests were performed on the different isotope biodistribution time points in each study and individual organ uptakes to assess any differences observed for statistical significance.

Pharmacokinetics were determined for ¹²⁵I and ¹¹¹In radioconjugates using a curve-fitting program (SAAM II; University of Washington, Seattle, WA) and assuming a two-compartmental model.

The stability of the iodine and indium label in vivo was assessed by the amount of free radionuclide present over time, as determined by ITLC. Blood collected from mice at each time point into heparinized tubes was pooled and spun at 1000 × gfor 15 min to obtain plasma. Two μl of each sample were spotted onto the origin of two standard ITLC strips. Strips were developed, and the percentage of free indium and iodine was calculated as described previously.

A single point immunoreactivity assay was also performed on each pooled plasma sample to determine the in vivo stability of the antibody in terms of binding ability. An aliquot of plasma, volume adjusted to standardize for % ID/g, was added to 12 × 10⁶ MCF-7 cells and incubated for 45 min at room temperature. Excess unbound antibody was removed at the end of the incubation period by washing twice with medium, and radioactivity in the washed pellet was measured with a gamma counter. An equal volume of blood as that added to the assay was kept as a standard, and binding of blood to cells was calculated as a percentage of antibody present in the standard.

To enable qualitative comparison of radiolabeled-antibody localization over time, a single mouse from the ¹²⁵I/¹¹¹In co-labeled antibody study was anesthetized on four separate occasions. At time points 4, 24, 72, and 120 h, the animal received i.p. injections of 0.2 ml of solution containing 18.7 mg/ml tribromoethanol dissolved in 12.5 μl of isopentyl alcohol and 1 ml of sterile water for injection. Images were obtained using a dual head gamma camera(“Biad”; Trionix Research Laboratories, Twinsbury, OH)linked to a Unix computer system with photopeaks set at 173 KeV and 247 KeV with 20% windows. A medium energy collimater was used to acquire images attributable predominantly to ¹¹¹In, with negligible contribution from ¹²⁵I. Ten-min images were obtained at each time point. Images were acquired in a 256 × 256-bit matrix, and a standard of known activity was included in the field of view.

Immediately after sacrifice, an MCF-7 tumor from one mouse at time points 24, 48, 72, and 120 h in the ¹²⁵I/¹¹¹In biodistribution study was excised, frozen, and sectioned on a cryomicrotome (Zeis Microm HM 5000, Melbourne, Australia). Sequential 5-μm sections were used for direct autoradiography and for H&E staining. Tissue sections were placed on silane-coated glass slides and allowed to dry. Glass slides were placed face down in contact with emulsion of Hyperfilm-MP film (Amersham Life Science, Little Chalfont,United Kingdom) or standard radiography film. The glass slides and film were enclosed in a cassette and stored at −70°C. An image intensifier (Hyperscreen; Amersham) was used for Hyperfilm samples and was located directly behind and in contact with the film(23). Slides and film were taken up at varying time intervals, and film was developed in a standard automatic film processor. It was not possible to distinguish between ¹²⁵I and ¹¹¹In using the method described.

Immunohistochemistry was performed to confirm the presence and distribution of Le^y antigen in tumor xenografts and to examine normal murine tissues for Le^yantigen expression. Xenografts and normal kidney, liver, lung, spleen,stomach, duodenum, jejunum, ileum, and large bowel were surgically removed, embedded in Tissue-Tek OCT Compound (Diagnostic Division,Elkhart, IN), and cut at 5-μm thickness with a cryomicrotome. Sections were fixed with acetone at 4°C for 10 min, and nonspecific binding was blocked using a commercial protein blocking agent (Lipshaw Immunon, Pittsburgh, PA). Five μg of hu3S193 diluted in 1% BSA-PBS were applied to sections and incubated at room temperature for 1 h(primary antibody), rinsed with PBS, and then the secondary antibody(biotinylated IgG1 goat antihuman 1:300; Sigma Immunochemicals)was applied for 30 min. After 10 min immersion in 0.3% peroxidase,streptavidin peroxidase (1:1000) was applied for 30 min, and finally the chromagen 3-amino-9 ethyl carbazole (Sigma Chemical Co., St. Louis,MO) for 20 min was used to develop pink coloration to antigen-positive cells. An isotype-matched huIgG1 control antibody (huA33) and no primary antibody controls were used for each tissue or time point analyzed.

ITLC of radiolabeled antibody prior to injection confirmed 94.5% bound ¹²⁵I (6.5% free) and >95.5% bound ¹¹¹In-CHX-A″-DTPA. The immunoreactivity of ¹²⁵I was 51.5% and for ¹¹¹In, 65%. K_avalues for ¹²⁵I and ¹¹¹In were 0.71 × 10⁻⁷m⁻¹ and 1.017 × 10⁻⁷m⁻¹, respectively, both within the range expected for a carbohydrate antibody(24). The number of antibody molecules bound per cell for each isotope was 7 × 10⁶ for ¹²⁵I and 4.62 × 10⁶ for ¹¹¹In. In the ⁹⁰Y-CHX-A″-DTPA study, 94.1% of radionuclide was bound to hu3S193; however, immunoreactivity was low at 20%. Scatchard analysis was not performed for ⁹⁰Y-labeled antibody.

¹²⁵I/¹¹¹In-labeled biodistribution study results are presented in Table 1. The comparison of %ID/gram of ¹²⁵I- and ¹¹¹In-labeled hu3S193 in MCF-7 and control tumors, blood, and group size at each time point are detailed. The%ID/g of both isotopes in MCF-7 xenografts peaked at 48 h after antibody injection, with the uptake of ¹¹¹In-labeled hu3S193 being 3-fold greater than ¹²⁵I (30.1% versus 10%,respectively) and reaching statistical significance(P = < 0.001). Uptake of hu3S193 within breast tumors was superior for the ¹¹¹In label at all time points studied, and of note ¹¹¹In uptake was >3-fold that of ¹²⁵I at t = 48 h to 360 h inclusive (Fig. 1). The blood levels of both isotopes closely approximated each other throughout the study, declining progressively over time. The tumor:blood ratio for ¹¹¹In-labeled hu3S193 peaked at 5.15 ± 1.65 at 168 h, compared with 2.83 ± 1.2 at 120 h for ¹²⁵I.

The biodistribution of radiolabeled hu3S193 in normal murine tissues was studied, and results at 48 and 72 h after injection of radiolabeled hu3S193 are presented in Fig. 2. The % ID/g in all normal tissues was <10%. Immunohistochemical analysis demonstrated Le^y expression only in the gastric glands of the stomach and intestinal glands of the jejunum and ileum; this result was consistent with the low uptake of hu3S193 in normal tissues. The uptake of ¹¹¹In was only significantly higher than ¹²⁵I in the kidney(Ps between < 0.001 and 0.006, depending on the time point studied).

Antibody uptake (%ID/g) in SW1222 control tumors was less than that seen in antigen-positive MCF-7 tumors at all time points except 4 h. The higher % ID/g at 4 h in control tumors most likely reflects differences in tumor vascularity between MCF-7 and SW1222 xenografts, with increased vascularity observed in the control SW1222 tumors. Both ¹¹¹In and ¹²⁵I uptake in control tumors reflected blood pool activity rather than specific tumor uptake, declining progressively over time with no evidence of antibody accumulation within SW1222 tumor xenografts (Figs. 1,2,3,4).

Results of the ⁹⁰Y-hu3S193 study are presented in Table 2, and a comparison to the ¹¹¹In-hu3S193 biodistribution study is shown in Fig. 3. ⁹⁰Y-labeled antibody uptake in MCF-7 tumors peaked at 22% ID/g (t = 48–72 h), which was slightly lower but not statistically significantly different(P = 0.087) from peak ¹¹¹In tumor uptake. At the last time point at which a comparison was made between the two isotopes(t = 360 h), the %ID/g within MCF-7 tumors was 4-fold higher for mice that received the ⁹⁰Y label, suggesting prolonged retention of ⁹⁰Y within the tumor. Blood clearance for both isotopes approximated each other. At 72 h, hematopoietic organs and kidneys were removed from four mice. Comparison with the results obtained from the ¹²⁵I/¹¹¹In biodistribution study at the same time points after antibody injection showed higher uptake of ⁹⁰Y in kidney, spleen, and bone marrow(%ID/g 9.6, 6.02, and 1.99, respectively). Differences were statistically significant in the kidney compared with ¹²⁵I and ¹¹¹In(P = 0.001 and 0.038, respectively) and in the spleen compared with ¹¹¹In(P = 0.009). Differences in bone marrow uptake between the three isotopes were not statistically significant.

The half-life of each isotope bound to hu3S193 was estimated from the%ID/g of blood obtained over the time points sampled. Assuming a two-compartmental model with a four-parameter fit, the mean T_1/2 α for ¹²⁵I-, ¹¹¹In- and ⁹⁰Y-labeled 3S193 was 9.3, 7.4, and 7.5 h and mean T_1/2 β 69.3, 69.3, and 79 h, respectively.

In vivo stability was analyzed in the ¹²⁵I-hu3S193 and ¹¹¹In-hu3S193 study. The percentage of bound ¹¹¹In, as determined by ITLC, was >99.5% for all time points up to 7 days at which time 98% bound ¹¹¹In was present. The percentage of bound ¹²⁵I was lower at baseline (93.5%) and decreased to 78% at 7 days. Immunoreactivity for ¹¹¹In-hu3S193 and ¹²⁵I-hu3S913 at t = 0 h, as determined by single-point binding assay, was 65 and 51.5%, respectively, and slowly decreased over time, not dropping markedly until 7 days after injection, at which time the %IR for ¹¹¹In-hu3S193 was 27.5% and ¹²⁵I-hu3S913 was 27%.

Localization of ¹¹¹In-hu3S193 in a single BALB/c mouse bearing an MCF-7 and control SW-1222 tumor over time, as determined by gamma camera imaging, is presented in Fig. 4. At 4 h, no tumor localization was evident, with the image showing generalized blood pool activity and slightly increased activity in the central portion of the scan corresponding to the cardiac area. At 24 h, minimal uptake was seen in the control tumor (left side of image), but there was definite localization of antibody to the MCF-7 Le^y-expressing xenograft (right side of image). Over time with decreasing blood pool activity, the uptake in the MCF-7 tumor was more clearly defined. The mouse was sacrificed after the last image was obtained at 120 h. Radioactivity was clearly localized to the MCF-7 tumor, and the tumor:blood ratio was determined to be 6.07:1.

The distribution of radioactivity via autoradiography, the location of viable tumor cells using hematoxylin staining, and antigen distribution as determined by immunoperoxidase techniques were compared in an MCF-7 xenograft removed at day 5 in the ¹²⁵I/¹¹¹In-hu3S193 biodistribution study. Representative results are shown in Fig. 5. Hematoxylin staining indicated that the tumor contained viable tissue peripherally and a small amount in the center of the section, with a large area of central necrosis and connective tissue (Fig. 5,B). Antigen (Le^y) was distributed evenly throughout the section in areas corresponding to viable tumor cells (Fig. 5,C). Control slides for antigen staining showed some nonspecific staining in the connective tissue stroma but were negative in the tumor itself (Fig. 5,D). Radioactivity,assessed by autoradiography, was distributed throughout regions of viable tumor including the central area of viable tumor cells, with little activity in the central necrotic area (Fig. 5 A),indicating specific localization of radiolabeled hu3S193 to tumor cells. A similar even distribution of radioactivity throughout viable tumor cells was seen in all other time points studied (results not shown), and given differences in tumor sections, in particular the quantity of necrotic or connective tissue, there was no qualitative difference in distribution of radioisotope over time.

The role of mAbs in the treatment of breast cancer is currently under active investigation, with a number of unmodified antibodies,radio-immunoconjugates, and antibody-chemotherapy conjugates being studied in the laboratory and in clinical trials. Among these is a humanized antibody directed against the p185^HER2growth factor receptor (encoded by the HER-2/neuoncogene), which is overexpressed in 25–30% of all breast cancers (25). Phase II and Phase III studies of this antibody in patients with metastatic breast cancer showed an objective response rate of 12% (25), and higher response rates have been observed when combined with chemotherapy (26). This antibody has been recently approved by the Food and Drug Administration for use in patients with metastatic breast cancer overexpressing the HER-2 receptor.

In addition, a number of other Le^y antibodies have been developed and have reached different stages in preclinical evaluation and clinical testing. These include ABL 364 (BR55-2), an IgG3 antibody derived from the BR55-2 hybridoma (7). Murine ABL 364 administered to patients with cytokeratin-positive tumor cells in their marrow was shown to reduce or eliminate Le^y antigen-cytokeratin (+) cells but not cytokeratin (+) cells in Le^y-negative patients(27). However, a dose escalation study of murine BR55-2 in patients with metastatic breast cancer reported only minor clinical responses in 28.6% of patients, with HAMA limiting repeated cycles of treatment (28). A Phase I study with the LMB-1 immunotoxin, incorporating the murine anti-Le^ymonoclonal antibody B3 and recombinant Pseudomonas exotoxin,has also been performed in patients bearing solid epithelial tumors. Responses were observed in 5 of 38 patients studied; however, 90% of patients developed neutralizing antibodies against LMB-1 after one cycle of treatment (29).

Studies in nude mice bearing human lung carcinoma xenografts of another anti-Le^y antibody, BR96, demonstrated that murine and chimeric BR96 were able to prevent or slow growth when administered soon after cell implantation but had only modest effects with staged tumors (22). In contrast, when chimeric BR96 was conjugated to doxorubicin, complete regressions of xenografted human lung, breast, and colon cancers grown s.c. in nude mice were seen(30). In Phase I and II clinical trials of chimeric BR96(IgG1) conjugated to doxorubicin in patients with Le^y-positive advanced stage carcinomas, limited response rates were observed, whereas gastrointestinal toxicity, which improved with implementation of steroid premedication, was demonstrated(31, 32).

We have examined the in vivo biodistribution and labeling stability of the recently humanized anti-Le^yantibody 3S193, comparing ¹²⁵I-, ¹¹¹In-, and ⁹⁰Y-labeled hu3S193 in an MCF-7 xenografted BALB/c nude mouse model. Excellent and specific tumor uptake was observed compared with a Le^y-negative control tumor, with the peak tumor(MCF-7) uptake for all three isotopes occurring at 48 h. The%ID/g for both ¹¹¹In and ⁹⁰Y were significantly higher than ¹²⁵I. In addition, both ¹¹¹In and ⁹⁰Y were retained within xenografts for a longer period of time than ¹²⁵I. Lower absolute levels of iodine uptake compared with ¹¹¹In in xenografts have been demonstrated in other studies (13), and this together with differences in tumor retention of radiolabel most likely represents differences in cellular processing of radiolabeled ¹²⁵I as opposed to both ¹¹¹In and ⁹⁰Y. Catabolism of ¹²⁵I linked to tyrosine results in the generation of [¹²⁵I]mono-iodotyrosine within lysosomes, which rapidly leaves the cell via cell-mediated transport systems (18, 33, 34). In contrast, endocytosed ¹¹¹In-DTPA-antibodies are delivered to lysosomes and hydrolyzed by lysosomal enzymes into small molecular weight metabolites that are retained within tumor lysosomes (13, 33, 34).

Cellular processing of ⁹⁰Y by tumor cells in vitro has been shown in other antibody-antigen systems to be similar to ¹¹¹In (35) and is also reflected in similarities in distribution in vivo(36). It has been reported that although the pattern of uptake is comparable, superior tumor retention is obtained with the use of ⁹⁰Y (36), a conclusion that this current study supports.

Within normal tissues, the uptake of ¹¹¹I and ⁹⁰Y in liver, spleen, and kidneys was higher than ¹²⁵I. The higher uptake values in spleen and liver is consistent with known sequestration of radiometals in reticuloendothelial organs (36). This uptake in normal organs is not a consequence of specific targeting attributable to an absence of Le^y antigen expression in these normal tissues of the mouse but is a reflection of blood flow and catabolism of the radiolabeled conjugates. The increased uptake of ¹¹¹In- and ⁹⁰Y-labeled antibody compared with ¹²⁵I in the kidneys reflects differences in metabolism and excretion of radiometals(37). However, the increased renal uptake of ⁹⁰Y compared with ¹¹¹In may also reflect the conjugation of ⁹⁰Y-CHX-A″-DTPA-hu3S193, given the observed low immunoreactivity of the labeling experiment presented, and warrants further investigation using chelating agents of enhanced stability for ⁹⁰Y immunoconjugates (37). The%ID/g of ⁹⁰Y in the bone, assumed to represent bone marrow, was approximately double that of ¹²⁵I and ¹¹¹In; however,the total amount present was negligible (2% ID/g), and differences were not statistically significant. Other studies have shown that accumulation of ⁹⁰Y within bone is a concern with evidence that ⁹⁰Y accumulates within the bone marrow (36) and indicates that ¹¹¹In may underestimate ⁹⁰Y bone marrow uptake.

The differences observed between isotopes in radiolabeling techniques and cellular processing were also reflected in the in vivostability results. The percentage of ¹²⁵I-bound radioactivity over time, as determined by ITLC, decreased rapidly,consistent with rapid release of free ¹²⁵I from lysosomes. In contrast, only 2% of free ¹¹¹In was detected over the same time period, again consistent with current literature suggesting retention of ¹¹¹In-DTPA-lysine catabolites by tumor lysosomes(13, 33, 34). Retention of antigen binding for up to 5 days in vivo was also observed for both ¹²⁵I- and ¹¹¹In-labeled hu3S193.

Although promising, clinical responses with unconjugated mAbs in the therapy of solid tumors have been poorer and less consistent than with hematopoietic neoplasms (38, 39, 40), and there remains room for improvement. Given the high frequency of epithelial tumors expressing the Le^y antigen and the high density of Le^y on the surface of tumor cells, hu3S193 has potential therapeutic applications. The in vivo studies presented demonstrate that ¹¹¹In- and ⁹⁰Y-radiolabeled hu3S193 have favorable biodistribution characteristics, with high %ID/g uptake of both antibody bound isotopes within Le^y-expressing MCF-7 breast xenografts. In contrast, the biological behavior of ¹²⁵I-hu3S193, as demonstrated in this report,would indicate that ¹³¹I would not be the preferred isotope for this antibody-antigen system. Tumor localization of ¹¹¹In-CHX-A″-DTPA hu3S193 was confirmed visually in imaging studies and by autoradiography, demonstrating good penetration of radionuclide to all areas of viable tumor cells. In addition to the direct targeting of the tumor by hu3S193, our previous studies have described the in vitro immune effector functions mediated by the IgG1 Fc portion of hu3S193 (8, 12). The antitumor effect of hu3S193 may be limited in the murine model used in the current study because of restricted murine serum complement activity. The potent in vitroantibody-dependent cellular cytotoxicity of hu3S193 would suggest that,in conjunction with potent complement-dependent cytotoxicity activity in the human, these humoral effects in a clinical setting would augment the therapeutic effects of targeted radiolabeled hu3S193.

hu3S193 has the theoretical advantage of reduced immunogenicity provided by humanization of an antibody, an assumption that has been supported by preliminary data with earlier, fully humanized antibodies(24, 41, 42), allowing multiple courses of treatment that have not been possible with murine or chimeric antibodies. The humanization of 3S913 has yielded a molecule with only 3–5% sequence homology to the parent murine antibody in the immunogenic variable domains. A similar humanization, through CDR grafting, to that of 3S193 has been described for an antihuman CD4 mAb using the same human framework (REI for variable light chain and KOL for variable heavy chain; Ref. 42). Preliminary clinical results in patients have been reported and show a lack of immunogenicity (42, 43).

The humanization strategy we adopted for hu3S193 and the results of clinical studies using mAbs with similar human frameworks provide evidence of the reduced potential for an immune response to hu3S193 in planned clinical trials.

Some Le^y antigen expression was observed in the gastric glands of the stomach and basal intestinal glands of the small bowel of the BALB/c mice used in the current study, a similar distribution pattern to that reported previously in immunohistochemical studies of normal human tissues with other anti-Le^y mAbs (6, 44). However, the current biodistribution study did not demonstrate increased specific targeting to the small bowel of the BALB/c mice (Fig. 2 B). The expression of Le^y antigen in normal human tissues (including the GI tract) is of concern in targeting strategies against this antigen; however, the accessibility of normal tissues to anti-Le^y antibodies remains unclear. Results of studies with murine and chimeric antibodies against Le^y suggest that gastric toxicity may be relevant in proposed clinical trials; however, significant toxicity has only been observed at protein doses >200 mg (31, 32). The optimal protein dose for effective targeting of Le^y tumors (tumor:blood ratios, normal tissue uptake) is yet to be defined with carefully designed biodistribution and biopsy-based trials. In this context, high protein doses of anti-Le^y antibodies reported to cause toxicities may not be required for effective tumor targeting.

The current model has demonstrated the stability of hu3S193, and its characteristics and properties are well suited to beginning trials in the clinic.