Radioimmunoscintigraphy of Human Met-Expressing Tumor Xenografts Using Met3, a New Monoclonal Antibody¹

Met, the protein product of the c-met proto-oncogene, was discovered nearly 20 years ago (1, 2, 3). Met is a receptor tyrosine kinase, like the receptors for insulin and epidermal growth factor. It acts as the cell surface membrane receptor for HGF/SF.³ Under normal conditions, Met mediates molecular signaling pathways responsible for a broad range of cellular phenomena, including differentiation, motility, proliferation, apoptosis, organogenesis, and angiogenesis (4). Transfection experiments and retrospective analysis of various human solid tumors have established that aberrant expression of Met-HGF/SF in neoplastic cells leads to emergence of an invasive/metastatic phenotype. Neoplasms with reportedly abnormal Met-HGF/SF expression encompass cancers originating in the head and neck, thyroid, lung, breast, stomach, liver, pancreas, colon and rectum, kidney, urinary bladder, prostate, ovary, uterus, skin, bone, muscle, and other connective tissues (4, 5, 6). Inherited and sporadic activating mutations in Met, associated with human renal papillary carcinomas and other cancers, are compelling arguments for the role of Met in human cancer (7).

The level of Met-HGF/SF expression by cancer cells correlates inversely with clinical outcome. This correlation has been most thoroughly examined for human breast carcinomas. Met overexpression in breast tumors is associated with breast cancer progression (8, 9, 10). High HGF/SF expression also correlates with poor survival in ductal breast carcinomas (11, 12). Four independent laboratories have reported aberrant expression of Met by about one-half to two-thirds of localized prostate cancers, but evidently by all osseous metastases (13, 14, 15, 16), suggesting that Met provides a strong mechanism of selection for metastatic growth in prostate cancer. A similar correlation holds for human nasopharyngeal carcinoma, a neoplasm infrequently encountered in the western hemisphere but common in southeast Asia (17).

We are evaluating molecular imaging and therapeutic approaches that target the Met-HGF/SF complex, using animal models as a prelude to agent development for clinical use (18, 19, 20). Successful agents would allow us to stratify and treat tumors based on their patterns of Met-HGF/SF expression, independent of their tissues of origin. In this study, we begin to test the validity of that concept.

In a previous study, we found that a mixture of mAbs reactive against Met-HGF/SF can be used for radioimmunoscintigraphy of tumors autocrine for human Met-HGF/SF (19). We have now examined the ability of anti-Met mAbs from a single clone—designated Met3—to image human Met-expressing tumors of four different tissue origins and to distinguish them according to their relative abundance of Met.

¹²⁵I was purchased as NaI [480–630 MBq (13–17 mCi) per microgram of iodine] from Amersham Corp. (Arlington Heights, IL). C-28 rabbit polyclonal antibody reactive with the COOH-terminal portion of human Met and H-235 rabbit polyclonal antibody reactive with β-tubulin were purchased from Santa Cruz Biotechnology, Inc. The Alexa 488-conjugated antimouse antibody was purchased from Molecular Probes. Immunodecoration reagents were purchased from Amersham Pharmacia Biotech.

S-114 cells (NIH 3T3 cells transformed with human HGF/SF and human Met; Ref. 21), SK-LMS-1/HGF cells (a human leiomyosarcoma cell line autocrine for human Met and human HGF/SF; Ref. 22), PC-3 cells (human prostate carcinoma cell line), and the human melanoma cell lines M14-Mel and SK-MEL-28 were all maintained in DMEM containing 10% FBS.

Female athymic nude (nu/nu) mice, about 6 weeks of age, received s.c. injections of S-114, SK-LMS-1/HGF, or PC-3 cell suspensions in the posterior aspect of their right thighs, or of melanoma cell suspensions in the right flank adjacent to the thigh. Each mouse received between 2 × 10⁵ and 5 × 10⁵ cells. Tumors developed for 1–6 weeks before imaging, reaching ≥0.5 cm in greatest dimension by external caliper measurement. Mice were housed in small groups and given ad libitum access to mouse chow and drinking water under conditions approved by the institutional animal care committees.

The cultured cell lines listed above were analyzed for relative abundance of Met by immunoblotting with minor modifications of the procedures described previously (23). In brief, cells were grown to near-confluency in DMEM containing 10% FBS. Cell lysates were prepared, clarified, and assayed for protein concentration. Normalized aliquots of cell lysates were subjected to SDS-PAGE, electrotransfer, and sequential immunodecoration with C-28 anti-Met polyclonal antibody and with anti-β-tubulin polyclonal antibody. Immune complexes were identified by enhanced chemiluminescence and visualized by exposure to X-ray film.

mAbs against the extracellular domain of human Met were produced and screened for reactivity as described previously (19). Antibodies from the hybridoma clone 2F6 were identified as exhibiting the highest affinity for Met by ELISA and the highest apparent affinity for the human Met extracellular domain by immunofluorescence (19). The antibodies from clone 2F6, used for the experiments described here, are designated Met3. For the experiments summarized in Fig. 5, the positive control mAb Met5 was prepared in similar fashion to Met3, and the irrelevant, negative control mAb LF7 was raised against the purified lethal factor subunit of Bacillus anthracis lethal toxin.⁴ Met3 is of antibody type IgG_2b, Met5 of type IgG₁, and LF7 of type IgG₁.

Immunohistochemical analysis of Met expression and distribution in formalin-fixed, paraffin-embedded sections of human tissues was performed as described elsewhere (16), modified as follows: tissue sections on microscope slides were incubated with Met3 and processed with the Ventana automated system. Slides were examined by conventional light microscopy.

Immunofluorescent analysis of Met expression in cultured cells was performed essentially as described previously (19), incubating fixed cell monolayers with Met3, followed by FITC-conjugated antimouse IgG and with C-28 polyclonal antibody, followed by rhodamine-conjugated antirabbit IgG, and visualizing staining patterns with appropriate fluorescence optics and filter sets.

FACS analysis of Met3 binding to cultured human prostate carcinoma cell lines was performed with a Becton Dickinson FACS Calibur instrument. Cultured cells were grown to near-confluency, detached and dissociated by chelation, and resuspended at about 10⁶ cells/0.1 ml in BSA-containing buffer. The cell suspensions were incubated with Met3 (10 μg/ml) for 30 min at 4°C, washed three times, incubated with secondary antibody (antimouse Alexa green; Molecular Probes) for 15 min at 4°C, and washed three times before analysis.

For nuclear imaging experiments, IgG fractions were purified from 2F6 (Met3) hybridoma cell line supernatant fractions by protein G affinity chromatography and adjusted to a final concentration of 2 mg/ml in 0.25 sodium phosphate buffer (pH 6.8–7.0). The purified IgG fractions were stored frozen in small aliquots (25–50 μg) and thawed just before radioiodination.

Met3 was radioiodinated by the procedure described previously (19). Briefly, to 25 μg of Met3 in 0.1 ml of 0.25 m sodium phosphate (pH 6.8) was added 74 MBq (2.0 mCi; 0.02 ml) of ¹²⁵I as sodium iodide and 20 nmol (0.01 ml) of chloramine-T. The reactants were mixed and agitated gently for 90 s at room temperature. The reaction was quenched by the addition of 42 nmol (0.02 ml) of sodium metabisulfite. I-125-Met3 was separated from nonreacted ¹²⁵I by ion exchange on a small column of Bio-Rad AG 1 × 8 resin, 50–100 mesh. The recovered product was stored at 4°C until used and injected within 24 h of labeling. Radiolabeling efficiency was determined in a Beckman Gamma 8000 counter, and the proportion of protein-bound ¹²⁵I in the final product was assessed by chromatography on ITLC-SG strips (Gelman) developed in 80% aqueous methanol. Assuming complete recovery of mAb from the labeling mixture, radiolabeling efficiency was >60%, and protein-bound radioactivity accounted for ≥90% of total activity in the final product. Purified IgG fractions of control antibodies Met5 and LF7 were radioiodinated in similar fashion.

Animals were imaged, and scintigrams were analyzed by methods we have reported previously (19, 24, 25, 26). In brief, each mouse received I-125-Met3, 50–100 μCi (1.8–3.7 MBq) in ∼50 μl i.v., by tail vein injection under light inhalation anesthesia. Just before each imaging session, each mouse was given up to 13 mg/kg xylazine and 87 mg/kg ketamine s.c. in the interscapular region. Posterior whole-body gamma camera images of each mouse were acquired beginning at 1–2 h after I-125-Met3 injection and again at 1 day, 3 days, and at least 5 or 6 days after injection. Sedated mice were placed singly or in pairs on top of an inverted camera head with a protective layer over the collimator and taped to the layer to maintain optimum limb extension. Images of ¹²⁵I activity were acquired on a Siemens LEM Plus mobile camera with a low-energy, high-sensitivity collimator. Acquisitions were obtained over a period of 15 min, during which we collected between 2 × 10⁵ and 3 × 10⁶ counts per total body image. Imaging with the control antibodies I-125-Met5 and I-125-LF7 was performed in similar fashion.

Relative activity was determined by computer-assisted ROI analysis for each tumor, for total body, and for appropriate background regions at each imaging time point. These data are expressed below as background- and decay-corrected activity ratios. Graphical and statistical analysis of the converted data used the program Excel (Microsoft).

We previously demonstrated that Met3 colocalizes with the commercially available polyclonal anti-Met antibody C-28 in cultured S-114 cells, a murine cell line transformed with human Met and human HGF/SF (19). In Fig. 1,A, we show that Met3 may also be used for immunohistochemistry of human tissues (e.g. prostate tissue) in formalin-fixed, paraffin-embedded tissue sections. In Fig. 1,B, we show that the pattern of staining for Met3 by immunofluorescence analysis in primary cultures of human prostate epithelial cells replicates that observed with C-28. Moreover, we have observed that Met3 binds to the surfaces of PC-3 and DU145 human prostate carcinoma cell lines, both of which express Met, but not to any significant level to the surface of LNCaP cells that express very little Met (Ref. 16; Fig. 1 C).

As shown in Fig. 2, the cell lines we selected for this study vary dramatically in their relative expression of Met when cultured in the presence of serum. Cell lysates normalized to cell protein were subjected to electrophoresis, electrotransfer, and immunodecoration with C-28 to assess the abundance of Met, and with anti-β-tubulin (as a control to verify comparable levels among the various cell lines of an irrelevant housekeeping gene product). Under these conditions, S-114 showed the highest abundance of Met, both as p170 precursor and mature p140 forms. The melanoma cell lines expressed very low levels of Met, with M14-Mel lower than SK-MEL-28. SK-LMS-1/HGF and PC-3 cells exhibited intermediate abundance of Met, with comparable levels of total Met (p170 plus p140), but with a lower ratio of p170 to p140 detected in PC-3 cells.

Fig. 3 shows serial total body gamma camera images of individual xenograft-bearing mice obtained between 1–2 h and 5–6 days after i.v. injection of I-125-Met3. A pair of simultaneously imaged host mice is depicted for SK-LMS-1/HGF xenografts. Activity is clearly visualized in the S-114 and SK-LMS-1/HGF xenografts at the earliest imaging session, with a faint asymmetry of hindlimb activity suggested initially in PC-3 xenografts. Tumor-associated radioactivity as a function of total body activity is most prominent in these three xenograft types by the third day after injection. Neither melanoma xenograft exhibited any qualitatively appreciable uptake or retention of radioactivity during the imaging sequence.

Fig. 4 shows graphical results of quantitative image ROI analysis, expressed in two forms. The top panel shows the estimated fraction of injected activity associated with xenografts of differing tissue origin as a function of time after injection. Each xenograft type exhibited the highest mean value for this function at the earliest imaging session, with respective maxima of 18.6 ± 2.1, 7.2 ± 2.2, and 5.4 ± 2.6% of the estimated injected activity for S-114, SK-LMS-1/HGF, and PC-3. The bottom panel displays the mean ratios of tumor:total body activity as a function of time after injection. For each xenograft type, the highest value for this function occurred at 3 days after injection, with respective mean values (±1 SD) of 0.32 ± 0.13, 0.15 ± 0.06, and 0.10 ± 0.04 for S-114, SK-LMS-1/HGF, and PC-3. M14-Mel or SK-MEL-28 accounted for ≤3% of injected or total body activity at any time after injection.

To confirm the specificity of the interactions between I-125-Met3 and Met-expressing tumor xenografts in vivo, we also imaged animals bearing SK-LMS-1/HGF xenografts with both positive and negative control mAbs, as shown in Fig. 5. Met5 (IgG1) is of a different isotype than Met3 (IgG2b) and binds to an epitope distinct from that recognized by Met3: both canine and human cells express the epitope recognized by Met5, whereas the epitope recognized by Met3 is expressed by human cells but not canine cells (data not shown). LF7, a monoclonal antibody reactive with anthrax lethal factor, was used as an irrelevant negative control antibody of the same isotype as Met5 (IgG1). The ratios of tumor:whole body activity as a function of time in the groups of animals imaged with I-125-Met3 and with I-125-Met5 were indistinguishable, whereas the mean tumor:whole body ratios for the group of animals imaged with I-125-LF7 were <0.05 at all time points.

We have recently reported that a mixture of monoclonal antibodies recognizing multiple epitopes of the human Met-HGF/SF receptor-ligand complex can be used for radioimmunoscintigraphy of autocrine tumor xenografts (19). Here, we extend our observations by demonstrating that Met3, the product of one hybridoma clone that recognizes a single epitope of the extracellular domain of human Met, is similarly effective for nuclear imaging. Our studies further suggest that Met3 could be useful for routine immunohistochemical analysis of formalin-fixed, paraffin-embedded sections of human tissue, for immunofluorescence analysis of primary human cell cultures, and for FACS-based analyses of human tumor cells, in particular for the evaluation of samples of normal and malignant human prostate tissues.

The data presented here, along with additional examples that we have not yet published, confirm that radiolabeled Met3 can be used to image Met-expressing human tumor xenografts of differing tissue origins. Moreover, the rank order of I-125-Met3 uptake and retention levels exhibited by different types of xenografts in vivo correlates directly with the rank order of relative Met abundance as assessed biochemically in the respective parent cell lines cultured in the presence of serum. Stated another way, even based on this limited data set, we might arbitrarily divide tumors into high, low, and intermediate Met3 uptake categories by nuclear imaging analysis and infer that those respective categories reflect high, low, and intermediate abundance of Met in the tumor cells.

The two tumor xenograft types that fall in the intermediate Met3 uptake category, SK-LMS-1/HGF and PC-3, show no statistically significant differences with regard to either of the ROI analysis functions we have shown here and by immunoblotting analysis of cultured cell lysates appear to have comparable total Met abundance (p170 plus p140). Nevertheless, both ROI analysis functions tended toward higher values in SK-LMS-1/HGF than in PC-3, perhaps because of the autocrine-mediated turnover of Met in the former. We, therefore, suspect that even minor differences in radiolabeled Met3 uptake and retention in vivo by cells with comparable total Met abundance may be attributable to differing rates of biological turnover of Met (23, 27). This possibility is supported by recent experiments comparing rates of I-125-anti-Met mAb clearance by additional types of xenografts in vivo with their responsiveness to HGF/SF stimulation in vitro.⁴

In conclusion, we have shown that the radioiodinated anti-Met mAb designated Met3 can be used to image human Met-expressing xenografts of different tissue origins. We further infer that, at least in this animal model, scintigraphy with radiolabeled Met3 distinguishes human tumor xenografts according to their levels of Met expression.