Radioimmunoscintigraphy of Tumors Autocrine for Human Met and Hepatocyte Growth Factor/Scatter Factor

Abstract

Inappropriate expression of the c-met-protooncogene product (Met) and/or of its ligand, hepatocyte growth factor/scatter factor (HGF/SF), has been correlated with poor prognosis in a variety of human solid tumors. We are developing animal models for nuclear imaging of Met and HGF/SF expression in tumors in vivo. We radioiodinated a mixture of monoclonal antibodies (MAbs) that bind to human HGF/SF and to the external ligand-binding domain of human Met, and then injected the I-125-MAb mixture intravenously into mice bearing tumors either autocrine for human HGF/SF and human Met or autocrine-paracrine for murine HGF/SF and murine Met. Serial total body gamma camera images were obtained, and regional activity was determined by quantitative region-of-interest (ROI) analysis. Tumors autocrine for human HGF/SF and Met demonstrated significantly more rapid uptake and more rapid clearance of the I-125-MAb mixture than tumors expressing one or both murine homologues, reaching a mean tumor to total body activity ratio of > 0.3 by 1 day postinjection. We conclude that radioimmunodetection of tumors autocrine for human HGF/SF and Met is feasible with an I-125-MAb mixture reactive against the ligand-receptor pair.

INTRODUCTION

Many different radiopharmaceuticals are available for imaging neoplasms. They range from classical agents such as [I-131]sodium iodide, [Tl-201]thallous chloride, and [Ga-67]gallium citrate to highly selective positron-emitting reporter gene detection systems [1],[2]. One successful approach to tumor imaging and therapy has been to develop radiolabeled molecules that bind to specific cell surface components, as exemplified by the clinical success of such agents as Octreo-Scan^® for imaging and potentially treating neuroendocrine neoplasms, CEAScan^® and OncoScint^® for imaging colorectal and ovarian cancers, and the agents Bexxar^® and Zevalin^® for detecting and treating certain lymphomas. As a novel variation of that strategy, we are developing radiopharmaceuticals that may distinguish neoplasms according to their genotype and invasive/metastatic potential rather than by tissue of origin. Here, we describe the use of monoclonal antibodies reactive with the molecules hepatocyte growth factor/scatter factor (HGF/SF) and its receptor, the c-met-protooncogene product (Met), which are broadly implicated in human solid tumors [3],[4].

Met, a protein tyrosine kinase receptor, is a transmembrane protein expressed in a wide variety of tissues, but mostly on the surface of epithelial cells. Its intracellular domain tyrosine kinase activity activates a complex cascade of biochemical reactions. Its extracellular domain binds the ligand HGF/SF, which is normally produced by mesenchymal cells. In normal tissues, the binding of ligand by Met effects important events in cellular development, including induction of cell proliferation, differentiation, invasion, and motility [3],[4]. In neoplastic cells, the aberrant expression of Met and HGF/SF leads to the emergence of an invasive/metastatic phenotype. This conclusion is supported both by transfection experiments and by retrospective analysis of many types of human solid tumors, including cancers of the breast, prostate, brain, colon, bone and soft tissues, kidney, pancreas, liver, and thyroid [3],[4].

We have raised and characterized monoclonal antibodies (MAbs) against both human HGF/SF and the extracellular domain of human Met, and we have recently reported that a mixture of at least three MAbs with different epitopic specificities is required to block the activation of Met by HGF/SF in vivo [5]. In this brief report, we present our early results of imaging tumors autocrine for human HGF/SF and Met with a radiolabeled mixture of MAbs reactive against the ligand-receptor pair.

RESULTS

Immunofluorescent Characterization of Anti-Met MAb

We characterized the anti-Met extracellular domain MAb 2F6 by immunofluorescent analysis on S-114 cells expressing human Met, as depicted in Figure 1. S-114 cells fixed in acetone/methanol were stained with both MAb 2F6 (in green, Panel A) and the rabbit polyclonal anti-Met C-terminal peptide antibody C-28 (in red, Panel B). Colocalization of staining (yellow) is evident in Panel C. A Nomarski image is provided (Panel D) to show the unstained location and characteristics of the cells in culture.

Image Analysis and Quantitation

Figure 2 shows serial total body gamma camera images of individual tumor-bearing mice obtained between 1 hr and 5 days following intravenous injection of the I-125-MAb mixture reactive with human HGF/SF and human Met. Activity is evident in the “human” tumors (SK-LMS-1 and S-114, both autocrine for human HGF/SF and human Met) as early as 1 hr postinjection and prominently thereafter. Activity is also clearly seen as early as 1 day postinjection in “murine” tumors (M-114, expressing murine HGF/SF and murine Met, and DA3, expressing murine Met alone). Nevertheless, mice bearing “human” tumors show more rapid clearance of radioactivity from the circulation than mice bearing “murine” tumors, as evidenced by their much lower levels of visceral activity at 3 and 5 days postinjection and more conspicuous thyroid activity (reflecting uptake of free radioiodide released by MAb deiodination). Even though the “human” and “murine” tumors generally appear to reach comparable levels of absolute activity over time, the proportion of nonthyroidal total body activity associated with “human” tumors, i.e., the tumor imaging contrast, appears to be greater than that associated with “murine” tumors at all imaging time points.

Figure 1.

Immunofluorescent analysis of MAb raised against human Met. S-114 cells fixed in acetone/methanol were labeled with anti-Met MAb 2F6 followed by FITC-conjugated anti-mouse IgG (green, Panel A) and with anti-Met rabbit polyclonal antibody C-28 (Santa Cruz Biotechnology) followed by rhodamine-conjugated anti-rabbit IgG (red, Panel B). Panel C confirms colocalization (yellow) of the antigens recognized by the MAb and polyclonal antibody. Panel D is a Nomarski-Differential Interference Contrast image of the cells.

To quantify these apparent differences and to determine whether they might be statistically significant, we examined the images from four mice bearing “human” tumors and from three mice bearing “murine” tumors by region-of-interest (ROI) analysis. The resulting data are summarized in Figure 3. Indeed, by t testing, the mean ratio of tumor activity to total body activity (including thyroid), designated Tt:WBt, was significantly higher for “human” than for “murine” tumors at all imaging time points (p #x003C; 0.02 at 1 hr; <i>p ≤ 0.001 after 1 hr), reaching mean values for these small groups of animals of 0.34 versus 0.11 at 1 day and of 0.37 versus 0.23 at 3 days postinjection. Mean retention of total body radioactivity, expressed as WBt:WB1h, was also significantly lower after 1 hr for “human” tumors (p ≤ 0.001). Finally, although the mean retention of tumor-associated activity (Tt:T1h) was lower in “human” than in “murine” tumors after 1 hr postinjection, this difference was not statistically convincing for the small number of animals studied (p = 0.3 at 1 day; p #x003C; 0.08 at 3 and 5 days).

Figure 2.

Total body images of tumor-bearing mice injected with an I-125-MAb mixture reactive with human HGF/SF and human Met. Each row of images contains serial total body scintigrams for a single tumor-bearing mouse injected with an I-125-MAb mixture reactive with the ligand-receptor pair human HGF/SF and human Met. The tumor hosted by each mouse is indicated to the left of its row. The time postinjection at which each image was acquired is indicated below each column. Images were obtained in posterior projection for the upper three rows, and in anterior projection for the mouse bearing DA3. The large arrows mark the transverse positions of respective tumors. Asterisks indicate the transverse positions of thyroids. The small arrow over the 1-day postinjection image for the mouse bearing DA3 indicates urinary bladder activity. Extracorporeal activity in the upper right corner of each scintigram for the mouse bearing M-114 represents a positional marker.

We chose to express ROI data as activity ratios rather than in a more traditional percent injected activity (%IA) format [11] in order to minimize the effects of variations in I-125-MAb intravenous injection efficiency on the data set. Technical factors make this variation potentially much greater in mice than in larger animal species with easier vascular access. In this way, we can use each animal's actual measured total body activity at the earliest imaging point as its own injection standard, rather than relying on a less accurate mean value for presumed injected activity. Moreover, assuming that no significant radionuclide excretion occurs during the first hour postinjection, the ratio of tumor activity to total body activity at 1 hr (T1h:WB1h) closely approximates %IA for a tumor at 1 hr, and the ratio Tt:WB1h similarly approximates %IA for a tumor at time t.

We have performed some preliminary negative and positive control studies to clarify the specificity of the I-125-MAb mixture's association with “murine” and “human” tumors, summarized as follows (data not shown):

Figure 3.

ROI comparison of tumors expressing human HGF/SF and Met versus tumors expressing murine HGF/SF and/or Met. Four mice bearing tumors autocrine for human Met and HGF/SF (three bearing S-114, one bearing SK-LMS-1) and three mice bearing tumors expressing murine HGF/SF and/or murine Met (two bearing DA3, one bearing M-114) were injected with an I-125-MAb mixture reactive against human Met and human HGF/SF. Tumor radioactivity (T) and whole-body radioactivity (WB) were quantified by ROI analysis of serial scintigrams obtained from 1 hr to 5 days postinjection, as depicted in Figure 2. Mean values (± 1 SD) are plotted for ratios of Tt:T1h (=ratio of T at time t to T at 1 hr postinjection), WBt:WB1h, Tt:WB1h, and Tt:WBt. Differences were significant between “human” and “murine” tumor-bearing mice for WBt:WB1h (p ≤ 0.001 after 1 hr) and for Tt:WBt (p #x003C; 0.02 at 1 hr; p ≤ 0.001 after 1 hr).</i>

1. The very ligand for Met, radioiodinated human HGF/SF, did not show convincing activity above that of blood pool in M-114 or DA3 tumors by 1 hr or at 24 hr postinjection.

2. An “aged” batch of the anti-Met and anti-HGF/SF I-125-MAb mixture (refrigerated for longer than 1 week and then repurified to remove liberated iodide) did not show convincing activity above blood pool in M-114 by 1 hr or at 24 hr postinjection, nor was “aged” I-125-anti-Met MAb alone an effective agent for imaging SK-LMS-1.

3. Tumor-imaging experiments using freshly labeled anti-Met MAbs and anti-HGF/SF MAbs separately indicate that both anti-Met and anti-HGF/SF contribute to the overall tumor-associated activity observed with the I-125-MAb mixture.

While each finding alone amounts to an imperfect control, taken together they argue that the levels and temporal patterns of tumor-associated activity that we observed in this study are somehow particular to the use of freshly radioiodinated anti-Met and anti-HGF/SF, and not to some nonspecific property of radioiodinated proteins in general.

Figure 4.

Potential mechanisms of radiolabeled MAb binding to tumor cells. Radiolabeled anti-Met MAb (*anti-Met) is depicted as binding directly to Met expressed on the tumor cell surface. Radiolabeled anti-HGF/SF MAb (*anti-HGF/SF) could either bind to free HGF/SF concentrated in the extracellular milieu, thereby surrounding tumor cells with radiolabeled soluble complexes, or form a ternary complex of MAb:HGF/SF:Met at the cell surface.

Finally, due to the prevalence of the Met-HGF/SF receptor-ligand pair in rapidly growing tumors, we have not yet identified a Met- and HGF/SF-nonexpressing tumor cell line to serve as a satisfactory “null-tumor” in vivo imaging control.

DISCUSSION

Our preliminary findings demonstrate that tumors autocrine for human HGF/SF and human Met can be imaged with an I-125-MAb mixture reactive against the ligand-receptor pair. Tumors expressing murine HGF/SF and/or murine Met can also be imaged with the radioiodinated MAb mixture, presumably because of epitopic crossreactivity, but the in vivo metabolism of the I-125-MAb mixture by “human” and “murine” tumors can be differentiated on kinetic and quantitative grounds. In brief, the “human” tumors that we have evaluated so far display rapid uptake and rapid clearance of the MAb mixture from the circulation, and constitute a significantly higher proportion of total body activity than the “murine” tumors from 1 hr to 5 days postinjection. Indeed, one would expect to observe just such differences between high-affinity, high-capacity tumors and those with lower affinity for binding and lower capacity for metabolizing a given radiotracer.

We began our imaging studies with a constituted mixture of MAbs reactive with the ligand-receptor pair, rather than a MAb with single epitopic specificity, because there is no a priori reason to pick one epitopic target over others in a tumor model that expresses both ligand and receptor. Moreover, we know that the anti-HGF/SF MAbs used in these studies map to different epitopes. As depicted in Figure 4, radiolabeled anti-Met MAbs should bind directly to Met molecules expressed on the tumor cell surface, while anti-HGF/SF MAbs may either bind to a locally concentrated pool of HGF/SF surrounding a cell or form a ternary complex with HGF/SF and Met, effectively targeting Met-expressing tumor cells in indirect fashion. We are currently determining how many and which of the epitopes targeted by this MAb mixture account for the observed imaging patterns. It may be that the neutralizing mixture of anti-HGF/SF MAbs somehow stabilizes Met so that anti-Met binds more readily than it would otherwise, or that any one of the MAbs in this mixture is sufficient on its own to image these tumors.

We present these preliminary findings from a small number of mice with the caveat that this is a highly contrived animal model using aggressive autocrine and paracrine tumor cell lines in animals with no endogenous human Met or human HGF/SF to obscure or compete for binding of the MAb mixture; we anticipate that nuclear images with such MAbs in human subjects who have spontaneous neoplasms will prove less dramatic.

If we can eventually develop radiolabeled MAbs capable of detecting Met- and/or HGF/SF-expressing tumors in humans, we may be able to offer patients with newly diagnosed cancers a novel sort of “metastatic risk stratification”—noninvasively assessing the likelihood as high or low that a given tumor will subsequently invade or metastasize. With such information, we might improve our ability to design appropriate monitoring and therapy protocols on an individual patient basis.

METHODS

Reagents

I-125 was purchased as NaI [480–630 MBq (13–17 mCi)/μg iodine] from Amersham (Arlington Heights, IL). C-28 rabbit polyclonal antibody reactive with the C-terminal portion of human Met was purchased from Santa Cruz Biotechnology.

Cell Lines and Tumor Induction

S-114 cells (NIH 3T3 cells transformed with human HGF/SF and human Met [6]) and M-114 cells (NIH 3T3 cells transformed with murine HGF/SF and murine Met) were grown in DMEM containing 8% calf serum. SK-LMS-1, a human leiomyosarcoma cell line autocrine for human Met and human HGF/SF [7], was maintained in DMEM containing 10% fetal bovine serum. DA3, a mouse mammary carcinoma cell line expressing murine Met [8], was grown in DMEM supplemented with 10% fetal bovine serum and antibiotics.

Female athymic nude (nu/nu) mice at about 6 weeks of age received subcutaneous injections of S-114, M-114, or SK-LMS-1 cell suspensions in the posterior aspect of their right thighs. Female BALB/c mice were injected with a DA3 cell suspension in their left inferior mammary fat pads. Each mouse received between 2 × 10⁵ and 5 × 10⁵ cells. Tumors developed for 3–6 weeks before imaging, reaching 1–2 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 and in compliance with the guidelines of the American Association of Laboratory Animal Science.

Production and Characterization of Monoclonal Antibodies

Neutralizing MAbs were raised against human HGF/SF and characterized as described in detail elsewhere [5].

MAbs against human Met were produced by injecting BALB/c mice intraperitoneally with 5 × 10⁶ 121-1TH-14 cells (NIH 3T3 cells transformed with human Met) in 0.5 ml phosphate-buffered saline (PBS), followed by three additional injections with the same quantity of cells. After 1 month, 1 × 10⁷ Okajima cells (human gastric carcinoma cells expressing Met [9]) in 0.5 ml PBS were injected intraperitoneally into each mouse. Spleen cells were collected 4 days after the final injection and fused with P3X63AF8/653 myeloma cells using standard techniques. Hybridoma cells were screened for reactivity to human Met by ELISA using 96-well plates coated with 0.5 μg/ml c-MET/Fc chimera (a fusion protein of human Met extracellular domain with human IgG1 heavy chain, R&D Systems, catalog number: 358-MT) in coating buffer (0.2 M Na₂CO₃/NaHCO₃, pH 9.6, 50 μl/well) overnight at 4 C. After blocking the plates with 200 μl per well of blocking buffer (PBS containing 1% BSA) for 1 hr at room temperature or at 4 C overnight, 50 μl of hybridoma supernatant were added to wells for 1.5 hr at room temperature. Plates were washed twice in washing buffer (PBS with 0.05% Tween 20), and alkaline phosphatase-coupled goat anti-mouse IgG (Sigma) was added (50 μl/well) at 1:2000 dilution for 1.5 hr at room temperature. After plates were washed four times with washing buffer, phosphatase substrate CP-nitrophenyl phosphate (Kirkegaard and Perry Laboratories) was added for 30 min and absorbance was measured at 405 nm. Hybridomas with strong reactivity with c-MET/Fc (OD value > 0.5, negative controls #x003C; 0.02) were recloned, and reactivity was again confirmed by ELISA.

To characterize the MAbs by immunofluorescence, S-114 cells (NIH 3T3 cells transformed with human HGF/SF and Met) and parental NIH 3T3 cells (serving as a negative control) in eight-well strips were fixed in either formaldehyde or acetone/methanol (50/50, v/v) for 10 min at room temperature, air-dried for 10 min, and blocked with blocking buffer (PBS with 1% BSA) for 30 min at room temperature. Purified anti-Met MAbs and control normal mouse IgG were diluted to 20 μg/ml with blocking buffer and added to either S-114 or control NIH 3T3 cells at 50 μl/well. After incubation at 37 C for 1 hr, strips were washed three times in washing buffer (PBS with 0.5% Tween 20). Cells were incubated with goat anti-mouse FITC conjugate at 1:20 dilution for 1 hr at 37 C, followed by three washes. Samples were observed by fluorescent microscopy, and the MAb showing strongest staining on acetone/methanol-fixed S-114 cells (designated 2F6), i.e., the one with highest apparent affinity for the human Met extracellular domain, was chosen for nuclear imaging.

IgG fractions were purified from 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 (50 μg) and thawed just prior to radioiodination.

For the experiments described here, equal volumes of the 2F6 anti-Met MAb and of a neutralizing mixture consisting of four anti-HGF/SF MAbs (designated A.1, A.5, A.7, and A.10) were combined to constitute a mixture reactive with the ligand-receptor pair.

Radioiodination and Injection of MAb Mixture

The final MAb mixture was radioiodinated following the recommendations of the radionuclide supplier. Briefly, to 25 μg of MAb mixture in 0.1 ml of 0.25 M sodium phosphate (pH 6.8) was added 74 MBq (2.0 mCi; 0.02 ml) of I-125 as sodium iodide and 20 nmol (0.01 ml) of chloramine-T. The reactants were mixed and agitated gently for 90 sec at room temperature. The reaction was quenched by the addition of 42 nmol (0.02 ml) of sodium metabisulfite. I-125-MAb was separated from nonreacted I-125 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 hr of labeling. Radiolabeling efficiency was determined in a Beck-man Gamma 8000 counter, and the proportion of protein-bound I-125 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 ≥ 85% of total activity in the final product.

Imaging Procedures and Analysis

Animals were imaged and scintigrams were analyzed by methods that we have previously reported [10],[11]. In brief, each mouse received the I-125-MAb mixture, 50–100 μCi (1.8–3.7 MBq) in ≤ 0.2 ml intravenously by tail vein injection under light inhalation anesthesia. Just prior to each imaging session each mouse was given up to 13 mg/kg xylazine and 87 mg/kg ketamine subcutaneously in the interscapular region. Anterior (for DA3 tumor-bearing mice) or posterior (for all other mice) whole-body gamma camera images of each mouse were acquired at 1 hr following I-125-MAb mixture injection and again at 1, 3, and 5 days postinjection. 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 I-125 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.

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 utilized the program Excel (Microsoft).