Int. 1. Cancer: 48, 457-462 (1991) 0 1991 Wiley-Liss, Inc.

Publication of the International Union Against Cancer Publication de I'Union lnternationale Contre 1e Cancer

BIODISTRIBUTION OF A MONOCLONAL ANTIBODY (RNL- 1) AGAINST THE NEURAL CELL ADHESION MOLECULE (NCAM) IN ATHYMIC MICE BEARING HUMAN SMALL-CELL LUNG-CANCER XENOGRAFTS' 0.c.BOER MAN^, E.P. MIJNHEERE~, J.L.V. BROERS3, G.P. vOOIJS2 and F.C.S. RAMAEKERS334 2Department of Pathology, University Hospital Nijmegen; and 3Department of Molecular Cell Biology, University of Limburg, P.O. Box 616, 6200 MD, Maastricht, The Netherlands. The purpose of this investigationwas to determine the targeting potential of the murine monoclonal antibody (MAb) RNL-I for human small-cell lung cancer (SCLC) in a nude mouse model. RNL-I is preferentially reactive with SCLC and lung carcinoids, and was classified as a cluster-I MAb as defined by the International Workshop on Small-Cell LungCancer Antigens. From the intercellular location of the target antigen and i t s reactivity with 3T3 cells transfected with nucleic acid sequences encoding for the neural cell adhesion molecule (NCAM), it was concluded that RNL-l is directed against NCAM. RNL- I was radiolabelled with either Iz5iodine or '"indium and injected into nude mice bearing NCI-HE2 SCLC xenografts. The biodistribution of the radiolabels was determined up to I20 hr post injection. Maximum tumour accretion for "'In-RNL-I was I1.8%ID/g and 6.5%ID/g for 1251-RNL-I.The accumulation of "'In-RNL- I could be visualized clearly by gamma scintigraphy without background subtraction techniques. Autoradiographs of whole-body sections from animals injected with '251-RNL-Ishowed that activity in the SCLC xenografts was mainly peripheral, suggesting that tumour uptake is dependent on the vascularization of the turnour tissue.

The subdivision of lung cancer into small-cell lung cancer (SCLC) and non-SCLC is of clinical significance, since initially SCLC, unlike the other forms of lung cancer, is highly sensitive to both chemotherapeutic agents and radiotherapy (Minna et al., 1985). However, in most cases recurring tumours develop, which in general are extremely therapyresistant. Therefore, there is continuous interest in the application of immunological techniques for the management of the disease. Monoclonal antibodies recognizing SCLC associated antigens have been developed in several laboratories (De Leij et al., 1985; Bemal and Speak, 1984; Waibel et al., 1987). In a recent International Workshop on Small Cell Lung Cancer Antigens (Souhami et al., 1988; Beverley et al., 1988) a cluster of antibodies, the so-called cluster- 1 antibodies, was distinguished which was characterized by the recognition of an antigen commonly present on the SCLC cell surface but absent in most non-SCLC. Later, this group of antibodies appeared to recognize NCAM epitopes (Patel et al., 1989). We have recently developed a monoclonal antibody (RNL- 1) recognizing NCAM, which was shown to detect SCLC (Broers et al., 1991). Since radiolabelled antibodies against SCLC cell surface antigens are potential tools for radio-immunodetection, and may have therapeutic potential if conjugated to drugs, toxins or radionuclides, we have investigated the possibilities of using the radiolabelled RNL-1 antibody to localize human SCLC xenografts in a nude mouse model. MATERIAL AND METHODS

Monoclonal antibodies The RNL-1 antibody (IgG, isotype) raised against the variant SCLC cell line NCI-H82 has been shown to be reactive with SCLC, lung carcinoids, renal cancers, and some sarcomas. The antibody was reactive in normal neuronal tissues, and endocrine glands, such as pancreatic islet cells, pituitary gland, and adrenal medulla. Expression was also observed in Leydig cells of the testis, in the thyroid and in smooth-muscle cells of the small intestine, colon and bladder (Broers et al., 1991). The anti-interferon MAb B140 (IgG, isotype), kindly provided by Dr. C. Pak (Centocor, Malvern, PA), was used as a control antibody in the animal experiments. Flow cytometric analysis The reactivity of RNL-1 with human blood cells was tested by flow cytometry. Blood mononuclear cells were isolated from buffy coats by density centrifugation on 1.077 g/ml Ficoll (Pharmacia, Uppsala, Sweden). Red blood cells remaining in the buffy-coat fraction were lysed with 20 volumes of lysing solution (155 mM NH,Cl, 10 mM KHCO,, 0.11 m~ EDTA). The white blood cells were washed with PBS. Following an immunofluorescence procedure in suspension (Broers et al., 1986), the cells were analyzed in an Ortho 50H flow cytometer (Ortho, Westwood, MA). Immuno-electron microscopy A suspension of 2.5 X lo6 NCI-H82 cells was subjected to a standard indirect immunoperoxidase procedure (Broers et al., 1986). The cells were then fixed in 3% glutaraldehyde (1 hr), washed in PBS, and postfixed in 1% osmium tetroxide (1 hr). Cells were dehydrated to 70% ethanol, stained with 0.5% uranyl acetate in 70% ethanol, further dehydrated towards 100% ethanol, and embedded in Epon 812. Ultra-thin sections were viewed with a Philips EM 300 electron microscope (Eindhoven, The Netherlands). Transfection studies 3T3 cells were transfected with a cDNA clone coding for the 125-kDa isoform of muscle NCAM (Gower et al., 1988). Reactivity of RNL-1 with transfected 3T3 cells was tested with an indirect immunofluorescence assay, using untransfected 3T3 cells as a control. Radiolabelling and quality control Ascitic fluid was generated by intraperitoneal (i.p.) inoculation of 5 X lo5 RNL-1-secreting hybridoma cells into pristane-primed BALBlc mice. RNL-1 was purified from ascitic

'Part of this work was presented at the Second International Workshop on Small Cell Lung Cancer Antigens, London 1990.

Cell line

The NCLH82 cell line (Carney et al., 1985) was grown in RPMI 1640-based culture medium (Gibco, Paisley, UK) supplemented with 13% newborn calf serum (Flow, Rickmansworth, UK).

4To whom correspondence and reprint requests should be addressed. Received: December 13, 1990 and in revised form February 1, 1991.

458

BOERMAN ET A L .

fluid using a protein A-Sepharose 4B column (Phannacia) (Ey et al., 1978). Antibody concentrations were quantitated by absorbance at 280 nm, assuming that 1 mg/ml reads an extinction of 1.38. The preparations were stored at - 20°C until use. The purified RNL-1 was radio-iodinated by either the iodogen method (Fraker and Speck, 1978) or the Bolton and Hunter method (1973). For the iodogen labelling procedure 100 pg of'RNL-1 in 50 mM phosphate buffer (PH 7.2) and 0.3 mCi 1251 (Amersham, Little Chalfont, UK) were added to glass tubes that had been pre-coated with 1,3,4,6-tetrachloro-3a,6a-diphenylglycouril(Pierce, Rockford, IL) (15 pg/lOO pl), and the reaction was allowed to proceed for 5 min. For the Bolton-Hunter labelling method, 400 pg of RNL1 in 0.1 M borate buffer, PH 8.5, was added to 1 mCi of dried ~251-B~lton-H~nter reagent (iodinated hydroxysuccinimide ester of 3-(4-hydroxyphenyl)propionic acid; Amersham) and incubated for 15 min at O"C, according to the manufacturer's instructions. Following both labelling reactions, the reaction mixtures were applied to a Sephadex G-25 column (Pharmacia) and eluted with PBS. The fractions containing the labelled antibody were pooled and used for in vivo studies within 24 hr. Purified anti-interferon MAb B140 was radiolabelled with 13'1 (Amersham), using the iodogen method as described above. RNL-1 was labelled with "'indium using the bicyclic anhydride of diethylenetriaminepentaacetic acid (cDTPA) as a bifunctional chelating agent (Hnatowich et al., 1983). Briefly, 3 mg of RNL-1 in 0.1 M NaHCO, (3 mg/ml) was incubated for 1 hr with a 5-fold molar excess of cDTPA. After removing unconjugated DTPA by Sephadex G-25 column chromatography, 100 p g of DTPA-conjugated RNL-1 in 0.1 M citrate buffer (PH 5.0) was incubated with 400 pCi "'InCI, (Amersham) for 20 min. The reaction mixture was eluted on Sephadex G-25 to remove unincorporated "'In, and the fraction of unbound '''In was determined by ITLC using 0.1 M citrate buffer, PH 5.0, as a solvent. Immunoreactivity of the labelled antibody preparations was checked in a competitive binding assay carried out as described (Boerman et al., 1990), using NCI-H82-coated 96-well plates as the solid phase. Briefly, after the plates were coated with 0.01% poly-L-lysine in PBS for 30 min, NCI-H82 cells (150,000 cellslwell) were spun onto the bottom of the wells, fixed in 0.025% glutaraldehyde in PBS (15 min, 20°C) and saturated with 0.5% BSA in PBS (1 hr, 20°C). After washing, the cells were incubated with 50 p1 of a fixed amount of L2sI-labelledRNL-1 (0.5 pglml) in the presence of various dilutions (10- ' to 10Ws mg/ml) of unlabelled RNL-1 for 4 hr at 20°C. After washing, the retained radioactivity was determined in a gamma counter. The affinity constant of the antibody was calculated as the reciprocal concentration of unlabelled MAb required for 50% displacement of the Iz5I-RNL-1 (Giacomini et al., 1985). Biodistribution studies Male nude BALBlc mice bearing subcutaneous (s.c.)human SCLC xenografts (NCI-H82) were used. Tumours were initiated by injecting 3 X lo7 NCI-H82 cells. When tumours had grown to approximately 1 cm in diameter, they were resected aseptically, minced into small pieces of 2 to 3 mm diameter and serially transplanted in anaesthesized animals. Biodistribution studies were initiated 4 weeks after transplantation, when tumours measured approximately 1.0 cm in diameter (0.5-2.0 8). NCI-H82 xenograft-bearing mice received a mixture of 10 pg of Iz5I-RNL-1 and 10 pg of 1311-B140in the tail vein. Another set of tumour-bearing mice received 10 p g of 'l'In-RNL-l. At 5 different time intervals after injection (4, 24,48,72 and 120 hr), groups of 3 mice were bled and killed under ether anaesthesia. Tumours and selected tissues were removed, washed, blotted dry, weighed and analyzed in a

or "'In activity. The results were gamma counter for IZsI, expressed as the percentage of the injected dose per gram of tissue (%ID/g), and the radioactivity in tumour as compared to that in normal tissues (tumourhon-tumour ratios). Radiolocalization indices (RI) were calculated for each tissue for each time point according to the formula: RI =

%ID/g of RNL-1 in tissue / %ID/g of RNL-1 in blood %ID/g of B140 in tissue / %ID/g of B140 in blood

Biodistribution experiments were done on 2 occasions to check the reproducibility of the determinations. Mice injected with "'In-RNL-1 were scanned 24, 48, 72 and 120 hr after injection with a single-head gamma camera with a parallel-hole, medium energy collimator (Siemens, Type Orbiter, Des Plains, IL). Digital acquisition was recorded on a Siemens Scintiview data processing system. Both 173keV and 247-keV gamma-ray peaks of II'In with symmetric 20% windows were used. Between 100,000 (24 hr) and 25,000 (120 hr) counts were acquired for the images. Animals were anaesthetized during imaging by i.p. injection of Nembutal. Whole-body section autoradiography The distribution of the radiolabel was visualized by wholebody-section autoradiography in tumour-bearing animals. Briefly, 2 and 3 days after injection of IZsI-RNL-1 animals were anaesthetized and snap-frozen as a whole in dryice-cooled isopentane. Frozen animals were embedded in 8% carboxymethyl cellulose and 30-pm-thick sections were cut at -2O"C, using a LKB-2250 PMV cryomicrotome (LKB, Stockholm, Sweden) as described by Rijntjes et al. (1979). Freeze-drying of the sections was carried out for 3 to 4 days at - 20°C and registration of radioactivity was made by exposure of desiccated sections on Kodak X-Omat AR films. RESULTS

Reactivity pattern of RNL-1 The reactivity of RNL-1 with the variant SCLC cell line NCI-H82 was studied immunocytochemically. Indirect immunofluorescence staining indicated that RNL- 1 is reactive principally with the tumour cell membrane (Fig. l a ) . Immunoelectron microscopy corroborated that the RNL- 1-defined antigen was localized on the cell surface, and especially at adherent junctions between adjacent cells, while cell-surface regions not in contact with each other showed a less intense staining reaction (Fig. 1b). Flow cytometric analysis indicated that RNL-1 was unreactive with various types of human blood cells (erythrocytes, monocytes, granulocytes, lymphocytes). RNL- 1 reacted positively in the immunofluorescence assay with the NCAM-transfected 3T3 cells, while the antibody did not react with untransfected 3T3 cells, confirming the suggestion that RNL-1 is directed against an NCAM epitope (experiments kindly performed by Dr. F. Walsh, London). Radiolabelling and quality control The method of radio-iodination of the purified RNL-1 antibody appeared to be critical and influenced its immunoreactivity markedly. According to the competitive binding assay on 0.025%-glutaraldehyde-fixed NCI-H82 cells, the antibody completely lost its immunoreactivity upon oxidative radioiodination (iodogen method). However, RNL-1 labelled by the Bolton-Hunter method was immunoreactive, suggesting that ester linkage of the 12sI-labelledBolton-Hunter reagent (80% incorporation) preserved the antigen-binding properties of the antibody. Fifty percent displacement of the IZsI-RNL-1 was obtained at an unlabelled antibody concentration of 6 pg/ml, indicating an affinity constant of 2.7 X lo7 M - I . Biodistribution studies Maximum tumour accretion of lZSI-RNL-1 was achieved at

TUMOUR TARGETING PROPERTIES OF A N ANTI-NCAM MONOCLONAL ANTIBODY

459

FIGURE1 - ( a )Immunofluorescencestaining pattern of RNL-1 in the NCI-H82 SCLC cell line. fbf Immuno-electron microscopy of NCI-H82 cells incubated with RNL-1. The antibody was detected by the immunoperoxidase technique. Scale bar: 10 km.

approximately 48 hr, reaching a level of 6.5 %ID/g (Table I, Fig. 2a). The activity in all of the normal mouse tissues decreased over time. The levels of activity in the normal tissues correlated to their blood contents. The lungs showed highest label uptake (2.5 %IDlg, 48 hr p.i.), and brain tissue showed the least uptake (0.2 %ID/g, 48 hr p.i.) (Table I, Fig. 2a). Blood clearance of the radiolabel was slow, and therefore tumour/blood ratio did not exceed 1.O during the first 3 days after injection. Comparing the biodistribution of the radio-iodinated RNL- 1 with the biodistribution of the anti-interferon MAb (I3lI-B 140) indicates that RNL- 1 specifically accumulates at the tumour xenografts. The radiolocalization indices (RI) for the tumour increased during the experiment from 1.8 at 4 hr to 4.3 at 120 hr (Fig. 3), while the RI of the normal mouse tissues ranged from 0.5 to 1.8 and were stable over time, except for the RI of the intestines, which decreased between 4 and 120 hr p.i. from 4.0 to 1.7. Tumour accretion was higher with the I ' 'In-labelled RNL- 1 in comparison with radio-iodinated RNL- 1. Tumour uptake reached 11.8% at 72 hr p.i. and remained at that level at least until 5 days after injection (Table IB, Fig. 2b). However, "'In

uptake in most non-tumour tissues, especially the spleen, liver, and kidney, was also higher as compared with the lz5I uptake, and persisted in these organs over time. Consequently, most tumoudnon-tumour ratios obtained with lzJI-RNL- 1 were much higher than those obtained with "'In-RNL-1. Blood activity for both radiolabels was in the same range. The selective tumour accretion of L1lIn-RNL-1 was confirmed by external scintigraphy, as shown in Figure 4. Tumours were already clearly visible at day 1, and apparently background activity was further reduced at days 2, 3 and 4. Imaging of 3 animals bearing different-sized tumours (0.27 g, 1.2 g , and 2.2 g) (Fig. 4a) with different tumour uptake levels, as determined 5 days after injection (14.5, 13.4 and 7.9 %ID/g respectively), showed that the smallest tumour is less photodense (Fig. 4b). The liver and probably the kidneys also showed enhanced lllIn uptake. Autoradiography The distribution of the activity 48 hr after the injection of 1251-RNL-1is shown in the autoradiography of the whole-body section shown in Figure 5. Highest activity is seen in the thyroid, bladder, and the tumour. Thyroid uptake is most likely a

TABLE I - BIODISTRIBUTION OF RADIOACTIVITY 4, 24, 48, 72 AND 120 HR AFTER INJECTION OF MICE BEARING NCI-HE2 XENOGRAFTS WITH '251-RNL-I' AND '111n-RNL-12 24 hr

4 hr

A.

Tumour Muscle

Spleen Liver Intestine Kidney Heart

Lung Brain

Blood B.

Tumour Muscle Spleen Liver Intestine Kidney Heart Lung Brain

Blood

4.2 ? 0.7 0.6 f 0.1 1.9 f 0.4 2.3 i 0.1 1.3 f 0.6 2.4 f 0.3 1.8 0.3 2.5 f 0.2 0.2 f 0.02 7.4 0.7

*

9.1 f 0.8 1.3 _t 0.1 9.5 f 0.9 20.0 f 0.2 3.6 f 1.6 13.1 2 0.7 4.0 f 0.4 4.0 ? 0.5 0.4 f 0.1 12.2 f 1.5

10.5 f 0.3 1.3 f 0.2 9.8 f 0.2 18.7 f 2.1 2.3 0.2 11.5 f 0.9 3.0 f 0.1 3.9 t 0.2 0.3 f 0.01 8.3 f 0.5

*

*

*

'Section A.-2Section B.-%ID/g 2

*

6.5 2.0 0.6 f 0.03 2.0 f 0.6 1.8 f 0.1 1 . 1 f 0.5 2.1 f 0.2 1.7 t 0.2 2.5 0.4 0.2 f 0.03 6.6 f 0.3

2.2 i 0.2 0.7 t 0.2 3.2 ? 0.2 3.6 ? 0.1 4.0 _t 1.9 5.0 ? 0.3 3.5 t 0.3 4.0 0.5 0.4 t 0.1 14.3 ? 0.8 3.8 ? 0.7 1.0 0.1 9.2 _t 1.9 22.0 i 2.9 2.8 t 0.3 12.2 f 1.5 6.5 f 0.8 8.2 i 0.7 0.6 t 0.1 28.7 f 0.8

48 hr

SD,

n

=

3

*

*

12 hr

5.6 0.5 1.0 0.8 0.6 1.4 1.1 1.8 0.1 5.3

120 hr

t 1.7 t 0.1 t 0.2 f 0.4 2 0.1 -t 0.2 f 0.2 t 0.2 ? 0.01 -t 0.9

4.1 rt 0.5 k 0.9 k 1.0 t 0.4 rt 1.1 rt 0.7 rt 1.4 -C 0.1 k 3.4 rt

11.8 f 0.2 1.0 f 0.2 9.2 1.9 16.6 2 1.3 1.8 t 0.5 9.4 f 0.9 2.6 f 0.3 3.3 f 0.5 0.3 2 0.03 5.2 2 1.5

11.6 0.9 -c 7.4 t 14.3 rt 1.0 rt 6.9 2.0 rt 2.6 t 0.2 rt 3.3 -C

*

1.1 0.1 0.4 0.4 0.1 0.3 0.2 0.6 0.01 1.2

* 1.6 *

0.1 0.3 1.2 0.2 0.5 0.2 0.04 0.03 0.6

460

BOERMAN ET A L

, to/g

vessels within the NCI-H82 xenograft. The distribution of activity within the tumour (Fig. 5 , c-j) was similar to the distribution of the blood vessels. The autoradiography of sections from animals frozen 72 hr p.i. were similar to those taken 48 hr p.i.

A

04 hrr I 24 hr. 0 48 hra I 72 hfa

n

120 hra

1- - - - .'

tumor murob aplwn 30

Ilver

kldnsy

lung

blood

Z IDIg

i

20

.........................................

16

.........................................

B

1

4hn 24hm 48 hra 72 hra

120 hra

tumor muack rp1-n

Ilnr

kMnoy

lung

blood

FIGURE2 - Tissue distribution of '251-labelled ( a ) and "'Inlabelled (b) RNL- 1 (IgG in NCILH82-xenograft-bearing nude mice 4.24, 48, 72 and 120 hr p i . The mean of 3 animals is presented.

.......

04 hra I 24 hrm 48 hra

12 hra 120 hrm

tumor

murclr

rplron

llnr

kldnoy

lung

FIGURE3 - Radiolocalization indices of lZ51-RNL-1compared with 13'I-B140 anti-interferon antibody in nude mice bearing NCLH82 xenografts ,

result of accumulation of free l Z 5 I formed after dehalogenation of the antibody in vivo, while bladder activity represents the '"I content of urine. Most strikingly, the tumour uptake is concentrated at the perimeter of the xenograft, while activity levels in the center of the tumour are much lower (Fig. Sa). Histological examination of the implants indicated that vascularization of the xenografts was minimal. A detail of the frozen tissue block (Fig. 5b) shows the distribution of the larger blood

DISCUSSION

We have presented the tumour-targeting properties of RNL1, a newly developed murine MAb raised against the variant SCLC cell line NCI-H82 (Broers et al., 1991). Because of its reactivity with normal and neoplastic tissues, RNL-1 was grouped among the cluster- 1 antibodies, as defined by the International Workshop on SCLC antigens (Souhami et al., 1988). Since the cluster-1 antibodies have optimal binding to small-cell carcinomas, neuroblastomas, sarcomas and some melanomas, they may be suitable vehicles for targeting drugs, toxins or radionuclides to these tumour types. Immunocytochemical studies on NCAM-transfected 3T3 cells confirmed the suggestion that RNL-1, like 15 other cluster-1 antibodies (Pate1 et al., 1989), recognizes an epitope on the 125 kDa isoform of the neural cell adhesion molecule. Flow cytometric analysis showed no reactivity of RNL-1 with most human blood cells, a prerequisite for in vivo use of MAbs. NCAM has been reported to be expressed on NK-cells (Abo etal., 1981); however, the interaction of RNL-1 with this (minor) sub-population of white blood cells was not detected primarily, because the flow cytometer was not set to detect NK-cells. Initially, our biodistribution studies with RNL- 1 were hampered because of the loss of immunoreactivity upon oxidative radio-iodination by the iodogen method. In contrast, radiolabelling by the Bolton-Hunter method or with lllIn using the cyclic anhydride, did not affect the immunoreactivity of the antibody. RNL- 1 radio-iodinated by the Bolton-Hunter method showed specific tumour accretion following i.v. injection into nude mice with NCI-H82 heterotransplants. The RNL-I antibody labelled according to the iodogen method injected in the nude mouse/NCI-H82 human SCLC model revealed similar tumour/blood ratios as obtained with the anti-interferon B 140 antibody (data not shown). These results suggest that tyrosine residues in the antibody molecule may be essential for antigen binding. Tumour accretion was higher with I1lIn-RNL-1 in comparison with 1z51-labelledRNL-I . This is consistent with previous studies (Brown et al., 1987; Andrew et al., 1988). Apparently, radio-iodinated antibodies are metabolized in the tumour and the free Iz5I label is rapidly removed via the blood (Halpern et al., 1987). In contrast, with lllIn-labelled antibodies virtually all of the activity is retained in the tumour. The comparison of the activity of both radionuclides in the normal tissues at consecutive time points after injection suggests that this phenomenon also occurs in non-tumour tissues. It has been suggested that the sub-cellular localization of the antigen is a critically important factor determining the accessibility of the tumour antigenic sites in tumour targeting (Matzku et al., 1988; Pervez et al., 1988, 1989), and thus its suitability for tumour targeting. Our immunocytochemical studies showed that on the NCLH82 cells the NCAM antigen is most abundantly expressed at adherent junctions between adjacent cells, and therefore NCAM may not be the most optimal target for antibodies. The limited accessibility of the RNL- 1-defined antigenic determinant at the NCI-H82 cells may account for the moderate tumour uptake and turnour/ non-tumour ratios obtained in our study as compared with other targeting studies in nude mouse/human tumour models (Smith et al., 1989; Jones et al., 1986; Sharkey et al., 1988). But nevertheless, in vivo the NCAM epitopes appear to be acces-

TUMOUR TARGETING PROPERTIES OF AN ANTI-NCAM MONOCLONAL ANTIBODY

46 1

FIGURE4 - External gamma scintigraphy following i.v. injection of 20 pCi 'l1In-RNL-1 in three nude mice bearing NCLH82 transplants of different sizes.

A

FIGURE5 - (a) Autoradiography of a whole-body section from a NCI-H82-xenograft-bearing nude mouse 48 hr after injection of L251-RNL-1 (20 FCi). The perimeter of the tumour shows more uptake of activity than the centre. Bladder (B) and thyroid (Th) also show enhanced uptake. (b)Histological appearance of an unstained cross-section through a tumour nodule, showing the distribution of the larger blood vessels. ( c - Autoradiographs of different levels in the tumour nodule shown in (b). Inset c shows the autoradiograph of the level in b.

sible to a degree allowing them to be used as a target in immunodetection procedures. In addition, the relatively low affinity of the antibody (2.7 x lo7 M-') might also limit the tumour accretion of the antibody. Further studies are needed to determine whether other SCLC antigens may be more suitable targets for antibodies in vivo. RNL-1 is reactive with a wide range of neuronal and endocrine tissues, including brain and peripheral nerve. No accumulation was seen in murine brain during our in vivo experiments. However, RNL-1 probably does not cross-react with murine NCAM, and the blood-brain barrier will also prevent the radiolabelled antibody from penetrating the brain tissue. Wilson et al. (1990) have shown in a rabbit model that the anti-NCAM antibody, LS2D6 17, specifically accumulates in antigen-positive normal tissues except brain tissue. The autoradiographs of the nude mice injected with radioiodinated antibodies not only visualized distribution of the radiolabel within the animal, but also the limited penetration of

the radiolabelled antibody into the tumour tissue, in agreement with other reports (Pervez et al., 1988; Fand et al., 1987). Apparently, after the antibody leaves the circulation and enters the interstitial fluid of the tumour, it principally binds to tumour cells in the immediate vicinity of the blood vessels and capillaries. Whether this is due to physiological barriers, such as increased interstitial pressure, or to avid binding of antigen remains to be determined. Our study shows the specific tumour accretion of radiolabelled RNL-1 in the human SCLC NCI-H82 nude mouse host model, suggesting the possible in vivo applicability of this antibody. However, it remains to be determined whether the expression of its target antigen (NCAM) in certain normal human tissues will limit its clinical usefulness. ACKNOWLEDGEMENTS

We thank Dr. D.N. Carney for providing the NCI-H82 cell line, andMrs. M. KleinRot, Mr. G. Schaart, Mr. 0. Moesker,

462

BOERMAN ET A L .

Mr. W. van den Broek, Mr. J. Koedam and Mr. G. Grutters for their skilled technical assistance. We are grateful to Dr. F. Walsh (London) and Dr. K . Pate1 (Bristol) for their assistance in determining the NCAM nature of the RNL-1-defined antigen, to Mr. C. Jacobs (Department of Nephrology, University of Nijmegen) for the flow cytometric analysis, and to Mr.

N. Rijntjes (Department of Clinical Pharmacy, University of Nijmegen) for the autoradiographic studies. Dr. R.M. Sharkey (Newark) is thanked for critically reviewing the manuscript. This work was supported by the Scientific Advisory Committee on Smoking and Health (Wetenschappelijke Adviesraad Roken en Gezondheid), The Netherlands.

REFERENCES

ABO,T. and BALCH,C.M., A differentiation antigen of human NK and K S.E., DICKSON,G. and WALSH,F.S., Alternative splicing generates a cells identified by monoclonal antibody (HNK-I). J . Immunol., 127, secreted form of NCAM in muscle and brain. Cell, 55, 955-964 (1988). 1024-1029 (1981). HALPERN,S.E. and DILLMAN,R.O., Problems associated with radioANDREW,S.M., PERKINS,A.C., PIMM,M.V. and BALDWIN,R.W., A immunodetection and possibilities for future solutions. J . b i d . Response comparison of iodine- and indium-labelled anti-CEA intact antibody, Modif., 6, 235-262 (1987). F(ab'), and Fab fragments by imaging tumour xenografts. Europ. J . nucl. HNATOWICH, D.J., CHILDS,R.L., LANTEIGNE, D. and NAJAFI,A,, The Med., 13, 59&604 (1988). preparation of DTPA-coupled antibodies radiolabeled with metallic radiBERNAL,S.D. and SPEAK,J.A., Membrane antigen in small-cell carci- onuclides: an improved method. J . immunol. Meth., 65, 147-157 (1983). noma of the lung defined by monoclonal antibody SMl. Cancer Res., 44, JONES,D.H., GOLDMAN, A , , GORDON, I., PRITCHARD, J., GREGORY, B.J. 265-270 (1984). and KEMSHEAD, J.T., Therapeutic application of a radiolabeled monoBEVERLEY, P.C.L., SOUHAMI, R.L. and BOBROW, L., Results of the cen- clonal antibody in nude mice xenografted with human neuroblastoma: tumoricidal effects and distribution studies. Int. J . Cancer, 35, 715-720 tral data analysis. Lung Cancer, 4, 15-36 (1988). (1986). BOERMAN, O.C., MAKKINK,W.K., THOMAS,C.M.G., HANSELAAR, W.H. and A.G.J.M., YEDEMA, C.A., KENEMANS, P. and POELS,L.G., Monoclonal MATZKU,S., TILGEN,W., KALTHOFF,H . , SCHMIEGEL, antibodies that discriminate between human ovarian carcinomas and be- BROCKER,E.B., Dynamics of antibody transport and internalization. Int. J . Cancer, Suppl. 2, 11-14 (1988). nign ovarian tumours. Europ. J . Cancer, 26, 117-127 (1990). E.J., Cancer of the lung. BOLTON,A.E. and HUNTER,W.M., The labelling of proteins to high MINNA,J.D., HIGGINS,G.A. and GLATSTEIN, specific radioactivities by conjugation to a 1251-containingacylating agent. In: V.T. DeVita, Jr., S . Hellman and S.A. Rosenberg (eds.), Cancer; Application to the radioimrnunoassay. Biochem. J . , 133, 529-536 (1973). principles and practices of oncology, 2nd Ed., pp. 507-597, Lippincott, Philadelphia (1985). G., LANE, BROERS,J.L.V., CARNEY, D.N., KLEINROT, M., SCHAART, G., ROSSELL, R.J., BEVERLEY, P.C., E.B., VOOIJS,G.P. and RAMAEKERS, F.C.S., Intermediate filament pro- PATEL,K., MOORE,S.E., DICKSON, J.T. and WALSH,F.S., Neural cell adhesion molecule teins in classic and variant types of small cell lung carcinoma cell lines: a KEMSHEAD, biochemical and immunochemical analysis using a panel of monoclonal (NCAM) is the antigen recognized by monoclonal antibodies of similar specificity in small-cell lung carcinoma and neuroblastoma. Int. J . Cancer, and polyclonal antibodies. J . Cell Sci., 83, 37-60 (1986). 44, 573-578 (1989). BROERS,J.L.V., MIJNHEERE, E.P., K L E ~ N ROT, M., SCHAART, G., PERVEZ,S . , EPENETOS, A.A., Moo], W.J., EVANS,D.J., ROWLINSON, SIJLMANS, A., BOERMAN, O.C. and RAMAEKERS, F.C.S., Novel antigens G . , DHOKIA,B. and KRAUSZ,T., Localization of monoclonal antibody characteristic of neuro-endocrine malignancies. Cancer, 67, 619-633 AUAl and its F(ab'), fragments in human tumour xenografts: an autora(1991). diograpbic and immunohistochemical study. Int. J . Cancer, Suppl. 3, BROWN,B.A., COMEAU, R.D., JONES,P.L., LIBERATORE, F.A., NEACY, 23-29 (1988). W.P., SANDS,H. and GALLAGHER, B.M., Pharmacokinetics of the mono- PERVEZ,S . , KIRKLAND, S.C., EPENETOS,A.A., MOOI, W.J., EVANS, clonal antibody B72.3 and its fragments labeled with either lZ5I of "'In. D.J. and KRAUSZ,T., Effect of polarity and differentiation on antibody Cancer Res., 47, 1149-1154 (1987). localization in multicellular tumour spheroid and xenograft models and its CARNEY, D.N., GAZDAR, A.F., BEPLER,G . , GUCCION, J.G., MARANGOS, potential importance for in vivo immunotargeting. Int. J . Cancer, 44, P.J., MOODY,T.W., ZWEIG,M.H. and MINNA,J.D., Establishment and 94CL947 ( 1989). identification of small-cell lung-cancer cell lines having classic and variant RIJNTJES,N.V.M., V A N DE PUTTE, L.B.A., V A N DER POL, M. and features. Cancer Res., 45, 2913-2923 (1985). GUELEN,P.J.M., Cryosectioning of undecalcified tissues for immunofluDE LEIJ,L., POPPEMA, S . , KLEINNULEND, J . , TERHAAR,A,, SCHWAN- orescence. J . immunol. Merh., 30, 263-268 (1979). P.E. and THE, T.H., Neuro-endocrine SHARKEY, DER,E., EBBENS,F., POSTMUS, R.M., PRIMUS,F.J., SHOCHAT, D. and GOLDENBERG, D.M., differentiation antigen on human lung carcinoma and Kulchitsky cells. Comparison of tumor targeting of mouse monoclonal and goat polyclonal Cancer Res., 45, 2192-2200 (1985). antibodies to carcino-embryonic antigen in the GW-39 human tumorEY,P.L., PROWSE,S.J. and JENKINS, C.R., Isolation of pure IgG,, IgG,, hamster host model. Cancer Res., 48, 1823-1828 (1988). and IgG,, immunoglobulins from mouse serum using protein A-sepbarose. SMITH,A., WAIBEL,R., WESTERA, G., MARTIN,A., ZIMMERMAN, A.T. Immunochemisrry, 15, 4 2 9 4 3 6 (1978). and STAHEL,R.A., Imrnunolocalisation and imaging of small-cell cancer FAND,I., SHARKEY, R.M., PRIMUS,F.J., COHEN,S.A. and GOLDEN- xenografts by the IgG,, monoclonal antibody S W A l l . Brit.J . Cancer, 59, 174-178 (1989). BERG,D.M., Relationship of radio-antibody localization and cell viability in a xenografted human cancer model as measured by whole-body auto- SOUHAMI, R.L., BEVERLEY, P.C.L. and BOBROW,L. (eds), Proceedings radiography. Cancer Res., 47, 2177-2183 (1987). of the First International Workshop on Small-Cell Lung-Carcinoma AntiFRAKER, P.J. and SPECK,J.C., Protein and cell membrane iodination with gens. Lung Cancer, Suppl. 4, 1-1 14 (1988). a sparingly soluble chloramide 1,3,4,6-tetrachloro-3a,6a-diphenyl- WAIBEL,R., O'HARA,C.J. and STAHEL,R.A., Characterization of an glucouril. Biochem. biophys. Res. Comm., 80, 849-857 (1978). epithelial and a tumor-associated human small-cell lung-carcinoma glycoGIACOMINI, P., NATALI,P. and FERRONE, S . , Analysis of the interaction protein antigen. Cancer Res., 47, 376&3770 (1987). between a high-molecular-weight melanoma-associated antigen and the WILSON,B . S . , PETRELLA,E., LOWE, S.R., LIEN, K., MACKENSEN, monoclonal antibodies to three distinct antigenic determinants. J . Immu- D.G., GRIDLEY,D.S. and STICKNEY, D.R., Radiolocalization of human nol., 135, 696-702 (1985). small-cell lung cancer and antigen-positive normal tissues using monoJ., MOORE, clonal antibody LS2D617. Cancer Res., SO, 3124-3130 (1990). GOWER,M.J., BARTON,C.H., ELSOM,V.L., THOMPSON,