Nucl. Med. Eiol. Vol. 18, No. 8, PP. 891-899, 1991 Int. J. Radiaf. Appl. Instrum. Parr B Printed in Great Britain. All rights reserved

0883-2897/9153.00+ 0.00 Copyright 0 1991Pergamon Press plc

Radioimmunolocalization of Human Colonic Cancer Xenografts; Aspects of Extensive Purification of Monoclonal Anti-CEA-antibodies A. SUNDIN]*, P. ENBLAD’, H. AHLSTRijM’,

J. CARLSSON3, E. MARIPUU’

and A. HEDIN” ‘Department of Diagnostic Radiology, Akademiska Sjukhuset, Uppsala University, S-751 85 Uppsala; rDepartment of Pathology, University Hospital, S-751 85 Uppsala; 3Division of Physical Biology, Department of Radiation Sciences, Uppsala University, Box 535, S-751 21 Uppsala; 4 Section of Hospital Physics, Department of Oncology, University Hospital, S-751 85 Uppsala and 5Pharmacia AB, S-751 82 Uppsala, Sweden (Received 12 September 1990; received for publication

13 Murch 1991)

Tumour-to-normal tissue ratios of i.p. injected ‘2sI-labelled monoclonal antibodies (MAbs), reacting with CEA were determined in nude rats xenografted with human colonic cancer cells (LS 174 T). Two MAbs, I-38Sl and 11-16, reactive with the GOLD I-epitope on CEA were tested. MAb I-38Sl was also tested after additional purification using anion exchange chromatography (thereafter named AEC 38). In the external activity measurements, MAb AEC 38 showed significantly better tumour-to-liver ratios than did MAb II-16 on all 4 days after injection. MAb I-38Sl gave intermediate ratios but was significantly better than II-16 only on day 3. The mean tumour-to-blood ratios were 3.0, 2.6 and 1.5 and the mean tumour-to-liver ratios were 6.6, 4.8 and 3.5 for MAbs AEC 38, I-38S1 and II-16 respectively. Gamma camera registrations in 3 animals on 4 days showed good imaging properties for all three MAbs and the patterns of tissue uptake were consistent with those seen in the external measurements. Furthermore, histopathological and immunohistochemical determinations were performed, showing that MAb II-16 gave about the same spatial binding as the previously analysed MAb I-38Sl. The results indicate that additional purification of MAbs using anion exchange chromatography may potentiate tumour uptake in this model

Introduction Diagnostic and therapeutic benefits of using radiolabelled monoclonal antibodies (MAbs) against tumour-associated antigens have, during the last few years, been gained both in experimental models and in clinical practice (Buchegger et al., 1989; Buraggi et al., 1987; Cheung et al., 1986; Epentos, 1984; Goldenberg et al., 1987; Mach et al., 1983; Sharkey et al., 1987). However, in order to become established as an alternative, or a complement to more conventional forms of therapy and imaging, MAb-guided methods will have to be improved by increasing uptake in the primary tumour and its secondaries, both in terms of absolute concentration and relative to other tissues (Epentos et al., 1986; Vaughan et al., 1987). An increase in the immunoreactivity of the MAbs may potentiate MAb uptake in the tumour. *Author for correspondence.

Most preparations of antibodies, despite being monoclonal, are heterogeneous in at least two First, they are contaminated with respects. immunoglobulin or other proteins from ascites or culture medium; secondly, MAbs have different isoelectric points. The fact that a microheterogenity in the distribution of charge in extensively purified MAbs exists was first described by Rosen et al. (1979). However, not all MAb preparations are similar in this respect. Some MAbs give 34 peaks, with regard to heterogenity seen by anion exchange chromatography, others only one. These peaks have, when tested in vitro, different immunoreactivity (unpublished results). The present investigation was undertaken to determine whether further purification with anion exchange chromatography of a well documented (Ahlstriim et al., 1987a,b, 1989; Hedin et al., 1982, 1986) anti-CEA MAb (I-38Sl) could increase the antibody uptake in human colonic cancer xenografted into nude rats. In the same tumour model, 891

A. SUNDIN

892

and in contrast to MAb I-38S1, we characterized the properties of another anti-CEA MAb, 11-16, which binds to a different epitope than I-38Sl although in the same epitope group (GOLD 1). This MAb has been studied earlier in vitro (Hammarstriim et al., 1989; Hedin et al., 1986; Larsson et al., 1987; Zoubir et al., 1990) but not in vivo.

Materials and Methods Tumour model Twenty male nude ruts (Rowett rnu/rnu, The Wallet&erg Laboratory, Lund, Sweden) 8-12 wk old and weighing 21&360 g were used. They were housed in an isolated room at a temperature of 25°C with a 12-h light, 12-h dark cycle. The animals were fed ad libitum with water and laboratory food (EWOS AB, Sweden). A carcinoembryonic antigen (CEA) producing human colonic adenocarcinoma cell line LS 174 (Tom et al., 1976) was used. The cells were grown in vitro in a nutrient mixture comprised of Ham’s F-10 (Flow Laboratories Swedish AB, Stockholm, Sweden) supplemented with 10% foetal calf serum, 2 mM L-glutamine, penicillin (20 UjmL) and streptomycin (20pg/mL). Trypsin-EDTA was added at 37°C for 15 min and the cells were suspended and centrifugated at 1200 rpm for 5 min. The pellet was suspended in culture medium without serum. Each rat was injected with 1.0 x 10’ cells in 0.1 mL of medium S.C. lateral, just above the knee in each hindleg (Ahlstrom et al., 1987a). Measurements of the ensuing tumour diameters were made prior to MAb-injection on days 10 (15 animals) and 14 (5 animals) after tumour cell injection, using callipers. Tumour volumes were calculated from the formula: V = 4/3 x rr x (A x B)3’2, where A = radius 1 and B = radius 2. For anaesthesia i.p. administered methohexital (Brietal, Eli Lilly & Company, Indianapolis, Ind., U.S.A.) 4.6mg/lOOg body weight was used. Monoclonal antibodies MAb I-38Sl and II-16 are mouse anti-CEA antibodies of the Ig Gl kappa isotype, produced against CEA from a human colorectal carcinoma and from a lung adenocarcinoma, respectively. MAb I-38Sl reacts with the D-CEA epitope (Hedin et al., 1982, 1986) and MAb II-16 with the H-CEA epitope (Hedin et al., 1986), belonging to the same CEA epitope group (GOLD 1) (Hammarstriim et al., 1989). Both MAbs precipitate CEA independently, indicating that the D and H epitopes are present at least twice in the CEA-molecule (Hedin et al., 1982, 1986). The procedures for hybridoma and MAb production are described elsewhere (Carlsson et al., 1985; Hedin et al., 1982). Briefly, the hybridoma cell line is derived from fusion of the mouse myeloma cell line SP 2/O-Ag 14 with BALB/c spleen cells from immunized mice. The cell line was grown in culture

et

al.

medium supplemented with 5% foetal calf serum, glutamine, sodium pyruvate and expanded to a density of about 2 x lo6 cells/ml, whereupon the medium was replaced with a serum-free one. The culture supernatant was harvested on the 12th day by filtration and then stored at -20°C. The purification of the MAbs was based on the method previously described by Carlsson et al. (1985) but was used here in a somewhat modified manner. The purification was effected using a Biopilot system (Pharmacia LKB Biotechnology group, Uppsala, Sweden). Thus, gel filtration was performed using Sephadex G 25 C (Pharmacia Fine Chemicals, Uppsala, Sweden). The gel was equilibrated by washing with Na-citrate buffer, pH 5.5, the culture supernatant added and the eluted fraction obtained transferred for cation exchange chromatography using an S-Sepharose HP-column (Pharmacia Fine Chemicals). The column was equilibrated by washing with 10 mM Na-citrate buffer, pH 5.5 and elution was performed using a linear salt gradient up to 0.3 M Na-chloride in 1OmM citrate buffer at pH 5.5. The volume of the gradient was 10 x VT. A chromatogram was obtained by measuring the fractions at absorbance 280 nm. Thereafter the antibodies were further purified by gel filtration on a Superdex 200 column equilibrated with 50mM phosphate buffer, pH 7.4. Fractions were then collected and samples subjected to SDS-PAGE. Before gel electrophoresis was carried out the samples were reduced with 2mercaptoethanol. For staining, Coomassie brilliant blue R-250 was used. To obtain the desired further purification, part of the I-38Sl MAbs was subjected to anion exchange chromatography using a FPLC system (Pharmacia LKB Biotechnology Group, Uppsala). First a change of buffer was made on a Sephadex G25 M column equilibrated with 75 mM Tri-HCl pH 8.0. Anion exchange chromatography was then carried out as the final stage of purification, using a Mono-Q column (Pharmacia Fine Chemicals). The Mono-Q column contains a hydrophillic resin in bead form with a particle size of 10 pm; the charged group on the gel is - CH, - N + (CH,), This was equilibrated with 75 mM Tris-HCl, pH 8.0. Bound MAbs were eluted using a linear salt gradient up to 0.3 M Na-chloride in 75 mM Tri-HCl at pH 8.0. One peak showing negligible heterogenity was pooled and the fractions, henceforth called MAb AEC 38, selected for subsequent experiments. After purification the MAbs were labelled with “‘1 (5.0-9.5 pg MAb/mL, 35-73 x lo6 cpm/mL) by the chloramine-T method (Hunter and Greenwood, 1962). Free iodine was separated from Ig-bound iodine by gel filtration on Sephadex G25 M (Pharmacia Fine Chemicals). The immunoreactivity of the MAb preparations was tested after labelling. CEA antigen was covalently coupled to CNBr-activated paper discs which were saturated with excess BSA. Labelled MAb in decreas-

Radioimmunolocalization

of human

concentrations was incubated with the discs overnight on a shaker, until equilibrium was reached. The percentage of bound over added MAb, in at least 50-fold excess of CEA antigen, was used as the value of immunoreactivity. ing

Administration of antibodies All nude rats were injected i.p., the first group of 10 animals with MAb I-38S1, the second of 10 animals with MAb AEC 38 and the third group of 5 animals with MAb 11-16. All animals received 2 pg MAb each, except the 3 rats, one from each group, receiving 10 pg. The thyroid glands were not suppressed with non-radioactive iodine before injection. Evaluation of MAb uptake External activity measurements were carried out once a day for 4 days, starting the day after MAb injection. We used a scintillation counter with a collimated sodium iodine crystal having a sensitive diameter of 10 mm, connected to a register which recorded the variations in activity over time. In the following this detector will be called the external detector. The animals were anaesthetized and measurements of activity (countslmin) were performed over the palpable tumours in both hind legs, liver, one front leg and, for a rough estimation of the degree of dehalogenation, also over the thyroid gland. One of the animals from each group which had been injected with 10 pg MAb was examined with a gamma camera to evaluate the diagnostic properties of the different MAbs and to study accumulation and clearance of MAbs in tumours and various tissues. A Phillips gamma processor with an LEGP collimator having a 128 x 128 matrix was used and static studies were carried out once a day for 4 days, starting the day after MAb injection. In the scintigraphic registrations, regular regions of interest (ROI), each ROI consisting of 36 cells, were placed over the tumours, liver, heart and thyroid gland. From these data (cpm/ROI) line charts were plotted, describing MAb uptake in the chosen tissues over time. Tissue measurements were performed on day 5 after MAb injection. The animals were anaesthetized with i.p. methohexital, blood was drawn from the heart and tumours and pieces from various organs were weighed and activity measured (counts/g x s) in a well-type gamma-counter. Some of the blood samples drawn from the animals on day 4 were centrifuged at 1000 rpm for 10 min and plasma transferred for gel chromatography on Sephadex G 25 M (Pharmacia Fine Chemicals) to estimate the high and low molecular fractions representing the amount of free and bound iodine, respectively. Tests with cultured cells LS 174 T cells were also plated in plastic culture dishes. A quantity of 2.8 f 0.3 x lo5 cells in each dish (2.9 x 104cells/cm2) was incubated for different periods of time (15 min, 1 and 24 h) with 15 PCi of

colonic

cancer

xenografts

x93

the three types of labelled MAbs in a humidified CO,-incubator at 37°C and 5% CO2 in air. Eighteen dishes were used, two for each MAb and period of incubation. Activity (cpm) was then measured, after suspending the cells in 0.1 M NaOH for 2 h at 37”C, in a liquid scintillation counter (1214 RackBeta “Exel”, LKB Wallac, Wallac Oy, Turku, Finland) using the standard “‘1 channel. Histopathology After fixation in 10% formalin buffer for 3 days at 4°C and in 70% alcohol five times during 2 days, tumour pieces were dehydrated for 10 h in 90% alcohol and for another 10 h in 99% alcohol. Infiltration was then performed using hydroxyethyl methacrylate (Historesin, LKB-Produkter AB, Bromma, Sweden) overnight at 4°C and then for 4 h at room temperature. After that the tumour pieces were embedded in methacrylate with hardener, allowed to polymerize for 12 h and then serial sectioned at 3 pm. Finally the sections were stained in Mayers’ haematoxylin and some also additionally stained with eosin-gelblich. Immunohistochemistry One xenograft from each rat, and for comparison, frozen samples from four human colorectal adenocarcinomas (one well differentiated, one moderately differentiated, one poorly differentiated and one mutinous) were immunohistochemically stained with MAb H-16. The avidin-biotin-peroxidase method was used (Vectastain ABC kit, Vector Laboratories, Burlingame, Calif., U.S.A.). 5-pm sections of the frozen tumour samples were placed on chromiumgelatin slides, air dried, fixed in acetone for 10 min at room temperature and rinsed in phosphate-buffered saline (PBS; 0.01 M, pH 7.4, 3 min x 2). Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide (30 min) and, after rinsing in PBS (3 x 2 min) non-specific background was reduced by incubation with 0.1% normal swine serum (15 min). The primary antibody, MAb H-16, was diluted with PBS containing 0.1% normal swine serum (1/SO, l/ 100 and l/400) and all subsequent reagents were diluted with PBS only. Biotinylated horse anti-mouse IgG antibody served as secondary antibody in a dilution of I;‘200 (30min) and after rinsing in PBS the sections were incubated with peroxidase-conjugated biotinavidin complex (l/200) for 30 min. Staining was developed in a solution of 3-amino-9-ethylcarbazole and 0.002% hydrogen peroxide (15 min). Next the sections were counterstained with Mayers’ haematoxylin (5 min), made blue with tap water (10 min) and mounted under a coverslip with glycerol gelatin. In order to exclude non-specific staining, primary antibody was replaced on parallel sections by 0.1% normal swine serum. Statistical method In order to estimate the variances involved. an ANOVA (model II) was applied. To test the hypoth-

A. SUNDIN ef al.

894

Table I. Comparison of culture dishes plated with (2.8 + 0.3) x 10’ LS 174 T cells. The radiolabelled MAb I-38S1, AEC 38 and II-16 were incubated for 15 min. 1 and 24 h. Mean binding values (cps) x IO6and maximum variations from two dishes are given

esis that group AEC 38 did not deviate from the other

two groups, a one-tailed Student’s t-test was used. To give a picture of the overall tendency in the data, 90% confidence intervals for the means were computed. In these experiments a bilateral tumour model was used. The activity values from each tumour are not considered separately. Instead, the mean value of the two tumours for each animal was determined and used in the statistical treatment of the radioactivity measurements. In the external measurements where the changes in radioactivity over time are considered, the 3 animals receiving the larger amount (10 pg) of MAb are not included in the comparison.

Incubation time MAb I-38Sl AEC 38 II-16

15 min

Ih

24 h

1.6 li: 0.1 1.7kO.2 2.1 k 0.6

3.6 f 0.2 2.9kO.l 3.8 f 0.5

2.7 f 0.5 2.0 + 0.2 2.8 f 0.2

human adenocarcinomas analysed which stained independently of their overall grade of differentiation. Tests with cultured cells

The immunoreactivity after labelling was 80% for MAb I-38Sl and II-16 and 70% for MAb AEC 38.

No significant differences were found between the three antibody preparations when tested on cultured LS 174 cells (Table 1). However, there was a tendency for MAb AEC 38 to bind slightly less than the others after 1 and 24 h.

Tumour growth

Evaluation of MAb uptake

Tumours developed at all sites becoming palpable and visible lo-14 days (15 and 5 animals, respectively) after injection. The tumours measured 0.7-2.8 (mean 1.4) cm in diameter with volumes between 0.2 and 4.5 (mean 1.6) cm3. No metastatic growth was noted. At dissection, the tumours showed macroscopically a capsule, solid tissue and, in the larger turnours, a central necrosis.

In the following the day of MAb injection is considered as day 0. In the external detector measurements it was found that most of the MAb uptake by the tumours took place during the first 24 h after MAb injection. For I-38Sl the maximum activity was reached on day 3; for MAb AEC 38, on day 2; and for 11-16, on day 4. Activity reached its peak over the liver on day 1, after which it decreased during the following days. When both the mean tumour and liver activities were considered, the former exceeded

Results Immunoreactivity

Histopathology and immunohistochemistry Microscopically the xenografts showed features resembling a well-moderately differentiated adenocarcinoma, as previously described by Ahlstriim et al. (1989). Immunohistochemically MAb II- 16 showedpositive staining for CEA in virtually all cells. The staining was localized predominantly to the apical border of the cells around the glandular lumina but was also present within the secreted mutinous material, though to a much lesser extent. Cytoplasmic and membranous staining was very discrete. Better differentiated regions with glandular formations showed more positive staining, compared with less differentiated parts lacking regular tubules. The binding of MAb II-16 showed the same pattern in the four

2.0, 1.5.

1

0 Fig. 1. External mean value for tumours to that 38 (0)

2 Days

Tumour/normal

tissues

4

measurements on 4 days, expressed as the each group, of the ratio of activity in the in liver. MAb I-38Sl (0) n = 5, MAb AEC n = 10 and MAb II-16 (A) n = 5.

Table 2. Tumour (T)-to-tissue ratios [ratios of the activity (cps/g tissue) in the tumours to that in other tissues] for the three MAbs, expressed as the mean value for each group and tissue and with one standard deviation eiven in oarentheses

T/blood T/liver T/muscle adjacent to tumour T/muscle foreleg T/spleen T/heart T/kidney T/small intestine T/colon T/testicle Tithvroid aland

3

MAb I-38Sl

MAb AEC 38

MAb II-16

2.6 (0.6) 4.8(1.4) 7.7 (2.6) 8.5 (4.6) 9.9 (2.0) 8.2(1.4) 7.3 (1.5) 9.1 (1.7) 9.7 (2.9) 12.1 (2.5) 0.016 (5.3 x IO--3)

3.0 (0.5) 6.6 (1.8) 12.4 (4.4) 12.9 (3.5) 10.6(1.9) 8.9(1.9) 7.4(1.3) 10.5(1.6) 13.9 (5.4) 12.4 (3.5) 0.043 (0.033)

1.6(0.1) 3.5 (0.1) 9.3 (4.5) 6.6 (2.1) 5.9 (0.9) 5.0 (0.7) 4.3 (0.7) 5.4(1.2) 7.4(1.8) 7.4Q.2) 0.014(9.3 X lo-‘)

Fig. 2. Gamma

camera

registrations

on days

14

895

of one rat injected

with 10 pg MAb 11-16.

Radioimmunoloealization of human colonic cancer xenografts 12.

a97

(b’

10

T/B

T/L

T/MT

T/MF

.

I

.

T/S

T/H

T/K

T/SB

T/C

-2-

T/TE

T/B

. .

l

6_

T/L

l

T/MT

T/MF

. .

.

T/S

. .

T/H

.*

T/K

l

T/SE

.

. .

T/C

. .

T/TE

.*

(c)

6. 4,

2 0.

P

-6. -6, T/B

T/L

T/MT

T/MF

T/S

.*

l

T/H

*

.*

T/K

. .

T/SE

T/C

T/TE

.

l

.

Fig. 3. Comparison of tumour-to-tissue ratios with 90% confidence intervals given for the mean differences between (a) MAb AEC 38 and I-38S1, (b) MAb AEC 38 and H-16, (c) I-38Sl and 11-16. T = tumour, B = blood, L = liver, MT = muscle adjacent to tumour, MF = muscle foreleg, S = spleen, H = heart, K = kidney, SB = small bowel, C = colon and TE = testicle. The result of the one-tailed significance testing marked with ‘5% and **I%.

that of the liver on day 2 for MAb AEC 38, on day 3 for MAb II-16 and somewhere between these days for MAb I-38Sl. The mean tumour-to-liver ratios were calculated (Fig. 1). MAb AEC 38 showed slightly higher ratios than I-38Sl and significantly higher than II-16 on all 4 days of measurement. In comparison with 11-16, MAb I-38Sl showed significantly better ratios only on day 3. The control activity over the front legs was low in all three groups and showed a slight time-related decrease. When the mean tumour-to-frontleg ratios were compared, only one significant difference was found, namely for MAb AEC 38 on day 4 in comparison with 11-16. The iodine uptake over the thyroid gland increased progressively over the 4 day period of measurement. In the gamma camera registrations the patterns of MAb uptake by the tumours and clearance from liver and blood seemed to be essentially the same as in the case of the external detector measurements. The only principal difference was for iodine uptake by the thyroid gland, where the activity in the scintigraphic studies was almost constant on the 4 days of measurements. The imaging properties were very good for all three MAbs (Fig. 2) with the tumours clearly visible on day 2. The mean tumour-to-tissue ratios for the various tissues measured are given in Table 2 and the confidence intervals for the mean differences between the three groups in Fig. 3. For MA\, AEC 38 the

tumour-to-tissue ratios were higher for all tissues than for I-38Sl. In the statistical testing, these ratios were significantly higher for 4 out of 10 tissues. When MAb AEC 38 was compared with II-16 the tumourto-tissue ratios were significantly better for 9 out of 10 tissues. For I-38Sl these ratios were significantly better for 6 out of 10 tissues in comparison with 11-16. The fraction of bound iodine in plasma on day 4 was about 15% for all three MAbs. No significant differences in tumour-to-tissue ratios were found when the animals given either 2 or 1Opg of MAb were compared.

Discussion The present investigation was concentrated on the problem of increasing the in viuo immunoreactivity of MAb I-38Sl by additional purification using anion exchange chromatography and thereby enhancing the uptake in the tumour. In addition, the binding properties of a second MAb (11-16) were characterized. The tumour-to-tissue ratios were better for all tissues, and for S out of 11 tissues significantly higher, for the further purified MAb AEC 38, in comparison with the same MAb before additional purification (I-38Sl). In comparison with 11-16, MA\, I-38Sl showed significantly higher ratios for 6 out of 10 tissues whereas after the anion exchange chromatography purification, giving MAb AEC 38, this was so

898

A.

SUNDIN et al.

for 9 out of 10 tissues. These results indicate that the additional purification of MAb I-38Sl increases its in uivo immunoreactivity giving an enhancement of tumour uptake in this model. The applied purification technique has not previously been described as a means to improve the immunoreactivity of MAbs, either in uiuo or in vitro. The immunohistochemical staining showed that the proportion of positive cells and intensity of positive staining for MAb II-16 was the same as previously demonstrated for MAb I-38Sl (Ahlstrom et al., 1989), i.e. virtually all cells were positive. No particular difference in staining pattern was observed between MAb I-38Sl and 11-16, though MAb I-38Sl showed rather more membranous staining. One reason for this slight difference might be that MAb I-38Sl and II-16 react with two different epitopes of the CEA molecule. This could also be one possible explanation for the differences in tumour-to-tissue ratios of the two MAbs. The staining patterns were the same, regardless of the overall grade of differentiation in the comparative study on the four human colonic adenocarcinomas. In this context, an interesting observation was made by Zoubir et al. (1990). When testing three anti-CEA MAbs on fourteen human colonic carcinomas, they found indications that the differences in immunohistochemical staining patterns are more closely related to the histology and CEA content of the tumour than to the particular MAb used. No significant differences were demonstrated in the comparison of the three MAbs in culture dishes where LS 174 T cells were incubated with the labelled MAbs. One such minor difference was that MAb II-16 showed the highest binding and I-38Sl bound somewhat better than AEC 38 at 1 and 24 h. Thus, although we found only small differences, the in vivo study showed MA\, II-16 to be clearly inferior to the other two MAbs. MAb AEC 38 showed rather poor binding in vitro after 1 and 24 h but gave the best tumour uptake. Furthermore, the immunoreactivity of the three MAb preparations tested with CEAcoupled paper discs showed rather lower values for MAb AEC 38 (70%) than for the other MAbs (80%). These findings are consistent with those of others (Mann et al., 1984), that in vitro criteria alone may not reflect the tumour localizing capacity in uivo. Many labelling procedures, however carefully performed, may damage the antibodies by altering their molecular structure thus reducing their immunoreactivity (Saccavini et al., 1986; Turner et al., 1988). However, Epenetos et al. (1986) have reported the same degree of immunoreactivity for both labelled and unlabelled MAbs-possibly due to their choice of labelling method. Thus, after labelling, the MAbs may become chemically heterogeneous, with some MAbs having little or no immunoreactivity. Therefore, even further purification of the MAbs after labelling by means of gel filtration, using for instance a Superdex column in a FPLC system or combined

anion exchange chromatography and gel filtration, would theoretically increase the immunoreactivity of the MAbs, even more than has been the case in the present investigation. To sum up, the purification of MAb I-38Sl by anion exchange chromatography improved the tumour uptake in our model. The use of this additional purification method in combination with further purification after labelling and the possibilities of modifying the technique may further potentiate the effect. However, further studies using in viuo models are required in order to devise optimal procedures. Acknowledgements-The

authors thank Veronica Asplund,

Erianne Christina Carlsson and Liliann Gille for technical assistance both with the animal and in vitro experiments and Anna Heibel for assistance concerning MAb purification. We thank Adam Taube for valuable help with the statistics and Jonas Berg for advice concerning the gamma camera registrations. The study was financially supported by the Swedish Cancer Society (grants 1176-B90-03XA, 90 : 307 and 2971-B91-OlXAB).

References Ahlstriim H., Carlsson L., Hedin A. and L&elms L.-E. (1987a) An experimental model for pharmacokinetic studies of monoclonal antibodies in human colonic cancer. Acta Oncol. 26, 447. Ahlstriim H., Carlsson L., Hedin A. and Liirelius L.-E. (1987b) Enhanced uptake of intra-arterially injected antiCEA monoclonal antibodies in human colonic cancer after mannitol infusion in an experimental model. Acfa Oncol. 26, 453. Ahlstrom H., Enblad P., Andersson A. and Liirelius L.-E. (1989) The spatial distribution of parenterally administered monoclonal antibodies against CEA in a human colorectal tumour xenograft. Acta Oncol. UI, 81, Buchegger F., Pfister C., Fournier K., Prevel F., Schreyer M., Carrel S. and Mach J.-P. (1989) Ablation of human colon carcinoma in nude mice by 131-I-labeled monoclonal anti-carcinoembryonic antigen antibody F(ab’), fragments. J. Clin. Invest. 83, 1449. Buraggi G., Turrin A., Bombardieri E., Deleide G., Gasparini M., Regalia E. and Scasselati G. (1987) Immunoscintigraphy of colorectal carcinoma with F(ab’), fragments of anti-CEA monoclonal antibody. Cancer Detect. Prevent. 10, 335. Carlsson M., Hedin A., Inganiis M., Harfast B. and Blomberg F. (1985) Purification of in vitro produced mouse monoclonal antibodies. A two-step procedure utilizing cation exchange chromatography and gel filtration. J. Immunol. Methods 19, 89. Cheung N.-K. V., Landmeier B., Neely J., Nelson A. D., Abramowsky C., Ellery S. and Adams R. B. (1986) Complete tumour ablation with iodine 131-labeled disialoganglioside G,,-specific monoclonal antibody against human neuroblastoma xenografted in nude mice. J. Nat1 Cancer Inst. 17, 739. Epenetos A. A. (1984) Attempted therapy of a human xenograft colonic cancer HT29 using 13I -I-labelled monoclonal antibodies. NATO Ado. Sci. Inst. (A) Life Sci. 82, 235. Epenetos A. A., Snook D., Durbin H., Johnson P. M. and Taylor-Papadimitriou J. (1986) Limitations of radiolabeled monoclonal antibodies for localization of human neoplasms. Cancer Res. 46, 3183. Goldenberg D. M., Sharkey R. M. and Ford E. (1987) Anti-antibody enhancement of iodine-131 anti-CEA

Radioimmunolocalization

of human

radioimmunodetection in experimental and clinical studies. J. Nucl. Med. 28, II-1604 Hammarstriim S., Shively J. E., Paxton R. J., Beatty B. G., Larsson A., Ghosh R. and Bormer 0. (1989) Antigenic sites in carcinoembryonic antigen. Cancer Res. 49, 4852. Hedin A., Hammarstriim S. and Larsson 8, (1982) Specificities and binding properties of eight monoclonal antibodies against carcinoembryonic antigen. Molec. Immunol. 19, H-1641. Hedin A., Zoubir F., Lundgren T. and Hammarstriim S. (1986) Epitope specificity and cross-reactivity pattern of a large series of monoclonal antibodies to carcinoembryonic antigen. Molec. Immunol. 23, 1053. Hunter W. M. and Greenwood F. C. (1962) Preparation of iodine-131 labelled human growth hormone of high specific activity. Nature (Lond.) 194, 495. Larsson A., Ghosh R. and Hammarstrdm S. (1987) Determination of intrinsic affinity constants of monoclonal antibodies against carcinoembryonic antigen. Molec. Immunol. 24, 569. Mach J.-P., Chatal J.-F., Lumbroso J.-D., Buchegger F., Forni M., Ritschard J. and Berche C. (1983) Tumour localization in patients by radiolabeled monoclonal antibodies against colon carcinoma. Cancer Res. 43, 5593. Mann B. D., Cohen M. B., Saxton R. E., Morton D. L., Benedict W. F., Korn E. L., Spolter L., Graham L. S.. Chang C. C and Burk M. W. (1984) Imaging of human

colonic

cancer

xenografts

x99

tumour xenografts in nude mice with radiolabeled monoclonal antibodies. Cancer 54, 1318. Rosen A., Ek K. and Aman 0. (1979) Agarose isolectric focusing of native human IgM and alfa-2-macroglobuline. J. Immunol. Methods 28, 1. Saccavini J. C.. Bohy J. and Bruneau J. (1986) Radiolabeling of monoclonal antibodies. Nucl. Med. Biol. 13, 191. Sharkey R. M., Pykett M. J., Siegel J. A., Alger E. A . Primus F. J. and Goldenberg D. M. (1987) Radioimmunotherapy of the GW-39 human colonic tumour xenograft with 131-I-Labeled murine monoclonal antlbody to carcinoembryonic antigen. Cancer Res. 47, 567? Tom B. H., Rutzky L. P., Jakstys M. M., Oyasu R.. Kayi: C. I. and Kahan B. D. (1976) Human colonic adenocarcinoma cells. I. Establishment and description of a new line. In Vitro 12, 180. Turner C. J., Sykes T. R., Longenecker B. M. and Noujaim A. A. (1988) Comparative radiolabeling and distribution of a tumour-directed monoclonal antibody. Yucl. Med Biol. 15, 701. Vaughan A. T. M., Anderson P., Dykes P. W., Chapman C. E., Bradwell A. R. and Path M. R. C. (1987) Limitations to the killing of tumours using radiolabeled antibodies. Br. J. Radio/. 60, 567. Zoubir F., Zeromski J.. Sikora J., Szmeja J.. Hedin A. and Hammarstram S. (1990) Tumour specificity 0; monoclonal antibodies to carcino-embryonic antigen. Tumour Blol. 11, 5.