Original Paper
Tumor Biol 1992;13:195-206
Departments of Cell Biology and Anatomy, Internal Medicine, and Radiology, University of Nebraska Medical Center, Omaha. Nebr., USA
Keyw ords Monoclonal antibody Human colon tumor xenografts Métastasés Radiolocalization
Quantitation of Metastatic Tumor Burden from Human Colon Tumor Xenografts using Radiolabelled Monoclonal Antibody 17-A Fragments Abstract In realistic models of human tumor xenograft metastasis, the metastatic foci arise in perivascular sites and rarely grow to sizes which are easily quantifiable by visual inspection. As an alternative approach, we have used monoclonal antibody (MAb) 17-1A F(ab')2 fragments labelled with radioiodine (l25I) to study the differential accumulation of label in xeno grafts and metastatic tumor sites as well as in noninvolved tissues of N1H Swiss nude mice receiving HT-29 human colon tumor cells. Images of the whole-body distribution and sites of localization were determined using a pinhole-collimated Angergamma camera. Radioactivity was determined in tissue samples using a well scintillation system, and pharmacokinetics were assessed during the initial 72 h after injection of antibody fragments. The half-life of l25I-F(ab')2 fragments in the blood. 8.6 h. was similar in nontumor-bearing control and tumor-bearing mice. The half-life in subcu taneous tumor xenografts was 30.1 h. The tumor xenograft to tissue activ ity ratios per unit weight (radiolocalization indices) at 72 h were: blood 90, lung 65. pancreas 50, muscle 35, spleen 20, liver and mesenteric lymph node 10. All subcutaneous xenografts were successfully imaged, and images of 5 of 9 mice (55 %) appeared to demonstrate the presence of met astatic tumor by differential and focal accumulation of MAb fragments after 48 or 72 h in the lung (2 cases) or abdominal cavity (3 cases). Necropsy and subsequent histological and biodistribution studies con firmed the presence of metastatic tumor in these sites and identified tumor in several additional sites. The smallest volume of metastatic tissue in liver or lung determined at necropsy which appeared to have been detected by imaging was about 20 mm3. Generally, for mice with meta static tumors, the radioactivity per unit weight of metastatic tumor-bear ing organs compared to tumor-free organs was 2- to 7-fold greater. The results indicate that a radiolocalization index of > 2 is generally necessary' for metastatic tumor detection by imaging although this is influenced by the extent of anatomical location of the tumor. It was possible to predict the tissue distribution of the fragments from the planar image for the amounts of radioactivity (approximately 1 mCi/kg body weight) em ployed in this study. These results demonstrate the utility of this approach to quantitate the metastatic burden arising from human colon tumor xenografts in this experimental model.
Received: September 26.1991 Accepted: Fcbruar>’ 25, 1992
Shantaram S. Joshi Departments of Cell Biology and Anatomy University of Nebraska Medical Center Omaha. N B68198-6395 (USA)
O 1992 S. Karger AG, Basel 1010-4283/92/ 0134-0195Î2.75/0
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Shantaram S. Joshia Sally L. M anna Margaret A. Tempered David A. Crouse™ Robert A. Stratbucker0 Merton A. Quaifec J. Graham Sharp™
In general, human tumor xenografts do not readily metastasize in nude mice [1], How ever, some human tumors, including the hu man colonic adenocarcinoma HT-29 [2], have been reported to metastasize in nude mice. We have noted that HT-29 growing sub cutaneously frequently produces liver and ab dominal micrometastases [3], Our observa tions suggested that these micrometastases failed to grow to gross métastasés because of an inflammatory or immune reaction mounted by the host [3], Consequently, we have refined this model by transplanting the xenografts subcutaneously into nude recipi ents whose natural killer cell responses were suppressed by injection of antibody to the ganglioside asialo GM1 [4], These mice ex hibited a greater yield of macroscopic liver and abdominal métastasés. As an alternative, we adopted a modified procedure described by Lafreniere and Rosenberg [5] which in volves intrasplenic injection of tumor cells followed by splenectomy. This technique pro duced macroscopic lung métastasés. How ever, the majority of métastasés in these mod els are associated with perivascular invasion of the tumor cells which is difficult to quanti fy. These metastatic foci rarely grow to sizes which are visible externally to the organ and, therefore, are not visually quantifiable as are métastasés in some animal models [6], Conse quently, this investigation was undertaken to determine if such métastasés could be de tected and quantified by radioimmunodetec tion techniques. These models should permit us to more closely duplicate the problems encountered in the clinical use of radiolabelled antibodies such as quantitation of tu mor burden and isotope localization, infor mation needed to make dosimetric estimates. Increasingly, monoclonal antibodies (MAbs) raised against epitopes expressed
196
preferentially or differentially on human tu mor cells are being used to target therapeutic agents such as radioisotopes, drugs or toxins to tumors in vivo [7-13], However, the ma jority of studies in animals have employed nude mice and have emphasized the detection of xenografts which are usually transplanted at sites with relatively low background levels of radiolabelled MAb localization [14, 15], Such experimental models do not accurately duplicate a major challenge of the clinical use of radiolabelled antibodies [16], Many tissues often incorporate significant levels of radioac tivity by nonspecific mechanisms [17], How ever, radiolabelled antibodies and antibody fragments have been used with moderate suc cess to image tumors in humans [16, 18-20] and animals [14, 21-25] including the use of MAb 17-1A to detect colon tumor métastasés [42], Consequently, we have evaluated the abil ity of l25I-labelled F(ab')2 fragments of the monoclonal antibody 17-1A to detect HT-29 colon tumor métastasés. Several studies have indicated the advantages of MAb fragments for imaging studies to localize tumors [20-28, 42], In addition, we have attempted to predict tissue radioactivity from planar images of these mice as an indicator of metastatic bur den which might be employed for future dosi metric studies.
Materials and Methods Cell Culture We have used a human colonic adenocarcinoma cell line, HT-29, obtained from American Type Cul ture Collections (ATCC designation HTB-38) for these studies. The derivation of the cell line was originally described by Fogh et al. [27], These cells were main tained as monolayer cultures in 25-cm2 culture flasks containing RPMI 1640 medium supplemented with 10% fetal calf serum, penicillin (100 IU/ml), strepto mycin (1 pg/ml) and /.-glutamine (2 rr\M).
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Métastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Introduction
Murine Monoclonal Antibody 17-1A Fragments F(ab')2 of the monoclonal antibody 171A were used in these studies. The antibody was a gift from Dr. Steplewsky of the Wistar Institute (Philadel phia, Pa., USA). This antibody, originally produced by Herlyn et al. [28], is known to react with 37-kD antigen molecules expressed on gastrointestinal adenocarci noma cells [29]. The purification and fragmentation of this antibody were conducted using standard protocols [30],
Radioiodination o f 17-1A F(ab)2 Fragments MAb 17-1A fragments were iodinated using the chloramine-T method [31]. To describe the method briefly, 10 mg of fragments in phosphate buffered saline (PBS) were mixed with 5 mCi of l25I (as sodium iodide) and 2 mg of chloramine-T in 0.1 ml PBS. The mixture was shaken gently for 7 min, and the reaction was stopped by adding 2 mg sodium metabisulfite in 0.1 ml PBS. The entire reaction mixture was then loaded onto a preequilibrated Sephadex G -10 column, and the column was eluted with phosphate buffer. Thirty i-ml fractions were collected, and the radioac tivity and protein content of each fraction were deter mined using a scintillation counter and a spectropho tometer, respectively. The radiolabelled protein peak was pooled and subjected to quality assurance as described below. Gel Electrophoresis The molecular integrity of the labelled antibody fragments was determined by gel electrophoresis of the labelled and unlabelled products followed by staining with Coomassie blue and autoradiography by methods described elsewhere [32]. Cell-Binding Assay An in vitro cell-binding assay was performed to demonstrate that the radiolabelled antibody retained antigen-binding capacity. A fixed quantity of radiola belled fragments (10 pg) was incubated with variable numbers of HT-29 cells (1 X 10s—50 X 105). Alterna tively, a fixed number (8 X 106) of HT-29 cells was incubated with variable amounts of antibody (250 pg). Incubations were carried out in quadruplicate for 30 min. After incubation, the cells were washed 3 times with PBS containing 1% fetal bovine serum and pelleted. The amount of radioactivity bound was de termined by scintillation counting. In vivo Imaging and Biodislribution of Radiolabelled Antibody Fragments Tumor-bearing or control mice were injected i.v. with radiolabelled fragments. Because these studies were performed over the course of several weeks, the amount of radioactivity injected varied, but an at tempt was made to maintain a constant specific activ ity of the injected antibody. The values expressed as 106 cpm/pg fragments were 0.20 ± 0.07 (range 0.120.25), 0.21 ± 0.08 (range 0.13-0.28), 0.28 ± 0.06 (range 0.13-0.33) and 0.20 ± 0.05 (range 0.12-0.24) at 1, 24, 48 and 72 h, respectively, corresponding to 20-40 pCi of iodine and 200-400 pg of antibody frag ments per mouse. Most mice were imaged daily until
197
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Human Colonic Tumor Xenografts These studies employed male and female outbred Swiss nude (nu/nu) mice, 6-8 weeks old, raised in our facility from breeding pairs kindly provided by Dr. Carl Hansen of NIH. Mice received tumor cither as 3 X 106—5 X 106 live tumor cells injected in 100 pi saline subcutaneously (s.c.) into the right rear flank, or intrasplcnically (i.s.) in 200 pi saline 3 days after intra venous (i.v.) injection of 50 pi (500 pg anti-asialoGM1 antibody (Waco, Dallas, Tex., USA). The s.c. route gives rise to a primary xenograft and largely abdominal métastasés [3]. The i.s. route does not usually lead to the establishment of a primary xeno graft and was suspected to be a better method of estab lishing lung métastasés [5], For the i.s. injections, ani mals were anesthetized using pentobarbital (50 mg/ kg), and the spleen was gently maneuvered into posi tion along an incision in the flank. It was then secured by gently placing (but not closing) a hemostat around the pedicle. Cells were then injected in 0.2 ml, and a small piece of gel foam sponge was used to stem any bleeding or leakage. One minute later the vessels of the splenic pedicle were ligated, the spleen was then re moved and the incision was closed in layers. The entire procedure was conducted ascptically and with care to ensure that the incision site and tissues did not become dry during the surgery. The mice were observed daily and used for further studies of metastasis detection by the radioimmunoconjugate after a period of 2-6 weeks, at which time the subcutaneous tumors were 0.5-1.0 cm in diameter. A group of handled, noninjected controls was prepared. For studies on the halflife of retained radioactivity, at least 3 mice from each group were autopsied after 1, 24, 48 or 72 h. Primary tumor xenografts were only available for study of radioactivity from the mice transplanted with s.c. tumors. Images were processed for detection of métas tasés in 6 mice (3 receiving s.c. tumor and 3 i.s. tumor) at 48 h, and in 8 mice (5 receiving s.c. tumor and 3 i.s. tumor) at 72 h. Multiple tissues from these mice were processed histologically at necropsy.
Necropsy Procedures After the final imaging, animals were killed and various organs and tissues were removed, observed grossly for tumor, weighed, and radioactivity was counted using a 7-counter. Whole organs (liver, lung, intestine, pancreas and spleen) were fixed in 10% neu tral buffered formalin for histological analyses. After blind coding, multiple sections cut from these organs using standard techniques and stained with hematoxy lin and cosin were examined for evidence of tumor metastases and other pathology. The presence of meta static foci determined histologically was compared with gross observations, imaging and the radioactivity in the same organ. Statistical Analyses Numerical values were expressed as the mean and SD and evaluated employing Student’s t test. The radioactive clearance data were subjected to log trans formation and fitted by least-squares analysis. The cpm/pixel data were further analyzed using a multivar iate analysis of variance. The criterion for statistical significance was p < 0.05.
198
Results Analysis o f Labelled Product
The pooled radiolabelled fragments sub jected to SDS polyacrylamide gel electropho resis with Coomassie blue staining and auto radiography exhibited a major and several minor molecular weight species. The approxi mately 96-kD F(ab')2 fragment predomi nated, but there was minor contamination with intact immunoglobulin (150kD) and also a very minor band of approximately 50 kD which was probably free heavy chain. Most of the radioactivity was attached to the F(ab')2 fragment and the bulk of the remain der was attached to whole immunoglobulin. The molecular integrity of the F(ab')2 frag ment was preserved after radioiodination. In vitro Binding o f 17-1A F(ab)2 Fragments to HT-29 Colonic Tumor Cells
A significant (p < 0.001) linear relation ship existed between the bound radioactivity and the amount of antibody fragment added to a fixed number of HT-29 tumor cells (fig. la). Similarly, a significant (p < 0.001) linear relationship existed between the bound radioactivity and the number of HT-29 tumor cells labelled with a fixed amount of antibody (fig. lb). These results demonstrate that the antigen-binding ability of the radiolabelled product was retained. Blood Clearance o f 17-1A F(ab)2 Fragments
The blood radioactivity was determined at various times after i.v. injection of the 17-1A F(ab')2 fragments into control nude mice and mice receiving a s.c. or i.s. inoculation of HT29 tumor cells (fig. 2). The clearance of ra dioactivity from the tumor in s.c. injected mice, and clearance from the blood in all groups was logarithmic, and the biological half-life (Ti/,) calculated by linear regression
Joshi/Mann/T empero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Metastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
autopsy. Initially, thyroid uptake of radioactivity was blocked by administering potassium iodide (16 mg/ml, 0.01 ml/g body weight intrapcritoncally 3, 6 and 18 h prior to radioconjugate administration). However, it was noted that thyroid uptake provided a convenient anatomical marker on the images, therefore this proce dure was discontinued. For imaging, the mice were anesthetized (sodium pentobarbital 50 mg/kg body weight) and taped prone on a styrofoam block. They were imaged using a Technicare 420 mobile Angergamma camera. An analogue film was recorded and the image stored digitally on magnetic media for later analysis. The digital images were displayed for analysis on a Picker PCS-512 com puter system. Regions of interest (ROI) were selected and outlined using a multiposition joystick which per mitted the definition of irregular areas or shapes for determining cpm/pixel. The images which demon strated tumors and suspected metastases at 48 and 72 h post-injection were analyzed independently by 4 investigators to determine cpm/pixel in sites of sus pected tumors as well as in heart, lung, liver, spleen (when present) and thigh muscle (as a remote back ground area). These values were expressed as ratios rel ative to the heart, and plotted as a function of cpm/mg for the same tissue or tumor relative to cpm/mg blood determined at necropsy.
F(ab')2, pg
Fig. 2. Clearance of l25I-labelled 17-1A F(ab'h MAb frag ments. ----- = Regression for blood radioactivity from control (o), s.c. injected (A) and i.s. injected (□) mice: half-life = 8.6 h: r = 0.98. -----= Regression for tumor (•) ra dioactivity; half-life = 30.1 h: r = 0.92.
based on all data points was 30.1 h in the tumor and 8.6 h in blood. Xenograft Localization o f the Radioconjugate
Nude mice injected s.c. with HT-29 tumor cells developed tumors at the site of injection. In order to confirm the in vivo localization of 17-1A F(ab')2 fragments, the ratio of radioac
samples, r = 0.99, p < 0.001. b Fixed amount (10 gg) of F(ab')2 vs. variable number of HT-29 cells. Quadrupli cate samples, r = 0.96, p < 0.001.
Hours after injection
tivity per gram of tumor versus various tis sues was determined (fig. 3). It can be seen that the ratios increased with time. This in crease presumably reflects the continuing clearance of the antibody fragment from the blood pool while that in the tumor is retained, as has been previously reported [33, 34], Our results suggest that the tumor/normal tissue ratio of antibody fragments is still increasing
199
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Fig. 1. In vitro binding of radiolabellcd 17-1A MAb fragments to HT-29 colonic adenocarcinoma cells, a Fixed number of HT-29 cells vs. variable amounts of F(ab'h MAb fragments. Quadruplicate
Fig. 3. Ratios of s.c. xenograft to organ radioactivity in HT-29tumor-bearing nude mice injected with l25I-labelled 17-1A F(ab')2 MAb fragments. o = Blood; • = lung; A = liver; ▲= pancreas; □ = mesenteric lymph node; ■ = spleen; X = muscle.
even at 72 h after antibody injection. Some of the radioactivity ratios were at the upper lim its of those previously reported [14. 21, 22], For example, at 72 h after antibody fragment injection, we observed tumor to blood ratios averaging 90:1, to lung 65:1. to pancreas 50:1, to muscle 35:1, to spleen 15:1, and to liver and mesenteric lymph node 10:1. There was some variation in size of the primary subcuta neous xenografts (0.5-1.0 cm diameter) but no correlation between their size and their content of radioactivity relative to blood lev els. Regardless of their size, all primary xeno grafts were readily imaged. In vivo Localization o f Metastatic Tumor
The primary tumor was clearly localized in all the nude mice receiving HT-29 tumor cells s.c., and metastases were strongly suspected by all observers in the lung or abdominal sites in 9 of 14 of the recipients of s.c. or i.s. tumor cells. Sample scans (gamma camera images) are shown in figure 4. When compared to the control (fig. 4a), the image in figure 4b shows both the primary xenografts and an increased localization of radioactivity in the liver. This
200
was shown by gross and histological examina tions of this sample to represent metastatic tumor (fig. 5c, d). Similarly, figure 4c shows an increased localization of radioactivity in the lungs of a recipient of i.s. tumor cells (no primary xenograft). This was shown by histo logical analysis to represent extensive meta static tumor (fig. 5b). In 6 mice receiving i.s. injection of tumor cells, 1 had lung metas tases, 1 liver metastases, 1 pancreas and mes enteric lymph node metastases and 2 both liver and lung metastases, and one of the latter also had mesenteric lymph node metastases. Table 1 summarizes the results of tissue and organ radioactivity determination in con trol nude mice without tumor compared to mice receiving HT-29 tumor cells but without evidence of metastasis in a particular tissue, and to suspected individual metastatic le sions. Suspected metastatic lesions were ini tially identified with scintigraphy, and con firmed either from gross observations of the dissected organ or by subsequent histopathological analysis. It can be seen that most meta static lesions had radioactivity levels 2-7 times the average level of radioactivity in that
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Metastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Hours after injection
Fig. 4. Sample Anger camera images of nude mice 72 h after injection with 125I-labcllcd 17-1A F(ab')2 MAb fragments, a Control mouse without tumor. The thyroid was not blocked, providing orientation, b HT-29-tumor-bearing mouse with s.c. xenografts (P) and liver metas tasis (Li). Thyroid activity was not blocked, c Mouse injected i.s. with HT-29 tumor cells
a
c
b
d Fig. 5. Histopathological analyses of HT-29 pri mary tumors and their metastasis in nude mice, a A primary HT-29 tumor growing in a subcutaneous site in a nude mouse. X 45. b Lung metastasis of HT-29 tumor cells in a nude mouse. X 45. c Liver metastasis
of HT-29 tumor cells evident largely as inflammatory cells with a small number of peripheral tumor cells. X 45. d Higher magnification of the periphery of a liver metastasis to demonstrate the presence of tumor cells. Some inflammatory cells are also present. X 180. HE.
201
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
showing lung metastasis (Lu). Thyroid radioactivity was blocked.
Table 1. Average levels of radioactivity in various tissues of HT-29-bearing nude mice suspected on the basis of imaging and gross examination to be free of metastatic disease, compared with individual levels of radioactiv ity in sites of suspected metastatic disease, and levels in nontumor bearing nude mice
Tissue
Radioactivity, % injected dose/mg, X 104 72 h (n = 7)
48 h (n = 7) control nontumor bearing
tumor bearing suspected free of metastasis3 metastasis
control nontumor bearing
tumor bearing suspected metastasis3 free of metastasis
Blood Muscle Lung
0.25 ±0.07 ND 0.20 ±0.04
0.36 ± 0.21 ND 0.39 ±0.19
0.11 ±0.04 0.11 ±0.05 0.23 ±0.06
0.12±0.02 0.12±0.02
Pancreas Spleen Liver
0.36 ±0.16 1.48 ±0.49 1.53 ± 0.33
0.35 ±0.15 1.28 ±0.36 l.29±0.61
0.11 ±0.04 1.09 ±0.13 1.03 ±0.12
0.21 ±0.07 0.33 ±0.06* 0.79 ±0.09
Mesenteric lymph node
0.98 ±0.22
1.38 ±0.92
0.16 ± 0.05
0.49 ±0.18
NA NA 0.96b (2.5) 0.78b (2.0) 2.59(7.4) -
2.18(1.7) 2.51 (2.0) 2.73(2.0) 3.21 (2.3)
0.10 ± 0.03
NA NA 2.64(22) 60.34(503) 0.45(2.1) -
3.22 (4.1) 0.92b (1.2) 1.67(3.4) 1.54(3.1)
* p < 0.05 in comparison with control. ND = No data; NA = not available. a Tissues suspected to contain metastatic tumor on the basis of imaging were carefully examined at necropsy for tumor whose presence was subsequently confirmed histologically. Note that in the case of abdominal metastasis the exact location, e.g. pancreas versus mesenteric lymph node, could not be determined from the image. Values in parentheses are ratios to metastasis-free levels. b Suspected on the basis of gross appearance (nodules) or tissue sectioning, not on an image.
202
Examples of the histopathological confir mation of metastatic tumor detected in liver and lung are presented in figure 5. Figure 5a shows a primary xenograft of the human co lonic adenocarcinoma cell line HT-29. Fig ure 5b shows metastatic tumor in the lung of a nude mouse injected i.s. with HT-29 tumor cells. Figures 5c and 5d demonstrate a liver metastasis of HT-29 tumor cells. On the basis of whole-body imaging and by counting the radioactivity, métastasés were predicted in these organs. The histological studies con firmed this prediction. Predictive Value o f Images
Images of control mice at various times were useful for comparison with images of
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Métastasés
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
particular tissue in metastasis-free mice. Two values, 1.3 and 1.7, were lower than this range and two were higher (22 and 503). The lowest value of 1.3 occurred for the liver of an image negative mouse which had gross tumor nod ules subsequently confirmed histologically to be tumor. The amounts of metastatic tumor in lung and liver varied greatly. The highest value of 503 was obtained for lung which was image positive and extensively involved with metastatic tumor (fig. 5c). In the mice without evidence of metastatic lesions, the tissue lev els of radioactivity were generally similar to those of the control with the exception of the spleen at 72 h, which was significantly lower than the control. The reasons for this differ ence are not known.
Fig. 6. Regression of relative (tumor/blood or tissue/blood) ra dioactivity levels determined from images vs. levels measured by scin tillation counting at autopsy. Slope = 10.52, r = 0.87.
tion of cpm/mg tumor, that together with T% values (fig. 2) permits dosimetric estimates to be made based solely on the image and a blood sample.
Discussion
We have demonstrated the in vivo ra dioimmunodetection of metastasis of human colonic tumor xenografts in two slightly dif ferent animal models. These studies validate the application of these models for the devel opment and evaluation of potential techno logical advances in radioimmunodetection and dosimetry. The model employs a human colonic tumor cell line HT-29 whose charac teristics are similar to those of many primary human colonic adenocarcinomas. It also ben efits from the utility of the nude mouse as a model for studying the behavior of various human tumors, and has already been estab lished [1. 17, 21, 22, 25]. It should be noted that since our objective was to model the likely clinical situation, we elected not to use irrelevant isotope-identical fragments as a control. The use of such a control facilitates
203
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
tumor-bearing mice. Images of tumor-bearing mice at 1 and 24 h showed localization in pri mary subcutaneous xenografts but were not useful for identification of metastatic tumor sites because of blood pool background. Therefore, the analysis of images for meta static tumor sites was limited to images of 7 mice at 48 h, and 7 mice at 72 h. The analysis of images for the prediction of tissue radioac tivity demonstrated that there were no signifi cant differences in the observers’ extraction of data for the tissues studied, with the excep tion of thigh muscle. The observers disagreed on the cpm/pixel values for this site probably because the limits of the images were diffuse and hard to define. The correlation of cpm/ pixel (mean values from all observers) with dpm/mg tissue/blood ratios is shown in fig ure 6. At lower levels of radioactivity the cor relation is linear and significant (p = 0.017). This result indicates that, at least within cer tain limits, the burden of tissue radioactivity can be predicted from the planar image as can the size of the tumor given certain assump tions as to its shape, for example, assuming it is an oblate spheroid. This information to gether with cpm/mg blood permits an estima
204
cols [18, 38, 39], The tumor to normal tissue ratios obtained using this antibody fragment were among the highest reported for epithelial tumors and, as such, offer promise for the use of such antibody fragments in detection [40. 41] of established metastases in this experi mental model. The reasons that the ratios of radioactivity in the tumor to liver and mesen teric lymph node (10:1) are much lower may be, in part, due to the presence of micrometastatic tumors in these tissues in some mice [3], The localization of radioactivity was greater in those sites identified by imaging or gross examination to be suspected sites of metastasis. In general, tumor to tissue ratios for the whole organ ranged from 2-7. This suggests that the metastases being detected represent about 5-20% of the mass of tissues. The exceptions were 2 lung metastases, which had high ratios, suggesting major metastatic lesions. This was confirmed histologically (fig. 5b). In 14 tumor-bearing mice which were imaged at 48 or 72 h, lung metastasis was suspected in 2 of 3 mice and abdominal metastasis in 3 of 6 mice. Metastatic lesions in 2 mice were detected by gross observation (small liver nodules) and were not evident in the images. In 2 other mice, tumor deposits were detected by sectioning the tissue. Thus, detection of metastatic lesions by imaging oc curred in 5 of 9 (55%) which is similar to results reported for studies in humans [18, 42] , Other investigators have selected a ratio of 2 [7] or greater [21] to indicate specific con centrations of radioactivity in tumors. All our primary xenografts and all except 2 metastatic lesions met this criterion. The metastatic tu mors observed at 48 h occurred in 2 of 3 mice receiving cells s.c. and in 3 of 4 mice receiving i.s. tumor cells. At 72 h metastases were found in 2 of 5 s.c. tumor cell recipients and 2 of 2 i.s. recipients. The frequency of metastasis in the two models, 4 out of 8 s.c. versus 5 out of 6 i.s., is not significantly different (p = 0.24).
Joshi/Mann/Tempero/Crousc/ Stratbuckcr/Quaife/Sliarp
Radiolocalization of Colon Tumor Metastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
the evaluation of specific radiolocalization but complicates the interpretation of the images obtained to detect sites of metastatic tumor. The in vitro binding of the 17-1A anti body fragments to HT-29 cells indicated that these antibody fragments should also be use ful in vivo as was shown to be the case [20, 36]. This approach may not be universally applicable because of antigenic variation be tween cells of the same tumor, and the possi bility that potentially useful antigens ex pressed by the tumor in vivo might not be expressed by the same tumor cells growing in vitro [14, 37], This emphasizes the need for suitable in vivo experimental models. Our observations confirm previous reports of the efficacy of F(ab')2 fragments of the MAb, 17-1A for the radioimmunodetection of human colonic tumors [42], The biological half-life of these fragments administered i.v. in our model was 8.6 h which is similar to the range of 5-16 h reported for MAb F(ab')2 and an irrelevant control [26], In therapeutic ap plications, there may be concern that the rapid clearance of antibody fragments would decrease the dose delivered to the tumor [43], However, the therapeutic ratio defined as the dose to the tumor versus the dose to critical normal organs would seem to be crucial. Po tentially, there may in fact be an advantage, in studies which propose to combine imaging with therapy of tumors, in employing a mix ture of fragments and whole immunoglobulin. The former would permit more rapid detec tion of tumor sites whereas the whole immu noglobulin would remain in the tumor longer and deliver a greater cumulative dose of ra dioactivity [26, 35], It is still far from clear which antibody fragments or whole immuno globulin are best. Additionally, the host im mune response to fragments may be expected to differ from that towards whole immuno globulin, and these differences might become significant in multiple administration proto
However, both models gave rise to macro scopic human tumor métastasés. These data also suggest that metastatic deposits as small as 20 mm3 could be detected as suggested by radioactivity and confirmed by histological examination. Consequently, this approach has merit for the quantitation of metastatic tumor burden.
Acknowledgements Wc thank Dr. Kashinath Patil and Dr. Greg Perry for their assistance with the statistical analyses, Dr. Shailendra Saxena and Mr. Richard Kelley for excel lent technical assistance, and Ms. Roberta Anderson for typing the manuscript. This work was partly sup ported by Nebraska Department of Health LB506 funds.
References
JAMA 1989:26 i: 744-746.
8 Ritz J, Schlossman SF: Utilization of monoclonal antibodies in the treatment of leukemia and lympho ma. Blood 1982;59:1-11. 9 Byers VS, Baldwin RW: Thera peutic strategies with monoclonal antibodies and immunoconjugates. Immunology 1988:65:329—335.
10 Goldenberg DM: Current status of cancer imaging with radiolabclled antibodies. J Cancer Res Clin Oncol 1987:113:203-208. 11 Blumenthal RD. Kashi R. Stephen R. et al: Improved radioimmuno therapy of colorectal cancer xeno grafts using antibody mixture against carcinoembryonic antigen and colon-specific antigen-P. Can cer Immunol Immunother 1991:32: 303-310. 12 Eary JF: Fundamentals of radioim munotherapy. Int J Rad Appl lnstrum [B] 1991;18:105-108. 13 Riva P. Marangala M. Tison V, et al: Treatment of metastatic colorectal cancer by means of specific mono clonal antibodies conjugated with iodine-131, a phase II study. Int J Rad Appl Instrum [B] 1991:13:109119. 14 Buchsbaum D, Lloyd R.Juni J. et al: Localization and imaging of radiola beled monoclonal antibodies against colorectal carcinoma in tu mor-bearing nude mice. Cancer Res 1988;48:4324-4333. 15 Sharkey RM. Wcadock KS, Natale A. et al: Successful radioimmuno therapy for lung métastasés of hu man colonic cancer in nude mice. J Natl Cancer Inst 1991:83:627-632. 16 Carrasquillo JA, Sugarbaker P. Colcher D, et al: Radioimmunoscintigraphy of colon cancer with io dine- 131 -labeled B72.3 monoclonal antibody. J Nucl Med 1988:29: 1022-1030.
17 Sharp JG, Coop KL, Cheng HF, ct al: Scintiscanning and tissue distri bution studies of radiolabelled anti tumor antibodies in rats bearing the corresponding intestinal tumor. Eur J Nucl Med 1982:7:28-34. 18 Rosen ST, Zimmer AM. GoldmanLekin R. ct al: Radioimmunodetec tion and radioimmunotherapy of cutaneous T cell lymphomas using l31I-labellcd monoclonal antibody: An Illinois Cancer Council Study. J Clin Oncol 1987;5:562-573. 19 Colcher D, Esteban JM, Carras quillo JA, et al: Quantitative analy ses of selective radiolabelled mono clonal antibody localization in met astatic lesions of colorectal cancer patients. Cancer Res 1987:47:11SS II 89. 20 Welt S, Chaitanya R, Divigi R. et al: Quantitative analysis of antibody localization in human metastatic co lon cancer: A phase I study of mono clonal antibody A33. J Clin Oncol 1990;8:1894-1906. 21 Stya M, Wahl RL. Natale RB. et al: Radioimmunoimaging of human small cell lung carcinoma xenografts in nude mice receiving several monoclonal antibodies. Natl Cancer Inst Monogr 1987;3:19-23. 22 McCabe RP. Peters LC. Haspel MV, et al: Prcclinical studies on the phar macokinetic properties of human monoclonal antibodies to colorectal cancer and their use for detection of tumors. Cancer Res 1988:48:43484353.
205
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
1 Dore JF, Bailly M. Bertrand S: Mé tastasés of human tumors in experi mental animals. Anticancer Res 1987:7:997-1004. 2 Kozlowski JM. Fidler 1J, Campbell D, et al: Metastatic behavior of hu man cell lines grown in nude mice. Cancer Res 1984:44:3522—3529. 3 Joshi SS, Jackson JD. Sharp JG: Comparison of the growth and me tastasis of four human intestinal tu mor cell line xenografts. Tumor Biol 1989:10:117-125. 4 Joshi SS, Mathews NB. Sinangil F. et al: Effects of differentiation in ducing chemicals on in vivo malig nancy and NK susceptibility of met astatic lymphoma cells. Cancer De tect Prev 1988:11:405-417. 5 Lafreniere R, Rosenberg SA: Suc cessful immunotherapy of murine experimental hepatic métastasés with lymphokine activated killer cells and recombinant interleukin-2. Cancer Res 1985:45:3735-3741. 6 Joshi SS. Sharp JG, Brunson K.W: Differential growth characterization of metastatic variant RAW117 mu rine lymphosarcoma cells. Oncology 1987:44:180-185. 7 Schlom J: Innovations in mono clonal antibody tumor targeting.
31 Hunter WM: Radioimmunoassay; in Weir DM (ed): Handbook of Ex perimental Immunology. Oxford, Blackwell, 1978, pp 14-33. 32 Joshi SS, Sharp JG, Gharpurc HM, ct al: Characterization of metastasisassociated antigens on RAW 117 lymphosarcoma cell lines. Clin Exp Metastasis 1987;5:89-104. 33 Pimm MV, Perkins AC, Armitage NC, et al: The characteristics of blood-borne radiolabels and the ef fect of anti-mouse IgG antibodies on localization of radiolabelled mono clonal antibody in cancer patients. J Nucl Med 1985;26:1011-1023. 34 Fogler WE, Kinger MR, Abraham KG, et al: Enhanced cytotoxicity against colon carcinoma by combi nations of noncompeting mono clonal antibodies to the 17-1A anti gen. Cancer Res 1988;48:63036308. 35 Hand PH, Colcher D. Salmon D, et al: Influence of spatial configuration of carcinoma cell populations on the expression of a tumor-associated glycoprotein. Cancer Res 1985;45: 833-840. 36 Mach JP. Buchegger F, Forni M, et al: Use of radiolabelled monoclonal anti-CEA antibodies for the detec tion of human carcinomas by exter nal photoscanning and tomoscinti graphy. Immunol Today 1981:2: 239-249.
37 Esteban JM. Seldom J,FrancoiseM, et al: Radioimmunothcrapy of athymic mice bearing human colon carcinomas with monoclonal anti body B72.3: Histological and auto radiographic study of effects on tu mors and normal organs. Eur J Can cer Clin Oncol 1987;23:643-655. 38 Esteban JM.Schlom J.GansowOA, et al: New method for the chelation of indium-111 to monoclonal anti bodies: Biodistribution and imaging of athymic mice bearing human co lon carcinoma xenografts. J Nucl Med 1987;28:861-870. 39 Sharkey RM, Kaltovich FA, Shih LB, et al: Radioimmunotherapy of human colonic cancer xenografts with 90Y-labelcd monoclonal anti bodies to carcinoembryonic antigen. Cancer Res 1988;48:3270-3275. 40 Buchegger F, Vacca A, Carrel S, et al: Radioimmunotherapy of human colon carcinoma by l3,I labeled monoclonal anti-CEA antibodies in a nude mouse model. Int J Cancer 1988;41:127-134. 41 LoBuglio AF, Saleh MN, Lee J, et al: Phase I trial of multiple large doses of murine monoclonal antibody C017-1A. I. Clinical aspects. J Natl Cancer Inst 1988;80:932-936. 42 Mach JP, Chatal JF, Lumbroso JD, et al: Tumor localization in patients by radiolabelled monoclonal anti bodies against colon carcinoma. Cancer Res 1983;43:5593-5600. 43 Colcher D. Zalutsky M, Kaplan W, et al: Radiolocalization of human mammary tumors in athymic mice by a monoclonal antibody. Cancer Res 1983;43:736-742.
206
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Métastasés
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
23 Endo K, Kamma H, Ogata T: Ra diolabeled monoclonal antibody 15 and its fragments for localization and imaging of xenografts of human lung cancer. J Natl Cancer Inst 1988;80:835-842. 24 Wahl RL, Barret J, Geatti D. ct at: The intraperitoneal delivery of radiolabelled monoclonal antibodies: Studies on the regional delivery ad vantage. Cancer Immunol Immunother 1988;26:187-201. 25 Herlyn D. Powe J, Alavi A, et al: Radioimmunodetection of human tumor xenografts by monoclonal an tibodies. Cancer Res 1983:43:2731— 2735. 26 Goodwin DA: Pharmacokinetics and antibodies. J Nucl Med 1987; 28:1358-1362. 27 Fogh J, Fogh JM, Orefo T: One hundred and twenty-seven cultured human cell lines producing tumors in nude mice. J Natl Cancer Inst 1977;59:221-226. 28 Herlyn M. Steplewski Z, Herlyn D: Colorectal carcinoma specific anti gens: Detection by means of mono clonal antibodies. Proc Natl Acad SciUSA 1979;76:1438-1442. 29 Gottlinger HG, Funke I, Johnson JP: The epithelial cell surface anti gen 17-1 A, a target for antibody me diated tumor therapy: Its biochemi cal nature, tissue distribution and recognition by different monoclonal antibodies. Int J Cancer 1986;38: 47-53. 30 Sears H, Herlyn D, Steplewski Z: Initial trial use of murine mono clonal antibodies as immunotherapcutic agents for gastrointestinal ad enocarcinoma. Hybridoma 1986;5: 109-115.
Tumor Biol 1992;13:195-206
Departments of Cell Biology and Anatomy, Internal Medicine, and Radiology, University of Nebraska Medical Center, Omaha. Nebr., USA
Keyw ords Monoclonal antibody Human colon tumor xenografts Métastasés Radiolocalization
Quantitation of Metastatic Tumor Burden from Human Colon Tumor Xenografts using Radiolabelled Monoclonal Antibody 17-A Fragments Abstract In realistic models of human tumor xenograft metastasis, the metastatic foci arise in perivascular sites and rarely grow to sizes which are easily quantifiable by visual inspection. As an alternative approach, we have used monoclonal antibody (MAb) 17-1A F(ab')2 fragments labelled with radioiodine (l25I) to study the differential accumulation of label in xeno grafts and metastatic tumor sites as well as in noninvolved tissues of N1H Swiss nude mice receiving HT-29 human colon tumor cells. Images of the whole-body distribution and sites of localization were determined using a pinhole-collimated Angergamma camera. Radioactivity was determined in tissue samples using a well scintillation system, and pharmacokinetics were assessed during the initial 72 h after injection of antibody fragments. The half-life of l25I-F(ab')2 fragments in the blood. 8.6 h. was similar in nontumor-bearing control and tumor-bearing mice. The half-life in subcu taneous tumor xenografts was 30.1 h. The tumor xenograft to tissue activ ity ratios per unit weight (radiolocalization indices) at 72 h were: blood 90, lung 65. pancreas 50, muscle 35, spleen 20, liver and mesenteric lymph node 10. All subcutaneous xenografts were successfully imaged, and images of 5 of 9 mice (55 %) appeared to demonstrate the presence of met astatic tumor by differential and focal accumulation of MAb fragments after 48 or 72 h in the lung (2 cases) or abdominal cavity (3 cases). Necropsy and subsequent histological and biodistribution studies con firmed the presence of metastatic tumor in these sites and identified tumor in several additional sites. The smallest volume of metastatic tissue in liver or lung determined at necropsy which appeared to have been detected by imaging was about 20 mm3. Generally, for mice with meta static tumors, the radioactivity per unit weight of metastatic tumor-bear ing organs compared to tumor-free organs was 2- to 7-fold greater. The results indicate that a radiolocalization index of > 2 is generally necessary' for metastatic tumor detection by imaging although this is influenced by the extent of anatomical location of the tumor. It was possible to predict the tissue distribution of the fragments from the planar image for the amounts of radioactivity (approximately 1 mCi/kg body weight) em ployed in this study. These results demonstrate the utility of this approach to quantitate the metastatic burden arising from human colon tumor xenografts in this experimental model.
Received: September 26.1991 Accepted: Fcbruar>’ 25, 1992
Shantaram S. Joshi Departments of Cell Biology and Anatomy University of Nebraska Medical Center Omaha. N B68198-6395 (USA)
O 1992 S. Karger AG, Basel 1010-4283/92/ 0134-0195Î2.75/0
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Shantaram S. Joshia Sally L. M anna Margaret A. Tempered David A. Crouse™ Robert A. Stratbucker0 Merton A. Quaifec J. Graham Sharp™
In general, human tumor xenografts do not readily metastasize in nude mice [1], How ever, some human tumors, including the hu man colonic adenocarcinoma HT-29 [2], have been reported to metastasize in nude mice. We have noted that HT-29 growing sub cutaneously frequently produces liver and ab dominal micrometastases [3], Our observa tions suggested that these micrometastases failed to grow to gross métastasés because of an inflammatory or immune reaction mounted by the host [3], Consequently, we have refined this model by transplanting the xenografts subcutaneously into nude recipi ents whose natural killer cell responses were suppressed by injection of antibody to the ganglioside asialo GM1 [4], These mice ex hibited a greater yield of macroscopic liver and abdominal métastasés. As an alternative, we adopted a modified procedure described by Lafreniere and Rosenberg [5] which in volves intrasplenic injection of tumor cells followed by splenectomy. This technique pro duced macroscopic lung métastasés. How ever, the majority of métastasés in these mod els are associated with perivascular invasion of the tumor cells which is difficult to quanti fy. These metastatic foci rarely grow to sizes which are visible externally to the organ and, therefore, are not visually quantifiable as are métastasés in some animal models [6], Conse quently, this investigation was undertaken to determine if such métastasés could be de tected and quantified by radioimmunodetec tion techniques. These models should permit us to more closely duplicate the problems encountered in the clinical use of radiolabelled antibodies such as quantitation of tu mor burden and isotope localization, infor mation needed to make dosimetric estimates. Increasingly, monoclonal antibodies (MAbs) raised against epitopes expressed
196
preferentially or differentially on human tu mor cells are being used to target therapeutic agents such as radioisotopes, drugs or toxins to tumors in vivo [7-13], However, the ma jority of studies in animals have employed nude mice and have emphasized the detection of xenografts which are usually transplanted at sites with relatively low background levels of radiolabelled MAb localization [14, 15], Such experimental models do not accurately duplicate a major challenge of the clinical use of radiolabelled antibodies [16], Many tissues often incorporate significant levels of radioac tivity by nonspecific mechanisms [17], How ever, radiolabelled antibodies and antibody fragments have been used with moderate suc cess to image tumors in humans [16, 18-20] and animals [14, 21-25] including the use of MAb 17-1A to detect colon tumor métastasés [42], Consequently, we have evaluated the abil ity of l25I-labelled F(ab')2 fragments of the monoclonal antibody 17-1A to detect HT-29 colon tumor métastasés. Several studies have indicated the advantages of MAb fragments for imaging studies to localize tumors [20-28, 42], In addition, we have attempted to predict tissue radioactivity from planar images of these mice as an indicator of metastatic bur den which might be employed for future dosi metric studies.
Materials and Methods Cell Culture We have used a human colonic adenocarcinoma cell line, HT-29, obtained from American Type Cul ture Collections (ATCC designation HTB-38) for these studies. The derivation of the cell line was originally described by Fogh et al. [27], These cells were main tained as monolayer cultures in 25-cm2 culture flasks containing RPMI 1640 medium supplemented with 10% fetal calf serum, penicillin (100 IU/ml), strepto mycin (1 pg/ml) and /.-glutamine (2 rr\M).
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Métastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Introduction
Murine Monoclonal Antibody 17-1A Fragments F(ab')2 of the monoclonal antibody 171A were used in these studies. The antibody was a gift from Dr. Steplewsky of the Wistar Institute (Philadel phia, Pa., USA). This antibody, originally produced by Herlyn et al. [28], is known to react with 37-kD antigen molecules expressed on gastrointestinal adenocarci noma cells [29]. The purification and fragmentation of this antibody were conducted using standard protocols [30],
Radioiodination o f 17-1A F(ab)2 Fragments MAb 17-1A fragments were iodinated using the chloramine-T method [31]. To describe the method briefly, 10 mg of fragments in phosphate buffered saline (PBS) were mixed with 5 mCi of l25I (as sodium iodide) and 2 mg of chloramine-T in 0.1 ml PBS. The mixture was shaken gently for 7 min, and the reaction was stopped by adding 2 mg sodium metabisulfite in 0.1 ml PBS. The entire reaction mixture was then loaded onto a preequilibrated Sephadex G -10 column, and the column was eluted with phosphate buffer. Thirty i-ml fractions were collected, and the radioac tivity and protein content of each fraction were deter mined using a scintillation counter and a spectropho tometer, respectively. The radiolabelled protein peak was pooled and subjected to quality assurance as described below. Gel Electrophoresis The molecular integrity of the labelled antibody fragments was determined by gel electrophoresis of the labelled and unlabelled products followed by staining with Coomassie blue and autoradiography by methods described elsewhere [32]. Cell-Binding Assay An in vitro cell-binding assay was performed to demonstrate that the radiolabelled antibody retained antigen-binding capacity. A fixed quantity of radiola belled fragments (10 pg) was incubated with variable numbers of HT-29 cells (1 X 10s—50 X 105). Alterna tively, a fixed number (8 X 106) of HT-29 cells was incubated with variable amounts of antibody (250 pg). Incubations were carried out in quadruplicate for 30 min. After incubation, the cells were washed 3 times with PBS containing 1% fetal bovine serum and pelleted. The amount of radioactivity bound was de termined by scintillation counting. In vivo Imaging and Biodislribution of Radiolabelled Antibody Fragments Tumor-bearing or control mice were injected i.v. with radiolabelled fragments. Because these studies were performed over the course of several weeks, the amount of radioactivity injected varied, but an at tempt was made to maintain a constant specific activ ity of the injected antibody. The values expressed as 106 cpm/pg fragments were 0.20 ± 0.07 (range 0.120.25), 0.21 ± 0.08 (range 0.13-0.28), 0.28 ± 0.06 (range 0.13-0.33) and 0.20 ± 0.05 (range 0.12-0.24) at 1, 24, 48 and 72 h, respectively, corresponding to 20-40 pCi of iodine and 200-400 pg of antibody frag ments per mouse. Most mice were imaged daily until
197
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Human Colonic Tumor Xenografts These studies employed male and female outbred Swiss nude (nu/nu) mice, 6-8 weeks old, raised in our facility from breeding pairs kindly provided by Dr. Carl Hansen of NIH. Mice received tumor cither as 3 X 106—5 X 106 live tumor cells injected in 100 pi saline subcutaneously (s.c.) into the right rear flank, or intrasplcnically (i.s.) in 200 pi saline 3 days after intra venous (i.v.) injection of 50 pi (500 pg anti-asialoGM1 antibody (Waco, Dallas, Tex., USA). The s.c. route gives rise to a primary xenograft and largely abdominal métastasés [3]. The i.s. route does not usually lead to the establishment of a primary xeno graft and was suspected to be a better method of estab lishing lung métastasés [5], For the i.s. injections, ani mals were anesthetized using pentobarbital (50 mg/ kg), and the spleen was gently maneuvered into posi tion along an incision in the flank. It was then secured by gently placing (but not closing) a hemostat around the pedicle. Cells were then injected in 0.2 ml, and a small piece of gel foam sponge was used to stem any bleeding or leakage. One minute later the vessels of the splenic pedicle were ligated, the spleen was then re moved and the incision was closed in layers. The entire procedure was conducted ascptically and with care to ensure that the incision site and tissues did not become dry during the surgery. The mice were observed daily and used for further studies of metastasis detection by the radioimmunoconjugate after a period of 2-6 weeks, at which time the subcutaneous tumors were 0.5-1.0 cm in diameter. A group of handled, noninjected controls was prepared. For studies on the halflife of retained radioactivity, at least 3 mice from each group were autopsied after 1, 24, 48 or 72 h. Primary tumor xenografts were only available for study of radioactivity from the mice transplanted with s.c. tumors. Images were processed for detection of métas tasés in 6 mice (3 receiving s.c. tumor and 3 i.s. tumor) at 48 h, and in 8 mice (5 receiving s.c. tumor and 3 i.s. tumor) at 72 h. Multiple tissues from these mice were processed histologically at necropsy.
Necropsy Procedures After the final imaging, animals were killed and various organs and tissues were removed, observed grossly for tumor, weighed, and radioactivity was counted using a 7-counter. Whole organs (liver, lung, intestine, pancreas and spleen) were fixed in 10% neu tral buffered formalin for histological analyses. After blind coding, multiple sections cut from these organs using standard techniques and stained with hematoxy lin and cosin were examined for evidence of tumor metastases and other pathology. The presence of meta static foci determined histologically was compared with gross observations, imaging and the radioactivity in the same organ. Statistical Analyses Numerical values were expressed as the mean and SD and evaluated employing Student’s t test. The radioactive clearance data were subjected to log trans formation and fitted by least-squares analysis. The cpm/pixel data were further analyzed using a multivar iate analysis of variance. The criterion for statistical significance was p < 0.05.
198
Results Analysis o f Labelled Product
The pooled radiolabelled fragments sub jected to SDS polyacrylamide gel electropho resis with Coomassie blue staining and auto radiography exhibited a major and several minor molecular weight species. The approxi mately 96-kD F(ab')2 fragment predomi nated, but there was minor contamination with intact immunoglobulin (150kD) and also a very minor band of approximately 50 kD which was probably free heavy chain. Most of the radioactivity was attached to the F(ab')2 fragment and the bulk of the remain der was attached to whole immunoglobulin. The molecular integrity of the F(ab')2 frag ment was preserved after radioiodination. In vitro Binding o f 17-1A F(ab)2 Fragments to HT-29 Colonic Tumor Cells
A significant (p < 0.001) linear relation ship existed between the bound radioactivity and the amount of antibody fragment added to a fixed number of HT-29 tumor cells (fig. la). Similarly, a significant (p < 0.001) linear relationship existed between the bound radioactivity and the number of HT-29 tumor cells labelled with a fixed amount of antibody (fig. lb). These results demonstrate that the antigen-binding ability of the radiolabelled product was retained. Blood Clearance o f 17-1A F(ab)2 Fragments
The blood radioactivity was determined at various times after i.v. injection of the 17-1A F(ab')2 fragments into control nude mice and mice receiving a s.c. or i.s. inoculation of HT29 tumor cells (fig. 2). The clearance of ra dioactivity from the tumor in s.c. injected mice, and clearance from the blood in all groups was logarithmic, and the biological half-life (Ti/,) calculated by linear regression
Joshi/Mann/T empero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Metastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
autopsy. Initially, thyroid uptake of radioactivity was blocked by administering potassium iodide (16 mg/ml, 0.01 ml/g body weight intrapcritoncally 3, 6 and 18 h prior to radioconjugate administration). However, it was noted that thyroid uptake provided a convenient anatomical marker on the images, therefore this proce dure was discontinued. For imaging, the mice were anesthetized (sodium pentobarbital 50 mg/kg body weight) and taped prone on a styrofoam block. They were imaged using a Technicare 420 mobile Angergamma camera. An analogue film was recorded and the image stored digitally on magnetic media for later analysis. The digital images were displayed for analysis on a Picker PCS-512 com puter system. Regions of interest (ROI) were selected and outlined using a multiposition joystick which per mitted the definition of irregular areas or shapes for determining cpm/pixel. The images which demon strated tumors and suspected metastases at 48 and 72 h post-injection were analyzed independently by 4 investigators to determine cpm/pixel in sites of sus pected tumors as well as in heart, lung, liver, spleen (when present) and thigh muscle (as a remote back ground area). These values were expressed as ratios rel ative to the heart, and plotted as a function of cpm/mg for the same tissue or tumor relative to cpm/mg blood determined at necropsy.
F(ab')2, pg
Fig. 2. Clearance of l25I-labelled 17-1A F(ab'h MAb frag ments. ----- = Regression for blood radioactivity from control (o), s.c. injected (A) and i.s. injected (□) mice: half-life = 8.6 h: r = 0.98. -----= Regression for tumor (•) ra dioactivity; half-life = 30.1 h: r = 0.92.
based on all data points was 30.1 h in the tumor and 8.6 h in blood. Xenograft Localization o f the Radioconjugate
Nude mice injected s.c. with HT-29 tumor cells developed tumors at the site of injection. In order to confirm the in vivo localization of 17-1A F(ab')2 fragments, the ratio of radioac
samples, r = 0.99, p < 0.001. b Fixed amount (10 gg) of F(ab')2 vs. variable number of HT-29 cells. Quadrupli cate samples, r = 0.96, p < 0.001.
Hours after injection
tivity per gram of tumor versus various tis sues was determined (fig. 3). It can be seen that the ratios increased with time. This in crease presumably reflects the continuing clearance of the antibody fragment from the blood pool while that in the tumor is retained, as has been previously reported [33, 34], Our results suggest that the tumor/normal tissue ratio of antibody fragments is still increasing
199
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Fig. 1. In vitro binding of radiolabellcd 17-1A MAb fragments to HT-29 colonic adenocarcinoma cells, a Fixed number of HT-29 cells vs. variable amounts of F(ab'h MAb fragments. Quadruplicate
Fig. 3. Ratios of s.c. xenograft to organ radioactivity in HT-29tumor-bearing nude mice injected with l25I-labelled 17-1A F(ab')2 MAb fragments. o = Blood; • = lung; A = liver; ▲= pancreas; □ = mesenteric lymph node; ■ = spleen; X = muscle.
even at 72 h after antibody injection. Some of the radioactivity ratios were at the upper lim its of those previously reported [14. 21, 22], For example, at 72 h after antibody fragment injection, we observed tumor to blood ratios averaging 90:1, to lung 65:1. to pancreas 50:1, to muscle 35:1, to spleen 15:1, and to liver and mesenteric lymph node 10:1. There was some variation in size of the primary subcuta neous xenografts (0.5-1.0 cm diameter) but no correlation between their size and their content of radioactivity relative to blood lev els. Regardless of their size, all primary xeno grafts were readily imaged. In vivo Localization o f Metastatic Tumor
The primary tumor was clearly localized in all the nude mice receiving HT-29 tumor cells s.c., and metastases were strongly suspected by all observers in the lung or abdominal sites in 9 of 14 of the recipients of s.c. or i.s. tumor cells. Sample scans (gamma camera images) are shown in figure 4. When compared to the control (fig. 4a), the image in figure 4b shows both the primary xenografts and an increased localization of radioactivity in the liver. This
200
was shown by gross and histological examina tions of this sample to represent metastatic tumor (fig. 5c, d). Similarly, figure 4c shows an increased localization of radioactivity in the lungs of a recipient of i.s. tumor cells (no primary xenograft). This was shown by histo logical analysis to represent extensive meta static tumor (fig. 5b). In 6 mice receiving i.s. injection of tumor cells, 1 had lung metas tases, 1 liver metastases, 1 pancreas and mes enteric lymph node metastases and 2 both liver and lung metastases, and one of the latter also had mesenteric lymph node metastases. Table 1 summarizes the results of tissue and organ radioactivity determination in con trol nude mice without tumor compared to mice receiving HT-29 tumor cells but without evidence of metastasis in a particular tissue, and to suspected individual metastatic le sions. Suspected metastatic lesions were ini tially identified with scintigraphy, and con firmed either from gross observations of the dissected organ or by subsequent histopathological analysis. It can be seen that most meta static lesions had radioactivity levels 2-7 times the average level of radioactivity in that
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Metastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
Hours after injection
Fig. 4. Sample Anger camera images of nude mice 72 h after injection with 125I-labcllcd 17-1A F(ab')2 MAb fragments, a Control mouse without tumor. The thyroid was not blocked, providing orientation, b HT-29-tumor-bearing mouse with s.c. xenografts (P) and liver metas tasis (Li). Thyroid activity was not blocked, c Mouse injected i.s. with HT-29 tumor cells
a
c
b
d Fig. 5. Histopathological analyses of HT-29 pri mary tumors and their metastasis in nude mice, a A primary HT-29 tumor growing in a subcutaneous site in a nude mouse. X 45. b Lung metastasis of HT-29 tumor cells in a nude mouse. X 45. c Liver metastasis
of HT-29 tumor cells evident largely as inflammatory cells with a small number of peripheral tumor cells. X 45. d Higher magnification of the periphery of a liver metastasis to demonstrate the presence of tumor cells. Some inflammatory cells are also present. X 180. HE.
201
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
showing lung metastasis (Lu). Thyroid radioactivity was blocked.
Table 1. Average levels of radioactivity in various tissues of HT-29-bearing nude mice suspected on the basis of imaging and gross examination to be free of metastatic disease, compared with individual levels of radioactiv ity in sites of suspected metastatic disease, and levels in nontumor bearing nude mice
Tissue
Radioactivity, % injected dose/mg, X 104 72 h (n = 7)
48 h (n = 7) control nontumor bearing
tumor bearing suspected free of metastasis3 metastasis
control nontumor bearing
tumor bearing suspected metastasis3 free of metastasis
Blood Muscle Lung
0.25 ±0.07 ND 0.20 ±0.04
0.36 ± 0.21 ND 0.39 ±0.19
0.11 ±0.04 0.11 ±0.05 0.23 ±0.06
0.12±0.02 0.12±0.02
Pancreas Spleen Liver
0.36 ±0.16 1.48 ±0.49 1.53 ± 0.33
0.35 ±0.15 1.28 ±0.36 l.29±0.61
0.11 ±0.04 1.09 ±0.13 1.03 ±0.12
0.21 ±0.07 0.33 ±0.06* 0.79 ±0.09
Mesenteric lymph node
0.98 ±0.22
1.38 ±0.92
0.16 ± 0.05
0.49 ±0.18
NA NA 0.96b (2.5) 0.78b (2.0) 2.59(7.4) -
2.18(1.7) 2.51 (2.0) 2.73(2.0) 3.21 (2.3)
0.10 ± 0.03
NA NA 2.64(22) 60.34(503) 0.45(2.1) -
3.22 (4.1) 0.92b (1.2) 1.67(3.4) 1.54(3.1)
* p < 0.05 in comparison with control. ND = No data; NA = not available. a Tissues suspected to contain metastatic tumor on the basis of imaging were carefully examined at necropsy for tumor whose presence was subsequently confirmed histologically. Note that in the case of abdominal metastasis the exact location, e.g. pancreas versus mesenteric lymph node, could not be determined from the image. Values in parentheses are ratios to metastasis-free levels. b Suspected on the basis of gross appearance (nodules) or tissue sectioning, not on an image.
202
Examples of the histopathological confir mation of metastatic tumor detected in liver and lung are presented in figure 5. Figure 5a shows a primary xenograft of the human co lonic adenocarcinoma cell line HT-29. Fig ure 5b shows metastatic tumor in the lung of a nude mouse injected i.s. with HT-29 tumor cells. Figures 5c and 5d demonstrate a liver metastasis of HT-29 tumor cells. On the basis of whole-body imaging and by counting the radioactivity, métastasés were predicted in these organs. The histological studies con firmed this prediction. Predictive Value o f Images
Images of control mice at various times were useful for comparison with images of
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Métastasés
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
particular tissue in metastasis-free mice. Two values, 1.3 and 1.7, were lower than this range and two were higher (22 and 503). The lowest value of 1.3 occurred for the liver of an image negative mouse which had gross tumor nod ules subsequently confirmed histologically to be tumor. The amounts of metastatic tumor in lung and liver varied greatly. The highest value of 503 was obtained for lung which was image positive and extensively involved with metastatic tumor (fig. 5c). In the mice without evidence of metastatic lesions, the tissue lev els of radioactivity were generally similar to those of the control with the exception of the spleen at 72 h, which was significantly lower than the control. The reasons for this differ ence are not known.
Fig. 6. Regression of relative (tumor/blood or tissue/blood) ra dioactivity levels determined from images vs. levels measured by scin tillation counting at autopsy. Slope = 10.52, r = 0.87.
tion of cpm/mg tumor, that together with T% values (fig. 2) permits dosimetric estimates to be made based solely on the image and a blood sample.
Discussion
We have demonstrated the in vivo ra dioimmunodetection of metastasis of human colonic tumor xenografts in two slightly dif ferent animal models. These studies validate the application of these models for the devel opment and evaluation of potential techno logical advances in radioimmunodetection and dosimetry. The model employs a human colonic tumor cell line HT-29 whose charac teristics are similar to those of many primary human colonic adenocarcinomas. It also ben efits from the utility of the nude mouse as a model for studying the behavior of various human tumors, and has already been estab lished [1. 17, 21, 22, 25]. It should be noted that since our objective was to model the likely clinical situation, we elected not to use irrelevant isotope-identical fragments as a control. The use of such a control facilitates
203
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
tumor-bearing mice. Images of tumor-bearing mice at 1 and 24 h showed localization in pri mary subcutaneous xenografts but were not useful for identification of metastatic tumor sites because of blood pool background. Therefore, the analysis of images for meta static tumor sites was limited to images of 7 mice at 48 h, and 7 mice at 72 h. The analysis of images for the prediction of tissue radioac tivity demonstrated that there were no signifi cant differences in the observers’ extraction of data for the tissues studied, with the excep tion of thigh muscle. The observers disagreed on the cpm/pixel values for this site probably because the limits of the images were diffuse and hard to define. The correlation of cpm/ pixel (mean values from all observers) with dpm/mg tissue/blood ratios is shown in fig ure 6. At lower levels of radioactivity the cor relation is linear and significant (p = 0.017). This result indicates that, at least within cer tain limits, the burden of tissue radioactivity can be predicted from the planar image as can the size of the tumor given certain assump tions as to its shape, for example, assuming it is an oblate spheroid. This information to gether with cpm/mg blood permits an estima
204
cols [18, 38, 39], The tumor to normal tissue ratios obtained using this antibody fragment were among the highest reported for epithelial tumors and, as such, offer promise for the use of such antibody fragments in detection [40. 41] of established metastases in this experi mental model. The reasons that the ratios of radioactivity in the tumor to liver and mesen teric lymph node (10:1) are much lower may be, in part, due to the presence of micrometastatic tumors in these tissues in some mice [3], The localization of radioactivity was greater in those sites identified by imaging or gross examination to be suspected sites of metastasis. In general, tumor to tissue ratios for the whole organ ranged from 2-7. This suggests that the metastases being detected represent about 5-20% of the mass of tissues. The exceptions were 2 lung metastases, which had high ratios, suggesting major metastatic lesions. This was confirmed histologically (fig. 5b). In 14 tumor-bearing mice which were imaged at 48 or 72 h, lung metastasis was suspected in 2 of 3 mice and abdominal metastasis in 3 of 6 mice. Metastatic lesions in 2 mice were detected by gross observation (small liver nodules) and were not evident in the images. In 2 other mice, tumor deposits were detected by sectioning the tissue. Thus, detection of metastatic lesions by imaging oc curred in 5 of 9 (55%) which is similar to results reported for studies in humans [18, 42] , Other investigators have selected a ratio of 2 [7] or greater [21] to indicate specific con centrations of radioactivity in tumors. All our primary xenografts and all except 2 metastatic lesions met this criterion. The metastatic tu mors observed at 48 h occurred in 2 of 3 mice receiving cells s.c. and in 3 of 4 mice receiving i.s. tumor cells. At 72 h metastases were found in 2 of 5 s.c. tumor cell recipients and 2 of 2 i.s. recipients. The frequency of metastasis in the two models, 4 out of 8 s.c. versus 5 out of 6 i.s., is not significantly different (p = 0.24).
Joshi/Mann/Tempero/Crousc/ Stratbuckcr/Quaife/Sliarp
Radiolocalization of Colon Tumor Metastases
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
the evaluation of specific radiolocalization but complicates the interpretation of the images obtained to detect sites of metastatic tumor. The in vitro binding of the 17-1A anti body fragments to HT-29 cells indicated that these antibody fragments should also be use ful in vivo as was shown to be the case [20, 36]. This approach may not be universally applicable because of antigenic variation be tween cells of the same tumor, and the possi bility that potentially useful antigens ex pressed by the tumor in vivo might not be expressed by the same tumor cells growing in vitro [14, 37], This emphasizes the need for suitable in vivo experimental models. Our observations confirm previous reports of the efficacy of F(ab')2 fragments of the MAb, 17-1A for the radioimmunodetection of human colonic tumors [42], The biological half-life of these fragments administered i.v. in our model was 8.6 h which is similar to the range of 5-16 h reported for MAb F(ab')2 and an irrelevant control [26], In therapeutic ap plications, there may be concern that the rapid clearance of antibody fragments would decrease the dose delivered to the tumor [43], However, the therapeutic ratio defined as the dose to the tumor versus the dose to critical normal organs would seem to be crucial. Po tentially, there may in fact be an advantage, in studies which propose to combine imaging with therapy of tumors, in employing a mix ture of fragments and whole immunoglobulin. The former would permit more rapid detec tion of tumor sites whereas the whole immu noglobulin would remain in the tumor longer and deliver a greater cumulative dose of ra dioactivity [26, 35], It is still far from clear which antibody fragments or whole immuno globulin are best. Additionally, the host im mune response to fragments may be expected to differ from that towards whole immuno globulin, and these differences might become significant in multiple administration proto
However, both models gave rise to macro scopic human tumor métastasés. These data also suggest that metastatic deposits as small as 20 mm3 could be detected as suggested by radioactivity and confirmed by histological examination. Consequently, this approach has merit for the quantitation of metastatic tumor burden.
Acknowledgements Wc thank Dr. Kashinath Patil and Dr. Greg Perry for their assistance with the statistical analyses, Dr. Shailendra Saxena and Mr. Richard Kelley for excel lent technical assistance, and Ms. Roberta Anderson for typing the manuscript. This work was partly sup ported by Nebraska Department of Health LB506 funds.
References
JAMA 1989:26 i: 744-746.
8 Ritz J, Schlossman SF: Utilization of monoclonal antibodies in the treatment of leukemia and lympho ma. Blood 1982;59:1-11. 9 Byers VS, Baldwin RW: Thera peutic strategies with monoclonal antibodies and immunoconjugates. Immunology 1988:65:329—335.
10 Goldenberg DM: Current status of cancer imaging with radiolabclled antibodies. J Cancer Res Clin Oncol 1987:113:203-208. 11 Blumenthal RD. Kashi R. Stephen R. et al: Improved radioimmuno therapy of colorectal cancer xeno grafts using antibody mixture against carcinoembryonic antigen and colon-specific antigen-P. Can cer Immunol Immunother 1991:32: 303-310. 12 Eary JF: Fundamentals of radioim munotherapy. Int J Rad Appl lnstrum [B] 1991;18:105-108. 13 Riva P. Marangala M. Tison V, et al: Treatment of metastatic colorectal cancer by means of specific mono clonal antibodies conjugated with iodine-131, a phase II study. Int J Rad Appl Instrum [B] 1991:13:109119. 14 Buchsbaum D, Lloyd R.Juni J. et al: Localization and imaging of radiola beled monoclonal antibodies against colorectal carcinoma in tu mor-bearing nude mice. Cancer Res 1988;48:4324-4333. 15 Sharkey RM. Wcadock KS, Natale A. et al: Successful radioimmuno therapy for lung métastasés of hu man colonic cancer in nude mice. J Natl Cancer Inst 1991:83:627-632. 16 Carrasquillo JA, Sugarbaker P. Colcher D, et al: Radioimmunoscintigraphy of colon cancer with io dine- 131 -labeled B72.3 monoclonal antibody. J Nucl Med 1988:29: 1022-1030.
17 Sharp JG, Coop KL, Cheng HF, ct al: Scintiscanning and tissue distri bution studies of radiolabelled anti tumor antibodies in rats bearing the corresponding intestinal tumor. Eur J Nucl Med 1982:7:28-34. 18 Rosen ST, Zimmer AM. GoldmanLekin R. ct al: Radioimmunodetec tion and radioimmunotherapy of cutaneous T cell lymphomas using l31I-labellcd monoclonal antibody: An Illinois Cancer Council Study. J Clin Oncol 1987;5:562-573. 19 Colcher D, Esteban JM, Carras quillo JA, et al: Quantitative analy ses of selective radiolabelled mono clonal antibody localization in met astatic lesions of colorectal cancer patients. Cancer Res 1987:47:11SS II 89. 20 Welt S, Chaitanya R, Divigi R. et al: Quantitative analysis of antibody localization in human metastatic co lon cancer: A phase I study of mono clonal antibody A33. J Clin Oncol 1990;8:1894-1906. 21 Stya M, Wahl RL. Natale RB. et al: Radioimmunoimaging of human small cell lung carcinoma xenografts in nude mice receiving several monoclonal antibodies. Natl Cancer Inst Monogr 1987;3:19-23. 22 McCabe RP. Peters LC. Haspel MV, et al: Prcclinical studies on the phar macokinetic properties of human monoclonal antibodies to colorectal cancer and their use for detection of tumors. Cancer Res 1988:48:43484353.
205
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
1 Dore JF, Bailly M. Bertrand S: Mé tastasés of human tumors in experi mental animals. Anticancer Res 1987:7:997-1004. 2 Kozlowski JM. Fidler 1J, Campbell D, et al: Metastatic behavior of hu man cell lines grown in nude mice. Cancer Res 1984:44:3522—3529. 3 Joshi SS, Jackson JD. Sharp JG: Comparison of the growth and me tastasis of four human intestinal tu mor cell line xenografts. Tumor Biol 1989:10:117-125. 4 Joshi SS, Mathews NB. Sinangil F. et al: Effects of differentiation in ducing chemicals on in vivo malig nancy and NK susceptibility of met astatic lymphoma cells. Cancer De tect Prev 1988:11:405-417. 5 Lafreniere R, Rosenberg SA: Suc cessful immunotherapy of murine experimental hepatic métastasés with lymphokine activated killer cells and recombinant interleukin-2. Cancer Res 1985:45:3735-3741. 6 Joshi SS. Sharp JG, Brunson K.W: Differential growth characterization of metastatic variant RAW117 mu rine lymphosarcoma cells. Oncology 1987:44:180-185. 7 Schlom J: Innovations in mono clonal antibody tumor targeting.
31 Hunter WM: Radioimmunoassay; in Weir DM (ed): Handbook of Ex perimental Immunology. Oxford, Blackwell, 1978, pp 14-33. 32 Joshi SS, Sharp JG, Gharpurc HM, ct al: Characterization of metastasisassociated antigens on RAW 117 lymphosarcoma cell lines. Clin Exp Metastasis 1987;5:89-104. 33 Pimm MV, Perkins AC, Armitage NC, et al: The characteristics of blood-borne radiolabels and the ef fect of anti-mouse IgG antibodies on localization of radiolabelled mono clonal antibody in cancer patients. J Nucl Med 1985;26:1011-1023. 34 Fogler WE, Kinger MR, Abraham KG, et al: Enhanced cytotoxicity against colon carcinoma by combi nations of noncompeting mono clonal antibodies to the 17-1A anti gen. Cancer Res 1988;48:63036308. 35 Hand PH, Colcher D. Salmon D, et al: Influence of spatial configuration of carcinoma cell populations on the expression of a tumor-associated glycoprotein. Cancer Res 1985;45: 833-840. 36 Mach JP. Buchegger F, Forni M, et al: Use of radiolabelled monoclonal anti-CEA antibodies for the detec tion of human carcinomas by exter nal photoscanning and tomoscinti graphy. Immunol Today 1981:2: 239-249.
37 Esteban JM. Seldom J,FrancoiseM, et al: Radioimmunothcrapy of athymic mice bearing human colon carcinomas with monoclonal anti body B72.3: Histological and auto radiographic study of effects on tu mors and normal organs. Eur J Can cer Clin Oncol 1987;23:643-655. 38 Esteban JM.Schlom J.GansowOA, et al: New method for the chelation of indium-111 to monoclonal anti bodies: Biodistribution and imaging of athymic mice bearing human co lon carcinoma xenografts. J Nucl Med 1987;28:861-870. 39 Sharkey RM, Kaltovich FA, Shih LB, et al: Radioimmunotherapy of human colonic cancer xenografts with 90Y-labelcd monoclonal anti bodies to carcinoembryonic antigen. Cancer Res 1988;48:3270-3275. 40 Buchegger F, Vacca A, Carrel S, et al: Radioimmunotherapy of human colon carcinoma by l3,I labeled monoclonal anti-CEA antibodies in a nude mouse model. Int J Cancer 1988;41:127-134. 41 LoBuglio AF, Saleh MN, Lee J, et al: Phase I trial of multiple large doses of murine monoclonal antibody C017-1A. I. Clinical aspects. J Natl Cancer Inst 1988;80:932-936. 42 Mach JP, Chatal JF, Lumbroso JD, et al: Tumor localization in patients by radiolabelled monoclonal anti bodies against colon carcinoma. Cancer Res 1983;43:5593-5600. 43 Colcher D. Zalutsky M, Kaplan W, et al: Radiolocalization of human mammary tumors in athymic mice by a monoclonal antibody. Cancer Res 1983;43:736-742.
206
Joshi/Mann/Tempero/Crouse/ Stratbucker/Quaife/Sharp
Radiolocalization of Colon Tumor Métastasés
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/28/2018 11:56:21 PM
23 Endo K, Kamma H, Ogata T: Ra diolabeled monoclonal antibody 15 and its fragments for localization and imaging of xenografts of human lung cancer. J Natl Cancer Inst 1988;80:835-842. 24 Wahl RL, Barret J, Geatti D. ct at: The intraperitoneal delivery of radiolabelled monoclonal antibodies: Studies on the regional delivery ad vantage. Cancer Immunol Immunother 1988;26:187-201. 25 Herlyn D. Powe J, Alavi A, et al: Radioimmunodetection of human tumor xenografts by monoclonal an tibodies. Cancer Res 1983:43:2731— 2735. 26 Goodwin DA: Pharmacokinetics and antibodies. J Nucl Med 1987; 28:1358-1362. 27 Fogh J, Fogh JM, Orefo T: One hundred and twenty-seven cultured human cell lines producing tumors in nude mice. J Natl Cancer Inst 1977;59:221-226. 28 Herlyn M. Steplewski Z, Herlyn D: Colorectal carcinoma specific anti gens: Detection by means of mono clonal antibodies. Proc Natl Acad SciUSA 1979;76:1438-1442. 29 Gottlinger HG, Funke I, Johnson JP: The epithelial cell surface anti gen 17-1 A, a target for antibody me diated tumor therapy: Its biochemi cal nature, tissue distribution and recognition by different monoclonal antibodies. Int J Cancer 1986;38: 47-53. 30 Sears H, Herlyn D, Steplewski Z: Initial trial use of murine mono clonal antibodies as immunotherapcutic agents for gastrointestinal ad enocarcinoma. Hybridoma 1986;5: 109-115.
