Localisation of malignant germ-cell tumours by external scanning after injection of radiolabelled anti-alpha-fetoprotein

Tumor scintigraphy by the method for subtracting the initial image with technetium-99m labeled antibody

The method for subtracting the initial image from the localization image was evaluated for radioimmunoscintigraphy of tumors with technetium-99m (Tc-99m) labeled antibodies. Monoclonal antibodies were parental mouse and mouse-human chimeric antibodies to carcinoembryonic antigen (CEA), designated F11-39 and ChF11-39, respectively, both of which have been found to discriminate CEA in tumor tissues from the CEA-related antigens. After reduction of the intrinsic disulfide bonds, these antibodies were labeled with Tc-99m. In vivo studies were performed on athymic nude mice bearing the human CEA-producing gastric carcinoma xenografts. Though biodistribution results showed selective and progressive accumulation of Tc-99m labeled antibodies at the tumor site, high radioactivity in blood was inappropriate for scintigraphic visualization of the tumors within a few hours. We examined the subtraction of the initial Tc-99m image from the Tc-99m localization image after a few hours. Subtracted images of the same count reflected the in vivo behavior of the Tc-99m radioactivity. The subtracted scintigrams revealed excellent tumor images with no significant extrarenal background. Visualization of the tumor site was dependent on antigen-specific binding and nonspecific exudation. These results demonstrate that a method of subtraction of the initial image may serve as a potentially useful diagnostic method for an abnormal site for agents with a low pharmacokinetic value.

Accessibility to intracellular antigens within nutritionally deprived human mammary epithelial cells

We have previously demonstrated immunolocalization of antikeratin antibodies in apparently random subpopulations of malignant cells in fresh surgical specimens of breast carcinoma (S. H. Dairkee and A. J. Hackett, 1988, J. Natl. Cancer Inst. 80, 1216-1220). The goal of the present study was to determine whether deficiencies in essential nutrients contribute toward cellular alterations in membrane integrity, consequently allowing antikeratin to bind to the cytoskeleton within live, unfixed cells. We have demonstrated here that in an in vitro model in which human mammary epithelial cells are subjected to an oxygen-glucose gradient, immunolocalization of antikeratin within the cells is observed in a dose-dependent manner in the depleted regions of the gradient, even though the cells appear to be morphologically unaltered. The potential use of antibodies to intracellular antigens for immunotargeting solid tumors and the use of this method in antibody-loading studies toward understanding functional aspects of specific cellular antigens, as well as determining differential response of various cell types under these culture conditions, are discussed.

Immunoscintigraphy with 131I-labelled H17E2 monoclonal antibody compared with conventional lymphangiography and computed tomography in the detection of metastases in patients with testicular germ cell tumours

131I-labelled H17E2 monoclonal antibody (MAb) was administered to 16 patients with germ cell tumours of the testis (GCT). Eleven patients had non-seminomatous GCT and five seminoma. The MAb was administered into the webs between the second and third toes of both feet in 12 patients and intravenously in four patients at a dose of 1.5-2mCi. 131I-labelled 2-118 MAb (non-specific) was administered subcutaneously into the webs between the second and third toes of both feet in two patients and intravenously in one patient with non-seminomatous GCT. All three patients had only computed tomography (CT) scan. Patients were scanned immediately after until 7 days post-injection. For comparison all patients had CT scan and eight out of 16 patients had conventional lymphangiography (LG). When the radiolabelled MAb was given subcutaneously, the immunoscan (IS) was true positive in 9/12 (75%) patients and true negative in 2/12 (16.5%) and equivocal in 1/12 (8.5%). The LG gave true positive results in 6/8 (75%) patients and true negative results in 2/8 (25%) and the CT scan true positive results in 8/12 (66.6%) patients, true negative results in 2/12 (16.3%) and false negative results in 2/12 (16.3%). There was an excellent correlation of IS images with the LG results (true positivity 100%). When the radiolabelled MAb was given intravenously, both IS and CT scan gave true positive results in four cases. Our findings showed that the true positivity of IS reached 93.8%, whereas that of CT scan 87.5%. In all three patients who had the 131I-labelled 2-118 non-specific MAb, the IS was false negative, whereas the CT scan was true positive. Thus, this procedure may offer information complementary to that provided by existing conventional imaging methods.

Influence of circulating antigen on the biodistribution and tumour localization of radiolabelled monoclonal antibody in a human tumour: nude mouse xenograft model

A monoclonal antibody, against a colorectal carcinoma tumour-associated antigen, was radioiodinated and its biodistribution studied in comparison with that of control immunoglobulin in nude mice with colon carcinoma xenografts. Tumour localization of the antibody in comparison with normal tissues was poor, and in absolute terms more control IgG than antibody was present per gram of tumour. This failure to achieve localization could not be ascribed to poor immunoreactivity of the antibody nor to the failure of the xenografts to express the appropriate antigen. Analysis of serum from mice with xenografts showed the presence of circulating tumour-derived antigen. This serum-borne antigen was found to form immune complexes both in vitro and in vivo with the monoclonal antibody, and this complex formation is probably the limiting factor in tumour localization of the antibody. This is one of only few examples where mice with human tumour xenografts have levels of circulating antigen sufficient to perturb biodistribution of antibody. These findings are relevant to the biodistribution of monoclonal antibodies in the clinical situation, since circulating antigen is often found in cancer patients.

The use of a monoclonal antibody against alpha-fetoprotein for the radioimmunodetection of hepatocellular carcinoma

The purpose of this study was to investigate the use of a radiolabeled mouse monoclonal antibody (and its F(ab')2 fragment) against alpha-fetoprotein in the scintigraphic diagnosis of hepatocellular carcinoma. Twenty-six southern African Blacks and one Caucasian with hepatocellular carcinoma and four patients with other malignant tumors of the liver were studied. Although six hepatocellular carcinomas appeared to selectively concentrate alpha-fetoprotein antibody, one of these tumors was not producing alpha-fetoprotein. Moreover, in another 18 patients with alpha-fetoprotein-producing hepatocellular carcinomas, uptake of alpha-fetoprotein antibody was at best only equal to that in nontumorous hepatic tissue, and three hepatocellular carcinomas that were not producing alpha-fetoprotein concentrated the antibody to the same extent as did the alpha-fetoprotein-producing tumors and hepatic tissue. All four tumors other than hepatocellular carcinoma concentrated alpha-fetoprotein antibody as well as did hepatic tissue. These findings suggest that the penetration of hepatocellular carcinomas by alpha-fetoprotein antibody is a passive and nonselective process. This conclusion is supported by an in vitro study in which a non-alpha-fetoprotein-producing hepatic metastasis took up as much radiolabeled alpha-fetoprotein antibody as did three of four alpha-fetoprotein-producing hepatocellular carcinomas. A likely explanation for the failure of alpha-fetoprotein monoclonal antibody to be selectively concentrated by hepatocellular carcinomas is that alpha-fetoprotein is an export protein and is not expressed on the cell membranes of malignant hepatocytes.

Radioimmunoscintigraphy of ovarian cancer with 131-iodine labeled OC-125 antibody fragments

Radioimmunoscintigraphy (RIS) with 131I labeled OC-125 F(ab')2 monoclonal antibody fragments was prospectively studied in 43 women for primary diagnosis and follow up of ovarian cancer. Total body planar photoscans with a scintillation camera were performed one to seven days after antibody application and results were compared with operation and or CT examination. By the region of interest technique the tumor to non tumor tissue ratio (T/N) was calculated in vivo. Sensitivity in primary diagnosis was 100% (10/10), specificity 33% (1/3). For local recurrency, sensitivity was 86% (19/22), for metastatic loci 80% (17/21). Specificity was 75% and 50%. T/N ratio was in the rage from 1.3 to 2.8. Sensitivity for ovarian cancer is high in primary diagnosis and follow up. By the region of interest technique it is possible to detect small recurrencies and to presume peritoneal carcinosis. Antibody accumulations in diseases different from ovarian cancer however diminish specificity.

Immunoscintigraphy with 131I-labelled HMFG2 and HMFG1 F(ab')2 in the pre-operative detection of clinical and subclinical lymph node metastases in breast cancer patients

Radiolabelled specific monoclonal antibodies (MAbs) HMFG2 and HMFG1 F(ab')2 and non-specific 11.4.1 and 4C4 F(ab')2 were injected into the webs between the 2nd and 3rd fingers of both hands in 31 patients with clinical diagnosis of breast cancer. We studied 10 patients with clinically obvious axillary lymph-node disease (group A) and 10 patients with clinically negative axilla (group B) using HMFG2, 5 patients with clinically negative axilla (group C) using HMFG1 F(ab')2 and 6 patients with clinically positive axilla (group D) using non-specific 11.4.1 and 4C4 F(ab')2 MAbs. In group A, 7 patients had true positive scans. There were also 3 false negative scans, due to problems related to proper iodination at the beginning of this study. In group B there were 4 true positive scans, 4 true negative, I false positive and I false negative. In group C there were 4 true negative scans. In one patient the radiolabelled antibody was arrested in the middle of the arm, because of lymphatic obstruction. In group D, there were 3 false negative scans with 11.4.1 antibody and 3 false negative scans with 4C4 F(ab')2 MAb. The results of immunoscintigraphy were in accordance with the histopathology and immunoperoxidase staining findings. These results indicate that this non-invasive approach can accurately detect metastatic involvement in the axillary lymph nodes and can be used for the diagnosis and staging of breast cancer.

Radioimmunodetection of cancer with monoclonal antibodies: current status, problems, and future directions

Early studies of immunoscintography with affinity-purified 131I-labeled polyclonal antibodies reactive against oncofetal antigens such as carcinoembryonic antigen (CEA) were moderately successful in detecting metastatic colorectal carcinoma. However, because of low tumor to background ratios of isotope, background subtraction techniques using 99Tc-labeled albumin were required to visualize small lesions. Antisera were often of low titer and lacked specificity. These problems could be overcome for the most part following the development of highly specific monoclonal antibodies (MoAb) against a variety of tumor-associated antigens. A number of clinical trials using 131I- or 111In-labeled MoAb to image tumors have demonstrated successful localization without the use of subtraction techniques. Variables limiting the usefulness of murine MoAb for diagnosis have included increased localization in liver and spleen, tumor vascularity and heterogeneity of antigen expression, and development of human antimurine globulins. Methods to overcome some of these problems are discussed. Radiolabeled MoAb appear useful as an adjunct to conventional diagnostic techniques both as a means to predict which antibodies might be useful for treatment and, in select patients, as a basis for treatment decisions.

Clinical evaluation of anti-alpha-fetoprotein radioimmunodetection

Anti-alpha-fetoprotein (AFP) radioimmunodetection was performed in response to clinical requests in 16 patients. In two patients, assessment of a known tumour was required; the anti-AFP scans were accurate and provided useful clinical information in both cases. In the remaining 14 patients the request was for localisation of suspected recurrent tumour. Accurate information was provided in four of these patients. In this latter group, various conventional methods of investigation had failed to disclose the site of recurrence. However, of a total of 21 sites reported as positive in these 14 patients, eight proved to be false positives. Two false negative results also occurred in this group and nine could not be evaluated. Although occasionally patients were usefully scanned, improvements are necessary before consistently reliable information can be obtained using this technique.

Radioimmunolocalization of human brain tumours: biodistribution of radiolabelled monoclonal antibody UJ13A

Monoclonal antibody UJ13A, radiolabelled with 131I, was intravenously administered to patients with primary brain tumours. The antigen recognised by UJ13A is present on most neuroectodermally derived tissue. The ratio of uptake in tumour to normal brain, assessed by scintigraphy, improved with time. Maximal tumour uptake occurred between 4 and 48 h. Dynamic and static scintigrams indicated some early sequestration of radiolabelled antibody by the liver. Tumours were surgically resected in seven patients at various intervals after antibody administration showing tissue to blood ratios increasing with time in all parts of the lesion (viable and necrotic tumour, cyst fluid), and in normal brain. The highest tissue to blood ratio in viable tumour was 5.1 at 16 days after injection. In tissues resected 2-3 days after injection there was relatively greater uptake in the viable tumour compared to necrotic tumour and cyst fluid. In contrast, tissues resected later (6-16 days) showed greater uptake in ischaemic tissue than viable tumour, suggesting diffusion was an important factor influencing tumour uptake. The amount of radioactivity per gram of tumour tissue was less than 0.005% of the injected dose. Future studies are needed using different antibodies, antibody fragments and additional methods of optimising delivery.

Radionuclide imaging of ovarian tumours with 123I-labelled monoclonal antibody (NDOG2) directed against placental alkaline phosphatase

A radiolabelled monoclonal antibody (NDOG2) directed against placental alkaline phosphatase (PLAP) was used in the radio-immunodetection of ovarian carcinoma. Tumour deposits were successfully visualized in 11 of 15 patients and the abnormalities demonstrated were classified as focal or diffuse. Of the 11 patients, eight showed focal abnormalities alone and three had a diffuse abnormality, of which two also showed a focal abnormality. False-positive results may occur not only due to uptake of 123I by gut mucosa and an inadequately blocked thyroid gland but also from activity in an incompletely emptied bladder. A false-negative result occurred due to high background activity in the liver masking a known, discrete tumour deposit.

123I radioiodinated antibody imaging of occult ovarian cancer

A monoclonal antibody HMFG2 labeled with iodine 123 (123I) was given to a patient who had ovarian cancer. The scan taken 18 hours after administration of the antibody demonstrated the presence and the position of residual tumor in the pelvis that was not previously detected by ultrasonography and computerized tomography (CT) scanning. The presence of tumor was confirmed by surgery and histologic as well as immunoperoxidase examination of resected tissues. The tumor mass found was less than 0.8 cm in diameter. It is concluded that, in the search for residual or early ovarian cancer in the pelvis, monoclonal antibody scanning using 123I-labeled HMFG2 can complement ultrasonography and CT scanning.

The localization of human breast carcinomas by radiolabelled monoclonal antibodies

Immune-deprived mice bearing HX99 human breast carcinoma xenografts were injected with a radiolabelled monoclonal antibody, LICR-LON-M8 (M8), to investigate the dependence of tumour localization on (i) tumour site and (ii) antibody radiolabel. No significant difference was found in the degree of localization of radio-iodinated M8 in subcutaneous, renal or intracranial xenografts, but a highly significant improvement in HX99 localization by M8 was recorded using an 111indium-DTPA conjugate of the antibody (111In-DTPA-M8), related to its rapid tumour uptake and blood pool clearance. Radio-iodinated or 111In-labelled M8 was given to 29 patients with breast cancer, 7 with primary tumours and 22 with metastases. Tumour localization was assessed by (i) examination of surgical specimens and (ii) antibody scans, which were compared with conventional X-rays and 99mTc-methylene diphosphonate (MDP) bone scans. Radiolabelled M8 localized preferentially in all primary tumours (radioactivity tumour: normal breast = 6.2 +/- 1.4 [mean +/- s.e.]). All ten patients with skeletal metastases had positive 111In-DTPA-M8 scans, but the correlation with X-rays and MDP scans showed a regional variation. Radio-iodinated M8 failed to identify metastases in any site. The favourable biodistribution of 111In-DTPA-M8 has led to the clear localization of breast carcinomas in patients and mice. In future such reagents may rationalize the clinical management of breast cancer.

Immunolocalization of neuroblastoma using radiolabeled monoclonal antibody UJ13A

The monoclonal antibody UJ13A, raised after immunization of mice with human fetal brain, recognized an antigen expressed on human neuroblastoma cell lines and fresh tumors. Antibody was purified and radiolabeled with iodine isotopes using chloramine-T. In preclinical studies, 125I-labeled UJ13A was injected intravenously into nude mice bearing xenografts of human neuroblastoma. Radiolabeled UJ13A uptake by the tumors was four to 23 times greater than that by blood. In control animals, injected with a similar quantity of a monoclonal antibody known not to bind to neuroblastoma cells in vitro (FD44), there was no selective tumor uptake. Nine patients with histologically confirmed neuroblastoma each received 100 to 300 micrograms UJ13A radiolabeled with 1 to 2.8 mCi 123I or 131I. Sixteen positive sites were visible on gamma scans 1 to 7 days after injection: 15 were primary or secondary tumor sites, and one was a false positive; there were two false negatives. In two of the 15 positive sites, tumor had not been demonstrated by other imaging techniques; these were later confirmed as areas of malignant infiltration. No toxicity was encountered.

The inhibitory effects of horse anti-rat AFP antiserum on the uptake of 2-deoxy-D-glucose by AFP-producing rat hepatoma cells

The inhibitory effects of horse antiserum against rat alpha-fetoprotein (AFP) on the uptake of 2-deoxy-D-glucose (2dG) by the AFP-producing rat ascites hepatoma AH66 cells was studied. AH66 cells cultured in medium containing 20% heat-inactivated antiserum had a 1.5-fold lower rate of sugar uptake than did AH66 cells which were cultured in medium containing 20% heat inactivated normal horse serum. The inhibition of 2dG uptake by antiserum was dependent on both the concentration and the exposure time of antiserum. Preincubation of AH66 tumor cells for 2 and 6 h with antiserum prior to the measurement of 2dG uptake resulted in a 70.1% and 58.2% decrease in 2dG uptake compared to control cells. Antiserum did not inhibit the rate of phosphorylation of 2dG by tumor cells. Kinetic constants for the uptake of 2dG in both AH66 cells treated with antiserum to AFP and in control cells were calculated from Lineweaver-Burk plots. The Km remained constant at approximately 1.2 mM, but the Vmax was twice as small for the cells treated with antiserum as for the control cells (571 vs 923 nanomoles/2 X 10(5) cells/min). These studies suggest that the inhibition of 2dG uptake by treatment with antiserum was the result of a decrease in the number of transport sites, or a decrease in the amount of carrier protein for the sugar which was present on the surface of the plasma membrane of the AH66 cells.

Nuclear imaging techniques with radiolabelled antibodies

The behaviour of radioimmunolocalisation with respect to neoplasia is reviewed. The present limitations are outlined and several new approaches to improve its clinical utility are discussed.

Improved tumour localisation using indium-111 labelled antibodies

Immunoglobulin G (IgG) antibodies to carcinoembryonic antigen (CEA) were labelled with radioactive indium (111In) or iodine (131I) and a comparison made of their value in locating CEA producing tumours. Eleven patients given 111In-anti-CEA had 31 tumours as judged by a combination of all techniques. Of these, 28 were detected by 111In-anti-CEA and 26 by conventional clinical techniques. Five of the patients also received 131I-anti-CEA. These patients had 15 tumour areas. Thirteen were detected by 111In and eight by 131I. 111In also produced a better signal to noise ratio in the scans and thereby showed lesions with greater certainty. In addition, the 111In isotope continued to accumulate in the tumour areas for considerably longer than 131I. Absorbed doses (whole body) were similar for both isotopes. The results show that antibody scanning is greatly improved by using 111In as the radiolabel in place of 131I and should allow the detection of smaller or deeper lesions.

Detection of thyroid tumours using radio-labelled anti-thyroglobulin

IgG antibody to human thyroglobulin was labelled with 131Iodine (131I) and used to locate deposits of thyroid follicular and papillary tumours with a gamma camera. Of twelve patients studied a total of 40 tumour 'areas' were detected by a variety of clinical and radiological techniques. Sixteen of these were detected using conventional 131I uptake scans whereas 34 were positive on the antibody scans. The difficulty of assessing diffuse pulmonary lesions (3 areas) and the possibility that free 131I from labelled antibody may have contributed to the antibody scan results in six areas left 31 definite areas for scan comparison. Twenty seven (87%) areas were positive on the antibody scan, nine (29%) were positive on conventional 131I scans whilst 24 (77%) areas were detected by a combination of clinical and other radiological criteria. Five areas were positive on the antibody scan alone but there was evidence, albeit indefinite, that these areas contained tumour. Four of the 31 areas were not detected by the antibody scans. The results indicate that anti-thyroglobulin scanning is more sensitive than conventional 131I-iodide scans and may contribute to the staging and management of thyroid cancer.

Targeting of iodine-123-labelled tumour-associated monoclonal antibodies to ovarian, breast, and gastrointestinal tumours

Two tumour-associated monoclonal antibodies, HMFG1 and HMFG2, were labelled with iodine-123 and used to detect primary and metastatic ovarian, breast, and gastrointestinal neoplasms by external body scintigraphy in twenty patients with advanced disease. Tumours became visible 3 min to 18 h after injection of labelled antibody. The presence of antibody in the tumours was confirmed by autoradiography and immunoperoxidase staining of surgically removed tissues. The mean tumour uptake of radiolabel was 0.6% of the injected amount. These antibodies can therefore localise specifically to tumours and successful imaging can thus be achieved. This method can complement existing diagnostic techniques and also provide a basis for a selective therapeutic approach to malignant disease.