Search

An enzyme-linked immunosorbent spot (ELISPOT) assay for interferon gamma production has been used to analyze specific T cell responses to a Flu 9-mer peptide, and a 9-mer peptide of carcinoembryonic antigen (CEA). Assays were performed on peripheral blood mononuclear cells (PBMC) of HLA-A2-positive patients with CEA-expressing carcinomas, both before and after vaccination with CEA-based vaccines, and from HLA-A2-positive healthy blood donors. The ELISPOT assay utilized aliquots of frozen PBMC, and assays were performed after 24 h in culture with peptide to rule out any artifacts due to long-term in vitro stimulation cycles. An internal standard was used for each assay to define reproducibility of the assay, and all samples from a given patient (pre- and post-vaccination, with both the Flu and CEA peptides) were analyzed simultaneously. The results indicated a trend towards healthy blood donors having higher levels of Flu-specific T cell precursors than do colon carcinoma patients, but these results were not statistically significant (P = 0.06). On the other hand, slightly higher CEA-specific T cell responses were observed in cancer patients with CEA-expressing carcinomas than in healthy blood donors. PBMC from two CEA-based vaccine clinical trials were analyzed for T cell responses to the same CEA peptide and to the Flu control peptide. The first trial consisted of three monthly vaccinations of CEA peptide (designated PPP) in adjuvant. The second trial consisted of cohorts receiving three monthly vaccinations of avipox-CEA recombinant (designated AAA) or cohorts receiving a primary vaccination with recombinant vaccinia-CEA followed by two monthly vaccinations with avipox-CEA (designated VAA). Few, if any, CEA-specific T cell responses were seen in the PPP vaccinations, while the majority of patients receiving the poxvirus CEA recombinants demonstrated increases in CEA-specific T cell responses and no increases in Flu-specific responses. CEA-specific IgG responses were also demonstrated in patients following recombinant CEA poxvirus vaccinations. Statistical analyses of the T cell responses to the same CEA peptide demonstrated a P value of 0.028 for the recombinant poxvirus vaccines, as compared with the peptide vaccine. There were no differences seen (P = 0.37) in Flu-specific responses after these two types of CEA vaccination. These results thus provide the first evidence that poxvirus recombinant-based vaccines are more potent in the initiation of tumor-antigen-specific T cell responses than vaccines employing peptide in adjuvant, when assays are conducted in an identical manner, and in defining responses to the same peptide. These results also demonstrate for the first time that an ELISPOT assay, performed over a 24-h period and without in vitro sensitization, can be successfully used to monitor immune responses to a tumor-associated antigen in cancer patients.

Monoclonal antibody (mAb) CC49, a murine IgG1, reacts with the tumor-associated glycoprotein-72 expressed on a variety of carcinomas. In clinical trials, radiolabeled CC49 has shown excellent tumor localization to a variety of carcinomas. To minimize the immunogenicity of CC49 mAb in patients, a humanized CC49 (HuCC49) was generated by complementarity-determining region (CDR) grafting. The relative affinity of HuCC49 was 2-3-fold less than that of the murine mAb. With the aim of improving tumor targeting, attempts have been made to enhance the avidity of the HuCC49 mAb. Previous research has yielded a single gene-encoded immunoglobulin, SCIgcCC49deltaCH1, which is a dimer of a single chain consisting of CC49 single-chain Fv linked to the NH2 terminus of the human gamma1 Fc through the hinge region. This molecule is comparable to the mouse-human chimeric CC49 in terms of in vitro antigen binding properties, cytolytic activity, and rate of plasma clearance in athymic mice bearing human tumor xenografts. Recently, a single gene encoding a single-chain immunoglobulin consisting of a HuCC49 diabody attached to human gamma1 Fc via the hinge region was constructed. The diabody, a bivalent antigen-binding structure, is made up of variable heavy (V(H))/variable light (V(L)) domains and V(L)/V(H) domains. In each of the variable domain pairs, the V(H) and V(L) domains are linked through a short linker peptide. Meanwhile, the two pairs are linked via a 30-residue Gly-Ser linker peptide to yield two antigen-binding sites by lateral and noncovalent association of the V(L) of one pair with the V(H) of the other. Transfectomas expressing the single-gene immunoglobulin secrete a homodimer of about Mr 160,000 that reacts to tumor-associated glycoprotein-72. This tetravalent humanized antitumor immunoglobulin molecule may potentially be an efficacious therapeutic and diagnostic reagent against a wide range of human carcinomas.

A chimeric immune receptor consisting of an extracellular antigen-binding domain derived from the CC49 humanized single-chain antibody, linked to the CD3zeta signaling domain of the T cell receptor, was generated (CC49-zeta). This receptor binds to TAG-72, a mucin antigen expressed by most human adenocarcinomas. CC49-zeta was expressed in CD4+ and CD8+ T cells and induced cytokine production on stimulation. Human T cells expressing CC49-zeta recognized and killed tumor cell lines and primary tumor cells expressing TAG-72. CC49-zeta T cells did not mediate bystander killing of TAG-72-negative cells. In addition, CC49-zeta T cells not only killed FasL-positive tumor cells in vitro and in vivo, but also survived in their presence, and were immunoprotective in intraperitoneal and subcutaneous murine tumor xenograft models with TAG-72-positive human tumor cells. Finally, receptor-positive T cells were still effective in killing TAG-72-positive targets in the presence of physiological levels of soluble TAG-72, and did not induce killing of TAG-72-negative cells under the same conditions. This approach is being currently being utilized in a phase I clinical trial for the treatment of colon cancer.

Monoclonal antibody (MAb) CC49 is a murine IgG1 that reacts with tumor-associated glycoprotein (TAG)-72, a pancarcinoma antigen. Clinical trials using radiolabeled CC49 for diagnostic imaging have demonstrated specific localization of more than 90% of carcinomas. The feasibility of adopting in vivo gene inoculation methods for antibody-based immunotherapy requires introduction and expression of two genes, encoding immunoglobulin (Ig) heavy and light chains, in a single cell to generate a functional antibody. To circumvent the problems inherent in this approach, we have constructed a single-gene encoding a single-chain immunoglobulin (SCIg) that, unlike previously developed SCIgs, contains all IgG domains. To construct the novel SCIg, the carboxyl end of the constant region of the chimeric (c) CC49 kappa chain is joined, via a 30 residue Gly-Ser linker peptide, to the amino terminus of the CC49 heavy chain. To our knowledge, neither a linker peptide this long nor a linkage between the constant light (C(L)) and variable heavy domains has been reported previously. Transfectomas developed by introducing the expression construct of the amplifiable gene in dihydrofolate reductase-deficient Chinese hamster ovary (CHO dhfr-) cells secrete a 160 kDa homodimeric molecule, SCIgcCC49. The in vitro antigen binding properties of SCIgcCC49 are comparable to those of cCC49 and SCIgcCC49deltaC(H)1, a single-chain Ig deficient in constant heavy chain-1 (C(H)1) and C(L) domains. The antibody-dependent cellular cytotoxicity (ADCC) of SCIgcCC49 and cCC49 were also comparable. This single-gene approach for generating an immunoglobulin molecule may facilitate in vivo gene inoculation as well as ex vivo transfection of patients' cultured tumor-infiltrating lymphocytes for immunotherapy protocols for a variety of diseases, including cancer.

While several murine monoclonal antibodies (MAbs) directed against carcinoma associated antigens have shown excellent tumor targeting properties in clinical trials, the use of radiolabeled MAbs for both diagnostic and therapeutic applications has been hindered by two factors: (a) the induction of host anti-immunoglobulin (Ig) responses and (b) slow plasma clearance of unbound radiolabeled MAb, resulting in bone marrow toxicity for therapeutic application, and long intervals between MAb administration and tumor detection for diagnostic applications. This report describes the development of the first recombinant Ig with properties designed to reduce or eliminate both of the above problems: a complementarity determining region (CDR)-grafted humanized (Hu) MAb with a CH2 domain deletion (delta CH2). The MAb chosen for engineering was CC49, which is directed against a pancarcinoma antigen designated TAG-72 that is expressed on the majority of colorectal, gastric, breast, ovarian, prostate, pancreatic and lung carcinomas. When characterized for antigen binding in solid phase competition radioimmunoassays, the HuCC49 delta CH2 MAb completely inhibited the binding of murine (mu) CC49 and HuCC49 for TAG-72. The relative affinity constants (Ka) of MAbs HuCC49 delta CH2, HuCC49 and muCC49 were 5.1 x 10(-9), 2.1 x 10(-9) and 2.3 x 10(-9), respectively. The plasma clearance of 131I-HuCC49 delta CH2 was significantly faster than that of intact 125I-HuCC49 after either i.v. or i.p. administration in athymic mice (p(2)0.05). Biodistribution studies in athymic mice bearing human colon carcinoma xenografts after i.v. or i.p. administration of 131I-HuCC49 delta CH2 and 125I-HuCC49 demonstrated the efficient tumor localization and substantially lower percent of the injected dose (%ID/g) of the HuCC49 delta CH2 in normal tissues. This is reflected in the significantly higher radiolocalization indices (%ID/g in tumor divided by %ID/g in normal tissue) observed with the HuCC49 delta CH2 for most normal tissues tested (p(2)0.05). The differential between the rate of plasma clearance of HuCC49 delta CH2 and HuCC49 was even more pronounced in SCID mice, which have been shown to be an appropriate model to study the metabolism of human IgG. These studies thus describe the development of a recombinant Ig molecule which, for the first time, combines 1) the properties of more rapid blood clearance than an intact humanized Ig molecule--without loss of antigen binding affinity--and 2) reduced potential for eliciting a human anti-murine antibody (HAMA) response in patients. These studies also demonstrate the potential utility of HuCC49 delta CH2 for i.p. as well as i.v. radioimmunodiagnosis and radioimmunotherapy in patients with TAG-72 positive tumors.

The ras p21 proto-oncogenes (ie, K-ras, H-ras, N-ras) encode a family of proteins vital to cellular signaling and function. Point mutations in these genes have been found in a wide diversity of human cancers, suggesting a strong association in the development of the malignant phenotype. Although the precise mechanisms leading to tumorigenesis are not fully understood, it has been proposed that point mutations in the ras p21 proto-oncogenes contribute to the transformation process through constitutive transduction of growth-promoting signals. These oncoproteins are distinct from normal ras p21 in both DNA and protein sequences at specific sites, typically positions 12, 13, 59, or 61. A large frequency of human cancers harbor point mutations in the ras gene at codon 12, where the normal Gly residue is substituted with either a Val, Asp or Cys residue. From an immunologic perspective, these "neo-determinants" may now represent unique and highly specific epitopes for T cell (CD4+ and/or CD8+) recognition in cancer immunotherapy. Evaluation of point-mutated ras as a T-cell epitope could be determined biologically with short synthetic peptides that precisely mimic those altered sites. Several laboratories have established approaches in both murine and human systems to evaluate the point-mutated ras p21 oncogene product as a potential tumor-specific target and characterization of the resulting cellular immune responses. It has been demonstrated that (1) active immunization of mice with the appropriate mutant protein or peptides leads to the production of cytotoxic CD4+ (Th1 subtype) or CD8+ T lymphocytes, which mediate MHC-restricted, antigen-specific lysis of tumor cells in vitro bearing endogenous mutant ras epitopes; and (2) in vitro stimulation of human lymphocytes from some normal individuals or carcinoma patients with mutant ras peptides results in the expansion of CD4+ and CD8+ precursors, which may exhibit cytotoxicity against autologous or MHC-matched, antigen-bearing target cells. Taken collectively, these preclinical findings provide the rationale for the development of potential immunotherapies directed against point-mutated ras oncogene products.

Background: Interleukin-2 (IL-2), a potent inducer of cellular immune responses, has been used for biological therapy of human cancer; however, the high doses of IL-2 required to mediate patients' immune responses can cause considerable systemic toxicity. The murine monoclonal antibody (MAb) CC49, which reacts with tumor-associated glycoprotein (TAG)-72, expressed on a variety of human carcinomas, has shown excellent tumor localization in recent clinical trials., Objectives: Development and characterization of a single-chain immunoglobulin-IL-2 (SCIg-IL-2) fusion protein which, by delivering IL-2 selectively to the tumor site, can serve as an effective reagent for CC49/IL-2 combination therapy., Study Design: A single-gene encoding the SCIg-IL-2 fusion protein derived from the chimeric (c) CC49 was designed, generated and inserted in an expression vector. The monomeric single-chain protein consisted of the CC49 heavy and light chain variable domains covalently jointed through a (GGGGS)3 linker peptide. The carboxyl end of the variable domain of the light chain was linked to the amino terminus of the human gamma 1 Fc through the hinge region, and the carboxyl end of the CH3 domain was linked to the amino terminus of the human IL-2 through a GGGSGGG linker peptide. The SCIg-IL-2, expressed from the murine myeloma cells transfected with the expression construct, was characterized for its antigen-binding specificity, antibody effector functions and IL-2 biological activity., Results and Conclusion: Transfection of murine myeloma cells with the single-gene expression construct SCIg-IL-2 expressed a single-chain protein of approximately 70 kD, which was secreted into tissue culture fluid as a homodimer of approximately 140 kD. SCIg-IL-2 competed completely with cCC49 for binding to the TAG-72 antigen, but approximately three- to four-fold more of the SCIg-IL-2 was required to achieve levels of competition similar to those observed with the murine or chimeric CC49. With human effector cells, the fusion protein mediated lysis of TAG-72-positive human carcinoma cells. Prior treatment of human effector cells with 100 U/ml of human IL-2 enhanced the fusion protein-mediated cytolysis from 32 to 65%. At doses of > or = 1 ng/ml, the stimulatory effect of SCIg-IL-2 on IL-2 dependent murine HT-2 cell proliferation was comparable to that of the recombinant human IL-2. The single-gene construct may also facilitate inoculation of the gene in animal tissue for in vivo expression of the fusion protein.

Monoclonal antibody (MAb) CC49 reacts with tumor-associated glycoprotein (TAG)-72, a human pancarcinoma antigen. In clinical trials, radiolabeled CC49 has shown excellent tumor localization; however, many of the patients receiving MAb CC49 develop a human antimouse antibody response. In an attempt to prevent this antiimmunoglobulin response, we have developed a humanized CC49 (HuCC49) by grafting the MAb CC49 hypervariable regions onto the variable light (VL) and variable heavy (VH) frameworks of the human MAbs LEN and 21/28' CL, respectively, while retaining those murine framework residues that may be required for the integrity of the antigen combining-site structure. The HuCC49 MAb was compared with native murine CC49 (nCC49) and chimeric CC49 (cCC49), using a variety of assays. SDS-PAGE analysis under nonreducing conditions showed that the HuCC49 MAb has virtually identical mobility to that of cCC49. Under reducing conditions, the HuCC49 yielded two bands of approximately 25-28 and approximately 50-55 kDa, characteristic of heavy and light immunoglobulin chains. In competition radioimmunoassays, HuCC49 completely inhibited the binding of 125I-labeled nCC49 to TAG-72, although 23- to 30-fold more HuCC49 was required to achieve a level of competition similar to those of cCC49 and nCC49. The relative affinity of HuCC49 was 2- to 3-fold less than those of the cCC49 and nCC49 MAbs, respectively. The plasma clearance in mice of HuCC49 was virtually identical to that of cCC49. Biodistribution studies demonstrated equivalent tumor-targeting of HuCC49 and cCC49 to human colon carcinoma xenografts. These studies thus suggest that HuCC49 and genetically modified molecules, such as sFv and domain-deleted immunoglobulins developed by using the HuCC49 variable region as a cassette, may be potentially useful in both diagnostic and therapeutic clinical trials in patients with TAG-72-positive tumors.

Human prostate-specific antigen (PSA) has been shown as an aid in the early detection of prostate cancer (W. J. Catalona et al., J. Am. Med. Assoc., 270: 948-954, 1993) and was approved in 1994 by the Food and Drug Administration for early detection of prostate cancer. Immunotherapies directed against PSA have been suggested in patients with metastatic prostate cancer. One of the essential questions is to define which nonhuman species express PSA for experimental studies. Using Southern blot analyses, genes related to human PSA have been detected in several nonhuman primate species, including chimpanzee, orangutan, gorilla, macaque, and rhesus monkey, but not in other mammalian species, including rabbit, cow, pig, dog, rat, or mouse. Immunohistochemical staining with anti-human PSA antisera detected strong staining in both human and monkey prostatic epithelial cells with no reactivity to rat prostate cells. Because the PSA gene is not present in the murine genome, a matched set of murine cell lines has been developed that may be useful to study the biochemical functions of PSA and as an experimental target for PSA-directed immunotherapy. To establish such cell lines, a C57BL/6 murine colon adenocarcinoma cell line, MC-38, was transfected with a retroviral vector containing cDNA encoding the human PSA gene. Genetic analysis of a PSA-secreting clone, PSA/MC-38, demonstrated that the PSA gene had been stably integrated into the MC-38 genome. The PSA/MC-38 cell line was found to secrete PSA into tissue culture medium, producing a protein of approximately M(r) 30,000. In vivo, PSA/MC-38 grew as a s.c. tumor in male and female mice. PSA/MC-38 tumors grew more rapidly in athymic mice than in syngeneic C57BL/6 mice, and in both mouse strains, the PSA/MC-38 tumors grew more slowly than control vector-transduced tumors. PSA was detected in the serum and tumors of PSA/MC-38 tumor-bearing mice. It is proposed that PSA/MC-38 cells may be used as a murine tumor model to test potential therapeutic vaccines and other experimental therapies directed against PSA.

Background: Numerous monoclonal antibodies (MoAb) have been developed and currently are being evaluated in both diagnostic and therapeutic clinical trials. Despite the major advances fostered by MoAb technology, several limitations inherent to the use of MoAb exist. For example, MoAb may not have the desired plasma pharmacokinetics and metabolic properties, and they may be immunogenic, thus reducing the possibility of numerous administrations., Methods: Recombinant DNA technology is used to develop (1) mouse-human chimeric antibodies in which the constant region of a murine antibody is replaced with the human constant region, (2) chimeric antibodies with domain-deletions or alterations in glycosylation, and (3) sFv molecules, i.e., recombinant proteins composed of a VL amino acid sequence of an immunoglobulin tethered to a VH sequence by a designed peptide., Results: This article reviews some of the genetic modifications that can be made with recombinant or chimeric immunoglobulin forms; two anti-TAG-72 MoAb, B72.3 and CC49, are used as examples. The immunoglobulin molecules that have been generated include those with alterations of subclass, domain deletions, and glycosylation, as well as those sFv molecules that have been prepared. The immunochemical and biologic properties of these novel immunoglobulin forms are described., Conclusions: Recombinant DNA technology makes feasible the development of novel immunoglobulin forms. These genetic modifications may result in more useful diagnostic reagents and in the production of more stable immunoconjugates with the characteristics of more efficient tumor cell killing.

Chimeric antibodies have been produced against a pancarcinomic tumor associated antigen, TAG-72, by fusing the genes for the variable region of mouse MAb B72.3 to the genes for the constant region of human IgG. In our efforts to optimize the pharmacokinetics of plasma clearance and the efficiency of tumor localization and penetrance of cB72.3, we have now developed truncated versions of immunoglobulin heavy chains. The domain-deleted antibodies are produced by transfecting cells that produce chimeric kappa chains with expression vectors that encode chimeric heavy chains lacking the sequences that encode the CH2 domain, CH3 domain, or both. Despite the absence of these domains, the transfectomas secrete H2L2 tetramers with appropriate antigenic specificity. All the domain-deleted immunoglobulins can be purified by chromatography on Protein G Sepharose which binds to a site on the Fab region that is retained in the domain-deleted antibodies. The CH2CH3 domain-deleted immunoglobulin produced in cell culture is analogous in size to enzymatically produced F(ab')2.

Monoclonal antibody (MAb) B72.3 reacts with TAG-72, a high-molecular-weight mucin expressed on several types of human carcinoma, and is currently being used in clinical trials for the diagnosis and therapy of human carcinoma. An expression construct containing cDNA encoding an immunoglobulin (Ig) heavy chain, with the variable region of murine MAb B72.3 and a human Ig constant region with a deletion of the CH2 domain, was generated. Immunoglobulin from the transfectoma with the highest expression of the TAG-72 immunoreactive antibody was designated MAb chimeric (c) B72.3 delta CH2. The pharmacokinetics of serum clearance of iodine-labeled MAbs cB72.3 delta CH2 and the intact cB72.3 were compared in athymic mice. By 24 hr, less than 1% of the cB72.3 delta CH2 was left in the plasma, while 36% of the cB72.3 still remained. The T1/2 alpha values of the cB72.3 delta CH2 and cB72.3 MAbs were 1.7 and 2.4 hr, respectively. The T1/2 beta values were 7.8 hr for the domain-deleted cMAb and 48.9 hr for cB72.3. Biodistribution studies in athymic mice bearing LS-174T xenografts showed a reduction in the percentage of injected dose per gram in tumor with 131I-cB72.3 delta CH2; however, the 131I-cB72.3 delta CH2 both localized to tumors faster and cleared from the blood faster than the 125I-cB72.3 MAb. Only trace amounts of the 131I-cB72.3 delta CH2 were detected in normal tissues, including kidney. The faster clearance rate, more rapid tumor targeting and lack of metabolic uptake in normal tissues demonstrated with the iodine-labeled CH2 domain-deleted cMAb may be an advantage for certain clinical protocols.

COL-1 is a murine IgG2a monoclonal antibody (MAb) with a high affinity (1.4 x 10(9) M-1) for carcinoembryonic antigen (CEA) and no detectable reactivity for CEA-related antigens, such as nonspecific cross-reacting antigen (NCA) and normal fecal antigen (NFA). 125I-labeled COL-1 IgG was shown to efficiently and specifically target the LS-174T human colon carcinoma xenograft in athymic mice. Dose titration studies in this same model with 131I-labeled COL-1 demonstrated reduction of tumor growth rate when 300 microCi of the immunoconjugate was used (0.005 > p > 0.001). Administration of higher levels as a single dose led to increased toxicity. Dose fractionation experiments with 131I-COL-1 demonstrated the ability to administer much higher levels of the immunoconjugate with little or no toxicity, which resulted in a greater therapeutic efficacy. For example, three fractions of 200 microCi of 131I-COL-1 given at weekly intervals (for a total of 600 microCi) resulted in the substantial reduction (p < 0.0005) of the growth of established tumors in 100% (7/7) of mice, and in no evidence of tumor growth in 71% (5/7) of mice, at the end of the 63-day observation period. These results thus demonstrate the potential therapeutic efficacy for radiolabeled COL-1 in clinical trials and demonstrate the principle of the advantage of dose fractionation protocols for this immunoconjugate.

The MC-38 C57BL/6 mouse colon adenocarcinoma cell line has been transduced with a retroviral construct containing cDNA encoding the human carcinoembryonic antigen (CEA) gene [Robbins PF, Kantor JA, Salgaller M, Horan Hand P, Fernsten PD, Schlom J (1991) Cancer Res 51: 3657]. Two clones, MC-38-ceal and MC-38-cea2, expressed high levels of CEA on their cell surface. A third CEA-expressing cell line, MCA-102-cea3, was similarly derived by transduction of the MCA-102 C57BL/6 mouse fibrosarcoma cell line and is described here. In this study, the three CEA-transduced murine tumor cell lines (MC-38-ceal, MC-38-cea2, MCA-102-cea3) were evaluated for their tumorigenic potential, as well as their ability to serve as in vivo model systems for active and passive immunotherapy studies. Parameters that were investigated include tumor growth rate, the antibody response of the host to CEA, and the CEA content of the tumors. The MC-38-cea2 model appeared to be the most appropriate for immunotherapy studies. Biodistribution studies, using an 125I-labeled anti-CEA mAb, demonstrated efficient tumor targeting of MC-38-cea2 tumors in C57BL/6 and athymic mice.

It has been demonstrated previously that the degree of glycosylation of a molecule may alter its pharmacokinetic properties and, in the case of an antibody, its metabolism and other biological properties. Transfectomas producing aglycosylated chimeric B72.3(gamma 1) pancarcinoma monoclonal antibody (mAb) were developed by introduction of the eukaryotic expression construct pECMgpB72.3 HuG1-agly, into SP2/0 murine myeloma cells producing the chimeric kappa chain of mAb B72.3. After cell cloning, one subclone with the highest binding to the TAG-72-positive human colon carcinoma was designated mAb aGcB72.3, and its biological and biochemical properties were compared with those of the chimeric B72.3(gamma 1), designated mAb cB72.3. Polyacrylamide gel electrophoresis showed that under non-reducing conditions, the molecular masses of the aGcB72.3 and cB72.3 mAbs were 162 kDa and 166 kDa respectively. The heavy chain of mAb aGcB72.3 had a slightly faster mobility than that of cB72.3, while the mobility of the light chains of the two chimeric mAbs was similar. No difference was observed in the isoelectric points of either chimeric mAb. Liquid competition radioimmunoassays demonstrated that the aGcB72.3 and cB72.3 mAbs have comparable binding properties to TAG-72. These studies demonstrate that aglycosylation of the chimeric IgG1 mAb B72.3 at the CH2 domain, as has been shown for other mAbs [Dorai H., Mueller B., Reisfeld R. A., Gillies S. D. (1991) Hybridoma 10:211; Morrison S. L., Oi V. T. (1989) Adv Immunol 44:65], eliminates antibody-dependent cell-mediated cytotoxicity activity, but does not substantially alter affinity or plasma clearance in mice. These studies also demonstrate for the first time (a) no difference in plasma clearance of an aglycosylated and a chimeric mAb in a primate after i.v. inoculation; (b) a difference (P less than or equal to 0.05) in mice in the more rapid peritoneal clearance of a chimeric mAb versus an aglycosylated chimeric mAb; (c) higher (0.05 less than or equal to P less than or equal to 0.1) tumor: liver ratios at 24, 72 and 168 h using 111In-labeled aglycosylated chimeric mAb versus chimeric mAb. Since the liver is the major site of metastatic spread for most carcinomas, slight differences in tumor to normal liver ratios may be important in diagnostic applications. These studies thus indicate that comparative analyses of a novel recombinant construct (i.e., aglycosylated) and its standard chimeric counterpart require documentation in more than one system and are necessary if one is ultimately to define optimal recombinant/chimeric constructs for diagnosis and therapy in humans.

A cell line derived from the mouse colon adenocarcinoma, MC-38, has been transduced with a retroviral construct containing complementary DNA encoding the human carcinoembryonic antigen (CEA) gene. MC-38, which forms tumors in syngeneic C57BL/6 mice, has been extensively studied as a target for active immunotherapy. Individual transduced clones that express high levels of cell surface CEA were isolated, and two clones, termed MC-38-ceal and MC-38-cea2, were extensively characterized. The levels of CEA found on the surface of these clones were considerably higher than that found in a moderately differentiated human colon carcinoma cell line (WiDr) and were comparable to those found on the human colon carcinoma cell lines GEO and CBS (among the highest CEA-expressing cells reported). Further analysis demonstrated that the CEA expressed in the MC-38-cea1 clone had a similar molecular weight to native CEA (Mr 180,000), but the MC-38-cea2 cell line expressed a single Mr 70,000 glycosylated immunoreactive product. Seven anti-CEA monoclonal antibodies were found to react with both clones. The CEA gene present in the MC-38-cea2 clone was partially sequenced and was found to contain a deletion of two of the three repeated domains present in CEA. These results provide a basis for future studies to map immunodominant epitopes of CEA and to develop a syngeneic model system that may aid in the design of reagents and protocols to study active and passive immunotherapy directed against a carcinoma expressing human CEA.

A monoclonal antibody, RAP-5, raised against the 21,000d ras-oncogene product, p21, was used in an immunoperoxidase staining procedure to study the expression of p21 in normal children's and foetal kidneys, pre-malignant lesions and benign and malignant childhood renal tumours with good, moderate or poor prognosis. ras-p21 was expressed in both normal and foetal kidneys but its distribution in renal tumours differed markedly (p less than 0.01). A quantitative liquid competition radioimmunoassay (RIA) was used to determine ras-p21 level in tissue homogenates. The results were expressed as pg of ras-p21 per microgram protein of tissue extract. There were significant differences in the levels of ras-p21 among various renal tissue extracts (p less than 0.05). Generally it emerged that the amount of ras-p21 was greater in both malignant renal tumours and foetal kidneys compared to normal kidneys, pre-malignant lesions and benign renal tumours.

We report here the generation and characterization of a recombinant/chimeric construct of murine gamma 1 monoclonal antibody (MAb) B72.3, containing the murine variable region and a human gamma 1 constant region [designated cB72.3(gamma i)]. cB72.3(gamma 1) was generated by first isolating functionally rearranged VH and VL genes of B72.3 from partial genomic libraries in phage vectors. Construction of mouse-human chimeric heavy and light chain genes was performed by inserting restriction fragments carrying VL and VH regions of B72.3 into unique sites of expression vectors which contains sequences encoding constant regions of human kappa and gamma 1, respectively. The expression constructs were subsequently electroporated into SP2/0 cells. The transfected SP2/0 murine cell line has been shown to synthesize cB72.3(gamma 1) at a level of 10-20 micrograms/ml. Reciprocal competition radioimmunoassays demonstrated that cB72.3(gamma 1), a previously described cB72.3(gamma 4), and native B72.3 (designated nB72.3) competed similarly. A rat anti-idiotype MAb made against nB72.3 was shown to bind equally well to cB72.3(gamma 1) and to the nB72.3. Immunochemical studies of the nB72.3, cB72.3(gamma 4), and cB72.3(gamma 1) revealed slight differences in size among the three MAb forms on sodium dodecyl sulfate gels and revealed a higher isoelectric point for the cB72.3(gamma 1). Antibody-dependent cell-mediated cytotoxicity experiments using human lymphokine-activated killer effector cells indicated better tumor cell killing by the cB72.3(gamma 1) than the nB72.3 or cB72.3(gamma 4). Dual label studies of coinjected cB72.3(gamma 1) and nB72.3 revealed that both MAbs could efficiently localize human tumor xenografts in athymic mice. Pharmacokinetic studies, analyzing the blood clearance of cB72.3(gamma 1), cB72.3(gamma 4), and nB72.3 in mice, showed that the nB72.3 beta phase of clearance was slower than that of other MAb forms. However, when the pharmacokinetic patterns of these three MAbs forms were analyzed in monkeys, the cB72.3(gamma 1) and the nB72.3 showed similar clearance curves, while the cB72.3(gamma 4) showed a much slower plasma clearance. In view of the binding properties of nB72.3 and its ability to localize a range of carcinomas in clinical trials, the studies reported here demonstrate that the cB72.3(gamma 1) may serve as a potentially useful diagnostic and/or therapeutic reagent.

Tumor targeting by monoclonal antibodies (MAbs) can be enhanced by (a) increasing the percentage of injected dose taken up by the tumor and/or (b) increasing the tumor:nontumor ratios. Several groups have demonstrated that one can increase tumor to nontumor ratios by the use of antibody fragments or the administration of second antibodies. Several other modalities are also possible: (a) the use of recombinant interferons to up-regulate the expression of specific tumor associated antigens such as carcinoembryonic antigen or TAG-72 on the surface of carcinoma cells and thus increase MAb tumor binding has proved successful in both in vitro and in vivo studies; (b) the intracavitary administration of MAbs. Recent studies have demonstrated that when radiolabeled B72.3 is administered i.p. to patients with carcinoma of the peritoneal cavity, it localizes tumor masses with greater efficiency than does concurrent i.v. administered antibody. Studies involving the comparative pharmacology of intracavitary administration of radiolabeled MAb in patients and several animal models will be discussed; (c) it has been reported that prior exposure of hepatoma to external beam radiation will increase radiolabeled MAb tumor targeting. We and others have not been able to duplicate this phenomenon with a human colon cancer xenograft model and radiolabeled MAbs to two different colon carcinoma associated antigens. The possible reasons for these differences will be discussed; (d) the cloning and expression of recombinant MAbs with human constant regions and subsequent size modification constructs will also undoubtedly alter the pharmacology of MAb tumor binding in both diagnostic and therapeutic applications.

The use of Mabs for the detection and treatment of human carcinoma lesions can still be regarded in its infancy. As with other new approaches to cancer therapy, several conceptual as well as real problems exist when designing clinical protocols for Mab-directed immunotherapy. From the Mab standpoint, studies using the intact IgG have shown that, in a majority of patients injected with IgG, human anti-mouse IgG antibodies develop that hamper the effectiveness of subsequent antibody administration. It is believed that the human anti-mouse antibody response is directed against the Fc region of the IgG molecule. The elimination of this region through fractionation of the Mab to obtain the minimum binding site could result in a less immunogenic molecule. Another approach aimed at reducing the immunogenicity of the Mab would be to clone the genes encoding for individual Mabs, reduce them via restriction endonuclease techniques, and insert human immunoglobulin constant regions. The resulting chimeric antibodies are believed to reduce the development of human anti-mouse antibodies. Effective Mab therapy of human tumor lesions may also be achieved through the recruitment of a portion of the host's immunologic defense system. An example is the use of anti-idiotype Mabs that use as immunogen a Mab to a tumor antigen. The anti-idiotype antibodies are selected for binding to the antigen binding, or idiotype, region of the first Mab. The binding sites of the new anti-idiotype Mabs should reflect the 'internal image' of the original antigen. The anti-idiotype antibodies may be used to immunize patients (i.e., vaccines) in an attempt to mount an active immune response against the antigen-positive tumor cells. Recent studies have shown a synergism between interferon-alpha and an anti-idiotype Mab for the in-vivo antitumor activity in a murine B-cell lymphoma experimental model. Whether an interferon-mediated increase in the tumor antigen or the Fc receptor was part of the synergism was not investigated. Mabs alone have also been shown to elicit cytotoxic activity in vitro and tumoricidal activity in vivo. Antibodies of the IgG2a isotype can direct macrophage-mediated cytotoxicity. These studies revealed the importance of the number of antibody sites per cell as well as the number of cells that bind the IgG2a Mab, thus suggesting a 'threshold' requirement for the demonstration of effective tumor cell lysis in vitro and in vivo.(ABSTRACT TRUNCATED AT 400 WORDS)

Splenic lymphocytes of mice immunized with membrane enriched fractions of human mammary carcinomas were fused with the NS-1 nonsecretory++ myeloma cell line. The resulting hybridomas were screened for the synthesis of immunoglobulins reactive with human mammary tumor associated antigens, and two IgG monoclonal antibodies (B1.1 and F5.5) were identified as being reactive with purified carcinoembryonic antigen (CEA). These antibodies were shown to bind to different epitopes on CEA based on their differential reactivities to five different purified CEA preparations, and their differential binding to the surface of tumor cells derived from various organ sites. Monoclonal B1.1 bound equally to the surface of human breast, colon, and melanoma cell lines. Monoclonal F5.5, on the other hand, did not react with the surface of melanoma cell lines, and showed a differential binding to breast carcinoma versus colon carcinoma cells. Monoclonals F5.5 and B1.1 were both used in immunoperoxidase studies with fixed tissue sections of a variety of histologic types of human mammary carcinomas and were shown to be reactive with 55% and 66%, respectively, of tumor masses.

The ras gene family of rodents and humans is highly conserved and consists of several distinct genes, i.e., rodent Harvey and Kirsten, and human Harvey, Kirsten and neuroblastoma. This gene family mediates transformation via (1) a point-mutation resulting in the change of one amino acid in the 21 kDA ras gene product (p21) or (2) increased expression of ras p21. Group-specific, type-selective and interspecies indirect binding liquid competition radioimmunoassays (RIAs), capable of providing truly quantitative analyses of the 21 ras oncogene and proto-oncogene products, have been developed. Using purified recombinant ras p21 from Escherichia coli expressing the full-length T24 mutant human Harvey-ras gene protein product as a standard in these RIAs, we have defined the absolute numbers of pg, fM and molecules of ras p21 in: (1) E. coli expressing the point-mutated or proto-ras p21 and (2) mammalian cell lines of human and murine origin. Two of the RIAs developed can be termed group-specific in that they have the ability to detect the point-mutated and proto forms of all 3 human ras genes (Harvey, Kirsten, and neuroblastoma), while the third RIA is type-selective, since it detects an antigenic determinant located primarily on the Harvey ras p21. All 3 RIAs are interspecies-specific since they are able to detect ras p21 in rodent as well as human cells. The adaptability of the RIAs to various assay conditions and ease of methodology make these immunoassays applicable to the study of several parameters associated with ras p21 expression. These assays, used in conjunction with specific cDNA probes to identify specific ras proto-oncogenes or point-mutated oncogenes being expressed, now provide truly quantitative analysis of ras p21 in mammalian cells to further the study of the association between ras p21 expression and transformation.

A virus, similar to the murine mammary tumor viruses (MuMTV) of the laboratory mouse Mus musculus, was identified in the milk of M. cervicolor popaeus mice. The virus was morphologically indistinguishable from the type-B MuMTV and was thus termed MC-MTV. Radioimmunoassays for the 52,000-dalton major envelope glycoprotein and the 28,000-dalton major internal protein of MuMTV demonstrated that MC-MTV shared some antigenic determinants with both of these MuMTV proteins. This reactivity was clearly different, however, from that observed with all MuMTV tested from M. musculus. MC-MTV had a density of 1.16 g/ml in sucrose and a virion-associated DNA polymerase with a divalent cation preference for Mg2+ over Mn2+. Radioimmunoassays clearly differentiated MC-MTV from the other viruses previously identified from M. cervicolor, i.e., M432, CERV-CI, and CERV-CII. These studies thus identified the first virus from another species that is immunologically related to the MuMTV of M. musculus. Particles similar to MC-MTV were also observed in a spontaneous M. cervicolor popaeus mammary tumor.

Monoclonal antibodies (MAbs) of predefined specificity have been generated by utilizing a synthetic peptide reflecting amino acid positions 10-17 of the Hu-rasT24 gene product as immunogen. These MAbs, designated RAP-1 through RAP-5 (RA, ras; P, peptide), have been shown to react with the ras gene product p21. Since the Hu-ras reactive determinants (positions 10-17) have been predicted to be within the tertiary structure of the p21 molecule, it was not unexpected that denaturation of cell extracts or tissue sections with Formalin or glutaraldehyde enhanced binding of the RAP MAbs. When paraffin-embedded Formalin-fixed tissue sections and the avidin-biotin complex immunoperoxidase method were used, the RAP MAbs clearly defined enhanced ras p21 expression in the majority of human colon and mammary carcinomas. The majority of all abnormal ducts and lobules from fibroadenoma and fibrocystic disease patients were negative, as were all normal mammary and colonic epithelia examined. The findings reported here form the basis for quantitative radioimmunoassays for a ras translational product and provide a means to evaluate ras p21 expression within individual cells of normal tissues and benign, "premalignant," and malignant lesions.

We report here both the range and patterns of reactivity of an IgG1 monoclonal antibody, B72.3, prepared against human, metastatic mammary carcinoma cells. When the avidin-biotin complex (ABC) immunoperoxidase technique was used on tissue sections, monoclonal B72.3 reacted with 19 of 41 (46%) primary mammary carcinomas and 13 of 21 (62%) metastatic lesions, either in axillary lymph nodes or at distal sites. Variable concentrations of antigen, recognized by B72.3, were observed among mammary tumors, as well as among different cell populations of a given tumor mass. Several patterns of antigen distribution were observed: membrane, diffuse cytoplasmic, focal and marginal. No reactivity was observed to normal mammary epithelium, stroma, or lymphocytes of the breast, nor to any cell types in a variety of other normal human tissues, melanomas, and sarcomas. Reactivity with all of four colon carcinomas was also observed. Assay of serial sections of mammary carcinomas with B72.3 and a monoclonal antibody directed against carcinoembryonic antigen demonstrated that these antigens were both distinct and non-coordinately expressed.

It has been hypothesized that epithelial and endothelial cells interact with the laminin component of basement membranes via a cell surface laminin receptor molecule. It has also been proposed that the expression of this molecule may be involved in the invasion of carcinoma cells from their tissue of origin and their subsequent penetration through blood vessel basement membranes. We report here the use of a monoclonal antibody, LR-3, to define the expression of laminin receptor in normal, dysplastic, and carcinomatous human tissues. Monoclonal antibody LR-3 is shown by immunoblotting to recognize the Mr 67,000 laminin receptor protein, to bind to the carcinoma cells, and to constitute approximately 0.1% of total cellular protein. Numerous normal human epithelial and endothelial cell types, as well as pulmonary macrophages, are shown to express laminin receptor to varying degrees. Selected human mammary carcinomas and colon carcinomas are shown to bind more monoclonal antibody LR-3 than normal or dysplastic counterparts. A monoclonal antibody to laminin receptor now makes possible the study of the role of laminin receptor in tumor cell metastases and in the differentiation and function of various normal human epithelial and endothelial cell types.

Using c-Ha-, c-Ki-, and c-N-ras-specific probes in a RNA-RNA hybridization assay we found enhanced expression of c-Ha-ras protooncogene in stomach adenocarcinomas relative to nonneoplastic epithelium, whereas little or no transcription of either c-Ki- or c-N-ras was detected. Enhanced levels of c-Ha-ras RNA expression were detected in all of the adenocarcinomas examined. Hybridization with c-Ha-ras was also detected in nonneoplastic gastric epithelium adjacent to carcinoma, although the labeling was less intense than that of carcinoma cells. More extensive analysis of the c-Ha-ras p21 expression was then carried out in formalin-fixed, paraffin-embedded tissue sections and extracts from surgically resected stomach tissues using monoclonal antibodies (MAbs) RAP-5 and Y13-259. The data obtained from the immunohistochemical studies were consistent with the results of in situ hybridization assay. Adenocarcinomas were much more reactive with MAb RAP-5 than benign and normal tissues, and the majority of carcinomas demonstrated increased expression of c-Ha-ras p21. Quantitative liquid competition radioimmunoassays using MAb Y13-259 also demonstrated significantly higher levels of c-Ha-ras p21 in extracts from stomach adenocarcinomas than those from normal mucosae. No strict correlation was found between ras p21 expression and the degree of tumor differentiation or histological type. Although advanced carcinomas generally demonstrated higher levels of ras p21 than early carcinomas, no correlation among advanced carcinomas and ras p21 levels was observed in relation to depth of tumor invasion to the muscularis propria, subserosa, or serosa. Benign lesions, in comparison, were much less reactive with MAb RAP-5 than carcinomas. Among the benign lesions tested, dysplastic lesions were more reactive than nondysplastic lesions. Normal stomach mucosa was generally nonreactive with the exception of parietal cells. Our results indicate that transformation of the stomach mucosa from benign to malignant phenotype is associated with an increase in c-Ha-ras p21 expression.

Splenic lymphocytes of mice, immunized with membrane-enriched fractions of metastatic human mammary carcinoma tissues, were fused with the NS-1 non-immunoglobulin-secreting murine myeloma cell line. This resulted in the generation of hybridoma cultures secreting immunoglobulins reactive in solid-phase radioimmunoassays with extracts of metastatic mammary carcinoma cells from involved livers, but not with extracts of apparently normal human liver. As a result of further screening of immunoglobulin reactivities and double cloning of cultures, 11 monoclonal antibodies were chosen that demonstrated reactivities with human mammary tumor cells and not with apparently normal human tissues. These monoclonal antibodies could be placed into at least five major groups on the basis of their differential binding to the surface of various live human mammary tumor cells in culture, to extracts of mammary tumor tissues, or to tissue sections of mammary tumor cells studied by the immunoperoxidase technique. Whereas a spectrum of reactivities to mammary tumors was observed with the 11 monoclonal antibodies, no reactivity was observed to apparently normal cells of the following human tissues: breast, lymph node, lung, skin, testis, kidney, thymus, bone marrow, spleen, uterus, thyroid, intestine, liver, bladder, tonsils, stomach, prostate, and salivary gland. Several of the antibodies also demonstrated a "pancarcinoma" reactivity, showing binding to selected non-breast carcinomas. None of the monoclonal antibodies showed binding to purified ferritin or carcinoembryonic antigen. Monoclonal antibodies of all five major groups, however, demonstrated binding to human metastatic mammary carcinoma cells both in axillary lymph nodes and at distal sites.

Recent studies in animal models involving antibody tumor targeting of hepatoma and melanoma and clinical trials involving hepatoma patients have suggested that preirradiation of tumors may enhance antibody tumor targeting. These reports led us to study the effect of external irradiation on monoclonal antibody (MAb) targeting of human carcinomas; as a model system, we used MAb B72.3 and the LS-174T human colon carcinoma xenograft in athymic mice. LS-174T tumors exposed to 300 cGy grew to approximately 93% the size of non-irradiated tumors, while those exposed to 600, 900, or 2,000 cGy were approximately 41% the size of control tumors. Splitting the 900 cGy into three 300-cGy fractions yielded a two-fold lower tumor volume compared with a single 900-cGy fraction. Histochemical evaluation of the carcinomas revealed a decrease in the number of mitoses per high power field consistent with early effects of radiation exposure. Using the avidin-biotin complex immunoperoxidase technique, carcinomas were assayed for expression of the tumor associated glycoprotein (TAG)-72, the high-molecular-weight mucin detected by MAb B72.3. No discernable variation was observed in the staining intensity among tumors in both the control and radiation treated group; that is, differences among tumors within each group were compatible with the known heterogeneous expression of TAG-72. Exposure of carcinomas to 300 or 900 cGy in a single fraction or 900 cGy split in three 300-cGy fractions did not yield a consistent or substantial enhanced localization of radiolabeled MAb B72.3 IgG or F(ab')2 to tumors. A 1.5-fold augmentation of MAb binding to tumors was observed in preirradiated mice; however, these results were not statistically significant. Inherent differences in tumors such as cell type of origin, size, spatial configuration, extent of vascularization and volume of interstitial space may contribute to variability of the effect of preirradiation of tumors on antibody binding. Our results suggest that consistent augmentation of radiolabeled antibody localization to tumors is not a universal phenomenon.

Eight different human breast carcinoma cell lines, as well as clonally derived cell lines, were used to study the expression of four different monoclonal antibody (MAb)-defined tumor-associated antigens (TAAs) and the ability of recombinant human (rHu)-interferon (IFN)-alpha A and rHu-IFN-gamma to increase tumor-associated and/or normal cell-surface antigen expression. The different breast tumor cell lines expressed a wide range of antigenic phenotypes with respect to four different cell-surface TAAs. The cell lines could be divided into high and low antigen-expressing groups based on their constitutive levels of the four TAAs. In general, those breast tumor cell lines that expressed high levels of carcinoembryonic antigen, the MAb DF-3-defined 290-kD antigen, and the 90-kD antigen reactive with MAb B6.2 were poor candidates for antigen augmentation by rHu-IFN-alpha A or rHu-IFN-gamma. In contrast, breast cell lines that constitutively express low levels of these TAAs were found to be highly responsive to the ability of either of the rHu-IFNs to enhance MAb binding to the cell surface. Treatment with either of the rHu-IFNs increased the level of surface binding of MAbs as much as fourfold. The relative abilities of the IFNs to increase MAb binding to the surface of human breast tumor cells thus appeared to depend on the degree of constitutive surface antigen expression of the breast tumor cells. The high-molecular-weight TAA, TAG-72, is known not to be expressed on most cell lines but is expressed on the majority of human carcinoma biopsies. Most breast tumor cell lines employed in this study did not express the TAG-72 high-molecular-weight TAA. However, rHu-IFN-alpha A did substantially increase the level of expression of TAG-72 on the surface of breast tumor cells that were isolated from a pleural effusion. These studies may thus provide an important insight into the criteria used in the selection of carcinoma patients for IFN-mediated antigen augmentation when employing MAb-guided radioimmunolocalization or MAb-guided therapy.

We have adapted T24P, a tumorigenic subline of the T24 human bladder cancer cell line, to grow in 5 mM butyrate. In the presence of butyrate, the adapted cells (T24P/B) grow more slowly than the unadapted cells (T24P/C), have a lower saturation density, increased serum requirement for growth, loss of ability to form colonies when plated at low cell density, and decreased ouabain sensitivity. Morphologically, T24P/B cells in butyrate are large and flattened with increased cytoplasm. When T24P/B cells are grown without butyrate, the morphological changes, growth rate, plating efficiency, and ouabain sensitivity return to those of T24P/C. While the saturation density increases, it does not return to levels of T24P/C, and the size of colonies never reaches that of the T24P/C colonies. Both T24P/C and T24P/B are tumorigenic in nude mice, however, the T24P/B tumors differ grossly and microscopically from those produced by T24P/C in that they contain large cystic structures filled with clear fluid and lined by transitional cell epithelium with flattened surface layers. Although the transformed phenotype and tumorigenicity of T24P are modified by adaptation to growth in butyrate, no significant changes in ras oncogene RNA or protein expression were identified.

Molecular studies have demonstrated increased expression of the Harvey (Ha) ras oncogene in human breast and colon carcinomas. With the use of a direct-binding liquid competition radioimmunoassay (RIA), capable of providing truly quantitative analysis of the 21,000-dalton (p21) ras oncogene and protooncogene products, absolute levels of Ha-ras p21 have been determined in human breast and colon carcinomas, benign lesions, and/or their respective normal tissues. Enhanced Ha-ras expression was documented in 66% of breast and 100% of colon carcinomas as compared with their normal counterparts, with levels in breast carcinomas ranging from 10.1 to 50.4 pg ras p21/micrograms protein and those in colon carcinomas ranging from 18.4 to 51.7 pg ras p21/micrograms protein. Some dysplastic lesions of the breast and colon also contained elevated Ha-ras p21. Relative levels of Ha-ras p21 expression, detected by competition RIA, correlated with percent Ha-ras p21-positive cells as determined by immunohistochemical assays. By use of liquid competition RIA and immunohistochemical assays, it has been shown that levels of ras p21 expression did not always correlate between primary and metastatic colon lesions of the same patient. The use of the quantitative RIA and semiquantitative immunohistochemical assays, in concert with cDNA probes for identification of specific ras point-mutated oncogenes or protooncogenes, may now provide the means for definitive quantitative analyses of ras p21 in human carcinomas and benign lesions.

Monoclonal antibodies RAP-5 and Y13-259, directed against the ras gene product [a protein with a molecular weight of 21,000 (p21)] have been used to evaluate ras p21 expression in malignant and benign mammary tissues as well as in the lesions of intermediate stature such as atypical hyperplasia using immunohistochemical assays. Invasive carcinoma demonstrated enhanced expression of ras p21, with generally decreasing expression in carcinoma in situ, atypical hyperplasia, and nonatypical hyperplasia, respectively. Heterogeneous expression of ras p21 was observed among primary as well as metastatic mammary carcinomas. Carcinomas from postmenopausal patients generally demonstrated higher levels of ras p21 than those from premenopausal patients, but no significant difference in ras p21 expression in carcinomas between estrogen-receptor rich and estrogen-receptor poor patients was found. Normal mammary epithelium in terminal duct lobular units from patients with hyperplasia generally demonstrated higher levels of ras p21 expression than did epithelium in large ducts. This demonstration of enhanced ras p21 expression by the epithelium of peripheral lobular portion of the breast is consistent with the previous hypothesis that these areas preferentially undergo malignant transformation. Analyses of the limited number of specimens available from patients with 15-yr follow-up revealed a generally higher level of ras p21 in hyperplasia from patients who subsequently developed carcinoma, as compared to those from patients without carcinoma development. However, no conclusions regarding the potential for malignant transformation could be drawn for any individual patient on the basis of ras p21 expression. Concomitant analyses of ras p21 expression in mammary carcinomas and benign lesions using liquid competition radioimmunoassay and immunohistochemical assay demonstrated the complementary nature of these alternative approaches. These results suggest that enhanced ras p21 expression may be involved in the early stages of mammary carcinogenesis but is probably not involved in the maintenance of the transformed phenotype.

Using a direct binding liquid competition radioimmunoassay, the amount of the ras oncogene protein product, p21, was quantitated in a variety of human tumors and adjacent apparently normal tissues. In 48 of 50 matched tumor and normal tissue biopsy specimens from 50 patients, more ras p21 was detected in the tumor than in its normal counterpart. Twenty-five of 28 breast tumors demonstrated more ras p21 than the average of the values obtained for fibroadenomas. Furthermore, in 17 of the 19 cases studied, over 20% more ras p21 was observed in breast carcinomas compared with their respective normal counterparts. More ras p21 was also demonstrated in the majority of tumors of the stomach, lung, colon and bladder compared with their respective adjacent normal tissues. Our data therefore indicate that ras p21 expression is quantitatively enhanced in many human tumors originating from several different tissue types.

A new method for the analysis of the binding of monoclonal antibodies to cell surface tumor-associated antigens utilizes 1- to 2-day primary cultures of human colonic carcinomas, adenomas, and normal epithelial tissue. The antibodies are added to the live cells which form monolayer epithelial patches of several hundred cells on the surface of the Petri dish by migration in a continuous sheet from a small explant. These epithelial patches are then fixed with methanol and processed in situ using the indirect immunoperoxidase assay. Three monoclonal antibodies (MAbs) prepared against membrane-enriched fractions of human metastatic breast cancer were assayed. MAb B1.1 bound to each of 11 benign and each of 18 malignant colonic tumors tested. MAb B6.2 displayed similar reactivity, binding to each of 7 adenomas and each of 15 carcinomas assayed. Both MAbs also bound to normal colonic epithelial cells in both the live cell studies presented here and in earlier studies (D. Stramignoni et al., Int. J. Cancer, 31: 543-552, 1983). MAb B72.3 bound only to tumor cells and not to normal epithelial cells in the live cell assay. This epitope was rapidly lost in culture. B72.3 reactivity on each of two carcinomas was decreased 9- to 36-fold when primary culture continued for 5-6 days. B72.3 bound to each of 20 tumors (15 carcinomas, 5 adenomas) when the cells were cultured for 1 or 2 days but on only 2 of 8 tumors when the cells were cultured for 3 to 8 days. The B72.3 epitope was more strongly expressed on the live cells in the explant and on those monolayer cells directly adjacent to the explant than on the cells more towards the edges of the patch colony. This implied that the cell flattening which occurred when cells migrated from the explant may have played some role in antigen loss. A very similar fraction of primary cultured carcinoma and adenoma cells bound each MAb, indicating that these MAbs in live cell assay do not distinguish between benign, noninvasive colonic tumors and invasive carcinomas. The live cell assay was compared to the standard assay utilizing sectioned, fixed tumors. In parallel assays of eight tumors the fraction of cells reactive in the indirect immunoperoxidase assay was consistently higher on live cells for each of these MAbs than on fixed tissue. Due to this greater sensitivity the live cell assay was able to detect reactive cells in two cases which were scored as negative (less than 1% positive cells) in the fixed tissue assay.

Monoclonal antibodies reactive with the surface of human breast carcinoma cells have been generated and characterized. The immunogens used were membrane-enriched fractions of metastatic carcinoma lesions. The various monoclonals were shown to react with previously known as well as with novel tumor-associated antigens (TAAs). The most specific of the latter group is monoclonal B72.3, which is reactive with a 220,000 to 400,000 high-molecular-weight glycoprotein complex found in 50% of human mammary carcinomas and 80% of human colon carcinomas. Monoclonal antibody B6.2, which recognizes a 90,000-d glycoprotein, was radiolabeled and shown to efficiently localize human carcinoma transplants in athymic mice via gamma imaging without the use of second antibody or background subtraction manipulations. F(ab')2 and Fab' fragments were shown to be more efficient for tumor localization than intact immunoglobulin. Whereas the phenomenon of antigenic heterogeneity of tumor cell populations has long been known to exist, this phenomenon was also shown to manifest itself as antigenic modulation, in which specific TAAs can modulate their expression on the cell surface concurrent with different phases of the cell cycle. A phenomenon known as antigen evolution, in which a specific cloned tumor cell population can gradually drift in antigenic phenotype, has also been demonstrated. Recombinant interferon has been employed to (1) enhance the expression of specific TAAs on the surface of tumor cells already expressing the antigen; and (2) induce the expression of specific TAAs on the surface of carcinoma cells not previously expressing the antigen. The clinical implications of such phenomena in gamma scanning for the detection of tumor masses and for tumor immunotherapy are discussed. Methods for circumvention of problems inherent in the clinical use of monoclonal antibodies are also addressed.

Radioimmunoassays were developed that can detect antigenic determinants common to mammary tumor viruses (MTV's) of four distinct Mus species: M. musculus, M. cervicolor, M. cookll, and M. caroll. The radioimmunoassays were based on the immunologic cross-reactivity observed between the murine mammary tumor viruses (MuMTV) of M. musculus and type B retrovirus isolated from M. cervicolor. Both of the glycoproteins of MuMTV (gp52, gp36) shared antigenic determinants with virions of M. cervicolor mammary tumor virus. Interspecies radioimmunoassays for gp52 and gp36 were developed and used to detect viruses in the milk of noninbred feral Mus species and MuMTV-related translational products in mammary tumors in these species. Type C and type D retroviruses, as well as the M432 retrovirus of M. cervicolor, did not react in either assay. Both interspecies immunoassays were therefore specific for the detection of distinct MuMTV-related antigenic determinants.

Murine monoclonal antibodies, prepared against human metastatic mammary tumor cells, were used to demonstrate differential expression of several tumor-associated antigens (TAAs) among various mammary carcinomas and within a given tumor mass. Using the immunoperoxidase technique on serial sections of 39 human primary mammary carcinomas, a spectrum of antigenic phenotypes of TAAs was observed: 13% of the tumors reacted with all of a panel of four monoclonal antibodies; while 10% of the mammary tumors scored negative with all four antibodies. The remaining 30 tumors could be divided into several additional groups based on their differential reactivity with some, but not all, of the monoclonal antibodies. Furthermore, variation among mammary carcinomas was also observed in the cellular localization of antigens. Antigenic phenotypic diversity of mammary tumor cell populations within a given tumor mass was also observed; this was noted with respect to (a) antigenic expression in one area of a tumor mass and not another and (b) a "patchwork" effect in which antigens were expressed on cells immediately adjacent to cells which scored negative. Antigenic phenotypic diversity was also observed in established mammary tumor cell lines grown in vitro. A differential loss of some cell surface TAAs was observed as a function of continued cell passage; consistent with this finding, MCF-7 mammary tumor cell lines obtained from four sources could be differentiated from each other by their pattern of cell surface TAA expression. Single-cell clones derived from the MCF-7 mammary tumor cell line exhibited at least four distinct antigenic phenotypes; a change in cell surface phenotype of some of the clones was seen during subsequent passage. This definition of phenotypic variation and modulation of TAA expression among, and within, human mammary carcinomas has implications towards both the design and the outcome of studies involving the in situ immunodetection and therapy of breast cancer.

With membrane-enriched fractions prepared from human metastatic breast tissue used as immunogen, a group of monoclonal antibodies (MAbs) were generated that recognized several distinct antigens on breast and other carcinomas. The antibodies were found to react with established human tumor cells in culture as well as in immunohistochemical protocols using sections of primary and metastatic lesions. One MAb, B72.3, demonstrated a high degree of selective reactivity for human carcinomas in that no reactivity was found with a large number of different normal tissues. These MAbs were used to demonstrate the heterogeneity of expression of the tumor antigens as well as the intrinsic cellular factors (i.e., cell-cycle kinetics and clonal variability) that modulate their expression. Additional studies showed that recombinant human leukocyte interferon can act as a potent regulator of surface antigen expression on human carcinoma cells. Analysis of these cells by radioimmunoassay or flow cytometry revealed that interferon treatment resulted in a higher percentage of the cell population that bind the MAb to the surface antigen. Furthermore, interferon treatment also increases the level of expression for particular tumor antigen throughout the tumor cell population. Experimental models were also developed to investigate the active localization of a radiolabeled MAb by a human tumor xenograft in athymic mice. The results demonstrate active uptake of an 125I-B6.2 by a transplantable human breast tumor that expressed significant quantities of the B6.2-reactive 90 kD tumor antigen. In contrast, a human melanoma cell line grown as a solid, subcutaneous tumor in athymic mice did not localize the labeled B6.2 and does not express the associated 90 kD antigen. We report the generation and characterization of anti-breast carcinoma MAbs. These immunologic probes were used to study relevant breast tumor antigens, the factors that influence their level of expression, and the ability of human tumors grown in athymic mice to localize radiolabeled antibodies.

Treatment of human breast or colon carcinoma cells with recombinantly derived human leukocyte (clone A) interferon (IFN-alpha A) increases the surface expression of specific tumor-associated antigens (TAAs) recognized by monoclonal antibodies (MAbs). The MAbs used, B1.1, B6.2, and B72.3, recognize three distinct TAAs, i.e., the Mr 180,000 carcinoembryonic antigen, a Mr 90,000, and a Mr 220,000 to 400,000 glycoprotein, respectively. The binding of the MAbs to the surface of tumor cells increased in a dose-dependent manner, with optimal levels of TAA enhancement at 100 to 1,000 units IFN-alpha A/ml. Higher concentrations of IFN-alpha A that were cytostatic or cytotoxic were also less effective in enhancing TAA expression. Human melanoma (A375) cells and normal fibroblasts (WI-38 and Flow 4000) do not express any of the three TAAs, either before or after interferon treatment. The ability of IFN-alpha A to increase the expression of TAAs on human carcinoma cells was also temporally dependent, with optimal enhancement occurring after 16 to 24 hr. The enhancement of specific TAAs at the surface of the carcinoma cells by IFN-alpha A was confirmed, using fluorescence-activated cell sorter analysis. These data demonstrate that the IFN-alpha A-mediated increase of surface antigen is a result of both an accumulation of more antigen per cell, and an increase in the percentage of cells expressing the antigen. The ability of recombinant interferon to enhance specific TAAs on human carcinoma cells may be exploited in designing protocols for the in situ detection and therapy of human carcinoma lesions by MAbs, as well as in further defining the role of specific TAAs in the expression of the transformed phenotype.