Phase 1 trial of the antiangiogenic peptide ATN-161 (Ac-PHSCN-NH(2)), a beta integrin antagonist, in patients with solid tumours.

To evaluate the toxicity, pharmacological and biological properties of ATN-161, a five -amino-acid peptide derived from the synergy region of fibronectin, adult patients with advanced solid tumours were enrolled in eight sequential dose cohorts (0.1-16 mg kg(-1)), receiving ATN-161 administered as a 10-min infusion thrice weekly. Pharmacokinetic sampling of blood and urine over 7 h was performed on Day 1. Twenty-six patients received from 1 to 14 4-week cycles of treatment. The total number of cycles administered to all patients was 86, without dose-limiting toxicities. At dose levels above 0.5 mg kg(-1), mean total clearance and volume of distribution showed dose-independent pharmacokinetics (PKs). At 8.0 and 16.0 mg kg(-1), clearance of ATN-161 was reduced, suggesting saturable PKs. Dose escalation was halted at 16 mg kg(-1) when drug exposure (area under the curve) exceeded that associated with efficacy in animal models. There were no objective responses. Six patients received more than four cycles of treatment (>112 days). Three patients received 10 or more cycles (> or =280 days). ATN-161 was well tolerated at all dose levels. Approximately, 1/3 of the patients in the study manifested prolonged stable disease. These findings suggest that ATN-161 should be investigated further as an antiangiogenic and antimetastatic cancer agent alone or with chemotherapy.

A phase I study of the oral combination of CI-994, a putative histone deacetylase inhibitor, and capecitabine.

BACKGROUND: This study was conducted to determine the toxicity profile, maximum tolerated dose (MTD) and pharmacokinetics of the putative histone deacetylase inhibitor CI-994 in combination with capecitabine. PATIENTS AND METHODS: Fifty-four patients were treated according to three different dosing schemes in which the capecitabine dose was fixed and the CI-994 dose was escalated. Capecitabine was administered in twice daily divided doses, and CI-994 was given as a single daily dose. In schedule A, 26 patients were treated with capecitabine 1650 mg/m2/day and CI-994 for 2 weeks of a 3-week cycle. In schedule B, six patients received capecitabine 1650 mg/m2/day for two 3-week cycles and CI-994 for 5 of 6 weeks. In schedule C, 22 patients were treated with capecitabine 2000 mg/m2/day and CI-994 for 2 of 3 weeks. RESULTS: At the MTD, the principal dose-limiting toxicity was thrombocytopenia. The pharmacokinetics of CI-994 were unaltered by capecitabine, and there was no correlation between body surface area and major pharmacokinetic parameters. Platelet count nadir was best predicted by the observed maximal concentration (C(max)) of CI-994. CONCLUSIONS: The recommended phase II dose is 6 mg/m2 (or 10 mg) of CI-994 in combination with capecitabine 2000 mg/m2/day for 2 weeks of a 3-week cycle.

11p deletions and breakpoints in squamous cell carcinoma: association with altered reactivity with the UM-E7 antibody.

The UM-E7 monoclonal antibody raised against the UM-SCC-I human squamous cell carcinoma (SCC) cell line identifies a cell surface antigen that is strongly expressed in normal tissues. The locus (MICI) controlling the expression of E7 and related cell surface antigens has been mapped to chromosome band 11p13. This band has been identified as a region of cancer-associated aberrations and as the probable locus of a tumor suppressor gene. Although E7 antigen expression is strong in normal keratinocytes, it varies among squamous carcinoma cell lines. Some SCC lines (12/26) exhibit weak expression of the E7 antigen, whereas other SCC cell lines (14/26) and 21 cell lines from other tumor types express the antigen strongly. On the basis of these observations and of mapping data, we postulated that low E7 antigen expression in a subset of SCC cell lines might be associated with chromosomal rearrangement or deletion involving the E7 locus on 11p. Fully evaluable karyotypes were prepared from 19 SCC cell lines, including 11 with weak and eight with strong E7 expression. Eight of the 11 lines with weak E7 expression had 11p abnormalities. Four of these contained 11p deletions, and four others had a breakpoint in 11p. In contrast, none of the cell lines in the group with strong E7 expression had an 11p deletion, although one had a rearrangement with an 11p breakpoint. In the four tumors with visible 11p deletions, the smallest region of overlap corresponded to the 11p13-p14 region. The mean log10 50% endpoint E7 titer in the group with 11p deletions or breakpoints was nearly two orders of magnitude lower than that of the lines with no 11p abnormality (1.95 +/- 0.53) (P less than 0.02). Our results indicate that the UM-E7 antibody identifies tumors with 11p13-p14 deletions and other 11p rearrangements and that the 11p region is a site of nonrandom chromosome rearrangement in a subset of human squamous cancers. The strong association of loss of antigen expression with visible 11p deletion or rearrangement in some tumors suggests that other tumors with this phenotype may contain submicroscopic lesions of 11p13-p14.

Serologic type conversion of a herpes simplex virus type 1 (HSV-1) to an HSV-2 epitope caused by a single amino acid substitution in glycoprotein C.

A monoclonal antibody to herpes simplex virus type 2 glycoprotein C (gC-2) did not recognize wild-type herpes simplex virus type 1 gC (gC-1) but did recognize a mutant gC-1 molecule. This conversion from a type 1 to a type 2 epitope was shown to be due to a single amino acid substitution in gC-1.

Radioimmunodiagnosis of human-derived squamous cell carcinoma.

Squamous cell carcinomas (SCC) of the head and neck are common and aggressive tumors. A murine IgG 2aK monoclonal antibody (A9) reactive with most SCCs was radiolabeled with I-131 and tested for localization and imaging ability in groups of nude mice bearing SCC tumors derived from two patients with head and neck cancer (UM-SCC-2 and UM-SCC-11B). The mice also received I-125 labeled isotype-matched control antibody UPC-10 for double-label counting. Animals were imaged at multiple time points post-injection and sacrificed 7-10 days later for assessment of the tissue distribution of the radiopharmaceuticals. In general, the tumors were visualized satisfactorily on gamma camera images of both the UM-SCC-2 and UM-SCC-11B tumor-bearing animals. While localization of the A9 antibody was mainly due to antibody specificity in the UM-SCC-2 tumor-bearing animals, there was also a large component of non-specific antibody localization in the UM-SCC-11B bearing animals. This may be due to greater central fluid space and or necrosis in the UM-SCC-11B tumors. This study demonstrates the feasibility of imaging human-derived squamous cell carcinomas with radiolabeled monoclonal antibody A9 and holds forth the possibility of therapy of this radiosensitive tumor in a like manner.

Altered expression in squamous carcinoma cells of an orientation restricted epithelial antigen detected by monoclonal antibody A9.

The monoclonal IgG2a antibody A9 was raised to the UM-SCC-1 human squamous cell carcinoma cell line. Hemadsorption assays using monolayer cultures as target cells revealed a restricted range of A9 reactivity with human cell lines. The A9 antibody was reactive with 29 of 34 squamous cell carcinoma cell lines and with 5 epithelial cancer cell lines of non-squamous origin. In contrast, no antibody binding was detected with twelve malignant melanoma, two fibrosarcoma, two malignant lymphoid, two transitional cell carcinoma, or four adenocarcinoma cell lines. Similarly, normal lymphoid cells, RBC, and fibroblasts from multiple donors were negative. The relative expression of the A9 antigen varied greatly among positive squamous cancer lines such that the 50% endpoint titer of A9 ascites fluid ranged from less than 10(1) for low antigen expressor lines to 10(6) for strong antigen expressors. Hemadsorption tests with secondary passage cultures of normal squamous cells failed to detect A9 binding to the cell surface. However, in experiments with intact colonies in primary cultures of normal squamous cells, antibody binding was observed at the periphery of individual colonies. Mechanical detachment of such colonies revealed A9 antigen bound to the plastic surface underneath the cells, but the newly exposed surface of the cells, like the upper surface, was negative. In contrast, when cells of similar cultures were detached by trypsinization prior to testing no antigen remained on the plastic, but approximately 30% of the trypsinized cells were positive. Similar experiments with the five surface-negative squamous cell carcinoma cell lines revealed that cryptic A9 antigen could also be detected underneath mechanically detached cells and on the surface of 100% of trypsinized cells. Trypsinization of melanomas and fibroblasts did not reveal A9 antigen. Immunoperoxidase assays on frozen sections of normal epidermis localized A9 antigen to the basal cells and the basement membrane region. In frozen sections of squamous cancers, individual tumor cells and particularly the growing edge of the tumors were strongly stained by A9, while in basal cell cancers only very slight staining could be detected. Thus, in normal epithelial cells the A9 antigen appears to be expressed only by the stem cell subset. In squamous cancer cells, the expression of the A9 antigen is not only increased but appears to have escaped the orientation restriction characteristic of normal epithelial cells.

Monoclonal antibody (G10) to a common antigen of human squamous cell carcinoma: binding of the antibody to the H type 2 blood group determinant.

The IgM monoclonal antibody G10 was raised against the human squamous cell carcinoma (SCC) cell line UM-SCC-1. In initial screening against cultured cells, G10 bound to 2 SCC lines (UM-SCC-1 and UM-SCC-13) and 1 pancreatic carcinoma line (UM-PAd-1) but not to cultured fibroblasts (WI-38), ovarian carcinoma cells (SK-OV-3), or malignant melanoma cells (SK-MEL-28 and MeWo). In subsequent tests against cultured cell lines, G10 gave positive reactions with 30 of 33 SCC lines but only 4 of 29 non-SCC lines. The non-SCC lines that bound G10 were UM-PAd-1, 2 transitional cell carcinoma lines (T24 and RT4), and 1 melanoma line (SK-MEL-22). When tested against cultures derived from normal skin or mucosa, G10 was reactive with the epitheloid squamous cells but not with the fibroblasts in each culture. The antigen defined by antibody G10 was stable to fixation with Formalin, and its distribution in tissue sections was examined with the use of immunoperoxidase assays. All SCC biopsy specimens examined in this way were reactive with antibody G10. In similar tests against sections of fixed normal tissues, G10 stained the superficial squamous cells of the epidermis and the basal and suprabasal layers of mucosal squamous epithelial cells from the esophagus. All layers of the laryngeal epithelium were positive. Endothelial cells and certain glandular cells were also positive for G10 binding. G10 agglutinated human red blood cells of all blood groups except those from individuals of the Bombay group (Oh) who lack the H blood group determinant. Against defined oligosaccharides, G10 bound strongly only to the monofucosyl H type 2 structure and was slightly cross-reactive with the synthetic difucosyl H type 2 or Y structure. These results are consistent with previous reports of blood group antigen tissue distribution and indicate that the H type 2 determinant is expressed by all or nearly all mucosal squamous cancers. Less frequent expression by cells of other tumor types may correlate with tissue-specific activation of the H gene-specified fucosyltransferase.

Wilms' tumor-aniridia association: segregation of affected chromosome in somatic cell hybrids, identification of cell surface antigen associated with deleted area, and regional mapping of c-Ha-ras-1 oncogene, insulin gene, and beta-globin gene.

Fusion of an auxotrophic mutant hamster cell with the skin fibroblasts of a child with the Wilms' tumor-aniridia association produced clones which, on the one hand, contained the child's normal chromosome 11 and, on the other, the chromosome 11 with the 11p13 deletion associated with the syndrome. Both hybrids were positive for human LDH-A by enzymatic assay. Clones containing the normal human chromosome 11 were killed by a cytotoxic monoclonal antibody to a cell surface antigen previously mapped to the 11p13----11pter region of chromosome 11. Clones with the abnormal 11 were not killed. Thus, we have produced hybrids from the same patient distinct from each other on the basis of their chromosome 11. These hybrids have been used to map the locus for a cell surface antigen to the deleted region on chromosome 11 of a patient with the Wilms tumor-aniridia association. The linkage between this antigen and the syndrome should be helpful in further study of the genetics of this disease. In addition, we have found that the c-Ha-ras-1 oncogene is distal to the p13 region of chromosome 11 and the position of insulin and beta-globin on the chromosome. Finally, by producing segregants of the hybrids containing the abnormal chromosome 11, we have provided evidence that chromosome 11-associated c-Ha-ras-1 is syntenic with chromosome 11 and not moved to a different portion of the genome.

Antibodies to human squamous cell carcinoma.

We are studying membrane antigens of human squamous cell cancer with the use of naturally occurring autologous antibodies from patients' sera, along with a set of other serologic reagents and monoclonal antibodies raised against cultured squamous cell lines. Twenty-eight squamous cell carcinoma cell lines have been established in our laboratory from tissues obtained from 23 patients. Antibody reactivity has been found against the autologous tumor cell line in 13 of 23 patients. One of these is of sufficient titer for detailed analysis. Four cell lines are available from this patient. UM-SCC-17A is derived from the primary laryngeal carcinoma, and UM-SCC-17B is derived from a lymph node metastasis removed during the same surgical procedure. Fibroblasts have been cultured from normal mucosa, and a B-lymphoblastoid line has been developed by Epstein-Barr virus transformation of the patient's peripheral blood lymphocytes. Antibody from this patient reacts with the UM-SCC-17A and -17B tumor cell lines but does not react with the normal fibroblasts (UM-NF-17).

Further studies on a hybrid cell-surface antigen associated with human chromosome 11 using a monoclonal antibody.

A monoclonal antibody has been obtained that recognizes an antigen encoded by human chromosome 11. We present evidence that this monoclonal antibody recognizes the same or a similar antigenic activity as that previously called a1. Genetic information necessary for a1 expression and recognition by the monoclonal antibody both map to 11p13 leads to 11pter. Mutants that have lost a1 are no longer recognized by the monoclonal antibody. The macroglycolipid fraction of human erythrocyte membranes which contains the a1 antigenic activity is able to convert antigen-negative Chinese hamster ovary cells into cells which are killed by the monoclonal antibody plus complement.

Genetically controlled variation of "acid" beta-galactosidase detected in Rattus norvegicus by isoelectric focusing.

Two genetically variant forms of rat "acid" beta-galactosidase were found to differ in isoelectric point and pH dependence, but not in thermostability or sensitivity to inhibition by p-mercuribenzoate (PMB). The results of two backcrosses and an intercross indicated that the isoelectric focusing phenotypes are controlled by two codominant alleles at a single autosomal locus, for which we propose the name Glb-1. No significant linkage between Glb-1 and albino (LG I), brown (LG II), or hooded (LG VI) was observed. Strain-specific differences in total levels of kidney beta-galactosidase were detected, but it is not yet known whether the variation is controlled by genes linked to Glb-1. Experiments in which organ homogenates were incubated with neuraminidase indicated that the genetically variant forms do not result from differences in sialylation, though sialylation does appear to be largely responsible for the presence of multiple bands within each phenotype and for differences in the banding patterns of beta-galactosidases derived from different organs. The beta-galactosidase present in the bands used for Glb-1 typing resembles human GM1 gangliosidase (GLB1) with respect to pH optimum, substrate specificity, and susceptibility to inhibition by PMB. It also appears that Glb-1 is homologous with the Bgl-e locus of the mouse. In rats as in mice the genetically variant bands of beta-galactosidase are active at acid pH and have relatively high isoelectric points. In both species these bands are readily detectable in kidney homogenates, and can be revealed in homogenates of liver or spleen following treatment with neuraminidase. The presence of the same beta-galactosidase bands in homogenates of rat kidney and small intestine as well as in neuraminidase-treated homogenates of liver and spleen suggests that the Glb-1 variants differ by one or more point mutations in the structural gene for "acid" beta-galactosidase.