Expression of EpCAM is up-regulated during regeneration of renal epithelia.

The human epithelial cell adhesion molecule (hEpCAM) is involved in epithelial morphogenesis and repair of epithelial tissues. We hypothesized that changes in hEpCAM expression in vivo correlate with regeneration of renal epithelia after ischaemia/reperfusion injury (IRi). Unilateral IRi was performed on kidneys of hEpCAM transgenic mice. Changes in hEpCAM expression were investigated by quantitative RT-PCR in renal cortex and medulla dissected by laser dissection microscopy and expression patterns of hEpCAM in regenerating kidneys were assessed by immunohistochemistry. The mechanism of hEpCAM promoter activation was investigated in vitro, by real-time bioluminescent imaging in HK-2 cells and in primary tubular epithelial cells (PTECs) subjected to hypoxia and reoxygenation. In vivo, the transcription of the human epcam gene significantly increased in the renal cortex during tubular re-epithelialization (p < 0.01). Moreover, the number of tubuli that expressed hEpCAM protein more than doubled in the renal cortex during regeneration. De novo expression of hEpCAM was detected in the S1 segments of proximal tubuli. Under hypoxic conditions in vitro, activity of the hEpCAM promoter was up-regulated two-fold in the HK-2 proximal epithelial cell line. Moreover, both in primary proximal epithelial cells and in HK-2 cells, hEpCAM protein expression was increased after hypoxia and reoxygenation. The significant up-regulation of hEpCAM during post-ischaemic renal regeneration in vivo and during in vitro hypoxia indicates that hEpCAM expression is associated with renal regeneration.

The histone deacetylase inhibitor valproic acid potently augments gemtuzumab ozogamicin-induced apoptosis in acute myeloid leukemic cells.

Gemtuzumab ozogamicin (GO) is a calicheamicin-conjugated antibody directed against CD33, an antigen highly expressed on acute myeloid leukemic (AML) cells. CD33-specific binding triggers internalization of GO and subsequent hydrolytic release of calicheamicin. Calicheamicin then translocates to the nucleus, intercalates in the DNA structure and subsequently induces double-strand DNA breaks. GO is part of clinical practice for AML, but is frequently associated with severe side effects. Therefore, combination of GO with other therapeutics is warranted to reduce toxicity, while maximizing therapeutic selectivity. We hypothesized that the histone deacetylase inhibitor valproic acid (VPA) sensitizes AML cells to GO. VPA-induced histone hyperacetylation opens the chromatin structure, whereby the DNA intercalation of calicheamicin should be augmented. We found that clinically relevant concentrations of VPA potently augmented the tumoricidal activity of GO towards AML cell lines and primary AML blasts. Moreover, VPA treatment indeed augmented the DNA intercalation of calicheamicin and enhanced DNA degradation. Importantly, synergy was restricted to CD33-positive AML cells and did not require caspase activation. In conclusion, the synergistic proapoptotic activity of cotreatment of AML cells with VPA and GO indicates the potential value of this strategy for AML.

Signaling through CD5 activates a pathway involving phosphatidylinositol 3-kinase, Vav, and Rac1 in human mature T lymphocytes.

CD5 acts as a coreceptor on T lymphocytes and plays an important role in T-cell signaling and T-cell-B-cell interactions. Costimulation of T lymphocytes with anti-CD5 antibodies results in an increase of the intracellular Ca2+ levels, and subsequently in the activation of Ca2+/calmodulin-dependent (CaM) kinase type IV. In the present study, we have characterized the initial signaling pathway induced by anti-CD5 costimulation. The activation of phosphatidylinositol (PI) 3-kinase through tyrosine phosphorylation of its p85 subunit is a proximal event in the CD5-signaling pathway and leads to the activation of the lipid kinase activity of the p110 subunit. The PI 3-kinase inhibitors wortmannin and LY294002 inhibit the CD5-induced response as assessed in interleukin-2 (IL-2) secretion experiments. The expression of an inactivated Rac1 mutant (Rac1.N17) in T lymphocytes transfected with an IL-2 promoter-driven reporter construct also abrogates the response to CD5 costimulation, while the expression of a constitutively active Rac1 mutant (Rac1-V12) completely replaces the CD5 costimulatory signal. The Rac1-specific guanine nucleotide exchange factor Vav is heavily phosphorylated on tyrosine residues upon CD5 costimulation, which is a prerequisite for its activation. A role for Vav in the CD5-induced signaling pathway is further supported by the findings that the expression of a dominant negative Vav mutant (Vav-C) completely abolishes the response to CD5 costimulation while the expression of a constitutively active Vav mutant [Vav(delta1-65)] makes the CD5 costimulation signal superfluous. Wortmannin is unable to block the Vav(delta1-65)- or Rac1.V12-induced signals, indicating that both Vav and Rac1 function downstream from PI 3-kinase. Vav and Rac1 both act upstream from the CD5-induced activation of CaM kinase IV, since KN-62, an inhibitor of CaM kinases, and a dominant negative CaM kinase IV mutant block the Vav(delta1-65)-and Rac1.V12-mediated signals. We propose a model for the CD5-induced signaling pathway in which the PI 3-kinase lipid products, together with tyrosine phosphorylation, activate Vav, resulting in the activation of Rac1 by the Vav-mediated exchange of GDP for GTP.

Epidermal growth factor-like repeats mediate lateral and reciprocal interactions of Ep-CAM molecules in homophilic adhesions.

Ep-CAM is a new type of cell adhesion molecule (CAM) which does not structurally resemble the members of the four major families (cadherins, integrins, selectins, and CAMs of the immunoglobulin superfamily) and mediates Ca(2+)-independent, homophilic adhesions. The extracellular domain of Ep-CAM consists of a cysteine-rich region, containing two type II epidermal growth factor (EGF)-like repeats, followed by a cysteine-poor region. We generated mutated Ep-CAM forms with various deletions in the extracellular domain. These deletion mutants, together with monoclonal antibodies recognizing different epitopes in the extracellular domain, were used to investigate the role of the EGF-like repeats in the formation of intercellular contacts mediated by Ep-CAM molecules. We established that both EGF-like repeats are required for the formation of Ep-CAM-mediated homophilic adhesions, including the accumulation of Ep-CAM molecules at the cell-cell boundaries, and the anchorage of the Ep-CAM adhesion complex to F-actin via alpha-actinin. Deletion of either EGF-like repeat was sufficient to inhibit the adhesion properties of the molecule. The first EGF-like repeat of Ep-CAM is required for reciprocal interactions between Ep-CAM molecules on adjacent cells, as was demonstrated with blocking antibodies. The second EGF-like repeat was mainly required for lateral interactions between Ep-CAM molecules. Lateral interactions between Ep-CAM molecules result in the formation of tetramers, which might be the first and necessary step in the formation of Ep-CAM-mediated intercellular contacts.

Chemokine scavenging by the human cytomegalovirus chemokine decoy receptor US28 does not inhibit monocyte adherence to activated endothelium.

The human cytomegalovirus has found smart ways to exploit the chemokine network in order to subvert immune attack. Chemokines trigger the arrest and firm adhesion of inflammatory cells to the vascular wall. Scavenging of chemokines by viral decoy receptors, such as US28, might prevent arrest of leukocytes to the vascular wall and impair an antiviral immune response. We determined the effect of chemokine scavenging by endothelium-expressed signaling mute US28 (US28R129A) on static monocyte adhesion. Despite the chemokine scavenging capacity of US28R129A, expression of this construct by endothelial cells was insufficient to disrupt leukocyte adhesion to cytokine-activated monolayers. Our results suggest that the concentrations of chemokines that trigger firm leukocyte adhesion are too high to be efficiently scavenged by viral chemokine decoy receptors like US28. From the results of this experimental model a role for US28 in viral immune evasion by chemokine scavenging would appear therefore unlikely.

Bispecific antibody-mediated target cell-specific costimulation of resting T cells via CD5 and CD28.

Induction of T-cell activation requires multiple signals provided by cell surface receptor interactions and/or cytokines. T-cell stimulation via the T-cell receptor/CD3 complex provides an important initial activation event which, when combined with the proper costimulatory signals, results in an activated effector T cell. In this report, we have investigated the effectiveness of epithelial glycoprotein-2- (EGP-2) positive tumor target cells to induce specific T-cell stimulation via CD3, CD5, and CD28 using various combinations of bispecific monoclonal antibodies (BsMab) directed against CD3, CD5, or CD28 on the one hand and the pancarcinoma-associated antigen EGP-2 on the other. Induction of T-cell activation was investigated by assessment of CD69 expression, induction of proliferation, and acquirement of cytolytic potential. EGP-2-specific induction of T-cell activation was observed using combinations of BsMab which simultaneous ligated CD3/CD5, CD3/CD28, or CD3/CD5/CD28 with EGP-2. Activation with CD3-, CD5-, or CD28-based BsMab alone did not result in significant induction of T-cell activation in the presence or absence of EGP-2-positive target cells. Simultaneous ligation via CD5/CD28 resulted in partial T-cell activation, including CD69 up-regulation and increased cytolytic activity. Stimulation via CD3 and CD5 or CD28 could be further increased by the addition of exogenously added recombinant Interleukin 2. In contrast, T-cell activation by simultaneous ligation of CD3/CD5/CD28 could not be further augmented by addition of exogenous interleukin 2, indicating that T-cell activation via the combination of CD3, CD5, and CD28 results in complete T-cell activation. Our results show that rapid and target cell-specific induction of T cells is possible using combinations of BsMab directed against different costimulatory molecules. Simultaneous costimulation via CD3/CD5/CD28 results in the most complete activation of T cells.

LAMA tumor in the rat as an experimental model for pre-B-cell leukemia.

A late pre-B-cell leukemia model in the rat, the LAMA tumor, is described. A mouse monoclonal antibody (HIS30) was developed against LAMA cells. HIS30 reacts with a membrane antigen in tumor tissue, whereas its reactivity with normal tissues is limited to the zona glomerulosa of the adrenal cortex and to the adrenal medulla. HIS30 was used for both the immunohistological detection of tumor cells in tissue sections and the immunolocalization of tumor cells in vivo. To enable in vitro studies with the LAMA model, an in vitro growing cell line (LAMA-K1) was established from the LAMA tumor. LAMA-K1 is immunophenotypically similar to the original tumor. Two tumor transplantation models were characterized. In the first model LAMA was implanted s.c., and local tumor growth occurred at the injection site, which was then followed by lymphatogenic and subsequently hematogenic tumor spread. In the second model i.v. transplantation caused direct hematogenic tumor dissemination. In both models early dissemination was especially prominent to the bone marrow, spleen, and liver. Later in the disease most visceral organs became involved, and partial paralysis of the animal was observed in the end stage of the disease. In combination with HIS30, the LAMA pre-B-cell tumor offers a model for both the investigation of in vivo transplanted tumor cells and for the in vivo detection of tumor cells by HIS30 in LAMA tumor-bearing rats.

Neuroendocrine differentiation antigen on human lung carcinoma and Kulchitski cells.

In the normal lung, a subset of cells with a histological appearance consistent with that of Kulchitski cells are the only lung cells reacting with a monoclonal antibody (MOC-1) raised against a human small cell lung carcinoma-derived cell line. Outside the lung, a subset of normal endocrine cells (in the adrenal, thyroid, ovary, and pancreas) as well as neural cells (brain and peripheral Schwann cells) also express the antigen detected by MOC-1 (named MOC-1-related antigen). Some of these positively reacting cells are ectodermally derived, whereas others are of proven endodermal origin, indicating that the MOC-1-related antigen is not a cell lineage-specific antigen. Instead, the common expression of the antigen by cells with a neural, endocrine, or neuroendocrine function suggests that the antigen related to a neuroendocrine differentiation state of these cells. The presence of the MOC-1-related antigen on several non-lung tumors mostly paralleled its normal tissue distribution, indicating that the antigen is generally retained upon malignant transformation. In lung carcinoma, the antigen proves to be present on almost all small cell carcinomas tested. In addition, adenocarcinoma and mixed adenosquamous carcinoma could also express the antigen, whereas pure squamous cell carcinoma generally did not. This finding will be discussed in relation to a proposed "common stem cell" histogenesis of lung carcinoma.

Resistance mechanisms in three human small cell lung cancer cell lines established from one patient during clinical follow-up.

Mechanisms for resistance were studied in three classic type, human small cell lung cancer cell lines, GLC14, GLC16, and GLC19, that were established from one patient during clinical follow-up. Clinically the tumor changed from sensitive (GLC14) to completely resistant to (chemo)therapy (GLC19) during this period. The stain with JSB-1 antibody, detecting the Mr 170,000 multidrug resistance associated glycoprotein, was most pronounced in GLC16 and absent in GLC19. Intracellular Adriamycin (Adr) concentrations were decreased in GLC16 and GLC19 versus GLC14. Glutathione levels were 12.9, 15.5, and 16.6 micrograms/mg protein; total sulfhydryl groups were 36.5, 45.7, and 48.8 micrograms/mg protein; and glutathione S-transferase activity was 13, 29, and 43 nmol I-chloro-2,4-dinitrobenzene/min/mg protein for GLC14, GLC16, and GLC19, respectively. Incubation with DL-buthionine-S,R-sulfoximine increased Adr and cisplatin induced cytotoxicity, whereas X-ray induced cytotoxicity remained the same. Catalase activity increased from 0.88 to 1.73 to 3.83 mumol H2O2/min/mg protein in, respectively, GLC14, GLC16, and GLC19. Compared to GLC14 and GLC16, Adr induced a higher amount of DNA strand breaks in GLC19. In none of the three cell lines could Adr induced DNA strand breaks be repaired. X-ray induced a comparable amount of DNA strand breaks in all three cell lines but all cell lines were capable of repairing the X-ray induced DNA strand breaks within 90 min. It is concluded that a number of different mechanisms are operative and that some but not all of the observed changes in mechanisms for drug resistance in these lines correlate with the clinical data.

Characterization of three small cell lung cancer cell lines established from one patient during longitudinal follow-up.

Three classic-type, small cell lung cancer cell lines (GLC-14, GLC-16, and GLC-19) have been established from one patient during longitudinal follow-up. During this period the tumor changed from sensitive to completely resistant to (chemo)therapy. A phenotypical and functional characterization of the different cell lines is given in combination with the matching clinical data. (a) The cell lines have been compared with the biopsies from which they were derived. There was a good match between the morphological, biochemical, and immunohistological findings in the cell lines as compared to those obtained in the biopsies. When the biopsy and cell line (GLC-14) obtained before the start of therapy were compared to the biopsies and cell lines (GLC-16 and GLC-19) acquired after the first and second reinduction therapy, respectively, no major changes could be observed. The only clear alteration was the loss of a neuroendocrine antigen (defined by monoclonal antibody MOC-51) in the posttherapy specimens. (b) The doxorubicin, melphalan, and etoposide sensitivity in vitro reflected the clinically observed development of resistance to treatment. The cell line (GLC-14) established before the start of therapy was more sensitive than the lines (GLC-16 and GLC-19) obtained after treatment. It is concluded that the cell lines described in this paper represent a well-characterized in vitro model in which the development of drug resistance in small cell lung cancer can be studied.

A homozygous deletion in a small cell lung cancer cell line involving a 3p21 region with a marked instability in yeast artificial chromosomes.

All types of lung carcinoma are characterized by a high frequency of loss of sequences from the short arm of chromosome 3, the smallest region of overlap containing D3F15S2 in band p21. Here we characterize a 440-kilobase segment from this region, which we found homozygously deleted in one of our small cell lung cancer-derived cell lines. The homozygous deletion maps between UBE1L and ZnF16, just centromeric to D3F15S2. Yeast artificial chromosomes with inserts originating from the deleted region are very unstable and readily lose parts of their insert.

Characterization of three new variant type cell lines derived from small cell carcinoma of the lung.

Three new, well growing cell lines (GLC-1, GLC-2, and GLC-3) have been established from small cell lung carcinoma (SCLC) and characterized. A subclone (GLC-1-M13) markedly different from its parent line GLC-1 was also isolated and characterized. Cytogenetic analysis of the cell lines revealed deletions in the short arm of chromosome 3 as a most consistent chromosomal aberration. The deleted region was not identical in all metaphases, 3p(21-23) being the shortest region of overlap. Despite their SCLC origin GLC-1, GLC-2, and GLC-3 do not show pronounced SCLC differentiation features. Neurosecretory granula were very rare (GLC-1) or completely absent (GLC-2 and GLC-3), whereas the SCLC-related enzyme and hormone markers L-3,4-dihydroxyphenylalanine decarboxylase, neuron-specific enolase, creatine kinase BB, and bombesin-like immunoreactivity were variably expressed. Although the subclone GLC-1-M13 was derived from the poorly differentiated GLC-1, it behaved according to the above criteria as a differentiated "classic" SCLC cell line. When assessed with specific monoclonal antibodies the different cell lines appeared to express different subsets of intermediate filament proteins, indicative for different stages and directions of differentiation: "undifferentiated" (GLC-1 and GLC-2); "neural tissue related" (GLC-2); "simple epithelium" related (GLC-1-M13); and a combination of simple and squamous epithelium related (GLC-3). We conclude that GLC-1, GLC-2, and GLC-3 represent dedifferentiated forms of SCLC, related to the recently described "variant" type of SCLC, whereas the clonal derivate GLC-1-M13 behaves like a differentiated "classic" SCLC cell line.

The epithelial glycoprotein 2 (EGP-2) promoter-driven epithelial-specific expression of EGP-2 in transgenic mice: a new model to study carcinoma-directed immunotherapy.

The human pancarcinoma-associated epithelial glycoprotein-2 (EGP-2), a M(r) 38,000 transmembrane antigen also known as 17-1A or Ep-CAM, is commonly used for targeted immunotherapy of carcinomas because it is strongly expressed by most carcinomas. EGP-2 is, however, also expressed in most normal epithelia. To evaluate anti-EGP-2-directed treatment-associated effects on tumors and on EGP-2-positive normal tissue, we generated EGP-2-expressing transgenic mice. A 55-kb DNA fragment consisting of the 14-kb genomic coding sequence of the human EGP-2 gene with approximately 10-kb-upstream and approximately 31-kb-downstream sequences was isolated and used to direct EGP-2 expression in an epithelium-specific manner. In the EGP-2 transgenic mice, EGP-2 appeared to be specifically expressed in all of those epithelial tissues that also express EGP-2 in humans, whereas all of the other tissues were negative. The specific in vivo localization of the i.v. administered anti-EGP-2 monoclonal antibody MOC31 was studied in EGP-2-positive and -negative tumors induced s.c. in this EGP-2 transgenic mouse model. Immunohistochemical analysis showed specific localization of MOC31 in the EGP-2-positive tumors but not in the EGP-2-negative tumors. No anti-EGP-2 monoclonal antibody localization was observed in any of the EGP-2-positive normal mouse tissues, which indicated a limited in vivo accessibility. In conclusion, an EGP-2 transgenic mouse model has been generated that expresses the EGP-2 antigen as in humans and, therefore, can serve as a model to evaluate the efficacy and safety of a variety of anti-EGP-2-based immunotherapeutic modalities in both tumors and normal tissue.

Structural heterogeneity of the cytoplasmic and outer membranes of Escherichia coli.

The cytoplasmic and outer membranes of gram-negative bacteria can be isolated from spheroplasts, and separated on sucrose density gradients. Lysis of spheroplasts causes extensive membrane fragmentation and since the characteristics of the fragments obtained by different lysis procedures need not be identical, the influence of the disruption method on membrane composition has been examined. Spheroplasts of Escherichia coli J5 were lysed by osmotic shock, which did not significantly separate the cytoplasmic and outer membranes, but resulted in mixed membrane vesicles. Lysis in the French press and by sonication caused extensive membrane fragmentation and separation. Sonication, however, also caused some fusion between fragments of the outer and the cytoplasmic membranes; this intermembrane fusion increased with sonication time. When the cytoplasmic and outer membranes were well separated and intermembrane fusion was minimal or absent, the cytoplasmic and outer membrane fragments were heterogeneous with respect to density and ovarll phospholipid, protein and lipopolysaccharide composition. In addition, cytoplasmic, but not outer, membrane fragments were also heterogeneous with respect to protein composition. It is concluded, therefore, that membrane fragments obtained from the cytoplasmic and outer membranes are heterogeneous independently of the lysis procedures used to obtain these fragments. Possible reasons for this heterogeneity are discussed.

How does lysozyme penetrate through the bacterial outer membrane?

Lysozyme fails to penetrate through the outer membrane of stationary phase cells of Escherichia coli when it is simply added to suspensions of plasmolyzed cells. Lysozyme penetrates the outer membrane only when these cells are exposed to a mild osmotic shock in the presence of EDTA and lysozyme. In the presence of Mg2+, the outer membrane is stabilized sufficiently so that there is no lysozyme penetration during osmotic shock. If Mg2+ is added after an osmotic shock has been used to cause lysozyme to penetrate a destabilized outer membrane, the outer membrane is stabilized once again. In this case however, cells are converted to spheroplasts by the lysozyme which has gained access to the murein layer prior to the addition of Mg2+. Mg2+ stabilizes the outer membranes of these spheroplasts sufficiently so that they remain immune to lysis even in the absence of osmotic stabilizers such as sucrose. These results are discussed in terms of current information on the structure of the murein layer and the outer membrane.

Nature of the regions involved in the insertion of newly synthesized protein into the outer membrane of Escherichia coli.

Outer membrane proteins are synthesized by cytoplasmic membrane-bound polysomes, and inserted at insertion sites which cover about 10% of the total outer membrane when cells grow with a generation time of 1 h. A membrane fraction enriched in outer membrane insertion regions was isolated and partly characterized. The rat at which newly inserted proteins are transferred from such insertion regions into the rest of the outer membrane was found to be very fast; the new protein content of insertion regions and that of the remaining outer membrane equilibrate completely within about 20 s at 25 degrees C. Given the rather rigid structure of the outer membrane and the multiple interactions between outer membrane components and the murein layer, lateral diffusion of newly inserted proteins from insertion sites to the remaining outer membrane is not likely to explain this rapid equilibration. Instead, the data support a model in which insertion regions move along the cell surface, leaving behind stationary, newly inserted outer membrane proteins.

Insertion of newly synthesized proteins into the outer membrane of Escherichia coli.

The insertion of newly synthesized proteins into the outer membrane of Escherichia coli has been examined. The results show that there is no precurser pool of outer membrane proteins in the cytoplasmic membrane because first, the incorporation of a [35S]methionine pulse into outer membrane proteins completely parallels its incorporation into cytoplasmic membrane proteins, and second, under optimal isolation conditions, no outer membrane proteins are found in the cytoplasmic membrane, even when the membranes are analysed after being labeled for only 15 s. The [35S]methionine present in the outer membrane after a pulse of 15 s was found in protein fragments of varying sizes rather than in specific outer membrane proteins. This label could however be chased into specific proteins within 30--120 s, depending on the size of the protein, indicating that although unfinished protein fragments were present in the outer membrane, they were completed by subsequent chain elongation. Thus, outer membrane proteins are inserted into the outer membrane while still attached to ribosomes. Since ribosomes which are linked to the cell envelope by nascent polypeptide chains are stationary, the mRNA which is being translated by these ribosomes moves along the inner cell surface.

Release of outer membrane fragments from normally growing Escherichia coli.

A complex containing lipopolysaccharides, phospholipids and proteine separated from the medium by gelfiltration on Sephadex G-200 or by centrifugation. Electron microscopy revealed that this material is released as vesicles and membrane fragements. To determine the origin of these fragments, they were compared to outer and cytoplasmic membranes with respect to keto-deoxyoctulosonic acid, phospholipid, and protein content, phospholipid composition, fatty acid composition, protein distribution on sodium dodecyl sulfate-polyacrylamide gels, buoyant density, and content of several membrane marker enzymes. The results of this comparison indicate that the membrane fragments found in the culture supernatant of normally growing Escherichia coli consist of practically unmodified outer membrane. Possible mechanisms as to the cause of the release of outer membrane fragments, and its relationship to cell-division, are discussed.

Clinical characterization of non-small-cell lung cancer tumors showing neuroendocrine differentiation features.

In most cases of small-cell lung carcinomas (SCLC) phenotypic features compatible with a neuroendocrine differentiation status can be identified by monoclonal (MOC) antibody-based immunohistological procedures. Similar features can be recognized only in a minority of non-SCLC tumors. During a period of 30 months, all diagnostic non-SCLC biopsies (141 cases) were prospectively analysed for the presence of markers indicative for neuroendocrine differentiation. In 31% of all cases, such a presence could be noticed. Neuroendocrine differentiation (50% to 100% positive-staining tumor cells) was recognized more frequently in adenocarcinoma when compared to large-cell and squamous-cell carcinoma (chi 2 = 9.31, 2 degrees of freedom, P less than 0.01). To investigate whether the clinical behavior of these "neuroendocrine" non-SCLC cases mimics SCLC, a multivariate analysis for prognostic factors was performed. Among other prognostic factors, biopsies containing more than 50% positive-staining tumor cells with the MOC antibody-1 (MOC-1) were recognized as negative prognostic factors.

Purging of epithelial tumor cells from peripheral blood stem cells by means of the bispecific antibody BIS-1.

Peripheral blood stem cell (PBSC) support in breast cancer patients allows high-dose chemotherapy, but tumor cell contamination of the PBSCs is a potential source of relapse. Specific carcinoma cell killing can be obtained by retargeting activated T cells with bispecific antibody BIS-1, directed against epithelial glycoprotein-2 and CD3. To purge epithelial tumor cells from the PBSCs of breast cancer patients, activation of T cells in PBSCs and T-cell retargeting by BIS-1 was studied. PBSCs, obtained by leukapheresis after chemotherapy and recombinant human granulocyte colony-stimulating factor, were cultured in the presence of PBS, interleukin-2, OKT3, or interleukin-2/OKT3 for induction of T-cell activation. Subsequently, lysis of epithelial tumor cell lines by activated T cells of PBSCs in the presence or absence of BIS-1 was assessed with the 51Cr-release assay or immunocytochemical staining. The effect on PBSC hematopoietic colony formation (HCF) was evaluated by the granulocyte macrophage colony-stimulating units assay. Prior to activation, PBSCs from breast cancer patients contained higher levels of CD8+ T cells than peripheral blood from healthy volunteers (P < 0.05). The potential of PBSCs to sustain tumor cell lysis was increased after all prior activations and was further enhanced by BIS-1. Maximal BIS-1 effect was observed after OKT3 activation of PBSCs for 72 h (P < 0.0005), inducing a >3 log depletion of tumor cells. HCF was not affected by prior OKT3 activation and/or BIS-1. In conclusion, specific tumor cell lysis by PBSCs can be obtained in vitro by OKT3 activation and BIS-1 retargeting of T cells, without affecting HCF. At present, studies are evaluating this format for future clinical application.