A Monoclonal Antibody Reactive with a 40-kDa Molecule on Fetal Thymocytes and Tumor Cells Blocks Proliferation and Stimulates Aggregation and Apoptosis¹

AKR mice are genetically predisposed to developing tumors of the thymus and develop thymic lymphomas at high frequency after 5 mo of age (1, 2). Lymphomagenesis occurs because of the production of endogenous retroviruses (2) and the intrathymic injection of recombinant (thymotropic) murine leukemia virus accelerates the appearance of tumors (3). The AKR murine leukemia viruses do not contain acute transforming oncogenes (4) and are thought to transform cells as a result of activating cellular proto-oncogenes in the vicinity of viral integration into chromosomal DNA (5).

E710.2.3 is a cloned murine CD4⁻CD8⁻ thymic T lymphoma cell line. It was isolated from a thymic tumor from an AKR/J mouse that was explanted and cultured in PMA-containing medium. When cultured by itself at low density, E710.2.3 does not proliferate unless it is stimulated with PMA. However, under these same conditions it is stimulated to proliferate by contact with thymocytes or splenocytes without PMA present. In contrast, E710.2.3 proliferates spontaneously when cultured at high density in the absence of PMA or other cells. When E710.2.3 is injected into syngeneic mice, it grows as a malignant tumor in lymphoid organs and the thymus (6). These results suggested that E710.2.3 expressed a cell surface molecule(s) that regulated its growth.

In an attempt to identify novel functional molecules that may be involved in the growth or survival of lymphomas and/or in normal thymocyte function, we generated hybridomas from hamsters injected with E710.2.3 and screened for Abs that inhibited the proliferation of E710.2.3. One mAb (DMF10.62.3) completely inhibited thymocyte-induced, PMA-induced, and spontaneous proliferation of E710.2.3. This study characterizes the expression, structure, and function of the molecule that is recognized by the mAb DMF62.3.

The following cell lines were used and are described in more detail in Tables I and II. RMA-S (7), WEHI-231 (8), B16 (9), MC57 (10), WOP-3027 (11), 293T (12), EL-4 (13), P815 (14), P388D1 (15), 721 (16), E36 (17), CHO (18), Jurkat (19), 143Btk⁻ (20), COS (21), BHK21 (22), HeLa (23), NFC105 (a thymic lymphoma obtained from Paul O’Donnel, Memorial Sloan-Kettering, New York, NY), G58.2 (a thymic lymphoma obtained from Paul O’Donnel, Memorial Sloan Kettering) and PBK101A2 (24) were maintained in RPMI supplemented with 10% FCS, 4 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg fungizone/ml (complete RPMI). E710.2.3 is a murine thymoma that has been previously described (6). This cell line was maintained in complete RPMI supplemented with 25 ng/ml PMA. LB27.4 (25) and A20 (26) were maintained in RPMI, supplemented as previously described (27). A3.1 (28), C2.3 (28), DO11.10 (29), 13G7.3.2 (30), RF33.70 (31), and DC2.4 (32) were maintained in DMEM supplemented with 10% FCS, 4 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg fungizone/ml. CTLL (33) and HT-2 (34) were maintained in RPMI supplemented as described above and also supplemented with rat Con A supernatant. LADp31 (L cells transfected with the p31 isoform of the invariant chain) were grown in medium as previously described (35). All cells were obtained from the laboratory of origin or the American Type Culture Collection (Manassas, VA) or were developed in our laboratory.

Anti-Fas was purchased from PharMingen (San Diego, CA). FITC-goat anti-hamster was purchased from Accurate Scientific and Chemical (Westbury, NY). Purified hamster IgG was purchased from Cappel (Durham, NC). Cytochalasin B, colchicine, trifluoroperazine, sodium azide, 2-deoxyglucose, and EDTA were obtained from Sigma (St. Louis, MO). Y-3 Ab (36) was produced from the hybridoma HB176, which was obtained from American Type Culture Collection. Phosphatidylinositol-specific phospholipase C (PI-PLC)³ was purchased from Oxford Glycosystems (Wakefield, MA).

Cells were stained with primary Abs (DMF10.62.3 or hamster IgG) for 30 min and with secondary Abs (FITC-goat anti-hamster) for 30 min at 4°C. Specific conditions for each experiment are described separately for each experiment. In some experiments dead cells were excluded by incubating cells with propidium iodide (PI) and then electronically gating PI-positive cells from the analysis. To determine whether adult CD4⁻CD8⁻ thymocytes express the molecule recognized by DMF10.62.3, a two-color staining was performed. DMF10.62.3 was detected using FITC as described above. CD4- and CD8-positive cells were detected using Cy-Chrome-conjugated anti-CD4 and CD8 Abs or PE-conjugated anti-CD4 and CD8 Abs (purchased from PharMingen). CD4- and CD8-negative cells were than analyzed for the expression of the molecule recognized by DMF62.3. Cells were analyzed by flow cytometry using a Becton Dickinson FACSCalibur (Mountain View, CA).

Timed pregnancies were set up using C57BL/10 mice (maintained at University of Massachusetts Medical Center animal facility), and embryos were sacrificed on fetal day 14. The fetal thymi were harvested in PBS using an Eppendorf tube glass plunger. Single-cell suspensions were incubated 20 min on ice with an anti-Fc γ receptor II/III (PharMingen) for 20 min to block Fc receptors. Cells were subsequently stained with either DMF10.62.3 or hamster IgG for 30 min followed by FITC-conjugated goat anti-hamster as described above along with allophycoyanin-conjugated anti-Thy 1.2 (PharMingen). In some experiments anti-CD25 conjugated to PE and anti-CD44 conjugated to Cy-Chrome (PharMingen) were also included. The stained cells were fixed overnight in 1% paraformaldehyde and subsequently analyzed by flow cytometry.

Armenian hamsters were injected i.p. with 10 million E710.2.3 and boosted 7–10 times before fusion. Fusions were performed using the fusion partner P3X63-AG8.653 as previously described (37). Supernatants from hybrids were screened first for binding ability to bind to E710.2.3A by FACS and then for their ability to inhibit proliferation of E710.2.3. Positive clones were subcloned and weaned into serum-free media. DMF10.62.3 is an IgG, and Ab was purified from serum-free supernatants using protein A or protein G immunoaffinity columns.

Cells (10⁵) were incubated with various concentrations of DMF10.62.3 or hamster IgG or without Ab in 200 μl of complete RPMI. To test the effect of inhibitors, 10⁵ cells were preincubated with inhibitor for 30 min, and then DMF10.62.3 mAb (10 μg/ml) was added in the continued presence of inhibitor for 6 h. To test the effect of paraformaldehyde on aggregation, cells were fixed in 1% paraformaldehyde for 10 min and washed, and then DMF10.62.3 was added for 6 h. Aggregation was scored visually as previously described (38, 39). Photomicrographs were taken at 6 h using a thermoelectrically cooled charged-coupled device camera (Princeton Instruments, Trenton, NJ).

Apoptosis was assayed using kits from R&D Systems (Minneapolis, MN) or PharMingen. Briefly, 2 × 10⁵ cells were incubated with various concentrations of Ab in 200 μl of medium. At the end of the incubation cells were washed twice in PBS, treated with PI and FITC annexin for 15 min, and then analyzed by flow cytometry. DNA fragmentation was assessed by agarose gel electrophoresis on 2% agarose gels as previously described (40).

E10.2.3 were washed free of PMA and cultured in complete RPMI for 48 h at low cell density (<10⁵/ml), to reduce background proliferation. Subsequently, 5 × 10³ cells were cultured for 72 h in flat-bottom microtiter plates with 25 ng/ml PMA or 5 × 10⁵ thymocytes in the presence or the absence of Abs. In experiments examining the effects of Abs on the spontaneous proliferation of cells, E710.2.3 (grown at high density, >10⁵/ml) or RMA-S cells were cultured for 36 h in the presence or the absence of different concentrations of Ab. [³H]thymidine (1 μCi/well) was added for the last 5 h, and the incorporation of label into DNA was measured in a beta scintillation counter (Wallac, Gaithersburg, MD).

Spleen, thymus, and bone marrow cells were prepared from adult (4–6 mo old) BALB/c or C57BL/6 mice. RBCs were removed from spleen cell suspensions using Tris ammonium chloride lysis. Unstimulated cells were stained immediately. Lymphoblasts were stimulated in culture with 1 μg/ml of Con A or 10 μg/ml of LPS. After 1–3 days of culture, cells were stained for the expression of DMF10.62.3 as described above.

E710.2.3 cells (5 × 10⁶) were starved for 1 h in methionine-free medium and then incubated for 2 h with [³⁵S]methionine at 0.5 mCi/ml. Labeled cells were lysed in immunoprecipitation buffer as previously described (41). Clarified lysates were precleared with hamster IgG, immunoprecipitated with DMF10.62.3 bound to protein A-Sepharose, and analyzed by SDS-PAGE on 14% gels.

The mAb DMF10.62.3 was obtained by immunizing hamsters with E710.2.3. When analyzed by immunofluorescence and flow cytometry, this Ab stained the surface of E710.2.3 brightly (Fig. 1,A). We then examined whether the molecule detected by DMF10.62.3 was expressed on other cell lines. By immunofluorescence analysis, DMF10.62.3 reacted with a number of murine cell lines (Table I), but was absent from others (Table II). Positive cell lines included some T cell lines (e.g., RMA-S), several B cell lymphomas (e.g., A20 and WEHI-231), and a macrophage cell line (C2.3). DMF10.62.3 also reacted with several immortalized cells of nonhemopoietic origin, including a stromal cell line (PBK101A2), a melanoma (B16), a sarcoma (MC57), and a polyoma-transformed fibroblast (WOP-3027). Several other immature (e.g., G58.2) and mature T cells (e.g., EL4), a macrophage (e.g., A3.1), a dendritic cell (DC2.4), and a fibroblast L cell line (LADp31) were negative for DMF10.62.3. Interestingly, the mAb also reacted with several human cell lines, including Jurkat, 293T, and 143Btk⁻ and also with a monkey SV40-transformed kidney cell line, Cos7. Some other human cell lines such as the B lymphoblastoid cell, 721, and the cervical carcinoma cell, HeLa, did not bind DMF10.62.3. Staining patterns of representative cell lines are shown in Fig. 1 for the DMF10.62.3-positive cells A20 (Fig. 1,B) and Jurkat (Fig. 1,C) and the DMF10.62.3-negative cell, RF33.70 (Fig. 1 D). These data indicate that expression of the molecule is broadly expressed on many, but not all, immortalized cell lines, and its expression is not species or cell lineage restricted.

Expression of the molecule recognized by DMF10. 62.3 was first examined on fetal thymocytes. When day 14 fetal thymocytes were stained with DMF10.62.3, the molecule was present on Thy 1.2-positive cells (Fig. 2,A). Interestingly, the molecule was present on both CD25⁺CD44⁺ fetal thymocytes as well as CD44⁺CD25⁻ fetal thymocytes (data not shown). However, staining of adult thymus (Fig. 2,C), adult spleen (Fig. 2,B), or adult bone marrow cells (Fig. 2 D) showed that the molecule recognized by DMF10.62.3 is not present on any of these cells at levels above those seen with control hamster IgG. Furthermore, the molecule could not be detected on adult CD4⁻CD8⁻ thymocytes after gating on CD4⁻CD8⁻ cells in a multiparameter analysis by flow cytometry or analysis of this population from RAG^−/− mice. Day 14 fetal liver cells were also nonreactive with DMF10.62.3 (data not shown).

To determine whether the molecule recognized by DMF10.62.3 was present on normal activated cells, splenic T cells were activated with the T cell mitogen Con A and stained for expression of the molecule recognized by DMF10.62.3. No significant staining above background was seen in unstimulated cells (Fig. 3,A) or at 24 h (Fig. 3,B), 48 h (Fig. 3,C), or 72 h (Fig. 3,D) after activation. In contrast, Con A treatment resulted in a significant increase in the expression of CD25 on these cells compared with that on unstimulated cells (Fig. 3). Similarly, when splenic B cells were activated with LPS, no staining with DMF10.62.3 was seen at 24 h (Fig. 3,E), 48 h (Fig. 3,F), and 72 h (Fig. 3,G), whereas these cells did express CD25, which is known to be up-regulated on activated B cells (42). The molecule recognized by DMF10.62.3 is not present on adult bone marrow cells (Fig. 2 D). These data indicate that the molecule recognized by DMF10.62.3 is present on some fetal thymocytes, but not on normal quiescent or activated cells of hemopoietic origin in adult animals.

When grown at low density and maintained in the absence of PMA, E170.2.3 cells proliferate slowly or not at all. However, they are stimulated to proliferate when cocultured with thymocytes (6) (Table III). DMF10.62.3 was initially identified by its ability to block this thymocyte-induced proliferation. As shown in Table III, DMF10.62.3 completely inhibits this response. We next investigated whether DMF10.62.3 would inhibit the response of E710.2.3 to other stimuli. As shown in Table III, the Ab also blocked PMA-induced proliferation of E710.2.3. Moreover, E710.2.3 spontaneously proliferated when grown at high density, and DMF10.62.3 inhibited this response (Fig. 4,A). Proliferation was significantly inhibited at 3 μg/ml, and complete inhibition was observed at 12.5 μg/ml. In contrast, hamster IgG had no effect on the response of E710.2.3 to any of these stimuli (Fig. 4,A and data not shown). Similarly, many of the mAbs from the original fusion bound E710.2.3, but did not inhibit its proliferation (e.g., DMF10.132; Table III). Therefore, DMF10.62.3 specifically inhibits the proliferation of E710.2.3 regardless of the stimulus used to induce proliferation.

The molecule recognized by DMF10.62.3 is present on a number of other cell lines. Therefore, it was of interest to determine whether the Ab had a similar effect on their spontaneous proliferation. DMF10.62.3 inhibited the proliferation of RMA-S (Fig. 4,B) as well as a number of other cell lines tested (data not shown). In contrast, the Ab had no effect on the spontaneous proliferation of RF33.70, which was negative for the DMF10.62.3 molecule (Fig. 4 C).

On visual observation of cultures of cells treated with Ab DMF10.62.3 the number of intact cells was noted to decrease. In addition, the cells no longer excluded the vital dye trypan blue (data not shown). This observation as well as the inhibition of proliferation suggest that the Ab was cytotoxic to the cells. Therefore, studies were performed to determine the mechanism by which DMF10.62.3 was inducing cell death. Cells can die by apoptosis or necrosis. One of the early changes seen in cells undergoing apoptosis is the externalization of phosphatidylserine on the plasma membrane, and this can be detected by staining with FITC-annexin. Early in the process, the apoptotic cells can exclude vital dyes, such as PI, and therefore can be identified as FITC-annexin positive and PI negative. Later in the apoptotic process membrane integrity is lost, and the FITC-annexin-positive cells become PI positive. In contrast, during necrosis cells lose membrane integrity and become simultaneously PI positive and FITC-annexin positive, without a FITC annexin-positive and PI-negative stage. As shown in Fig. 5, a percentage of E710.2.3 cells undergoes spontaneous apoptosis in culture (10.9–15% annexin⁺, PI⁻; Fig. 5, D, H, and I). However, as little as 1 μg/ml of DMF10.62.3 caused a significant increase in apoptosis in 1 h (28.9% annexin⁺, PI⁻; Fig. 5,A), and this increased over time (48.6% annexin⁺, PI⁺ by 3 h; Fig. 5,C). Higher concentrations of DMF10.62.3 (15 μg/ml) stimulated apoptosis more quickly in time (37.1% annexin⁺, PI⁺ by 1 h) and in more cells (Fig. 5, E–G). In contrast, treatment with similar amounts of hamster IgG had no significant effect above that of medium alone (Fig. 5, D, H, and I). Apoptosis was also verified by visualizing DNA fragmentation by agarose gel electrophoresis (data not shown).

Because the molecule recognized by DMF10.62.3 is expressed on other cells, and this Ab inhibits their proliferation (where tested), we next examined whether it also stimulated them to undergo apoptosis. DMF10.62.3 caused significant apoptosis of the murine cells lines RMA-S, CTLL, LB27.4, and A20 and the human cell lines Jurkat and 143Btk⁻ (Table IV). Apoptosis was induced using 15 μg/ml of DMF10.62.3 and increased with higher concentrations of Ab. In contrast, DMF10.62.3 did not cause apoptosis in RF33.70, which is negative for the molecule. The level of apoptosis induced by DMF10.62.3 varied among different cell lines and appeared to be dependent on the level of surface expression as well the percentage of cells within the population expressing the molecule (Table IV). For example, most E710.2.3 and RMA cells express the molecule at high levels, and DMF10.62.3 induces high levels of apoptosis in both of these cell lines. In contrast, few A20 and LB27.4 cells express the molecule and at lower levels, and DMF10.62.3 induces lower levels of apoptosis in these cells (Table IV). The stimulation of apoptosis by DMF10.62.3 appears to be independent of Fas, as E710.2.3 and RMA-S cells do not express Fas (Table IV).

DMF10.62.3 was also noted to induce homotypic aggregation of E710.2.3 in culture. At 6 h significant aggregation was observed with cells treated with 5 μg/ml or more of Ab. In contrast, no aggregation was observed in cultures treated with hamster IgG or medium (Fig. 6). This aggregation was blocked by treatment with various agents, including cytochalasin B, which disrupts actin microfilaments; trifluoperazine, which inhibits calmodulin-dependent processes; sodium azide and 2-deoxyglucose, which inhibit ATP synthesis; and EDTA, which chelates Ca²⁺ and Mg²⁺. In contrast, aggregation was not affected by colchicine, which inhibits microtubule formation. The aggregation was also inhibited by incubation at 4°C and by treatment with paraformaldehyde (Table V). These results indicate that the aggregation is an active process and is not simply agglutination. DMF10.62.3 also caused homotypic aggregation of some of the other cell lines (e.g., RMA-S and CTLL) that express the DMF10.62.3 molecule. However, little aggregation above the background was seen for some other DMF10.62.3-positive cell lines (e.g., Jurkat, LB27.4, A20, and 143Btk⁻; however, some of these latter cells spontaneously form clusters in culture). No aggregation was seen with RF33.70, which does not bind DMF10.62.3.

To characterize the molecule bound by DMF10.62.3, a set of biochemical studies was performed. To determine its molecular mass, E170.2.3 cells were labeled for 2 h with [³⁵S]methionine. Immunoprecipitates from labeled cells were analyzed by SDS-PAGE under reducing conditions. The mAb DMF10.62.3 immunoprecipitated an ∼40-kDa molecule from E710.2.3 under reducing conditions.(Fig. 7, lane 2). The electrophoretic mobility of this protein was not altered under nonreducing conditions (data not shown). This band was not seen in immunoprecipitates with normal hamster IgG (Fig. 7, lane 1) or in immunoprecipitates with an anti-MHC class I Ab, Y-3 (Fig. 7, lane 3). A 40-kDa was also identified in lysates of surface-labeled E710.2.3 and RMA-S cells (data not shown).

Several cell surface molecules, such as Thy-1 and Ly-6 A/E, are linked to the cell surface via GPI anchors (43). This surface linkage is sensitive to treatment with PI-PLC (44). To determine whether the molecule recognized by DMF10.62.3 was GPI linked, RMA-S cells, which express the molecule on the cell surface, were treated with PI-PLC. PI-PLC treatment did not reduce DMF10.62.3 staining, but did decrease staining for the GPI-linked molecule Thy-1 (data not shown), suggesting that the molecule recognized by DMF10.62.3 is not GPI linked.

The goal of this study was to identify molecules that influenced the growth of E710.2.3 cells. We isolated an Ab, DMF10.62.3, that blocks the proliferation and induces homotypic aggregation and apoptosis of E710.2.3. It appears to react with a novel cell surface protein.

E710.2.3 was chosen for this analysis because in vitro it is stimulated to proliferate upon contact with normal cells such as thymocytes and lymphocytes (6). This finding suggested the presence of a cell surface receptor that stimulated this cell line to proliferate upon interaction with a ligand that is relatively broadly expressed. Is DMF10.62.3 reactive with this putative stimulatory receptor? Although it blocks the proliferation of E710.2.3 stimulated by thymocytes, it also inhibits the response of these cells to PMA and the spontaneous proliferation that occurs when they are cultured at high density. Moreover, DMF10.62.3 blocks the proliferation of many other cell lines that are not responsive to the thymic and lymphoid stimulator cells. Therefore, it seems most likely that DMF10.62.3 acts against some other molecule whose cross-linking inhibits the proliferation of E710.2.3.

Interestingly, DMF10.62.3 stimulates E710.2.3 to undergo apoptosis, and this might account at least in part for its ability to inhibit cell proliferation. A similar effect is observed with other cell lines that react with DMF10.62.3, although the extent of apoptosis induced varies, and this appears to correlate with the level of expression of DMF10.62.3. In the past several years a number of cell surface molecules have been identified that upon binding to their ligands or Abs induce death by apoptosis. A number of these belong to the TNF receptor family. The molecule detected by DMF10.62.3 appears to be different in m.w. and/or pattern of expression from all the previously described members of this family, including Fas (45, 46), TNF receptors (47), CD27 (48), OX-40 (49), CD30 (50), CD40 (51, 52), 41-BB (53, 54), DR3 (55, 56, 57, 58), DR4 (59), DR5 (60), DcR1 (61), LARD (62), ATAR (63), HVEM (64), CAR1 (65), and TAC-1 (66). Therefore, we conclude that DMF10.62.3 is not reactive with a known member of the TNF receptor family. Apoptosis is also induced in some cell lines upon withdrawal of growth factors such as cytokines. However, the m.w. of the molecule detected by DMF10.62.3 is different from those of known cytokine receptors expressed on thymocytes, and DMF10.62.3 induces apoptosis in tumor cell lines that are not cytokine dependent. In fact, we are unaware of any previously described molecule that has the same m.w. and pattern of expression as the DMF10.62.3 ligand. It therefore appears that DMF10.62.3 identifies a novel death-inducing molecule.

DMF10.62.3 stimulates homotypic aggregation of E710.2.3 cells and some other cell lines. This process requires metabolic activity and is therefore not simply agglutination. Presumably, cross-linking of the molecule detected by DMF10.62.3 stimulates several effects in cells, including up-regulation of adhesion molecules and/or their ligands.

E710.2.3 has a phenotype that most closely corresponds to very immature thymocytes (CD4⁻CD8⁻). Therefore, it was possible that the DMF10.62.3 mAb might detect a molecule that is involved in early T cell development. In fact, the molecule detected by the DMF10.62.3 mAb is expressed on day 14 fetal thymocytes. This is of interest because it is clear that cells at this stage receive signals from the thymic microenvironment that stimulate them to proliferate and differentiate or die, but the receptors and ligands involved in this process are poorly understood. However, by staining and in apoptosis and aggregation assays, DMF10.62.3 does not appear to react with any adult thymocytes, including very immature thymocytes, which are thought to be similar to day 14 fetal cells. Therefore, it seems unlikely that this molecule functions in normal postnatal thymocyte development; however, we cannot exclude the possibility that it is expressed and functions on a very small subset of thymocytes in adult animals or on cells that are rapidly eliminated upon its expression. It remains possible that it plays a role on fetal thymocytes.

Other normal cells, such as bone marrow and splenocytes, do not react with DMF10.62.3. The molecule detected by DMF10.62.3 is expressed on many immortalized cells. It is not, however, simply a common activation Ag, because it is not present on mitogen-stimulated T or B cells. Because it is expressed on fetal cells and tumors, but not other normal cells, as far as has been examined, this molecule may be an oncofetal Ag and might play a role in cell immortilization, or since it is expressed on cultured cell lines may be the result of long term in vitro culture.

The isolation of the DMF10.62.3 Ab provides a tool to further characterize the structure and function of this interesting new molecule. The expression of this molecule on immortalized, but not normal, cells and its ability to induce apoptosis raise the possibility that it could be a target for Ab-based immunotherapy of tumors.

We thank Bob Schreiber for advice on the generation of hamster mAbs.