INTRODUCTION

CD20 is a cell surface protein with a molecular mass of 35 kDa expressed in mature B lymphocytes (1-4). Monoclonal antibodies raised against CD20 affect the growth of B lymphocytes. Thus, most of the antibodies inhibit cell proliferation, whereas some are stimulatory. These observations led to the consideration that CD20 is involved in the regulation of cell growth in lymphocytes. The primary structure of CD20 has been determined by molecular cloning (5-7), and the predicted amino acid sequence indicated that CD20 is a transmembrane protein with four transmembrane domains with both C- and N-terminals located in the cytoplasm. Hence, the structure of CD20 resembles those of ion channels and ion transporters. Indeed, when expressed in fibroblasts, CD20 functions as a calcium-permeable cation channel (8). In lymphocytes, CD20 is phosphorylated by protein kinases including calmodulin-dependent protein kinase. Furthermore, CD20 associates with src family tyrosine kinases including p53/56^lyn, p56^lck, and p59^fyn (9). Since an addition of monoclonal antibody against CD20 induces tyrosine phosphorylation of several proteins (10), CD20 may also participate in the protein tyrosine kinase cascade. Nevertheless, the regulatory mechanism modulating the activity of CD20 is largely unknown and the ligand that activates CD20 has not been identified.

To investigate the function of CD20 as a calcium-permeable channel, we stably expressed CD20 in Balb/c 3T3 fibroblasts (11). CD20 expressed in these cells functioned as a calcium-permeable channel and modulated the growth characteristics of these cells. Thus, CD20 expression accelerated cell cycle progression through the G₁ phase and enabled the cells to progress to the S phase in medium containing low extracellular calcium (11). Insulin-like growth factor-I (IGF-I)¹ is a progression factor that induces G₁ progression (12). As described by Stiles et al. (12), IGF-I exerts its action in a cell cycle-dependent manner. When IGF-I is added to quiescent Balb/c 3T3 cells, it cannot induce cell cycle progression (12, 13). In contrast, cells progress toward the S phase in response to IGF-I when they are first exposed to platelet-derived growth factor followed by epidermal growth factor (14). These are referred to as primed competent cells (14, 15). Therefore, IGF-I exerts its progression activity specifically in primed competent, but not quiescent, cells. However, when CD20 is expressed in quiescent Balb/c 3T3 cells expressing CD20, at least some of the cells progress toward S phase in response to IGF-I (11). This result suggests that IGF-I can elicit progression, even in quiescent cells with the aid of CD20, and implies that a CD20-like protein is critical for the progression activity of IGF-I. If so, it is possible that the function of CD20 expressed in Balb/c 3T3 cells is modulated by IGF-I. In the present study, we investigated this notion. The results indicate that IGF-I, by acting on the IGF-I receptor, activates the channel activity of CD20.

EXPERIMENTAL PROCEDURES

Recombinant human IGF-I was supplied by Fujisawa Pharmaceutical Co. Ltd. (Osaka, Japan). Na[¹²⁵I] was obtained from ICN Biomedicals (Costa Mesa, CA). [³H]Thymidine and [³²P]dCTP were obtained from Dupont NEN. mAb against CD20 (CBL456) was purchased from Cymbus Bioscience Ltd. (Southampton, UK).

Balb/c 3T3 cells (clone A31) and Raji cells (B lymphoblastoid cell line) were provided by the RIKEN cell bank (Tsukuba, Japan). Balb/c 3T3 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (Life Technologies, Inc.). Raji cells were cultured in RPMI 1640 medium containing 10% fetal calf serum. These cells were cultured under humidified conditions of 95% air and 5% CO₂ at 37 °C.

The inducible CD20 expression vector (CD20-pMEP4) was stably transfected into Balb/c 3T3 cells by electroporation as described previously (11). CD20 expressing quiescent Balb/c 3T3 cells were obtained by incubating confluent cells in Dulbecco's modified Eagle's medium containing 0.5% platelet-poor plasma and 80 µM ZnCl₂ for 24 h. After the treatment with ZnCl₂, all of the cells expressed CD20 (11).

DNA synthesis was assessed by measuring [³H]thymidine incorporation into trichloroacetic acid-precipitable materials. Cells were cultured in 24-well plate (Falcon, Lincoln Park, NJ). Queiscent cells were incubated with 0.5 µCi/ml [³H]thymidine for 24 h in the presence of 1 nM IGF-I. The level of [³H]thymidine incorporation was measured as described by McNiel et al. (16).

Constitutively active G_i2 mutant (Gip2) expression vector, Gip2-pcDNA I was generously provided by Dr. H. Bourne of the University of California, San Francisco. Balb/c 3T3 cells were co-transfected with CD20-pMEP4 and Gip2-pcDNA I using a transfection reagent DOTAP (Boehringer Manhein, GmbH, Germany). Twenty-four hours after exposure to DNA, cells were selected for 3 weeks of culture in the presence of 100 µg/ml hygromycin B. Hygromycin-resistant colonies were independently picked up and screened by Northern blotting for high expression of Gip2 and CD20.

The cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_c) was monitored using fura-2, as described previously (11). Briefly, cells cultured on glass coverslips were incubated with 2 µM fura-2 acetoxymethyl ester (Dojin Laboratories, Kumamoto, Japan) for 20 min at room temperature (20-26 °C), then placed on a flow-through chamber mounted on the stage of TMD microscope (Nikon, Tokyo, Japan). The perifusion medium comprised 137 mM NaCl, 5 mM KCl, 1.25 mM CaCl₂, 1 mM MgCl₂, 5 mM glucose, and 10 mM Hepes/NaOH (pH 7.4). Dual wavelength microfluorometry of the fura-2 fluorescence was performed using CAM-230 (Nihon Bunko, Tokyo, Japan). The emission signals excited at both 340 and 380 nm and the ratio of these signals (340/380 ratio) was recorded. In some experiments, the cytoplasmic free Ca²⁺ concentration was calibrated as described elsewhere (17). Statistical significance was evaluated by analysis of variance.

The perforated-patch (18) and the cell-attached patch clamp techniques were applied for the voltage-clamp studies. Micropipettes were pulled from borosilicate glass capillaries and heat-polished at the tip. They had resistance values between 4 and 8 megohms after filling with a pipette solution. High resolution membrane currents were recorded using an EPC-9 patch-clamp amplifer (HEKA, Lambrecht, Germany) controlled by "E9SCREEN" software on an Atari computer. All voltages were corrected for a liquid junction potential between the bath and pipette solutions. Voltage ramps were of 300-ms duration, covering a range of 100 to +100 mV. Capacitance and series resistance were canceled before each voltage ramp using the automatic neutralization routine of the EPC-9. For studies using the perforated whole-cell configuration, the micropipettes for electrical recording were filled with a solution containing 143 mM cesium-aspartate, 4 mM MgSO₄, 1 mM EGTA, 200 µg/ml nystatin (Sigma; 200 mg/ml; dissolved in dimethyl sulfoxide) and 10 mM HEPES (pH 7.2, adjusted by adding CsOH). The bath contained 140 mM N-methyl-D-glucamine-methanesulfonic acid, 10 mM Ca(OH)₂, and 10 mM HEPES (pH 7.4). Statistical significance was evaluated by Student's t test.

Single-channel currents were recorded as described by Hamill et al. (19). The signal was stored on video tape after analogue/digital conversion (Sony PCM 501 ES, modified by Shoshin EM Corp., Okazaki, Japan). For studies using the cell-attached configuration, the micropipettes were filled with a solution containing 110 mM BaCl₂ or CaCl₂, 200 nM tetrodotoxin, and 10 mM HEPES (pH 7.4, adjusted by adding Ba(OH)₂ or Ca(OH)₂). In some experiments, Cl was replaced with aspartate. The bath solution contained 137 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1.25 mM CaCl₂, 5 mM glucose, and 10 mM HEPES (pH 7.4, adjusted with NaOH). Single channel recordings were analyzes by using the TAC program (HEKA). The total number of functional channels (N) in the patch were estimated by observing the number of peaks detected on the amplitude histogram. As an index of channel activity, NP_o (number of channels multiplied by the open probability) was calculated as

(Eq. 1)

where T is the total record time, n is the number of channels open, and t_n is the recording time during which n channels are open. Therefore, NP_o can be calculated without making assumptions about the total number of channels in a patch or the open probability of a single channel. All electrophysiological experiments were performed at 20-26 °C.

Cells were harvested, and the total RNA was isolated using Isogene and quantified spectrophotometrically. Total RNA (20 µg) was resolved by electrophoresis on 1.2% agarose gels containing 2.2 M formaldehyde, 20 mM MOPS (pH 7.0), 8 mM sodium acetate, and 1 mM EDTA, and transferred to a nylon membrane (Hybond-N⁺⁺, Amersham Corp.). by means of capillary blotting in 10 × sodium citrate buffer. Hybridization was performed with a probe labeled with [³²P]dCTP by random priming according to the manufacture's instructions (Pharmacia Biotech Inc.). The hybridized membrane was exposed to Kodak XAR film (Eastman Kodak Co.).

RESULTS

We showed that IGF-I stimulates DNA synthesis when added to quiescent cells expressing CD20, whereas it has no stimulatory effect on DNA synthesis in untransfected quiescent cells (12). As shown in Fig. 1, IGF-I stimulated [³H]thymidine incorporation in CD20-expressing quiescent cells and the effect of IGF-I was inhibited by a monoclonal antibody (mAb) against CD20 in a dose-dependent manner. This monoclonal antibody also inhibited serum-induced DNA synthesis in Raji cells that express native CD20 (data not shown). It should be noted that mAb against CD20 did not affect DNA synthesis induced by IGF-I (Table I). Furthermore, mAb against CD20 did not cross-react with IGF-I assessed by Western blotting (data not shown).

Table I. Effect of mAb against CD20 on IGF-I-induced DNA synthesis in parental Balb/c 3T3 cells Primed competent Balb/c 3T3 cells were incubated for 24 h with 1 nM IGF-I in the presence and absence of 1 µg/ml mAb. Values are the means ± S.E. for four determinations. Addition [3H]Thymidine incorporation cpm IGF-I 36,107  ± 2,162 IGF-I + mAb 35,128  ± 2,632

IGF-I induces oscillatory changes in [Ca²⁺]_c in primed competent cells (17) by activating IGF-operated calcium-permeable channels (15, 20). In quiescent cells, however, IGF-I does not affect [Ca²⁺]_c (16). To test whether or not CD20 is modulated by IGF-I, we first monitored the changes in [Ca²⁺]_c in response to IGF-I using quiescent cells expressing CD20. Fig. 2A shows the typical changes in [Ca²⁺]_c in CD20-expressing cells monitored by measuring the fluorescence of a Ca²⁺ indicator, fura-2. As we reported (11), increasing the extracellular calcium concentration from 10 µM to 2 mM slightly increased [Ca²⁺]_c. When 1 nM IGF-I was added to the cells, [Ca²⁺]_c further increased (43 out of 43 cells). The increase in [Ca²⁺]_c induced by IGF-I was monophasic in 39 out of 43 cells and the net increase in [Ca²⁺]_c was 168 ± 46 nM (means ± S.E., n = 39), which was statistically significant (p < 0.01). The [Ca²⁺]_c of mock-transfected control and untransfected cells did not change significantly under these conditions. In some cells (4 out of 43 cells), IGF-I induced oscillatory changes in [Ca²⁺]_c (Fig. 2B). The effect of IGF-I was reproduced by 100 nM insulin (30 of 30 cells) (Fig. 2C). The IGF-I-induced elevation of [Ca²⁺]_c was dependent on extracellular calcium and its removal (Fig. 2A) or the addition of lanthanum (data not shown) abolished the elevated [Ca²⁺]_c. Elevation of [Ca²⁺]_c induced by IGF-I were completely abolished by the mAb against CD20 and [Ca²⁺]_c returned to the level of that in cells incubated low-calcium containing medium (6 out of 6 cells) (Fig. 2D). These results suggest that IGF-I stimulated Ca²⁺ influx through the CD20 channel in CD20-expressing cells.

Patch-clamp experiments were performed to record the changes in calcium permeability of the membrane induced by IGF-I. We previously reported that IGF-I increases the open probability of a calcium permeable cation channel in primed competent, but not quiescent cells (15, 20). To distinguish CD20 from the native IGF-operated Ca²⁺-permeable cation channel, we used quiescent cells.

First, we examined changes in Ca²⁺ conductance induced by IGF-I using a nystatin-perforated whole-cell patch clamp. The current-voltage (I-V) relationship was obtained by applying voltage ramps from 100 to 100 mV. Fig. 3A shows a comparison of the I-V curves obtained before and 3 min after the addition of IGF-I. These inward currents were generated by Ca²⁺, since Ca²⁺ is the only cation permeant in this condition. At a holding potential of 100 mV, inward current in cells treated with IGF-I was 446 ± 51% (means ± S.E., n = 42) of that in control cells, which was statistically significant (p < 0.01). The I-V curve was not changed as a function of time in the absence of IGF-I (Fig. 3B). The Ca²⁺ current was completely abolished by the mAb against CD20 (Fig. 3B) and, in addition, both La³⁺ and Co²⁺ inhibited the Ca²⁺ currents (data not shown). Similar results were obtained when Ca²⁺ was replaced with Ba²⁺ as a charge carrier. When 100 nM insulin was added, the calcium current was similarly enhanced (Fig. 3C). It is notable that we could not identify IGF-I-stimulated Ca²⁺ permeability using the conventional whole-cell patch clamp procedure. This may be due to the fact that, in whole cell configuration, soluble components of the cytosol are quickly dialyzed by the solution filling the pipettes.

Next, we recorded inward currents in cell-attached patches to further characterize the action of IGF-I on Ca²⁺ influx in cells expressing CD20. Fig. 4 shows a typical record obtained from a cell-attached patch on a CD20-expressing cell with a high concentration of barium (110 mM) in the pipette and a holding potential of 40 mV. An inward current was observed when IGF-I was present in the pipette solution (200 out of 218 patches) (Fig. 4B), whereas little channel activity was evident in a patch without IGF-I (Fig. 4A). The IGF-I-activated inward current was not detected when mAb against CD20 (2 µg/ml) was present in the pipette solution (none of 32 patches) (Fig. 4C). IGF-I may activate CD20 channels either directly or indirectly via a soluble second messenger in the cytoplasm. To elucidate whether the CD20 channel is regulated by a diffusible second messenger, IGF-I was added to the bath solution while recording from cell-attached patches with 110 mM Ba²⁺ in the pipette. IGF-I therefore had access to its receptors located only outside the patch. In this condition, CD20 channels were not activated (none of 28 patches) (Fig. 4D). This finding suggested that IGF-I augments channel activity by a direct mechanism that does not involve a diffusible second messenger. The I-V curve displayed a slope conductance of 7.0 ± 0.6 pS (means ± S.E., n = 8) and extraporation of the data indicates a reversal potential of +20 mV (Fig. 5). Again, a high concentration of insulin (100 nM) reproduced the effect of IGF-I (data not shown). Barium was used as a charge carrier since calcium channels are generally more permeable to this ion, which often permits a better resolution of differences in single-channel amplitudes. However, when barium was substituted for calcium ions, there was no significant difference in the IGF-I-induced current. Moreover, current amplitudes were the same whether chloride or aspartate was the anion in the pipette solution, which confirmed that the currents were carried by an influx of cations (Ca²⁺ or Ba²⁺), rather than by an outward anion flux, which would display negative reversal potentials. Potassium cannot be taken into account since this ion was not in the pipette solution and a high concentration of barium inhibits K⁺ permeability. No voltage dependence was detected and inhibitors of voltage-dependent calcium channel such as nifedipine and verapamil had no effect on the IGF-I-activated currents.

To examine the involvement of pertussis toxin (PTX)-sensitive G protein in IGF-I-induced activation, we studied the effect of IGF-I in PTX-treated cells. When cells were pretreated with PTX (21), the addition of IGF-I in the pipette did not affect the activity of CD20 channel (none of 30 patches) (Fig. 6A). Likewise, IGF-I did not increase calcium current in PTX-treated cells (none of 18 cells) (Fig. 6B). Additionally, IGF-I did not elevate [Ca²⁺]_c in PTX-treated cells (none of 40 cells) (Fig. 6C). Conversely, 50 pM mastoparan, an activator of G_i/G_o class of G proteins (22), markedly stimulated the activity of the CD20 channels (43 out of 43 patches) (Fig. 7A) whereas Mas 17, an analogue which does not activate the G proteins, was ineffective (none of 14 patches) (Fig. 7B).

To further assess the role of a G protein in the regulation of CD20 channel, we transfected Balb/c 3T3 cells expressing CD20 with the cDNA for Gip2, a constitutive active form of G_i2 protein (23). When Gip2 cDNA was expressed, the activity of CD20 channel measured in the cell-attached patch was markedly augmented without the addition of any ligand (32 of 32 patches) (Fig. 8A). When transmembrane calcium current was measured in the whole-cell-perforated patch, calcium permeability was greatly elevated (20 out of 20 cells) (Fig. 8B).

DISCUSSION

In the present study, we examined whether the calcium-permeable cation channel activity of CD20 is affected by IGF-I in CD20-expressing Balb/c 3T3 cells. The notion that CD20 channel is activated by IGF-I in Balb/c 3T3 cells was supported by results obtained by three independent methods: monitoring changes in [Ca²⁺]_c, measurement of whole-cell calcium current and the single channel analysis. It was most explicitly demonstrated by using the perforated patch clamp. As shown in Fig. 4, addition of IGF-I to the bath solution increased the inward calcium current, which was reversed by an anti-CD20 monoclonal antibody. Therefore, it is clear that IGF-I activates the channel activity of CD20 expressed in Balb/c 3T3 cells. The following observations supported the notion that IGF-I exerts its effect via the IGF-I receptor. First, we showed previously that, at a concentration of 10⁹ M, IGF-I predominantly acts on the IGF-I receptor but not on either the IGF-II receptor or the insulin receptor (15). In CD20-expressing cells, 1 nM IGF-I activated the channel activity of CD20. Second, a high concentration of insulin, which does not bind to the IGF-II receptor (24), activated the CD20 channel. Given that Balb/c 3T3 cells do not express significant number of the insulin receptor, a high concentration of insulin bound to the IGF-I receptor and modulated the activity of CD20. At present, the precise mechanism by which the IGF-I receptor activates calcium-permeable CD20 channels is not certain. The IGF-I receptor structurally resembles the insulin receptor (24-26), and it consists of - and -subunits. Ligand binding to the -subunit results in the activation of the intrinsic tyrosine kinase located in the -subunit. It is generally accepted that receptor-associated tyrosine kinase is needed for the signal transduction. Yet, the downstream signal leading to the channel activation is not clear at present. Since a single channel current of CD20 was not activated by IGF-I added outside the patch, the regulatory mechanism by which IGF-I receptor activated the channel may be direct, not involving a soluble second messenger. In this regard, previous studies done in our laboratory showed that IGF-I-mediated calcium influx via the calcium-permeable cation channel is blocked by pertussis toxin (27). In addition, IGF-I-mediated calcium entry is blocked by GDPS and conversely, augmented by GTPS (28). These results indicated that IGF-I, by acting on the IGF-I receptor, modulates the calcium-permeable channel by a mechanism involving pertussis toxin-sensitive G protein. Similarly, IGF-I stimulates calcium influx in Chinese hamster ovary cells by G_i-dependent mechanism (29), although the precise mechanism by which IGF-I activates the channel via G protein is not yet identified. As shown in Fig. 6A, pertussis toxin also abolished the activation of CD20 by IGF-I. Mastoparan, which directly activates pertussis toxin-sensitive G proteins (22), stimulated the CD20 channels. Additionally, the CD20 channel was markedly activated by the co-expression of cDNA for Gip2, a constitutive active form of G_i2. Hence, CD20 can be directly activated by _i2 subunit. Taken together, the mechanisms by which IGF-I activates CD20 and the IGF-operated channel may be similar. Recently, Luttrell et al. (30) demonstrated that activation of MAP kinase by IGF-I is attenuated by pertussis toxin. Their results support the notion that a pertussis toxin-sensitive G protein is involved in the signaling system activated by IGF-I. Despite the fact that CD20 is not a natural effector molecule of the IGF-I signaling system, it may provide a good model system with which to study the regulation of the calcium-permeable channel by the IGF-I receptor. Further study is needed to elucidate the mechanism by which CD20 is regulated by the IGF-I receptor. The present results also provide some insight into the nature of the IGF-operated calcium-permeable channel. Despite of the fact that CD20 channel is ectopically expressed in Balb/c 3T3 fibroblasts, the channel is activated by the IGF-I receptor. This raises an interesting possibility that the putative IGF-operated channel and CD20 may share some structural homology.

CD20 is a cell-surface protein expressed in B lymphocytes. Although it has several functions in the signal transduction system in B cells (8-10), information regarding the ligand that activates CD20 is not available. The present results may provide some insight into the regulation of CD20 functions. An obvious candidate ligand that may activate CD20 is interleukin 4, since this cytokine and insulin share the signaling molecules, insulin receptor substrate-1 and -2 (31, 32). However, interleukin 4 does not activate CD20 channels in Raji cells.² It is possible that a ligand that acts on a receptor system functionally resembling the IGF-I receptor activates the channel activity of CD20. Alternately, the ligand activating PTX-sensitive G proteins may activate the CD20 channel.