Growth Inhibition of a Human Myeloma Cell Line by All-transRetinoic Acid Is Not Mediated Through Downregulation of Interleukin-6 Receptors but Through Upregulation of p21^WAF1

Human myeloma OPM-2 cells were cultured in RPMI 1640 medium (GIBCO, Grand Island, NY) with 10% fetal bovine serum (Intergen, Purchase, NY), 10 mmol/L HEPES buffer, pH 7.2, 100 U/mL penicillin, and 100 μg/mL streptomycin (GIBCO) at 37°C in 5% CO₂ in humidified air. Cells in logarithmic growth phase were exposed to various concentrations of ATRA or Dex (Sigma Chemical Co, St Louis, MO) for 2 to 3 days. ATRA, dissolved in dimethyl sulfoxide (Sigma) at 40 mmol/L, was diluted with culture medium before use. The residual dimethyl sulfoxide was previously shown not to affect the growth of myeloma cells.1 Dex was dissolved in phosphate-buffered saline at 8 mmol/L and diluted with medium for use in cultures.¹²⁵I-labeled recombinant human (rh) IL-6, 1,200 and 2,234 Ci/mmol in two separate batches, was obtained from Amersham, Searle Pharmaceuticals (Arlington Heights, IL).

An expression vector designed to achieve constitutive expression of IL-6Rα in OPM-2 cells was constructed by insertion of the IL-6Rα cDNA in the Rc/CMV vector (Invitrogen), which contained the cytomegalovirus (CMV) promoter. The IL-6Rα cDNA was removed from the pBSF2R.23b clone,10 a generous gift from Dr T. Kishimoto (Osaka University, Osaka, Japan), by partialXhoI digestion. The 2.3-kb XhoI fragment containing IL-6Rα cDNA was purified by electrophoresis in low-melting agarose,HindIII linkers attached, and ligated into the HindIII site of the Rc/CMV vector. Transformations were performed using the TOP 10 Escherichia coli strain, and colonies containing the IL-6Rα cDNA insert identified by colony hybridization. Restriction enzyme mapping was performed to identify clones containing the IL-6Rα cDNA in correct orientation with respect to the CMV promoter in the Rc/CMV vector. OPM-2 cells were transfected with the expression vector by electroporation (200 V, 1,080 μF, 1 second; 10 × 10⁶ cells; 20 μg of BglII linearized Rc/CMV IL-6Rα plasmid DNA). Cells were then plated in 96-well plates (2,000 cells/well) in RPMI medium containing 500 μg/mL of G418. From a total of two plates, six G418-resistant clones were isolated and designated as clones A1, C2, C3, C4, C5, and D6. The remaining transfected cells, grown in batch in RPMI medium containing G418 without further cloning, were designated as “pooled” transfectants.

Cell-surface IL-6Rα was detected by indirect immunofluorescence. Briefly, 1 to 2 × 10⁶ myeloma cells, pretreated in minimal essential medium with 10 mmol/L HEPES buffer, pH 7.2, 0.3% bovine serum albumin (BSA), and 0.1% sodium azide at 4°C for 10 minutes, were sequentially incubated for 45 minutes in 70 μL of biotinylated goat anti-human IL-6R antibody (15 μg/mL medium; R & D Systems, Minneapolis, MN) and avidin-fluorescein isothiocyanate conjugate (15 μg/mL medium; Sigma). Goat IgG was used as antibody control. After incubation at 4°C with constant mixing, 0.43 mL of ice-cold medium was added, and cells were collected by centrifugation over a cushion of 0.5 mL of medium containing 5% BSA. The fluorescent signals were detected using a Becton Dickinson FACScan flow cytometer (Becton Dickinson, San Jose, CA) and acquired with both linear and logarithmic amplifiers. The mean channel number of a fluorescence intensity histogram in linear scale minus that of antibody control provided a relative measure of cell-surface IL-6Rα density.

¹²⁵I-IL-6 binding assay was performed as previously described.1 Briefly, myeloma cells were washed twice with warm culture medium and once with the binding medium after incubation at 4°C for 10 minutes in the binding medium, consisting of Dulbecco’s modified Eagle medium with 0.3% BSA, 10 mmol/L HEPES, pH 7.2, and 0.1% sodium azide. 0.3 to 1 × 10⁶ cells were then incubated for 150 minutes with a graded concentration of¹²⁵I-IL-6 in 60 μL of binding medium in an ice-water bath with constant mixing. 0.44 mL of ice-cold binding medium was then added and the whole volume was layered over a 0.5-mL cushion of the binding medium containing 5% BSA and centrifuged at 3,000 rpm for 10 minutes. With the supernatant removed, the tube was cut just above the pellet, and the total (specific plus nonspecific) cell-bound radioactivity was determined. The nonspecific binding was measured by the addition of 100× excess of unlabeled IL-6 and was subtracted from the total cell-bound radioactivity to yield specific cellular binding. The binding data were analyzed by Scatchard method.

Cellular extracts were prepared from 10 to 20 × 10⁶myeloma cells in 1 mL of extraction phosphorylation lysis buffer consisting of 50 mmol/L HEPES, pH 7.4, 150 mmol/L NaCl, 1.5 mmol/L MgCl₂, 10 mmol/L sodium pyrophosphate, 1 mmol/L EDTA, 100 mmol/L NaF, 100 μmol/L sodium orthovanadate, 1% Triton X-100 (Boehringer Mannheim, Mannheim, Germany), 10% glycerol, 20 μg/mL aprotinin, and 0.5 mmol/L phenylmethylsulfonyl fluoride (PMSF). Protein concentrations were determined by the Bradford method using the Biorad Protein Assay dye reagent (Biorad, Cambridge, MA). The immunoprecipitation and immunoblotting assays were performed as previously described.11 12 Briefly, the cell lysates with equal protein loads were immunoprecipitated with rabbit anti-human gp130 (Santa Cruz Biotech, Santa Cruz, CA) using protein-G sepharose (Pharmacia LKB Biotechnology, Piscataway, NJ). After five washes with phosphorylation lysis buffer containing 0.1% Triton X-100, proteins were analyzed by sodium dedocyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinyl difluoride filters (Immobilon; Millipore, Bedford, MA). The residual binding sites on the filters were blocked by incubating with TBST (10 mmol/L Tris HCl, pH 8.0, 15 mmol/L NaCl, 0.05% Tween 20) containing 20% BSA for 1 hour at room temperature. The filters were subsequently incubated with antiphosphotyrosine monoclonal antibody (4G10; Upstate Biotechnology, Lake Placid, NY) or anti-human gp130 and developed with an enhanced chemiluminescence (ECL) kit following the manufacturer’s recommended procedure (Amersham, Arlington Heights, IL).

Total cell extracts of OPM-2 and C5 cells were prepared with a high-salt RIPA buffer consisting of 50 mmol/L Tris, pH 7.4, 300 mmol/L NaCl, 1 mmol/L EDTA, 1% NP40, 0.25% sodium deoxycholate, 100 mmol/L NaF, 5 μg/mL aprotinin, 1 μg/mL leupeptin, and 1 mmol/L Na₃NO₄. Retinoblastoma protein from extracts (200 μg protein) was immunoprecipitated by incubation with anti-RB antibody (OP28; Calbiochem, San Diego, CA) for 1 hour at 4°C, followed by overnight incubation with Protein A/G sepharose (Santa Cruz Biotech). The precipitated protein was washed with RIPA buffer (50 mmol/L Tris, pH 7.4, 150 mmol/L NaCl, 1 mmol/L EDTA, 1 % NP40, 0.25% sodium deoxycholate, 100 mmol/L NaF, 5 μg/mL aprotinin, 1 μg/mL leupeptin, 100 μg/mL PMSF, and 1 mmol/L Na₃VO₄) and separated in 5% SDS-PAGE. Immunoblots, prepared as described above, were incubated with a second anti-RB antibody (sc-50-G; Santa Cruz Biotech) followed by peroxidase-conjugated anti-goat Ig (Santa Cruz Biotech) and visualized using ECL reagents.

Nuclear extracts were prepared from approximately 10 × 10⁶ cells. Cells were first washed and incubated in 400 μL of buffer A (10 mmol/L HEPES, pH 7.9, 1.5 mmol/L MgCl₂, 10 mmol/L KCl, 0.5 mmol/L dithiothreitol, 1 μg/mL aprotinin, 1 μg/mL leupeptin, 100 μg/mL PMSF, and 1 mmol/L Na₃VO₄) on ice for 10 minutes. Cells were vortexed for 10 seconds and followed by a 10-second centrifugation. The pellets were resuspended in 50 μL of buffer C (20 mmol/L HEPES, 320 mmol/L NaCl, 1.5 mmol/L MgCl₂, 0.2 mmol/L EDTA, 0.5 mmol/L dithiothreitol, 25% vol/vol glycerol, 1 μg/mL aprotinin, 1 μg/mL leupeptin, 100 μg/mL PMSF, and 1 mmol/L Na₃VO₄), incubated on ice for 20 minutes followed by a 2-minute centrifugation. The extracted proteins were separated by 12% SDS-PAGE and transferred to nitrocellulose membranes as described above. Immunoblots were sequentially stripped and probed with anti-p21 (sc-397-G; Santa Cruz Biotech), anti-p27 (sc-528-G), anti-CDK2 (sc-163-G), and anti-CDK4 (sc-260-G), followed by an appropriate anti-goat or anti-rabbit Ig second antibody (Santa Cruz Biotech), and visualized with ECL reagents.

The clonogenic growth assay was performed as previously described.1 Briefly, 800 cells in 0.2 mL of Iscove’s medium (Sigma), containing 30% prescreened, heparinized, normal human plasma, 0.8% methylcellulose, and 5 × 10⁻⁵mol/L 2-mercaptoethanol, were plated in 48-well plates. Colony growth was scored after 2 to 3 weeks of incubation. Conventional flow cytometric DNA analysis with propidium iodide staining was performed as previously described.1 Fluorescence intensity histograms were analyzed with ModFit program (Variety Software House, Inc, Sunnyville, CA).

Flow cytometric analysis of cellular IL-6Rα expression showed increased expression of IL-6Rα in transfectants. Figure 1 displays the fluorescence intensity histograms of a representative analysis. The respective mean channel numbers (MCN) in linear scale of the unstained, antibody control (background), and stained cells were 30, 57, and 274 for the parental OPM-2 cells, and 25, 51, and 670 for “pooled” transfectants. Thus, the IL-6Rα density, expressed as MCN corrected for the background, was 217 and 619 for the parental cells and “pooled” transfectants, respectively. The analysis of the isolated stable transfectant clones are summarized in Table 1 (column A,a). IL-6Rα expression increased 1.5- to 6-fold in the transfectant clones except clone A1, where it was close to that of the parental OPM-2 cells, indicating loss or loss of function of the transfected IL-6Rα cDNA in this clone. A repeat experiment showed similar results.

Radioligand binding assays were also performed on parental, clone A1, C4, and C5 cells. Scatchard analysis of the binding data yielded an IL-6Rα density of 237 binding sites/cell and a dissociation constant of 1.5 × 10⁻¹⁰ mol/L for the parental line. The respective values were 431 sites/cell and 1.2 × 10⁻¹⁰ mol/L for clone A1; 2,742 sites/cell and 2.2 × 10⁻¹⁰ mol/L for clone C4; and 2,385 sites/cell and 2.4 × 10⁻¹⁰ mol/L for clone C5, consistent with the relative IL-6Rα density detected by flow cytometry. The binding site estimate for parental OPM-2 cells was comparable to our previous measurement.1 The estimates of the dissociation constant, however, were higher in the present study.1 The reason for this discrepancy is not clear. Nevertheless, the dissociation constants were clearly of similar magnitude for the parental and transfected clones, indicating that the expressed IL-6Rα from the transfected IL-6Rα cDNA in transfectants were as avid as the endogenous IL-6Rα in specific ligand binding.

The expression of IL-6Rα was analyzed before and after the exposure of cells to ATRA at 10⁻⁶ mol/L for 3 days. These conditions were previously shown to result in downregulation of IL-6Rα, but not of gp130, in OPM-2 cells.1Flow cytometric analysis (Table 1, column A,b) showed that while IL-6Rα expression of the parental and A1 clones was reduced after ATRA exposure to 40% to 58% of the untreated control, it was increased to 134% to 220% of the control in the remaining transfectant clones. The effect of ATRA on clone C5 was also analyzed by the radioligand binding assay. The binding sites per cell and the dissociation constant were 2,663/cell and 2.8 × 10⁻¹¹ mol/L before and 3,783/cell and 3.7 × 10⁻¹¹ mol/L after ATRA treatment, an increase in binding sites to 142% of the untreated control. This magnitude of increase was almost identical to that measured by flow cytometry shown above (139% of untreated control). There were no significant changes in the dissociation constant with ATRA treatment. The increase in IL-6Rα expression with ATRA treatment most likely reflected the effect of ATRA on CMV promoter.13 It is of interest to note that, for clone A1, whose transfected IL-6Rα cDNA was apparently lost or nonfunctional, IL-6Rα expression was also reduced with ATRA exposure, reflecting the expression of the endogenous IL-6Rα gene.

The binding of IL-6 to IL-6Rα initiates homodimerization and tyrosine phosphorylation of gp130, an early step in IL-6 signal transduction.14 To determine the functionality of the expressed IL-6Rα in parental and transfected cells, IL-6–induced gp130 phosphorylation was determined. Ten to 20 × 10⁶cells were washed twice with serum-free RPMI medium, and “starved” in 1 mL of serum-free medium at 37°C for 1 hour. Recombinant human IL-6 (Immunex, Seattle, WA) was then added to the final concentration of 100 ng/mL. After 10 minutes of incubation, cells were harvested and cell lysates prepared. Lysates of OPM-2 and C5 cells pre-exposed to 10⁻⁶ mmol/L of ATRA for 3 days were similarly prepared. The phosphorylation of gp130 was analyzed after immunoprecipitation. The results (Fig 2) showed that the parental OPM-2 and C5 transfectant clones contained comparable quantities of gp130 protein (anti-gp130 immunoblots, lanes 1, 2, 5, and 6). The level of gp130 was not greatly affected by pretreatment with ATRA (lanes 3, 4, 7, and 8). The addition of IL-6 triggered a rapid tyrosine phosphorylation of gp130 in parental and C5 cells (antiphosphotyrosine immunoblots, lanes 2 and 6). In OPM-2 cells pretreated with ATRA, however, IL-6–induced gp130 phosphorylation was greatly reduced (lane 4), consistent with ATRA-induced downregulation of functional IL-6Rα. In contrast, substantial IL-6–induced gp130 phosphorylation was observed in C5 cells exposed to ATRA (lane 8), indicating that the expressed IL-6Rα were functionally equivalent to the endogenous receptors. Similar results were obtained in a repeat study.

Table 1 (columns B) summarizes the results of the clonogenic growth assays of myeloma cells in the presence or absence of 6.1 × 10⁻⁷ mmol/L of ATRA, a concentration previously shown to cause 99% suppression of the growth of parental OPM-2 cells (IC₉₉). The growth of parental cells was suppressed by 82%, while the growth of transfected clones was similarly suppressed, ranging from 91% to 98%. Similarly, ATRA at 9.9 × 10⁻⁸ mmol/L (IC₅₀) suppressed the growth of parental and transfectant clones to a similar degree, ranging from 60% to 81% (data not shown). Analysis of the cell-cycle distribution showed that exposure of cells to ATRA at IC₉₉ for 2 days resulted in a reduction of cells in S phase and an accumulation of cells in G0/G1 phases in the transfected clones studied (A1, C4, and C5), similar to the parental OPM-2 cells (Table 2). ATRA at IC₁₀, however, was less effective on C4 and C5 clones. The results indicate a block at the G1-S phase transit (Table 2).

The drug sensitivity of the representative transfectant C5 clone was further analyzed in detail. The dose-response analysis yielded an IC₅₀ of 3.0 × 10⁻⁸ mol/L for ATRA, compared with 9.9 ± 4.6 × 10⁻⁸ mol/L (n = 5) for the parental OPM-2 line. Similar to the parental line, the inhibition of C5 cells by ATRA was not reversed by IL-6: the colony growth was 14 ± 8/well (n = 3) with ATRA at IC₅₀, as compared with the control of 155 ± 47 colony/well, and was not reversed by the addition of 10 ng/mL of IL-6 (0.3 ± 0.6/well). The sensitivity to Dex was also analyzed, showing an IC₅₀ of 9.1 × 10⁻⁸mol/L, compared with 5.6 ± 1.5 × 10⁻⁸ mol/L (n = 6) for the parental line. The inhibitory effect of dexamethasone was again reversible: the colony growth in culture with Dex at IC₉₀ was 1 ± 1/well (n = 3), compared with the control of 59 ± 7/well. The addition of IL-6 at 10 ng/mL to the Dex-treated culture yielded 74 ± 9 colony/well, reversing the inhibition by Dex. Similarly, the drug sensitivity profile of clone A1 was identical to that of the parental OPM-2 cells (data not shown). Thus, clone C5 and A1 cells, the latter serving, in effect, as vector control, were fully responsive to IL-6 in reversing the inhibitory effect of Dex, indicating that the IL-6 postreceptor transduction pathway was not altered by the transfection processes and remained functionally intact.

The above findings clearly indicate that the growth inhibition by ATRA is not mediated by modulation of IL-6R. Thus, we investigated the possibility that ATRA affects the IL-6 postreceptor pathway. We first examined its effect on pRB phosphorylation, as IL-6 has previously been shown to promote myeloma cell growth through pRB phosphorylation.15 Myeloma cells were exposed to 10⁻⁶ mol/L of ATRA. At 48 and 72 hours, aliquots of cells were obtained and total cellular proteins were extracted and immunoprecipitated. Western blot analysis (Fig 3) showed that pRB was constitutively phosphorylated (p-RB) in OPM-2 cells. Incubation with ATRA caused pRB dephosphorylation and dephosphorylated pRB was the predominant form after 72 hours of ATRA exposure. We next conducted Western blot analysis, screening for the expression of cell-cycle regulatory proteins, p21, p27, CDK2, and CDK4, on the nuclear extracts of cells before and after a 24-hour ATRA exposure. Figure 4 showed that p21 was upregulated by ATRA after 24 hours of incubation, while the expression of CDK2, CDK4, and p27 was not substantially affected. Kinetic studies of p21 upregulation showed that p21 was not or only minimally expressed in untreated cells, but was upregulated within 24 hours of ATRA exposure, and was markedly increased in 48 hours (Fig5).The activation of p21 and dephosphorylation of pRB is consistent with our previous finding that ATRA caused G1 arrest in OPM-2 cells.1 To determine whether the activation of p21 and pRB dephosphorylation by ATRA is dependent on the cellular expression of IL-6Rα, we compared the effect of ATRA on OPM-2 parental line and C5 clone. Figures 3 through 5 showed that ATRA upregulated p21 and dephosphorylated pRB in OPM-2 and C5 cells to a similar extent and in a similar time course. Thus, the upregulation of p21 and consequent dephosphorylation of pRB is independent of the presence or absence of functional IL-6R. Similar results were obtained on repeat studies. To determine whether reconstitution of functional IL-6Rα allows for the reversal of ATRA effect by IL-6, we exposed C5 cells to 10⁻⁶ mmol/L ATRA, 100 ng/mL of IL-6, or both for 48 hours. Aliquots of cells were analyzed by DNA flow cytometry. Cellular extracts were also prepared and analyzed by Western blot for the expression of p21 and pRB. Figure 6 shows that similar to our previous findings with the parental OPM-2 cells,1 ATRA caused a G1 block of C5 cells, which was not reversed by simultaneous treatment with IL-6. IL-6 neither affected p21 expression and pRB phosphorylation nor reversed the upregulation of p21 and dephosphorylation of pRB by ATRA in C5 clones. These results further support that the action of ATRA is independent of IL-6R.

To test the hypothesis that ATRA inhibited the growth of myeloma cells by downregulation of IL-6Rα expression, we constructed an IL-6Rα expression vector, in which IL-6Rα cDNA was linked to the CMV promoter. Transfection of OPM-2 myeloma cells with this vector yielded stable transfectants expressing high levels of IL-6Rα. In all six stable transfectants, except clone A1, high levels of cellular IL-6Rα were shown both by immunofluorescence with specific anti-human IL-6Rα antibody and ¹²⁵I-IL-6 binding assay. The expressed IL-6Rα in these transfectants was shown to bind ¹²⁵I-IL-6 with a dissociation constant comparable to that of the parental OPM-2 cells. In contrast to the parental cells, ATRA treatment resulted in an increase in IL-6Rα expression in these transfectants. Thus, with the exception of clone A1, the transfected IL-6Rα cDNA in transfectants was effective in reconstituting IL-6Rα under ATRA treatment. To vigorously assess the functional integrity of the expressed IL-6Rα in transfectants, clone C5, which had the highest IL-6Rα expression, was analyzed in detail. It was shown that the gp130 content of C5 cells was comparable to that of the parental OPM-2 cells. In both OPM-2 and C5 cells, IL-6 induced rapid tyrosine phosphorylation of gp130, an early step in IL-6 signal transduction. Pretreatment of OPM-2 cells with ATRA greatly reduced the IL-6–induced phosphorylation of gp130, correlated with the downregulation of cellular IL-6Rα expression and the reduced binding of IL-6 shown previously.1 In contrast, C5 cells, pretreated with ATRA, continued to bind ¹²⁵I-IL-6 with high affinity and the addition of exogenous IL-6 caused intense gp130 phosphorylation (Fig 2). Clearly, the expressed IL-6Rα in C5 cells in the presence of ATRA were functionally equivalent to endogenous receptors, capable of binding ligand and initiating gp130 phosphorylation. Comparing the drug sensitivity of the parental and C5 clones, we found that both clones were equally sensitive to the growth inhibitory effect of ATRA and Dex with comparable IC₅₀. In both parental and C5 clones, IL-6 entirely reversed the inhibition by Dex but not by ATRA. Because evidence strongly indicated that Dex acted through downregulation of IL-6 expression, as discussed above, the retained Dex-sensitivity in C5 cells and its reversal by exogenous IL-6 strongly suggested that the IL-6 postreceptor signal transduction pathway was not altered by transfection and remained functionally intact. With functional IL-6Rα reconstituted and an intact postreceptor transduction pathway, the clonogenic growth of these transfectants, however, was still strongly inhibited by ATRA. Thus, in OPM-2 myeloma cells, the growth inhibitory effect of ATRA is clearly independent of the downregulation of IL-6Rα.

Studies on the postreceptor pathway showed that, in the parental OPM-2 cells, ATRA upregulated p21, but not CDK2, CDK4, or p27 expression, and caused dephosphorylation of pRB. Our previous studies have shown that the increased level of p21 was sufficient to inhibit CDK2 activity by 60%.16 Thus, upregulation of p21 and consequent dephosphorylation of pRB may be a key mechanism of ATRA effect in the G1 arrest and growth inhibition of myeloma cells. The role that IL-6R modulation may play in the activation of p21 and pRB by ATRA was further investigated by comparing the effect of ATRA in the parental OPM-2 and C5 clones. As shown above, the pattern of p21 upregulation and pRB dephosphorylation was identical between OPM-2 and C5 clones, indicating that their activation is independent of IL-6R expression. Reconstitution of functional IL-6R in C5 clone also failed to impart the reversibility of ATRA effect by IL-6, providing further evidence that the ATRA effect is not mediated through downregulation of IL-6R.

The exact mechanism of ATRA growth inhibition, however, requires further investigation. A disruption and, thus, a loss of IL-6 growth signaling remains a possibility, as a block at a site further downstream of IL-6 postreceptor signal pathway may still be present, even though the initial step of gp130 phosphorylation in IL-6 signal transduction is apparently unaffected by ATRA, as shown above. Because multiple alternative signaling paths exist subsequent to gp130 activation,17 it is also conceivable that ATRA may “redirect” IL-6 signaling away from growth, perhaps, toward differentiation signaling. Whether a loss of IL-6 growth signaling can lead to p21 upregulation is not known, although IL-6 has been shown to suppress Dex-induced p21 expression in IL-6–responsive myeloma cells.18 Alternatively, ATRA activation of p21 could be separate and independent of its effect on IL-6 signal transduction, and represents the triggering of a negative growth signal. This appears likely as it has been shown in U937 myelomonoblastic cells that ATRA induction of p21 was dependent on a retinoic acid response element in the p21 promoter.19 Thus, p21 could be directly upregulated through the family of RAR and RXR nuclear receptors. Furthermore, ATRA has been shown to upregulate STAT-1 in U937, NB4 promyelocytic, and MCF-7 breast cancer cells,20-22 raising the possibility that ATRA may additionally upregulate p21 through the induction of STAT1 that subsequently activates a STAT-responsive element in the p21 promoter.23 Further studies on the effect of ATRA on the IL-6 postreceptor signaling should be facilitated by the use of stable transfectants with constitutive IL-6Rα expression.

The present study is also of particular interest and relevance to the general concept of growth regulation through modulations of the expression of cytokine/growth factor receptors. There are many examples of changes in cytokine/growth factor receptors accompanying growth modulation by various agents. In myeloma, in addition to ATRA, interferons have also been shown to inhibit myeloma growth while downregulating IL-6R in some IL-6–dependent cell lines.24,25 In a variety of cell types, the inhibition of cell growth by retinoic acid is associated with modulations of a number of cytokine/growth factor receptors.26 For example, the growth arrest of an epidermoid carcinoma cell line27,28 and sensitive glioma cells29 is associated with reduced expression of the epidermal growth factor receptor (EGFR). On the other hand, RA-induced cellular differentiation and growth inhibition in embryonal carcinoma,30 neuroblastoma,31 and HL-60 leukemia cell lines32 is accompanied by an upregulation of transforming growth factor β receptors (TGFR) and an induced-susceptibility to the growth inhibition by TGF-β, suggesting the activation of negative growth regulatory pathways. Other cytokines and hormones, such as TGF-β,33 tumor necrosis factor-α,34 interferon-γ,35IL-2,36 and dihydrotestosterone,37 can also modulate various cytokine/growth factor receptors, including receptors for hematopoietic growth factors, TGF-β, and EGF, in cell types ranging from hematopoietic stem cell lines to prostate and hepatocellular carcinoma cell lines. Thus, despite the keen awareness of the complex, pleiotropic effects of cytokines and hormones and the recurring question of whether the alterations in receptors might be the consequence of cellular phenotypic changes associated with cell growth and differentiation,26 the modulation of cytokine/growth factor receptors has been widely implicated as a simple and yet elegant mechanism of the growth regulation by these agents.25,26 38This concept, to our knowledge, has not been critically analyzed, and the present study represents the first attempt at formally testing this model. In our case, the failure of the reconstituted functional IL-6Rα to reverse the growth inhibition by ATRA is clearly inconsistent with the simple model of growth regulation through quantitative modulation of cytokine/growth factor receptors. Our study points out the need for careful validation of this concept in various systems, where modulations of cytokine/growth factor receptors have been implicated as the mechanisms of growth regulation by biologic and pharmacologic agents.

We thank Dr Ronald Hoffman for encouragement and support.