Et Al., 1989; Reibnegger Et Survival Responses

Home , My5

Role of the AKT kinase in expansion of multiple myeloma clones: eects on cytokine-dependent proliferative and survival responses

Jung-hsin Hsu1, Yijiang Shi1, Liping Hu1, Myrna Fisher1, Thomas F Franke2 and Alan Lichtenstein*,1

1Department of Medicine and Pathology, West LA VA-UCLA Medical Center and Jonsson Comprehensive Cancer Center, UCLA, Los Angeles, California, CA 90073, USA; 2Department of Pharmacology, Columbia University, New York, NY 10032, USA

IL-6 is an established growth factor for multiple bone marrow of myeloma patients and levels correlate myeloma tumor cells, stimulating proliferative and with disease activity (Bataille et al., 1989; Reibnegger et survival responses. Recent work indicates that IL-6 can al., 1991; (2) Monoclonal anti-IL-6 antibodies can activate the AKT kinase in myeloma cells. Thus, to test induce anti-tumor responses in patients (Klein et al., a potential role for AKT in IL-6-induced cellular 1991); Bataille et al., 1995), and; (3) IL-6 transgenic responses, we transfected myeloma cell lines with an mice are prone to developing MM (Suematsu et al., active `E40K' or dominant negative `PH' AKT construct 1992), while IL-6 knockout mice are resistant (Hilbert using an adenoviral vector. Transfection of the E40K et al., 1995). The source of IL-6 is from the marrow into myeloma cells resulted in enhanced tumor cell microenvironment which stimulates MM cells in a growth and expression of the PH dominant negative paracrine fashion (Klein et al., 1989). AKT resulted in both inhibition of the IL-6-dependent The crucial signal transduction pathways that induce proliferative response and a decrease in S phase IL-6 stimulatory eects on MM cells are under intensive distribution. While transfection of E40K protected study. Previous investigations have implicated the ERK myeloma cells from dexamethasone-induced apoptosis, (Ogata et al., 1997), Jun kinase (Chauhan et al., 1997; the dominant negative PH had no eect on the ability of Xu et al., 1998), and STAT (Catlett-Falcone et al., 1999) IL-6 to protect these cells from dexamethasone. These signaling cascades. However, recent studies suggest a results clearly demonstrate that AKT activation is signal pathway involving the AKT kinase may also be critical for the IL-6 proliferative response. In addition, important in expansion of myeloma clones. AKT is although the level of AKT activation can regulate encoded by c-akt, the cellular homologue of the viral sensitivity to dexamethasone-induced apoptosis, addi- oncogene, v-akt, originally identi®ed in an acute tional cytokine-induced AKT-independent pathways can transforming retrovirus causing T cell lymphomas in mediate IL-6 protection against dexamethasone. mice (Bellacosa et al., 1991). AKT is activated as a Oncogene (2002) 21, 1391 ± 1400. DOI: 10.1038/sj/ major target of phosphoinositol 3-kinase (PI 3-kinase)- onc/1205194 generated signals (Franke et al., 1995; Stokoe et al., 1997). Activation requires phosphorylation of AKT on Keywords: multiple myeloma; AKT; interleukin-6 threonine and serine residues (Klippel et al., 1997; Franke et al., 1995). A newly described membrane bound kinase, 3-phosphoinositide-dependent kinase Introduction (PDK1), phosphorylates AKT on threonine residues. The phospholipid second messengers of the PI 3-kinase In multiple myeloma (MM), enhanced proliferation reaction bind to the pleckstrin homology (PH) domain and increased resistance to apoptotic stimuli account of AKT (Franke et al., 1997) and alter its conformation for the expansion of the malignant clone. It is now so that threonine residues become accessible to PDK1 clear that interleukin-6 (IL-6) can function as a speci®c (Klippel et al., 1997). PDK1 is a constitutively active MM tumor growth factor due to its ability to stimulate kinase that is membrane-bound due to its own PH both proliferative and survival responses in these cells domain binding to low basal levels of phospholipids. In (Anderson et al., 1989; Lichtenstein et al., 1995). contrast, AKT is located in the cytosol of unstimulated Further support for IL-6 as a tumor growth factor in cells but translocates to the cell membrane following vivo includes: (1) IL-6 levels are elevated in plasma and stimulation (Andjelkovic et al., 1997). Thus, the upstream PI 3-kinase-generated phospholipid messengers are crucial as they recruit AKT to the membrane (near PDK1) and also facilitate threonine phosphorylation by *Correspondence: A Lichtenstein, Hematology-Oncology, VA West PDK1. Following threonine phosphorylation, AKT LA Hospital, W111H, 11301 Wilshire Blvd, Los Angeles, CA 90073, then autophosphorylates itself at serine residues (Toker USA; E-mail: [email protected] Received 13 August 2001; revised 30 October 2001; accepted 26 and Newton, 2000). Activated AKT may promote November 2001 oncogenesis by virtue of its ability to activate several AKT in myeloma J-h Hsu et al 1392 distinct downstream pathways which mediate proliferative or survival responses. In fact, the ability of AKT to transform cells (Cheng et al., 1997), of AKT antisense to inhibit transformation (Cheng et al., 1996), the presence of ampli®cation/over-expression of AKT in some solid tumors (Cheng et al., 1992; Miwa et al., 1996) and the fact that the PTEN phosphatase, which inhibits AKT activation, is a known tumor suppressor gene (Wu et al., 1998), all implicate AKT in oncogenesis. Support for a role of AKT speci®cally in myeloma pathogenesis comes from recent studies which demonstrated that IL-6 can activate AKT in myeloma cells (Tu et al., 2000), that some myeloma cell lines harbor PTEN loss-of-function mutations with heightened AKT activation (Hyun et al., 2000), and that the inhibition of myeloma cell growth in vitro could be induced by selective inhibition of PI 3-kinase, the upstream activator of AKT (Tu et al., 2000). Furthermore, in a recent study, we demonstrated frequent AKT activation in situ in myeloma cells of patient bone marrow (Hsu et al., 2001). However, there Figure 1 IL-6 induces wortmannin-sensitive activation and has been no previous assessment of the direct eects of activity of AKT. (a) AKT in vitro kinase assay of AF-10 cells treated with IL-6 (0.5 ng/ml) for 15 (15') or 30 (30') min or for AKT activation or inhibition on myeloma cell growth. 15 min following pre-treatment with wortmannin (0.1 mM) We, thus, initiated the current study to test the eects (15'+W). Immunodetection of phosphorylated GSK-3 is shown. of dominant active or dominant negative AKT Immunoprecipitates were also immunoblotted for total AKT constructs on MM cell responses to IL-6. (lower panel). (b) Western blot for phosphorylated AKT (P-AKT) or total AKT (AKT) in AF-10 cells treated with IL-6 (0.5 ng/ml) or with IL-6 following pre-treatment with wortmannin (IL- 6+W). Duration of IL-6 treatment is shown in minutes. Control Results cells (c) incubated for 15 min in absence of IL-6

AKT activation in multiple myeloma cells transfected with inhibiting AKT activation and AKT phosphorylation active or dominant negative constructs as it had no eect on IL-6-induced phosphorylation of The AF-10 myeloma cell line was used to investigate the p42 and p44 ERK in the same cells (Tu et al., 2000). role of AKT activation in IL-6-induced proliferative The above ability of wortmannin to prevent IL-6- response. This cell line consistently demonstrates an induced AKT activation in AF-10 cells correlated with a approximate twofold increase in its proliferative rate signi®cant (P50.05) inhibitory eect of the drug on IL- when exposed to the MM growth factor IL-6 (Tu et al., 6-induced stimulation of proliferation. Whereas the 2000). In addition, we have previously detected AKT mean IL-6-dependent proliferation index (assayed at activation in AF-10 cells following stimulation with IL- 72 h) in control AF-10 cells was 2.2+/70.4 (mean+/7 6(Tuet al., 2000). A more systematic evaluation of of four experiments), wortmannin at 0.1 mM signi®- AKT activation in this cell line was undertaken to cantly (P50.05) inhibited IL-6-induced cell growth evaluate its kinetics and wortmannin-sensitivity. As (proliferation index of 1.3+/70.2). These data are shown in one of several in vitro kinase assays (Figure similar to the previously demonstrated inhibitory eect 1a), IL-6 activated AKT activity within 15 min of of wortmannin on IL-6-dependent AF-10 cell growth incubation and activity decreased by 30 min. Pre- (Tu et al., 2000) and support a role for PI 3-kinase in treatment with the PI 3-kinase inhibitor wortmannin the IL-6 proliferative response. To speci®cally investi- (at 0.1 mM) abrogated AKT activity. IL-6 and wort- gate the role of AKT in IL-6-stimulated proliferation, mannin had no eect on expression of total AKT we exploited the susceptibility of myeloma cells to (bottom panel). These data are consistent with AKT adenovirus transfection (Teoh et al., 1998; Prince et al., activation occurring downstream of IL-6-induced PI 3- 1998). We transfected cells with an active AKT kinase activation. Similar results were obtained by construct that exhibits enhanced activity or a dominant immunoblotting with phospho-speci®c anti-AKT anti- negative AKT that inhibits endogenous activity. As an body, which recognizes AKT when phosphorylated at active AKT, we utilized the AKT-E40K mutant that serine 473 (Figure 1b, one of ®ve separate experiments). possesses enhanced activity due to a point mutation In the immunoblot analysis, phosphorylated AKT was resulting in an increased anity of the PH domain for ®rst detected by 5 min, increased by 15 min, and was the second messenger phospholipids (Aoki et al., 1998). completely abrogated by pre-treatment with 0.1 mM In addition, the E40K could potentially respond to wortmannin. Wortmannin, again, had no eect on physiological stimuli, like IL-6 (Aoki et al., 1998). An expression of total AKT (bottom panel). Furthermore, adenoviral vector expressing the enhanced green this concentration of wortmannin was speci®c for ¯uorescent protein (EGFP) was used which allowed us

Oncogene AKT in myeloma J-h Hsu et al 1393 to con®rm the sensitivity of AF-10 cells to adenoviral dominant-negative (DN) to inhibit endogenous AKT transfection. At an MOI of 10, transfection of AF-10 activity (Soonyang et al., 1997). Again, using MOIs as cells with either E40K or control vector resulted in low as 10 and 25, we con®rmed successful expression in 490% transfection eciency. Cells were infected with 490% of cells (by EGFP ¯uorescence) in both groups the `E40K' adenovirus at either 10 or 25 MOI, or the of transfected cells. Following adenoviral infection with control virus only expressing EGFP at an MOI of 25. an MOI of 10, AF-10 MM cells were treated with or As shown in Figure 2a, the EGFP control-transfected without IL-6. Anti-HA immunoblots con®rmed expres- control cells were still capable of modest AKT sion of the truncated PH construct and the ability of phosphorylation when exposed to IL-6 for 20 min. In PH to inhibit IL-6-induced AKT activity was demon- AF-10 cells transfected with the `E40K' construct, a strated by in vitro kinase assay (Figure 2b). Similar much higher level of total and phosphorylated AKT eective inhibition of IL-6-induced AKT kinase was expressed and the amount expressed correlated with activity was achieved with an MOI of 25 (not shown). the MOI used (compare E MOI 25 to E MOI 10 in Figure 2a). IL-6 was capable of only a modest further Role of AKT in IL-6-dependent proliferation increase in AKT phosphorylation in cells transfected at MOI 10. By densitometric analysis, IL-6 induced a 2.3- AF-10 cells were ®rst transfected with the active E40K fold increase in AKT phosphorylation (mean of three construct (vs EGFP control) and, 24 h later, treated experiments) at 20 min in AF-10 cells transfected with with or without IL-6. At varying time points after E40K at MOI 10. After transfection with E40K at MOI addition of IL-6, cell recovery was determined by 25, there was little eect of IL-6, possibly because of a Trypan blue and MTT assays. As shown in Figure 3, higher basal AKT phosphorylation. there was a signi®cant (P50.05) increase in cell To inhibit endogenous AKT activity, we infected recovery induced by transfection of the E40K con- AF-10 cells with a virus expressing the HA-tagged struct. At 48 h, there was a 64% increase (3.6 vs pleckstrin homology (PH) domain of AKT using the 2.26105, means of four separate experiments) and, at same adenoviral vector. In other models, this N- 72 h of culture, a 43% increase (4.8 vs 3.356105 cells). terminal regulatory domain of AKT functions as a The MTT assay gave comparable results with a 55% increase in OD (E40K4EGFP control) at 48 h and 48% increase at 72 h (data not shown). There was no dierence in the per cent viability (by dye exclusion assay) between the two groups at either time point. Interleukin-6 increased cell growth in both control and E40K-transfected cells. Although cell recovery by Trypan blue or MTT assays was signi®cantly (P50.05) higher in IL-6-treated E40K-expressing cells compared to IL-6-treated control cells at both 48 and 72 h, the proliferation indices (PIs) were comparable. For example, from the results shown in Figure 3 in one of four separate experiments, the PIs for EGFP and E40K transfected cells at 48 h were 1.82 and 1.72, respectively and, at 72 h, were 1.85 and 1.75. Similar results were seen when MOIs of 25 were used in this experiment, i.e., signi®cantly higher cell recovery in E40K-transfected cells and comparable IL-6 proliferation indices (data not shown).

Figure 2 AKT activation in transfected AF-10 cells. (a) AF- 10 cells transfected with the E40K construct at MOI 10 or 25 or with the control vector (at MOI 25). Twenty-four hours later, cells were treated with IL-6 (0.5 ng/ml) for 5 (5') or 20 (20') min or without IL-6 (7) for 20 min. Western blot for phospho- speci®c AKT (P-AKT) or total AKT (AKT) was performed. (b) Figure 3 Expression of E40K enhances myeloma cell growth. AF-10 cells were transfected with the HA-tagged PH dominant AF-10 cells were transfected with E40K (E) or control vector (C), negative AKT (PH) or control vector (at MOI 10) and, 24 h later, and 24 h later, 105 cells were cultured without or with IL-6 cells were treated or untreated with IL-6 for 20 min, after which (0.5 ng/ml) (C+IL-6; E+IL-6) for an additional 48 or 72 h. AKT in vitro kinase assay (upper part) and anti-HA immunoblot Total viable cells were then recorded. Results are means+s.d. (lower part) was performed determinations from four separate experiments

Oncogene AKT in myeloma J-h Hsu et al 1394 We next transfected AF-10 myeloma cells with the dependent AKT kinase activity in 8226 cells (Figure PH dominant negative AKT (or EGFP control), treated 5b) and IL-6-induced AKT phosphorylation in OCI- them with or without IL-6, and, 72 h later, recorded the My5 cells (Figure 5d). increase in viable cell recovery. There was no signi®cant E40K-transfected MM cells were then exposed to eect of the dominant negative AKT on viable cell dexamethasone, anti-fas or doxorubicin, and assayed at recovery (compared to the control adenoviral infection) varying intervals for induction of cytotoxicity and in MM cells not exposed to IL-6 (not shown). However, apoptosis. There was no eect of E40K expression on expression of the PH AKT in MM cells inhibited IL-6- sensitivity to cytotoxicity (MTT assay) induced by anti- induced cell expansion in culture. As shown in Figure fas or doxorubicin. By MTT assay, the LD50 4a, AF-10 cells infected with control adenovirus (dark (concentration inducing 50% cytotoxicity) for anti-fas bars) are not aected in their ability to respond to IL-6, (assayed at 24 h) was 0.2 mg/ml and 0.3 mg/ml for demonstrating a proliferation index of approximately 2, control-transfected and E40K-transfected 8226 cells, which is comparable to non-infected cells (Tu et al., respectively. For control-transfected and E40K-trans- 2000, and data not shown). However, AF-10 cells fected OCI-My5 cells, it was 0.3 and 0.4 mg/ml, expressing the PH dominant-negative protein (light gray respectively. The LD50 for doxorubicin (assayed at bars) were inhibited in their ability to respond to IL-6 72 h) was 0.4 and 0.3 mM for control-transfected and with respect to increased cell growth. Signi®cant E40K-transfected 8226 cells and 0.2 and 0.1 mM for inhibition (P50.05) was demonstrated at all MOIs similarly transfected OCI-My5 cells. These small between 10 and 50. dierences in LD50s are not signi®cant. A comparable Cell cycle analysis (Figure 4b) con®rmed that, while LD50 between control- and E40-transfected cells was expression of the PH dominant-negative AKT protein also seen when sensitivity to doxorubicin was assayed did not aect cell cycle distribution in AF-10 cells not at 48 h (not shown). An absence of eect of E40K was exposed to IL-6 (compare a to c), it prevented the IL- also seen when we speci®cally assayed maximal 6-induced increase in proportion of cells in S phase. As apoptosis (anti-fas, 0.5 ± 1 mg/ml at 24 h, serum starva- shown, the percentage of cells in S phase in control- tion at 72 h, or doxorubicin, .5 ± 1 mM at 72 h) induced infected cells increased from 32% (panel a) to 52% by these conditions using ¯ow cytometry for expression (panel b) by 24 h of exposure to IL-6. In contrast, in of annexin-V (not shown). However, as shown in PH-transduced AF-10 cells, no IL-6-induced increase Figure 6, expression of E40K signi®cantly protected was observed (S=27% in media (panel c) versus 25% both MM cell lines against cytotoxicity and apoptosis in IL-6-treated cells (panel d)). This latter experiment induced by 1076 m dexamethasone (Figure 6a,b). was repeated twice with identical results. Comparable results were present when apoptosis was assayed by evaluating a sub-G1 peak in propidium iodide (PI)-stained cells (not shown). In addition, the Role of AKT in IL-6-dependent protection against protection against dexamethasone was also present at a apoptosis lower concentration of drug. At 561077 M, dexa- To test the role of AKT in IL-6-induced protection methasone induced 32% cytotoxicity (mean+s.d. of against apoptosis, we utilized the 8226 and OCI-My5 three experiments) of control-transfected 8226 cells but MM cell lines. These MM targets express functional only 6% of cytotoxicity of E40K-transfected cells and IL-6 receptors and exogenous IL-6 protects them induced 27% cytotoxicity of control-transfected OCI against apoptosis (Xu et al., 1998; Lichtenstein et al., cells but only 15% cytotoxicity of E40K-transfected 1995) but does not stimulate proliferation. Thus, we cells. When IL-6 (100 U/ml) was added to E40K ± can isolate cytokine-induced eects on apoptosis with- versus control-transfected cells which were challenged out obfuscating eects on proliferation. In addition, with dexamethasone (1076 M), a comparable degree of both lines demonstrate IL-6-induced AKT activation cytokine-induced protection was seen in both cell types. (Tu et al., 2000) and both are easily transfected with For example, in control-transfected 8226 cells, the our adenoviral vectors (490% transfection eciency degree of dexamethasone-induced apoptosis was de- at MOI of 10). Figure 5a,d demonstrates the ectopic creased from 30 to 14% by addition of IL-6. In E40K- expression of AKT and phosphorylated AKT in E40K- transfected 8226 cells, the per cent apoptosis decreased transfected 8226 and OCI-My5 cells, respectively. IL-6 from 14 to 6% when IL-6 was added. had a very minimal eect on further increasing AKT Even though the IL-6 protective eect in E40K- phosphorylation in E40K-transfected 8226 cells (ap- transfected MM cells was equivalent to control- proximately 1.46control when MOI 10 was used, transfected cells, the above data suggested that expres- Figure 5a) and no eect on E40K-transfected OCI- sion of high amounts of activated AKT could protect My5 cells (Figure 5d). In contrast, IL-6 was capable of MM cells against dexamethasone. However, the level of a modest increase in AKT phosphorylation in both activated AKT induced by IL-6 in non-transfected cells control-transfected cell lines. Figure 5c,e demonstrate is much lower than that induced by E40K transfection. eective expression of the PH AKT dominant negative Thus, to test whether AKT activity was critical for IL-6- construct in 8226 and OCI-My5 cells, respectively. This induced protection in these two MM cell lines, we was shown by immunoblot detection of the HA-tagged transfected the cells with the dominant negative PH or truncated AKT protein. The PH dominant negative control adenovirus. Cells were then challenged with was functional in these MM cells, inhibiting IL-6- dexamethasone and treated with or without IL-6. As

Oncogene AKT in myeloma J-h Hsu et al 1395

Figure 4 PH dominant negative AKT prevents IL-6-induced cell growth. (a) AF-10 cells were transfected with PH (light gray bars) or control (dark black bars) vector at MOI 10, 25 or 50, and, 24 h later, incubated with or without IL-6 (0.5 ng/ml) for an additional 72 h. Proliferation indices were then recorded. Data are means+s.d. of three separate experiments. (b) AF-10 cells were transfected with PH or control vectors and, 24 h later, incubated with or without IL-6 for an additional 24 h. Cell cycle analysis was then performed. Per cent of cells in S phase shown shown in Figure 7, expression of the PH dominant myeloma cell lines, there are IL-6-induced AKT- negative AKT had no eect on the ability of IL-6 to independent survival mechanisms generated. protect either 8226 or OCI-My5 cells from dexamethasone. This was true for MTT assays that measured cytotoxicity or ¯ow cytometric analysis that measured Discussion per cent apoptosis by annexin-V expression (per cent apoptosis shown above the bars in Figure 7). Thus, The results of this study demonstrate IL-6-induced although experiments with the E40K construct demon- activation of AKT in multiple myeloma cells and strate that AKT activation can potentially regulate support a role for the kinase in cytokine-dependent sensitivity to dexamethasone, at least in these two myeloma cell expansion. Ectopic expression of an

Oncogene AKT in myeloma J-h Hsu et al 1396

Figure 5 AKT activation in transfected 8226 and OCI-My5 myeloma cells. (a) 8226 cells transfected with E40K (at MOI 5, 10 or 25) or control (EGFP) (at MOI 25) vectors and, 24 h later, treated or not treated with IL-6 (0.5 ng/ml) for 20 min. Cell extracts were then tested by immunoblot assay for expression of phosphorylated AKT (AKT-P) or total AKT (AKT). (b) 8226 cells transfected with PH or control vectors at MOI 10 and, 24 h later, treated with or without IL-6 for 20 min. AKT in vitro kinase assay then performed by assaying phosphorylation of the AKT substrate GSK-3. Bottom panel shows equal amounts of total AKT were present in immunoprecipitates. (c) 8226 cells transfected with HA-tagged PH dominant negative (PH) or control vector (MOI 10) and anti-HA immunoblot performed 24 h later. (d) OCI-My5 cells transfected with PH, control (EGFP) or E40K vectors (at MOI 10) and, 24 h later, treated without (0) or with IL-6 for 15 or 30 min. Cell extracts were then tested for expression of phosphorylated AKT (AKT-P) or total AKT (AKT). (e) OCI-My5 cells transfected with PH or control vector (c) and immunoblot assay for HA-tagged PH performed 24 h later

active AKT gene with enhanced kinase activity resulted Utilizing immunohistochemistry with a phospho- in a signi®cant increase in viable cell recovery and speci®c anti-AKT antibody, we have previously inhibition of endogenous AKT activation/activity with identi®ed a very frequent incidence of AKT activation a dominant negative gene, prevented an IL-6-induced in myeloma cells in patient bone marrow (Hsu et al., increase in cell cycling. Although enhanced AKT 2001). The current study as well as our previous report activity signi®cantly protected myeloma cells from (Tu et al., 2000) and an additional paper by Chen et dexamethasone-induced apoptosis, inhibition of endo- al., 1999 are all consistent in con®rming the ability of genous activity did not aect the ability of IL-6 to IL-6 to activate AKT. Since myeloma patients are protect these cells from dexamethasone. known to express high levels of IL-6 in bone marrow

Oncogene AKT in myeloma J-h Hsu et al 1397 (Bataille et al., 1989), it is reasonable to indict this cytokine as the stimulus of AKT activation in patient marrow. However, it is possible that other cytokines are equally important. IL-6 is one of several cytokines that stimulate cells through the gp 130 signaling protein. It is, thus, not surprising that other gp 130- signaling cytokines can also activate AKT (Oh et al., 1998). As many of these other gp 130-dependent cytokines can also stimulate myeloma cell responses in vitro (Zhang et al., 1994), and are present in vivo, they may also contribute to AKT activation in patient cells. In addition, IGF-1, a known myeloma growth factor and also present in myeloma marrow, potently stimulates AKT activation in myeloma cells (Tu et al., 2000; Hyun et al., 2000). The molecular mechanism that mediates IL-6-induced AKT activation in myeloma cells is unclear. In AF-10 cells, activation is wortmannin-sensitive and abrogated by a dominant-negative PI3-kinase construct (Tu et al., 2000). In addition, IL-6 treatment of hepatocytes rapidly induces tyrosine phosphorylation of p85, the regulatory sub-unit of PI3-kinase (Chen et al., 1999). However, since the gp 130 protein does not associate with the SH2 domain of p85 after cytokine-treatment of Figure 6 Transfection of E40K protects against dexamethasone. cells, nor does it contain consensus binding sites for p85 (a) 8226 or OCI-My5 cells were transfected with control (con) or SH2, intermediate signal proteins must be operating. E40K vectors (MOI 10) and then, 24 h later, treated with or without dexamethasone (1076 M) for an additional 72 h. Per cent Possibilities include ras, JAK-1, STAT-3, or vav. All are cytotoxicity was then determined as described in Materials and potentially activated by IL-6 (Ogata et al., 1997; Catlett- methods. Data are means+s.d. of three separate experiments. (b) Falcone et al., 1999; Lee et al., 1997) and all are capable The same cells described above in a were assayed for apoptosis by of activating PI3-kinase/AKT downstream (Pfeer et annexin-V expression. Groups are control-transfected cells treated without (C) or with (C/dex) dexamethasone or E40K-transfected al., 1997; Oh et al., 1998; Shigematsu et al., 1997; cells treated without (E) or with (E/dex) dexamethasone. Data are Franke et al., 1995; Foschi et al., 1997). It will be means+s.d. of three experiments particularly clinically relevant if ras is an important upstream activator as up to 40% of myeloma tumor cell populations harbor oncogenic ras mutations. Interleukin-6 is known to stimulate myeloma cell proliferation as well as survival when challenged with anti-fas, serum starvation, doxorubicin or dexamethasone. Some experimental evidence indicates these two cytokine-dependent eects are dissociable and, thus, possibly mediated by distinct mechanisms (Lichtenstein et al., 1995). To ®rst assess the role of AKT in MM cell proliferation, we enhanced or inhibited AKT activation/activity by gene transfer using adenoviral vectors. The active E40K construct signi®cantly increased cell growth in vitro, indicating a role for AKT in MM cell expansion. Although IL-6 was capable of further AKT activation in E40K-transfected cells, this was minimal and possibly due to stimulation of the remaining endogenous AKT rather than the E40K protein. In any case, the ability of IL-6 to stimulate MM cell expansion in E40K-transfected cells Figure 7 PH dominant negative AKT does not prevent IL-6- was not greater than the IL-6 eect in control- induced protection against dexamethasone. 8226 or OCI-My5 transfected cells when comparing proliferation indices. cells were transfected with control (C) or PH dominant negative (PH) vectors (MOI 10) and, 24 h later, treated with dexametha- An additional possibility for this lack of increased sone (1076 M) in the presence or absence of IL-6 for 72 h. Per proliferation index is that, at high levels of activated cent cytotoxicity was determined and shown in bar graphs as AKT such as induced by E40K transfection, additional means+s.d. of three experiments. Per cent apoptosis was also IL-6-dependent pathways may become rate-limiting for determined by annexin-V expression and is shown in numbers induction of cell growth and they are equally above each bar. The per cent apoptosis in 8226 and OCI-My5 control cells not treated with dexamethasone was 4 and 6% stimulated in E40K and control-transfected cells. For respectively example, prior work (Ogata et al., 1997) con®rms

Oncogene AKT in myeloma J-h Hsu et al 1398 participation of the raf-mek-erk pathway in IL-6- induced by IL-6 (Tu et al., 2000). A second theoretical induced proliferation of MM cells. possibility is that IL-6 can induce additional non-AKT Although we could not show the active AKT anti-apoptotic pathways (like STAT3 (Catlett-Falcone construct enhances IL-6-stimulated growth, the results et al., 1999)), which IGF-1 can not. of experiments utilizing the dominant negative PH In summary, these results indicate activation of AKT con®rm that AKT mediates this cellular AKT occurs in multiple myeloma plasma cells. response. Interruption of AKT function in these cells Because of its central location, activating diverse in the absence of IL-6 had no eect on cell growth in downstream proliferative and anti-apoptotic path- vitro or on cell cycle analysis (32% S phase in control ways, AKT is a promising target for future and 27% S phase in PH-expressing cells, (Figure 4)). molecular-based therapy. In addition, the frequent These data are consistent with the fact that there is activation in myeloma tumor cells compared to non- little basal AKT activation in AF-10 cells. However, malignant cells suggests a therapeutic window may the PH construct prevented the IL-6-induced increase exist in patients. in cell numbers as well as the increase in fraction of cells in S phase (Figure 4). In a previous study (Tu et al., 2000), we showed a similar inhibition of IL-6- induced proliferation of AF-10 cells when cultured Materials and methods with wortmannin, the PI 3-kinase inhibitor. There was no increase in apoptotic death in PH-expressing cells Myeloma cells following IL-6 stimulation, either assayed by Trypan The MM cell lines 8226, OCI-My5, and AF-10 were kind blue staining, cell morphology, or DNA content gifts from Drs J Epstein (University of Arkansas), H analysis. Collectively, these data clearly indict PI 3- Messner (University of Toronto), and James Berenson kinase and AKT as critical eectors of the IL-6- (UCLA). The lines were maintained in vitro as described induced proliferative response. Two possible AKT previously (Lichtenstein et al., 1995; Xu et al., 1998; Tu et substrates that might mediate this proliferative eect al., 2000). are the p70S6kinase and 4E-BP1 translational repres- sor. When phosphorylated downstream of AKT Reagents (possibly via the mTOR intermediate (Sekulic et al., Recombinant IL-6 was purchased from R&D Systems 2000)), both p70 and 4E-BP1 would promote transla- (Minneapolis, MN, USA). Phospho-speci®c and total anti- tion of proteins required for cell cycle transit (Jeeries AKT antibodies were purchased from Biosource, Inc. and et al., 1994; Rosenwald et al., 1995). In preliminary New England Biolabs. All other reagents were purchased experiments we have identi®ed IL-6-induced activation from Sigma Chemical Co. (St. Louis, MO, USA) unless of p70S6kinase activity and 4E-BP1 phosphorylation in otherwise speci®ed. AF-10 cells, which is wortmannin sensitive and inhibited by the PH dominant negative AKT. Expression constructs and transfections Initial experiments with the E40K active AKT The N-terminus AKT construct HA-AKT(PH), which demonstrated a clear AKT-dependent resistance to includes the PH domain but lacks the kinase domain, and the apoptotic eects of dexamethasone but not to anti- the HA-E40K construct, which contains a point mutation in fas or doxorubicin. As dexamethasone is a frequently its PH domain, have been previously described (Soonyang et used and often eective drug in myeloma patients, the al., 1997). The PH functions as a dominant-negative inhibitor high levels of AKT activation we found in patient of endogenous AKT (Soonyang et al., 1997), and the E40K marrow tumor cells (Hsu et al., 2001) may play a role possesses enhanced AKT activity (Aoki et al., 1998). Both in dexamethasone-resistance. To test if AKT played a contain hemagglutinen (HA) epitope tags (Soonyang et al., role in the well documented ability of IL-6 to protect 1997). A recombinant adenovirus that encodes HA-AKT(PH) myeloma cells against dexamethasone, we used the or HA-E40K was generated by homologous recombination in bacteria as previously described (He et al., 1998). The HA- same PH dominant negative vector, transfecting tagged constructs were ®rst cloned into pAdTrack-CMV dierent myeloma cell lines, where we could isolate which also contains the green ¯uorescent protein (GFP) gene. cytokine-dependent eects on apoptosis. The results The resultant construct was linearized and transformed with demonstrate that IL-6-induced protection against the supercoiled adenoviral vector, pAdEasy-1, into E. coli dexamethasone in these cells was due to AKT- and recombinants selected in kanamycin and screened by independent pathways. However, this does not rule restriction endonuclease digestion. The recombinant adeno- out the possibility that AKT may mediate cytokine- virus was then transfected into 293 cells. Control pAdTrack- induced protection of MM cells against dexamethasone CMV recombinant virus was generated in identical fashion in other MM cell lines or when other cytokines are but without transgene insertion. Myeloma cells were used. For example, IGF-1 treatment of 8226 or OCI- transduced with adenovirus at varying MOIs for 2 h. Adenovirus was then washed away and cells were resus- My5 cells protects them against dexamethasone and pended in media with low serum concentration (1% FCS) to such protection is inhibited by co-culture with the PI3- minimize proliferation. At 24 and 48 h, EGFP ¯uorescence kinase inhibitors wortmannin and Ly294002 (Tu et al., demonstrated that 485% of cells were successfully trans- 2000). One dierence between these cytokine treat- fected when using MOIs between 10 and 100. Expression of ments is that AKT activation induced by IGF-1 in HA-AKT(PH) was also con®rmed by Western anti-HA these MM cell lines is much more impressive than that immunoblotting.

Oncogene AKT in myeloma J-h Hsu et al 1399 Western blot analysis Flow cytometry Protein was extracted and separated by 12.5% SDS ± PAGE For cell cycle analysis of adenoviral vector-transfected cells, as previously described (Lichtenstein et al., 1995; Tu et al., MM cells were stained with hypotonic propidium iodide 2000). Proteins were transferred to polyvinylidene di¯uoride (50 mg/ml in 0.1% sodium citrate) and 0.1% Triton X-100 for membranes and phosphorylated as well as total AKT proteins 1 h at 48C. Cells were kept in the dark at 48C before analysis. were detected as previously described (Lichtenstein et al., Cell cycle distribution was determined by analysing 10 000 ± 1995; Tu et al., 2000). Relative expression of phosphorylated 15 000 events on a FACScan ¯ow cytometer (Becton AKT was determined by densitometry. Dickinson, San JoseÂ , CA, USA). The DNA data were ®tted to a cell cycle distribution analysis by use of the MODFIT program for MAC V2.0. AKT kinase assay The AKT in vitro kinase assay utilized a non-radioactive kit Apoptosis assay purchased from New England Biolabs. AKT was ®rst immunoprecipitated from cell extracts and then incubated Neo-expression of membrane annexin-V was used to with GSK-3 fusion protein in the presence of ATP and kinase determine per cent apoptosis. Since transfected cells expressed buer. AKT-dependent GSK-3 phosphorylation was then green ¯uorescence due to expression of EGFP, phycoerythrin measured by immunoblotting using a phospho-GSK-3 anti- (PE)-conjugated annexin-V (Pharmingen) was employed. body that recognizes GSK-3 when phosphorylated. Cells were stained with PE-annexin-V and 7-amino-actino- mycin-D (7-AAD) according to the vendor's directions. Early apoptotic cells were identi®ed by ¯ow cytometry as PE- Treatment of cells and determination of cell numbers annexin-V-positive and 7-AAD-negative. MM cells were cultured at 26105/ml in 6 or 12-well plates in complete media. IL-6 was added to some groups at a Statistics concentration of 5 ng/ml. After 24 ± 72 h of culture, cells were harvested and cytotoxicity (an assessment of proliferation as The t-test was used to determine signi®cance of dierences well as survival) determined by Trypan blue staining or MTT between groups. assays as previously described (Aparicio et al., 1998). IL-6- induced expansion of cell numbers was described as a proliferation index, calculated as number of viable cells (or OD in MTT assays) in experimental groups divided by number of cells (or OD) in the control group (no IL-6). To test sensitivity to cytotoxicity, cells were treated with increasing concentrations of anti-fas antibody (clone CH11, Acknowledgments Upstate Biotech.), doxorubicin, or dexamethasone. At varying The authors thank the UCLA ¯ow cytometry core time points, per cent cytotoxicity was determined as follows: laboratory of the Jonsson Comprehensive Cancer Center (1 ± OD (or viable cell count) or exp group/OD (or viable cell for assistance and Dr Jay Persselin for his continued count) of control group6100). The LD50 (concentration of support. A Lichtenstein was supported by research funds drug inducing 50% cytotoxicity) was determined from of the Veteran's Administration, including the Research extrapolation of results of MTT assays, whereby percentage Enhancement Awards Program entitled `Cancer Gene cytotoxicity was plotted against log10 concentration. To Medicine', a year 2000 Senior Research Award from the speci®cally test sensitivity to apoptosis, ¯ow cytometry was Multiple Myeloma Research Foundation, and Grant performed on annexin-V-stained cells (below). #CDA DAMD17-0-1-0214, awarded by the US Army.

References

Anderson K, Jones R, Morimoto C, Leavitt P and Barut B. ChauhanD,KharbandaS,OgataA,UrashimaM,TeohG, (1989). Blood, 73, 1915 ± 1922. Robertson M, Kufe D and Anderson K. (1997). Blood, 89, Andjelkovic M, Alessi DR, Meier R, Fernandez A, Lamb 227 ± 236. NJ, Frech M, Cron P, Cohen P, Lucocq JM and ChenR-H,ChangM-C,SuY-H,TsaiY-TandKuoM-L. Hemmings BA. (1997). J. Biol. Chem., 272, 31315 ± 31324. (1999). J. Biol. Chem., 274, 23013 ± 23019. Aoki M, Batista O, Bellacosa A, Tsichlis P and Vogt P. Cheng JQ, Godwin AK, Bellacosa A, Taguchi T, Franke T, (1998). Proc. Natl. Acad. Sci. USA, 95, 14950 ± 14955. Hamilton TC, Tsichlis PN and Testa JR. (1992). Proc. Aparicio A, Gardner A, Tu Y, Savage A, Berenson J and Natl. Acad. Sci. USA, 89, 9267 ± 9274. Lichtenstein A. (1998). Leukemia, 12, 220 ± 229. Cheng JQ, Ruggeri B, Klein WM, Sonoda G, Altomare DA, Bataille R, Jourdan M, Zhang X-G and Klein B. (1989). J. Watson DK and Testa JR. (1996). Proc. Natl. Acad. Sci. Clin. Invest., 84, 2008 ± 2011. USA, 93, 3636 ± 3643. Bataille R, Barlogie B, Lu Z-Y, Rossi J-F, Lavabre-Bertrand Cheng JQ, Altomare DA, Klein MA, Lee WK, Kruh GD, T, Beck T, Wijdened J, Brochier J and Klein B. (1995). Lissy NA and Testa JR. (1997). Oncogene, 14, 2793 ± 2799. Blood, 86, 685 ± 691. Foschi M, Chari S, Dunn MJ and Sorokin A. (1997). EMBO Bellacosa A, Testa JR, Staahl SP and Tsichlis PN. (1991). J., 16, 6439 ± 6448. Science, 254, 274 ± 277. Franke TF, Yang S-I, Chan TO, Datta K, Kazlauskas A, Catlett-Falcone R, Landowski T, Oshiro MM, Turkson J, Morrison DK, Kaplan DR and Tsichlis PN. (1995). Cell, Levitzki A, Savino R, Ciliberto G, Moscinski L, 81, 727 ± 738. Fernandez-Luna JL, Nunez G, Dalton W and Jove R. Franke TF, Kaplan DR, Cantley LC and Toker A. (1997). (1999). Immunity, 10, 105 ± 115. Science, 275, 628 ± 630.

Oncogene AKT in myeloma J-h Hsu et al 1400 HeTC,ZhouS,daCostaLT,YuJ,KinzlerKWand Prince HM, Dessureault S, Gallinger S, Krajden M, Suther- Vogelstein B. (1998). Proc. Natl. Acad. Sci. USA, 95, land DR, Addison C, Zhang Y, Graham FL and Stewart 2509 ± 2514. AK. (1998). Exp. Hemat., 26, 27 ± 35. Hilbert D, Kopf M, Mock B, Kohler G and Rudiko S. Reibnegger G, Krainer M, Herold M, Ludwig H, Watcher H (1995). J. Exp. Med., 182, 243 ± 248. and Huber H. (1991). Cancer Res., 51, 6250 ± 6255. HsuJ-H,ShiY,KrajewskiS,RennerS,FisherM,ReedJC, Rosenwald IB, Kaspar R, Rousseau D, Gehrke L, Leboulch Franke T and Lichtenstein A. (2001). Blood, (in press). P, Chen JJ, Schmidt EV, Sonenberg N and London IM. HyunT,YamA,PeceS,XieX,ZhangJ,MikiT,GutkindJS (1995). J. Biol. Chem., 270, 21176 ± 21180. and Li W. (2000). Blood, 96, 3560 ± 3568. Sekulic A, Hudson CC, Homme JL, Yin P, Otterness DM, Jeeries HBJ, Reinhard C, Kozma SC and Thomas G. Karnitz L and Abraham RT. (2000). Cancer Res., 60, (1994). Proc. Natl. Acad. Sci. USA, 91, 4441 ± 4445. 3504 ± 3513. Klein B, Zhang X-G, Jourdan M, Content J, Houssiau F, ShigematsuH,IwasakiH,OtsukaT,OhnoY,ArimaYand Aarden L, Piechaczk M and Bataille R. (1989). Blood, 73, Niho Y. (1997). J. Biol. Chem., 272, 14334 ± 14343. 517 ± 526. Soonyang Z, Baltimore D, Cantley LC, Kaplan DR and Klein B, Wijdenes J, Zhang X, Jourdan M, Boiron J, Franke T. (1997). Proc. Natl. Acad. Sci. USA, 94, 11345 ± Liautard J, Merlin M, Clement C, Morel-Fournier B, Lu 11350. Z, Manoni P, Sany J and Bataille R. (1991). Blood, 78, Stokoe D, Stephens LR, Copeland T, Ganey PRJ, Reese 1198 ± 1203. CB, Painter GF, Holmes AB, McCormick F and Hawkins Klippel A, Kavanaugh WM, Pot D and Williams LT. (1997). PT. (1997). Science, 277, 567 ± 572. Mol. Cell. Biol., 17, 338 ± 347. Suematsu S, Matsusaka T, Matsuda T, Ohno S, Miyazaki J- Lee I-S, Liu Y, Narazaki M, Hibi M, Kishimoto T and Taga I, Yamamura K-I, Hirano T and Kishimoto T. (1992). T. (1997). FEBS Lett., 401, 133 ± 143. Proc. Natl. Acad. Sci. USA, 89, 232 ± 235. Lichtenstein A, Tu Y, Fady C, Vescio R and Berenson J. TeohG,ChenL,UrashimaM,TaiY-T,CeliL,ChenD, (1995). Cell Immunol., 162, 248 ± 255. Chauhan D, Ogata A, Finberg RW, Webb IJ, Kufe DW Miwa W, Yasuda J, Murakami Y, Yashima K, Sugano K, and Anderson K. (1998). Blood, 92, 4591 ± 4597. Sekine T, Kono A, Egawa S, Yamaguchi K, Hayashizaki Toker A and Newton AC. (2000). J. Biol. Chem., 275, 8271 ± Y and Sekiya T. (1996). Biochem. Biophys. Res. Com., 225, 8274. 968 ± 974. Tu Y, Gardner A and Lichtenstein A. (2000). Cancer Res., Ogata A, Chauhan D, Urashima M, Teoh G, Hatziyanni M, 60, 6763 ± 6770. Vidriales VMB, Schlossman R and Anderson K. (1997). J. Wu X, Senechal K, Neshat MS, Whang Y and Sawyers C. Immunol., 159, 2212 ± 2220. (1998). Proc. Natl. Acad. Sci. USA, 95, 15587 ± 15591. Oh H, Fujio Y, Kunisada K, Hirota H, Matsui H, Kishimoto Xu F-H, Sharma S, Gardner A, Tu Y, Raitano A, Sawyers C T and Yamauchi-Takihara K. (1998). J. Biol. Chem., 273, and Lichtenstein A. (1998). Blood, 92, 241 ± 249. 9703 ± 9710. Zhang XG, Gu JJ, Lu ZY, Yasukawa K, Yancopoulos GD, Pfeer L, Mullersman JE, Pfeer SR, Murti A and Yang CH. Turner K, Shoyab M, Taga T, Kishimoto T, Bataille R (1997). Science, 276, 1418 ± 1425. and Klein B. (1994). J. Exp. Med., 179, 1337 ± 1344.

Oncogene