Lenalidomide Promotes p53 Degradation by Inhibiting MDM2 Auto-ubiquitination in Myelodysplastic Syndrome with Chromosome 5q Deletion

Abstract

Allelic deletion of the RPS14 gene is a key effector of the hypoplastic anemia in patients with myelodysplastic syndrome (MDS) and chromosome 5q deletion [del(5q)]. Disruption of ribosome integrity liberates free ribosomal proteins to bind to and trigger degradation of MDM2, with consequent p53 transactivation. Herein we show that p53 is overexpressed in erythroid precursors of primary bone marrow del(5q) MDS specimens accompanied by reduced cellular MDM2. More importantly, we show that lenalidomide acts to stabilize MDM2, thereby accelerating p53 degradation. Biochemical and molecular analyses showed that lenalidomide inhibits the haplodeficient PP2Acα phosphatase resulting in hyperphosphorylation of inhibitory serine-166 and serine-186 residues on MDM2, and displaces binding of RPS-14 to suppress MDM2 auto-ubiquitination; whereas PP2Acα over expression promotes drug resistance. Bone marrow specimens from del(5q) MDS patients resistant to lenalidomide over-expressed PP2Acα accompanied by restored accumulation of p53 in erythroid precursors. Our findings indicate that lenalidomide restores MDM2 functionality in the 5q- syndrome to overcome p53 activation in response to nucleolar stress, and therefore may warrant investigation in other disorders of ribosomal biogenesis.

Introduction

The 5q- syndrome is a pathologically and cytogenetically distinct subtype of myelodysplastic syndrome (MDS) (1, 2). Affected individuals have an isolated interstitial deletion involving chromosome 5q [del(5q)] accompanied by a refractory macrocytic anemia, erythroid hypoplasia, megakaryocytic dysplasia and low risk for progression to acute leukemia. The commonly deleted region (CDR) resolved by deletion mapping involves a 1.5 megabase segment extending between bands 5q32 and 5q33 containing 44 genes (3, 4). Although haploinsufficiency of several genes in the CDR are believed to contribute to the disease phenotype, allelic deletion of the ribosomal processing gene, RPS14, a component of the 40S ribosomal subunit, is a key determinant of ineffective erythropoiesis (2, 5–8). Using an RNA interference screen of the CDR genes, Ebert et. al. showed that only inactivation of RPS14 impaired erythroblast proliferation and viability, while over-expression of RPS14 was sufficient to rescue erythropoiesis in primary del(5q) MDS specimens (5). Disruption of ribosome assembly as a result of deletion or mutation of genes encoding ribosomal proteins (RP) leads to nucleolar stress and sequestration of the human homologue of the E3 ubiquitin ligase mouse double minute 2 protein (MDM2) by free RP, triggering its autologous degradation and consequent p53 stabilization (9–11). In normal CD34⁺ cells, shRNA suppression of RPS14 gene expression activates p53 and corresponding expression of its target genes in an erythroid lineage restricted fashion (12). The RP-MDM2-p53 pathway has emerged as a critical effector of the erythroid hypoplasia characteristic of del(5q) MDS and congenital anemias involving RP gene mutations such as Diamond-Blackfan anemia and Schwachman-Diamond syndrome (13–15). A murine model of the human 5q- syndrome generated by allelic deletion of the syntenic genes in the human CDR showed that p53 inactivation completely rescues the hematopoietic phenotype, indicating that the molecular pathogenesis of the 5q- syndrome is p53-dependent (16).

The thalidomide analogue, lenalidomide, is highly active in del(5q) MDS, restoring effective erythropoiesis in more than two-thirds of patients (17, 18). We recently reported that gene dosage of two dual specificity phosphatases encoded within or adjacent to the CDR, Cdc25C (cell division cycle 25C) and PP2Acα (protein phosphatase 2A catalytic domain alpha), underlies the selective suppression of del(5q) clones (19). Lenalidomide inhibits the Cdc25C and PP2A phosphatases, which are key regulators of the G2/M checkpoint, resulting in sustained G2/M arrest and induction of apoptosis in cells with reduced gene expression (19). Given the importance of the RP-MDM2-p53 pathway in the hematologic phenotype of del(5q) MDS, we investigated the regulatory effects of lenalidomide on p53. We show that erythroid precursors in del(5q) MDS over-express p53, and that lenalidomide stabilizes MDM2 by inhibiting its auto-ubiquitination to promote p53 degradation. Inhibition of the haplodeficient PP2A phosphatase by lenalidomide results in hyperphosphorylation of inhibitory serine residues on MDM2 and abrogates binding to RPS14 to suppress MDM2 autologous ubiquitin ligase activity, whereas PP2Acα over-expression promotes resistance. Moreover, development of resistance to lenalidomide in del(5q) MDS patients was associated with up-regulation of PP2Acα, with consequent restoration of p53 activation and hypoproliferative anemia. Our findings indicate that lenalidomide restores MDM2 functional activity in del(5q) MDS to overcome p53 activation in response to nucleolar stress. This pathway may represent a novel therapeutic target in patients with acquired resistance to lenalidomide and warrants investigation in other disorders of ribosomal protein deficiency.

Results

Lenalidomide stabilizes MDM2 and decreases p53 accumulation in del(5q) Namalwa cells and primary MDS specimens

Nucleolar stress arising from allelic deletion of RPS 14 disrupts ribosome assembly, resulting in stabilization of p53 in del(5q) MDS (5) (9, 12, 13, 20). MDM2 is a key negative regulator of p53 in response to ribosomal stress that is essential to rescue primitive erythroid progenitors from p53-mediated apoptosis (21). Given the importance of MDM2 neutralization in the hematologic phenotype of ribosomopathies, we investigated the pattern of MDM2 and p53 expression in Namalwa cells that harbor a chromosome 5q deletion (22). Expression of MDM2 in untreated cells was low, accompanied by corresponding up-regulation of p53 (Figure 1a). Treatment with lenalidomide for 24 hours, however, resulted in a dose dependent induction of MDM2 protein accompanied by decreased p53 accumulation (Figure 1a). This effect can be detected after both 48 and 72 hours of treatment with lenalidomide (Figure 1b and c).

Figure 1. Lenalidomide induces MDM2 expression and down-regulates p53 in del(5q) Namalwa cells and primary MDS specimens.

(a), Lenalidomide (Len) induces MDM2 expression and down-regulates p53 protein in a concentration dependent manner. Namalwa cells were exposed to various concentrations of lenalidomide for 24 hours and Western blotted for p53 and MDM2. (b–c) Cells were also exposed to various concentrations of lenalidomide for either 48hours (b) or 72 hours (c) at which time protein lysates were prepared and immunoblotted with anti-MDM2 antibody. (d) BM-MNCs isolated from four MDS patients with del(5q) were treated with 5 µM lenalidomide (Len) or vehicle (DMSO) for 24 hours or 48 hours before Western blot analysis with either anti-MDM2 or anti-p53 antibodies (upper panel). Densitometric analysis was performed demonstrating MDM2 stabilization and accelerated p53 degradation by lenalidomide (lower panel bar graph). A total of eight MDS patients with del(5q) were tested and data from 4 patients are illustrated.

To determine if lenalidomide exerts similar effects on MDM2 and p53 in primary del(5q) MDS bone marrow specimens, we performed similar analyses using BM-MNCs isolated from eight individuals with del(5q) MDS. P53 expression was demonstrable in all patient specimens by Western blot analysis prior to lenalidomide exposure, accompanied by little or no discernable MDM2 (Figure 1d upper panel). Treatment with lenalidomide induced MDM2 protein accumulation after 24 and 48 hours of drug exposure, while immunodetection of p53 was reduced or eliminated. DMSO, the vehicle control for lenalidomide had no effect, indicating a drug specific effect of lenalidomide on MDM2 and p53. The ratio of MDM2 to p53 was analyzed using densitometric analysis showing a significant reduction in p53 and corresponding stabilization of MDM2 induced by treatment with lenalidomide (Figure 1d lower panel, bar graph).

Lenalidomide inhibits p53 target gene activation and disrupts RPS14/MDM2 association

To determine if RPS14 interacts with MDM2 in the setting of ribosomal stress and investigate the effects of lenalidomide, Namalwa cells were treated with low concentrations of actinomycin D (Act.D), which induced p53 expression. As shown in Figure 2a, Act.D treatment increased RPS14 binding to MDM2 (Figure 2a, lane 2, middle panel), while treatment with lenalidomide induced MDM2 protein expression (Figure 2a, top panel) and either reduced (lane 3) or abolished RPS14 association with MDM2 (lane 4) at low and high concentrations, respectively. Interestingly, lenalidomide interference with RPS14 binding to MDM2 shows relative specificity in Namalwa cells which harbor a complex karyotype with chromosome 5q deletion. Association of RPS19, RPL5 and RPL11 (data not shown for RPL5 and RPL11) with MDM2 is also increased following Act. D exposure, but lenalidomide did not alter binding to MDM2 either at low (5 µM) or high concentrations (20 µM) (Figure 2a, lower panel). More importantly, expression of p21 and PUMA, downstream targets of p53, were similarly reduced (Figure 2b), suggesting that nucleolar stress induced by perturbation of the RP-MDM2-p53 pathway may be overcome by lenalidomide treatment in del(5q) MDS. To determine if lenalidomide’s modulation of RPS14/MDM2 interaction and corresponding protein stabilization is specific for del(5q) progenitors, we preformed similar experiments in U937 cells lacking the chromosome 5q deletion, Act.D treatment increased RPS14 and RPS19 binding to MDM2 in U937 cells, however, lenalidomide had no modulatory effect (Figure 2 c), suggesting that the inhibitory effect of lenalidomide is restricted to del(5q) progenitors.

Figure 2. Effect of lenalidomide on RPS14 and MDM2 interaction and the regulation of p53 transcriptional targets.

(a) Namalwa cells were treated with actinomycin-D (Act. D) for 16 hours to induce nucleolar stress following treatment with or without lenalidomide at 5 or 10 µM for 48 hours (lanes 3, 4,). Whole cell lysates were prepared and immunoprecipitated with anti-MDM2 antibody before Western blotting with either anti-MDM2, anti-RPS14 or anti-RPS19 antibodies as indicated. (b) Whole cell lysates from (a) were Western blotted with either anti-p53, anti-p21 or anti-PUMA antibodies as indicated. Anti-actin antibody was used as loading controls (bottom panel). (c) U937 cells were treated with Act. D for 16 hours following treatment with or without lenalidomide at 5 or 10 µM for 48 hours (lanes 3, 4,). Whole cell lysates were prepared and immunoprecipitated with anti-MDM2 antibody before Western blotting with either anti-MDM2, anti-RPS14 or anti-RPS19 antibodies as indicated. The results are representative of four separate experiments.

Lenalidomide modulates MDM2-p53 association via inhibition of PP2A

To investigate how lenalidomide modifies the RP-MDM2-p53 cascade, we evaluated phosphorylation of p53 and MDM2, key post-translational modifications in the regulation of both proteins (23, 24). Although the precise kinase(s) involved in the phosphorylation of these regulatory sites remains unclear, phosphorylation and dephosphorylation occur rapidly, resulting in either activation or inactivation of the proteins. PP2A, a lenalidomide sensitive dual specificity phosphatase, has been implicated in the regulation of p53 by dephosphorylating thr55 and ser46, thereby preventing proteasome-degradation and consequent induction of the cyclin-dependent kinase inhibitor p21 (WAF1/Cip1) (19, 25, 26). We investigated whether lenalidomide inhibition of the haplodeficient PP2Acα perturbs p53 stability. Treatment of Namalwa cells with lenalidomide promoted p53 degradation with dose proportional increased phosphorylation of thr55 and ser46, consistent with PP2A inhibition (Figure 3a). Similarly, in response to appropriate cell cycle stimuli, PP2A is recruited to the p53-MDM2 complex to dephosphorylate serine/threonine residues on MDM2 and facilitate MDM2-p53 disassociation (27, 28). This prompted us to investigate whether inhibition of PP2Acα by lenalidomide promotes MDM2 specific phosphorylation. Treatment of Namalwa cells with lenalidomide not only enhanced total serine phosphorylation of MDM2, but specifically increased phosphorylation of key regulatory sites at ser166 and ser186 (Figure 3b), residues involved in the regulation of MDM2 auto-ubiquitination. We did not observe similar changes in phosphorylation of threonine residues. Phosphorylation of MDM2 at ser166 and ser186 is mediated by Akt, which is responsible for its nuclear localization and blocking p19ARF binding, thus increasing p53 degradation (29). This suggests that lenalidomide may enhance phosphorylation of ser166 and ser186 through inhibition of PP2A phosphatase activity. Indeed, treatment of cells with lenalidomide for 48 hours promoted MDM2 nuclear localization demonstrated by MDM2 specific immunostaining (Figure 3c). Quantitative analysis of immunofluorescent images indicates that more than 61% of cells treated with lenalidomide showed increased MDM2 nuclear localization compared to untreated cells which displayed only 23% MDM2 nuclear localization (Figure 3d). To evaluate the effect of lenalidomide on the endogenous interaction of PP2Acα and MDM2, we immunoprecipitated PP2Acα followed by Western blot analysis of MDM2 after lenalidomide exposure. We found that PP2Acα binds MDM2 following lenalidomide treatment in a concentration dependent manner (Figure 3e), suggesting that lenalidomide modifies p53 stability by modulating MDM2 phosphorylation via inhibition of PP2A phosphatase activity in PP2Acα-haplodeficient del(5q) cells. To confirm that lenalidomide’s effects on p53-MDM2 interaction are PP2A dependent, we over-expressed PP2Acα in Namalwa cells and assessed changes in MDM2 induction and p53 stabilization after lenalidomide treatment. Namalwa cells expressing an HA-tagged PP2Acα construct were treated with lenalidomide for 48 hours and MDM2 and p53 protein expression were analyzed. In comparison to pcDNA3 vector control, PP2Acα over-expression abrogated lenalidomide-mediated MDM2 induction, thereby stabilizing p53 (Figure 3f, left panel). More importantly, increased specific phosphorylation on both p53 and MDM2 induced by lenalidomide was abolished by over-expression of PP2Acα (Figure 3f, right panel). Because PP2A dephosphorylates protein kinase B (PKB)/Akt to suppress kinase activity, inhibition of the haplodeficient phosphatase by lenalidomide in del(5q) progenitors may also activate Akt signaling and in turn, contribute to MDM2 hyperphosphorylation. Lenalidomide treatment of Namalwa cells increased phosphorylation of the regulatory PP2A substrate, thr308 on Akt as well as ser9 on the PKB/Akt substrate, GSK3β (Figure 3g), indicating that lenalidomide activation of PKB/Akt may also contribute to MDM2 phosphorylation. These data confirm the PP2Acα dependent modulation of p53-MDM2 interaction by lenalidomide, and provide a plausible mechanism for development of lenalidomide resistance in del(5q) MDS through the over-expression of PP2Acα.

Figure 3. Analysis of p53 and MDM2 phosphorylation and PP2A-MDM2 association in response to lenalidomide treatment in del(5q) cells.

(a) Treatment with lenalidomide (Len) increases phosphorylation of ser46 and thr55 of p53. Namalwa cells were treated with lenalidomide at various concentrations for 48 hours, the cell lysates were prepared and analyzed by Western blotting with antibodies to MDM2, phospho-ser46 of p53, phospho-thr55 of p53 and total p53. (b) Treatment with lenalidomide increases phosphorylation of MDM2 at ser166 and ser186. Namalwa cells were treated with lenalidomide at various concentrations for 48 hours, the cell lysates were immunoprecipitated using anti-MDM2 antibody. The samples were separated by SDS PAGE and then Western blotted with either anti-MDM2 (top panel) or antibodies specific to phospho-ser166, phospho-ser186, total phospho-serine or total phospho-threonine of MDM2 as indicated (left panel). Densitometric analysis was performed on Western blots shown in left panel and the level of MDM2 phosphorylation was normalized with total MDM2 as indicated (right panel). (c) Lenalidomide promotes MDM2 nuclear translocation. Namalwa cells were treated with lenalidomide or DMSO at a concentration of 10 µM for 48 hours, as indicated. Cells were cytospinned and stained with anti-MDM2 antibody, followed by AlexaFlour-594 secondary antibody. Nuclei were counter stained with DAPI (original magnification 400×). (d) Quantitative analysis of the MDM2 nuclear translocation in lenalidomide treated cells compared to DMSO treated controls. (e) Lenalidomide enhances association of PP2A with MDM2. Namalwa cells were treated with DMSO or lenalidomide at various concentrations as indicated. Protein lysates were prepared and immunoprecipitated with anti-PP2A followed by Western blot analysis with anti-MDM2 antibody. The findings are representative of three separate experiments. (f) Over-expression of PP2Acα down-regulates MDM2 expression induced by lenalidomide. Namalwa cells were transfected with plasmids containing either pcDNA3 vector control or HA-tagged PP2Acα and then treated with lenalidomide at a concentration of 20 µM for 48h. Protein lysates were prepared and Western blotted with antibodies specific to MDM2, p53, HA-PP2A and β-actin (left panel). Protein lysates were also used for immunoprecipitation with anti-MDM2 or p53 antibodies following Western blot analysis with specific phosphorylation antibodies to MDM2 or p53 as indicated (right panel). (g) Protein lysates were prepared from Namalwa cells treated with or without lenalidomide at concentrations of 0.2, 2 or 20 µM and Western blot analysis performed using antibodies specific to either phospho-Akt (thr308) or phospho-GSK3β (ser9) as indicated.

Lenalidomide inhibits auto-ubiquitination of MDM2 and promotes MDM2-p53 association

As an ubiquitin protein ligase, MDM2 regulates proteasomal degradation of p53, and modifies its own stability through autologous ubiquitination (30–32) The latter activity is modulated by phosphorylation of regulatory sites on MDM2, suggesting that PP2A inhibition by lenalidomide may stabilize MDM2 by antagonizing self-ubiquitination. To explore this, MDM2 and His-tagged ubiquitin plasmids were co-transfected into Namalwa cells for 48 hours followed by treatment with lenalidomide to determine whether lenalidomide alters MDM2 ubiquitination. Although all ubiquitinated proteins were immobilized using nickel beads, specific MDM2 degradation is determined by Western blot analysis using MDM2 specific antibodies. Here we show that MDM2 self-ubiquitination is dependent upon ubiquitin co-expression (Figure 4a) and transfection of both MDM2 and His-tagged ubiquitin accelerated MDM2 ubiquitination, as evidenced by the appearance of multiple specific MDM2 bands (Figure 4a, lane 2). In contrast, exposure to both low [0.2 µM] and high [20 µM] concentrations of lenalidomide triggered a concentration-dependent reduction in MDM2 ubiquitination (Figure 4a, lanes 3–4), indicating that lenalidomide stabilizes MDM2 by inhibiting its autologous ubiquitination.

Figure 4. Role of PP2Acα in lenalidomide stabilization of MDM2 in del(5q) cells.

(a) Lenalidomide inhibits MDM2 auto-ubiquitination. Namalwa cells were untransfected or co-transfected with plasmids containing MDM2 and His-tagged ubiquitin following treatment with or without lenalidomide at various concentrations. His-tagged ubiquitinated proteins were purified by Ni-NTA beads and specific ubiquitinated MDM2 was detected by Western blotting with antibodies to MDM2. (b) Lenalidomide enhances MDM2 binding to p53. Namalwa cells were treated with DMSO or lenalidomide at various concentrations for 48 hours and immunoprecipitated with anti-MDM2 prior to Western blotting with either anti-MDM2 (upper panel) or anti-p53 (lower panel). (c) U937 cells were mock infected (lane 1), infected with lentiviral vectors encoding non-target shRNA control (lane 2), shCdc25C (lane 3), shPP2Acα (lane 4), shCdc25C and shPP2Acα (lane 5) treated without (lanes 1–5) or with 10 µM of lenalidomide for 48 hours (lanes 6–10). Protein lysates were prepared and Western blot analysis was performed using antibodies to either MDM2, p53, Cdc25C or PP2A as indicated. (d–f) Protein lysates or mRNA was prepared with either Namalwa cells or U937 cells treated with or without 10 µM lenalidomide for 24 or 48 hours as indicated. Q-PCR (d and e) and Western blot analysis (f) were performed to examine the level of PP2Acα gene expression in response to lenalidomide treatment. (g) Namalwa cells were treated with either control siRNA or siRNA specific to MDM2 before treatment with or without 10 µM of lenalidomide as described in the Methods. Protein lysates were prepared and Western blot analysis was performed using antibodies specific to MDM2 or p53 as indicated.

Lenalidomide treatment of Namalwa cells markedly increased binding between MDM2 and p53 (Figure 4b). Although Namalwa cells harbor an exon 7, codon 248 mutation analogous to that found in the Li-Fraumeni syndrome (33), our findings suggest that this mutation is not critical for p53 interaction with MDM2 in response to lenalidomide treatment. Together, these results suggest that MDM2 rather than p53 is the critical target responsible for lenalidomide-induced p53 degradation, likely mediated by phosphorylation dependent inhibition of MDM2 auto-ubiquitination as shown in Figure 4a.

Reduced gene dosage of PP2Acα is a key determinant of MDM2 stabilization by lenalidomide in del(5q) progenitors

Our previous studies showed that reduced gene dosage of PP2Acα and Cdc25C is responsible for the lenalidomide induced anti-proliferative effect on del(5q) cells (19). To discern the role of each of the lenalidomide sensitive haplodeficient phosphatases in drug stabilization of MDM2, we introduced a recombinant lentiviral-based system carrying shRNA specific for either Cdc25C or PP2Aca gene transcripts into U937 cells that have a non-del(5q) karyotype, followed by exposure to lenalidomide. Our findings show that reduced expression of PP2Acα, is indispensible for lenalidomide induced MDM2 stabilization (Figure 4c lane 9). PP2Acα knockdown led to a 2-fold induction of MDM2 by densitometric analysis, Cdc25C suppression yielded no change in MDM2 expression with lenalidomide treatment compared to controls (Figure 4c lane 8). These findings were supported by analysis of Cdc25C and PP2Acα protein expression levels showing complete shRNA knockdown of the Cdc25C phosphatase (Figure 4c, lanes 3 and 8), whereas, 60% suppression of PP2Acα was sufficient to induce MDM2 with lenalidomide exposure (lanes 5 and 9). Although isolated knockdown of Cdc25C did not stabilize MDM2 with lenalidomide treatment, it enhanced MDM2 stabilization in response to lenalidomide treatment in the setting of the dual knockdown compared to either gene alone (Figure 4c, lane 10). Importantly, single knockdown of either Cdc25C or PP2Acα resulted in P53 activation, thereby mimicking the findings in del(5q) MDS (Figure 4c), suggesting that Cdc25C and PP2Acα haplodeficiency may contribute to p53 activation in del(5q) MDS. These findings are consistent with our previous observation that PP2A directly associates with the MDM2 complex (Figure 3e), whereas Cdc25C does not (data not shown). Together, these data confirm and provide a functional rationale for a critical role of reduced PP2Acα gene dosage in cellular susceptibility to lenalidomide stabilization of MDM2. We previously reported that lenalidomide directly inhibits Cdc25C phosphatase activity, but indirectly inhibits PP2A phosphatase activity (19). To exclude the possibility that lenalidomide alters PP2Acα gene expression, Q-PCR and immunoblot analysis were performed in both Namalwa and U937 cells treated with or without lenalidomide. These studies show that lenalidomide has no direct effect on PP2Aca expression at both the mRNA and protein levels in either Namalwa or U937 cells (Figure 4 d, e and f). Furthermore, lenalidomide stabilization of MDM2 and degradation of P53 in Namalwa cells was abrogated by specific suppression of MDM2 by treatment with MDM2 siRNA, indicating that MDM2 is indispensible for lenalidomide’s action to overcome p53 activation in del(5q) cells (Figure 4 g).

Secondary resistance to lenalidomide in del(5q) MDS is associated with PP2Acα and p53 up regulation

Although treatment with lenalidomide is effective in the majority of patients with del(5q) MDS, more than half develop drug resistance within three years (3, 17). To investigate mechanisms underlying acquired resistance to lenalidomide in del(5q) MDS patients, we studied sequential bone marrow specimens obtained prior to treatment, upon achievement of transfusion independence and cytogenetic response, and at the time of treatment failure in 22 patients with low/intermediate-1 risk disease. Among these, 11 patients achieved transfusion independence with lenalidomide treatment and subsequently experienced recurrence of disease despite continuation of lenalidomide treatment (i.e., secondary or acquired resistance). One patient had primary resistance to lenalidomide and 10 patients continue to maintain transfusion independence (median duration 22.5 months; range, 7.5–68 months). Immunohistochemical staining for p53, PP2Acα and Cdc25C protein was performed on bone marrow biopsy sections, and results compared to bone marrow biopsies from 6 age-matched lower risk non-del(5q) MDS patients and 6 age-matched normal controls. Immunohistochemical staining for p53 was almost exclusively restricted to erythroid precursors and was significantly higher in del(5q) MDS specimens compared to normal controls (P=0.002) and non-del(5q) MDS (P=0.016) specimens (Figure 5a). Mean cellular expression of p53 was significantly decreased at the time of hematologic and cytogenetic response when compared to normal controls (P=0.04), but increased more than threefold at the time of treatment failure (P=0.003) (Figure 5b). Mean cellular expression of PP2Acα protein declined at the time of response to lenalidomide treatment (P=0.091), but significantly increased at the time of treatment failure (P=0.003) (Supplementary Figure S1). Quantitative-PCR analysis showed that the latter change in protein expression was associated with a greater than threefold increase in PP2Acα gene mRNA at treatment failure (Supplementary Figure S1, lower right panel). These findings confirm our laboratory findings that PP2Acα gene expression level is a critical determinant of lenalidomide’s capacity to promote escape from p53 arrest in del(5q) MDS erythroid precursors.

Figure 5. Secondary resistance to lenalidomide in del(5q) MDS is associated with p53 and PP2Acα over-expression.

(a) Immunohistochemical comparison of p53 expression in bone marrow from healthy donors and patients with del(5q) (left panel). The mean relative expression of p53 compared between del(5q) (n=18), controls (n=6) and non-del(5q) bone marrow biopsies (n=6) in right panel by bar graph; differences were compared using paired t-test. (b) Immunohistochemical comparison of p53 expression in bone marrow cells isolated at the time of response to lenalidomide and time of lenalidomide treatment failure. The mean relative expression of p53 is compared in del(5q) patients at baseline, time of response and time of treatment failure (paired t test is used for statistical analysis) shown in right panel by bar graph.

We next investigated the relationship between bone marrow changes in PP2Acα expression and duration of transfusion independence (TI). The magnitude of reduction in PP2Acα expression as measured by the difference between response and baseline RE (relative expression), was significantly associated with duration of transfusion independence. Median duration of TI was not reached (mean, 1507+days) in patients with a reduction in PP2Acα relative expression (n=9) compared to a median TI duration of 679 days (mean 658 days) in patients with no reduction in PP2A relative expression at time of response vs. baseline (n=4) (P=0.006, log rank). There was no significant correlation between reduction in Cdc25C or p53 and duration of TI.

The magnitude of reduction in PP2Acα expression as measured by the difference between response and baseline RE, was significantly associated with a longer duration of transfusion independence (P=0.021; hazard ratio =0.95). These data support our laboratory findings that over-expression of the PP2Acα gene suppressed MDM2 induction and restored p53 stabilization upon lenalidomide treatment in Namalwa cells (Figure 3f), indicating that expression level of PP2Acα is a key determinant of drug induced p53 degradation, and that up-regulation of this key haplodeficient drug target promotes resistance to lenalidomide in del(5q) progenitors.

Cdc25C is over-expressed in BM-MNCs from lenalidomide resistant MDS

Although reduced Cdc25C gene dosage may not singularly influence lenalidomide stabilization of MDM2, Cdc25C was up-regulated in bone marrow cells in patients with secondary drug resistance. We found that changes in Cdc25C sub-cellular localization after lenalidomide treatment support the emergence of drug resistance. Immunostaining showed a mixture of both nuclear and cytoplasmic distribution prior to treatment, whereas at the time of treatment response, protein expression was significantly diminished (P<0.05) and was almost exclusively restricted to the cytoplasm, consistent with drug-induced nuclear exclusion (Figure 6a upper panel). However, Cdc25C expression was significantly increased at the time of treatment failure (P< 0.001) and was largely limited to a nuclear distribution, findings consistent with escape from drug inhibition (Figure 6a upper panel). Q-PCR showed approximately a twelve-fold increase in gene message at the time of treatment failure, consistent with transcriptional up-regulation (Figure 6a lower right panel). Consistent with our previous results, intracellular tracking of the Cdc25C phosphatase is regulated by cytoplasmic binding to 14-3-3 proteins (19), whereas nuclear retention of the phosphatase is facilitated by a nuclear localization sequence (NLS) and nuclear export signal (NES).

Figure 6. Secondary resistance to lenalidomide in del(5q) MDS is associated with Cdc25C over-expression.

(a) Immunohistochemical comparison of Cdc25C expression in bone marrow cells isolated prior to treatment (baseline), time of response to lenalidomide and time of treatment failure (upper panel). The relative expression of Cdc25C is compared using paired t test at designated time points shown in lower left panel by bar graph. Increased expression of Cdc25C in bone marrow cells from patients at treatment failure is shown in the lower right panel. RNA was purified from BM-MNC from patients prior to treatment and at the time of lenalidomide treatment failure. Expression of Cdc25C was analyzed by Q-PCR to quantitate transcript levels and normalized to GAPDH. (b) TP53 and PP2Acα DNA copy number quantification of BM-MNC from patients prior to treatment and at the time of treatment failure. Genomic DNA was prepared from BM-MNCs isolated from three patients as described in the Methods section. Commercial DNA plasmids (IDT, Inc.) were used as the templates for the standard curve. All reactions were performed in triplicate.

To determine if secondary resistance to lenalidomide arises from acquired gene mutations in p53 or target phosphatases, bone marrow DNA obtained at the time of treatment failure was analyzed by Sanger sequencing. Among the 5 specimens analyzed, we detected no somatic mutations within the TP53 DNA binding domain (exons 4–9) or the Cdc25C nuclear export signal domains (exon 11) (data not shown). To exclude gene amplification of TP53 and PP2Acα with emergence of drug resistance, we compared relative DNA sequence copy number of both genes in 3 sequential bone marrow specimens from lenalidomide treated patients. We utilized quantitative microsatellite analysis (QuMA) for rapid measurement of relative DNA sequence copy number of a test locus for TP53 and PP2Acα relative to a pooled reference and assessed using Q-PCR amplification of loci carrying simple sequence repeats (34). DNA copy number of both TP53 and PP2Acα decreased at the time of treatment failure compared with pretreatment specimens, indicating that amplification of p53 and PP2Acα DNA copy number is not involved in lenalidomide induced drug resistance in del(5q) MDS (Figure 6b). The decrease in PP2Acα copy number is consistent with clonal expansion at treatment failure, i.e., increasing from a mixture of normal and del(5q) cells at baseline, to 100% del(5q) by metaphase karyotyping at treatment failure.

Discussion

The RP-MDM2-p53 pathway is a critical effector of the hypoplastic anemia in patients with del(5q) MDS and congenital anemias arising from RP gene mutations (13–15). Both pharmacological inhibition of p53 activity in del(5q) MDS progenitors and TP53 inactivation in the syngeneic murine model of the human 5q- syndrome are sufficient to rescue the hematologic phenotype, emphasizing the key role of p53 in the molecular pathogenesis of the syndrome (12, 16). We confirmed that p53 is over-expressed in a lineage restricted manner in erythroid precursors of primary human bone marrow del(5q) MDS specimens (12), and show that treatment with lenalidomide restores MDM2 stability to promote p53 degradation in both a cell line model and primary del(5q) MDS specimens, accompanied by suppression of downstream p53 effector genes.

MDM2, a RING finger E3-ubiquitin ligase and key negative regulator of p53, is indispensable for the rescue of primitive erythroid progenitors from p53-mediated apoptosis (21). MDM2^−/− null mice with homozygous wild type TP53 develop a progressive cytopenia arising from reduced proliferative potential of hematopoietic progenitors (21). Introduction of a single copy of MDM2 is sufficient to rescue the hematopoietic defect. Our finding that lenalidomide not only disrupts RPS14 association with MDM2, but also stabilizes MDM2 to promote p53 proteasomal degradation in del(5q) progenitors, suggests that strategies to induce or enhance MDM2 activity may restore effective erythropoiesis in disorders affecting integrity of ribosomal biogenesis.

Small ribosomal proteins such as RPL-5 and -11 are liberated upon ribosomal perturbation, bind to the central acidic domain of MDM2, and similar to ARF, trigger auto-ubiquitination and consequent stabilization of p53 (9, 11, 35). Inactivating mutations of the RPL5 and RPL11 genes, however, account for up to 20% of Diamond-Blackfan anemia genotypes, suggesting that other ribosomal proteins may compensate or compete for binding to MDM2 (36). Our studies provide the first evidence that perturbation of ribosome integrity by treatment with actinomycin-D promotes the binding of RPS14 to MDM2 that is blocked by lenalidomide exposure, whereas binding of RPS19, or RPL5 and RPL11 is unaffected. Whether RPS14 binding to MDM2 occurs in del(5q) MDS or triggers MDM2 autologous degradation is unknown, and awaits further investigation.

MDM2 ubiquitinates p53 and modifies its own stability through auto-ubiquitination. This balance is controlled by post-translational modifications such as phosphorylation of p53 on inhibitory NH₂ terminal serine and threonine residues, and multiple sites on MDM2 that suppress its auto-ubiquitination and block interaction with its negative regulator ARF (37–40). PP2A, whose catalytic domain, PP2Acα, is encoded within the proximal del(5q) CDR, dephosphorylates these regulatory sites to uncouple MDM2-p53 association and restore p53 activation (23, 24, 27, 28, 30, 41). We found that shRNA suppression of PP2Acα in the non-del(5q) U937 cells to levels commensurate with haplodeficiency was necessary for lenalidomide inhibition of PP2A phosphatase activity and MDM2 stabilization with consequent p53 degradation, indicating that gene dosage of PP2Acα is a critical determinant of drug sensitivity (Figure 4c). Moreover, siRNA suppression of MDM2 completely abrogated drug induced suppression of p53, indicating that MDM2 is indispensible for lenalidomide’s effect on p53 dynamics in del(5q) cells (Figure 4g). Our investigations show that treatment of del(5q) progenitors with lenalidomide resulted in concentration-dependent hyperphosphorylation of p53 at thr55 and -ser46 (Figure 3a), and corresponding residues on MDM2 (Figure 3), which abolished MDM2 auto-ubiquitination (Figure 4). Of interest, hyperphosphorylation of serine residues −166 and −186 on MDM2 upon lenalidomide exposure promoted nuclear translocation of MDM2 and p53 degradation. These specific serine residues are phosphorylated by PKB/Akt, whose kinase activity is suppressed by PP2A dephosphorylation of Thr308 and Ser473 (29); (42, 43). Indeed, lenalidomide treatment promoted retention of phospho-Thr308 on Akt as well as Ser9 on its downstream target, GSK3β (Figure 3g), indicating that lenalidomide activates Akt signaling in del(5q) progenitors to reinforce MDM2 phosphorylation. Phosphorylation of GSK3β deactivates the protein kinase and thereby complements the effects of MDM2 stabilization by promoting cell cycle progression (44).

Over-expression of PP2Acα, in contrast, antagonized lenalidomide’s action, preventing induction of MDM2 and restoring p53 accumulation (Figure 3f). These findings are in agreement with our previous studies showing that lenalidomide is a relatively weak indirect inhibitor of PP2Acα, and indicate that over-expression of this key allelic deficient drug target is alone sufficient to promote resistance to lenalidomide in del(5q) progenitors. Given the weak effects in normal diploid progenitors, combined treatment with other agents that inhibit p53 transcriptional activity such as corticosteroids, may augment its effects, and perhaps merits consideration in ribosomopathies.

Our studies of primary bone marrow specimens from del(5q) MDS patients treated with lenalidomide confirm its effects on p53 dynamics in vivo, and for the first time show that transcriptional up-regulation of PP2Acα may underlie acquired drug resistance. Immunodetection of p53 was almost exclusively restricted to erythroid precursors, and was significantly higher in del(5q) MDS specimens compared to normal controls (P=0.002) or non-del(5q) MDS (P=0.016) (Figure 5), consistent with perturbation of ribosomal biogenesis (5, 16). Relative expression of PP2Acα and p53 coordinately decreased upon response to lenalidomide treatment to levels approaching that of normal controls (P=0.04). These findings provide a mechanistic rationale to explain drug induced escape from p53 arrest necessary for transition to G2/M, where del(5q) cells arrest as a consequence of lenalidomide inhibition of the haplodeficient phosphatases, PP2Acα and Cdc25C. Moreover, the magnitude of reduction in cellular PP2Acα was directly proportional to the duration of transfusion independence with lenalidomide treatment (P=0.006), whereas no similar relationship was found for Cdc25C or p53, confirming the importance of PP2Acα expression level as a determinant of response durability. Similarly, development of resistance to lenalidomide was associated with over-expression of PP2Acα (P=0.003), accompanied by restored p53 accumulation in erythroid precursors without evidence for acquisition of mutations within the TP53 DNA binding domain or gene amplification.

Using a more sensitive deep sequencing technique, Jädersten et. al. identified small clones harboring mutations in the DNA-binding domain of p53 in a subset of patients with del(5q) MDS, which was associated with an increased risk of disease progression with expansion of the mutant clone (45). Our findings that lenalidomide destabilizes p53 raises questions as to whether checkpoint abrogation with lenalidomide treatment might modify potential for expansion of p53 mutant clones or effect DNA repair in patients receiving concomitant treatment with DNA-damaging agents.

Although we previously identified Cdc25C as an alternate enzymatic target of lenalidomide in del(5q) MDS, isolated shRNA suppression of Cdc25C had no effect on MDM2 after lenalidomide exposure. Nonetheless, dual knockdown of Cdc25C and PP2Acα further enhanced drug induced MDM2 stabilization compared to suppression of PP2Acα alone, suggesting a cooperative role for Cdc25C deficiency. Of interest, single knockdown of either Cdc25C or PP2Acα activated P53 in the absence of discernable changes in MDM2 level, suggesting that Cdc25C and PP2Acα haplodeficiency may contribute to p53 transactivation in del(5q) MDS. Ito et. al. recently reported that the teratogenic effects of the lenalidomide analogue thalidomide, arises from its binding to and inhibition of the E3 ubiquitin ligase activity of cereblon (46). The precise mechanism of inhibition remains unclear, however, these findings raise the possibility of a shared class effect of thalidomide derivatives on these ubiquitous molecular targets. Nonetheless, our findings provide insight for development of novel therapeutic strategies for del(5q) MDS and congenital ribosomopathies. Given the broad functions of PP2A in cell cycle surveillance and cell fate decisions, PP2A-selective inhibitors may have limited utility and present potentially excessive risk. Nevertheless, small molecules interfering with free RP binding to the MDM2 central acidic domain could prove highly effective and specific, and perhaps provide an alternate strategy to overcome acquired resistance to lenalidomide in patients with del(5q) MDS.

Materials and Methods

Cells and reagents

The Namalwa cell line, a Burkitt lymphoma cell line with chromosome 5q deletion (22), and U937, a non -del(5q) human leukemic monoblastic cell line were cultured in RPMI 1640 supplemented with 10% fetal bovine serum and penicillin-streptomycin. Lenalidomide was provided by Celgene Corporation (Summit, New Jersey) and was dissolved in dimethylsulfoxide (DMSO). Antibodies against human MDM2 (SPM14, Santa Cruz Biotechnology), MDM2 (2A10, Abcam), p53 (BD Pharmingen), p53 phospho-ser46 (Santa Cruz Biotechnology, Inc), p53 phospho-thr-55 (Santa Cruz Biotechnology, Inc.), HA tag (Roche), MDM2 phospho-ser166 (Cell Signaling), MDM2 phospho-ser 186 (Abcam), MDM2 phospho-serine, MDM2 phospho-threonine (Abcam), phospho-Akt (Thr308) and phosphor-GSK-3β (Ser9) (Cell Signaling) were used for Western blotting or immunoprecipitation. MDM2 (3G9) antibody and plasmid encoding wild type MDM2 and His-tagged ubiquitin were kindly provided by Dr. Jiandong Chen (Moffitt Cancer Center). Plasmid encoding PP2Acα was kindly provided by Dr. W. Stratford May (University of Florida Shands Cancer Center, Gainesville).

Patients and preparation of bone marrow specimens

Patients with del(5q) treated or untreated (n=30) were recruited from the Malignant Hematology clinic at the H. Lee Moffitt Cancer Center & Research Institute and the Taussig Cancer Center at Cleveland Clinic. MDS diagnoses and karyotype were confirmed by central review and classified in accordance with World Health Organization (WHO) criteria. After obtaining written informed consent, bone marrow mononuclear cells (BM-MNC) were isolated from heparinized bone marrow aspirates by Ficoll-Hypaque gradient centrifugation, as previously described (19).

Western blotting analysis

Cell lysates prepared from BM-MNCs from del(5q) MDS patients or from Namalwa cells that were treated with lenalidomide at concentration of 10 µM unless otherwise indicated for 24h, and 48h. DMSO was added as vehicle control. Cells were harvested and lysed at 4°C for 30 min in 1% NP-40, 10 mM Tris, 140 mM NaCl, 0.1 mM phenylmethylsulfonyl fluoride, 10 mM iodoacetamide, 50 mM NaF, 1 mM ethylenediamene tetra acetic acid, 1 mM sodium orthovanadate, 0.25% Na Deoxycholate, 100 µl protease inhibitor cocktail I and cocktail II (Sigma). Cell lysates were centrifuged at 12 000 g for 15 min to pellet cell debris and the supernatant containing protein lysate was collected. The protein concentration was determined using the Bio-Rad (Bradford) protein assay (Bio-Rad, Hercules, CA, USA). Separation of 30 µg of total protein was performed on 8.5% or 10% SDS–polyacrylamide gels, and transferred to a nitrocellulose membrane before Western blotting with the primary antibodies indicated. The specific proteins were detected by the enhanced chemiluminescent substrate (Pierce).

RNA Isolation and quantitative real-time PCR (Q-PCR)

Total RNA was purified from BM-MNCs isolated from del(5q) MDS patients using RNeasy Mini Kit according to the manufacturer’s instructions (Qiagen, Valencia,CA, USA). Reverse transcription (RT) reactions were performed using iScriptTM cDNA Synthesis kit (BIO-RAD, Hercules, CA, USA). cDNA was synthesized by adding 1 µg of total RNA, 4 µl of 5× iScript Reaction Mix, and 1 µl of iScript Reverse Transcriptase for a total volume of 20 µl. The reaction was incubated at 25°C for 5 min, 42°C for 30 min, and 85°C for 5 min. Oligonucleotide primers for amplifying TP53 were P53-F (5’- GTA CAT CTG GCC TTG AAA CC-3’) and p53-R (5’-AGC TGC CCA ACT GTA GAA AC-3’). Primers for amplifying MDM2 were MDM2-F (5'-GTC AAT CAG CAG GAA TCA TCG -3') and MDM2-R (5'- CCT TTT GAT CAC TCC CAC CTT -3'). Primers for amplifying PP2Acα -F(5'- TCT CAC TGC CTT GGT GGA T -3'’- and PP2Acα -R (5'- CCC TCA TGG GGA ACT TCT T -3'). Primers for amplifying GAPDH were GAPDH-F (5’-GAA GGT GAA GGT CGG AGT-3’), and GAPDH-R (5’-GAA GAT GGT GAT GGG ATT TC-3’). Q-PCR reactions were performed by means of iQ SYBR Green Supermix of Bio-Rad. Each reaction (25 µl) contained 12.5 µl of iQSYBR green supermix, 0.25 µl of forward primer (20µM), 0.25 µl of reverse primer (20µM), 11 µl of RNase-free water, and 1.0 µl of cDNA. The following cycles were performed 1×3 min at 95°C, 40 amplification cycles (15 s 95°C, 60 s 56°C), 1 × 1 min 95°C, 1 × 1 min 55°C and a melting curve (80 × 10 s 55°C with an increase of 0.5°C per 10 s). A negative control without cDNA template was run with every assay. The optimal melting point of dsDNA I and the efficiency of the reaction were optimized beforehand. Transcript copy number per subject was calculated by normalization to GAPDH expression.

Preparation of Cdc25C, PP2A, MDM2 RNAi and Lentivirus

Preparation of lentiviral vectors containing shRNA is as previously described (19). Briefly, Nucleotide sequences for short hairpin RNA (shRNA) were described and designed for ShRNA-Cdc25C as follows: sense strand, 5’-GAAGAGAATAATCATCGTGTT –3’ and antisense strand, 5’-GAAGAGAATAATCATCGTGTT-3’, for shRNA-PP2Acα were designed as follows: sense strand, 5’-TGGAACTT GACGATACTCTAA-3’, antisense strand, 5’-TGGAACTTGACGATACTCTAA-3’ A scrimbord RNAi sequence was used as a nonspecific control. Oligonucleotides were designed that incorporated these sequences within a short hairpin structure, using the stem loop sequence 5’- CTCGAG-3’, which were then cloned into lentiviral plasmids (pLKO.1-puro purchased from Sigma). Lentiviral particles were generated by transfection of lentiviral plasmids and packaging mix (purchased from Sigma) into HEK-293-T cells using lipofectamine 2000 reagent (Invitrogen). Supernatant containing viral particles were harvested between 36–72 hours. The supernatant was purified and used for Cdc25C and PP2A knocking down experiments. Non-target shRNA were used as negative control, and lentiviral vectors containing GFP were used to evaluate the infection rate. For lentiviral infection, 0.5 × 10⁶/ml of U937 cells were incubated with recombinant lentiviruses at MOI=1:5 in the presence of 8µg/ml of polybrene for 48 hours before treatment with lenalidomide.

For experiments using MDM2-siRNA, MDM2 siRNA and scrambled siRNA control were purchased from Santa Cruz Biotecnology. Namalwa cells were plated at 2×10⁵ cells /well in a 6-well plate overnight. Transfection was then performed by adding a mixture of 0.2 µM siRNA-MDM2 or siRNA-control and 10 µl of lipofectin-2000 into each well in RPMI medium-10%FBS. After 48h of incubation, Western blot analysis of cell lysates was performed using anti-MDM2 or anti-p53 antibodies.

Analysis of DNA copy number by Q-PCR

Genomic DNA was prepared using the QIAamp DNA mini kit (Qiagen,Inc,Valencia,CA, USA) from MDS patient BM specimens. Copy number estimation by Q-PCR was carried out in duplicate 25 µl reactions with 12.5 µl of iQSYBR green supermix, 0.25 µl of forward primer (20µM), 0.25µl of reverse primer (20 µM), 20 ng genomic DNA. PCR cycling condition was performed at 95°C for 15 sec, 56°C for 1min for 40 cycles. The primers used for Q-CPR were as follows: Cdc25C (NC_000005), forward, 5'- AGA GCA AGA CCC TGT CTC AA -3', reverse, 5'- TCT CAT CCT TCC TTC ACA GC -3'; PPP2cα (NC_000005), forward, 5'- ATT GCC CAG TCT TGT CTC G -3', and reverse 5'- TTC AGG CTG GGC ACT GTA T -3'; p53 (NC_000017), forward 5'- TGT CAT CTC TCC TCC CTG CT -3', and reverse, 5'- TCT GAG TCA GGC CCT TCT GT -3'. We used plasmids, which we designed gene products flanking forward and reverse primers for each gene and inserted into vectors pIDTSMART-KAN (IDT, Inc,), as the templates for the standard curve. A 10-fold serial dilution series of the plasmids ranging from 1×10 to 1×10⁷ copies/ µl was used to construct the standard curves for Cdc25C, PPP2cα and TP53. 2 µl of the plasmids were added to each well. The copy number was calculated using the following equation (47):

$$\text{DNA}(\text{copy})=\frac{6.02\times {10}^{23}(\text{copy}/\text{mol})\times \text{DNA amount}(\mathrm{g})}{\text{DNA length}(\mathrm{d}\mathrm{p})\times 660(\mathrm{g}/\text{mol}/\mathrm{d}\mathrm{p})}$$

C_T values in each dilution were measured in duplicate using Q-PCR to generate the standard curves for PPP2cα and p53, respectively. Absolute quantification determines the exact copy concentration of target gene by relating the C_T value to a standard curve.

Sequencing

Genomic DNA was extracted from paraffin-embedded blocks using RecoverAll™. Total Nucleic Acid Isolation Kit [Ambion, Applied Biosystems (ABI), Austin, TX] and was used for PCR amplification. Primers for TP53 exons 4–9 have been previously described (48). Cdc25C nuclear export signal primers are forward,5’-GGTGGACAGTGAAATGAAATATTTGGG-3’ and reverse, 5’-TTCTGGCTATGAGGGTTGCTGGAT-3’. DNA was amplified using iProof™ High Fidelity DNA Polymerase (Biorad, Hercules, CA). The PCR products were separated on 1.8% agarose with ethidium bromide and purified with Wizard® SV Gel and PCR Clean UP System (Promega, Madison, WI). PCR products were sequenced with BigDye® Terminator v3.1 Cycle Sequencing Kit (ABI) and a 3130×1 Genetic Analyzer (ABI).

Immunohistochemical staining

Antibodies to Cdc25C, PP2Acα and p53 were used for immunohistochemical staining. The rabbit anti-human antibody recognizing Cdc25C (Novus Biologicals, Littleton, CO) was used at a 1:125 concentration. The rabbit anti-human antibody recognizing PP2Acα (Cell Signaling, Danvers, MA) was used at a 1:25 concentration. The mouse monoclonal primary antibody for p53 (Ventana) was used at a prediluted strength according to the manufacturer’s instructions. Immunohistochemical staining was performed as follows: the paraffin embedded blocks from BM core biopsies were cut into 4-µm sections and deparaffinzed on the automated system with EZ prep solution (Ventana). Heat-induced antigen retrieval method was used in cell conditioning 1 (Ventana). Incubation with primary and secondary antibodies was based on preset conditions. Immunohistochemical stains were performed conditionally per manufacture’s instructions by Ventana Discovery XT automated system (Ventana Medical System, Tucson). The detection system used was the Ventana OmniMap Kit. The stained slides were dehydrated and cover slipped for a pathologist’s evaluation. Percentage and strength of staining (scored 0, 1+, 2+ 3+ or between) were documented in Excel spread sheet.

Immunofluorescence

Namalwa cells (3 ×10⁵/ml) were treated with lenalidomide (10uM) or DMSO for 48 hours, centrifuged onto microscope slides and fixed with methanol/acetone (3:1) in −20 °C for 30 min. An anti-MDM2 antibody (1:200 dilution) and secondary goat anti-mouse Ig AlexaFluor (594) (Sigma) was used to visualize the translocation of MDM2. Nuclei were stained with DAPI. Immunofluorescence was detected using a Leitz Orthoplan 2 microscope and images were captured by a charge-coupled device (CCD) camera with the Smart Capture Program (Vysis, Downers Grove, IL). On each slide, 100 cells were counted for MDM2 nuclear translocation. Nonspecific binding with secondary antibody alone was not detected (data not shown). Immunoflorescent images were analyzed using Image Pro Plus 6.2 (Mediacybernetics Inc., Maryland) as previously described (19), and the percentage of cells with increased nuclear signal was recorded.

In situ ubiquitination assay

Namalwa cells in 9-cm plates were transfected with 4 µg of His tagged 6-ubiquitin expression plasmid and 4 µg of MDM2 plasmid using LipofectAMINE 2000 (Invitrogen). Thirty-two hours after transfection, cells were cultured with 40 µM MG132 for 4hours. Cells from each plate were collected into two aliquots. One aliquot (10%) was used for conventional Western blotting to confirm expression and degradation of transfected proteins. The remaining cells (90%) were used for purification of His-tagged ubiquitin proteins by adding 30 µl Ni²⁺-nitrilotriacetic acid (NTA) beads (Qiagen) for each plate. The cell pellet was lysed in buffer A (6 M guanidinium-HCl, 0.1 M Na2HPO4/NaH2PO4, 0.01 M Tris-Cl [pH 8.0], 5 mM imidazole, 10 mM β-2-mercaptoethanol) and incubated with NTA beads for 4 hours at room temperature. The beads were washed with buffer A, buffer B (8 M urea, 0.1 M Na2PO4/NaH2PO4, 0.01 M Tris-Cl [pH 8.0], 10 mM β -mercaptoethanol), and buffer C (8 M urea, 0.1 M Na2PO4/NaH2PO4, 0.01 M Tris-Cl [pH 6.3], 10 mM β-2-mercaptoethanol), then, bound proteins were eluted with buffer D (200 mM imidazole, 0.15 M Tris-Cl [pH 6.7], 30% glycerol, 0.72 M β -mercaptoethanol, 5% SDS). The eluted proteins were analyzed by Western blotting for MDM2 (3G9) and p53 antibodies.

Statistical analysis

Relative expression (RE) of p53, Cdc25C, and PP2Acα was assessed using the product of the percentage of positive cells/high powered field (HPF) in the bone marrow specimens and staining intensity. The mean RE for p53 was compared between del(5q) patients, non-del(5q) specimens and normal controls using an independent t-test. The mean RE of each parameter was compared at baseline, time of response and time of failure using a paired t test. Kaplan-Meier estimate was used to calculate duration of transfusion independence in patients who achieved a reduction PP2Acα relative expression at time of response compared to those who did not; log rank test was used to compare the difference between the two groups.