Lenalidomide treatment promotes CD154 expression on CLL cells and enhances production of antibodies by normal B cells through a PI3-kinase–dependent pathway

Chronic lymphocytic leukemia (CLL) is the most common adult leukemia, and is characterized by an elevated frequency of infections, secondary malignancy, and autoimmune complications compared with the general population. Current treatment options for CLL are palliative and further exacerbate the immune deficiency seen in this disease. Nonetheless, CLL represents an “immunoresponsive” disease as evidenced by extended disease remission and potential cure with reduced intensity allogeneic stem cell transplantation (reviewed in Gribben¹ ). This suggests that strategies that restore immune function have potential to effectively eliminate CLL.

The immune defect in CLL is characterized by both humoral and cellular immune defects. Although detailed studies of normal B cells in CLL patients have not been performed due to the difficulty in isolating these cells, hypogammaglobulinemia is often present at diagnosis and becomes worse with disease progression. A profound cellular immune defect^2-4  is present in CLL with significant alterations in genes involved in differentiation, cytoskeleton formation, vesicle trafficking, and cell death.⁴  Coculture of CLL cells with normal T cells produces the same T-cell defects observed in CLL patients,⁴  suggesting a direct role of the leukemia cells in contributing to the T cell–dependent cellular immune deficiency. The clinical manifestation of the humoral and cellular immune defects in CLL patients includes hypogammaglobulinemia,^5,6  poor response to both polysaccharide-based^7-9  and protein-based¹⁰  vaccines, and a high predisposition to infections^11,12  that represents a leading cause of death.

To date, attempts to reverse the immune defects in CLL have been limited. Most promising has been adenovirus-delivered CD154 gene therapy that in small numbers of patients reversed cellular and humoral tumor tolerance. CD154 is the surface ligand of CD40 and is expressed on activated T cells, natural killer cells, and dendritic cells, but not normal B cells. Activation of T cells promotes increased surface expression of CD154, thereby promoting both activation and antigen presentation in normal and transformed B cells. Congenital mutations in the CD154 gene promote profound cellular and humoral immune deficiency. Although mutations of the CD154 gene have not been described in CLL, these patients have diminished CD154 expression on T cells after CD3 ligation.¹³  Transduction of murine or human CD154 into primary CLL cells ex vivo with adenovirus gene therapy vectors, followed by systemic reintroduction, has been pursued clinically.¹⁴  Surface expression of CD154 on CLL cells after gene therapy treatment promotes expression of costimulatory molecules including CD40, CD80, and CD86 on neighboring bystander CLL cells, thereby making them better costimulants for T-cell activation. As a consequence, increases in interferon-gamma, interleukin-12, and total CD4 T-cell counts were observed after CD154 gene therapy.¹⁴  Response to CD154 gene therapy by residual normal B cells was also demonstrated by improvement in both hypogammaglobulinemia and development of antibodies to the CLL tumor-specific antigen ROR1.¹⁵  Extending from immune activation, a favorable activation phenotype in the CLL cells occurs after CD154 gene therapy or CD40 ligation. This phenotype includes up-regulation of BID, DR5, and p73, thereby enhancing sensitivity of these tumor cells to both tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) and fludarabine-based therapies.^16,17  The CD154 gene therapy approach for CLL represents an exciting proof of concept to reverse the disease-induced immune defect. However, as with other gene therapy approaches, CD154 gene therapy is inefficient, cumbersome, and has not produced durable remissions, perhaps due to the inability to administer therapy over an extended period of time. Identifying an alternative pharmacologic strategy that mimics this approach would represent a major therapeutic advance for CLL and other related lymphoproliferative disorders.

Lenalidomide is an oral therapeutic agent with clinical activity in multiple myeloma,¹⁸  lymphoma, myelodysplasia,¹⁹  and CLL.^20,21  The exact mechanism of action of lenalidomide is uncertain, although it has been reported to promote cellular and innate immune activation,^22-25  up-regulation of SPARC,²⁶  and interference with tumor cell microenvironment.²⁷  In addition, lenalidomide increases costimulatory molecules CD40, CD80, and CD86 on CLL cells.^28,29  Coculture of CLL cells and autologous T cells with lenalidomide reverses the T-cell immune synapse defect present in this disease.³⁰  Improvement in T-cell function with lenalidomide is dependent upon CLL cell interaction,³⁰  similar to CD154 gene therapy. The subsequent observation of robust in vivo production of polyclonal hypergammaglobulinemia with concurrent disease response in a CLL patient receiving lenalidomide prompted us to examine whether this agent promotes CD154 gene expression in CLL cells. Herein, we demonstrate that lenalidomide promotes expression of functional CD154 on CLL cells through 2 distinct mechanisms of transcriptional activation and posttranscriptional mRNA stabilization. Most importantly, extended duration lenalidomide treatment promotes a CD154 gene therapy phenotype with multiple downstream effects on CLL cells, suggesting novel applications for its clinical use in this disease.

Blood was obtained from patients with CLL as described by National Cancer Institute Working Group criteria³¹  under an Ohio State University Institutional Review Board–approved protocol with informed consent according to the Declaration of Helsinki. The case description and other samples obtained in vivo are derived from a phase 1 clinical trial administering lenalidomide to patients with relapsed and refractory CLL at doses ranging from 2.5 to 7.5 mg daily. All patients enrolled on this trial had received rituximab with exception of the index case. CLL cells were isolated by Ficoll centrifugation using Rosette-Sep reagent (StemCell Technologies) and maintained in culture as previously described.²⁹  Lenalidomide (Revlimid; Celgene) was obtained and extracted as previously described.²⁸  Pokeweed mitogen (PWM), 1,9-dideoxyforskolin, okadaic acid, and transcriptional inhibitor actinomycin D (ActD) were from Sigma-Aldrich.

Percentage and mean fluorescence intensity of CD154 and CD19 on CLL cells relative to a specific isotype control were determined by flow cytometry (EPICS-XL; Beckman-Coulter) and analyzed using WinMDI (Windows Multiple Document Interface for Flow Cytometry; The Scripps Research Institute). CD19 antibody was from BD Biosciences and CD154 was from Serotec. For acid washing of cells before analysis, cells were exposed to acidic phosphate-buffered saline (pH 4.1) for 3 minutes at room temperature (RT) and then washed twice with normal pH before exposure to antibodies.³² 

Immunoblot assays were performed with the multiple antigen detection immunoblotting method as described.²⁹  Antibodies for Akt, phospho-Akt, c-Rel, Brg-1, nuclear factor of activated T cells c1 (NFATc1), BID, p73, actin, and ROR1 were purchased commercially from Cell Signaling Technology and R&D Systems. Recombinant human ROR1 extracellular domain (ECD) protein, approximately 50 kDa, was expressed in human embryonic kidney (HEK 293F) cells (Invitrogen) and purified by ion exchange and size exclusion chromatography. CLL B cells were washed and cell surface proteins biotinylated using Sulfo-NHS-LC-Biotin (Pierce).³³  Lysates were immunoprecipitated overnight at 4°C using either 3 μg of rabbit anti-CD154 or isotype-matched antibody (Santa Cruz Biotechnology) conjugated to rabbit ExactaCruz immunoprecipitation matrix (Santa Cruz Biotechnology). Precipitated materials were boiled, separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), transferred to 0.2 μm nitrocellulose membranes (Bio-Rad Laboratories), exposed to horseradish peroxidase (HRP)–conjugated streptavidin (Pierce) for 2 hours, and then imaged using a chemiluminescence substrate (SuperSignal; Pierce). An indirect enzyme-linked immunoabsorbance assay (ELISA) protocol was used to detect the presence of anti-ROR1 immunoglobulin in human sera. Polystyrene microtiter plates were incubated overnight at 4°C with 50 μL of ROR1-ECD (at 1 μg/mL in borate buffer), washed 3 times with 0.1% Tween/borate, and incubated in blocking buffer (0.5% bovine serum albumin, 1% goat serum/borate buffer) for 1 hour at RT. Plates were washed 3 times with 0.1% Tween/borate and serial dilutions of human serum in blocking buffer were added to individual wells and incubated for 2 hours at RT. Plates were washed 6 times with 0.1% Tween/borate and incubated with HRP-conjugated goat anti–human immunoglobulin G (IgG; Southern Biotechnology) for 1 hour. Plates were washed 6 times and SureBlue TMB Substrate was added to develop antibodies labeled with HRP. Plates were analyzed using a luminometer. Samples were run in triplicates.

RNA was extracted from 10⁷ cells using TRIzol reagent (Invitrogen). cDNAs were prepared using a SuperScript First-Strand Synthesis System (Invitrogen) as previously described.²⁹  Real-time polymerase chain reaction (PCR) was performed using predesigned TaqMan Gene Expression Assays and ABI Prism 7700 sequence detection system (Applied Biosystems).

Chromatin immunoprecipitation (ChIP) was performed by the MagnaEZ-ChIP assay kit (Upstate Biotechnology) as previously described.³⁴  For immunoprecipitation, 5 μg of each antibody or negative control rabbit IgG was used. SYBR-Green semiquantitative PCR was performed as described.²⁹ 

The pGL3-basic CD154-luc reporter constructs (−1280) and the pGL3-basic empty vector have been previously described.³⁵  Transient transfections of CLL B cells were conducted using the Amaxa Nucleofector protocol (Lonza). Luciferase and renilla assays were performed according to the manufacturer's directions (Promega). Luciferase activity values were normalized to transfection efficiency monitored by the cotransfected renilla expression vector.

A previously described procedure was followed.³⁵  Effector cells included 5 × 10⁴ autologous T cells (AutoTxr) from the healthy target B-cell donor, lenalidomide-treated CLL B cells (CLLBxr), or vehicle control-treated CLLBxr. Effector cells were irradiated (20 Gy) and placed in culture with 5 × 10⁴ target, purified B cells from the healthy donor, in the absence or presence of PWM (5 μg/mL). Control samples included 5 × 10⁴ target, unirradiated B cells alone, with and without PWM, and 5 × 10⁴ irradiated CLL B cells, in the absence of target B cells. All the experimental conditions were assayed in triplicate. Supernatants were collected after 7 days of coculture and tested for immunoglobulin production by a standard enzyme-linked immunoabsorbance assay (ELISA) protocol. IgG and IgM standards were purchased from Thermo Fisher Scientific Inc, whereas anti-IgG and -IgM antibodies were from Southern Biotech.

CLL B cells (10⁷) were treated with lenalidomide 0.5μM or vehicle control for 0, 12, 24, and 48 hours. At the end of each time point, cells were treated with 10 μg/mL ActD for varied times. Calculation of mRNA half-lives was carried out by plotting data on a semilogarithmic graph and calculating the first-order decay slope using linear regression. Values of independent experiments were used to calculate the average half-lives (± SD). Alternatively, data of several experiments were plotted on a single graph to calculate the linear regression slope and 95% confidence limits.

Cell lysates were prepared using a kinase buffer (20mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 2mM dithiothreitol, 10mM NaF, 1mM sodium vanadate, and 10mM β-glycerophosphate containing 1% Nonidet P40) plus a protease inhibitors. Cell lysate (500 μg) was incubated with 1 μg of rabbit polyclonal antibody specific for inhibitor of NF-κB kinase subunit β (IKK-β) for 1 hour at 4°C and then with anti–rabbit-specific agarose beads (eBioscience) overnight. Beads were washed 5 times with kinase buffer and then incubated in 20 μL of kinase buffer containing 20μM adenosine triphosphate (ATP), 10 μCi (0.37 MBq) of [γ³²P]ATP, and 0.5 μg of glutathione S-transferase (GST)–IκBα (1-54) substrate at 30°C for 30 minutes. The reaction mixture underwent SDS-PAGE (10%). In the radioactivity assay, the gel was dried and phosphorylation of GST-IκBα was measured by autoradiography. To normalize kinase activity, the proteins were transferred to nitrocellulose membrane and probed by Western blotting with anti–phospho-IκBα, total IκBα, and total IKK-β (Santa Cruz Biotechnology). Protein level on the blots was detected by addition of chemiluminescent substrate (SuperSignal; Pierce) followed by quantification by ChemiDoc system with Quantify One software (Bio-Rad Laboratories).

Myelin basic protein was labeled with ³²P by cyclic adenosine monophosphate–dependent kinase using ³²P-ATP according to the protocol provided by New England Biolabs. The protein phosphatase activity of total cell lysate was determined by measuring the released free ³²P. Cell lysates were prepared using a low-detergent lysis buffer (1% Nonidet P-40, 10mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 150mM NaCl, 10% glycerol, 1mM phenylmethylsulphonyl fluoride, 5nM benzamidine, and 10 μg/mL leupeptin). PP2A was immunoprecipitated from the cell lysate using an anti-PP2A antibody (Upstate Biotechnology) and incubated with ³²P-myelin basic protein at 30°C for 10 minutes. The reaction was terminated by adding trichloroacetic acid. The mixture was centrifuged and the supernatant was analyzed on a multipurpose liquid scintillation counter (Beckman Coulter). For cell-free assay, purified PP2A, an ABC trimer, was purchased from Upstate Biotechnology and dissolved in 20mM 3-[N-morpholino]propanesulphonic acid (pH 7.4), 0.1M NaCl, 60mM 2-mercaptoethanol, 1mM MgCl₂, 1mM ethyleneglycoltetraacetic acid, 0.1mM MnCl₂, 1mM dithiothreitol, 10% glycerol, and 0.1 mg/mL serum albumin. Neither Nonidet P40 nor bovine serum albumin was used. In each reaction, 0.5 units of the enzyme were used.

All statistical analyses were performed by biostatisticians in the Center for Biostatistics at The Ohio State University. Linear mixed models were used for modeling treatment effects, including patient random effects. Hypothesis testing used contrasts that directly answered primary questions, for example, interaction contrasts were used to directly test the inhibitory or synergistic hypotheses found in the experiments that test mechanisms of action. Random effects associated with these contrasts were included in the random effects portion of the models to avoid biased low estimates of the standard errors of the contrasts. The Holm method was applied to adjust for multiplicity and control the overall family-wise type I error rate at α = .05. SAS/STAT software (Version 9.1.3; SAS Institute Inc) was used for all statistical analyses.

A 78-year-old woman was diagnosed with CLL in June 2007 and ultimately developed symptomatic disease in June 2008. Genetic studies showed IgV_H unmutated status together with a complex karyotype including del(17p13.1). A 12-week treatment of subcutaneous alemtuzumab was administered from June 2008 to August 2008. Progressive disease with symptomatic lymphadenopathy was noted 2 months later. A pretreatment bone marrow biopsy had 25% to 30% CLL cells with no plasma cells noted. Lenalidomide was administered in November 2008 as part of an institutional review board–approved phase 1 clinical trial. The lenalidomide dose was escalated from 2.5 to 7.5 mg daily during the first month of therapy. Asymptomatic neutropenia developed during cycle 2 and a dose of 5 mg of lenalidomide daily has been maintained continuously since this time. Over 5 months, a slow reduction in blood lymphocyte count (Figure 1A) and nodal disease was noted. Concomitant with this, a polyclonal IgM, IgA, and IgG expansion was observed (Figure 1B). A bone marrow biopsy at 5 months demonstrated persistent but diminished CLL infiltration (10%) with 5% to 10% polyclonal plasma cells. The patient remains on therapy with continued improvement in her CLL.

Normal B cells generally do not express CD154, whereas CLL cells have increased CD154 expression at an mRNA level^35,36  but rarely express this at the protein level. We hypothesized that lenalidomide was promoting the production of polyclonal IgG and IgM in part through induction of functional CD154 surface expression on CLL cells, similar to that previously reported with CD154 gene therapy.¹⁵  Lenalidomide ex vivo treatment for 48 hours induces an increase in surface CD154 expression (P < .001) that was noted in 22 (73%) of the 30 patients examined (Figure 2A). The increase in CD154 was both time dependent (n = 5, P = .006, Figure 2B), with the optimal time of CD154 surface expression being 48 hours, and dose dependent (n = 7, P = .008 linear trend for CD154 expression as the dosage increases; supplemental Figure 1, available on the Blood website; see the Supplemental Materials link at the top of the online article), with up-regulation of CD154 in some patients at lenalidomide concentrations as low as 10nM but maximal at approximately 0.5 to 1μM as measured by flow cytometry at the 48-hour time point. Figure 2C demonstrates a representative patient with increased surface CD154 expression on CLL cells. CD154 expression on CLL cells was not from a soluble source, as demonstrated by lack of antigen loss with acid washing (data not shown) and the presence of both the 39-kDa and 32-kDa transmembrane domains of CD154 from lenalidomide-treated cells (n = 5 independent experiments, 2 representative patients are shown), but not vehicle control-treated cells, after biotinylated cell- surface protein pull down (Figure 2D).

CD154 regulation in T cells has been shown to occur via increased as well as transcript stabilization (reviewed in Vavassori and Covey³⁷ ). In CLL cells, lenalidomide induced CD154 expression at the mRNA level (Figure 3A; P < .001 compared with vehicle control), and this effect was dose dependent (supplemental Figure 2). This enhanced mRNA expression was in part through lenalidomide-induced CD154 mRNA stabilization, as demonstrated by a significant reduction in degradation after actinomycin D treatment (Figure 3B; P < .001 compared with vehicle control). Although increased CD154 gene transcript levels with lenalidomide treatment could be due solely to enhanced mRNA stability, increased transcriptional activation could also contribute. We therefore investigated this using luciferase constructs driven by the CD154 promoter. Lenalidomide significantly increased CD154-driven luciferase activity in primary CLL cells compared with vehicle control (Figure 3C; P = .001). This effect was concomitant with enhanced nuclear expression of c-Rel and NFATc (Figure 3D), and enhanced promoter binding of c-Rel (P < .001) and NFATc1 (P < .001) to the CD154 gene promoter as identified by chromatin immunoprecipitation assay (Figure 3E).

CD154 gene expression on CLL cells has been shown to reverse the T-cell defect and also result in the increased production of immunoglobulin by residual normal B cells.^14,38  Given our case observation, we sought to determine whether lenalidomide-induced CD154 expression could cause normal B cells to generate IgM and IgG. To do this, normal B cells were stimulated with pokeweed mitogen and cocultured with CLL cells pretreated with lenalidomide (48 hours) followed by irradiation (Figure 4A-B). Under these conditions, B cells significantly increased their production of IgM (n = 5, P = .006) and IgG (n = 11, P = .001) relative to B cells cocultured with CLL cells pretreated with vehicle only. Furthermore, significant (n = 7, P = .014) normal B-cell proliferation was also observed in this experiment (supplemental Figure 3). These results were consistent with results from autologous T cells, a known positive costimulatory control.³⁵ 

To confirm that increased antibody production by normal B cells was induced by the expression of CD154 on CLL cells, we first determined whether there was a correlation of CD154 induction on CLL cells with production of IgG and IgM. A significant correlation of increase in CD154 expression with lenalidomide and rise in IgM synthesis was observed (Pearson correlation coefficient; r = 0.94, P = .036). A similar finding correlating increase in CD154 expression on CLL cells and rise in IgG synthesis was observed (r = 0.93, P = .001). To provide additional evidence that the interaction between CD154 (on lenalidomide-treated CLL cells) and CD40 (on normal B cells) has a fundamental role in stimulating normal B cells to secrete immunoglobulin, we repeated these costimulatory studies including a CD40 blocking antibody. Treatment of normal B cells with a blocking CD40 antibody dramatically suppressed the production of IgM by these cells in the presence of lenalidomide-treated CLL cells (Figure 4C, n = 6, P = .001). Collectively, these data demonstrate that the lenalidomide-induced surface expression of CD154 on CLL cells is functional and can promote immunoglobulin production by normal B cells. In addition, the increased IgG production also suggests CD154-mediated class switching is occurring in normal pokeweed mitogen–stimulated B cells after coculture with lenalidomide-treated CLL cells or normal T cells.

The phosphoinositide-3 (PI3)–kinase/AKT axis is a common proximal signaling pathway by which nuclear factor κB (NF-κB) is activated in normal B cells.³⁹ Figure 5A demonstrates that lenalidomide treatment of CLL cells promoted AKT phosphorylation. This kinase also phosphorylates and activates the IKK complex.⁴⁰  In Figure 5B and C, we demonstrate such activation of IKK, together with increased phosphorylation of the NF-κB inhibitor IκB and its degradation after treatment with lenalidomide. Consistent with this, pretreatment of CLL cells with PI3-kinase signaling pathway inhibitors prior to lenalidomide treatment prevented lenalidomide-induced CD154 mRNA transcription (Figure 5D). Furthermore, pretreatment of CLL cells with PI3-kinase pathway inhibitors also prevented lenalidomide-induced CD154 mRNA stabilization (Figure 5D) and abrogates IgM production of normal B cells cocultured with lenalidomide-treated CLL cells (Figure 5E). Collectively, these data suggest that lenalidomide-mediated activation of the PI3-kinase pathway is essential for downstream NF-κB activation, CD154 mRNA stability, and CLL cell costimulatory activity.

In addition to promoting antibody production on normal B cells, CD154 gene therapy results in the increased expression of BID and DR5 concomitant with sensitization to TRAIL.¹⁷  In addition, ex vivo CD40 ligation has been shown to activate p73 and enhance fludarabine-mediated killing even in the absence of functional p53.¹⁶  Lenalidomide treatment of CLL cells demonstrates similar increases in protein expression of BID, DR5, and p73 (Figure 6A-B). As with preclinical models of CD154 gene therapy,^16,17  lenalidomide pretreatment of CLL cells for 5 days enhanced TRAIL- and fludarabine-mediated killing of CLL cells (P < .001 and P = .027, respectively; data not shown). The relevance of this observation in patients was indicated by serial samples from representative CLL patients receiving lenalidomide, in which we observed similar increases in CD154 mRNA stabilization (Figure 6C, n = 5, P = .045 compared with untreated control). BID and p73 induction was also noted in 6 of 9 patients examined (Figure 6D). In addition, our case report patient (Figure 1), the only subject who did not receive rituximab B cell–depleting therapy prior to lenalidomide, had evidence of development of a measurable ROR1 antibody (Figure 6E) after lenalidomide treatment for 6 months. This finding was also confirmed with a ROR1-specific ELISA (data not shown). Together, these findings indicate that lenalidomide reproduces the CLL phenotype of CD154 gene therapy in CLL and can promote an in vivo tumor humoral response to tumor cells.

Several decades of laboratory research have characterized the profound immune defect that exists in CLL. Investigators have tried a variety of therapeutic approaches to reverse this defect,^41-45  the most successful of which has been the introduction of CD154 to CLL cells via gene therapy. Although CD154 gene therapy has generated promising preliminary results, this therapeutic approach is cumbersome, expensive, and not adaptable to extended administration. Herein, we demonstrate the development of polyclonal hypergammaglobulinemia with increasing titers of ROR1 tumor-specific antibodies in a CLL patient receiving lenalidomide, providing in vivo human data that this agent can reverse humoral tumor tolerance. We then show that lenalidomide promotes up-regulation of CD154 antigen on the surface of CLL cells at concentrations readily attainable in the clinic. The increase in CD154 antigen expression on CLL cells with lenalidomide treatment leads to biologic responses similar to those observed with CD154 gene therapy, including increased IgM and IgG production by normal B cells and up-regulation of BIM, DR5, and p73. Consistent with these events, lenalidomide enhances TRAIL- and fludarabine-mediated cytotoxicity. We also show that the CD154 gene therapy phenotype, including increased CD154 transcription and mRNA stabilization, increased immunoglobulin production, and anti-ROR1 antibodies, can be observed in CLL patients receiving lenalidomide. The relevance of these results to the therapy of CLL and related diseases is significant. However, to fully appreciate the benefit of this application, new strategies of lenalidomide administration and combination strategies that build upon previously reported observations with CD154 gene therapy-based approaches^14,38  will be required.

Our studies demonstrate CD154 induction is functional in CLL cells, and also detail 2 distinct convergent mechanisms by which lenalidomide increases CD154 gene expression on CLL cells. Lenalidomide enhances mRNA gene transcription with increased binding of the activating c-REL/NFATc1 complex to the CD154 promoter, and also enhances CD154 promoter activity in primary CLL cells similar to that observed in lymphoma.⁴⁶  In addition, lenalidomide also enhances stabilization of the mRNA transcript as observed in T cells.³⁷  Although CD154 mRNA stabilization is known to lead to increased protein expression in activated T cells,³⁷  this effect has not been previously reported in normal or transformed B cells. In T cells, mRNA transcript stabilization involves a variety of proteins including HnRNP, polypyrimidine tract–binding protein, and HuR. The time course of CD154 surface expression and loss at 72 hours seen with lenalidomide is similar to what has previously been observed with T cells.⁴⁷  Studies examining the ability of lenalidomide or related drugs to stabilize RNA are limited, although in colon cancer cell lines, thalidomide has been shown to diminish COX2 mRNA stability via shuttling of HuR.⁴⁸  Further study of the mechanism of CD154 mRNA stabilization by lenalidomide in CLL and the kinetics of this process warrant further study that is ongoing.

Notable to these 2 mechanisms of CD154 up-regulation is the dependence on the PI3-kinase signaling pathway. Our data demonstrate that lenalidomide enhances PI3-kinase pathway activity as measured by increased phosphorylation of AKT and also downstream IKK. Disruption of PI3-kinase signaling with biochemical inhibitors prevents up-regulation of CD154 through both the transcriptional and the mRNA stabilization mechanisms. PI3-kinase inhibition also prevents B-cell costimulatory activity and subsequent increased production of IgM and IgG by normal B cells. How lenalidomide activates the PI3-kinase/AKT pathway in transformed CLL cells is of great interest. In transformed myeloid cells, lenalidomide indirectly inhibits the proapoptotic phosphatase PP2A, which can prevent dephosphorylation of AKT.⁴⁹  We demonstrate that lenalidomide has similar PP2A inhibitory activity in primary CLL cells, as measured by direct phosphatase activity assays in CLL cells (supplemental Figure 4A) and results in increased AKT phosphorylation (supplemental Figure 4B). The PP2A activating agent, 1,9-dideoxyforskolin, when coincubated with lenalidomide, antagonizes lenalidomide-induced CD154 mRNA and protein (supplemental Figure 4C-D). In addition to PP2A, other mechanisms activating the PI3-kinase/AKT pathway may contribute to CD154 induction. Given the divergent source of activation of PI3-kinase pathway through catalytic domain isoforms, upstream targets directly influenced by lenalidomide might be identified. Studies elucidating the mechanism(s) by which the PI3-kinase pathway is activated by lenalidomide are ongoing.

Our data provide rationale for novel combination treatment strategies with lenalidomide using principles from the previous CD154 gene therapy experience in CLL. The observation that lenalidomide can reverse humoral tolerance to known tumor-specific antigens such as ROR1, which is selectively expressed in CLL and is under investigation as a target for therapeutic monoclonal antibodies,⁵⁰  is especially intriguing. We hypothesize that the reason our index case was the only one to see recovery of immunoglobulins and development of an anti-ROR1 antibody was due to all of the other subjects having received prior rituximab that depletes normal B cells. A preliminary study of lenalidomide in previously untreated CLL patients by the M. D. Anderson Cancer Center has demonstrated that an overall increase in immunoglobulin levels is observed with extended treatment.⁵¹  This finding is supportive of our hypothesis that normal B cells are required to observe this effect. Prospective studies evaluating the relationship of induction of CD154 on CLL cells by lenalidomide in previously untreated patients with subsequent recovery of immunoglobulin levels after lenalidomide treatment are warranted. Given that many CLL therapies deplete normal B cells, application of lenalidomide early in the course of the disease in combination with targeted therapies that spare normal immune cells may be warranted to optimize lenalidomide-induced tumor humoral response. This will be particularly true with anti-CD20 antibodies that deplete both normal and tumor B cells if the humoral B-cell contribution to lenalidomide's tumor mechanism of action in CLL. In addition, these data point to the potential role of this agent to reverse the hypogammaglobulinemia commonly present in CLL that contributes significantly to morbidity from infection. Given the broad activity that lenalidomide has shown across many different B-cell malignancies, it will be vital to explore this agent's ability to reverse humoral tumor tolerance in other diseases and also to enhance vaccine-based approaches. Indeed, given the diverse mechanisms of action of lenalidomide in CLL and other diseases, clinical trials incorporating detailed laboratory correlative studies will be required to attain the full potential of this exciting agent.

This work was supported by the Cancer and Leukemia Group B Foundation, D. Warren Brown Foundation, Specialized Center of Research from the Leukemia & Lymphoma Society, and P50-CA140158, 1K12 CA133250, P01 CA95426, and P01 CA101956 from the National Cancer Institute.

L.A. and A.J.J. are Paul Calabresi Scholars. R.B. is a Howard Hughes Medical Institute Research Training Fellow.

National Institutes of Health

Howard Hughes Funding

Contribution: R.L. planned the research, performed experiments, analyzed data, drafted the first and subsequent drafts of the paper, and approved the final version of the paper; L.A. planned the research, led the CLL clinical trial, reviewed drafts of the paper, and approved the final version of the paper; Q.L., S.E.M., R.B., C.A.R., L.L.S., A.R., and N.M. were involved in planning components of the research, performed experiments, reviewed drafts, and approved the final version of the paper; K.A.B. planned the research, accrued patients to the CLL clinical trial, reviewed drafts of the paper, and approved the final version of the paper; W.B. wrote the clinical trial, planned the research, reviewed drafts of the paper, and approved the final version of the paper; A.L., X.M., and D.J. were involved in planning components of the research, performed all the statistical analysis, reviewed drafts, and approved the final version of the paper; C.-S.C., R.F., L.V.P., and C.R. were involved in making necessary reagents essential to the hypothesis of this paper, reviewed drafts, and approved the final version of the paper; and A.J.J. and J.C.B. planned every aspect of the proposal, supervised the research, analyzed data, reviewed drafts, obtained funding for the research work, and approved the final version of the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Research funding was received from Celgene for the clinical trial described in this paper.

Correspondence: John C. Byrd, B302 Starling-Loving Hall, 320 W 10th Ave, Columbus, OH 43210; e-mail: john.byrd@osumc.edu; or Amy J. Johnson, OSUCCC Bldg, Rm 455C, 410 W 12th Ave, Columbus, OH 43210; e-mail: amy.johnson@osumc.edu.