Lenalidomide inhibits the proliferation of CLL cells via a cereblon/p21^WAF1/Cip1-dependent mechanism independent of functional p53

Lenalidomide is a second-generation immunomodulatory drug (IMiD)^1-3  that has both direct tumoricidal, as well as immunomodulatory activity in patients with multiple myeloma.⁴  This drug also has clinical activity in patients with chronic lymphocytic leukemia (CLL), even though it is not directly cytotoxic to CLL cells in vitro.^5,6  As such, its clinical activity in CLL is presumed to be secondary to its immune modulatory activity.⁷  Indeed, lenalidomide indirectly modulates CLL-cell survival in vitro by affecting supportive cells, such as nurse-like cells,⁸  found in the microenvironment of lymphoid tissues. Lenalidomide also can enhance T-cell proliferation¹  and interferon-γ production⁹  in response to CD3-crosslinking in vitro and dendritic-cell–mediated activation of T cells.¹⁰  Moreover, lenalidomide can reverse noted functional defects of T cells in patients with CLL.^11,12  Finally, lenalidomide can also induce CLL B cells to express higher levels of immunostimulatory molecules such as CD80, CD86, HLA-DR, CD95, and CD40 in vitro,^5,13  thereby potentially enhancing their capacity to engage T cells in cognate interactions that lead to immune activation in response to leukemia-associated antigen(s).¹⁴ 

However, lenalidomide may also have direct antiproliferative effects on CLL cells that account in part for its clinical activity in patients with this disease. This drug can inhibit proliferation of B-cell lymphoma lines¹⁵  and induce growth arrest and apoptosis of mantle-cell lymphoma cells.¹⁶  Although originally considered an accumulative disease of resting G_0/1 lymphocytes, CLL increasingly is being recognized as a lymphoproliferative disease that can have high rates of leukemia-cell turnover, resulting from robust leukemia cell proliferation that is offset by concomitant cell death. Indeed, CLL cells can undergo robust growth in so-called “proliferation centers” within lymphoid tissues, in response to signals received from accessory cells within the leukemia microenvironment. In vivo heavy-water labeling studies have demonstrated that some patients can have relatively high rates of leukemia-cell turnover, generating as much as 1% of their total leukemia-cell population each day, presumably in such tissue compartments.¹⁷  Inhibition of leukemia-cell proliferation could offset the balance between CLL-cell proliferation and cell death, resulting in reduction in tumor burden over time. Herein, we examined whether lenalidomide could inhibit the growth of CLL cells that are induced to proliferate, an effect that potentially could contribute to its noted clinical activity in patients with this disease.

Lenalidomide was provided by Celgene Corporation (San Diego, CA) and solubilized in dimethylsulfoxide (DMSO; Sigma, St. Louis, MO), which was used as a vehicle control in all experiments. Between 0.01 and 30 μM of lenalidomide was added every 3 days to long-term cultures, unless otherwise indicated.

Blood samples were collected from CLL patients at the University of California San Diego Moores Cancer Center who satisfied diagnostic and immunophenotypic criteria for common B-cell CLL, and who provided written, informed consent, in compliance with the Declaration of Helsinki¹⁸  and the Institutional Review Board of the University of California San Diego. Peripheral blood mononuclear cells were isolated by density centrifugation with Ficoll-Hypaque (Pharmacia, Uppsala, Sweden), resuspended in 90% fetal calf serum (FCS) (Omega Scientific, Tarzana, CA) and 10% DMSO for viable storage in liquid nitrogen. Alternatively, viably frozen CLL cells were purchased from AllCells (Emeryville, CA) or Conversant Biologics (Huntsville, AL). Samples with >95% CD19⁺CD5⁺ CLL cells were used without further purification throughout this study.

HeLa cells were obtained from the American Type Culture Collection (Manassas, VA). CD154-expressing HeLa cells (HeLa_CD154) were generated as described.¹⁹  Fibroblasts_CD154 were provided by Dr Ralph Steinman.²⁰  For experiments using HeLa_CD154 cells, CLL cells were plated at 1.5 × 10⁶ cells per well (per mL) in a 24-well tray on a layer of irradiated HeLa_CD154 (8000 Rad) cells at a CLL:HeLa_CD154 cell ratio of 15:1 in RPMI-1640 medium supplemented with 10% FCS, 10 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (Life Technologies, Grand Island, NY), penicillin (100 U/mL)-streptomycin (100 µg/mL) (Life Technologies), 5 ng/mL of recombinant human interleukin (IL)-4 (R&D Systems, Minneapolis, MN), and 15 ng/mL recombinant human IL-10 (PeproTech Inc, Rocky Hill, NJ). In earlier studies, we noted that expression of CD154 on the supportive cells, combined with exogenous IL-4 and IL-10, provided for optimal CLL-cell proliferation (see supplemental Figure 1A-B on the Blood Web site). These cells were also stained with fluorochrome-conjugated monoclonal antibodies specific for CD19, CD5, or ROR1 to confirm via flow cytometry that the proliferating cells were CLL cells (supplemental Figure 1C). For coculture on Fibroblasts_CD154, 0.8 to 1 × 106 CLL cells were plated on 6 × 10⁵ cells/well mitomycin-C–treated fibroblasts (10 µg/mL; 3 hours) in 1 mL per well in the media described above. For long-term cultures, half of the media was renewed every 3 days. For CpG stimulation, CLL cells were plated at 1.5 × 10⁶ cell/mL in RPMI-1640 supplemented with FCS, N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, penicillin, and streptomycin as above, to which we added 2.5 µg/mL CpG (InvivoGen, San Diego, CA), 10 ng/mL rhIL-2, and 15 ng/mL rhIL-10 (both from PeproTech Inc).

Descriptions of CLL-cell proliferation and viability measurements, flow cytometry, cell-cycle analysis, gene expression and TP53-mutation analysis, transfection, immunoblot and statistical analysis are provided in the supplemental “Methods.”

We induced CLL cells to undergo proliferation by coculturing them with accessory cells made to express CD154 in the presence of exogenous IL-4 and IL-10. CLL cells cocultured with CD154-expressing HeLa cells (Figure 1A-B) or human fibroblasts (Figure 1C-D) were induced to proliferate, as detected by the reduction in green fluorescence of dividing cells labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE). Moreover, the CD154-expressing cell lines supported CLL-cell proliferation through several rounds of division, as shown by the dilution of CFSE fluorescence over time (Figure 1A-D). Induction of CLL-cell proliferation was associated with changes in the distribution of cells in the different phases of the cell cycle, decreasing the numbers of cells in G₀/G₁ and concomitantly increasing the numbers of cells in S or G₂/M phases of the cell cycle, without affecting the fraction of cells in sub-G₁ (Figure 1E-H). Consistent with the induced cellular proliferation, we noted increased numbers of viable CLL cells over time in cultures with exogenous cytokines and CD154-expressing cells, but not in control cultures exposed to supportive cells not expressing CD154 (Figure 1I).

The capacity to induce CLL-cell proliferation allowed us to examine the impact of lenalidomide on dividing CLL cells in vitro. First, CLL cells were cocultured on CD154-expressing accessory cells in media containing IL-4 and IL-10 for 2 days. Subsequent to this, we added lenalidomide to the concentrations listed and cultured the cells for 7 days. We observed a dose-dependent inhibition of CLL-cell proliferation with lenalidomide (Figure 2A), which led to a significant decrease in the fraction of dividing CLL cells, starting at concentrations as low as 0.3 µM (Figure 2B). The inhibition of proliferation was confirmed by cell-cycle analysis using propidium iodide (PI) staining (Figure 2C). The fraction of cells in the G₀/G₁ phases of the cell cycle were significantly elevated in a dose-dependent fashion with increasing amounts of lenalidomide. The increases in the fraction of cells in the G₀/G₁ phase were accompanied by a significant decrease in the fraction of cells in the S or G₂/M phase in cells treated with lenalidomide, relative to cultures without the added drug. We also observed comparable cytostatic effects, measured via CFSE labeling, viable cell counts, 5-bromo-2′-deoxyuridine incorporation, or cell-cycle analysis, when lenalidomide was added at the initiation of the cocultures at drug concentrations ≥3 µM (supplemental Figure 2A-F). Using this regimen, we observed that lenalidomide could inhibit the induced proliferation of CLL cells of almost all patients examined (19 of 22), significantly inhibiting the induced increase in numbers of CLL cells noted after 6 days in coculture (supplemental Figure 2G). There were no distinctive characteristics that were shared by the 3 nonresponsive samples. On the other hand, we did not observe direct cytotoxic effects of lenalidomide on the proliferating CLL cells, as noted in other studies on resting cells,^5,6  even when the drug was added to the CLL cells at the initiation of coculture (supplemental Figure 2H) or after 2 days of coculture (Figure 2D). Of note, the level of CD154 expressed on accessory cells was not altered by treatment with lenalidomide (supplemental Figure 3A-B), suggesting that the effect of the drug was not due to an indirect effect on the accessory cells used to induce CLL-cell proliferation. Finally, lenalidomide also inhibited CLL-cell proliferation induced by coculture with accessory cells expressing CD154, in the absence of IL-4 and IL-10, suggesting that the inhibitory effects of lenalidomide on CLL-cell proliferation is not mediated through inhibition of the signaling induced by cytokines, such as IL-4 and/or IL-10 (supplemental Figure 3C). Consistent with this notion, lenalidomide also displayed a cytostatic effect on CLL cells stimulated to divide following treatment with CpG oligonucleotides in the presence of IL-2 and IL-10 (supplemental Figure 4), suggesting that the inhibition of proliferation was not restricted to inhibition of CD40-signaling.

We observed that CLL cells exposed to lenalidomide accumulate in the G₀/G₁ phase of the cell cycle when stimulated with CD154 and IL-4/IL-10. Prior studies on the B-cell line Namalwa showed that lenalidomide could upregulate the expression of p21,¹⁵  a protein that could inhibit the activity of cyclin-dependent kinases involved in G₁/S progression.^21,22  We, therefore, monitored the expression levels of p21 in CLL cells stimulated to divide and exposed to lenalidomide. We observed an upregulation of p21 messenger RNA levels after 24 hours of exposure to lenalidomide (Figure 3A). A dose-dependent increase of p21 protein was also observed, regardless of whether there was a detectable expression of p53 (Figure 3B-C), suggesting that p21 may be upregulated via a p53-independent mechanism. To test this hypothesis, we measured the effect of lenalidomide on CLL cells lacking functional p53, of which 99.5% had del(17p) on one allele and an inactivating mutation in exon 5 (at protein codon 174 [R→W]) in the retained allele. In contrast to CLL cells with wild-type p53, we observed that these CLL cells could not be induced to express higher levels of p53 or p21 following exposure to γ-irradiation (Figure 3D). However, such CLL cells could be induced to express p21 by treatment with lenalidomide, indicating that lenalidomide-induced upregulation of p21 in CLL cells does not require functional p53.

Next, we monitored the expression levels of p21 in a larger group of patient samples after treatment with 3 µM lenalidomide. We observed that lenalidomide induced upregulation of p21 in most samples tested (13 of 16) (Figure 3E-F). Of note, the 3 samples that failed to upregulate p21 expression were induced to proliferate in the coculture system, but were not sensitive to the inhibition mediated by 3 µM lenalidomide (2.6 ± 0.7 vs 2.4 ± 0.7-fold expansion from day 1 to day 6 in control-treated cells vs lenalidomide-treated cells, respectively). Furthermore, we observed that p21 was upregulated in a fashion that correlated with the inhibition of proliferation induced by lenalidomide (Figure 3G), suggesting that p21 is involved in the growth inhibitory effects of lenalidomide.

We used small interfering RNA (siRNA) technology to silence p21 in primary CLL cells and monitored the activity of lenalidomide on the transfected cells stimulated to divide by coculture with CD154-expressing cells and exogenous cytokines. We observed reduced expression of p21 following silencing (Figure 4A-B). Upon exposure to lenalidomide, the p21-silenced cells still showed an increase in p21 protein, but its levels were reduced compared with that of control-treated cells (Figure 4C-D). CLL cells transfected with siRNA specific for p21 were significantly less sensitive to the growth inhibitory effects of lenalidomide than CLL cells transfected with control siRNA (Figure 4E-F). These results indicate that expression of p21 contributes to the cytostatic activity of lenalidomide on CLL cells.

The only known direct molecular target of lenalidomide is CRBN,²³  that is part of the Cul4A-DDB1 E3 ligase complex. CRBN also plays a role in the cytostatic and/or cytotoxic activity of lenalidomide in multiple myeloma cells^23,24  and in the neoplastic B cells of patients with activated B-cell diffuse large B-cell lymphoma.²⁵  To evaluate the role of CRBN in the cytostatic activity of lenalidomide on CLL cells, we used siRNA technology to silence CRBN in primary CLL cells. We observed a significant reduction of CRBN protein levels in CLL cells 48 hours after transfection with CRBN siRNA, but not with control siRNA (Figure 5A-B). CLL cells transfected with CRBN siRNA expressed substantially lower amounts of p21 following treatment with lenalidomide than CLL cells transfected with control siRNA (Figure 5C-D), indicating that CRBN contributed to the lenalidomide-induced expression of p21. Furthermore, we observed that the CRBN-silenced cells were less susceptible to the cytostatic activity of lenalidomide than control-treated CLL cells (Figure 5E-F). These results indicate that the cytostatic activity of lenalidomide involves induction of p21 in CLL cells in a CRBN-dependent manner.

Lastly, we monitored for changes in the expression of Ikaros family zinc finger proteins 1 and 3 (IKZF1 and IKZF3), transcription factors found to be degraded in a CRBN-dependent manner and required for the activity of lenalidomide on multiple myeloma cells and T cells.^26-28  For this, we transfected CLL cells with either control siRNA or CRBN-specific siRNA, cocultured the transfected cells on CD154-expressing supportive cells in media without (control) or with various concentrations of lenalidomide for 24 hours, after which we prepared cell lysates which were examined via immunoblot analysis. We observed reduced expression levels of IKZF1 and IKZF3 in control CLL cells treated with lenalidomide at concentrations as low as 0.3 µM (Figure 5G-H). However, CLL cells silenced for CRBN failed to experience enhanced degradation IKZF1 or IKZF3 upon treatment with lenalidomide, even at concentrations as high as 30 µM (Figure 5G-H). These results suggest that changes in expression levels of IKZF1 and IKZF3 may be involved in the CRBN-dependent mechanism-of-action of lenalidomide on CLL cells.

To assess whether lenalidomide could induce p21 in CLL cells at concentrations that could be achieved in vivo, we examined CLL cells of patients before and after starting therapy with low-dose lenalidomide (eg, 5 mg by mouth daily). CLL blood mononuclear cells were isolated from patients (n = 5) before and 15 days after starting treatment with lenalidomide at 2.5 mg per day for the first week, and 5 mg per day for the second week. For each of these samples, the proportion of CD19⁺CD5⁺ cells exceeded 91%. Patients providing samples CLL1, CLL2, CLL3, CLL4, or CLL5 experienced a reduction from pretreatment values in their blood absolute lymphocyte counts of 41%, 22%, 40%, 88%, or 55%, respectively, following lenalidomide therapy. We examined the CLL cells for expression of p21 by immunoblot analysis (Figure 6A–B). For all patient samples examined, we observed induced-expression of p21 in the CLL cells of patients following initiation of therapy with lenalidomide. These results indicate that lenalidomide can induce the expression of p21 in CLL cells of patients receiving doses as low as 5 mg per day.

This study describes, for the first time, a mechanism by which lenalidomide directly affects CLL cells, potentially contributing to its noted clinical activity in patients with CLL. Recent studies have found that patients with CLL may have a high rate of leukemia-cell turnover, with cell death rates balancing out the birth rates of new cells, which may represent approximately 1% of the leukemia-cell clone each day.¹⁷  Agents that interfere with the capacity of leukemia cells to grow could tilt the balance in favor of cell-death, resulting in clearance of leukemia cells, even when such agents lack direct cytotoxic activity. Lenalidomide potentially could act through such a mechanism. We found that lenalidomide could cause a dose-dependent inhibition of CLL-cell proliferation that resulted in the accumulation of cells in the G_0/1 phase of the cell cycle. This activity was dependent upon the induced expression of the cell-cycle regulator p21 by a mechanism dependent on CRBN, a known molecular target of lenalidomide. Importantly, we observed these effects at concentrations as low as 0.3 µM, below the 0.6 µM estimated concentration of the drug found in the plasma of patients treated daily with 5 mg of lenalidomide.²⁹  Consistent with this, we found that the CLL cells of patients treated with lenalidomide at such doses were induced to express p21 in vivo. As such, the capacity of lenalidomide to inhibit the proliferation of CLL cells may account in part for its observed clinical activity.

We found that lenalidomide could induce p21 in leukemia cells harboring del(17p) and a single dysfunctional mutant allele of p53, indicating that the cytostatic activity of this drug did not require functional p53. This may account in part for the observed clinical activity of lenalidomide in patients with CLL cells with del(17p).^30-32  It also could support the notion of treating patients with lenalidomide to mitigate the risk of disease progression. Our results are consistent with the p53-independent mechanism by which lenalidomide upregulates p21 in the Namalwa cell line,³³  into which a mutation prevents the binding of p53 to the p21 promoter. In these cells, the upregulation of p21 is mediated by a lysine-specific demethylase 1-dependent epigenetic mechanism, leading to a reduction in histone methylation and an increase in histone acetylation of the p21 promoter, which facilitates the access to DNA of transcription factors such as sp1, sp3, Egr1, and Egr2.³³  Silencing experiments in multiple myeloma cells identified the contribution of sp1, sp3, Egr1, and Egr2 to the upregulation of p21 induced by lenalidomide or pomalidomide in plasma cells, even though none of these factors alone appeared to be physiologic regulators of p21 expression.³³  Rather, it is suggested that a combination of these factors, and possibly others from the abundant array of p21 regulators,³⁴  may be involved. A comparable complex network of factors also may be responsible for the lenalidomide-induced upregulation of p21 in CLL.

We found that the cytostatic activity of lenalidomide and its capacity to induce p21 in CLL cells was dependent upon CRBN. We observed that the silencing of CRBN in primary CLL cells mitigated the cytostatic effect of lenalidomide and inhibited its capacity to induce leukemia-cell expression of p21. This is consistent with the notion that reducing the expression levels of a drug target reduces its activity, and is also consistent with the data presented by Zhu et al, showing that the suppression of CRBN renders multiple myeloma cell lines less responsive to lenalidomide.²⁴  CRBN is the adaptor protein part of a Cul4A-DDB1 E3 ubiquitin ligase complex and is the subunit responsible for substrate recognition.³⁵  Our results suggest that the interaction of lenalidomide with this E3 ligase complex via binding to CRBN increases p21 expression, which ultimately contributes to cell-cycle arrest of CLL cells.

Recent studies identified the transcription factors IKZF1 (Ikaros) and IKZF3 (Aiolos) as being substrates of CRBN, which are rapidly degraded upon exposure to lenalidomide, and required for its antiproliferative and immunomodulatory activity.^26-28  IKZF1 and IKZF3 are transcriptional regulators involved in hematopoiesis and in the development of the adaptive immune system.³⁶  Work performed in mutant mouse models showed that IKZF1 is required at different levels in hematopoiesis, such as in the generation of chronic lymphoid precursors, while IKZF3 appears more restricted to B cells and is required for the generation of peritoneal, marginal, and recirculating B cells, as well as for the generation of long-lived plasma cells.³⁶  These transcription factors may act as transcriptional repressors by forming complexes with proteins from the mSin3 family of co-repressors and histone deacetylases.³⁷  Disruption of these complexes via lenalidomide-induced degradation of IKZF1 and IKZF3 could alleviate the repression of p21 transcription that may be mediated by factors such as Runx-1.³⁸  This would be in line with the recently described mechanism of IL-2 de-repression in T cells following lenalidomide exposure that was dependent on IKZF1/IKZF3 degradation.^26,27  Additional studies are required to evaluate whether these proteins are responsible for the cytostatic activity of lenalidomide in CLL cells.

Collectively, this study sheds new light on the mechanism-of-action of lenalidomide, which may inhibit CLL-cell proliferation. Such activity might mitigate the activity of drugs that are most active against cycling cells, accounting in part for the unfavorable results of clinical trials coupling lenalidomide with cytotoxic agents such as fludarabine monophosphate.^39-41  This mechanism could also account in part for the lack of disease progression generally observed during lenalidomide maintenance therapy, regardless of whether the patient has leukemia cells that harbor del(17p) and/or have dysfunctional p53.⁴²  Since lenalidomide is not directly cytotoxic to CLL cells, this drug might tilt the balance between leukemia-cell growth and cell death,¹⁷  conceivably resulting in a faster decline in tumor burden of patients who have relatively high rates of leukemia-cell turnover. Clinical trials combining heavy-water-labeling assessment of leukemia-cell turnover, and subsequent treatment with lenalidomide, would be required to formally test this hypothesis.

Contribution: J.-F.F. designed and performed research, analyzed data, and wrote the paper; E.M.G. performed research, analyzed data, and revised the paper; S.G. designed and performed research, and analyzed data; D.F., I.S.B., M.S., and B. Cui performed research and analyzed data; L.G.C., B. Cathers, and A.L.-G. designed research, supervised the study, and revised the paper; D.M. designed research, supervised the study, and revised the paper; T.J.K. designed research, supervised the study, provided patient samples, and wrote the paper.

Conflict-of-interest disclosure: L.G.C., S.G., B. Cathers, and A.L.-G. are employees of Celgene Corporation. T.J.K. and D.M. received financial support from Celgene Corporation. The remaining authors declare no competing financial interests.

The current affiliation for I.S.B. is Genoptix, a Novartis Company, Carlsbad, CA 92008 and for D.M. is Inception Sciences, San Diego, CA 92121

Correspondence: Thomas J. Kipps, University of California, San Diego Moores Cancer Center, 3855 Health Sciences Dr, Rm 4307, La Jolla, CA 92093-0820; email: tkipps@ucsd.edu.

The authors thank Andrew Abriol Santos Ang for his excellent technical assistance.

This work was supported in part by research funding from the Celgene Corporation (T.J.K. and D.M.); the National Institutes of Health grant for the CLL Research Consortium (P01-CA081534) (T.J.K.); the Blood Cancer Research Fund (J.-F.F.); Le Fond de la Recherche en Santé du Québec (J.-F.F.); and the Leukemia and Lymphoma Society (SCOR grant #7005-14) (T.J.K.).