Targeted nanoparticle delivery overcomes off-target immunostimulatory effects of oligonucleotides and improves therapeutic efficacy in chronic lymphocytic leukemia

Abstract

Several RNA-targeted therapeutics, including antisense oligonucleotides (ONs), small interfering RNAs, and miRNAs, constitute immunostimulatory CpG motifs as an integral part of their design. The limited success with free antisense ONs in hematologic malignancies in recent clinical trials has been attributed to the CpG motif-mediated, TLR-induced prosurvival effects and inefficient target modulation in desired cells. In an attempt to diminish their off-target prosurvival and proinflammatory effects and specific delivery, as a proof of principle, in the present study, we developed an Ab-targeted liposomal delivery strategy using a clinically relevant CD20 Ab (rituximab)-conjugated lipopolyplex nanoparticle (RIT-INP)- and Bcl-2-targeted antisense G3139 as archetypical antisense therapeutics. The adverse immunostimulatory responses were abrogated by selective B cell-targeted delivery and early endosomal compartmentalization of G3139-encapsulated RIT-INPs, resulting in reduced NF-κB activation, robust Bcl-2 down-regulation, and enhanced sensitivity to fludarabine-induced cytotoxicity. Furthermore, significant in vivo therapeutic efficacy was noted after RIT-INP-G3139 administration in a disseminated xenograft leukemia model. The results of the present study demonstrate that CD20-targeted delivery overcomes the immunostimulatory properties of CpG-containing ON therapeutics and improves efficient gene silencing and in vivo therapeutic efficacy for B-cell malignancies. The broader implications of similar approaches in overcoming immunostimulatory properties of RNA-directed therapeutics in hematologic malignancies are also discussed.

Figure 1

TLR9-driven immunostimulation of free G3139 is associated with limited modulation of Bcl-2 expression. (A) Effect of free G3139 on Bcl-2 and Mcl-1 protein levels. Top panel shows 2 representative Western blot results of n = 10 B-CLL patient cells. Primary CLL B cells were incubated with 1, 2, and 5μM G3139 for 48 hours and then collected and lysed for Western blot analysis. Bottom panel shows average Western blot band intensities determined by densitometry. Data are presented as relative percentages compared with untreated cell controls (for Bcl-2, n = 10, mean ± SEM; for Mcl-1, n = 4, mean ± SEM). (B) Relative CLL B-cell viability normalized to medium control, determined by annexin V/propidium iodide staining (n = 10). Primary CLL B cells were incubated with 1, 2, and 5μM G3139 for 48 hours. Relative cell viability was defined as the percentage of annexin V⁻ and propidium iodide⁻ cells relative to the untreated control group. (C) Fold changes of surface markers relative to medium controls in CLL B cells after G3139 treatment. Primary CLL B cells were incubated in the presence of 1, 2, and 5μM G3139. After 48 hours, expressions of CD40, CD80, and CD86 were measured by flow cytometry. The data are presented according to MFI (n = 10, mean ± SEM). (D) Effect of knock-down of TLR9 on the immunostimulatory properties of G3139. Top panel is the fold changes of surface markers relative to medium controls after TLR9 down-regulation by siRNA. Primary CLL B cells were transfected by TLR 9 siRNA using electroporation and incubated for 48 hours, followed by further treatment with 1μM G3139. After 24 hours, the expressions of CD40, CD80, and CD86 were measured by flow cytometry. The data are presented as MFI (n = 4, mean ± SEM). Variable cell death (5%-20%) after transfections was observed. Bottom panel is a representative Western blot showing down-regulation of TLR9 expression by siRNA. (E) Subcellular distribution analysis of TLR9 in CLL B cells. Purified CLL B cells were fixed and stained intracellularly with anti-TLR9, anti–EEA-1, or anti–LAMP-1 antibodies as described. Images were acquired using an Olympus FV1000 confocal microscope. Scale bar indicates 10 μm. The white arrows indicate colocalization (yellow dots).

Figure 2

Design, characterization, and optimization of RIT-INPs. (A) Design of G3139-loaded RIT-INPs for CLL B cell–specific targeting to overcome immunostimulatory effects. (B) Atomic Force Microscopy images of NPs and RIT-INPs. Shown are G3139 encapsulated in NPs (1) and G3139 encapsulated in RIT-INPs (2). Particle suspensions were dried on a mica substrate. All measurements were recorded in both height and amplitude modes. Height images are presented here. (C) Confocal micrographs of uptake of cy3-G3139 (1μM) loaded in RIT-INPs under various molar ratios of RIT/total lipids (approximately 1/500-1/12 000) in Raji cells after 4 hours of incubation. DIC indicates differential interference contrast (bright-field) images. Scale bar indicates 10 μm (D) Flow cytometry analysis of binding/uptake of RIT-INP carrying FAM-G3139 in Raji cells compared with free FAM-G3139 and nontargeted NP carrying FAM-G3139. Raji cells were treated with FAM-G3139–loaded RIT-INPs with various molar ratios of RIT/total lipids. Data are presented according to mean fluorescence intensity (MFI) changes (n = 3, mean ± SD). (E) Binding of free FAM–labeled G3139 and various NP-formulated FAM-ODN to Raji (CD20⁺) and Jurkat (CD20⁻) cells. (F) Binding of free FAM-G3139 and various NP-formulated FAM-G3139 to CLL B cells. The CLL B cells were incubated with free FAM-G3139 or FAM-G3139 in HER-INP or RIT-INP with the concentration of 1μM at 37°C for 1 hour and washed twice with cold PBS. (G) Inhibition of RIT-INP binding to Raji cells by excess RIT or alemtuzumab (anti-CD52). Untreated cells (bold line), cells treated with anti-CD20 ILP (thin solid line), and cells blocked with rituximab or alemtuzumab (broken line) were assessed by flow cytometry.

Figure 3

Differential compartmentalization, immunostimulatory effects, target down-modulation, and cytotoxicity by G3139 and RIT-INP–G3139 in CLL B cells. (A) RIT-INPs mediate the early endosomal compartmentalization of G3139. Purified CLL B cells were incubated with free FAM-G3139 (2μM) or RIT-INP–FAM-G3139 (2μM) for 1 hour. Cells were washed, fixed, and stained with anti–EEA-1 or anti–LAMP-1 antibodies for confocal observation. The white arrows indicate the colocalization (yellow dots). Scale bar indicates 10 μm. (B) Fold changes of costimulatory molecules relative to medium control in CLL B cells. Primary CLL B cells were incubated in the presence of free G3139 (1μM), and RIT-INP–G3139 (1μM). The data are based on MFI. Results are shown as means of n = 10 independent experiments. (C) Effect of RIT-INP–G3139 on Bcl-2 protein level. Top panel is a comparison of relative Bcl-2 protein level (n = 5, mean ± SEM). Primary CLL B cells were incubated with free G3139 or HER-INP- or RIT-INP–formulated G3139 and G3622 at 2μM for 48 hours. Average Western blot band intensities were determined by densitometry and data are presented as relative percentages compared with untreated cell controls. Bottom panel is a presentative Western blot analysis of Bcl-2 protein in CLL B cells. (D) Relative percentage of CLL B-cell viability normalized to medium controls. CLL B cells were treated with various conditions at 37°C for 48 hours. The percentage of viable cells was determined by annexin V/propidium iodide staining and was analyzed by flow cytometry (n = 5, mean ± SEM). (E) Improved cytotoxicity of fludarabine after treatment by G3139-loaded RIT-INPs. Relative cell viability was determined by propidium iodide staining of free G3139 and various G3139-containing INPs at 1μM for 24 hours, followed by fludarabine (1μM) for another 48 hours. Results are presented as means of n = 4 independent experiments. (F) Induction of NF-κB–binding activity detected by the EMSA assay. CLL B cells were incubated with free G3139 and INP-loaded G3139 at 1μM for 4 hours. The CLL cells were stimulated with 500 ng/mL of CD40L for 1 hour as a positive control. (G) Western blot analysis of NF-κB (p65) phosphoration. CLL B cells were incubated with free G3139 and INP-loaded G3139 at 1μM for 3 hours, followed by lysis for Western blot analysis. (H) Differential cytokine inductions of IL-6, TNF-α, and IFN-γ on CLL B cells. Primary CLL B cells were treated under the indicated conditions for 48 hours. Supernatants were collected for ELISA analysis.

Figure 4

Rituximab signaling is not compromised in RIT-INP formulation. (A) Time-dependent changes of p38 in CLL B cells activated by free G3139 and G3139-containing RIT-INPs. The p38 MAPK activation in CLL B cells treated by free G3139 (1μM) and RIT-INP–encapsulated G3139 (1μM) was monitored by Western blotting. (B) Partial inhibition of costimulation of free G3139 by cross-linking of rituximab with anti-Fc. Fold changes of costimulatory molecules on CLL B cells with treatment of G3139 (1μM), RIT (10 μg/mL), and anti-Fc (50 μg/mL) after 48 hours were measured by flow cytometry. Results are shown as means of n = 3 independent experiments. (C) Partial inhibition of costimulation of free G3139 with cross-linking of empty INP. Similarly, CLL B cells were treated by free G3139 plus empty RIT-INP for 48 hours and then the expressions of costimulatory molecules were measured by flow cytometry (n = 3, mean ± SEM). (D) The effect of CLL B cells cotreated by free G319 and empty RIT-INP on Bcl-2 down-regulation. Comparison of protein levels of CLL B cells treated by free G3139 (1μM) and G3139 (1μM) plus empty RIT-INP and RIT-INP–G3139 (1μM) were measured by Western blotting after 48 hours.

Figure 5

Evaluation of in vivo therapeutic efficacy of RIT-INP–G3139. (A) Enhanced Bcl-2 down-regulation of RIT-INP–G3139 in Raji cells. Raji cells were treated with free G3139 (2μM) or various formulated G3139 (2μM) and G3622 (2μM) for 48 hours. Cells were collected and lysed for Western blot analysis. (B) Relative percentage of Raji cell viability of RIT-INP–G3139 normalized to medium control. The percentage of viable cells was determined by annexin V/propidium iodide staining and was analyzed by flow cytometry. Results are presented as means of n = 3 independent experiments. (C) Survival curve for NOD-SCID Raji mice treated with various formulations of G3139 at 5 mg/kg. NOD-SCID mice engrafted with 2 × 10⁶ Raji cells on day 0. The 10-day treatment was initiated via IP injection on day 4. For another in vivo study (D-G), mice were engrafted by Raji cells and treated with free G3139 (5 mg/kg), RIT-INP–G3139 (5 mg/kg), or HER-INP–G3139 (5 mg/kg) on day 10 after inoculation and total 3 treatments were given for each group. On day 15, mice were killed for analysis. (D) Percentage of CD20⁺ cells from BM assessed by flow cytometry. (E) Immunohistochemical staining of Bcl-2 on BM from the Raji xenograft model. The arrows indicate the intracellular Bcl-2 stained Raji cells (brown color) among the control and treated groups. (F) Fold changes of costimulatory molecules (CD40 and CD86) in cells from BM. Data were normalized to the PBS-treated group (n = 4, mean ± SEM). (G) Cytokine production of IL-6 (n = 4) and IFN-γ (n = 4) in serum from mice treated with PBS, free G3139 (5 mg/kg), and RIT-INP–G3139 (5 mg/kg) or HER-INP–G3139 (5 mg/kg). Cytokine release was determined by ELISA.

Figure 6

B cell–specific delivery of the payload and lack of immunostimulation by RIT-INP in whole blood from CLL patients. (A) Nonselectivity of free FAM-ODN (1μM) in PBMCs. (B) Selectivity of RIT-INP–encapsulated FAM-ODN (1μM) in PBMCs. (C) Fold changes of costimulatory molecules (CD40 and CD86) in whole blood from CLL patients. Expression was determined by flow cytometry. Data were normalized to PBS-treated blood. (D) Cytokine production of IL-6 (n = 4) and TNF-α (n = 4) in serum from patient whole blood treated with PBS, free G3139 (25 μg/mL), and RIT-INP–G3139 (25 μg/mL) or HER-INP–G3139 (25 μg/mL). Cytokine release was determined by ELISA.

Figure 7

Proposed mechanisms of the reduced activation of G3139 in RIT-INPs. The recognition of G3139 by TLR9 mainly in late endosomes activates NF-κB pathway due to the CpG motifs in the G3139 sequence, thereby up-regulating its downstream target antiapoptotic proteins and inducing cytokine release. We present a new strategy of using INPs to control endosomal compartmentalization of ONs using the preserved signaling from antibodies on INPs. The early endosomal retention of liposomal nanoparticles and the cross-linking effect of rituximab lead to the reduced activation of G3139 in RIT-INPs, thus enhancing their gene-silencing effects in leukemic cells in vitro and in vivo.