CD22ΔE12 as a molecular target for RNAi therapy

Abstract

B-precursor acute lymphoblastic leukemia (BPL) is the most common form of cancer in children and adolescents. Our recent studies have demonstrated that CD22ΔE12 is a characteristic genetic defect of therapy-refractory clones in pediatric BPL and implicated the CD22ΔE12 genetic defect in the aggressive biology of relapsed or therapy-refractory pediatric BPL. The purpose of the present study was to evaluate the biologic significance of the CD22ΔE12 molecular lesion in BPL and determine if it could serve as a molecular target for RNA interference (RNAi) therapy. Here we report a previously unrecognized causal link between CD22ΔE12 and aggressive biology of human BPL cells by demonstrating that siRNA-mediated knockdown of CD22ΔE12 in primary leukemic B-cell precursors is associated with a marked inhibition of their clonogenicity. Additionally, we report a nanoscale liposomal formulation of CD22ΔE12-specific siRNA with potent in vitro and in vivo anti-leukemic activity against primary human BPL cells as a first-in-class RNAi therapeutic candidate targeting CD22ΔE12.

Introduction

CD22, a member of the Siglec (sialic acid-binding Ig-like lectins) family of regulators of the immune system, is a negative regulator of multiple signal transduction pathways critical for proliferation and survival of B-lineage lymphoid cells. The inhibitory and apoptosis-promoting signaling functions of CD22 are dependent on recruitment of the Src homology 2 domain–containing inhibitory tyrosine phosphatase (SHP)-1 to the immunoreceptor tyrosine-based inhibitory motifs (ITIMs) of its cytoplasmic domain upon phosphorylation by the Src family tyrosine kinase LYN. Without the inhibitory function of CD22, signaling pathways in B-lineage lymphoid cells would remain in “overdrive”, contributing to abnormal proliferation. B-precursor acute lymphoblastic leukemia (BPL) is the most common form of cancer in children and adolescents. BPL cells express a dysfunctional CD22 due to deletion of Exon 12 (CD22ΔE12) arising from a splicing defect associated with homozygous intronic mutations (Uckun et al. 2010). CD22ΔE12 results in a truncating frame shift mutation yielding a mutant CD22ΔE12 protein that lacks most of the intracellular domain including the key regulatory signal transduction elements, such as the ITIMs that provide docking sites for the SH2 domains of SHP1, and all of the cytoplasmic tyrosine residues (uckun et al., 2010). Our recent studies have demonstrated that CD22ΔE12 is a characteristic genetic defect of therapy-refractory clones in pediatric BPL and implicated the CD22ΔE12 genetic defect in the aggressive biology of relapsed or therapy-refractory pediatric BPL (Ma et al., 2012). The purpose of the present study was to evaluate the biologic significance of the CD22ΔE12 molecular lesion in BPL and determine if it could serve as a molecular target for RNA interference (RNAi) therapy. Here we report a previously unrecognized causal link between CD22ΔE12 and aggressive biology of BPL cells by demonstrating that siRNA-mediated knockdown of CD22ΔE12 in primary BPL cells is associated with a marked inhibition of their clonogenicity. We report a nanoscale liposomal formulation of CD22ΔE12-specific siRNA with promising in vitro and in vivo anti-leukemic activity against primary human BPL cells as an RNA interference (RNAi) therapeutic candidate targeting CD22ΔE12. These results establish the CD22ΔE12 oncoprotein as a molecular target for effective RNAi therapy against BPL.

Materials and Methods

Leukemia cells

We used 6 ALL xenograft clones that were derived from spleen specimens of xenografted NOD/SCID mice inoculated with leukemia cells from 3 newly diagnosed (including one patient with Ph⁺ BPL) and 3 relapsed pediatric BPL patients. The secondary use of leukemic cells for subsequent laboratory studies did not meet the definition of human subject research per 45 CFR 46.102 (d and f) since it did not include identifiable private information, and it was approved by the IRB (CCI) at the Children’s Hospital Los Angeles (CHLA) (CCI-10-00141; CCI Review Date: 7-27-2010; IRB Approval: 7-27-2010). Human Subject Assurance Number: FWA0001914. We also used the ALL-1 (Ph⁺ adult BPL) and RAJI (Burkitt’s leukemia/B-cell ALL) cell lines. RAJI is a radiation-resistant Burkitt’s leukemia/B-ALL cell line. ALL-1 is a chemotherapy-resistant BCR-ABL⁺ t(9; 22)/ Ph⁺ pre-pre-B ALL cell line. Both cells lines are CD22ΔE12-positive and have homozygous intronic CD22 gene mutations at Rs10413526 (C>G) and Rs4805120 (A>G) that are associated with CD22ΔE12 (Uckun et al. 2010, Ma et al. 2012).

Transfections and PCR assays

Transfections of ALL cells with small interfering RNA (siRNA) were accomplished using standard methods and procedures. PCR and RT-PCR assays were also performed according to standard procedures. See Supporting Information for a more detailed description of these methodologies.

In vitro assays

We used standard assays and procedures, including Western blot assays, RT-PCR, colony assays, flow cytometric apoptosis assays, immunophenotyping, confocal imaging, and genomic PCR. See Supporting Information for a more detailed description of these methodologies (Uckun et al. 2010; Uckun et al., 2012).

Preparation of a nanoscale liposomal CD22ΔE12-siRNA formulation

We prepared a liposomal nanoformulation (LNF) of the CD22ΔE12-siRNA duplex using the standard thin film evaporation method and used standard methods to characterize it. See Supporting Information for a more detailed description of these methodologies.

Transgenic (Tg) mice

Pronuclear microinjection of the human CD22ΔE12 transgene (Accession#LM652705), founder generation, and genotyping analysis of tail DNA, PCR-based clonality assays (Accession # LM652707), immunophenotyping by multi-parameter flow cytometry, Western blot analyses, fluorescent in situ hybridization (FISH) and spectral karyotyping (SKY) were performed using standard methods. See Supporting Information for a more detailed description of these methodologies.

NOD/SCID mouse xenograft model of human BPL

The anti-leukemic activity of CD22ΔE12-siRNA LNF was studied in a NOD/SCID mouse model of human BPL.

Pharmacokinetic (PK) Studies

The research was conducted according to Institutional Animal Care and Use Committee (IACUC) Protocols that were approved by the IACUC of CHLA. We first used the CD22ΔE12-siRNA LNF prepared using the Cy3-labeled CD22ΔE12-siRNA as a PK formulation in PK experiments. The Cy3-based fluorescence of the plasma samples was measured with an excitation at 530 nm and fluorescence emission intensity at 590 nm using the Synergy HT Biotek fluorescence microplate reader and the Gen5 software (BioTek Instruments Inc., Winooski, VT). For validation employing a different assay platform, we used a quantitative RT-PCR technique and measured CD22ΔE12-specific siRNA levels in plasma samples after administration of CD22ΔE12-siRNA LNF prepared with unlabeled CD22ΔE12-siRNA. Pharmacokinetic modeling and pharmacokinetic parameter estimations were carried out using non-linear fitting of plasma concentrations-time profiles of the combined PK dataset (i.e., fluorescence-based plasma levels and qRT-PCR based plasma levels) to 2 equations (JMP 10 Software, SAS, Cary, NC; 2 parameter exponential, 3 parameter exponential). An appropriate pharmacokinetic model was chosen on the basis of lowest sum of weighted squared residuals, lowest Akaike’s Information Criterion value, lowest SE of the fitted parameters, and dispersion of the residuals. See Supporting Information for a more detailed description of these methodologies.

Study approval

The animal research in mice was conducted according to Institutional Animal Care and Use Committee (IACUC) Protocols 280-12 and 293-10 that were approved by the IACUC of CHLA. All animal care procedures conformed to the Guide for the Care and Use of Laboratory Animals (National Research Council, National Academy Press, Washington DC 1996, USA). Leukemia cells isolated from deidentified patient specimens were used in the described experiments. The secondary use of leukemia cells for subsequent laboratory studies did not meet the definition of human subject research per 45 CFR 46.102 (d and f) since it did not include identifiable private information, and the corresponding research protocol CCI-10-00141 was approved by the CHLA IRB (CCI) (Human Subject Assurance Number: FWA0001914).

Results

CD22ΔE12 as a molecular target for RNA interference (RNAi) therapy in human B-precursor leukemia

We first sought to determine how RNAi with CD22ΔE12-specific siRNA would affect the in vitro clonogenicity and proliferation of aggressive patient-derived human BPL xenograft clones isolated from spleens of NOD/SCID mice that had developed overt leukemia with massive splenomegaly following inoculation with primary leukemia cells from the respective BPL patients (Uckun et al., 2013). Immunophenotyping by multiparameter flow cytometry confirmed the co-expression of multiple BPL-associated human lymphoid differentiation antigens, including high levels of CD10, CD19, and CD34 (Figure 1A) that are recognized markers of putative leukemic stem cells in BPL. Transfection with CD22ΔE12-siRNA, but not scrambled (scr)-siRNA, caused selective (albeit partial) depletion of CD22ΔE12-mRNA as well as CD22ΔE12-protein in aggressive BPL xenograft cells (Figure 1 B&C). This CD22ΔE12-knockdown was associated with a marked inhibition of their clonogenicity in vitro (Figure 1 D&E).

Figure 1. RNAi Knockdown of CD22ΔE12 in Aggressive Human BPL Xenograft Cells.

[A] Depicted are representative FACS-correlated two-parameter displays of BPL xenograft cells isolated from spleens of NOD/SCID mice that developed overt leukemia after inoculation with primary ALL cells from BPL patients. Cells were stained by direct immunofluorescence for human lymphoid differentiation antigens CD10, CD19, CD34, CD45, HLA-DR/DP/DQ, and HLA-A,B,C. The labeled cells were analyzed on a LSR II flow cytometer. Xenograft cells co-expressed CD45, HLA-DR/DP/DQ and HLA-A,B,C antigens and had an immature B-cell precursor immunophenotype characterized by co-expression of B-lineage progenitor antigens CD10/CALLA, CD19, and CD34. [B & C] Cells were transfected with a Nucleofector II device (Amaxa GmbH; protocol VCA-1002/C005) with either CD22ΔE12-siRNA (50 nM) or scr-siRNA (50 nM). Controls included untreated cells. Depicted in B1 and C1 are images of 1% agarose gels showing the RT-PCR products obtained from the RNA samples of CD22ΔE12-siRNA transfected xenograft cells using the P7 primer pair to amplify a 182-bp region (c.2180–c.2361) of the CD22 cDNA extending from Exon 11 to Exon 13 and spanning the entire Exon 12. Deletion of Exon 12 results in a smaller CD22ΔE12-specific PCR product of 63-bp size with this primer set migrating slightly below the 100-bp size marker. The P10 primer set was used as a positive control of RNA integrity to amplify a 213-bp region (c.433–c.645) of the CD22 cDNA present in both wildtype CD22 and CD22ΔE12 mRNA species (B2, C2). The 63-bp CD22ΔE12-mRNA was abundant in the total RNA isolated from the BPL xenograft clones. RT-PCR products were sized using the 1 Kb Plus DNA ladder from InVitrogen. Gel images were taken with an UVP digital camera and UV light in an Epi Chemi II Darkroom using the LabWorks Analysis software (UVP, Upland, CA). Whole cell lysates of xenograft cells were subjected to CD22 and alpha-Tubulin (TUB) Western blot analysis. The positions of truncated mutant CD22ΔE12 protein (B3, C3) and TUB (B4, C4) are indicated with arrowheads. BPL xenograft cells expressed a CD22ΔE12-associated truncated CD22 protein Instead of the 130/140-kDa intact CD22 protein. Transfection with CD22ΔE12-siRNA, but not scrambled (scr)-siRNA, caused selective depletion of CD22ΔE12-mRNA as well as CD22ΔE12-protein in aggressive BPL xenograft cells. [D&E] We evaluated the effects of CD22ΔE12-siRNA vs. scr-siRNA (50 nM) on blast colony formation in day 7 methylcellulose (0.8%) cultures of xenograft cells derived from 6 BPL patients (3 newly diagnosed -including one patient with Ph⁺ BPL- and 3 relapsed pediatric BPL patients). Colony formation was examined using an inverted Nikon Eclipse TS100 microscope. Images were taken using a Digital Sight DS-2MBW Nikon camera. Depicted in D1 is the image of a representative blast colony and 2 adjacent blast clusters in day 7 cultures of scr-siRNA transfected xenograft cells from a representative case. No blast colonies or clusters were observed in cultures of CD22ΔE12-siRNA transfected cells in this particular case (D2). The bar graphs shown in E depict the mean ± SE values for blast colonies per 100,000 cells plated for all ALL xenograft clones (N=6). The mean (± SE) numbers of blast colonies per 100,000 cells plated were 103±33 in untreated controls, 146±40 after scr-siRNA transfection, and only 20±12 after CD22ΔE12-siRNA transfection (Linear contrast P-value = 0.0007).

CD22ΔE12-siRNA liposomal nanoformulation causes selective CD22ΔE12 mRNA depletion and results in loss of the in vitro as well as in vivo clonogenicity of human B-precursor leukemia cells

In order to achieve RNAi in all leukemic cells in a targeted BPL cell population, we prepared a nanoscale liposomal formulation of CD22ΔE12-siRNA duplex with an encapsulation efficiency of 96.3±0.7% and a favorable stability profile both during extended storage at 4°C as well as incubation with mouse or human serum (Figure S1, Figure S2). We examined the ability of the LNF to effectively deliver CD22ΔE12-siRNA into CD22ΔE12⁺ human leukemia cells by using confocal microscopy. The LNF was capable of delivering the Cy3-labeled CD22ΔE12-siRNA into CD22ΔE12⁺ human ALL cell lines RAJI (Burkitt’s leukemia/B-cell ALL) and ALL-1 (BCR-ABL⁺ BPL), as documented by detection of the internalized CD22ΔE12 siRNA as red-fluorescent foci in the perinuclear cytoplasm as well as the nucleus in all of the cells examined within 4 h of exposure (Figure 2A). The documented nuclear delivery of CD22ΔE12-siRNA confirmed the prerequisite endosomal escape of the siRNA molecules from the internalized liposomes and showed that the CD22ΔE12-siRNA LNF should enable both cytoplasmic post-transcriptional knockdown as well as nuclear transcriptional/co-transcriptional knockdown of the CD22ΔE12 gene expression. In agreement with these observations, treatment of RAJI as well as ALL-1 cells with CD22ΔE12-siRNA LNF (49.5 nM siRNA, 117 μM lipid) resulted in a selective depletion of the target CD22ΔE12-mRNA by 48 h (Figure 2B). Neither the scrambled-siRNA containing LNF (90.6 nM siRNA, 117 μM lipid) nor the siRNA-free LNF (0 nM siRNA, 117 μM lipid) that were included as control LNF affected the CD22ΔE12-mRNA levels in RAJI or ALL-1 cells. CD22ΔE12-depletion by the CD22ΔE12-siRNA formulation was associated with a loss of in vitro clonogenicity in both cell lines, while the control formulations did not affect the in vitro clonogenic growth of either cell line (Figure 2C). Treatment with CD22ΔE12-siRNA LNF (49.5 nM siRNA, 117 μM lipid) did not affect the clonogenic growth of the CD22ΔE12-negative human DAUDI Burkitt’s leukemia/lymphoma cell line (Figure 2D). Thus, the cytotoxic effects of CD22ΔE12-siRNA LNF are dependent on the presence of its molecular target in the treated cell population.

Figure 2. Cellular Uptake, RNAi Activity and Cytoxicity of CD22ΔE12-siRNA liposomal nanoformulation (LNF).

[A]. We used 5′ Cy3-labeled CD22ΔE12-siRNA to prepare the LNF of CD22ΔE12-siRNA for cellular uptake and trafficking experiments using confocal imaging. Depicted are confocal images demonstrating the LNF-mediated CD22ΔE12-siRNA delivery into the BCR-ABL⁺ human BPL cell line ALL-1 (depicted in A1) and the Burkitt’s leukemia cell line RAJI (depicted in A2). After incubation with the LNF of Cy3-labeled siRNA (49.5 nM) for 1, 2, or 4h, the internalized Cy3-labeled CD22ΔE12-siRNA was detected and localized using the tetramethyl rhodamine (TRITC) filter sets. Cells were also stained with an anti-alpha-Tubulin MoAb and subsequently incubated with green-fluorescent Alexa Fluor 488 dye-labeled goat anti-mouse IgG (secondary Ab). Cells were then washed with PBS and counterstained with the blue fluorescent DNA-specific nuclear dye 4′,6-diamidino-2-phenylindole (DAPI). Slides were imaged using the PerkinElmer Spinning Disc Confocal Microscope and the PerkinElmer UltraView ERS software (Shelton, CT) or the Volocity V5.4 imaging software (PerkinElmer, Shelton, CT). Depicted are representative images of cells incubated with the CD22ΔE12-siRNA LNF for 4h. Similar results were obtained after 1h and 2 h incubations as well as after a 1h treatment with the LNF, wash and 24h incubation in the absence of the LNF. Arrowheads point to location of Cy3-labeled siRNA molecules inside the DAPI-stained (blue) nucleus and tubulin-containing (green) perinuclear cytoplasm. [B] Depicted are images of a 1% agarose gel showing the RT-PCR products obtained from the RNA samples of CD22ΔE12-siRNA LNF-treated ALL-1 and RAJI cells using the P7 primer pair to amplify a 182-bp region (c.2180–c.2361) of the CD22 cDNA extending from Exon 11 to Exon 13 and spanning the entire Exon 12. RT-PCR products were sized using the 1 Kb Plus DNA ladder from InVitrogen. Deletion of Exon 12 results in a smaller CD22ΔE12-specific PCR product of 63-bp size with this primer set, Gel images were taken with an UVP digital camera and UV light in an Epi Chemi II Darkroom using the LabWorks Analysis software (UVP, Upland, CA). Cells were treated for 48 h with the CD22ΔE12-siRNA LNF 4A (49.5 nM siRNA, 117 μM lipid), the scrambled-siRNA containing LNF 4C (90.6 nM siRNA, 117 μM lipid) or the siRNA-free LNF 4B (0 nM siRNA, 117 μM lipid). [C & D] Effects of CD22ΔE12 depletion by 4A on the in vitro clonogenicity of the CD22ΔE12⁺ human leukemia cell lines ALL-1 and RAJI (combined dataset depicted in C) vs. the CD22ΔE12⁻ human leukemia/lymphoma cell line DAUDI (depicted in D). Concentrations of formulations used were: 4A treatment: 49.5 nM siRNA, 117 μM lipid; 4B treatment: 0 nM siRNA, 117 μM lipid; 4C treatment: 90.6 nM scrambled siRNA, 117 μM lipid. Untreated cells (CON) as well as cells (0.5×10⁵ cells/plate, in duplicate) treated with CD22ΔE12-siRNA formulation 4A or control formulations 4B and 4C were examined for colony formation in methylcellulose supplemented (0.8%) cultures. Colony formation was examined using an inverted Nikon Eclipse TS100 microscope. Treatment with 4A (but not control formulations 4B or 4C) abrogated leukemic cell clonogenicity. Depicted are bar graphs of mean (% of untreated control/CON) colony formation for ALL-1 and RAJI cells after treatment with 4A vs. control LNF 4B or 4C. The day 7 colony counts in untreated control cultures were 333±19 colonies/0.5×10⁵ cells for ALL-1 cells and 72±3 colonies/0.5×10⁵ cells for RAJI cells. No colonies were detected in cultures of 4A-treated ALL-1 or RAJI cells. By comparison, there were 336±20 colonies/0.5×10⁵ ALL-1 cells and 53±3 colonies/0.5×10⁵ RAJI cells, respectively, in cultures of 4B-treated cells. Likewise, there were 353±65 colonies/0.5×10⁵ ALL-1 cells and 64±5 colonies/0.5×10⁵ RAJI cells, respectively, in cultures of 4C-treated cells. The values for mean% of control colony formation were not significantly reduced for 4B- (87.2±8.4%, P=0.2) or 4C-treated cells (97.8±9.7%, P=0.8). Colony formation by DAUDI cell line was not affected by CD22ΔE12-siRNA LNF. Depicted in [D] are the mean numbers of colonies per 0.5×10⁵ cells for DAUDI cells after various treatments.

We also used flow cytometric quantitative apoptosis assays to further evaluate the potency of the CD22ΔE12-siRNA LNF against RAJI and ALL-1 cells. As shown in Figure 3A & B, 96h treatment with CD22ΔE12-siRNA LNF (but not the scr-siRNA LNF) caused apoptosis in both cell lines at a 40 nM concentration. Apoptosis of target leukemia cells was documented by the significantly lower percentages of Annexin V-FITC⁻PI⁻ live cells located in the left lower quadrant of the corresponding two-color fluorescence dot plots within the P1 lymphoid window. Furthermore, there was a marked shrinkage and altered SSC as well as decreasing numbers of remaining cells in the P1 lymphoid window in the corresponding FSC/SSC light scatter plot from the 10,000 cells analyzed. The magnitude of apoptosis was concentration-dependent with no residual viable leukemia cells remaining at 400 nM (Figure 3A). Confocal images of treated cells confirmed the apoptotic destruction, as documented by an extensive loss of cell integrity, loss of tubulin, cellular shrinkage and nuclear fragmentation (Figure 3C).

Figure 3. CD22ΔE12-siRNA LNF causes apoptotic destruction of CD22ΔE12⁺ human leukemia cells.

[A & B] Cells were treated for 96 h at 37°C with CD22ΔE12-siRNA LNF (4A, 40 nM – 400 nM in [A], 40 nM in [B]), siRNA-free control LNF (4B, 0 nM siRNA), or scr-siRNA LNF (4C, 40 nM). Controls included sham-treated cells (CON) cultured x 96 h. Panel A depicts the results for the ALL-1 cell line, whereas Panel B shows data for both ALL-1 and RAJI. Cells were analyzed for apoptosis using the standard quantitative flow cytometric apoptosis assay with the Annexin V-FITC Apoptosis Detection Kit (Sigma, Catalog # APOAF-50TST). The labeled cells were analyzed on a LSR II flow cytometer. The anti-leukemic potency of CD22ΔE12-siRNA LNF is evidenced by the significantly lower percentages of Annexin V-FITC⁻PI⁻ live cells located in the left lower quadrant of the corresponding two-color fluorescence dot plots within the P1 lymphoid window as well as a marked shrinkage and altered SSC as well as decreasing numbers of remaining cells in the P1 lymphoid window in the corresponding FSC/SSC light scatter plot from the 10,000 cells analyzed. [C] Confocal images of RAJI and ALL-1 cells analyzed in [B] showing nuclear (Blue) destruction and loss of tubulin (Green) cytoskeleton after treatment with the CD22ΔE12-siRNA LNF, 4A.

We next used blast colony assays in 3 independent experiments to compare the anti-leukemic activity of CD22ΔE12-siRNA LNF (49.5 nM × 24h incubation at 37°C) alone and in combination with standard chemotherapy drugs vs. chemotherapy drugs alone against ALL xenograft clones derived from pediatric BPL patients. At this nM concentration, CD22ΔE12-siRNA LNF was significantly more potent than the chemotherapy drugs DEX at 25 μM (P=0.033) and PEG-ASP at 10 IU/mL (P=0.037) and it was at least as effective as micromolar concentrations of ADR and VCR (Figure 4). Notably, the combinations of CD22ΔE12-siRNA LNF with DEX, PEG-ASP, ADR and VCR were significantly more effective than the chemotherapy drugs alone. These results demonstrate that chemotherapy-resistant leukemic clones are not cross-resistant to CD22ΔE12-siRNA LNF and siRNA-mediated CD22ΔE12-depletion significantly impairs the ability of leukemic clones to resist the standard chemotherapy drugs.

Figure 4. CD22ΔE12-siRNA augments the potency of standard chemotherapy drugs against patient-derived B-precursor ALL xenograft cells.

We examined the potency of CD22ΔE12-siRNA LNF (49.5 nM) alone and in combination with standard chemotherapy drugs vs. chemotherapy drugs alone against ALL xenograft clones derived from pediatric BPL patients. Depicted are the cumulative data from 3 independent experiments. Each bar represents the mean ± SE values from 14 samples (4 replicate samples from Xenograft clone #1, 6 replicate samples from Xenograft clone #2, and 4 replicate samples from Xenograft clone #3). Cells were treated with the respective reagents for 24h at 37oC. Immediately after treatment, cells (1×10⁵ cells/mL) were suspended in alpha-MEM supplemented with 0.9% methylcellulose, 30% fetal calf serum, and 2 mM L-glutamine. We used the standard chemotherapy drugs commonly used in BPL therapy, including Vincristine (VCR, 11 μM), Adriamycin/Doxorubicin (ADR, 17 μM), PEG-Asparaginase (PEG-ASP, Oncospar; 10 IU/mL), Dexamethasone (DEX, 25 μM). The chemotherapy drugs were obtained from the Pharmacy of the Children’s Hospital Los Angeles. The following concentrations were used in treatments: 49.5 nM CD22ΔE12-siRNA LNF (117 μM lipid), 90.6 μM scr-siRNA LNF (117 μM lipid), 11 μM VCR, 10 IU PEG-ASP, 17 μM DOX, 25 μM DEX. Controls included untreated cells and in some experiments cells treated with siRNA-free LNF (0 nM siRNA, 117 μM lipid). 4 or 6 replicate 1 mL samples containing 0.5×10⁶ cells/sample were cultured in 35 mm Petri dishes for 7 d at 37°C in a humidified 5% CO₂ atmosphere. On day 7, colonies containing ≥20 cells were counted using an inverted Nikon Eclipse TS100 microscope. [B] depicts the statistical comparisons between various treatment groups using non-parametric Wilcoxon/Mann-Whitney tests on ranked values for each pair of treatments. CD22ΔE12-siRNA LNF alone significantly reduced the colony counts compared to DEX (P = 0.033) and PEG-ASP (P = 0.037). Significant reductions in colony counts were observed compared to control for CD22ΔE12-siRNA LNF (P = 0.0005), ADR (P = 0.039) and VCR (0.016), but not for DEX (P = 0.15) or PEG-ASP (0.56). CD22ΔE12-siRNA LNF in combination with ADR (P<0.0001), DEX (P = 0.0001), PEG-ASP (P = 0.0012) or VCR (P = 0.0022) was more effective in abrogating colony counts than treatment of ADR, DEX, PEG-ASP or VCR alone. CD22ΔE12-siRNA LNF in combination with ADR (P<0.0001), DEX (P = 0.0075), PEG-ASP (P = 0.011) or VCR (P = 0.0044) was more effective in abrogating colony counts than treatment of CD22ΔE12-siRNA LNF alone.

We next sought to determine how treatment with the CD22ΔE12-siRNA LNF would affect the leukemia-initiating in vivo clonogenic leukemic cell fraction (viz.: candidate leukemic stem cell population) in xenograft specimens derived from patients with relapsed BPL. After a 24 h treatment with the CD22ΔE12-siRNA LNF (660 nM), cells were injected intravenously into healthy NOD/SCID mice. Notably, 14 of 15 mice challenged with BPL xenograft cells that were untreated or treated with unformulated CD22ΔE12-siRNA or control liposome formulations 4B and 4C developed overt leukemia between 78 days and 103 days. Necropsy revealed massive splenomegaly at the time of death (Spleen size: 2.9±0.1 cm) (Figure 5A–C) and histopathological examinations showed evidence of disseminated leukemia with leukemic infiltrates in multiple organs, including bone marrow, brain, liver, and kidney (Figure 5D, E). By comparison, the spleen size of NOD/SCID mice challenged with CD22ΔE12-siRNA LNF-treated xenograft cells was significantly smaller (1.8±0.3 cm, P=0.0014) (Figure 5A–C) and histopathologic examinations revealed leukemic cell engraftment in only one of 4 mice examined (P=0.016) (Figure 5D, E vs. F). Thus a single in vitro exposure to CD22ΔE12-siRNA LNF to deliver CD22ΔE12-siRNA into leukemic B-cell precursors in the BPL xenograft specimens was capable of effectively abrogating the in vivo clonogenicity of their leukemia-initiating subpopulations.

Figure 5. CD22ΔE12-siRNA loaded nanoscale liposomal formulation 4 A (660 nM × 24h) abrogates the ability of in vivo clonogenic BPL xenograft cells to engraft and initiate leukemia in NOD/SCID mice.

[A & B] We observed massive splenomegaly in NOD/SCID mice that invariably developed disseminated leukemia after iv injection of B-precursor ALL xenograft cells that were either untreated or treated for 24h at 37°C with unformulated CD22ΔE12-siRNA (660 nM) or control formulations 4B (same lipid as in 4A but no siRNA) and 4C (1.2 μM scr-siRNA). Test mice were inoculated with xenograft cells that were treated with the CD22ΔE12-siRNA LNF 4A (660 nM × 24h). In the preparation of the two mouse diagrams in A, the artist used color photographs taken at the time of necropsy. Scans of the photos were used to accurately trace the contours of organ structures and mouse outlines. S, spleen. The spleen images in B were obtained using an iPhone 4S equipped with an 8-megapixel iSight camera. [C & D] The cumulative data on spleen size are shown in C. Statistical comparisons of diffuse organ infiltration data are shown in D. One mouse inoculated with untreated xenograft cells developed CNS leukemia without evidence of other organ involvement except for few clusters of leukemic blasts in the bone marrow. One of the control mice receiving 4C-treated xenograft cells showed only bone marrow involvement. Notably, 14 of 15 mice challenged with BPL xenograft cells that were untreated or treated with unformulated CD22ΔE12-siRNA or control liposome formulations 4B and 4C developed overt leukemia between 78 days and 103 days. Necropsy revealed massive splenomegaly at the time of death (Spleen size: 2.9±0.1 cm) and histopathological examinations showed evidence of disseminated leukemia with leukemic infiltrates in multiple organs, including bone marrow, brain, liver, and kidney. By comparison, the spleen size of NOD/SCID mice challenged with 4A-treated xenograft cells was significantly smaller (1.8±0.3 cm, P=0.0014) and histopathologic examinations revealed leukemic cell engraftment in only one of 4 mice examined (P=0.016). These findings provide direct experimental evidence that CD22ΔE12-directed RNAi in vivo initiated by a 24h in vitro exposure to the CD22ΔE12-siRNA formulation 4A severely damages the in vivo clonogenic fraction in xenograft cell populations derived from patients with aggressive BPL and abrogates their ability to engraft and initiate leukemia in NOD/SCID mice. [E & F] Depicted are representative results obtained using xenograft cells derived from a relapsed pediatric B-precursor ALL patient. The histopathology images of H&E stained tissue slides in E1–E4 (CON) vs. F1–F4 (4A) illustrate the presence of leukemic infiltrates in the organs of a mouse from the CON group that received untreated xenograft cells vs. their absence in the organs of a mouse from the CD22ΔE12-siRNA LNF/4A pretreatment group.

In vivo pharmacokinetics and anti-leukemic potency of the CD22ΔE12-siRNA liposomal nanoformulation

We studied the pharmacokinetics (PK) of CD22ΔE12-siRNA LNF in NOD/SCID mice with xenografted human leukemia in an attempt to determine if effective antileukemic concentrations of can be achieved at nontoxic dose levels. A PK formulation was prepared using Cy3-labeled CD22ΔE12-siRNA and plasma siRNA levels were determined using fluorescence measurements. Validation experiments were performed using LNF of unlabeled CD22ΔE12-siRNA and a CD22ΔE12-siRNA specific qRT-PCR, as described in Materials and Methods. The composite plasma concentration–time curve of CD22ΔE12-siRNA after the IV injection of a nontoxic, 500 pmols (25 nmol/kg) bolus dose as well as the calculated PK parameter values are shown in Figure 6A. A 3-parameter exponential decay pharmacokinetic model was fit to the plasma concentration–time curves, and the calculated pharmacokinetic parameter values are shown in Figure 6A. Peak plasma concentrations of CD22ΔE12-siRNA LNF in excess of 50 nM, which is highly effective against human BPL cells could be easily achieved at this dose level. The predicted maximum plasma concentration (C_max) was 81±14 nM and the systemic exposure (area under the curve [AUC])_0–24h at this dose level was 297 nM.h (AUMC 0–24h = 2364 nM · h²) (Figure 6A). The estimated mean residence time (MRT) (AUMC/AUC) was 8 h and the plasma half-life (=MRTx0.693) was 5.5 h.

Figure 6. In Vivo Pharmacokinetics and Anti-Leukemic Potency of CD22ΔE12-siRNA LNF.

[A] Plasma concentration-time profile of CD22ΔE12-siRNA LNF in NOD/SCID mice with xenografted human BPL after a single iv bolus injection of CD22ΔE12-siRNA LNF (25 nmols/kg). Depicted is the plasma concentration-time curve that was generated using the combined dataset generated using two different PK formulations, as described in Materials and Methods. The PK parameters are also shown. [B] Depicted are the EFS curves of NOD/SCID mice that were inoculated i.v. with xenograft cells (4×10⁵ cells/mouse) derived from primary leukemia cells of two pediatric patients with BPL. Sixteen control mice were either left untreated or treated with the liposomal control nanoformulation of scr-siRNA (25 nmol/kg/d × 3 d, d1–3) or an empty control LNF. Test mice were treated with the CD22ΔE12-siRNA LNF (Low-dose regimen = 2.5 nmol/kg/d × 3 d, d1–3, N=10; High-dose regimen = 25 nmols/kg/d × 3d, d1–3, N=9). Each of the 16 control mice developed fatal leukemia within 125 d with a massive splenomegaly (Spleen size: 3.4 ± 0.1 cm for untreated mice, 3.2 ± 0.1 cm for mice treated with control LNF; Nucleated spleen cell count: 564 ± 203 × 10⁶ (Log₁₀ = 8.6 ± 0.2) for untreated control mice and 828 ± 98 × 10⁶ (Log₁₀ = 8.9 ± 0.1) for mice treated with control LNF. Their spleen size and cellularity were markedly greater than those of non-leukemic control NOD/SCID mice that had not been inoculated with any leukemia cells (Spleen size 1.3 ± 0.1 cm, (T-test, Unequal variances, Log₁₀ transformed data, P<0.0001); Nucleated spleen cell count = 4.6 ± 0.4 × 10⁶ (Log₁₀ = 6.7 ± 0.04), P<0.0001). CD22ΔE12-siRNA LNF significantly improved the EFS outcome at both dose levels tested. [C] Life table statistics for the EFS curves shown in [B].

We next evaluated the in vivo anti-leukemic activity of the CD22ΔE12-siRNA LNF in NOD/SCID mouse xenograft models of relapsed BPL. All untreated control NOD/SCID mice challenged with an intravenous inoculum of xenograft cells derived from two different relapsed BPL patients developed rapidly progressive disseminated leukemia and either died or were electively killed in moribund condition due to their advanced leukemia within 125 d with a median EFS time of only 116 d. Treatment with the LNF of scr-siRNA (25 nmol/kg/d × 3 d, d1–3) or an empty LNF not containing any siRNA did not alter this aggressive course of disease and all mice on this control treatment regimen died of overt leukemia with a median EFS of 107 d (Figure 6B). In contrast, the median EFS time was more than doubled for the 19 test mice that were treated with the LNF of CD22ΔE12-siRNA at two different dose levels (EFS >250 d on the low-dose regimen = 2.5 nmol/kg/d × 3 d, d1–3, N=10); EFS >250 d on the high-dose regimen = 25 nmols/kg/d × 3d, d1–3, N=9) (P-value: <0.0001) (Figure 6B &C). Mice treated at the higher dose level of CD22ΔE12-siRNA LNF had a better EFS outcome than mice treated at the lower dose level: Their proportion remaining alive free of leukemia-associated morbidity at 200 days (viz.: 89%) was above the higher confidence interval value for the lower dose group (95%CI = 22–82%). Whereas none of the control mice was alive after 125 d, 76±15% of NOD/SCID mice that were treated with the LNF of CD22ΔE12-siRNA at the higher dose level (25 nmols/kg/day × 3 d) remained alive with no evidence of leukemia-associated morbidity at 250 d after inoculation with an invariably fatal dose of highly aggressive human BPL xenograft cells (Figure 6B & C). These results illustrate that the mutant CD22ΔE12 mRNA would be an excellent molecular target for RNAi therapy against BPL.

Forced expression of CD22ΔE12 in transgenic (Tg)-mice causes fatal CD19⁺CD24⁺CD45R/B220⁺CD127/IL7-R⁺sIgM⁻ BPL at a median age of 190 days indicating that CD22ΔE12 alone is sufficient for malignant transformation and clonal expansion of B-cell precursors in mice (Figure S3–S8). By using a combination of FISH and SKY, the approximate chromosomal band locations of the human CD22ΔE12 transgene in BPL cells from CD22ΔE12-Tg mice were identified on 18D (detected only in 11 of 31 metaphases) and 19C2 (Figure S5, Figure S8). SKY revealed a predominant clone with a normal diploid karyotype and 2 subclones with abnormal karyotypes: 40,XX [21]/41,XX, +3[3]/40,XX, Der (13)T(10D2,13D2)[7] (Figure S5 A–D). PCR-based clonality assays confirmed the clonal origin of the CD22ΔE12-Tg BPL cells and showed that the IgH locus of their immunoglobulin genes was aberrantly rearranged with a noncanonical incomplete V_H(D)J_H recombination (Figure S5 E–G) (Accession # LM652707). The presence of the CD22ΔE12 transgene was further confirmed by genomic PCR using a panel of 10 separate primer pairs recognizing different segments of the transgene (Figure S8 B–E). By using RT-PCR with a 5′ CD22ΔE12 E10 primer (E10-F) and a 3′ vector backbone primer (CD22ΔE12Tg-R2) to amplify a segment of the transgene message spanning CD22ΔE12 exons 10, 11, 13, and 14, the expression of the CD22ΔE12 transcript was confirmed in BPL cells from leukemic CD22ΔE12-Tg mice as well as splenocytes from pre-leukemic CD22ΔE12-Tg mice (Figure S8 F1). The expression of the CD22ΔE12 protein in CD22ΔE12-Tg BPL cells was confirmed by Western blot analysis (Figure S8 F2) as well as confocal fluorescence microscopy examination of frozen sections of bone marrow from leukemic mice (Figure S8 G).

Inoculation of NOD/SCID mice with 2×10⁶ leukemia cells from CD22ΔE12-Tg mice invariably caused fatal BPL with a median leukemia-free survival time of only 26d (Figure S9). Leukemic NOD/SCID recipients developed massive splenomegaly with very high spleen cell counts (Figure S9B), replacement of normal splenocytes by CD19⁺CD45R/B220⁺CD127/IL7-R⁺ BPL (Figure S9C) and leukemic infiltration of bone marrow, kidney/perirenal adipose tissue, CNS, and liver (Figure S9 D1–D4), reminiscent of the fatal BPL in NOD/SCID mice caused by inoculation of CD22ΔE12⁺ human BPL cells. A 24hr pretreatment with CD22ΔE12-siRNA LNF (200 nM) markedly impaired the ability of the CD22ΔE12-Tg mouse BPL cells to initiate leukemia in NOD/SCID mice reminiscent of its effects on leukemia-initiating human BPL xenograft cells in NOD/SCID mice. No detailed in vitro experiments were performed with the CD22ΔE12-Tg BPL cells as they don’t grow in vitro and >99% die spontaneously at 36–48 h. Notably, 10 of 10 mice challenged with 1×10⁶ CD22ΔE12-Tg BPL cells that were untreated or treated with the scr-siRNA containing control liposome formulation developed overt leukemia on day 12 after leukemic cell inoculation. Necropsy revealed massive splenomegaly at the time of death (Spleen size/cell count [in millions]: 2.5±0.3 cm/626±53 for mice receiving untreated BPL cells and 2.3±0.2 cm/227±20 for mice receiving BPL cells treated with scr-siRNA LNF). In contrast, none of the 5 mice receiving CD22ΔE12-Tg BPL cells that were treated with CD22ΔE12-siRNA LNF had evidence of overt leukemia on day 12. Their spleen sizes were normal (Figure S10 A) and their spleen cell counts were significantly lower than those of control mice challenged with untreated or scr-siRNA LNF-treated BPL cells (Figure S10 B). These findings provide direct experimental evidence that CD22ΔE12-directed RNAi in vivo initiated by a 24h in vitro exposure to the CD22ΔE12-siRNA formulation 4A severely damages the in vivo clonogenic fraction of the mouse BPL cells derived from CD22ΔE12-Tg mice BPL and abrogates their ability to engraft and initiate leukemia in NOD/SCID mice. We next set out to determine if CD22ΔE12-siRNA LNF could improve the EFS outcome of NOD/SCID mice challenged with a very aggressive BPL clone derived from leukemia cells of CD22ΔE12-Tg mouse with BPL. Notably, CD22ΔE12-siRNA LNF (2.5 nmol/kg/day × 3 days) significantly improved the EFS outcome of NOD/SCID mice challenged with these aggressive mouse BPL cells derived from CD22ΔE12-Tg mice (P<0.0001 vs. untreated mice or mice treated with scr-siRNA LNF) (Figure S10 C&D). Calculation of the 95% confidence Intervals at 3 time points (viz.: 25, 30 and 35 days after inoculation of leukemia cells) showed increased survival for the CD22ΔE12-siRNA LNF treatment compared to the upper 95% confidence interval for the combined group of scr-siRNA LNF treated and untreated control mice (Survival at 25, 30 and 35 days for CD22ΔE12-siRNA LNF treated mice were 90% (Upper 95% CI for the combined control = 48%), 90% (Upper 95% CI for the combined control = 43%) and 40% (Upper 95% CI for the combined control = 28%), respectively.

Discussion

The major challenge in the treatment of BPL is to cure patients who have relapsed despite intensive frontline chemotherapy (Reaman, 2004; Seibel, 2008; Seibel, 2011; Hastings, 2014; Gaynon, 2005). There is an urgent and unmet need to identify new drug candidates capable of destroying chemotherapy-resistant leukemic B-cell precursors (BCP). Our overarching objective in this project was to design an innovative and effective strategy that would utilize the CD22ΔE12 genetic lesion as a molecular target to gain a therapeutic advantage against chemotherapy-resistant aggressive BPL. Specifically, we sought to validate CD22ΔE12 as an effective molecular target for RNA interference (RNAi) therapy in BPL. The potent in vitro and in vivo anti-leukemic activity of the liposomal nanoformulation of CD22ΔE12-siRNA against human leukemia cells from relapsed BPL patients provided the first preclinical proof-of-concept for a potentially paradigm-shifting innovative strategy for treatment of relapsed BPL patients, whereby the chemotherapy resistant leukemic clones could be destroyed using nanoformulations of CD22ΔE12-specific siRNA as a new class of RNAi therapeutics. The ability of CD22DE12-siRNA LNF to augment the anti-leukemic potency of standard chemotherapy drugs further underlines the translational potential of this strategy.

Nanoparticles represent particularly attractive delivery systems for anti-sense oligonucleotides and small interfering RNA (siRNA) and may provide the foundation for rational design and formulation of RNAi-triggering nanomedicines (Guo, 2011; Haque et al., 2012; Karve et al., 2012; Farrell et al., 2011; Shu et al., 2014; Gokhale et al., 2002; Uckun et al., 2013). A nanoscale liposomal formulation (LNF) of CD22ΔE12-specific siRNA was developed as an RNAi therapeutic candidate against BPL and its ability to cause CD22ΔE12 depletion in primary BPL cells and abrogate their clonogenicity both in vitro and in vivo was confirmed. We used a mixture of the cationic lipid 2,3-dioleoyloxypropyltrimethylammonium chloride (DOTAP) for complexation with the polyanionic siRNA cargo as well as cell membrane penetration and the helper neutral lipid 1,2-dioleoyl-sn-glcero-3-phosphoethanolamine (DOPE) as a membrane phase inversion inducer capable of facilitating endosomal escape of the internalized siRNA molecules. Both of these lipid components have been previously used in clinical nanoformulations with favorable patient safety profiles (Chang et al., 2012; Uckun, 2012). Further development and optimization of this RNAi therapeutic candidate targeting CD22ΔE12 may facilitate a paradigm shift in therapy of relapsed BPL. Nanoparticles can also be functionalized with a tumor targeting moiety such as a ligand or a monoclonal antibody (MoAb) directed against a surface receptor on cancer cells in order to achieve optimal tumor targeting and site-specific drug delivery to further reduce their toxicity and improve their efficacy (Guo, 2011). One potential candidate surface receptor for such tumor targeting is CD19 (Myers et al., 2014). In future studies, we will explore if the antibody-mediated targeting of the CD22ΔE12-siRNA LNF will facilitate a selective uptake by CD19⁺ BPL cells and thereby further enhance its anti-leukemic potency.

The expression of CD22 on leukemic B-cell precursors has motivated the development and clinical testing of CD22-directed MoAb, recombinant fusion toxins and antibody-drug conjugates as therapeutic agents against BPL in children (D’Cruz, 2013). However, these therapeutic modalities all target the surface epitopes of CD22 and do not discriminate between normal B-cells expressing intact CD22 and BPL cells expressing CD22ΔE12. Due to the presence of CD22 on normal human B-cells and B-cell precursors, lymphotoxicity with reduced B-cell numbers and possible hypogammaglobulinemia with an increased risk of infections would be anticipated side effects of CD22 directed MoAb and MoAb based therapeutics in clinical settings. In contrast, RNAi therapeutics targeting CD22ΔE12 would only kill BPL cells while leaving normal B-cell precursors and B-cells with an intact CD22 encoded by a wildtype CD22 mRNA unharmed.