In vivo targeting of B-cell lymphoma with glycan ligands of CD22

Abstract

Antibody-mediated cell depletion therapy has proven to provide significant clinical benefit in treatment of lymphomas and leukemias, driving the development of improved therapies with novel mechanisms of cell killing. A current clinical target for B-cell lymphoma is CD22, a B-cell-specific member of the sialic acid binding Ig-like lectin (siglec) family that recognizes alpha2-6-linked sialylated glycans as ligands. Here, we describe a novel approach for targeting B lymphoma cells with doxorubicin-loaded liposomal nanoparticles displaying high-affinity glycan ligands of CD22. The targeted liposomes are actively bound and endocytosed by CD22 on B cells, and significantly extend life in a xenograft model of human B-cell lymphoma. Moreover, they bind and kill malignant B cells from peripheral blood samples obtained from patients with hairy cell leukemia, marginal zone lymphoma, and chronic lymphocytic leukemia. The results demonstrate the potential for using a carbohydrate recognition-based approach for efficiently targeting B cells in vivo that can offer improved treatment options for patients with B-cell malignancies.

Figure 1

Liposomal nanoparticles displaying glycan ligands of CD22 for targeting and killing B-cell lymphoma. (A) Schematic illustration of a chemotherapeutic-loaded liposomal formulation comprising ^BPCNeuAc lipids for active targeting to CD22. (B) Synthesis of ^BPCNeuAc-pegylated lipids.

Figure 2

^BPCNeuAc liposomes are bound and internalized by CD22-expressing cells. (A) FACS analysis for binding of fluorescently (NBD) labeled liposomes to wild-type CHO cells or CHO cells expressing recombinant human CD22. Cells were incubated with the naked (blue) or ^BPCNeuAc (green) liposomes or left untreated (gray) at 37°C for 1.5 hours before analysis. (B) CHO cells expressing CD22 were compared for their binding to the fluorescent (green) naked or ^BPCNeuAc liposomes. CD22 was detected with anti–human CD22 (red), and the nuclei were visualized by staining with DAPI. (C) Fluorescence microscopy analysis of the colocalization of ^BPCNeuAc liposomes with early endosomes. CHO-CD22 cells were incubated with fluorescent liposomes (green) as described. Early endosomes were visualized by staining with an Alexa Fluor 555–labeled anti-EEA1 (red). (D) Binding of ^BPCNeuAc liposomes to Daudi human B lymphoma cells. Daudi cells were incubated in mouse serum with liposomes containing 0% to 5% ^BPCNeuAc lipids or without liposomes (gray) before FACS analysis. (E) ^BPCNeuAc liposomes rapidly bind to Daudi cells. Fluorescent naked or ^BPCNeuAc liposomes were added to an aliquot of Daudi cells in mouse serum and incubated at 37°C for the indicated time before FACS analysis. Data are presented as mean channel of fluorescence (MCF) plus or minus SD. (n = 3). (F) Competitive binding of ^BPCNeuAc liposomes to Daudi cells in the presence of the free ^BPCNeuAc ligands. Fluorescent naked or ^BPCNeuAc liposomes were incubated with Daudi cells with the presence of the monovalent ^BPCNeuAc ligands at indicated concentration. Data were analyzed by FACS and shown as normalized MCF plus or minus SD. (n = 3). (G) Cytotoxicity of dox-loaded liposomes toward Daudi B cells. Cells were subjected to free dox, dox-loaded naked liposomes, or ^BPCNeuAc liposomes for 1 hour at 37°C. Cells were washed and incubated at 37°C for an additional 48 hours before measuring cell viability. Data shown are means of triplicate plus or minus SD. Representative data from 1 of 3 independent experiments are shown. Data were fitted using the Prism nonlinear regression software.

Figure 3

Pharmacokinetics and siglec specificity of ^BPCNeuAc liposomes. (A) ^BPCNeuAc liposomes selectively bind to Daudi cells in mouse blood. In vitro (top panels): Daudi cells were spiked into an aliquot of mouse whole blood followed by addition of fluorescent naked or ^BPCNeuAc liposomes. Cells were stained with anti–human CD19 to distinguish Daudi cells from other cells in the mouse blood. In vivo (bottom panels): after intravenous injection of Daudi cells, mice were injected with fluorescent naked or ^BPCNeuAc liposomes. After 2 hours, a blood sample was drawn and the binding of liposomes to Daudi cells was analyzed by FACS. The numbers in the quadrants represent percentages of CD19 Daudi cells that bound or did not bind to liposomes. Shown are data from 1 of 3 independent experiments. (B-C) Dox-loaded liposomes were injected intravenously to the tumor-free SCID mice (3 mice per group) without or with pretreatment of clodronate to deplete tissue macrophages. A sample of blood was withdrawn at 0.5, 2, and 25 hours after liposome injections. The plasma concentration of dox was measured, and data are presented as percentage remaining of the initial injected drug plus or minus SD. (D) FACS analysis for binding of naked or ^BPCNeuAc liposomes to siglec-expressing CHO lines and Daudi, A20, and TSn cell lines that express hCD22, mCD22 and hSn, respectively. Binding is shown as MCF plus or minus SD (n = 3). Binding degree of ^BPCNeuAc liposomes to CHO-mSn, CHO-hCD22, Daudi, and TSn cell lines was significant in comparison to the same cell line that was treated with the naked liposomes (*P < .01). (E) Comparison of ^BPCNeuAc or ^BPANeuAc liposomes in binding to cell lines expressing hCD22, mCD22, hSn, and mSn. Binding of liposomes is expressed as MCF plus or minus SD (n = 3). (F) Top panel shows that the structures of the trisaccharide ligands designed to be specific for human CD22 are based on the parent compound NeuAcα2-6Galβ1-4GlcNAc, varying the biphenyl substituent at C-9 (R1). Bottom panel shows that ^BPANeuAc liposomes exhibit a long circulation time in vivo. A sample of blood was withdrawn from mice (n = 3) that received dox-loaded naked, ^BPCNeuAc, or ^BPANeuAc liposomes at 0.5, 2, and 25 hours after liposome injections. The plasma concentration of dox was detected using a fluorometer. Data are presented as percentage remaining of the initial dose ± SD. (G) Pharmacokinetics analysis for naked and ^BPCNeuAc liposomes in wild-type C57BL/6 and Sn knockout mice. Data are presented as percentage remaining of the initial dose ± SD (n = 3).

Figure 4

Efficacy of CD22-targeted liposomes in a xenograft model of the disseminated Daudi human B lymphoma. (A) Top panel shows timeline for the in vivo efficacy study. Bottom panel: tumor-bearing mice that received PBS (untreated, n = 10), dox-loaded naked liposomes (n = 8), or ^BPCNeuAc liposomes (n = 8) containing 2% or 5% ^BPCNeuAc ligands were monitored for the onset of hind-leg paralysis for up to 100 days. Survival rate is presented in a Kaplan-Meier plot with indication of numbers of long-term survivor animals. (B) Estimation of residual Daudi lymphoma cells in the bone marrow. Shown are histograms of bone marrow cells isolated from tumor-bearing mice followed by staining with isotype (solid) or anti–human CD19 (line) antibodies to identify infiltrated Daudi cells. Shown are results from 1 of 3 representative mice that received the indicated treatment. Percentages of lymphoid gated CD19⁺ Daudi cells in the bone marrow are indicated. N.D. (not detected) refers to less than a 0.4% background observed for an IgG isotype control.

Figure 5

CD22-targeted liposomes bind to and kill malignant cells from patients with B-cell lymphomas or leukemias. (A) Binding of ^BPCNeuAc liposomes to patient B cells. Shown are representative samples from 1 healthy donor and 6 patients with HCL, CLL, or splenic MZL. Lymphocytes were gated based on the forward- and side-scatter characteristics. B-cell populations detected with anti–human CD19 were analyzed for binding of anti–human CD22 (top row), naked liposomes (middle row), and CD22-targeted ^BPCNeuAc liposomes (bottom row). CD22-bright HCL (red circle) was distinguished from normal B cells (black circle). (B) Correlation of binding of ^BPCNeuAc liposomes with CD22 or CD20 expression on the B-cell lymphomas. The diagonal lines represent linear regression that was analyzed using Prism software and the values of goodness of fit (r²) are indicated (n = 25). (C) Cytotoxicity of the dox-loaded ^BPCNeuAc liposomes toward malignant B cells. Top panel shows results of the viability of blood lymphocytes evaluated by the standard MTT assay after treatments of dox-loaded naked or ^BPCNeuAc liposomes with dox concentrations at 10 or 40μM. Cells left untreated (Unt) were defined as the maximal cell viability. Complete cell killing was determined from the Triton X-100 lysed cells (Tri). Bottom panel shows percentages (means of triplicate ± SD) of the viable blood lymphocytes after treatment with dox-loaded naked or ^BPCNeuAc liposomes. *P < .05 compared with control treatments of naked liposomes. Representative data from 1 of 4 samples are shown.