Siglecs as targets for therapy in immune cell mediated disease

Acute myelogenous leukemia (AML) depletion therapy: targeting CD33

The primary goal for treatment of AML is to deplete the tumor cells without killing the host. CD33 was identified as a target for immunotherapy upon the demonstration that the receptor was expressed on myeloblasts of 90% of all patients suffering from AML. Gemtuzumab ozogamicin (GO, Mylotarg™), a calicheamicin-conjugated humanized murine anti-CD33, was approved for cell depletion therapy in AML patients in 2000 (Table 2)^{36, 37}. Binding and endocytosis of GO by AML cells is followed by intracellular release of calicheamycin and disruption of DNA synthesis causing cell death. It has been recently reported that anti-CD45 enhances the in vitro, efficacy of GO, but not calicheamicin alone, and significantly extends survival rates in murine models of human AML compared to treatment with either anti-CD45 or GO alone.³⁸ This effect may be explained by enhanced uptake of GO considering the observation of anti-CD45 dependent increased internalization of an unconjugated anti-CD33, but additional contributions from CD45 signaling or activation of Fc receptor signaling have not been ruled out. CD33 is expressed on many myeloid cells (e.g. monocytes, neutrophils, other granulocytes and myeloid precursors), resulting in severe myelosuppression and neutropenia in all patients. However, since CD33 is not expressed on pluripotent hematopoietic stem cells,³⁹ these cells replenish the myeloid cell compartment over time. Thus, despite the rather broad cell type expression of CD33, the side effects resulting from temporary depletion of normal CD33-expressing myeloid cells are manageable.

Although Siglec-9 exhibits a similar expression profile to CD33, there is no evidence at present to suggest that it would be a better target for cell depletion therapy²⁶. Siglec-7 is also expressed on AML cells. As a proof of principle for Siglec-7 directed immunotherapy, fluorescent dye-loaded immunocolloidal particles bearing anti-Siglec-7 antibodies conjugated to the surface have been constructed. These nanoparticles bound to and were taken up by Siglec-7 expressing mouse embryonic fibroblasts in a Siglec-7 dependent manner, suggesting another drug targeting approach.⁴⁰ In an alternative CD33-targeting approach, natural killer cells bearing a recombinant chimeric anti-CD33 bearing T cell receptor was able to elicit specific lysis of an AML cell line⁴¹. However, this interesting approach does not readily translate to a clinical path for treatment of AML since it involves recombinant proteins in NK cells.

B cell depletion therapy targeting CD22 (Siglec-2)

CD22 (Siglec-2) has restricted expression on mature B cells, which lose CD22 expression upon differentiation to plasma cells.¹⁵ Since its identification as a marker for B cell malignancies^{3, 4, 6}, CD22 has been pursued as a target for cell depletion therapy, with ongoing clinical trials for three anti-CD22 antibodies currently in clinical development (Table 2). In the meantime, B cell depletion therapy has become a well-established treatment for NHL as a result of the clinical success of Rituxan, an anti-CD20 B cell-specific antibody, approved by the FDA in 1997. Rituxan is a native antibody that relies on CDC and ADCC-mediated immune responses to ablate NHL cells. Despite the success of Rituxan, there is ample room for improvement. While first-time treatment of NHL patients with Rituxan plus standard chemotherapy (CHOP) achieves 90% response rates, and 4 year survival of >80% of the patients, most patients eventually relapse, and 50-60% of relapsed patients do not respond to Rituxan^42-45. Consequently, there is a need for improved methods of treatment of NHL patients, particularly for agents that can work synergistically with Rituxan. Currently there are three anti-CD22 antibodies in clinical trials for treatment of B cell malignancies, two that are immunotoxins (BL22 and CMC-544) and one that is a native antibody (Epratuzumab).

BL22 is an anti-CD22 conjugated to a Pseudomonas exotoxin. In phase I/II clinical trials investigating the efficacy of BL22 against refractory hairy cell leukemia, 80% of patients showed complete or partial remission⁴⁶. Recently, BL22 cytotoxicity was compared with a similar anti-CD19 immunotoxin in a panel of human lymphoma cell lines, and shown to have 10-100-fold lower IC₅₀ values, despite 4-9-fold lower levels of CD22 expression compared to CD19.⁴⁷ Although both conjugates were endocytosed, the improved cytotoxicity of BL22 correlated with efficient endocytosis by CD22, aided by rapid replenishment of cell surface CD22 from intracellular pools.⁴⁷ These results underscore the utility of CD22 as a target that can carry toxic cargo into the cell.

CMC-544 is a humanized IgG₄ anti-CD22 antibody conjugated to the chemotoxin calicheamicin. Its construction is analogous to the anti-CD33 based Mylotarg™ approved for the treatment of AML, where calicheamycin is conjugated to the antibody via an acid-labile bond, requiring endocytosis into acidic compartments of the cell to release the active agent. Thus, like BL22, it has been designed to optimally use the endocytic activity of CD22. CMC-544 has demonstrated dramatic efficacy in murine models of human NHL^{48, 49} and ALL⁵⁰. It shows strong synergy with Rituxan in a disseminated model of NHL⁴⁸, and shows superior activity to Rituxan in regression of established subcutaneous ALL tumors⁵⁰. CMC-544 is currently in Phase II/III trials for treatment of NHL and diffuse large B cell lymphoma.

Epratuzumab is a humanized IgG₁ anti-CD22 antibody that is being pursed in clinical trials for treatment of NHL and Systemic lupus erythamatosis (SLE)⁵¹. As a single agent, Epratuzumab produces only a modest reduction of B cells^{52, 53}, perhaps not surprising since it is a native antibody that is rapidly taken up by B cells⁵⁴. In SLE patients, Epratuzumab was found to deplete CD27⁻ B cells, which represent naïve and transitional B cells, as opposed to memory B cells and plasmablasts⁵⁵, although the mechanism is not known. Combination therapy with Rituxin and Epratuzumab have been ongoing, and favorable results from an international, multicenter phase 2 study of patients with follicular NHL or small lymphocytic lymphoma were reported recently.⁵² This study included patients with relapsed/refractory, indolent NHL, following previous chemotherapy; the majority of patients achieved at least an objective response, with many of these also achieving durable, complete responses. A recent study describes treatment of a B lymphoma xenograft in mice with an ⁹⁰Y-conjugated Epratuzumab in combination with the unconjugated anti-CD20 Veltuzumab.⁵³ While ⁹⁰Y-Epratuzumab alone demonstrated efficacy, the therapy was greatly improved by the inclusion of Veltuzumab. The mechanism of action and possible synergism of this treatment has not been fully elucidated, although it is expected that the Epratuzumab is efficiently endocytosed by CD22, causing an accumulation of radioisotope in the cell. In an antibody-tethered combination therapy approach, the Fab fragments of anti-CD22 and whole CD20 antibodies were covalently linked⁵⁶. Interestingly, this conjugate does not undergo endocytosis, despite engaging CD22. Presumably the retention of CD20 on the cell surface is dominant over the endocytic behavior of CD22. This phenomenon suggests a different mode of action from treating with the two antibodies separately, which is supported by the more potent inhibition of cell proliferation by the covalent conjugate compared to administration of Epratuzumab and Veltuzumab as separate entities.

Although Epratuzumab is not efficient at depletion of normal B cells, a panel of antibodies that bind to different CD22 epitopes have recently been evaluated, revealing that antibodies that block ligand binding cause dramatic depletion of both normal B cells and B lymphoma cells⁵⁷. Since the antibodies do not cause apoptosis of B cells in vitro, it is postulated that blocking ligand binding affects the half-life and turnover of the cells. The results suggest a novel approach to B cell depletion using antibodies that block the ligand-binding site of CD22.

In recent years B cells have been increasingly appreciated for their role in initiation and maintenance of autoimmune disease, suggesting the potential for B cell depletion therapy in breaking the inflammatory cycle in these diseases^58-61. Approval of Rituxan for treatment of rheumatoid arthritis in combination with methotrexate has had a major impact on treatment of the disease and stimulated investigation on the role of B cells in various autoimmune diseases. Epratuzumab is currently under investigation for treatment of SLE⁵⁵, and it is likely that other CD22-targeted therapeutics will be investigated in autoimmune disease as they progress in clinical trials.

Targeting Siglec-8 on eosinophils

Eosinophils express Siglec-8 in a highly cell-type restricted manner, with basophils showing only weak expression (Table 1). While there is no murine ortholog, Siglec-F has been documented as a functional paralog, due to its restricted expression on eosinophils, and the unique specificity of both Siglec-8 and Siglec-F for a sulfated-sialylated glycan ligand (6-sulfo-sialyl Lewis X)^{62, 63}. Genetic ablation of Siglec-F in mice results in lung, blood and bone marrow eosinophilia when challenged with an allergen, which is consistent with the observed upregulation of Siglec-F and its ligands upon allergen challenge in wild-type mice.⁶⁴ Above-normal concentrations of eosinophils and enhanced eosinophil responses in the knock-out mice established Siglec-F as a negative regulator of eosinophil activation in vivo. Anti-Siglec-8 antibodies, in the presence of secondary antibodies, induce apoptosis of eosinophils by triggering signaling through a caspase-dependent pathway⁶⁵. Interestingly, it was shown that eosinophil sensitivity to apoptosis was increased by cytokines such as IL-5, which normally promote eosinophil survival. It is significant that this mechanism involves the cell signaling function of Siglec-8, suggesting that native antibodies could be used for eosinophil depletion in vivo by a mechanism that does not involve CDC or ADCC. Antibodies to murine Siglec-F similarly cause apoptosis of murine eosinophils, and in vivo can induce a marked depletion of eosinophils from the peripheral blood of mice with eosinophilia⁶⁶. The treatment has no significant effect on other leukocytes. Siglec-F antibodies were also used to successfully treat eosinophilic inflammation in a mouse model of oral egg ovalbumin-induced inflammation in the gastro-intestinal mucosa.⁶⁷ Both eosinophil numbers and gastro-intenstinal mucosal damage were significantly reduced in anti-Siglec-F treated mice. Thus, Siglec-8 may represent an attractive target for the treatment of hypereosinophilia, as well as allergic disorders involving eosinophils. Considering the routine use of intravenously-administered immunoglobulins (IVIg) to treat inflammation, it is noteworthy that autoantibodies including anti-Siglec-8 are found in these preparations, suggesting therapeutic relevance for autoimmune and allergic disorders such as Churg-Strauss syndrome.⁶⁸ There is also evidence for the use of anti-Siglec-8 antibodies to target mast cells for the inhibition of FcεRI-dependent mediator release. While anti-Siglec-8 does not induce apoptosis of mast cells, it was shown to inhibit calcium flux, release of histamine and prostaglandin-D2, and bronchial smooth muscle contraction upon stimulation with anti- FcεRI.⁶⁹ Interestingly, this inhibition was found to be dependent on the ITIM of Siglec-8.

Restricted expression of sialoadhesin on tissue macrophages

Sialoadhesin (Siglec-1) exhibits highly restricted expression on subsets of resident tissue and inflammatory macrophages and activated monocytes^{24, 70, 71}. While there has been little attempt to date to directly target sialoadhesin, this siglec has been implicated for its potential to target these cells. Macrophages are believed to be critical effector cells in the inflammation associated with autoimmune disease, and the efficacy of Rituxan in rheumatoid arthritis is believed to result from the ablation of B cells, which are responsible for activation of these macrophages⁵⁹. Similarly, the secretion of a macrophage-activating factor, versican, from tumor cells promotes an inflammatory microenvironment that aids in tumor metastasis.⁷² Thus, targeting sialoadhesin-bearing macrophages might have impact in the treatment of inflammatory responses that promote rheumatoid arthritis and tumor metastasis.

Sialoadhesin is also gaining interest for its role in viral infections^{24, 25, 71}. An increase in Sialoadhesin expression on CD14+ monocytes has been shown to correlate with HIV-1 viral load in humans, both by RT-PCR analysis and flow cytometry. Sialoadhesin binds HIV-1 directly, and is responsible for trans-infection of other cells^{24, 71}. One recent report suggests targeting of sialoadhesin as a vaccine strategy to deliver antigens for presentation to T cells, taking advantage of the localization of macrophages on spleen and lymph nodes, and the rapid endocytic activity of sialoadhesin⁷³.

Targeting Siglec-9 for inflammatory disorders

Siglec-9 is primarily expressed on monocytes, neutrophils, and dendritic cells. It was recently found that anti-Siglec-9 antibodies induce apoptosis in neutrophils,⁷⁴ and that this cytotoxicity is enhanced in the presence of pro-inflammatory cytokines, such as GM-CSF.⁷⁵ While in the absence of cytokines the effect is caspase-dependent, the cytokine-enhanced effect is caspase-independent, but involves reactive oxygen species (ROS). These results suggest that in vivo, the cell-killing effect may be more potent in hyperinflammatory microenvironments. This phenomenon may already play a role in the routine treatment of autoimmune diseases with IVIg. Autoantibodies, including anti-Siglec-9, have been identified in IVIg, and intact, but not anti-Siglec-9-depleted, IVIg was shown to induce neutrophil cytotoxicity in vitro.⁷⁶ Consistent with the previous study using anti-Siglec-9, IVIg-induced neutrophil cytotoxicity was enhanced in the presence of pro-inflammatory cytokines.⁷⁶ These findings have important implications for the clinical use of IVIg, by providing a deeper understanding of the mechanisms of action, and for the potential of targeting Siglec-9 specifically in the treatment of hyperinflammation.