Abstract
SM03, an anti-CD22 recombinant IgG1 mAb, is currently in a phase III clinical trial for the treatment of rheumatoid arthritis (NCT04312815). SM03 showed good safety and efficacy in phase I systemic lupus erythematosus and phase II moderate to severe rheumatoid arthritis clinical trials. We propose the success of SM03 as a therapeutic to systemic autoimmune diseases is through the utilization of a novel mechanism of action unique to SM03. CD22, an inhibitory coreceptor of the BCR, is a potential immunotherapeutic target against autoimmune diseases. SM03 could disturb the CD22 homomultimeric configuration through disrupting cis binding to α2,6-linked sialic acids, induce rapid internalization of CD22 from the cell surface of human B cells, and facilitate trans binding between CD22 to human autologous cells. This in turn increased the activity of the downstream immunomodulatory molecule Src homology region 2 domain-containing phosphatase 1 (SHP-1) and decreased BCR-induced NF-κB activation in human B cells and B cell proliferation. This mechanism of action gives rationale to support the significant amelioration of disease and good safety profile in clinical trials, as by enabling the “self” recognition mechanism of CD22 via trans binding to α2,6 sialic acid ligands on autologous cells, SM03 specifically restores immune tolerance of B cells to host tissues without affecting the normal B cell immune response to pathogens.
Introduction
CD22 is a 135-kDa transmembrane sialic binding protein (also known as Siglec-2) that preferentially binds to sulfated sialoside NeuAcα2–6Galβ1–4[6-SO4]GlcNAc, and to a lower extent NeuAcα2–6Galβ1–4GlcNAc (α2,6-linked sialic acid). CD22, an Ig-type lectin, shares its homologous sequence and structural features to the Siglec family (1, 2). CD22 is primarily expressed in B cells, where it is traditionally understood to function as an immune checkpoint to assist B cells to identify self-associated molecular patterns, thus modulating BCR activation. This is done through the phosphorylation of the well-characterized ITIMs in the cytoplasmic tail domain of CD22 upon BCR cross-linking (2, 3). The cytoplasmic domain then recruits and phosphorylates the protein tyrosine phosphatase SHP-1 (Src homology region 2 domain-containing phosphatase 1), leading to a cascade of events that act to inhibit BCR downstream activation (4).
CD22 utilizes its particular molecular structure of a rigid rod and its unique orientation of the N terminus extracellular Ig-like domain, containing the ligand binding site that faces away from the membrane (5), to preferentially facilitate trans binding to multiple ligands expressing α2,6-linked sialic acid residues present on glycoproteins expressed by a variety of immune cell types. These include T and B lymphocytes (T and B cells), neutrophils, monocytes, erythrocytes, activated endothelial cells, and others (1, 6, 7). Other tissues in human organs also express α2,6-linked sialic acid, such as human lung epithelial cells (not goblet cells); splenic cords and follicles; ileum stoma; heart stroma and blood vessels; skin epithelial cell secretions in eccrine sweat glands; colon stromal cells; brain white matter; and liver sinusoids and stromal cells, to name a few (8). Binding of CD22 to the trans ligand and coengagement of the BCR can facilitate the CD22 downstream signaling response, leading to the suppression of BCR activation and/or induction of B cell tolerance (9–12). On resting B cells, CD22 prominently forms CD22 homomultimers by cis ligand binding (3, 13) and appears as clusters on the cell surface in “distinct islands” separated from the BCR (14), whereby disruptions of cis ligand binding increased lateral mobilization and therefore regulation of BCR signaling on the cell surface (14). CD22 also demonstrates ligand binding–dependent endocytosis and recycling from the cell surface via a clathrin-coated vesicle–dependent mechanism (15). Collectively, much research has been dedicated to the understanding of the physiological function of CD22 and implicates a potential therapeutic target against autoimmune diseases through immunomodulation (11).
B cells play a central role in modulating the immune response, and there has recently been a rise in the interest of biologics that use the strategy of B cell depletion therapies, which have seen moderate success, especially in immunotherapy. There have been encouraging results regarding their efficacy and safety in clinical trials entailing anti-CD20 Abs (rituximab, ofatumumab, and ocrelizumab), and anti–BAFF-R/BCMA/TACI/APRIL receptor agents (belimumab, telitacicept, and atacicept) (16), and anti-CD22 (17) toward the treatment of a variety of immunological disorders such as rheumatoid arthritis (RA) (18), relapsing multiple sclerosis (19), systemic lupus erythematosus (SLE) (20), neuromyelitis optica spectrum disorders (21), and pemphigus vulgaris (22).
Epratuzumab (EMAB) is the most studied anti-CD22 Ab that has reached a clinical trial to date, although it has been unable to meet the endpoint in phase III clinical trials. Nevertheless, it appears that there were observable biological activities arising from EMAB treatment. It was proposed that the mechanism of action (MOA) of EMAB worked by 1) the modulation of BCR activation via colocalization of CD22 to BCRs, leading to the triggering of intracellular phosphorylation events, followed by reductions in Ca2+ influx and decreased downstream inflammatory signaling and gene activation (23–25); 2) physical depletion of components of the BCR complex from the B cell surface through internalization (23); and 3) trogocytosis, in which immune-related B cell Ags such as CD22, CD19, CD21, and CD79b were reduced, leaving the resultant B cells with reduced numbers of surface Ags (26), with the last being the indicated key mechanism proposed for EMAB to modulate B cell activities in SLE. Because CD19 is overexpressed in the B cells of SLE patients, a reduction in surface CD19 on B cells after trogocytosis was regarded as the MOA responsible for the elicitation of a therapeutic response against SLE (26). Although CD19 is relevant toward BCR signaling (27), thus far there is no study indicating that the downregulation of CD19 expression on SLE B cells could be a viable MOA toward immunosuppression.
A careful look into the physiological function of CD22 has led us to believe that trogocytosis alone might not be sufficient in exerting the full biological/therapeutic potential of anti-CD22 Abs for treating autoimmune disease; moreover, based on the clinical results, this mechanism might not be sufficient in achieving significant therapeutic responses.
Abs that are capable of enhancing or restoring the biological function of CD22, entailing the discrimination between “self” from “non-self” via a unique MOA, could perhaps be therapeutically more relevant and useful in treating B cell–associated autoimmunity such as RA and SLE. In this study, we introduce a novel and unique MOA for SM03, a recombinant Ig (IgG1) mAb that selectively targets and binds CD22. SM03 is currently in a phase III clinical trial (NCT04312815) and has demonstrated therapeutic efficacies for treating RA in Chinese patients with a good safety profile in a proof-of-concept phase II study (28). In this study, we report that by binding to a unique epitope, SM03 is capable of restoring the normal immunomodulatory function of CD22 in patients with autoimmune disease without affecting the normal B cell immune response to pathogens.
Materials and Methods
CD22 recombinant protein construction and epitope mining
The CD22 domain 2–4 sequence and extracellular domain CD22 conjugated with murine Fc (CD22-muFc) were cloned into a pET bacterial expression vector preceded with a pelB sequence (29) followed by transformation into bacterial host cell BL21(DE3) pLysS (Promega, Madison, WI), and 1 mM isopropylthio-β-galactoside was added to induce expression. After incubation for 4 h at 37°C with orbital shaking at 250 rpm, the bacteria were pelleted by centrifugation at 6000 rpm for 15 min. Cells were lysed in lysis buffer (50 mM Tris-HCl [pH 8] containing 0.1 M NaCl, 5 mM EDTA, 0.1% NaN3, 0.5% Triton X-100, 100 mM PMSF, and 1 mM 10 DTT) with a glass homogenizer. French press treatment was used before 2.5 mM MgCl2 was added to chelate the EDTA. Then, 0.01 mg/ml DNase and 0.1 mg/ml lysozyme were added and the mixture was incubated at room temperature for 20 min. Inclusion body was pelleted by centrifugation at 6000 rpm for 15 min and washed three times. The pelleted inclusion bodies were dissolved in 6 M guanidine hydrochloride. CD22 domain 2–4 was engineered with a 15 His tag, and refolding and purification of the CD22 domain 2–4 were done using metal chelate affinity chromatography, whereas extracellular domain CD22-muFc refolding and purification were done using protein A (GE Healthcare, Chicago, IL), which was loaded with Ni Sepharose 6 fast flow resin (GE Healthcare, Chicago, IL) according to the manufacturer’s specifications.
The epitope mapping was conducted using Pepscan technologies. Briefly, to reconstruct linear epitopes of human CD22, a library of peptide-based epitope mimics was synthesized using solid-phase Fmoc synthesis (proprietary). By grafting the library with a proprietary hydrophilic polymer formulation (Pepscan, Lelystad, the Netherlands), and then a reaction with t-butyloxycarbonyl-hexamethylenediamine (BocHMDA) using dicyclohexylcarbodiimide (DCC) with N-hydroxybenzotriazole (HOBt) followed by successive cleavage of the Boc groups using trifluoroacetic acid (TFA), an amino functionalized polypropylene support was obtained. The synthesis of peptides on the amino functionalized solid support was through standard Fmoc peptide synthesis utilizing custom-modified JANUS liquid handling stations (PerkinElmer).
Binding of SM03 to the library of overlapping synthetic peptide on the solid support could be quantified with an automated ELISA-type readout. Briefly, the peptide arrays were incubated with SM03 overnight at 4°C. After washing, the peptide arrays were incubated with a 1:1000 dilution of goat anti-human (HRP) conjugate (SouthernBiotech, Birmingham, AL) for 1 h at 25°C. The samples were washed again before adding the peroxidase substrate 2,2′-azino-di-3-ethylbenzothiazoline sulfonate and 20 mg/ml 3% H2O2. After 1 h, a charge-coupled device (camera and an image processing system) was used to quantify the color development. The analysis was done using a box-and-whisker plot, linear intensity profiles, and heat map where the data values taken by a variable in a two-dimensional map were represented as colors for data visualization.
As under high-stringency conditions, screening of the anti-CD22 Ab SM03 did not yield detectable binding (data not shown). This suggested that SM03 would only bind to a discontinuous and conformational epitope. Pepscan’s Chemical Linkage of Peptides onto Scaffolds (CLIPS) technology (Pepscan, Lelystad, the Netherlands) was then employed on the human CD22 sequence for the elucidation of the possible conformational or discontinuous epitopes on human CD22 recognized by the SM03. In brief, the CD22 sequence was synthesized on a proprietary minicard and chemically converted into 25 spatially defined CLIPS constructs (Pepscan, Lelystad, the Netherlands). Synthesis of the CLIPS reaction takes place between bromo groups of the CLIPS scaffold and thiol side chains of cysteines introduced into a range of peptide structures mimicking conformational and discontinuous binding sites. CLIPS templates were coupled to cysteine residues, or side chains of multiple cysteines in the peptides were coupled to one or two CLIPS templates to generate CLIPS peptide libraries that could mimic secondary structure elements, such as loops, α helixes, and β strands.
Only constructs containing the right amino acid sequence in the correct conformation could best bind SM03. The CLIPS library contained constructs representing both parts of the discontinuous epitope in the correct conformation: constructs presenting the incomplete epitope would bind SM03 with lower affinity, whereas constructs not containing the epitope would not bind at all. The CLIPS template would bind to the solid-phase bound peptides of the peptide arrays (455 wells/plate with 3-ml wells) via the side chains of two cysteines. The peptide arrays were entirely covered in solution and gently shaken in the solution for 30–60 min. Finally, using excess H2O, the peptide arrays were washed thoroughly and sonicated in disrupt-buffer containing 1% SDS/0.1% 2,2′-(ethylenedioxy)diethanethiol in PBS (pH 7.2) at 70°C for 30 min, before sonication in H2O for another 45 min. Intensity profiles were recorded with combinatorial epitope mimics.
Kinetic analysis by biomolecular interaction analyzer ForteBio Octet RED96e
Kinetics analysis using the biomolecular interaction analyzer ForteBio Octet RED96e (ForteBio) was performed by coupling purified SM03 Abs to anti-human IgG Fc capture (AHC) dip and read biosensors (ForteBio) using standard amine chemistry. First, biosensors were equilibrated for 10 min in kinetics buffer (PBS plus 0.02% Tween 20, 0.1% BSA, and 0.05% sodium azide) and microplates were filled with 200 µl of sample in kinetic buffer and agitated at 1000 rpm. Association of recombinant CD22-His tag (Sino Biological) was observed by immersing biosensors in wells containing decreasing titrations of 32.6, 10.8, 3.66, 1.22, 0.40, and 0.13 nM of CD22-His tag for 900 s with 1000 rpm agitation and monitoring interferometry. Dissociation was measured after transfer of the biosensors into kinetic buffer for 900 s at 1000 s of agitation. The observed on/off rates (kon/koff) were analyzed using a 1:1 binding global fit model, the equilibrium binding constant, KD, was calculated using ForteBio Octet analysis software (ForteBio), and the sensorgram was presented using Prism (GraphPad Software).
CD22 α2,6-linked sialic acid binding ELISA
The ability of SM03, EMAB, and m971 to block CD22 α2,6-linked sialic acid binding was assessed by ELISA. First, α2,6- and α2,3-linked sialic acids (GlycoNZ) were coated onto 96-well polystyrene strips (Costar, 2580) at 5 µg/ml in PBS (pH 7.4) at 4°C overnight. CD22-muFc was pretreated at 4 µg/ml with Ab in 1% BSA (Millipore) in PBS at 4°C overnight. The coated strips were blocked with 1% BSA in PBS for 1 h at room temperature (RT), and the pretreatment was then incubated with the strips for 1 h at RT. CD22-muFc was detected by anti–murine Fc-HRP (Jackson ImmunoResearch) at 1:6000 in 1% BSA in PBS. Colorimetric signal was generated with tetramethylbenzidine single solution (Life Technologies) and stopped with 2 N sulfuric acid. Measurements were obtained with a Sunrise absorbance microplate reader (Tecan).
Ab internalization
The amounts of Abs internalized at time points 0, 2.5, 5, 10, and 20 min were measured by flow cytometry. SM03, EMAB, and control IgG (HG1K IgG1, Sino Biological) were labeled with an EZLabel protein FITC labeling kit (BioVision) per manufacturer’s instructions. Next, 2 × 105 cells were treated on ice with 10 µg/ml FITC-labeled Abs in wash buffer (2% FBS [Life Technologies] in PBS) for 1 h. Excess Ab was washed with wash buffer, and cells were incubated in a 37°C water bath for time points 0, 2.5, 5, 10, and 20 min and immediately placed back on ice. Cells were divided into PBS wash and acid wash groups. For the PBS wash group, cells were washed with wash buffer before flow cytometry. For the acid wash group, cells were incubated in 0.133 M citric acid (Sigma-Aldrich) and 0.06 M NaHCO3 (Sigma-Aldrich) for 4 min at RT. Cells were subsequently neutralized by addition of 19 vol of wash buffer and then washed once more prior to flow cytometry. Flow cytometry was carried out on BD FACSLyric flow cytometry system (BD Biosciences) with gating strategy demonstrated in Fig. 2B (top), and analysis was performed with FlowJo.
CD22 internalization
RAMOS cells (3 × 105) were treated on ice with 10 µg/ml SM03 or IgG isotype control in wash buffer for 1 h. Excess Ab was washed with wash buffer, and cells were incubated in a 37°C water bath for time points 0, 5, 10, and 20 min and immediately placed back on ice. Cells were then incubated in 0.133 M citric acid (Sigma-Aldrich) and 0.06 M NaHCO3 (Sigma-Aldrich) for 4 min at RT. Cells were subsequently neutralized by addition of 19 vol of wash buffer, then washed once more before staining with anti-CD22 Ab S-HCL-1 with allophycocyanin conjugate (Life Technologies) at 5 μl per test for 1 h on ice prior to flow cytometry. Flow cytometry was carried out on a BD FACSLyric flow cytometry system (BD Biosciences), and analysis was performed with FlowJo.
CD22 internalization and trans ligand binding with immunocytochemistry
The effect of CD22 internalization induced by SM03, EMAB, and isotype control (HG1K IgG1, Sino Biological) on CD22 trans binding to α2,6-linked sialic acids was visualized by immunocytochemistry. Cells (1 × 105) were treated on ice with 10 µg/ml Ab in 2% FBS in PBS for 1 h. Excess Ab was washed and cells were incubated in a 37°C water bath as for flow cytometry. After incubation in a water bath, cells were washed with PBS (pH 7.4) and coated onto 12-mm round glass coverslips (Warner Instruments) on ice at 1 × 106 cells/ml. Cells were incubated on ice for 1 h. The cells were then fixed with 4% paraformaldehyde in PBS (Sigma-Aldrich) for 10 min at RT onto Superfrost plus microscope slides (Thermo Fischer Scientific) according to Tsang et al. (30). Paraformaldehyde was washed off with PBS (pH 7.4), and coverslips were blocked with 1% BSA in PBS for 1 h at RT. Cells were stained with Neu5Acα2–6Galβ1–4Glcβ-sp2-PAA-fluo (FITC-conjugated α2,6-linked sialic acid GlycoNZ) at 5 µg/ml at 4°C overnight. The next day, Abs were detected by anti-human IgG conjugated with Alexa Fluor 647 (Jackson ImmunoResearch) at 1:500 in 1% BSA in PBS. Coverslips were mounted onto Superfrost microscope slides (Thermo Scientific) with Duolink in situ mounting medium with DAPI (Sigma-Aldrich). Cells were visualized using an LSM 880 confocal microscope (Carl Zeiss) using a ×60 oil immersion objective lens. Images were adjusted and normalized using ImageJ software.
CFSE analysis
Human PBMCs (ZenBio) were used for B cell purification using the B cell isolation kit II (human; Miltenyi Biotec) according to the manufacturer’s instructions. B cells were stained with a CFSE cell division tracker kit (BioLegend) according to the manufacturer’s instructions before 1 × 106 CFSE-stained B cells were seeded to six-well plates. Cells were then treated with anti-IgM (Jackson ImmunoResearch) at 2 μg/ml and SM03 or IgG (HG1K IgG1, Sino Biological) at 1 μg/ml for 72 h before staining with a Zombie NIR fixable viability kit (BioLegend) before washing three times with 2% FBS in PBS and run using flow cytometry. Flow cytometry was carried out on a BD FACSLyric flow cytometry system (BD Biosciences) with gating strategy demonstrated in Fig. 3A (top), and analysis was performed with FlowJo.
RAMOS NF-κB reporter cell line analysis
An NF-κB luciferase vector was constructed by cloning an NF-κB consensus transcriptional response element upstream of the minimal promoter region of pNL3.3[secNluc/minP] (Promega). RAMOS cells (7.5 × 106) were transfected via electroporation with 5 μg/ml vector DNA before seeding to 96-well plates at 5 × 105 cells per well. Cells were cotreated with anti-IgM (Jackson ImmunoResearch) at 2 μg/ml or LPS (PeproTech) at 1 μg/ml SM03, IgG (HG1K IgG1, Sino Biological), or SM03 Fab at increasing concentrations for 24 h before measuring using a Nano-Glo luciferase assay system (Promega) according to the manufacturer’s conditions and using a multimode microplate reader (Varioskan Flash, Thermo Scientific).
Cell treatment for western blot
RAMOS cells (American Type Culture Collection) cultured in RPMI 1640 (Life Technologies) with 10% FBS (Life Technologies) were seeded at 1 × 106 to 24-well plates, and cells were cotreated with anti-IgM (Jackson ImmunoResearch) at 4 µg/ml, Neu5Acα2–6Galβ1–4Glcβ-sp2-PAA (GlycoNZ) at 0.25 µg/ml, and SM03, SM03-F(ab′)2, and isotype control (as mentioned above) at 10 μg/ml for 1 h. Cells were lysed with RIPA buffer (50 mM Tris Cl [pH 7.4] [Sigma-Aldrich], 150 mM NaCl [Sigma-Aldrich], 5 mM EDTA [Sigma-Aldrich], 1% Triton X-100 [Sigma-Aldrich], 1% sodium deoxycholate [Sigma-Aldrich], 0.1% SDS [Sigma-Aldrich]) with 1× Halt protease and phosphatase inhibitor cocktail (Thermo Scientific). Samples were boiled in NuPAGE LDS sample buffer (Invitrogen) with 5% 2-ME (Bio-Rad) before running on SDS-PAGE and Western blot (Bio-Rad). SHP-1, p–SHP-1 (Cell Signaling Technology), and β-tubulin (Thermo Scientific) at 1:1000 was used to probe, and anti-rabbit secondary Ab with HRP conjugate or anti-mouse secondary Ab with Alexa Fluor 488 (Jackson ImmunoResearch) at 1:5000 was used for the secondary probe. Pierce ECL Western blotting substrate (Thermo Scientific) was used to show chemiluminescent signals for p–SHP-1 and SHP-1 (Supplemental Fig. 1). Detection of both chemiluminescent signal and fluorescent signal was done by using a gel documentation system (Bio-Rad ChemiDoc MP). Analysis was conducted using Image Lab software (Bio-Rad).
Clinical trial patient selection, intervention method, and assessment method
The clinical trial registered as NCT04192617 was conducted between December 31, 2014 and February 3, 2016 as a 24-wk phase II randomized, double-blind, multi-dose, placebo-controlled study across 11 medical centers in China. RA patients enrolled were at minimum 18 y old, at least 12 mo since RA onset, and fulfilled the classification criteria for RA according to the American College of Rheumatology (ACR) (31). Inclusion criteria of patients with active disease identified were as follows: 1) 8 out of 66 swollen joints, 2) 8 out of 68 tender joints, 3) erythrocyte sedimentation rate (Westergren method) ≥28 mm/h, 4) C-reactive protein >6 mg/l, or 5) morning stiffness ≥45 min. Exclusion criteria were patients with 1) a history of inflammatory joint disease other than RA, 2) prior exposure to disease-modifying antirheumatic drugs other than MTX within 4 wk of study entry, 3) known allergy to Abs of murine or human origin, 4) receipt of therapy with other mAbs 6 mo before enrollment, and 5) pregnant or lactating women.
The clinical trial complied with the SPIRIT statement, as well as reviews by independent committees and Institutional Review Boards of participating hospitals (32). The study was conducted adhering to the International Conference on Harmonization Guidelines for Good Clinical Practice and the Declaration of Helsinki and was reported conforming to the CONSORT statement (33).
A validated central randomization tool, DAS for IWRS, was used to generate centralized randomization service and a randomization schedule enacted for the study. Patients must have received MTX (≥7.5 mg/wk) continuously for ≥12 wk and a stable dose between ≥7.5 and ≤20 mg/wk for ≥4 wk prior to intervention with SM03 with stable MTX background. Moreover, nonsteroidal anti-inflammatory drugs and low-potency analgesics were allowed when clinically needed, with a dosage ≤10 mg/d prednisone (or equivalent) and stable within 4 wk prior to randomization. Disease-modifying antirheumatic drugs other than MTX, systemic immune inhibitors, biologics, and cytotoxic drugs were not allowed 4 wk prior to the randomization and were prohibited throughout the study.
Intervention, as described in Results, were administered at random to three groups entailing: placebo (the control group), 600 mg of SM03 at weeks 0, 2, 12, and 14 (the low-dose group), and 600 mg of SM03 at weeks 0, 2, 4, 12, 14, and 16 (the high-dose group) in a ratio of 1:1:1.
Patients were recorded regularly for vital signs, 12-lead electrocardiogram, and adverse events (AEs). Samples for laboratory tests such as blood routine, erythrocyte sedimentation rate, rheumatoid factor, urine routine, liver function, and kidney function were obtained at weeks 0, 2, 4, 8, 12, 16, and 24. Physician’s global assessment and patient’s global assessment of pain were recorded using the visual analog scale. Disease activity score using 28 joint counts (DAS28) and Health Assessment Questionnaire–Disability Index (range 0–3, with higher scores indicating greater disability) were also recorded (33).
Safety assessments were performed during scheduled visits. Throughout the study, the incidence and severity of infusion-related reactions, AEs, and treatment-emergent AEs were monitored, where AE coding was performed according to the Medical Dictionary for Regulatory Activities (MedDRA), version 18.1. Safety assessments were based mainly on the occurrence, frequency, and severity of AEs. When necessary, patients with AEs were withdrawn from the study. All patients who received at least one dose of the study drug and had at least one follow-up were included in the safety assessments results.
Statistical analysis of clinical results
The statistical analyses of efficacy were prespecified and adhered to the intention-to-treat approach. The full analysis set contained patients who were randomly assigned, received study medication, and had the study medication’s efficacy evaluated.
The per protocol set included patients who completed the treatment and underwent efficacy evaluation, met the eligibility criteria and did not have a major violation of the study protocol, and showed good compliance. The last observation carried forward was used for the efficacy endpoints for patients who dropped out of or discontinued the study.
Mean, SD, median, and 95% confidence intervals or range were used to express continuous variables. Frequencies and percentages were used to express categorical variables. Statistical analysis was done using SAS 9.3 (SAS Institute, Cary, NC). Demographic and baseline variables of the two groups were analyzed using ANOVA or a χ2 test. The efficacy endpoints were analyzed using Fisher exact tests or a χ2 test. Temporal changes in clinical parameters, including morning stiff time, tender and swollen joint counts, and visual analog scores, were analyzed by ANOVA for the multiple group comparison, and to determine whether p values were ≤0.05, a least significant difference t test was performed. A χ2 test or Fisher exact test was used for comparing the incidence of AEs between the two groups in the analysis of safety events. All statistical tests were two-sided and α was set at 0.05, and p values ≤0.05 were considered statistically significant.
Statistical analyses of in vitro assays
Statistical analyses were performed using and GraphPad Prism version 7 (GraphPad Software). The respective tests used are indicated in the figure legends.
Results
SM03 binds to an epitope distinct from that of EMAB and prevents the binding of CD22 to α2,6-linked sialic acid binding
SM03 targets the epitope residing in the domain 2–4 of human CD22 (Fig. 1A); furthermore, SM03 demonstrated dose-responsive high binding affinity to truncated CD22 that only expressed CD22 domain 2–4 (Fig. 1A). Moreover, SM03 binds to a discontinuous conformational epitope within domain 2 of the human CD22 Ag containing the sequence (Fig. 1A, 1B). SM03 recognizes a discontinuous epitope on domain 2 composed of peptide stretches that are separated at the level of primary structure (Fig. 1A, 1B). The epitopes 161CLLNFSCYGYPIQ173 (red) and 198VFTRSELKFSPQWSHHGKIVTC219 (blue) are shown to come together in three-dimensional space, as seen in the three-dimensional rendering of the binding epitopes on CD22 domain 1–3 (Fig. 1A, 1B). This differs from EMAB, as studies indicate the epitope of EMAB to be near N231 at domain 2–3 (5). Moreover, binding affinity of both Abs were compared using the Octet analysis system where SM03 and EMAB were loaded at 20 μg/ml and CD22 (extracellular domain)-His tag protein was associated with either Ab at titrated concentrations (Fig. 1C). Results indicated SM03 and EMAB to have a KD of 0.48 and 1.92 nM, respectively (Fig. 1C). Moreover, SM03 partially blocked the binding between CD22 and its ligand α2,6-linked sialic acid, where both were able to block α2,6-linked sialic acid significantly more than EMAB (dose-dependent) or isotype control M971, a CD22-binding Ab that binds to a different domain than SM03 or EMAB (34) (Fig. 1D).
SM03 showed rapid internalization and facilitated trans binding of CD22 to α2,6-linked sialic acid glycans
SM03 internalized rapidly upon binding to CD22 across three B cell lymphoma cell lines, that is, RAMOS, RAJI, and DAUDI cells (Fig. 2A, 2E). We found that upon binding, SM03 induced CD22 internalization; indeed, we observed this effect in as short as 5 min of internalization (Fig. 2E). Using the flow cytometry analysis protocol described by O’Reilly et al. (15), ∼40% out of all of the SM03-bound CD22 was internalized at 10 min (Fig. 2B). SM03 internalized faster than EMAB, where internalization could be observed as early as 2.5 min (Fig. 2B). In 20 min, SM03 showed a 60–80% internalization compared with EMAB, which had only 30% internalization in the B cell lymphoma cell line RAMOS (Fig. 2B). To confirm that SM03 binding enhances CD22 internalization, the rates of reduction of surface CD22 levels upon pretreatment with SM03, IgG, or control were measured on RAMOS cells: significantly higher levels of surface CD22 reduction were observed in B lymphoma cells pretreated with SM03 compared with those pretreated with IgG (isotype control) or control (Fig. 2C). We then hypothesized that SM03 could either block or disrupt the cis-ligand binding to promote trans-ligand binding between CD22 and α2,6-linked sialic acid (situated in autologous cells), although initially, SM03 upon binding to CD22 was unable to promote trans binding of CD22 to α2,6-linked sialic acid (Fig. 2D, 2E). SM03 facilitated more trans binding to α2,6-linked sialic acid after internalization, as opposed to EMAB or IgG isotype control where little to no trans binding was observed with the latter two regardless of internalization (Fig. 2D).
SM03 induced downstream modulation of BCR- and TLR4-induced B cell activation
To delineate whether the immunomodulatory effects could be seen in nonlymphoma B cells, we isolated B cells from PBMCs of healthy human samples (donors) purchased from Zen-Bio via negative selection and analyzed the ability of SM03 to modulate BCR-induced proliferation; indeed, we observed a significant decrease in proliferation upon SM03 treatment (Fig. 3A). NF-κB is a vital transcription factor that promotes autoimmunity and inflammation by mediating the activation and differentiation of autoimmune and inflammatory T cells. Utilizing an NF-κB reporter B cell lymphoma cell line (35), SM03 Fab was shown to decrease NF-κB signaling and transfection in a dose-dependent manner (Fig. 3B). Other than BCR activation, SM03 was also seen to modulate TLR4 receptor–induced B cell activation (36, 37), and SM03 significantly decreased the TLR4-induced NF-κB response by LPS ligand induction in the NF-κB reporter cell line (Fig. 3C), demonstrating that SM03 is able to modulate the innate response where TLR4 activation has been associated with RA (37).
We then hypothesized that the mechanism of facilitating cis breaking and trans binding between CD22 and α2,6-linked sialic acid by SM03 could influence CD22 immunomodulation through downstream signaling. Upon BCR activation of cells by anti-IgM, SHP-1, a downstream signaling molecule of CD22, was phosphorylated (38) (Fig. 3D, Supplemental Fig. 1) This is concurrent with intrinsic CD22 modulation promoted by BCR activation (39). Although SM03 by itself did not upregulate SHP-1 phosphorylation, upon cotreatment with an α2,6-linked sialic acid probe, we found a significantly heightened phosphorylation of SHP-1 than control (Fig. 3D, Supplemental Fig. 1). The increase of SHP-1 phosphorylation was unlikely to have resulted from the engagement of the SM03 constant Fc region with FcγRIIA and/or FcRN receptors (Fig. 3D), as similar levels of SHP-1 activation were observed when RAMOS cells were cotreated with pepsin-digested SM03 F(ab′)2 fragments (which do not carry Fc regions) and α2,6-linked sialic acid probe.
Summary of the clinical responses in RA and SLE patients
Thus far, a total of five clinical trials have been conducted and completed for SM03 mAb on the indications of non-Hodgkin’s lymphoma (NHL), SLE, and RA. A total of 228 patients (1242 person-times) received an SM03 i.v. drip at the dose ranging from 60 to 480 mg/m2 (NHL) and from 400 to 600 to 900 mg (SLE and RA), and the observational period ranged from 6 wk (NHL phase I) to 12 or 24 wk (SLE and RA, phase I/II) to 52 wk (NHL phase IIa). Optimal clinical benefits were observed at the dose above 360 mg/m2 and 600 mg. The phase III randomized, double-blind, placebo-controlled clinical trial is ongoing at present for SM03 in the treatment of moderate-to-severe active RA (SM03-RAIII, NCT04312815) (Table I).
Identification No. . | Stage . | Indications . | First Visit of First Patient . | Dose/Exposure . |
---|---|---|---|---|
SM03-NHL-I- V3.3 | I | Non-Hodgkin’s lymphoma | April 20, 2007 | 60–120–240–360–480 mg/m2, 21 cases |
CTR20131130 | IIa | Non-Hodgkin’s lymphoma | July 4, 2012 | 360, 480 mg/m2, 15 cases |
CTR20130117 | I | Systemic lupus erythematosus | July 24, 2011 | 240 mg/m2, 600–900 mg, 29 cases |
CTR20131127 | I | Rheumatoid arthritis | August 14, 2012 | 600 mg twice, 8 cases |
CTR20140856 | II | Rheumatoid arthritis | December 31, 2014 | 0, 600 mg twice, 600 mg three times, 156 cases |
A total of 228 subjects were exposed |
Identification No. . | Stage . | Indications . | First Visit of First Patient . | Dose/Exposure . |
---|---|---|---|---|
SM03-NHL-I- V3.3 | I | Non-Hodgkin’s lymphoma | April 20, 2007 | 60–120–240–360–480 mg/m2, 21 cases |
CTR20131130 | IIa | Non-Hodgkin’s lymphoma | July 4, 2012 | 360, 480 mg/m2, 15 cases |
CTR20130117 | I | Systemic lupus erythematosus | July 24, 2011 | 240 mg/m2, 600–900 mg, 29 cases |
CTR20131127 | I | Rheumatoid arthritis | August 14, 2012 | 600 mg twice, 8 cases |
CTR20140856 | II | Rheumatoid arthritis | December 31, 2014 | 0, 600 mg twice, 600 mg three times, 156 cases |
A total of 228 subjects were exposed |
We previously reported a potential ameliorating effect of SM03 on SLE in an open-label, multiple-center, multiple–ascending dose, phase I study in 29 SLE patients, in which we observed a decrease in the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) score and the British Isles Lupus Assessment Group index in five patients at 17.8 and 17.2%, respectively (40). In a phase I, open-label, single-center, multiple-dose study of SM03 injection given at doses of 600 mg (∼360 mg/m2) for 2 wk in eight patients with active RA, we reported that four out of eight patients reached ACR criteria ACR20 within 12 wk, whereas one patient reached ACR70 and a European League Against Rheumatism good response; furthermore, these results were consistent with DAS28 data. In the latest 24-wk phase II randomized, double-blind, multi-dose, placebo-controlled clinical trial to study efficacy and safety of SM03 in moderately-to-severely active RA patients in China (28), 156 patients on background methotrexate were treated with a cumulative dose of 3600 mg (high dose, 600 mg for six infusions at weeks 0, 2, 4, 12, 14, and 16) or 2400 mg SM03 (low dose, 600 mg for four infusions at weeks 0, 2, 12, and 14), or placebo (Table II). A significant increase in patients who reached ACR20, ACR50, and ACR70 was observed with both low and high dose compared with placebo at 24 wk, whereas a high dose also demonstrated a significantly better European League Against Rheumatism and DAS28 response (28). In particular, components of the ACR criteria, including tender joint count, swollen joint count, physician’s global assessment, patient’s global assessment, patient’s assessment of pain, and the Health Assessment Questionnaire–Disability Index, displayed significant amelioration of disease at both 12 and 24 wk of treatment (Table II).
Component (median) . | Placebo (n = 47) . | High Dose (600 mg three times) (n = 49) . | Low Dose (600 mg twice) (n = 51) . | |||
---|---|---|---|---|---|---|
Baseline . | 24 wk (changea) . | Baseline . | 24 wk (changea) . | Baseline . | 24 wk (changea) . | |
Tender joint count (0–68) | 24 | 14 (−10) | 21 | 6 (−15) p = 0.01, versus Plc | 28 | 15 (−7) p = 0.03, versus Plc |
Swollen joint count (0–66) | 20 | 9 (−11) | 17 | 2 (−15) p < 0.001, versus Plc | 18 | 3 (−15) p < 0.001, versus Plc |
Physician’s global assessment of disease activity (VAS) | 7 | 5 (−2) | 7 | 3 (−4) p < 0.001, versus Plc | 7 | 4.8 (−2.2) p = 0.006, versus Plc |
Patient’s global assessment of disease activity (VAS) | 7 | 5 (−2) | 7 | 3.9 (−3.1) p < 0.001, versus Plc | 7.2 | 4.5 (−2.7) p = 0.002, versus Plc |
Patient’s assessment of pain (VAS) | 7 | 5 (−2) | 7 | 4 (−3) p < 0.001, versus Plc | 7 | 4.5 (−2.5) p = 0.003, versus Plc |
HAQ-DI | 24 | 15 (−9) | 24 | 6 (−18) p < 0.001, versus Plc | 24 | 10 (−14) p = 0.038, versus Plc |
C-reactive protein level (mg/l) | 21.1 | 14 (−6.9) | 22.3 | 11.9 (−10.4) | 25.5 | 14.2 (−11.3) |
Component (median) . | Placebo (n = 47) . | High Dose (600 mg three times) (n = 49) . | Low Dose (600 mg twice) (n = 51) . | |||
---|---|---|---|---|---|---|
Baseline . | 24 wk (changea) . | Baseline . | 24 wk (changea) . | Baseline . | 24 wk (changea) . | |
Tender joint count (0–68) | 24 | 14 (−10) | 21 | 6 (−15) p = 0.01, versus Plc | 28 | 15 (−7) p = 0.03, versus Plc |
Swollen joint count (0–66) | 20 | 9 (−11) | 17 | 2 (−15) p < 0.001, versus Plc | 18 | 3 (−15) p < 0.001, versus Plc |
Physician’s global assessment of disease activity (VAS) | 7 | 5 (−2) | 7 | 3 (−4) p < 0.001, versus Plc | 7 | 4.8 (−2.2) p = 0.006, versus Plc |
Patient’s global assessment of disease activity (VAS) | 7 | 5 (−2) | 7 | 3.9 (−3.1) p < 0.001, versus Plc | 7.2 | 4.5 (−2.7) p = 0.002, versus Plc |
Patient’s assessment of pain (VAS) | 7 | 5 (−2) | 7 | 4 (−3) p < 0.001, versus Plc | 7 | 4.5 (−2.5) p = 0.003, versus Plc |
HAQ-DI | 24 | 15 (−9) | 24 | 6 (−18) p < 0.001, versus Plc | 24 | 10 (−14) p = 0.038, versus Plc |
C-reactive protein level (mg/l) | 21.1 | 14 (−6.9) | 22.3 | 11.9 (−10.4) | 25.5 | 14.2 (−11.3) |
Source: NCT04192617 clinical study report.
Mean change from baseline at week 24.
HAQ-DI, Health Assessment Questionnaire–Disability Index; Plc, placebo; VAS, visual analog scale.
Although the NHL phase I clinical trial had several patients who were withdrawn from the study due to AEs (16.7%), other clinical trials investigating autoimmune diseases (phase I SLE and phase II RA) had low numbers of dropouts (3.4% or one patient and 1.8% or two patients, respectively) (41) (Table III). Drug-related AEs were at rates of 13.9, 20.7, and 12.6% for NHL, SLE, and RA, respectively. Serious AEs were higher in NHL at 8.3%, but very low for SLE and RA at 3.4 and 0.9%, respectively. Most AEs for SM03 in the treatment for SLE and RA are infections (17.2 and 13.6%, respectively), as well as elevated transaminase and decreases in neutrophil and platelet numbers. SLE patients were also observed to have the AEs of fever, elevated urine protein and IgG, as well as prolonged coagulation time. Overall, we report good efficacy and safety data in clinical trials of SM03 in a phase I clinical trial in RA and SLE patients as well as a phase II clinical trial in RA patients (28, 40).
Disease Indication . | Non-Hodgkin's Lymphoma . | Systemic Lupus Erythematosus . | Rheumatoid Arthritis . |
---|---|---|---|
Clinical trial phase | Phase I (n = 36) | Phase I (n = 29) | Phase II (n = 103) |
Treatment-emergent AEs, n (%) | 24 (68) | 14 (48.3) | 45 (43.7) |
Treatment-related AEs, n (%) | 5 (13.9) | 6 (20.7) | 13 (12.6) |
Discontinuation due to AEs, n (%) | 6 (16.7) | 1 (3.4) | 2 (1.8) |
SAEs n (%) | 3 (8.3) | 1 (3.4) | 1 (0.9) |
Infusion reaction, n (%) | — | — | 1 (0.9) |
AE descriptions | Most AEs were grade 1–2; the high-frequency AEs included fever, reduction of WBCs and neutrophils, elevated transaminase, tiredness, reduced platelet numbers, decreased hemoglobin, abnormal coagulation factors, increased lactate dehydrogenase, increased creatinine, increased urine WBCs and or RBCs, and tumor progression | Most AEs were mild and moderate; the most common AEs included infection and various laboratory abnormalities, as well as fever, elevated transaminases, decreased neutrophils and platelets, elevated urine protein and IgG, and prolonged coagulation time | Most AEs were mild and moderate; the most common AEs included infection and various laboratory abnormalities, as well as infection, increased or decreased neutrophils, increased transaminase, and reduced platelet numbers |
Disease Indication . | Non-Hodgkin's Lymphoma . | Systemic Lupus Erythematosus . | Rheumatoid Arthritis . |
---|---|---|---|
Clinical trial phase | Phase I (n = 36) | Phase I (n = 29) | Phase II (n = 103) |
Treatment-emergent AEs, n (%) | 24 (68) | 14 (48.3) | 45 (43.7) |
Treatment-related AEs, n (%) | 5 (13.9) | 6 (20.7) | 13 (12.6) |
Discontinuation due to AEs, n (%) | 6 (16.7) | 1 (3.4) | 2 (1.8) |
SAEs n (%) | 3 (8.3) | 1 (3.4) | 1 (0.9) |
Infusion reaction, n (%) | — | — | 1 (0.9) |
AE descriptions | Most AEs were grade 1–2; the high-frequency AEs included fever, reduction of WBCs and neutrophils, elevated transaminase, tiredness, reduced platelet numbers, decreased hemoglobin, abnormal coagulation factors, increased lactate dehydrogenase, increased creatinine, increased urine WBCs and or RBCs, and tumor progression | Most AEs were mild and moderate; the most common AEs included infection and various laboratory abnormalities, as well as fever, elevated transaminases, decreased neutrophils and platelets, elevated urine protein and IgG, and prolonged coagulation time | Most AEs were mild and moderate; the most common AEs included infection and various laboratory abnormalities, as well as infection, increased or decreased neutrophils, increased transaminase, and reduced platelet numbers |
–, no patients; AE, adverse event; SAE, severe adverse event.
Discussion
In many autoimmune diseases, a deregulation of immune tolerance is often observed. Treatment by targeting the regulation of B cell tolerance to self-antigens is of particular interest due to the prevalence of autoantibodies directly contributing to the pathology of many inflammatory diseases (42–44). The BCR is vital for maintaining the mechanism of regulating B cell tolerance, including receptor editing, deletion, and anergy, and thus CD22, a regulator of BCR, would be a favorable target (45). Furthermore, studies have shown that deletion of CD22 influences the susceptibility to develop autoantibodies that are prevalent in autoimmune diseases (46, 47).
EMAB had shown good efficacy in phase I and II clinical trials of treatment against SLE (initial open label, ALLEVIATE-1, ALLEVIATE-2, and EMBLEM) (48), indicative of clinically meaningful improvements in patients with moderate-to-severe SLE as measured by the British Isles Lupus Assessment Group–based primary endpoint and partial depletion of the circulating B cell population (34). Moreover, EMAB treatment also led to improvements in health-related quality of life and a reduction in mean glucocorticoid dose (49). In the EMBLEM phase IIb randomized controlled trial (RCT) study investigating the efficacy and safety of various EMAB doses in a 12-wk dosing regimen at 2400 mg cumulative dose (cd) (600 mg weekly), 2400 mg cd (1200 mg every other week), and combined 2400 mg cd groups in moderate to severe SLE, statistically significant improvement in both EMAB 2400 mg cd groups compared with placebo was observed using the British Isles Lupus Assessment Group–based combined lupus assessment (BICLA) index as a composite endpoint (50, 51).
However, two follow-up phase III RCT studies (EMBODY 1 and EMBODY 2) for EMAB treatment against moderate to severe SLE failed to meet the primary endpoint; that is, patients treated with the two different dose regimens of EMAB at 600 mg weekly (2400 mg cd) and 1200 mg every other week (2400 mg cd) for 12 wk were not significantly different compared with placebo (52). Recently, a post hoc analysis of the aforementioned phase III trials has implicated a hypothesis that in anti-SSA–positive patients with lupus and features of Sjögren’s syndrome, EMAB treatment showed improvement in SLE disease activity (53), whereas the pathogenesis of Sjögren’s syndrome is also associated with increased B cell activation that led to tissue infiltration and destruction (54). Postulates on other factors that might have contributed to the inability to meet the endpoints in phase III studies include the following: a suboptimal dosing regimen, an unaccounted high background response rate in placebo receiving standard of care of corticosteroids and immunosuppressants, a high rate of dropouts, as well as potential heterogeneity of the patient population that were unresponsive to EMAB (48, 55). Overall, EMAB appears to have insufficient strength to immunomodulate SLE in patients.
Although discouraging, the clinical trial results from EMBODY 1 and EMBODY 2 do not necessarily refute the therapeutic potential of anti-CD22 Abs against immunological disorders. Anti-CD22 Abs that bind to alternate epitopes of EMAB, and probably via a different MOA other than the proposed “trogocytosis,” might have other clinical applications. An example could be seen with anti-CD20 Ab biologics, for example, rituximab, ofatumumab, and ocrelizumab, which each bind to distinct epitopes of CD20, yet have demonstrated clinical applications on different disease indications (56–59).
CD22, a coreceptor of BCR, functions to regulate detrimental autoantigen signaling through the ability to differentiate between Ags on pathogens (“non-self”) or autologous (“self”) cells that express α2,6-linked sialic acid ligands, where CD22 would bind to the latter and modulate unwanted BCR signaling (3, 38, 60). CD22 was previously demonstrated to exist in preformed homomultimeric nanoclusters, where the formation is dependent on cis binding of α2,6-linked sialic acid ligand on neighboring CD22 molecules (14). A masking effect within the nanocluster of CD22 would restrict modulation of the BCR. On the other hand, when autoantigens on autologous cells (self) were engaged with the BCR, preferential trans-ligand binding of CD22 to the α2,6-linked sialic acid ligand of the autologous cell would impart an increased immunomodulation of B cell activation, which could lead to immune tolerance against the autoantigen (9–12). A skewed balance between cis-ligand and trans-ligand binding of CD22 on B cells will affect the immunomodulatory function of CD22, leading to autoimmunity. Indeed, deregulation of CD22 α2,6-linked sialic acid ligand binding by impairing sialic acid O-acetyl esterase (SIAE), which functions to enable CD22 sialic acid binding, was found to be associated with autoimmune disorders (61). Moreover, transgenic mice with SIAE deletions are prone to develop SLE-like autoimmune diseases, highlighting the important role played by CD22 and α2,6-linked sialic acid ligand binding in B cell immunomodulatory functions (12, 61, 62).
SM03 is demonstrated to bind with high affinity and specifically to a conformational epitope of CD22 at domain 2 that is completely different to that of EMAB (Fig. 1A, 1B) (5). Furthermore, SM03 demonstrated a 4-fold stronger binding affinity than EMAB (Fig. 1C). We also found that SM03 was able to partially block α2,6-linked sialic acid binding to CD22 (Fig. 1D) as compared with EMAB, which did not block but rather increased α2,6-linked sialic acid binding to CD22, implicating a completely different binding mechanism between SM03 and EMAB. Although the binding epitope does not overlap with the α2,6-linked sialic acid binding site, it is likely that the proximity of the epitope of SM03 to the α2,6-linked sialic acid binding site is sufficient to impart steric hindrance for CD22 cis ligation (5).
SM03 induced a rapid rate of internalization of the CD22 molecule across two human B lymphoma cell lines upon binding (Fig. 2A, 2B, 2E). The rate of internalization for SM03 was two to three times as fast as that of EMAB (Fig. 2B). Furthermore, we have confirmed that binding of SM03 directly influences the rate of internalization of CD22 (Fig. 2C). Enhanced endocytosis of CD22 will not only bring the cis-ligated CD22 into the endosome, the low pH environment of which would free up α2,6-linked sialic acid engagement, thereby disrupting the inactive homomultimeric cluster and allowing the resurfaced CD22 (via recycling) to engage with the trans ligand (15, 63), but, moreover, enhanced endocytosis of CD22 could potentially reroute SM03 to target the immunomodulatory endosomal TLRs (including TLR3, TLR7, and TLR9), which have been indicated in the pathology of SLE and other autoimmune diseases (64).
SM03 induced CD22 trans-ligand (α2,6-linked sialic acid) binding in human B lymphoma cell lines, whereas EMAB failed to do so (Fig. 2D). We hypothesized that in autoimmune patients, B cells that target autoantigens of autologous cells (with surface expression of CD22 ligands) exhibited a defective CD22 cis-ligand to trans-ligand binding conversion, where under normal circumstances the trans-ligand bound CD22 would suppress the BCR-elicited immune response against the autologous cell. Binding of SM03 would facilitate this cis-to-trans conversion, allowing more CD22 to engage with ligand binding on autologous cells in the trans-binding configuration, thereby activating the immunomodulatory function of CD22 in mitigating immune responses or regaining tolerance to self-tissues. The reasons for the loss of CD22 cis-to-trans conversion are unknown, but the subject is interesting to explore further.
Age-dependent factors could deregulate the B cell cis or trans binding ability (65, 66). This could be induced by changes of glycans on immune cells (B cells or autologous cells) or impairment of function or expression of enzymes such as β-galactoside α2,6-sialyltransferases or SIAE that are responsible for production of these sugar structures (61). In particular, loss-of-function mutation in SIAE has also been reported to have higher prevalence in patients with autoimmune diseases, such as RA, type 1 diabetes, and SLE (61); these alterations could lead to a global reduction in sialic acid prevalence, thus lowering the level of trans bindings of CD22. Single-nucleotide polymorphism mutation of mouse CD22 ligand-binding domain Arg130Glu led to deregulation of BCR–CD22 association, resulting in the reduction of CD22-exerted immunomodulation against BCR activation (67, 68). Furthermore, a correlation between decreased sialic acid ligand production with aging is implicated in literature (65, 66). These factors could induce a model where either a modified sialic acid ligand structure on B cells would favor cis-ligand binding or modified sialic acid ligand structures on autologous cells would discourage trans-ligand binding. This could lead to loss of tolerance against some autoantigens and initiate autoimmune diseases such as RA, SLE, pemphigus vulgaris, Sjögren’s syndrome, systemic sclerosis, vasculitis, Wegener’s granulomatosis, antiphospholipid syndrome, and Goodpasture’s syndrome, among others (69).
We further demonstrated that upon BCR activation (BCR cross-linked by anti-IgM), SM03 not only facilitated CD22 cis-ligand to trans-ligand binding conversion, but also upregulated the activation of SHP-1, a CD22 downstream signal for initiating immunosuppression, upon cotreatment with exogenous α2,6-linked sialic acid glycan (Fig. 3D). A similar effect was observed upon cotreatment with SM03-F(ab′)2 fragments, which lack the Fc region (Fig. 3D). This implicates that SM03-induced SHP-1 activation relies more on the Ab epitope binding to CD22, allowing α2,6-linked sialic acid glycan trans binding; furthermore, the involvement of Fc receptor engagement/coengagement-induced immunomodulation and SHP-1 activation may not be a necessary condition for the immunomodulatory mechanism of SM03 (9, 70, 71).
Moreover, SM03 could also reduce the proliferation of B cells derived from PBMCs (Fig. 3A), implicating an additional immuno-modulatory function of SM03 on BCR-activated B cells upon binding to CD22 (37, 72).
In addition to the facilitation of cis-to-trans conversion, SM03 might have other immunomodulatory functions by directly binding to the CD22 Ag. We demonstrated that SM03 downregulated both anti-IgM–induced (ligand of BCR) and LPS-induced (ligand of TLR4) NF-κB activation in the RAMOS reporter cell line (Fig. 3B, 3C). It is therefore inferred that SM03 could modulate the immune system via the TLR signal transduction pathway, as CD22 was involved in TLR transduction (60). It is possible that the enhanced endocytosis induced by SM03 could in turn inhibit endosomal TLRs, which are known to be receptors for autoantigenic nucleic acids and innate autoimmunity (73).
The collective clinical data for SM03 implicate clinical efficacies and good safety profiles against NHL, SLE, and RA patients (Tables II and III) (28, 40, 41). In a RCT phase II trial for RA, SM03 demonstrated clinical responses comparable to those of other marketed biologics targeting TNF-α (such as infliximab [Remicade] etanercept [Enbrel], adalimumab [Humira], and certolizumab [Cimzia]), IL-6 (such as tocilizumab [Actemra]), and CD20 (such as rituximab [Rituxan]) (74) when measured in ACR20 (65.3%), ACR50 (17%) and ACR70 (44.9%), respectively. It is noteworthy that unlike anti-CD20 Abs, which completely ablate circulating B cells, only a 10–30% reduction in the population for peripheral B cells was observed in SM03-treated patients (28, 40). The results indicated that the B cell-specific SM03 is effective for treating autoimmune diseases, not by the elimination of B cells, but rather via a different MOA leading to the modulation of the immune system. Similar biological responses were also observed in a phase I study for SLE (Table II). Interestingly, when the collective clinical data for SM03 for phase I for NHL, phase I for SLE, and phase I and II for RA were reviewed, SM03 stands out as a product with a good safety profile (Table III), particularly on the incidence of infection and serious infection, when compared with that of other pan-immunosuppressive biologics targeting TNF-α, IL-6, and CD20 (74).
Our proposed MOA not only addresses the clinical efficacy of SM03 in modulating B cell activation against autoantigens, it also gives a rationale behind the enhanced safety profile of SM03. The targeting of CD22 by SM03 would facilitate trans-ligand binding for self recognition, wherein the potentially self-immunoreactive B cells would be suppressed by the trans-ligand binding of CD22. In contrast, although SM03 would enable trans-ligand binding of B cells, it would not compromise the ability of B cells to recognize and trans-ligand crosslink to pathogens that contain no surface α2,6-linked sialic acid ligands. Thus, the pathogen-specific B cells maintain normal B cell responses against these pathogens, as manifested clinically with an improved safety profile.
Both RA and SLE are autoimmune diseases reflecting the irregularity of the immune system when it fails to recognize autoantigen as self and elicit a response. SM03 therefore could be a viable therapeutic candidate for these indications, as it coincides particularly well with the MOA of regaining B cell tolerance toward self-antigens by enforcing trans binding of CD22 on autologous tissues (62). Age-dependent factors may deregulate the B cell cis or trans binding ability, which in turn could lead to loss of tolerance against some autoantigens and initiate autoimmune diseases such as RA, SLE, Sjögren’s syndrome, systemic sclerosis, vasculitis, Wegener’s granulomatosis, antiphospholipid syndrome, and Goodpasture’s syndrome (69). It would be interesting to explore whether SM03 could be used for these diseases.
Acknowledgements
We thank Pepscan Presto BV for their contribution in mapping and establishing the discontinuous epitope for SM03.
Footnotes
This work was supported by SinoMab Bioscience Ltd.
The online version of this article contains supplemental material.
K.L.W. designed and conducted the study, designed and/or performed experiments, analyzed the data, and wrote the manuscript. S.O.L. designed the study, designed the experiments, and edited the manuscript. Z.L. and D.W. designed and conducted the study, analyzed the data, and edited the manuscript. N.S. and C.H.C. designed the experiments and edited the manuscript. F.M. and K.K.L. designed and/or performed experiments, analyzed the data, and edited the manuscript. All authors critically reviewed the manuscript.
Abbreviations used in this article:
- ACR
American College of Rheumatology
- AE
adverse event
- cd
cumulative dose
- CD22-muFc
CD22 conjugated with murine Fc
- CLIPS
Chemical Linkage of Peptides onto Scaffolds
- DAS28
disease activity score using 28 joint counts
- EMAB
epratuzumab
- MOA
mechanism of action
- NHL
non-Hodgkin’s lymphoma
- RA
rheumatoid arthritis
- RCT
randomized controlled trial
- RT
room temperature
- SHP-1
Src homology region 2 domain-containing phosphatase 1
- SIAE
sialic acid O-acetyl esterase
References
Disclosures
K.L.W., Z.L., F.M., D.W., C.H.C., N.S., and K.K.L. are current employees and/or shareholders of SinoMab BioScience Ltd.; S.O.L. is an officer and director of SinoMab BioScience Ltd. SinoMab BioScience Ltd. has filed certain patent applications pertaining to SM03.