A series of surface plasmon resonance studies was conducted to characterize the monovalent and bivalent interaction of narsoplimab with human MASP-2 using a BiaCore 3000 (GE Healthcare Life Sciences) instrument. MASP-2 antigen was immobilized on the glass surface by direct amine coupling, and the association and dissociation kinetics of intact narsoplimab or monovalent Fab were assessed at 25°C using 10 mM HEPES, 150 mM NaCl, 0.01% Tween-20, pH 7.4 as the mobile phase. Experimental data were globally fitted to a 1:1 interaction model to estimate association (k_a) and dissociation (k_d) rate constants, which were subsequently used to calculate the equilibrium dissociation constant (K_D ) for the narsoplimab–MASP-2 interaction (K_D = k_d/k_a).

To obtain K_D values for the binding of intact narsoplimab or monovalent Fab to MASP-2 by ELISA, polystyrene plates were coated overnight at 2°C to 8°C with recombinant human MASP-2 as antigen. Any remaining sites were blocked with bovine serum albumin, then serially diluted narsoplimab or Fab was added to the antigen-coated wells and incubated for 1 hour. After washing the wells to remove unbound material, a horseradish peroxidase-conjugated anti-human IgG detection antibody was added to all wells. Incubation was followed by washing to remove unbound conjugate, then 3,3′,5,5′-tetramethylbenzidine (TMB) substrate was added to the wells for color development. The development step was stopped with acid and the extent of TMB substrate conversion was quantitated colorimetrically by measuring the absorbance at 450 nm. The K_D was estimated by fitting the binding dose-response data to a 4-parameter logistic model using GraphPad Prism version 6.02.

The specificity of narsoplimab for human MASP-2 relative to four other serine proteases of the complement system (C1r, C1s, MASP-1, and MASP-3) in human serum was assessed by solid-phase ELISA following the procedure described above. Briefly, the serine proteases were immobilized on polystyrene plates, then allowed to react with different concentrations of narsoplimab. Bound antibody was revealed by peroxidase-labelled detection antibody and quantified by absorbance measurements. Apparent K_D values were estimated by nonlinear regression using a four-parameter logistic model.

To assess selectivity, the effects of narsoplimab on lectin, classical, and alternative pathway-dependent activation of C5b-9 were compared using the Wieslab^® Complement System Screen ELISA kit (Svar Life Science, Malmö, Sweden). Human serum samples, diluted into pathway-specific assay buffers following the manufacturer’s instructions, were preincubated with different concentrations of narsoplimab followed by exposure of serum antibody mixtures to the appropriate pathway-specific activation wells provided by the kit. C5b-9 activation was evaluated using alkaline phosphatase-labeled antibody specific for the neoantigen exposed upon C5b-9 formation.

Traditional methods to evaluate lectin pathway activity use highly diluted serum samples (typically 1% serum). While simple and robust, the concentrations of all components controlling the complement activation response in assays using highly diluted serum are approximately 100-fold lower than their respective in vivo concentrations, and evaluations of MASP-2 inhibitors such as narsoplimab in these assay conditions may not appropriately characterize their pharmacologic activity under physiologic conditions. To assess pharmacologic activity under near-physiologic conditions in vitro, ELISA and flow cytometry-based test methods using minimally diluted serum samples from mice, rabbits, cynomolgus monkeys, and rats as test matrix were developed.

In separate 96-well microtiter plates, minimally diluted (90%) serum samples from mice, rabbits, cynomolgus monkeys, and rats were pre-incubated with serially diluted narsoplimab (up to 500 nM). For ELISA assays, lectin-induced complement deposition reactions were initiated by addition of serum antibody preincubation mixtures to the mannan-coated assay plates. For flow cytometry assays, deposition reactions were initiated by the addition of 2.5 μL of mannan-coated beads (~250,000 beads) to antibody-serum preincubation mixtures. Following incubation at the indicated reaction time and temperature, reactions were stopped and developed for C3 or C4 activation as described below. During method development, time-course studies were conducted to analyze reaction kinetics of the lectin pathway activation response for the test matrix from each species, and a reaction time during the linear phase of the reaction that provided a robust complement activation signal in the absence of narsoplimab was selected for pharmacologic assessments. Lectin pathway activation was very rapid and proceeded to completion within minutes when evaluated at 37°C (data not shown). Therefore, assays were conducted at 4°C to achieve reaction kinetics compatible with standard benchtop plate-based assay procedures.

At the end of the incubation period, the reactions were stopped with ice-cold buffer and the plates (ELISA) or beads (flow) were washed twice. After the wash steps, lectin-induced complement activation was quantified using a primary antibody suitable for the relevant test species. For C3 activation, the wells or beads were incubated for 1 hour at room temperature with rabbit anti-human C3c antibody (which recognizes C3b and C3c) for all species. For C4 activation by ELISA, the wells were incubated for 1 hour at room temperature with biotinylated anti-mouse C4 antibody for mouse serum and with biotinylated chicken anti-human C4 antibody for monkey, rabbit, and human serum. For C4 activation by flow cytometry, the beads were incubated for 1 hour at 4°C with biotinylated anti-mouse C4 antibody for mice, with biotinylated chicken anti-human C4 antibody for rabbits, and with mouse anti-human C4d monoclonal antibody (which recognizes C4b as well as its degradation product C4d) for humans and monkeys.

The functional potency of MASP-2 inhibition (pharmacologic activity) was estimated by determining the half-maximal inhibitory concentration (IC₅₀) for the narsoplimab inhibition dose-response curve for activation of C3 or C4. IC₅₀ values were estimated by plotting the absorbance values (ELISA) or mean fluorescent intensity values (flow cytometry) against the antibody concentration and then fitting the data to a 4-parameter logistic regression.

To characterize the pharmacokinetics and pharmacodynamics of narsoplimab in primates, cynomolgus monkeys were treated with a single narsoplimab dose of 0.05, 0.15, 0.5, 1.5, or 5 mg/kg administered by intravenous (IV) bolus injection. Serial serum samples were collected predose and approximately 1, 24, 96, 168, 336, and 672 hours after narsoplimab administration. Additional serum samples were collected at 0.083, 0.25, 3, 6, 48, 240, and 504 hours after narsoplimab administration for monkeys exposed to 5 mg/kg. Narsoplimab serum concentrations were measured using a human IgG4 sandwich ELISA kit (Cayman Chemical, Ann Arbor MI) following the manufacturer’s instructions.

The pharmacodynamic response to narsoplimab treatment was evaluated by quantifying changes in ex vivo lectin pathway activity in response to narsoplimab treatment using an ELISA method similar to that employed for in vitro pharmacologic activity assessment (see Section 2.3). To calculate the pharmacodynamic response (% lectin pathway inhibition), the residual ex vivo lectin pathway activity (C4 activation by mannan) in serum samples collected after narsoplimab administration was quantified and expressed relative to the ex vivo lectin pathway activity in serum samples from the same animal at baseline using the following formula: % lectin pathway inhibition = 100 * (lectin pathway activity [baseline] – lectin pathway activity [postdose])/lectin pathway activity (baseline).

To characterize the pharmacokinetic/pharmacodynamic relationship in cynomolgus monkey, the complete set of drug concentration and pharmacodynamic response (% lectin pathway inhibition) data were fitted using a 4-parameter logistic regression model to estimate the EC₅₀ value.

Animal welfare complied with the U.S. Department of Agriculture’s (USDA) Animal Welfare Act (9 CFR Parts 1, 2 and 3). The Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, National Academy Press, Washington, D.C., was followed. The facilities maintained an Animal Welfare Assurance statement with the National Institutes of Health, Office of Laboratory Animal Welfare. To ensure compliance, each study protocol was approved by the Institutional Animal Care and Use Committee (IACUC) before the initiation of treatment. Each study was designed to use the fewest number of animals possible consistent with the objective(s) of the study, the scientific needs of the Sponsor, and contemporary scientific standards.

Narsoplimab bound to human MASP-2 with high affinity, with K _D values of 0.073 nM and 0.062 nM as determined by surface plasmon resonance ( Table 1 ) and ELISA ( Table 2 ), respectively. The affinity of narsoplimab-related monovalent Fab for human MASP-2 was approximately 100-fold lower: the K _D was 10.3 nM by surface plasmon resonance ( Table 1 ) and 14.9 nM by ELISA ( Table 2 ). While narsoplimab and monovalent Fab associated with human MASP-2 at similar rates (k_a, 3.5 vs 2.6 × 10⁶ M^-1s^-1), narsoplimab dissociated approximately 100-fold more slowly than monovalent Fab (k_d, 2.47 × 10^-2 vs 2.57 × 10^-4 s^-1; Table 1 ), indicating a substantial contribution of avidity to the tight interaction of narsoplimab with MASP-2.

To shed light on the mechanism of MASP-2 inhibition by narsoplimab, we analyzed co-crystal structures for narsoplimab Fab bound to MASP-2 and C4 bound to MASP-2 (Protein Data Bank code 5JPM) ( Figure 2 ). From the co-crystal structures, it is evident that narsoplimab and C4 bind to a common site bridging the complement control protein (CCP) domains CCP1 and CCP2 of MASP-2. Hence, narsoplimab appears to inhibit MASP-2 activity by competing with C4 substrate for MASP-2 binding. Interestingly, the active proteolytic site of MASP-2 that cleaves and activates C4 is not directly impacted by narsoplimab binding.

The specificity of narsoplimab for human MASP-2 relative to four other serine proteases of the complement system (C1r, C1s, MASP-1, and MASP-3) was assessed by solid-phase ELISA. Narsoplimab bound with high-affinity to zymogen and enzymatically active forms of human MASP-2 (K _D = 0.062 nM and 0.089 nM, respectively). By contrast, the interaction with C1s exhibited >5000-fold lower affinity (K _D ~500 nM); no discernable binding to C1r, MASP-1, or MASP-3 was observed at concentrations up to 500 nM ( Table 3 ).

To assess selectivity, the effects of narsoplimab on lectin, classical, and alternative pathway-dependent activation of C5b-9 were compared using the Wieslab^® total complement screening kit. These tests showed that narsoplimab inhibited lectin pathway-dependent activation of C5b-9 in human serum with an IC₅₀ of ~1 nM whereas narsoplimab concentrations up to 500 nM had no observable effect on classical pathway-dependent or alternative pathway-dependent activation of C5b-9 in human serum ( Figure 3 ). These data demonstrate that narsoplimab binds to MASP-2 with high specificity and selectively inhibits the lectin pathway without affecting the classical pathway or alternative pathway of the complement system.

ELISA-based and flow cytometry-based test methods were developed to assess lectin pathway-dependent activation of C3 or C4 under near-physiologic conditions and to characterize the pharmacologic activity of narsoplimab in mice, rabbits, cynomolgus monkeys, and humans. To conduct these evaluations at near-physiologic (90%) serum concentrations, it was necessary to reduce the reaction temperature to 4°C in order to achieve reaction kinetics compatible with standard benchtop plate-based assay procedures. Under these conditions, reaction kinetics were characterized by an initial lag phase followed by a near-linear time-course over 30 to 60 minutes (data not shown). For each species, a reaction time during the linear phase of the reaction that provided a robust lectin pathway activation signal in the absence of narsoplimab, with minimal background complement activation in the presence of excess narsoplimab, was selected for further pharmacologic assessments. Background complement activation was presumably due to classical or alternative pathway activity, particularly for C3 activation.

Optimized conditions for ELISA-based and flow cytometry-based lectin pathway functional assays were applied to in vitro pharmacologic characterization of narsoplimab in serum from mice, rabbits, cynomolgus monkeys and humans. Representative examples of inhibition of lectin pathway functional activity assessed by flow cytometry are shown in Figure 4 , and the data from all experiments are summarized in Table 4 . Using the flow cytometry assay, 50% inhibition of C4 activation in human, mouse, rabbit, and cynomolgus monkey sera was achieved at 3.4 ± 2.3, 13 ± 1.3, 9.2 ± 2.6, and 33 ± 13 nM, respectively. Where tested, the IC₅₀ value for inhibition of C3 activation and inhibition of C4 activation were comparable ( Table 4 ). ELISA assays yielded IC₅₀ values comparable to the flow cytometry assays, although ELISA assays tended to have greater variability due to the lower signal-to-noise ratios compared with flow cytometry assays ( Supplementary Table 1 ). In rat serum, narsoplimab concentrations up to 500 nM achieved <50% inhibition of lectin pathway activation using the ELISA assay, indicating that narsoplimab does not have appreciable pharmacologic activity in rats (data not shown).

To characterize the pharmacodynamic response to narsoplimab in primates, the time-course and dose-response of lectin pathway inhibition measured ex vivo was evaluated in cynomolgus monkeys treated with varying doses of narsoplimab. Narsoplimab concentrations were also determined to define the PK profile and to characterize the PK/PD relationship in cynomolgus monkeys. Narsoplimab exposure increased over the dose range evaluated ( Figure 5A ), with approximately dose-proportional increases in exposure from 0.05 to 1.5 mg/kg and a less-than-proportional increase at 5 mg/kg. The terminal half-life was approximately 3 to 8 days.

As expected, increasing doses were associated with corresponding increases in the magnitude and duration of the pharmacodynamic response ( Figure 5B ). While no appreciable pharmacodynamic response (% inhibition of lectin pathway activity measured ex vivo in post-dose serum samples relative to lectin pathway activity measured at baseline) was seen after 0.05 or 0.15 mg/kg administration of IV narsoplimab, clear dose-related pharmacodynamic responses were observed in the dose range of 0.5 to 5 mg/kg IV. Single dose IV administration of 0.5 or 1.5 mg/kg of narsoplimab to cynomolgus monkeys resulted in approximately 60% and 90% inhibition, respectively, of ex vivo lectin pathway activity immediately after administration, followed by a gradual decline of inhibition over time. A single dose administration of 5 mg/kg of narsoplimab IV in cynomolgus monkeys resulted in nearly complete inhibition of lectin pathway activity that lasted for more than 1 week followed by a slow and gradual decline of inhibition over time.

Further analysis of pharmacokinetic and pharmacodynamic data revealed a well-defined relationship between narsoplimab serum concentration and pharmacodynamic response, with an ex vivo EC₅₀ value of ~6 µg/mL ( Figure 5C ). This ex vivo EC₅₀ value in cynomolgus monkeys was in good agreement with the in vitro IC₅₀ value for inhibition of lectin pathway activity in the same species (33 ± 13 nM, ~5 µg/mL), indicating that in vitro functional potency data predict in vivo target engagement as a function of narsoplimab serum concentrations in primates. Based on the substantially greater functional potency of narsoplimab in humans relative to monkeys (IC₅₀ of 3.4 nM [~0.5 µg/mL] for humans compared to 33 nM [~5 µg/mL] for cynomolgus monkeys), it is expected that pharmacodynamic responses can be achieved in humans at approximately 10-fold lower drug concentrations and a correspondingly lower dose.