Targeting the Initiator Protease of the Classical Pathway of Complement Using Fragment-Based Drug Discovery

Abstract

The initiating protease of the complement classical pathway, C1r, represents an upstream and pathway-specific intervention point for complement-related autoimmune and inflammatory diseases. Yet, C1r-targeted therapeutic development is currently underrepresented relative to other complement targets. In this study, we developed a fragment-based drug discovery approach using surface plasmon resonance (SPR) and molecular modeling to identify and characterize novel C1r-binding small-molecule fragments. SPR was used to screen a 2000-compound fragment library for binding to human C1r. This led to the identification of 24 compounds that bound C1r with equilibrium dissociation constants ranging between 160-1700 µM. Two fragments, termed CMP-1611 and CMP-1696, directly inhibited classical pathway-specific complement activation in a dose-dependent manner. CMP-1611 was selective for classical pathway inhibition, while CMP-1696 also blocked the lectin pathway but not the alternative pathway. Direct binding experiments mapped the CMP-1696 binding site to the serine protease domain of C1r and molecular docking and molecular dynamics studies, combined with C1r autoactivation assays, suggest that CMP-1696 binds within the C1r active site. The group of structurally distinct fragments identified here, along with the structure-activity relationship profiling of two lead fragments, form the basis for future development of novel high-affinity C1r-binding, classical pathway-specific, small-molecule complement inhibitors.

Keywords: complement inhibitors; fragment-based drug discovery; surface plasmon resonance.

Figure 1

(A) Complement is activated by three canonical pathways known as the classical pathway (CP), lectin pathway (LP), or alternative pathway (AP). Activation of the classical pathway is controlled by the C1 complex (i.e., C1qC1r₂C1s₂). The pattern recognition protein C1q binds to target surfaces resulting in the autoactivation of the zymogen C1r proteases (shown here as ‘Pro-C1r’) into C1r enzymes, which then proteolytically cleave and activate C1s within the C1 complex. The lectin pathway is activated by lectin pathway-specific pattern recognition proteins in complex with mannan-binding associated serine proteases (MASPs), while the alternative pathway is constitutively activated at low levels by a spontaneous hydrolytic event known as tick-over. Both the classical and lectin pathways converge at the cleavage of C2 and C4 to generate the classical/lectin pathway C3 convertases, C4b2b. Alternative pathway activation results in the formation of C3 convertases in the form of C3bBb. C3 convertases cleave the central molecule of the cascade, C3, into C3a and C3b, resulting in an amplification loop that produces increasing quantities of surface bound C3b. At high surface concentrations of C3b, C3 convertases bind an additional C3b molecule, resulting in a switch of substrate specificity to C5. Cleavage of C5 by these C5 convertases (i.e., C4b2bC3b and C3bBbC3b) results in the release of the anaphylatoxin C5a and the formation of the pore-like lytic structure called the membrane attack complex (i.e., C5b–C9). (B) Fragment-based drug discovery schematic.

Figure 2

Direct binding of compounds to full-length C1r by SPR. A 2000-compound library was screened at 500 µM final compound concentration for solubility in SPR buffer and for non-specific binding to a blank sensor chip surface (i.e., ‘clean screen’). A total of 1619 compounds were soluble and exhibited low non-specific binding capacity in our SPR assay system. The ability of each of these compounds to bind directly to C1r was measured by injecting a 500 µM final compound concentration over immobilized full-length C1r. A molecular weight corrected theoretical maximal binding response (R_max) for each compound was calculated and compounds that exhibited superstoichiometric binding (i.e., > 2 × R_max) were eliminated from further consideration. In total, 95 compounds exhibited ≥ 60% R_max (green circles).

Figure 3

Dose-dependent binding of full-length C1r by selected hit compounds. Dose-dependent C1r binding for 24 compounds was measured by injecting a two-fold variable concentration series of each compound ranging from 7.8 to 500 µM. Steady-state affinities were calculated from the resulting sensorgrams. A representative set of sensorgrams are shown along with the associated steady-state K_D values. The corresponding steady-state fits are shown in Figure S1. K_D values are reported as the mean ± S.D. calculated from three independent injection series.

Figure 4

Inhibition of the classical pathway by selected hit compounds. All 24 hit fragments were tested for their ability to block C4 activation in an ELISA-based assay under conditions specific for the classical pathway. Each compound was tested in triplicate at a single concentration of 500 µM. Positive hits (four in total) were defined as any compound that significantly reduced C4b deposition relative to a non-binding control compound (CMP-685), as judged by an unpaired t-test (* p < 0.05).

Figure 5

Selectivity and mechanistic analysis of CMP-1611 and CMP-1696. (A) Chemical structure of CMP-1611. (B) Chemical structure of CMP-1696. (C) Dose-dependent inhibition by CMP-1611 and CMP-1696 in a classical pathway-specific ELISA. Data were fit with GraphPad Prism using non-linear regression with a log(inhibitor) vs. response model. For CMP-1611, an IC₅₀ value of 660 µM with an associated 95% confidence interval of (560–790 µM, n = 9) was calculated. For CMP-1696, an IC₅₀ value of 520 µM with an associated 95% confidence interval of (410–680 µM, n = 7) was calculated. The CMP-778 inhibitory response could not be fit to a dose–response inhibition model. (D) Complement pathway selectivity of CMP-1611 and CMP-1696. Compounds were assessed for their ability to inhibit activation of complement via the lectin and alternative pathways using single doses of 500 µM compound in triplicate. To ensure only lectin pathway activation, 2% (v/v) C1q-depleted serum (CompTech) was used and mannan was used as the activator. To match serum sources and amounts for this assay, the classical pathway assays were repeated here using serum from CompTech at 2% (v/v) final concentration. Alternative pathway activation assays were performed using 20% (v/v) serum (CompTech), alternative pathway buffers, and C3b detection (see Methods and Materials for details). CMP-1611 had no effect on the lectin or alternative pathway, whereas CMP-1696 blocked lectin but not alternative pathway activation. (E) C1r, C1r-CUB1, and C1r-CCP2-SP, were immobilized on an SPR sensor chip and binding responses for 500 µM CMP-1611 and CMP-1696 or 10 µM Futhan were each injected in duplicate over all surfaces. Binding responses were corrected for the molecular weight of each analyte and the immobilization level and molecular weight of each surface ligand. Measures of statistical significance in (D) were obtained by comparison of vehicle control using an unpaired t-test (* p < 0.05).

Figure 6

CMP-1696 structure activity relationship. (A) CMP-1696 was redocked onto C1r-CCP2-SP (PDB: 1MD8, grey surface representation) and the top nine scored poses are shown. All CMP-1696 poses dock into the S1 subsite (cyan) near the catalytic triad (blue). (B) C1r autoactivation assay. C1r proenzyme undergoes time-dependent autoactivation at 37 °C. Autoactivation was measured using a synthetic substrate for C1r enzyme. The reaction progress of vehicle control (dashed line) or in the presence of 10 mM CMP-1696 (solid line) was monitored for 1 h. (C) Molecular dynamics (MD) simulations of CMP-1696/C1r-CCP2-SP. Root mean square deviation (RMSD) in nm for each of the CMP-1696 poses measured over the 10 ns molecular dynamics simulation. (D) MM/PBSA energy calculations for each pose in the 10 ns MD simulations indicate that pose 1 is the most energetically favorable. (E) Hydrogen bonding interactions at the start of the MD simulation are shown as dashed lines. (F) A 50 ns MD simulation (Video S1) was carried out for pose 1 and MM/PBSA was used to calculate total energy. Subcategorized energy contributions are also shown where vDW is van der Waals forces and SASA is solvent-accessible surface area.