Enabling the formation of native mAb, Fab′ and Fc-conjugates using a bis-disulfide bridging reagent to achieve tunable payload-to-antibody ratios (PARs)

Either as full IgGs or as fragments (Fabs, Fc, etc.), antibodies have received tremendous attention in the development of new therapeutics such as antibody–drug conjugates (ADCs). The production of ADCs involves the grafting of active payloads onto an antibody, which is generally enabled by the site-selective modification of native or engineered antibodies via chemical or enzymatic methods. Whatever method is employed, controlling the payload–antibody ratio (PAR) is a challenge in terms of multiple aspects including: (i) obtaining homogeneous protein conjugates; (ii) obtaining unusual PARs (PAR is rarely other than 2, 4 or 8); (iii) using a single method to access a range of different PARs; (iv) applicability to various antibody formats; and (v) flexibility for the production of heterofunctional antibody-conjugates (e.g. attachment of multiple types of payloads). In this article, we report a single pyridazinedione-based trifunctional dual bridging linker that enables, in a two-step procedure (re-bridging/click), the generation of either mAb-, Fab′-, or Fc-conjugates from native mAb, (Fab′)2 or Fc formats, respectively. Fc and (Fab′)2 formats were generated via enzymatic digestion of native mAbs. Whilst the same reduction and re-bridging protocols were applied to all three of the protein formats, the subsequent click reaction(s) employed to graft payload(s) drove the generation of a range of PARs, including heterofunctional PARs. As such, exploiting click reactivity and/or orthogonality afforded mAb-conjugates with PARs of 6, 4, 2 or 4 + 2, and Fab′- and Fc-conjugates with a PAR of 3, 2, 1 or 2 + 1 on-demand. We believe that the homogeneity, novelty and variety in accessible PARs, as well as the applicability to various antibody-conjugate formats enabled by our non-recombinant method could be a suitable tool for antibody–drug conjugates optimisation (optimal PAR value, optimal payloads combination) and boost the development of new antibody therapeutics (Fab′- and Fc-conjugates).


Synthetic chemistry section
General experimental details for synthetic chemistry Chemicals were purchased from Sigma-Aldrich, Santa Cruz Biotechnology, AlfaAesar or Axispharm, and were used as received unless otherwise stated. Solvents were used as supplied. All reactions were monitored by thin-layer chromatography (TLC) on pre-coated silica gel plates. Flash column chromatography was carried out with either pre-loaded Biotage® SNAP column chromatography cartridges or pre-loaded GraceResolvTM flash cartridges on a Biotage® Isolera Spektra One flash chromatography system. All reaction mixtures were stirred magnetically unless stated otherwise. All reactions involving moisture sensitive compounds or procedures were carried out in flamedried flask under an atmosphere of argon. Room temperature (RT) is defined as 16-23 °C. Reactions at 0 °C were cooled with an ice/water bath. Removal of solvent and concentration in vacuo was carried out on a Büchi rotary evaporator followed by evaporation under high vacuum. 1 H NMR spectra were obtained at 400, 500, 600 or 700 MHz. 13 C NMR spectra were obtained at 100, 125, 150 or 175 MHz. All results were obtained using Bruker NMR instruments, the models are as follows: Avance III 400, Avance 500, Avance III 600, Avance Neo 700. The chemical shifts (δ) for 1 H and 13 C are quoted relative to residual signals of the solvent on the ppm scale. 1  where necessary. Infrared spectra were obtained on a Perkin Elmer Spectrum 100 FTIR Spectrometer operating in ATR mode. Infra-red spectra were recorded on a Bruker ALPHA FT-IR spectrometer operating in ATR mode, with frequencies given in reciprocal centimetres (cm -1 ). Melting points were taken on a Gellenkamp apparatus and are uncorrected. High and low resolution mass spectra were recorded on a VG70 SE mass spectrometer, operating in modes ESI, EI, or CI (+ or -) depending on the sample, at the Department of Chemistry, University College London.
The combined organic layers were washed with brine (15 mL), dried (MgSO4), and concentrated under vacuum.

Protein LC-MS general protocol
Protein conjugates were prepared for analysis by desalting into distilled water (Zeba™ Spin, ThermoFisher Scientific,  Table 1). The column effluent was continuously electrosprayed into the capillary ESI source of the Agilent 6510 QTOF mass spectrometer and ESI mass spectra were acquired in positive electrospray ionisation (ESI) mode using the m/z range 1,000−8,000 in profile mode with Quad AMU set to 500. The raw data was converted to zero charge mass spectra using a maximum entropy deconvolution algorithm, over the appropriate regions as identified via the LC trace, with the software, MassHunter (version B.07.00).

Protein A purification
The sample was applied to a NAb TM protein A column (Thermo Scientific) and incubated at RT with end-over-end

General protocol for the reduction of mAb, Fab or Fc
A solution of 10 mM TCEP was prepared by dissolving TCEP-HCl (2.86 mg) in 5 × BBS (1 mL). mAb, (Fab')2 or Fc were prepared in BBS pH 8, 2 mM EDTA (resulting solution ~20 µM), followed by addition of 10 molar equivalents of TCEP solution (10 mM in 5 × BBS pH 8). The mixture was incubated for 90-120 min at 37 °C under constant agitation (300 rpm). The buffer was then exchanged for BBS pH 8, 2 mM EDTA (1 x Zeba spin 7 MCWO unsalting procedure, followed by 1 x Zeba spin 7 MCWO buffer-swap procedure. Alternatively, if sample volume was superior to 120 µL, buffer-swap was realised in 3 washing cycles with Vivaspin 10 MCWO). NB.: The 5 × BBS was employed to maintain the pH of the solution at 8, as TCEP-HCl is strongly acidic.  (For LC-MS analysis, sample was diluted down in H2O to ~5 µM final concentration. PNGase F was added (0.6 µL) and resulting sample was incubated 12 h at RT before analysis).
(For LC-MS analysis, sample was diluted down in H2O to ~5 µM final concentration. PNGase F was added (0.6 µL) and resulting sample was incubated 12 h at RT before analysis).
(For LC-MS analysis, sample was diluted down in H2O to ~5 µM final concentration).
(For LC-MS analysis, sample was diluted down in H2O to ~5 µM final concentration. PNGase F was added (0.6 µL) and resulting sample was incubated 12 h at RT before analysis).

PAR 2 + 1
3. LC-MS spectra NB: Given that sensitivity of high m/z was low for mAb conjugates analyses (see mass envelop raw data), the mass of half-antibody species (no desired but present in small proportions due to disulfide scrambling during rebridging) were included as "expected mass", in order to reinforce confidence in the automatic deconvoluted data generated by the software for mAb species. Manual deconvolution check was realised too. PNGase enzyme was used for deglycosylation in every mAb-and Fc-conjugates analysis (even if not necessarily shows up on MS spectra).