Cytokine binding by polysaccharide-antibody conjugates

Abstract

Cytokine-neutralizing antibodies are used in treating a broad range of inflammatory conditions. We demonstrate that monoclonal antibodies against interleukin-1β and tumor necrosis factor-α were still active when conjugated to high molecular weight polysaccharides. These polysaccharides are hydrophilic, but their size makes them unable to circulate in the bloodstream when delivered to tissues, opening up the possibility of localized treatment of inflammatory conditions. To explore this new class of protein-polysaccharide conjugates, we covalently modified interleukin-1β and tumor necrosis factor-α monoclonal antibodies with high molecular weight hyaluronic acid and carboxymethylcellulose. Rigorous purification using dialysis with a 300 kDa-cutoff membrane removed unconjugated monoclonal antibodies. We characterized the composition of the constructs and demonstrated using molecular binding affinity measurements and cell assays that the conjugates were capable of binding proinflammatory cytokines. The binding affinities of both the unconjugated antibodies for their cytokines were measured to be approximately 120 pM. While all conjugates had pM-level binding constants, they ranged from 40 pM for the hyaluronic acid-(anti-interleukin-1β) conjugate to 412 pM for the carboxymethylcellulose-(anti-interleukin-1β) conjugate. Interestingly, the dissociation time constants varied more than the association time constants, suggesting that conjugation to a high molecular weight polysaccharide did not interfere with the formation of the antibody-cytokine complex but could stabilize or destabilize it once formed. Conjugation of cytokine-neutralizing antibodies to high molecular weight polymers represents a novel method of delivering anticytokine therapeutics that may avoid many of the complications associated with systemic delivery.

Figure 1

(A) Non-denaturing polyacrylamide gels, stained with Alcian blue, showing the results from analyzing HA samples with varying concentration with and without conjugated mAb. (i) 0.5 mg/mL HA, (ii) 0.25 mg/mL HA, (iii) 0.13 mg/mL HA, (iv) 0.06 mg/mL HA, (v) 0.03 mg/mL HA, (vi) 0.1x HA-anti-IL1β mAb, (vii) 0.1x HA-anti-TNF-α mAb, (viii) 0.1x HA+mAb. (B) CMC concentration calibration curve fit to the equation y=0.0398x-0.0388, R²=0.9965, (C) HA concentration calibration curve fit to the equation y=0.0346x-0.0062, R²=0.9964.

Figure 2

Calibration curve from fluorescence immunosorbent assay of mAb concentration in polysaccharide-mAb conjugates. Data were fit to y=1.1572x+0.771 (R²=0.9878).

Figure 3

Binding affinity analysis using Fortebio Octet system. (Top) Association, dissociation curve, and best-fit isotherms of different conditions: (A) anti-IL-1β mAb, (B) HA-anti-IL-1β mAb conjugate, (C) CMC-anti-IL-1β mAb conjugate. (Bottom) Plot of the residuals from the best-fit curve.

Figure 4

Binding constants (K_D) measured and calculated using ForteBio Octet System. The results are shown as the average of the five separated measurements of each condition, and the error bars are shown as the standard deviation. (*) p<0.001.

Figure 5

Imaging cytometry result of PMA-differentiated THP-1 macrophages stained with p65 subunit of nuclear factor-kappa B (NF-κB) and nucleus after stimulated by various different conditions. (A) Representative imaging cytometry image of THP-1 macrophages treated with 100 ng/mL of IL-1β. The nuclei appear as bright spots surrounded by bright rings, which represents the stained cytosolic NF-κB. (B-D) Histogram representation of the NF-κB translocation value of THP-1 macrophages treated with (B) culture media, (C) 100 ng/mL IL-1β, and (D) CMC-anti-IL-1β + 100 ng/mL IL-1β. A dashed line represents the 90^th percentile of the unstimulated THP-1 macrophage population, and this translocation value of 320 was used as a threshold to identify responsive THP-1 cells in the stimulated populations. The number beside the dashed line in each histogram is the fraction of stimulated cells determined by the threshold.

Figure 6

Fraction of the THP-1 cells with translocation values above the threshold defined as the 90^th percentile in the unstimulated control group. (i) 100 ng/mL LPS, (ii) culture medium, (iii) IL-1β, (iv) 0.1 wt% CMC-anti-IL-1β + 100 ng/mL IL-1β, (v) 0.1 wt% HA-anti-IL-1β + 100 ng/mL IL-1β, (vi) 100 μg/mL anti-IL-1β + 100 ng/mL IL-1β, (vii) 100 ng/mL TNF-α, (viii) 0.1 wt% CMC-anti-TNF-α + 100 ng/mL TNF-α, (ix) 0.1 wt% HA-anti-TNF-α + 100 ng/mL TNF-α, (x) 100 μg/mL anti-TNF-α mAb + 100 ng/mL TNF-α, (xi) 1 mg/mL CMC, and (xii) 1 mg/mL HA. Error bars represent the standard deviation of four separate samples.

Scheme 1

Preparation of polysaccharide-antibody conjugates. Polysaccharides were partially activated at carboxylic acid sites followed by coupling reaction with pendant amines on the mAb (H₂N-mAb). Carboxymethylcellulose (CMC) and hyaluronic acid (HA) were used as two representative polysaccharides.