Low-Cost, Large-Scale Production of the Anti-viral Lectin Griffithsin

Abstract

Griffithsin, a broad-spectrum antiviral lectin, has potential to prevent and treat numerous viruses including HIV, HCV, HSV, SARS-CoV, and SARS-CoV-2. For these indications, the annual demand for Griffithsin could reach billions of doses and affordability is paramount. We report the lab-scale validation of a bioprocess that supports production volumes of >20 tons per year at a cost of goods sold below $3,500/kg. Recombinant expression in engineered E. coli enables Griffithsin titers ∼2.5 g/L. A single rapid precipitation step provides > 90% yield with 2-, 3-, and 4-log reductions in host cell proteins, endotoxin, and nucleic acids, respectively. Two polishing chromatography steps remove residual contaminants leading to pure, active Griffithsin. Compared to a conventional one this process shows lower costs and improved economies of scale. These results support the potential of biologics in very large-scale, cost-sensitive applications such as antivirals, and highlight the importance of bioprocess innovations in enabling these applications.

Keywords: SARS-CoV-2; antiviral; biologics manufacturing; downstream recovery; griffithsin.

FIGURE 1

(A) Estimated cost of goods sold and scale of production for various protein drug products. Gray circles: monoclonal antibodies (mAbs), based on an analysis of publicly available financial data summarized in Supplementary Table S2. Brown diamond: results of an in silico-modeled tobacco-based process for Griffithsin (GRFT) by Alam et al. (2018), which does not include formulation or packaging costs. Black squares: in silico-models of a GRFT process based on E. coli fermentation and a conventional, chromatography-based downstream purification, modeled at various scales. Blue triangles: in silico models of GRFT processes based on E. coli fermentation and a downstream purification based on the precipitation step reported here, modeled at various scales. The green shaded area indicates an estimated target range for large-scale, low-cost deployment of an anti-SARS-CoV-2 biologic (>20,000 kg/yr at <$10.00/g). (B) The GRFT processes and target region from (A), replotted on zoomed-in axes for clarity. Cost per 24 mg dose is also shown, assuming packaging in 200 mg multi-dose vials.

FIGURE 2

E. coli fermentation-based processes for GRFT antiviral manufacturing, modeled to produce approximately 24,000 kg of GRFT per year in a filled and finished antiviral (here, a multidose vial with a formulation suitable for nebulization). (A) Process with conventional chromatography-based purification. (B) Process with precipitation-based purification. (C) Costs of Goods Sold per gram of formulated and packaged GRFT for the processes from (A) (left) and (B) (right), shown by process section and cost category. (D) Estimated capital expenses for the processes from (A,B), shown by process section. CD - cycle duration, AEx - Anion exchange, WFI = water for injection, DSP - downstream purification, including primary recovery, Entity - final packaged product (e.g., a filled vial). Fermentation includes seed train and primary fermentation operations. Materials include all raw materials (e.g., chemicals, water) and consumables (e.g., membranes and resins).

FIGURE 3

COGS and throughput sensitivities for the two processes shown in Figure 2. Variables along the Y-axis were varied one at a time and in each case the model was then adjusted to maximize throughput and minimize cost without any changes in equipment or plant layout. Light gray bars show high conditions (increasing the variable), dark gray bars show low conditions (decreasing the variable). Black line shows baseline values for each model. (A) Conventional purification. (B) Precipitation-based purification. Filler - the vial filler used in formulation. AEx - anion exchange, CEx - cation exchange, TFF - tangential flow filtration.

FIGURE 4

(A) Expression of GRFT in 96-well plate microfermentations using E. coli strain DLF_Z0025 containing GRFT under the control of the low-phosphate inducible promoters phoA, phoB, and yibD. (B) Expression of GRFT in a two-stage fermentation. E. coli strain DLF_Z0025 containing GRFT under the control of the low-phosphate inducible phoA promoter was cultured in a 1 L bioreactor. Biomass levels are shown by gray triangles and the final GRFT titer (∼2.7 g/L) is shown by a green square, corresponding to a GRFT expression level of ∼20% of total cell protein.

FIGURE 5

Results of Design of Experiment (DoE) studies to optimize the precipitation step. Three key variables are shown: temperature, ammonium sulfate concentration (% saturation) and pH. Two outputs were evaluated: yield (A) and separation factor (B). Gray dashed lines are included for perspective. (C) A summary of these outputs over each experiment in the two rounds of DoE. (D) Fluorescently stained SDS-PAGE gel, converted to grayscale and with brightness values inverted for clarity. Lane 1, Ladder; Lane 2, untreated E. coli lysate containing GRFT (diluted 1:40); Lanes 3 and 4, supernatant following precipitation and supernatant following diafiltration into chromatography running buffer (each diluted 1:30); Lane 5, flow-through fraction from the final endotoxin removal chromatography step (diluted 5:8).

FIGURE 6

Binding kinetics of purified GRFT vs. purified HIV gp140. Measurements were accomplished with SPR. Lines show the mean response at each concentration and shaded areas show standard deviations. Each replicate set was analyzed independently by fitting with a heterogeneous ligand model, followed by calculation of summary statistics.