The extract solution is filter-pressed at 25–30 psig to remove the aggregated protein impurities. Filtering has a process time of 1 h and requires a filter area of 3 m2 to handle the 590 kg/batch of the process stream. At this stage, the process loses a further 8% of the Griffithsin but removes all the RuBisCO and 87% of the TMV coat protein impurities. The filtrate from this step is transferred to a second mixing and storage tank, mixed with bentonite clay and magnesium chloride, and stored at 4◦C for a 12-h period. This stage is the bottleneck operation for the downstream process. After the 12-h incubation, the solution is filtered through a second 0.3µm filter press and a 0.2µm inline sterilizing filter. These operations remove the remaining protein impurities leaving a Griffithsin extract with greater than 99% purity but at the cost of losing 6% of the Griffithsin. The second plate-and-frame filter has a filter area of about 3 m2 and will process all of the extract in 1 h. There is approximately 222 g of Griffithsin per batch at the end of the filtration phase. Following the filtrations steps, the Griffithsin extract solution is collected in a storage tank and further purifified using an AxiChrom column with Capto MMC resin to remove residual color and potential non-proteinaceous impurities. To accommodate the 222 g of Griffithsin in solution, 4.9 L of MMC bed resin is needed at a 45 mg/mL binding capacity . The order of the operations for this chromatography step are: Equilibrate, load, wash, elute, and regenerate. In total, chromatography requires 10 h with the load step taking the longest, at 8 h, because approximately 600 L of solution are processed. Chromatography is necessary to decolorize the extract at the expense of losing 4% of the Griffithsin,vertical garden indoor giving a remaining Griffithsin mass of 210 g per batch. The 10 L of eluant process fluid is sent through a viral clearance filter and transferred into a pool/storage tank.

Subsequently, the extract is sent through anultrafiltration/diafiltration cycle to remove salts introduced in the chromatography column. After ultrafiltration, the product is transferred into a storage tank to be mixed with the final formulation components. The concentrated Griffithsin is diluted to give a concentration of 10 g/L Griffithsin in 10 mM Na2HPO4, 2.0 mM KH2PO4, 2.7 mM KCl and 137 mM NaCl at pH 7.4. The final volume of the DS is 21 L per batch. As shown by Figure 4, each batch in the downstream requires 39 h of process time which includes all SIP and CIP operations. As batches move from the upstream portion of the facility every 3.44 days, the remaining time left over in the downstream is set as slack time in the model that may be dedicated toward repair, maintenance, etc. The assumptions and results developed in SuperPro were used to calculate the economics of the process described. Table 2 shows the total operating costs segregated individually for upstream and downstream components. Figure 5 displays process category cost contributions graphically, including percentages of total costs. In upstream operations, the largest cost components are utilities and labor , representing 61% and 32% of total upstream costs, respectively. In downstream operations, labor-dependent costs are the highest contributors at 30% of total downstream costs, followed by consumables at 27% of total downstream costs. Overall, the upstream component represents nearly 57% of the total Griffithsin production cost, which is calculated as just over $106/g protein. For a microbicide dose of 3 mg, the per-dose manufacturing cost is $0.32, excluding any CMO fee. An environmental health and safety assessment was also conducted for this case study following the method of Biwer and Heinzle and the results are found in Supplementary Tables 2–4 in Supplementary Materials. Overall, the process uses chemicals that are not harmful to people or the environment, as can be seen by the low magnitude of input and output Environmental Factor values in Supplementary Table 4. The biggest causes for concern are TMV in the residual biomass, and sodium hydroxide and phosphoric acid used in clean-in-place operations, if released to the environment; however we included costs for a thermal or chemical deactivation step for the TMV-contaminated biomass and pH neutralization for the acid and base cleaning agents which would eliminate the environmental impact of these components.

It should also be noted that the upstream nutrient compounds can be more efficiently recycled to increase nutrient utilization by the plants and reduce water/soil impact. Waste compounds in the downstream process are disposed of through wastewater and bio-waste treatment. An aggregate disposal cost of $0.01 per liter of non-TMV-contaminated aqueous streams and $0.1 per kg of bio-waste is assigned in SuperPro for expenses related to wastewater disposal and thermal/chemical deactivation of bio-waste streams. Compounds introduced during or after the post-inoculation step in the upstream facility are considered as bio-waste since they may contain TMV. This includes spent nutrient solution in the post-inoculation step and retentate streams from plate-and frame and dead-end sterilizing filtration skids. Disposal of TMV-contaminated materials poses low environmental risk. There is extensive industrial experience in disposing of TMV contaminated materials, which can be rendered non-infective by treatment with bleach, heat or detergents, diluted and disposed of as municipal waste . The facility modeled can annually produce 20 kg of the potent antiviral Griffithsin for use in microbicide products. The host used in our modeling was Nicotiana benthamiana. This species was selected because of its aforementioned productivity, but also because our previous report on technoeconomic modeling of Nicotiana-produced therapeutic and industrial products prefaces the work reported herein. In addition, the use of Nicotiana for production of clinical trial materials is also familiar to FDA and other regulatory agencies, thus facilitating Nicotiana’s acceptance in regulation-compliant manufacturing . The API is manufactured in the host Nicotiana benthamiana using tobacco mosaic virus as the expression vector. The upstream plant growth and Griffithsin production operations are adapted from the facility layout detailed byHoltz et al. . Over 158,000 plants are housed in vertically stacked hydroponic grow racks, fitted with high-efficiency LED lights.Each batch of 14,450 plants grows over the course of 38 days and yields a total of 578 kilograms of biomass. Ninety-five batches are seeded and grown annually, with one batch reaching harvest every 3.44 days. The downstream Griffithsin extraction and purification process is scaled up from the pilot industrial scale process presented by Fuqua et al. .

An expression rate of 0.52 grams of Griffithsin per kilogram of biomass and a downstream recovery of 70% were used in the base case and give a combined yield of 0.370 grams of Griffithsin per kilogram of harvested biomass. Sterile filtration and CIP/SIP systems facilitate compliance with cGMP guidelines. Downstream processing commences upon the completion of an upstream batch and takes 39.3 h. The stable final formulation is >99% Griffithsin as the API with negligible endotoxin levels. In the model,vertical garden indoor system the upstream costs account for nearly 57% of the total cost of Griffithsin production. Containing both upstream and downstream losses of the protein could significantly reduce COGS. Approximately 12% of the protein API is non-liberated from the homogenized biomass and 18% is lost during downstream polishing steps. Based on the data and assumptions employed in the current analysis, the unit production cost of Griffithsin is estimated to be $0.32 per dose . The model was based on published designs for a commercial scale facility and pilot-scale data on Griffithsin production adapted to the facility described. This type of modeling is useful for determining ranges of API selling price, production capacity and expression level requirements for commercial supply and profitability. In this study we modeled the manufacturing of Griffithsin through a contract manufacturing organization instead of a greenfield build of a new facility because we assumed that that would be the most prudent approach to launching a new product. If the product manufactured using the process modeled is used directly as a vaginal rinse or rectal enema, the additional costs post manufacturing would include transportation, storage, insurance, distribution, marketing, etc., none of which were modeled in this manufacturer-level analysis. If the Drug Substance produced via the process analyzed is further formulated , or used as a component of another device , those costs and other product-specific costs would be additive and were also excluded from our manufacturer-level analysis. The cost of goods calculated by the current model reflflects the manufacturer’s cost of production. We are less certain about the wholesale price of the drug because there is no standard “offff- the-shelf ” profifit margin that can be added to toll manufacturing cost to arrive at a standardized answer. Often scale up to commercial launch volumes of a product requires additional process development and optimization, validation batches, etc., which lead to negotiated transfer prices depending on volume, duration of engagement, license fees, export duties, and other factors, all of which would impact the cost of bulk Griffithsin. Nevertheless, for this discussion we assumed a manufacturer’s fee of 20% of COGS for a total production cost of bulk Griffithsin Drug Substance of $0.38/dose. Additive formulation, storage, distribution, insurance, marketing, sales margins and other costs could lead to a consumer-level use cost of $1-2/dose . This technoeconomic analysis emphasized Griffithsin’s use in microbicides because such products arguably represent the most price-constrained applications of this new drug. We cannot define the target retail price of a Griffithsin microbicide; there is no market reference price for microbicides since no commercial microbicides yet exist. For perspective, the user cost of a Griffithsin microbicide can be benchmarked against pre-exposure prophylaxis with traditional male condoms and PrEP with microbicides containing antiretroviral drugs as a newer alternative. Analyses have been conducted on the cost of prevention modalities and the cost savings to the healthcare system enabled by preventing HIV transmission, with prevention being far more cost effective than treatment in most scenarios.

Walensky et al. conducted an analysis of the cost-effectiveness of a Tenofovirbased PrEP microbicide in South African women. In their cost modeling of a vaginal gel, they multiplied the product cost of $0.32/dose times 2 and by 7.2 to arrive at a product use cost of approximately $5/woman month. However, the price of the microbicide gel used in the study was assumed and region-adjusted and hence pricing in other countries may be different. Terris-Prestholt et al. estimated Tenofovir gel prices of $0.25–0.33 per dose, provided that the gel was used in combination with a condom , from which an additive cost of use of $7–$12/person-month can be derived. Assuming the same average use rate of a Griffithsin containing microbicide applied singly without a condom and priced at $1.00–$2.00 per dose, the cost of use would be $7– <$15/person-month. Whether a higher cost of use discourages adoption of Griffithsin-based microbicides by men and women remains to be shown. A market study by Darroch and Frost of the Alan Guttmacher Institute consisted of detailed interviews of a cross-section of 1,000 sexually active women aged 18–44 in the continental United States. Their statistically rigorous survey identified levels and predictors of women’s concerns about STDs and interest in microbicides, as well as their preferences regarding method characteristics and likelihood of usage versus price of product, with survey sample results extrapolated to the national level. The results showed that of the estimated 12.6 million women aged 18–44 interested in microbicides and concerned about STDs, including HIV, 11.5 million would still be interested in the method even if it were not 100% effective, and 11.0 million would remain interested even if the microbicide did not protect against STDs other than HIV. The same study found that women’s predicted use of a microbicide was affected by price, but interest was still high at $2 per application, or roughly up to 5-times the average price of a male condom. The survey concluded that more than seven million sexually active women in the USA would be interested in a vaginal microbicide even if the product only protected against HIV, was only 70–80% effective and cost them $2 per application . That conclusion was arrived at in 1999; the $2 per application cost back then would be $3.05 in 2018. One can conclude from these results that there is interest in effective yet inexpensive, self-administered HIV and STD prevention modalities even if such products might cost more than conventional prevention methods.