Introduction

Vaccine production has proven critical for the prevention, control and eradication of infectious diseases, generating profound impact for modern societies towards combating pandemic outbreaks. Traditional methods of vaccine development employ the utilization of live attenuated or inactivated pathogen forms. The first are repeatedly passed through a culture medium or animal cells until the pathogen is weakened, while still remaining immunologically active1. The second are produced by inactivating the viral pathogen by chemical treatment or heat1. Traditional vaccines present limitations in terms of clinical safety and performance. Attenuated vaccines pose significant risk to individuals with weakened immune systems, while inactivated vaccines fail to provide the desired long-lasting immunity, requiring booster shots for immunity to be achieved. Moreover, spanning the basic research of traditional vaccines up to their industrial scale up, hard to solve technical challenges and barriers arise2. These limitations have speared the development of novel technologies to develop either antigens or monoclonal antibodies. Recombinant DNA and RNA technologies were successfully enforced to combat the recent COVID pandemic. Recombinant DNA technology involves the insertion of a gene encoding the pathogen or immunoglobulin sequence of choice into a cell factory organism, i.e. a yeast or bacterium culture, which in turn produces the antigen or the antibody, respectively3. RNA technology on the other hand, leverages stoichiometric biochemical reactions to produce the messenger RNA (mRNA) encoding the antigen or the antibody which will in turn be translated in vivo by the recipient’s cells to the corresponding protein i.e. the in vitro transcribed (IVT) mRNA product that carries the genetic instructions for the patient’s own cells to produce the viral antigen or the immunoglobulin chains4. In the case of the autologously produced viral antigen, the acceptor immune system recognizes it as foreign and consequently mounts an immune response combating the targeted virus. Finally, in the case of the mRNA encoding the immunoglobulin chains of monoclonal antibodies, these are again translated in vivo to produce a fully operational antibody against the virus.

While both mRNA and DNA technologies have established proof of therapeutic effectiveness, they differ widely in their approach to how the therapeutic protein of choice is obtained, as such stipulating different manufacturing processes. Recombinant DNA manufacturing technology is considered time-consuming and expensive, requiring specialized laboratory facilities and trained personnel5. IVT mRNA manufacturing technology on the other hand, is presented as fast, flexible and inexpensive hence the mRNA sequence should be straight forwardly produced and modified in order to target different combinations of antigens and or mutations. On the downturn, as demonstrated by the COVID vaccination practice, mRNA vaccine supply chains require cold transportation and storage owing to the instability and sensitivity of the RNA molecule to nucleases and environmental conditions, respectively. Moreover, manufacturing of IVT mRNA based vaccines has not yet been standardized thus emancipating several quality risks6. Due to the plug-and-play functional advantages, both IVT mRNA and DNA gene-based technologies have captured the global scientific interest. Moreover additional applications of these technologies pave the way for novel approaches to treat cancer, allergic and autoimmune diseases, proteinic imbalances, genetic and neurodevelopment failures.

Despite the profound impact of these technologies, there exist no scientific announcement that compares them in terms of manufacturing in technical and economic level. To substantiate such effort as fruitful, the transfer of both processing technologies to in-silico flowsheet platforms through the development of Digital Shadows is required. Digital Shadows lie in the context of pharmaceutical Industry 5.07 as enabling tools suitable to model a system in a fiat cyber-physical environment8, delivering data flow abstractions of processing performance. They have been  also employed in this study to simulate and then analyse each technology’s technical merits and production costs at given operating conditions of choice, thus investigating root cause deviations and evaluating the cost-effectiveness scenarios of each proposed solution9,10. In this paper, we construct Digital Shadows to compare recombinant manufacturing of monoclonal antibodies and antigens versus IVT mRNA production processes. In order to address this unmet scope, we build novel algorithmic threads to inspect the strengths and weaknesses of each production process line, through stepwise multi parametric analysis. Our efforts are herein streamed to provide to the expert readers the first comprehensive, advanced technoeconomic study comparing these ground breaking technologies and then attempt to determine which of the respective proposed techniques and protocols would be suited for any of the two types of antiviral vaccines considered (antigens or monoclonal antibodies). By this work, we provide enabling tools of strategic decision planning to complement the rational process development and scalable transfer to industrial production of next generation vaccines.

Materials and methods

Recombinant DNA monoclonal antibody production process

Recombinant monoclonal antibody vaccines termed as study Case A.1, are produced by eukaryotic Chinese ovarian cell cultures inside batch reactors which are assumed to be derived by a CHO cell line bank supplied by external vendors that undertake the task of recombinant plasmid synthesis. The said conventional bioproduction is herein presented as  a scalable bioproduction protocol, consisting of four (4) stages.

Batch culture

The cells are grown and multiplied in a batch bioreactor. The culture temperature is settled at the ideal value of 35 °C for proper protein i.e., antigen or monoclonal antibodies production, while agitation conditions are set not higher than 80 rpm. For every batch cycle an initial cell density of 1.7 × 105 cells/mL generated via inoculation is evidently assumed, and the main bioreactor’s volume is considered to be 2 m3. The cell growth obeys first-order Monod kinetics, while the protein production profile is described by the Leudeking-Piret model’s Eq. (2)11:

$$\frac{d{C}_{CHO}}{dt}\frac{1}{{C}_{CHO}}=\mu ={\mu }_{max}\frac{{C}_{gluc}}{{K}_{s}+{C}_{gluc}}$$
(1)
$$\frac{d{C}_{T}}{dt}=\alpha \left(\frac{d{C}_{CHO}}{dt}\right)+\beta {C}_{CHO}$$
(2)

where μ is the specific growth rate (hr−1), μmax the maximum specific growth rate (hr−1), Ks the substrate saturation constant (g L−1), and a and β are the Leudeking-Piret constraints respectively. The bioreactor’s peak cell and produced antigen or monoclonal antibody densities approach the values of 3 × 106 cells/mL and 0.9 g/L respectively driven by a culture growth cycle of 5 days12,13,14. The upstream operating conditions for the presented production scenario are summarized  in Table 1. The amino acids consumption from the culture’s substrate calculation was based on the concentration time profiles derived from proven bibliographic experimental data15.

Table 1 Operating conditions of upstream processing for study Case A.1 DNA recombinant monoclonal antibody production.
Full size table

Primary capture

The broth is separated from the cells via a centrifuge, and  additional filtration steps take place for enhanced purification. The primary capture of the peptidic construct from the broth is performed by Protein A chromatography, through which impurities, such as host cell proteins, are removed19. The columns required were sized based on the Kozeny-Blake equation (Eq. (3))20. Kozeny Blake is used to size every chromatographic column in the current study, while adhering to the same set of assumptions to minimize bias. In Eq. (3), ΔΡ is the pressure drop inside the packed column in (psi), μ is the viscosity of the mobile phase in (cP), K0 is an empirical constant, dp is the particle diameter of the resin beads (in the case of porous membranes, the porous characteristic size) in (μm), ε is the porosity, L is the bed height in (cm), and u is the mobile phase linear velocity in (cm/s). Batch chromatographies demonstrate poorer column utilization than continuous ones, and for a given resin capacity they present lower separation yields21,22. Considering though that Protein A has demonstrated high efficiency and performance in capturing monoclonal antibodies even in batch operations19, a 95% functional yield was evidently assumed.

$$\Delta {\rm P}=\frac{\mu {\rm K}_{0}{\left(1-\varepsilon \right)}^{2}Lu}{{d}_{p}^{2}{\varepsilon }^{3}}$$
(3)

Virus inactivation and polishing

The captured protein is subsequently imposed to the virus inactivation step whereas acid is added in a blended tank to prevent the action of any virus jeopardizing the product’s safety23. The peptidic products are further polished via a cation exchange and a hydrophobic interaction chromatography. These non-affinity chromatographies aid the clearance from viruses, aggregates, leached ligands, and host cell proteins, and for these separation and polishing steps the bioproduct is purified, obtaining an assumed operational yield of 85%24.

Formulation

The polished peptidic construct having undergone the various ultrafiltration and diafiltration processes is subsequently formulated via freeze drying in − 50 ℃25. The lyophilization model is based on three process steps: freezing, sublimation (primary drying) and desorption (secondary drying). The freezing step includes the temperature drop under constant pressure, the sublimation is performed by pressure decrease under the same temperature of the previous step and the water desorption is administered by the removal of the gaseous moisture via temperature raise. The final product is a peptidic based construct obtained as dry powder that can be stored in preservation (2–8 °C) and shipped in room temperature conditions. In Fig. 1 the above mentioned four processing stages of recombinant DNA monoclonal antibody vaccine production are presented. Temperature and flow sensors (S-XXX blocks) have been strategically installed throughout all processes, providing comprehensive data, and enabling precise control at every step. Flow sensors are clearly marked along the associated flows, whereas temperature sensors are installed on the respective unit blocks. Furthermore, Table 2 provides a comprehensive list of all the Kozeny-Blake parameters for the chromatographic columns of study Case A.1, as far as the calculated pressure drop and the column sizing used for determining resin requirements.

Figure 1
figure 1

Process flow diagram of DNA recombinant production of monoclonal antibody vaccines i.e. study Case A.1 (the four processing stages are herein presented with unit blocks); R-101, R-102, R-103 fermentation reactors; D-201 disc-stack centrifuge; F-201, F-202, F-301, F-302, F-303 ultrafiltration filters; PR-201 protein A chromatography; T-301 jacketed mixer; M-201, M-301, M-302, M-303 flow connectors; CE-301 cation exchange chromatography; HI-301 hydrophobic interaction chromatography; FD-401 lyophilizer.

Full size image
Table 2 Kozeny-Blake and column parameters for chromatographies in study Case A.1 (Pressure drop ΔP converted from psi to bar).
Full size table

IVT mRNA-based production processes

The IVT mRNA production refers to the formation of an mRNA vaccine encoding any monoclonal antibody construct i.e. study Case A.2 or antigen i.e. study Case B.2. The chosen transcribed genetic material expresses the peptidic construct of choice. The DNA sequence or the gene of choice is inserted into genetically recombined plasmids, which are then multiplied within a continuous prokaryotic culture inside a perfusion bioreactor. Previously we have described how these stages can be modeled for a complete design of production line for the formulation of a typical mRNA vaccine in a lipid carrier20. For the scope of this paper, we considered that the genetic active ingredient is encapsulated within cationic polymer carriers instead of lipid-based nanocarriers to avoid the side effects of lipid carriers. The proposed predominantly IVT mRNA production is also implemented into four (4) stages:

Perfusion culture

The genetically modified prokaryotic cell culture grows in a perfusion bioreactor following first-order Monod kinetics. An Escherichia-coli perfusion culture is considered, as prokaryotic cultures are more flexible for continuous production than eukaryotic ones. Prokaryotes demonstrate higher specific growth rates, and pose minor requirements for oxygen and carbon sources and higher tolerance in environmental contaminants. Table 3 includes the upstream operating conditions considered for the perfusion bioreactor.

Table 3 Operating conditions of IVT mRNA upstream processing.
Full size table

Figure 2 below presents the perfusion bioreactor's modeling sheet in the Aspen Plus V10 engine input set up, as a detailed breakdown of all rational elemental processing steps.

Figure 2
figure 2

Process flow diagram of Stage 1, common process step for IVT mRNA and recombinant protein vaccine production20. T-101 pre-culture tube; R-101 fermentation reactor; R-102 perfusion bioreactor which includes: M-101 flow connector; P-101, P-102, P-103 pumps; UF-101 ultrafiltration filter; SP-101 flow separator.

Full size image

Cell lysis

The collected prokaryotic cells are then mixed with alkaline solution and inserted in a centrifugal tank filled with glass beads for their membrane to be disrupted, and thus the modified plasmids to be released20. The plasmids are collected from the solution and purified after passing through a hydrophobic interaction, an anion exchange, and a size exclusion chromatographic step. All chromatographic columns are sized using the Kozeny-Blake equation. To assist the continuous operation of the upstream processing, the iterated operation steps in the cell lysis stage and all the downstream processing stages, are performed continuously. Continuous Simulating Moving Bed (SMB) is therefore the chosen model that describes the chromatographies described in the process, and it is designed to include four columns in series with four integrated valves that determine the flow direction after a specific period of time, called switch time20, schematically presented by the process flow diagram of Fig. 3.

Figure 3
figure 3

Process flow diagram of Stage 2 of IVT mRNA vaccine production20. V-201 horizontal vessel; M-201, M-202, M-203, M-204 flow connectors; UF-201, UF-202, UF-203, UF-204 ultrafiltration filters; HIC-201 hydrophobic interaction chromatography; AEX-201 anion exchange chromatography; SEC-201 size exclusion chromatography.

Full size image

Linearization and in-vitro transcription

The collected purified DNA plasmids first turn into linear DNA sequential molecules after being subjected to a solution filled with restriction enzymes. The linear DNA then gets in vitro transcribed to the complementary linear mRNA using RNA bacteriophage polymerase (T7) in the respective reaction. In-vitro transcription is performed inside three batch reactors of 10 L each. The reactors’ cycle, including cleansing and sterilization is considered to last 4 h, but their operation is approached as continuous after utilizing auxiliary vessels and timing schedules to equilibrate their batch operation20. The linearized plasmids and the in vitro transcribed mRNA are purified via an anion exchange SMB chromatography, an affinity and a HPLC SMB chromatography, respectively. The above-mentioned process steps are herein given by the stepwise flow model presented in Fig. 4.

Figure 4
figure 4

Process flow diagram of Stage 3 of IVT mRNA vaccine production20 C-301, C-302, C-303, C-304 containers; P-301, P-302, P-303, P-304 pumps; B-301 jacketed mixer; UF-301, UF-302, UF-303, UF-304, UF-305 ultrafiltration filters; M-301, M-302, M-303, M-304, M-305, M-306, M-307 flow connectors; SP-301, SP-302 flow separators; AEX-301 anion exchange chromatography; R-301, R-302, R-303 reactors; AFC-301 affinity chromatography; HPLC-301 high performance liquid chromatography.

Full size image

Table 4 summarizes the employed Kozeny-Blake parameters, the pressure drop, and the column sizing for the resin requirements determination of the chromatographic columns of the identical Cases A.2 and B.2.

Table 4 Kozeny-Blake and column parameters for SMB chromatographies in Cases A.2 and B.2 (AFC-301 is not included because sizing is based exclusively on bibliographic data20).
Full size table

DEAE-dextran nanoparticles formation

For the delivery of the mRNA sequences, protection from degradation is required, where the latter is commonly induced by the action of various RNAse enzymes. The transcribed mRNA is therefore encapsulated by DEAE-dextran cationic polymeric carriers40, which enable the formulation to self-assemble, encapsulating the genetic material and protecting the RNA nanocrystals from degradation41. DEAE-dextran is considered to be manufactured from basic raw materials instead of being procured by external suppliers, and the production is performed in parallel with the peptidic synthetic production. Refer to Fig. 6 for additional information on the process flow.

Recombinant protein production process

Recombinant protein antigenic vaccines i.e. study Case B.1 are produced by bacterial cell factories. This production process shares certain similarities with the IVT mRNA production. The primary distinction though lies in the absence of stages related to the conversion of DNA to mRNA i.e. the linearization and the in-vitro transcription stage. The upstream production process for these vaccines is identical to that of IVT mRNA vaccines (Fig. 2), with the only difference being the production capacity. Recombinant protein expression necessitates a dosage capacity that is one fold higher than that of IVT mRNA antigenic vaccine (Case B.2). Therefore a perfusion culture of Escherichia coli is employed within a bioreactor capable to produce the same doses with study Case B.2, given that the titer of recombinant protein is 0.9 g/L of culture42. Considering a batch failure rate of 4% and 80% downstream yield43, a bioreactor working volume of 0.69 m3 is required to approach the desired annual production capacity of 105,630 g. Design specifications and the reactor’s key parameters are found in Table 3 of "IVT mRNA-based production processes" section.

The downstream process involves a single protein recovery step via cell lysis, one purification step44, and one formulation step45. These steps align with those used for IVT mRNA-based processes, following Butani et al.'s and Chawla and Yang’s protocols46,47 which involve one anion exchange chromatography (AEX), one hydrophobic interaction chromatography (HIC), and one size exclusion chromatography (SEC). Detailed descriptions of each are provided in "IVT mRNA-based production processes" section. The final phase of the process includes formulation. The product is processed in a lyophilizer, similar to the approach used for recombinant DNA monoclonal antibody vaccines. The downstream process flow for the production of recombinant protein antigenic vaccines is illustrated by Fig. 5. Table 5 summarizes the Kozeny-Blake parameters associated with the chromatographic blocks, the calculated pressure drop and their column sizing.

Figure 5
figure 5

Downstream process of recombinant protein vaccine production (Case B.1). (V-201 horizontal vessel; M-201, M-202, M-203, M-204 flow connectors; UF-201, UF-202, UF-203, UF-204 ultrafiltration filters; HIC-201 hydrophobic interaction chromatography; AEX-201 anion exchange chromatography; SEC-201 size exclusion chromatography; FD-401 lyophilizer).

Full size image
Table 5 Kozeny-Blake and column parameters for SMB chromatographies in Case B.1
Full size table

DEAE-dextran production process

Dextran is produced by prokaryotic cultures. Leuconostoc mesenteroides is the bacterial species most commonly utilized as cell factory for dextran production48, and is also the chosen  species for this study. The culture is kept by a perfusion reactor under the operational conditions analyzed by Table 6.

Table 6 Operating conditions of upstream processing (Dextran production).
Full size table

Dextran exists dissolved in the culture’s broth, and first for its recovery the cells should be harvested from the broth50. For the biomass elimination, a continuous disk-stack centrifuge is employed to support upstream operation. The dissolved dextran is precipitated by mixing it with ethanol in a batch blending tank48, and the process is again approached as continuous by utilizing auxiliary residence vessels20. The recovered dextran is then collected by another disk stack centrifuge. To ensure efficient self-assembly, that encapsulates the genetic material and  creates stable polymeric nanoparticles, dextran should be positively charged40. Cationic charge is herein achieved by reacting with dimethyl-aminoethyl chloride, where three dextran monomers and equal number of dimethyl-aminoethyl chloride molecules react to form one DEAE-dextran monomer in a high pH (basic) environment40. The reaction is described by the following stoichiometry:

$${\text{3 C}}_{{6}} {\text{H}}_{{{1}0}} {\text{O}}_{{5}} + {\text{ 3 C}}_{{4}} {\text{H}}_{{{1}0}} {\text{NCl }} + {\text{ NH}}_{{4}} {\text{OH}} \to {\text{C}}_{{{3}0}} {\text{H}}_{{{6}0}} {\text{N}}_{{3}} {\text{O}}_{{{15}}}{^{{{3} + }}} + {\text{ NH}}_{{3}} + {\text{ H}}_{{2}} {\text{O }} + {\text{ 3 Cl}}^{ - }$$

The quantity of the carrier required to be produced in order to complement the mRNA production capacity was calculated based on the 27:1 (DEAE-dextran : mRNA) molar ratio rule, which generates the same number of the cationic polymer’s amine groups and the genetic material’s phosphate groups (N/P ratio equals one)40. The charged DEAE-dextran is recovered from the basic broth via disk-stack centrifugation, and it is further dried in a spray-dryer to obtain its final form51. The complete process set up flowsheet is exhibited by Fig. 6.

Figure 6
figure 6

Process flow diagram of the dextran production with unit blocks. M-401, M-402, M-403, M-404 flow connectors; R-401 perfusion bioreactor; P-401, P-402, P-403, P-404, P-405, P-406, P-407 pumps; F-401, F-402 filter; SP-401 flow separator; D-401, D-402, D-403 disc-stack centrifuge; C-401, C-402 container; R-402 reactor; DR-401 spray dryer.

Full size image

Technoeconomic analysis

All study cases were evaluated via a discounted cash flow analysis, aiming initially at the estimation of the Minimum Selling Price per Dose (MSPD) (€) and then the determination of the profitability of each production scenario. The manufacturing units are equity-financed operating in the time horizon of thirty (30) years, out of which the first 4 are devoted to construction and start-up purposes. The annual cash flows of the units are calculated according to the following equation:

$${C}_{n}=\left({R}_{n}-CO{M}_{d,n}-{d}_{n}\right)\left(1-t\right)+{d}_{n}$$
(4)

where, Rn is the annual revenues, dn the depreciation rate and t the income tax rate. Also, COMd,n is the annual cost of manufacturing calculated according to Eq. (5), where FCI is the capital investment, CL the cost of labor, CU the utility costs, CR the waste costs, CRM the raw material costs and CC the consumable costs.

$$CO{M}_{d,n}=0.18FCI+2.73{C}_{L}+1.23({C}_{U}+{C}_{R}+{C}_{RM}+{C}_{C})$$
(5)

The CR cost of waste treatment was assumed negligible, while the CL cost of labor was found using as basis the number of operators20. The fixed capital investment was found by calculating the purchase cost of each unit block device using Eqs. (6) and (7). Whereas the Cp,0 cost of provision of a specific equipment in a specific chronological year bearing a characteristic size equal to X0 is known, the Cp,1 cost of the same equipment for the same year with a characteristic size of X1 is calculated according to Eq. (5).

$${C}_{p,1}={C}_{p,0}{\left(\frac{{X}_{1}}{{X}_{0}}\right)}^{a}$$
(6)

The α index is the Lang factor and it is equipment-dependent. Every calculated Cp,1 equipment purchase cost is extrapolated to its associated value Cp,2 that refers to the chronological year of interest by reference to the Chemical Engineering Plant Cost Indexes according to Eq. (7).

$${C}_{p,2}={C}_{p,1}\frac{CEPC{I}_{2}}{CEPC{I}_{1}}$$
(7)

Equipment depreciation is considered, as well, according to the Modified Accelerated Cost Recovery System (MACRS), and the tax rate’s attributed value is 0.21. All the cash flows are discounted to their Net Present Values (NPV) via the following formula:

$$NPV=-{C}_{0}+\sum_{0}^{n}\frac{{C}_{n}}{{\left(1+r\right)}^{n}}$$
(8)

where, C0 is the total capital investment and r the discount rate. The total capital investment is found as a function of the fixed capital investment, and specifically is approximated with the Percentage of Delivered-Equipment Cost for a solid–fluid processing plant52 (Table 7).

Table 7 Percentage of delivered-equipment cost for total capital investment estimation.
Full size table

Risk analysis and MSPD estimation

Recombinant DNA monoclonal antibody i.e. study Case A.1, recombinant protein i.e. study Case B.1, and IVT mRNA study Cases A.2 and B.2 encumber different levels of product risk. Elaborating on the previous statement, the eukaryotic culture and the recombinant proteins are less adaptable to environmental changes which may lead to rejected batches. On the other hand the long-term effects of IVT mRNA vaccines are not clarified and therefore may potentially lead to withdrawals. To present an unbiased comparison, these risks must be quantified and included in the technoeconomic analysis. In this study, the discount rate r was used as the parameter that determines the risk quantification of each production scenario. Decreasing the discount rate results in a higher NPV and  an increased overall profits threshold, and also generates slower profit distribution over time. A higher discount rate, conversely, leads to faster earning of profits, but  decreases the NPV and the overall profit threshold. The futures uncertainty constitutes current profits more valuable, but it also comes with a cost. The discount rate affects two conflicting objectives, (i) the maximization of the economic performance indices, and (ii) the proper time distribution of the profits. This creates a multi-objective optimization problem with the discount rate r as the control parameter. To calculate the optimal value of the decision parameter, a weight of prioritization to each objective must be given herein utilizing the Tchebycheff norm53 (Eq. (9)):

$${U}_{PI-R}=max \left({w}_{PI}\frac{P{I}^{0}-PI}{P{I}^{0}}, {w}_{R}\frac{{R}^{0}-R}{{R}^{0}}\right)$$
(9)

In Eq. (9), wPI is the weight of prioritization for the objective (i), and the metric of interest is chosen to be the Profitability Index (PI), which is calculated using Eq. (10),

$$PI=\frac{\sum_{0}^{n}\frac{{C}_{n}}{{\left(1+r\right)}^{n}}}{{C}_{0}}$$
(10)

PI0 is the value of the PI when the second objective is not considered, and namely it represents its Utopic Value. The wR is the weight of prioritization for the objective (ii), which is quantified using the parameter R. Parameter R represents the ratio of the cumulative discounted cash flow in the halfway point of the unit's lifespan divided by the total discounted profits in the final operational year, with the latter essentially being the Net Present Value (NPV). Respectively, R0 represents the Utopic Value of the second objective. The weights of prioritization, wPI and wR, should sum to one, and the optimal discount rate, ropt, for a given pair of weights is calculated by minimizing the Tchebycheff norm (Eq. (11)),

$${U}_{min}\left({r}_{opt}\right)=min({U}_{PI-R})$$
(11)

Each result of this multi-objective optimization problem represents a Pareto-optimal point, and collectively these points form the respective frontiers for each production scenario, useful for targeted decision-making purposes.

Results

Monoclonal antibody recombinant DNA i.e. study Case A.1, recombinant protein i.e. study Case B.1, IVT mRNA technologies i.e. study Case A.2 and study Case B.2, were separately assessed. Case A.1 was compared to Case A.2 and Case B.1 was compared to Case B.2 for equal number of doses produced. Recombinant DNA (Case A.1) and IVT mRNA (Case A.2) monoclonal antibody vaccine production bear a capacity of 650,000 doses per year each. To achieve this, 104 kg of monoclonal antibodies per year (as indicated in Table 1) within the upstream bioreactor are necessary to be produced for study Case A.1. Similarly, for Case A.2 (as per Table 3), 8110 g per year needs to be produced overall. Recombinant DNA monoclonal antibody vaccine dose weighs between 100 and 300 mg, depending on the patient’s body mass and/or the stage of the disease17. On the other hand, for the expression of the same quantity of the protein by the transcribed mRNA vaccine the required dose is one fold lower compared to Case A.118. Case B.2 antigenic vaccines encumber a capacity of 270,432,000 doses per year, achieved by changing only the dose capacity from study Case A.2 model. Therefore the upstream process of the recombinant protein antigenic vaccine production in study Case B.1 must achieve a protein capacity of approximately 105 kg per year, considering the mentioned batch failure rates and downstream yields.

The equipment sizing and purchase costs, the raw material provisions and cost requirements are summarized herein by Tables 8, 9, 10, 11, 12, 13. The Chemical Engineering Plant Cost Index (CEPCI) chosen value corresponds to the year 2021(https://www.chemengonline.com/2021-cepci-updates-september-prelimandaugust-final/ by online web on 6.5.2023) when the first mRNA vaccines began to enter the market.

Table 8 Equipment sizing and cost of study Case A.1 (monoclonal antibody recombinant DNA production).
Full size table
Table 9 Raw material pricing of study Case A.1 (monoclonal antibody recombinant DNA production).
Full size table
Table 10 Equipment sizing and cost of study Cases A.2 and B.2 (IVT mRNA processes, including dextran production).
Full size table
Table 11 Raw material pricing of study Cases A.2 and B.2 (IVT mRNA processes, including dextran production).
Full size table
Table 12 Equipment sizing and cost of study Case B.1 recombinant protein vaccine production.
Full size table
Table 13 Raw material pricing of study Case B.1 recombinant protein vaccine production.
Full size table

The utilities costs (CU) for all cases are associated with the operation of the main reactors in each of the upstream processes. The costs of labor (CL) were calculated using the number of operators, and 190,476 € per operator was considered as acceptable61. The consumable costs primarily consist of UF/DF membranes, which are assumed to have a lifespan of one batch in all cases, with a cost of €25,000 per batch56,62,63.

Figure 7a and b demonstrate the contributions of each stage of the monoclonal antibody recombinant DNA production Case A.1 to the overall capital investment and the yearly raw material consumption. It becomes evident that upstream processing incurs the least capital expenses but accounts for nearly the entire annual cost of raw materials. Upstream processing of eukaryotic cultures is a demanding task, considering that it requires expensive culture broth and its recycling is unachievable. Chromatographic protein A resin, required to capture primarily the monoclonal antibodies from the culture broth, represents an unavoidable, costly resource process step, one that requires disposal and replacement protocols after 30 cycles of use. Chromatographic resins in the virus inactivation and polishing steps are less costly than the Protein A resin, but still require replacement after 30 chromatographic cycles.

Figure 7
figure 7

Contribution of monoclonal antibody DNA recombinant vaccine production stages on (a) the capital investment (b) the yearly raw material consumption cost (Stage 4 not displayed owing to minor contribution).

Full size image

Figure 8a and b demonstrate the respective contribution of each stage to the overall capital investment and raw material consumption of the IVT mRNA-based vaccine production processes. The cell lysis (Stage 2), the linearization and the in-vitro transcription (Stage 3) stage count for approximately 40% and 45% of the overall capital budget respectively. The production’s upstream process includes one perfusion bioreactor block which is half the size of the bioreactor used for the monoclonal antibody recombinant DNA vaccine production i.e. study Case A.1. Moreover, while one preculture step with one bioreactor is considered for IVT mRNA production, for monoclonal antibody recombinant DNA vaccine production two preculture steps are assumed to be operated by two bioreactors accordingly. Prokaryotic cultures do not require demanding preculture steps, as they offer the ability to grow cells with high rates with less need for precise tuning of environmental parameters. Dextran production does not significantly burden the overall capital budget. The relevant equipment is sized to reach a production target of just 11 kg of dextran yearly, which is correlated to 1.45 times fold the required mRNA production rate48. Regarding the raw material cost contribution, the linearization and in-vitro transcription stage (Stage 3) are responsible for almost all (99.7%) of the yearly raw material expenses, due to unavoidable necessities regarding the procurement of the costly linearization and clean cup enzymes. The minimal raw material costs in the upstream process are primarily due to the recycling of the broth inside the perfusion reactor chamber as well as the use of a less costly culture’s carbon source. Dextran production’s requirements for raw material are also restricted, as its production does not require expensive resources or high capacity.

Figure 8
figure 8

Contribution of IVT mRNA vaccine production stages (Cases A.2 and B.2) (a) the capital investment (b) the yearly raw material consumption cost (Stage 1 and 4 not displayed due to minimal contribution).

Full size image

Figure 9a,b demonstrate the respective contribution of each stage to the overall capital investment and raw material consumption of the recombinant protein antigenic vaccine production i.e. study Case B.1. Results demonstrate that Stage 3 purification steps account for the majority of capital and raw material costs.

Figure 9
figure 9

Contribution of recombinant protein production stages (Cases B.1) (a) the capital investment (b) the yearly raw material consumption cost (Stage 2 and 4 not displayed due to minor contribution).

Full size image

The discounted cash flow analysis pinpoints the estimation of the required MSPD of each production scenario, projecting meaningful insights regarding their investment value. The optimized discount rate for each production case is decided upon the weight of prioritization attributed to each conflicting objective (i) and (ii) (see Risk Analysis and MSPD section). Figure 10a and b demonstrate the Pareto curves of the efficient points and the Utopia Points for the monoclonal antibody vaccine productions for study Cases A.1 and A.2. These are generated by attributing 700 € as selling price, arbitrarily per pharmaceutical dose utilizing nine (9) weight pairs. Figure 10c,d demonstrate the curves for the dose price range between 700 € and 1000 €. Figure 11a and b demonstrate the Pareto curves of the efficient points and the Utopia Points for the antigenic vaccines for Cases B.1 and B.2 respectively, generated by attributing 3 € arbitarily as selling price, while Fig. 11c,d demonstrate the curves for the dose price range between 3 € and 9 €. The numeric values of the discount rate r, i.e. the decision parameter range from 0.1 and 0.4. For each price selection, different Utopia Points are generated, in lieu of changes in the values of the PI, while the R parameter remains unaffected by the price. The ratio R of the cumulative discounted cash flow in the half-life of the unit to the investment NPV depends solely on the discount rate. This explains the PI changes linked to variations in the weights of prioritization, while R remains constant for each weight pair (see Figs. 10c,d, 11c,d).

Figure 10
figure 10

Pareto fronts for monoclonal antibody recombinant DNA and IVT mRNA vaccine manufacturing processes (a) Case A.1, 700 € price per dose, Utopia Point indicated (b) Case A.2, 700 € price per dose, Utopia Point indicated (c) Case A.1, various prices per dose (d) Case A.2, various prices per dose.

Full size image
Figure 11
figure 11

Pareto fronts for antigenic recombinant protein and IVT mRNA vaccine manufacturing processes (a) Case B.1, 3 € price per dose, Utopia Point indicated (b) Case B.2, 3 € price per dose, Utopia Point indicated (c) Case B.1, various prices per dose (d) Case B.2, various prices per dose.

Full size image

Since the monoclonal antibody recombinant DNA vaccine production by CHO cells is a validated and well-establised technique, the weight of prioritization for the objective (i), namely for the maximization of the Profitability Index (PI), is chosen to be 90%. Consequently the weight for the objective (ii) is set at 10%. This assumption generates a discount rate equal to 0.103 by the Tchebycheff norm, which in turn leads to a MSPD of 130 €/dose for study Case A.1. Since  passive treatments using mRNA vaccines are not yet commercialized, the investment carries significant risks in the long-run, and thus the weight for the objective (i) is reduced to 60%. Given this weight, the Tchebycheff norm proposed a discount rate of 0.11 which in turn leads to a MSPD of 365 €/dose for study Case A.2.

When assessing the risk associated with antigenic vaccines, a similar assumption is applied. Specifically, the consistent track record of the recombinant protein (Case B.1) over the years sets the priority for objective (i) at 90%. This results in a MSPD of 0.07 €/dose for Case B.1, in contrast to the corresponding IVT mRNA case i.e. study Case B.2, where the MSPD is 0.88 €/dose (Fig. 11).

Table 14 summarizes the information generated from the discounted cash flow analysis. Monoclonal antibody recombinant DNA production presents significant technical requirements in the upstream process mainly attributed to the nature of the eukaryotes and the batch operation. On the other hand, IVT mRNA vaccine production emancipates flexible, smart cell factories, but the downstream processing of the obtained genetic material is demanding. Dextran production leads to a 40% decrease of the yearly raw material expenditure compaired to lipid carriers20. Recombinant protein antigenic vaccine production (Case B.1), involves significantly reduced cost of raw materials. Regarding the upstream, the use of bacterial cell factories constitutes the procurement of specialised reagents unecessary, while for the downstream no complex processes, like in vitro transcription and linearization, that consume costly enzymes such as T7 polymerases and ECORIs, are performed. This results in superior economic performance when compared to Case B.2.

Table 14 Summary of technoeconomic analysis results.
Full size table

The Pareto frontiers, depicted by Figs. 10 and 11, illustrate that for the recombinant DNA vaccine production (Case A.1), lower end product prices are necessary to achieve similar profitability  to that of the IVT mRNA vaccine production (Case A.2). In addition, the price increase per dose for Case A.1 poses a more significant impact on the profitability indeces, due to the lower capital investment requirements. However, this does not necessarily mean that the recombinant DNA method will always demonstrate superior economic performance. External factors, such as resource availability and accessibility, may impact the volatility of the model. Regarding the antigenic vaccines, the recombinant protein production (Case B.1) demonstrates lower MSPD and higher economic performance due to the lower cost requirements in raw material. Apparently, the results of the technoeconomic analysis represent a pure mathematical conclusion where steady state conditions of production and no changes in the production plan are assumed for the 30-year lifespan horizon. This study has demonstrated that for the monoclonal antibody vaccines, for a given number of doses produced, the DNA recombinant technology (Case A.1) converts more efficiently the investment capital to profits, thus demonstrating higher economical yield. Likewise, for antigen containing vaccines, the recombinant method demonstrates superiority in terms of economic performance, featuring lower MSPD and higher efficiency records.

Moreover, the MSPD can be viewed as a range rather than fixed values, since prices of raw material can vary depending on the supply / demand and market trends. Therefore, a window of pricing instead of strictly defined obtained values should be concidered and this has been specifically applied to the most cost intensive raw material in each case study. In the context of monoclonal antibody recombinant DNA vaccine production i.e. Case A.1, the cost associated with the CHO cell line procurement is identified as the most expensive resource depicted by Fig. 7b. Conversely, for recombinant protein vaccine production i.e. Case B.1, the cost linked to the provision of SEC resin is dominant, as demonstrated by Table 13. Regarding IVT mRNA vaccine production i.e. Case A.2 and Case B.2, the restriction enzyme required to linearize the obtained plasmids is the costliest resource (Table 11). The CHO cell line and the SEC chromatographic resin pricing may demonstrate variability, but it is unlikely to exhibit significant deviations over time. These raw materials are in mainstream use for biopharmaceutical manufacturing for more than 30 years and therefore stable cost budgeting is expected. On the contrary, linearization restriction enzymes were submited to unusual price raise, owing to increased demand during manufacturing of mRNA vaccines against the COVID-19 pandemic. Therefore the expected projections of the price range for establishing the CHO cell bank or the SEC resin are assumed to fluctuate by a 10% range, while for the linearization restriction enzyme by a 50% range. Figure 12a–d illustrate the relationship between the MSPD and the discount rate for each production scenario, considering this expected price fluctuation analysis.

Figure 12
figure 12

MSPD (€) as a function of the discount rate and (a) of the CHO cell line price for the monoclonal antibody recombinant DNA production (Case A.1), (b) of the linearization restriction enzyme price for the IVT mRNA vaccine production (Case A.2), (c) for the SEC resin for recombinant protein production (Case B.1), and (d) of the linearization restriction enzyme price for the IVT mRNA vaccine production (Case B.2).

Full size image

As projected by Fig. 12, the cost of raw materials pose minimal impact on the estimated MSPD in the case of the recombinant protein antigenic vaccine manufacturing i.e. Case B.1. This observation is attributed to the lower raw material requirements in comparison to the capital investment. Furthermore, the price fluctuation of the linearization restriction enzyme poses less impact on the IVT mRNA monoclonal antibody vaccine’s Case A.2 MSPD, despite this fluctuation being notably higher compared to the corresponding one of Case A.1.

Discussion

Recombinant production methods result to highly stable and therefore advantageous products, encompassing a proven track record of clinical safety and efficacy, minimizing the risk of unknown side effects, carrier-related allergic reactions and withdrawals. These vaccines require minimal maintenance to preserve their stability and functionality hence they can be stored at temperature between 2 to 8 °C for long periods, and can be maintained before use at room temperature with no decomposition. Consequently, this feature ameliorates any respective logistical challenges and minimizes the risks related to post-production regulatory withdrawals. From Chinese Ovary Cell and bacterial cell factories to the necessary chromatographic resins, all the components required for production are accessible for the pharmaceutical industry, with negligible chances of future shortages and extreme price fluctuations, even in case when industrial scale quantities are required. However, it is uncertain whether relying on this production scenario could be successful under the capacities required in case of pandemic outbreaks i.e. whether the economies of scale required should be reached in short time. One of the major drawbacks of the recombinant DNA vaccine production platform is its complicated and therefore difficult-to-automate sub-processes, which constitute any standardization efforts a challenge. As a result, this technology requires additional personnel for process supervision and control purposes. Consequently, the above-mentioned intrinsic barriers constitute the platform less appropriate for encountering pandemic bursts or even following up their mutations, whereas in such cases delay of response can cause devastating results to global health systems and to patients. Furthermore, cellular protein production carries high risk of material contamination, particularly in the upstream processing. The cells used are complex and require strict sanitization measures, which increases the likelihood of contamination and costs. In the event of contamination, the cell line and its products must be discarded, resulting in additional delays and hard to predict financial losses. In addition to contamination risks, recombinant DNA and protein production methods feature low production yield due to the nature of the related live processes. For the monoclonal antibody recombinant DNA platform, in the best-case study scenario for every kilogram that is produced in the upstream line, only 600–700 g are obtained in the required by the regulatory agencies’ purity. This means that raw and side materials must be purchased at high quantities and handled by expert personnel to reach the required production capacity, whilst in any case valuable cellular product will still remain underutilized and ultimately discarded.

On the other hand, IVT mRNA manufacturing protocols hold strong competitive advantages related to certain manufacturing characteristics. Even though the mRNA vaccine itself is highly sensitive to environmental conditions, the production process is easier to standardize, automate, adapt and operate in continuous mode due to the synthetic chemical nature of its sub-processes. Regarding the biochemical section of the related processes, i.e., the utilization of prokaryotic cultures as cell factories, offers a less complex, more effective alternative that poses minimal requirements of retaining sterility compared to eukaryotics, thus reducing significantly the risks of cross contamination and the respective quality-related batch rejections, leading to overall higher yields and less product losses. This, in turn, lowers raw material processing resources and reduces product development time contributing to the delivery of products readily distributed to broad markets.

The studied production cases emancipate different levels of investment risk. The utilization of our Pareto norm enabled the quantification of this risk by recruiting the discount rate of the investment as the lead decision parameter. The optimal value of the discount rate for the IVT mRNA production, according to the Pareto norm, equals to 0.110, while for the rest equals to 0.103.

According to the simulation results, recombinant DNA manufacturing demonstrates superior profitability indeces in comparison to IVT mRNA manufacturing for equal number of produced doses. The IVT mRNA vaccine manufacturing approach poses increased raw material requirements due to its intricate processes, i.e. linearization and in vitro transcription.

In the case of monoclonal antibodies produced by recombinant DNA technology, superior economic performance is showcased, with a MSPD of 130 €, lower compared to the IVT-mRNA vaccine bearing a price value of 365 € per dose. Within the context of our second comparative framework, the recombinant protein antigenic vaccine showcased superiority in both economic performance indices and MSDP (0.07 €/dose). This is attributed to a combination of factors, primarily the lower capital investment, and the notably reduced raw material requirements. These advantages offset concerns related to potential batch failures or variations in downstream yields. In addition, this case incorporates a combination of the most cost-effective strategies for both the upstream and downstream processes. This approach leverages the use of bacterial cell factories, avoiding the use of costly raw materials such as those required for the linearization and the related in vitro transcription processes.

Conclusions

For the case of monoclonal antibody products, our simulations revealed higher Profitability Index for recombinant DNA technology compared to IVT mRNA manufacturing. Although recombinant DNA monoclonal antibodies, due to their inferior potency profile, require higher dose i.e. approximately 10 times that of the IVT mRNA, we find that these elevated mass flow quantities are not reflected analogously on the final production cost. It is furthermore exhibited that IVT mRNA manufacturing demonstrates higher dependencies on raw materials leading to higher end product MSPD values.

For the study case of antigenic vaccines, recombinant DNA technology demonstrates higher economic performance requiring reduced capital resources, hence the utilisation of bacterial cell factories and the nature of the related downstream processes, ameliorate the use of costly raw material. Moreover, recombinant vaccine manufacturing encompasses proven, well-grounded protocols for both the upstream and downstream process development. Regarding the clinical field, this intervention does not rely on the human body to translate a pro drug to peptides or proteins, unlike the treatments based on IVT mRNA vaccines that stimulate the anticipated immune response only after translation is performed in vivo with all the related risks and pharmacokinetic variations. Whereas demand of supply is stable allowing for planned production, recombinant manufacturing appears advantageous hence it fulfills all the technical and financial expectations delivering proven track records of safety and efficacy, minimizing the risk of clinical ambiguity.

Whether large production capacities are required immediately such as in cases of global pandemic outbreaks, IVT mRNA significantly shortens the time required from development to clinical application and from benchtop to scale manufacturing, fostering rapid market penetration. In addition, when product flexibility is required to follow up viral mutations, IVT mRNA manufacturing,  due to its synthetic processes and the reduced requirements for installation of large scale production equipment, offers unparalleled advantages. Clinical trials and actual field practice in the future will demonstrate whether mRNA technologies will manage to offer non-inferior therapeutic results compaired to their DNA recombinant established alternatives.