See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/356474248

Messenger RNA (mRNA) Vaccine Large Scale Manufacturing – Process Modeling

and Techno-Economic Assessment (TEA) using SuperPro Designer.

Preprint · November 2021

DOI: 10.13140/RG.2.2.34118.80966

CITATIONS READS

2 10,527

2 authors:

Rafael Da Gama Ferreira Demetri Petrides

Intelligen Inc. Intelligen, Inc.
46 PUBLICATIONS 202 CITATIONS 153 PUBLICATIONS 1,143 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Demetri Petrides on 05 December 2021.

The user has requested enhancement of the downloaded file.

 Messenger RNA (mRNA) Vaccine
Large Scale Manufacturing
Process Modeling and Cost Analysis

Using SuperPro Designer®

by

Rafael da Gama and Demetri Petrides

November 2021

This is the ‘ReadMe’ file of a SuperPro Designer example that analyzes the industrial production of
messenger RNA (mRNA) vaccines such as those developed against COVID-19 by Moderna and Pfizer /
BioNTech. This example comes with a detailed and a simplified model of the process. The flowsheet of
the simplified model is appended to the bottom of this document. You may test-drive both the detailed
and simplified models by downloading the functional evaluation edition of SuperPro Designer from the
downloads page of our website (www.intelligen.com). All the files of this example can be found in the
Examples \ Pharmaceuticals \ mRNA Vaccine folder (after the program is installed). The default
installation path of the SuperPro Designer Examples folder follows below.

C:\ Users \ Public \ Public Documents \ Intelligen \ SuperPro Designer \ v12 \ Process Library \ Examples

If you have any questions regarding this example and SuperPro Designer in general, please send an
email message to dpetrides@intelligen.com

INTELLIGEN, INC.
Simulation, Design, and Scheduling Tools
For the Process Manufacturing Industries
www.intelligen.com
Introduction

Messenger RNA, or mRNA for short, is a single-stranded RNA molecule that contains the information of a
DNA sequence in the form of a series of ribonucleotide bases (adenine, uracil, guanine, and cytosine), and
which can be converted into one or more proteins in the cell by the ribosome. mRNA is synthesized in the
cell by the enzyme RNA polymerase using a DNA molecule as a template, in a process called
transcription; in this context, mRNA is sometimes referred to as the transcript. The coding region of the
mRNA molecule is referred to as the coding sequence (CDS); in this region, each triplet of nucleotides is
called a codon and corresponds to a specific amino acid. When mRNA is processed by the ribosome,
these amino acids are recruited and linked to each other through peptide bonds, forming a protein; this
process is called translation. By definition, a DNA sequence that may be transcribed into mRNA and then
translated into a protein is called a gene.

In eukaryotic organisms, in addition to the CDS, the mRNA strand includes a modified guanosine nucleotide
at the 5’ end of the RNA, called a 5’ cap; a poly-adenosine sequence at the 3’ end, called a poly-A tail;
and untranslated regions (UTRs) on both sides of the CDS, as shown in Figure 1.

Figure 1: Structure of a typical eukaryotic mRNA. Source: [1].

In nature, the addition of the 5’ cap (“capping”) is facilitated by a set of enzymes during transcription [2]; it
is crucial for recognition of the mRNA by the ribosome (and therefore for translation) and helps to protect
the mRNA from ribonucleases (i.e., increases its stability) [3–9]. The inclusion of the poly-A tail happens at
the end of transcription, in a process catalyzed by another set of enzymes [2]. The poly-A tail is important
for mRNA stability, protecting the transcript from ribonucleases, and aids translation as well. The 5’ UTR
and the 3’ UTR contain regulatory sequences that affect both translation efficiency [3–11] and mRNA
stability [7–11]; for instance, certain nucleotide sequences in the 5’ UTR can interact with the first transfer
RNA (initiator tRNA) and increase the translation rate [2].

The principle of mRNA vaccines is that, by providing the gene of a pathogenic protein (antigen) in the form
of mRNA to human cells, these cells can synthesize the antigenic protein themselves and thus generate
an immune response that protects the individual from the actual pathogen [3–5,10] (Figure 2). mRNA
vaccines present numerous advantages vis-à-vis DNA vaccines and conventional vaccines such as those
based on inactivated/attenuated pathogens, pathogen fragments or proteins:

1
 • The manufacturing process of mRNA is essentially cell-free and independent of the mRNA
sequence, making it easy and fast to develop and mass produce a new vaccine. mRNA vaccines
may therefore be considered as a platform technology, which makes them particularly
advantageous for tackling pandemics [4,6,8–13]. In contrast, to produce an inactivated or
recombinant protein vaccine, it is necessary to develop specific microbial/cell culture processes
with particular microbial strains or cell lines that usually require a series of long culture steps to
generate the product, as well as custom product recovery and purification steps. Moreover, cell-
free manufacturing has a lower risk of microbial contamination [6].
• Given that the antigenic protein is synthesized from the mRNA in the host cell by the host’s own
cell machinery, the protein generated has exactly the same structural characteristics (including
post-translational modifications) that the actual pathogen’s protein would have in the host [4,7,11].
This is not generally the case when the antigenic protein is industrially manufactured using bacteria
or yeast cells, for instance.
• mRNA does not have to enter the cell nucleus to be decoded, as opposed to DNA in the case of
DNA vaccines.
• mRNA does not pose a risk of DNA integration into the genome, which may exist for DNA vaccines
[4,6,8–10,12,13].

Figure 2: Schematic representation of how mRNA vaccines work. In this drawing, the antigen produced by
the vaccine is a coronavirus Spike protein. Source: [14].

2
mRNA does present, however, certain challenges for its use in vaccines:

• The uptake of naked mRNA (i.e., mRNA that is not associated with some sort of delivery carrier
such as lipid nanoparticles) by cells is very low under normal circumstances [3,6,8–10].
• Naked mRNA is highly unstable in vivo, being rapidly degraded by ribonucleases.
• Exogenous naked mRNA is perceived by the organism as antigenic itself and may elicit a strong
immune reaction. Although this may be useful for vaccination to a certain extent, it may reduce the
translation efficiency of mRNA and consequently hinder the development of an effective immune
response against the antigenic protein [3,4,6,8,10,12,15].

The idea that the injection of exogenous mRNA would lead to the production of the corresponding protein
in vivo was first reported in 1990 in a landmark paper by Wolff and colleagues [4,7]. In this work, mRNA
was injected into muscle tissue of mice, and the presence of the corresponding protein was detected
afterwards [16]. It was only in late 2020, however, that the first two mRNA vaccines were approved (for
emergency use) by regulatory authorities both against COVID-19: one developed by Pfizer / BioNTech and
the other by Moderna. Both vaccines showed remarkable effectiveness against the disease, reaching
efficacy levels of approximately 95% after two vaccine doses in phase III clinical trials [3,8,10,12,17]. The
relatively long interlude between the first mRNA in vivo experiments and the first mRNA vaccine approval
is explained by the substantial challenges posed by mRNA listed above. The main strategies employed by
vaccine developers to tackle these issues have been:

• Nucleoside modification - it has long been observed that replacing ordinary nucleosides with
naturally occurring modified nucleosides (e.g., replacing uridine with pseudo-uridine) reduces the
immune response to exogenous mRNA and increases its stability [3,4,15,6–13]. In fact, both
COVID-19 vaccines from Pfizer and Moderna employ modified N1-methylpseudouridine instead of
uridine. The vaccine from the German company CureVac, on the other hand, relies on ordinary
uridine [8,12,17].
• Capping - the 5’ cap is added to the mRNA strand either during in vitro transcription (IVT), by
including a cap analog in the IVT reaction mixture or after IVT, by using specialized enzymes (e.g.,
2′-O-methyltransferase). Proprietary cap analogs with high capping efficiencies have been
developed for this purpose [6,9]. Co-transcriptional capping tends to give lower yields than post-
transcriptional capping but may be more cost-effective [6].
• Polyadenylation - the addition of the poly-A tail at the 3’ end of the mRNA strand can be done
enzymatically (mimicking the natural process to some extent), or a sequence of deoxythymidine
nucleotides can be appended to the DNA template, so that the poly-A tail is naturally produced
during transcription [5–7,9]. The drawback of the latter approach lies in the production of the DNA
template; this template is usually in the form of plasmid DNA (pDNA) replicated in bacteria, and a
long stretch of identical nucleotides in pDNA makes it structurally unstable [9].

3
• Sequence optimization - the sequences of the UTRs and of the coding region can be optimized to
improve translation efficiency and mRNA stability. UTRs of natural mRNAs such as those of alpha-
or beta-globin are commonly used as the basis for the UTRs of mRNA vaccines [9,10,15,17]. In
the CDS, the degeneracy of the genetic code (i.e., the fact that multiple codons code for the same
amino acid) makes it possible to replace rare codons with frequent codons in human cells; this
procedure is called codon optimization and it generally improves mRNA stability and translation
efficiency [3–7,9,10,12,17].
• Purification (removal of double-stranded RNA) - it has been observed during mRNA manufacturing
that a certain amount of double-stranded RNA (dsRNA) is formed. dsRNA is immunogenic and its
removal is desirable for therapeutic applications [4,6,7,10,17,18]. Purification steps such as
hydrophobic interaction chromatography are capable of distinguishing and separating dsRNA from
single-stranded RNA (ssRNA) [18].
• Development of efficient mRNA vectors - a crucial step for the maturation of mRNA vaccines was
the development of efficient ways to transport mRNA into cells. mRNA itself is polyanionic and thus
negatively charged, as is the surface of the mammalian cell membrane; consequently, they repel
each other electrostatically. This repulsion, together with mRNA’s considerable size, make mRNA’s
natural uptake by cells extremely low [3,6,9,19,20] (less than 1 in 10,000 molecules [3]). Moreover,
naked mRNA is highly susceptible to degradation by ribonucleases, and can provoke severe
immune reactions [3,4,6,8,10,12,15]. Although various mRNA vectors have been developed to
address these issues, including peptides [3,6–8,10], polymers [3,6–8], and dendritic cells [3,6,7,10],
lipid nanoparticles (LNPs) have been the most successful so far. In fact, the two mRNA COVID-19
vaccines that have been approved and all the leading mRNA vaccine candidates for COVID-19
undergoing clinical trials utilize this type of delivery system [3,4,22,8–10,12,13,17,19,21]. LNPs are
self-assembled spherical structures typically composed of four different lipids (a cationic / ionizable
lipid; a PEGylated lipid; a phospholipid; and cholesterol) that are capable of encapsulating mRNA
molecules [3,4,21,22,7–10,12,17,19,20]. LNPs are engulfed by cells through a process called
endocytosis, forming vesicles inside the cell called endosomes. When the lipids of the LNP merge
with the endosome membrane, they release the mRNA to the cytosol, where it can be translated
[3,9,11,13,19,22].

4
Figure 3: Schematic representation of a lipid nanoparticle encapsulating mRNA. Lipid nanoparticles are
typically composed of four lipids: an ionizable / cationic lipid; a PEGylated lipid (PEG-lipid); a phospholipid
(e.g., distearoylphosphatidylcholine a.k.a. DSPC); and cholesterol. Source: [23].

Manufacturing of mRNA Vaccines

In the context of mRNA vaccine manufacturing, mRNA is synthesized enzymatically in a cell-free reaction
called in vitro transcription (IVT). The IVT reaction mixture contains ribonucleotides, a viral RNA
polymerase (e.g., T7 RNA polymerase), and a DNA template (usually plasmid DNA that was previously
linearized with a restriction enzyme) [6,7,10]. In the case of co-transcriptional capping, the cap analog is
added to the reaction mixture as well [6,9,24,25]. A source of magnesium ions (e.g., MgCl2) and a polyamine
such as spermidine are also included; magnesium is required for RNA polymerases to work [26] and
polyamines have been shown to enhance transcription [26,27]. After transcription, a deoxyribonuclease
(DNase) may be added to break down the DNA template and thus facilitate its subsequent removal [5–
7,25]. By the end of IVT, the reaction mixture contains a variety of impurities: buffer components, enzymes,
unused nucleotides and cap analogs, truncated RNA / RNA fragments, dsRNA, DNA template, Mg2+,
spermidine, etc. As such, the single-stranded mRNA must be recovered and purified; typical unit operations
employed for that purpose are crossflow filtration [6,25,28], chromatography [4,6,25,28] and lithium chloride

5
precipitation [4,6,28]. Chromatography is particularly critical to separate and remove dsRNA from single-
stranded mRNA [6,7,10,15,18]; common types of chromatography applied to mRNA purification include:

• Anion-Exchange Chromatography (AEC) - (single-stranded) mRNA, truncated RNA, dsRNA, DNA,

and proteins all have different surface charges and therefore can be separated by anion-exchange
columns [6,18,25,28]. AEC is usually operated in capture mode; due to mRNA’s relatively large
size, however, the nucleic acid binds very strongly to the stationary phase and a heating system is
usually necessary to release the mRNA from the column [6,18]. An advantage of AEC columns is
that they generally have large binding capacities [6].
• Hydrophobic Interaction Chromatography (HIC) & Reverse Phase Chromatography (RPC) - single-
stranded mRNA, truncated RNA, dsRNA, DNA and proteins have distinct degrees of affinity for
hydrophobic ligands and therefore can bind to HIC and RPC resins. Given that certain impurities
(such as proteins) can bind strongly and foul HIC columns, HIC is usually employed as a polishing
step [6,18].
• Oligo-dT Affinity Chromatography (AC) - this type of chromatography takes advantage of mRNA
having a poly-A tail that can non-covalently bind to a stretch of deoxythymidine (dT) nucleotides
due to the natural affinity of adenine for thymine. This is in fact the same A-T interaction found in
the DNA double-helix. As such, oligo-dT chromatography is used in capture mode and removes
most of the impurities present in the IVT mixture [6,18], except for RNA fragments that contain the
poly-A tail and possibly dsRNA and DNA that contain a poly-A tail as well.
• Size-Exclusion Chromatography (SEC) - the differences in length and/or hydrodynamic volume
among mRNA, DNA, proteins and small impurities allow the utilization of SEC columns to purify
mRNA [6]. SEC however suffers from low capacity and tends to dilute the product solution, thus
requiring a concentration step afterwards.

After purification, the mRNA is formulated with a delivery system. Typically, the mRNA is encapsulated in
LNPs in a self-assembly process promoted by rapid mixing of two solutions: an aqueous acidic buffer
containing the mRNA, and an ethanol solution containing several lipids [3,8,10,17,19,22,29,30]. This
process is often performed using microfluidic devices, which allow the production of LNPs with a significant
control over particle size and a high degree of homogeneity [10,19,20,22,30,31]. Particle size affects the
transport of LNPs into the organism and their uptake by cells [9,10,21,22,30], therefore influencing the
efficacy of the vaccine.

Following mRNA encapsulation, it is necessary to remove ethanol and introduce the encapsulated mRNA
in an appropriate formulation buffer. This is typically done through ultrafiltration / diafiltration (UF/DF), which
also facilitates concentration of the product solution [8,24,29,32]. The vaccine is subsequently frozen at
temperatures that range from -20 °C to -80°C for long-term stability [8,12,13,33]. To facilitate storage and
distribution, however, manufacturers are currently evaluating the possibility of freeze-drying mRNA
vaccines [8,10,12,34].

6
Process Description

This example includes two SuperPro Designer files:

• mRNA_Detailed
• mRNA_Simplified

Both files model essentially the same process, the industrial production of mRNA vaccines such as those
developed against COVID-19 in 2020. The process starts with the conditioning of plasmid DNA (the DNA
template), followed by the cell-free synthesis of mRNA in a rocking bioreactor. A crossflow filter is then used
to remove small impurities and to condition the mRNA in an appropriate buffer for the subsequent affinity
(oligo-dT) chromatography step. Oligo-dT chromatography is operated in capture mode, removing most of
the remaining impurities. Next, the mRNA solution is sent to a hydrophobic interaction chromatography
(HIC) step for polishing. A second crossflow filter is used to exchange the buffer and adjust the mRNA
concentration for the subsequent nanoencapsulation stage. In this stage, the mRNA strands are
encapsulated within lipid nanoparticles (LNPs) by mixing the aqueous (acidic) solution containing the mRNA
molecules with an ethanol solution containing several lipids. This blending process is performed using
microfluidic mixers. Lastly, the resulting LNP suspension goes through another crossflow filtration step to
replace the ethanol/water mixture with an appropriate formulation buffer.

File mRNA_Detailed models the complete manufacturing process comprehensively, including buffer
preparation and storage steps. File mRNA_Simplified is a simpler version of the same process lacking
buffer preparation and storage procedures. It should be noted that the process description, scheduling, and
cost analysis presented next are based on the mRNA_Detailedfile. The mRNA_Simplified version is
recommended for students that have access to the academic departmental license of SuperPro Designer
which has a limit of 25 unit procedures per flowsheet.

For reporting and analysis purposes, the process has been divided into three sections:

• RNA Synthesis (green icons)

• RNA Purification (dark red icons)
• Nanoencapsulation (blue icons)

Flowsheet sections in SuperPro are simply sets of related unit procedures (processing steps). For
information on how to specify flowsheet sections and edit their properties, please use the Help tool (Help
Index Section). The contents of each process section are described in greater detail next. The flowsheets
of the two models are appended to the bottom of this document.

RNA Synthesis

7
The process starts with the linearization of circular plasmid DNA (pDNA); linear pDNA serves as the
template for RNA synthesis and is henceforth called Template DNA. It is assumed that the circular pDNA
is acquired from another company or produced at another plant of the same company. Please note that
SuperPro Designer includes a detailed example analyzing the production of pharmaceutical grade pDNA.
It can be found in the Examples \ Pharmaceuticals \ pDNA folder. The default installation path of the
Examples folder is C:\ Users \ Public \ Public Documents \ Intelligen \ SuperPro Designer \ v# \ Process
Library \ Examples.

pDNA linearization is performed in a rocking bioreactor (P-01 / RBR-101), with 1.5 g of pDNA in a 15-L
reaction volume, using a restriction enzyme to cut the circular DNA. The reaction is conducted at 37 °C and
takes 4 h. pDNA linearization is modeled by the following mass stoichiometry.

1 Circular Plasmid DNA → 1 Template DNA (1)

A conversion of 99% was assumed. After pDNA linearization, the restriction enzyme is inactivated by
heating the solution to 65 °C and incubating it for 30 min.

Next, a crossflow filter (P-03 / DF-101) is employed to remove the restriction enzyme and exchange the
restriction buffer with nuclease-free water for injection (WFI). The filtration membrane has a molecular
weight cutoff (MWCO) of 100 kDa, thus retaining the Template DNA. The solution is first concentrated by
a factor of 10 (i.e., to 1 g/L of template DNA), diafiltered with 10 WFI volumes, and then flushed with extra
WFI to maximize DNA recovery.

Next, the Templated DNA solution is sent to the In Vitro Transcription (IVT) step (P-06 / RBR-102), which
is also performed in a rocking bioreactor. The reaction mixture has a total volume of 30 L and includes:

• Tris-HCl, 40 mM (buffer, pH 8)
• Template DNA, 50 mg/L
• Nucleotides (adenosine, guanosine, cytidine and pseudo-uridine triphosphate, abbreviated as ATP,
GTP, CTP and ψTP, respectively), 5 mM each
• A synthetic 5’-cap analog (CleanCap® AG), 4 mM
• T7 RNA Polymerase, 8 kU/mL
• Magnesium Acetate (a source of magnesium ions for T7 RNA polymerase), 16.5 mM
• Murine Ribonuclease Inhibitor (a protein that prevents degradation of the RNA formed), 1 kU/mL
• Dithiothreitol (a reducing agent that prevents oxidation of enzymes), 10 mM
• Spermidine (a polyamine that increases the yield of transcription), 2 mM
• Inorganic Pyrophosphatase (an enzyme that breaks down pyrophosphate into phosphate ions, thus
preventing feedback inhibition from the pyrophosphate released during RNA polymerization as well
as the precipitation of magnesium ions with pyrophosphate), 2 U/mL

8
The reaction is conducted at 37 °C and takes 4 h. RNA synthesis was modeled by the following molar
stoichiometry.

1105 ATP + 1315 CTP + 1061 GTP + 802 ψTP + 1 CleanCap® AG → 1 mRNA + 4283
(2)
Pyrophosphate

The coefficients above are based on the actual sequence of the Pfizer / BioNTech vaccine, which is publicly
available [35]. The mRNA formed in this reaction is complete with a 5’ cap and a 3’ poly-A tail (assuming
that the Template DNA contains a deoxythymidine sequence corresponding to the poly-A tail). A final
concentration of 5 g/L of mRNA is achieved, which corresponds to a reaction conversion of 94%.

RNA Purification

Primary recovery begins with a crossflow filtration step (P-08 / DF-201). The filtration membrane has a
molecular weight cutoff (MWCO) of 100 kDa, retaining mRNA molecules but letting smaller impurities such
as salts, nucleotides and enzymes pass through. The IVT mixture is diafiltered with 10 volumes of
phosphate buffer (NaH2PO4 29.3 mM, Na2HPO4 20.7 mM, NaCl 250 mM, EDTA•Na2 5 mM) and the
membrane is flushed with extra buffer to minimize product loss. The rejection coefficient (RC) of the
membrane during diafiltration is assumed to be 0.998 for mRNA, 1.0 for Template/Plasmid DNA and 0.0
for the other components. As a result, DNA is found in the retentate along with mRNA, and the mRNA
recovery in this step is 98%. After diafiltration and flushing, the retentate is diluted to a concentration of
0.5 g/L of mRNA in preparation for the subsequent oligo-dT chromatography step.

Oligo-dT chromatography (P-12 / C-201) is performed in an 8-L monolithic column, in 8 cycles per batch.
The column binds intact mRNA strands as well as single-stranded mRNA fragments containing the poly-A
tail. In principle, exposed strands of Template DNA or dsRNA may bind to the column as well. Everything
else, however, flows through. It was assumed that the binding capacity of the resin is 3 g/L and the overall
product recovery yield is 95%. Elution is done with a total eluant volume of 8 bed volumes (BVs) and the
product is recovered 2.5 BVs which is collected in a storage tank (P-22 / ET-201). The collected eluate has
a volume of 160 L and mRNA concentration of 0.87 g/L.

The eluate is then sent to a hydrophobic interaction chromatography (HIC) step (P-27 / C-202) to further
purify the mRNA (in particular, to remove mRNA fragments and dsRNA). HIC is performed in a single 8-L
monolithic column, in 5 cycles per batch. The mRNA solution is in-line mixed with a high salt buffer
(NaH2PO4 29.3 mM, Na2HPO4 20.7 mM, NaCl 2 M, EDTA•Na2 5 mM) prior to loading onto the column.
Raising the ionic strength of the mobile phase promotes binding of mRNA to the column, and in-line mixing
prevents mRNA precipitation. It was assumed that the binding capacity of the resin is 4 g/L and the overall
product recovery yield is 80%. Elution is carried out with a linear buffer gradient ranging from 1.6 M NaCl
to 0. A total eluant volume of 50 BVs is charged to the column, and the product is recovered in 20 BVs of

9
collected solution. The collected eluate has a volume of approximately 800 L and mRNA concentration of
0.14 g/L.

The mRNA solution is concentrated and diafiltered with a low salt buffer using another crossflow filter (P-
38 / DF-202). The solution is first concentrated to 5 g/L of mRNA and then, diafiltered with 10 volumes of
storage buffer (citric acid 0.12 mM, sodium citrate 0.88 mM). The membrane is flushed with extra buffer to
minimize product loss. The rejection coefficient (RC) of mRNA was assumed 1.0 during concentration and
0.995 during diafiltration. The overall product recovery yield of this step is 95%. After flushing, the mRNA
solution present in the retentate tank is diluted to a concentration of 0.22 g/L with additional storage buffer.

Nanoencapsulation

During nanoencapsulation, an aqueous solution of mRNA is rapidly mixed with an ethanol solution of lipids,
which leads to the formation of lipid nanoparticles (LNPs) with mRNA strands inside. First, the mRNA
solution is mixed with a concentrated citrate buffer, forming an acidic solution (pH 4.5) containing 0.20 g/L
of mRNA, 31 mM of citric acid and 19 mM of sodium citrate (P-42 / BT-306). The ethanol solution is prepared
by dissolving a mixture of four lipids – cationic (ionizable) lipid, cholesterol, DSPC, and PEG-lipid – in pure
ethanol, according to the following molar ratio: 50.0:46.1:10.1:1.8, respectively (P-47 / BT-302). This ratio
corresponds to a lipid mixture weight composition of 46.3%, 42.7%, 9.4% and 1.6%, respectively. The total
lipid concentration of the ethanol solution is set to 25 mM. The mRNA solution and the lipid solution are
then mixed using four parallel microfluidic systems modeled by a batch generic box procedure (P-46 /
MFX-301). The amount of ethanol solution mixed with the mRNA solution is such that 6 cationic lipids are
added for each mRNA phosphate group (the mRNA molecule contains 4285 nucleotides). This proportion
(6:1) is called the N:P ratio. Consequently, the flow rate ratio of aqueous solution to organic solution is
approximately 3.14 (it usually falls between 1:1 to 5:1). The maximum throughput of each microfluidic mixer
is 12 L/h, so that approximately 15 h are necessary to process all the material using four parallel units. The
formation of LNPs containing mRNA in procedure P-46 is represented by the following mass stoichiometry:

55.95 Cationic Lipid + 26.04 Cholesterol + 11.71 DSPC + 6.30 PEG-Lipid + 3.96 mRNA
(3)
→ 103.96 mRNA-LNP

where mRNA-LNP corresponds to the LNP containing mRNA. The conversion of this reaction, which
corresponds to the encapsulation efficiency, is assumed to be 95%.

After encapsulation, it is necessary to remove the organic solvent and replace the acidic buffer with a
suitable formulation buffer. First, the mRNA-LNP suspension is diluted with phosphate-buffered saline
(PBS) 10× to raise its pH (P-50 / BT-307); then, the suspension is sent to a crossflow filter (P-55 / DF-301)
where it is concentrated to 26.25 g/L of mRNA-LNP (which corresponds to 1.0 g/L of mRNA). The
concentrated mRNA-LNP suspension is diafiltered with 10 volumes of PBS buffer, and finally, the

10
membrane is flushed with extra buffer to minimize product loss. The rejection coefficient (RC) of mRNA-
LNP was assumed 1.0 during concentration and 0.995 during diafiltration. The product recovery of this step
is 95%. By the end of the FLUSH operation, the mRNA-LNP suspension has a total volume of 150 L and
mRNA-LNP concentration of 16.78 g/L (corresponding to 0.64 g/L of mRNA).

Next, the mRNA-LNP suspension is filter-sterilized (P-59 / DE-306) and collected in a blending tank (P-60
/ BT-308). In this tank, the mRNA concentration is adjusted to 0.50 g/L with a sucrose-containing buffer
(PBS with sucrose 34% w/w). However, to perform this concentration adjustment and visualize the amount
of product in terms of mRNA (rather than mRNA-LNP), a pseudo-decomposition of the mRNA-LNP
component was introduced in P-60. This fake decomposition is represented by a chemical reaction whose
stoichiometry is exactly the reverse of eq. (3), with a conversion of 100%. As a result, the product stream
(S-157) contains mRNA and LNP lipids as separate components, rather than mRNA-LNP. A total of 101 g
of mRNA is produced per batch in 202 L of product solution.

Process Scheduling and Cycle Time Analysis

The overall batch time for this process is approximately 64 h. This is the time elapsed from the start of a
given batch (i.e., the setup of the plasmid linearization step) to the end of that batch (the formulation of the
mRNA vaccine drug substance). However, the batch cycle time is set to 24 h; this is possible because all
individual procedures in this process are shorter than 24 h. The user can specify the recipe cycle time
through the Recipe Scheduling Information dialog (Tasks Recipe Scheduling Information…).

To visualize the process schedule, the user may click on Charts Equipment Occupancy Multiple
Batches. This will generate the Equipment Occupancy Chart (EOC) presented in Figure 4. This chart
displays the utilization of each equipment item over time. Six consecutive batches of the mRNA vaccine
are shown, each one associated with a different color. The three process sections – RNA Synthesis, RNA
Purification, and Nanoencapsulation – are also indicated on the chart.

Another view of the process schedule is provided by the Operations Gantt chart (in the MS Project style).
That chart displays detailed scheduling information for one or multiple batches. The Gantt chart for a single
batch is generated by selecting Charts Gantt Charts Operations GC. Figure 5 displays a portion of
that Gantt chart, illustrating the scheduling of operations in the RNA Synthesis section. The golden bar
indicates the duration of the entire recipe, whereas the dark blue and cyan bars represent the duration of
procedures and operations, respectively.

The Gantt chart enables users to visualize the execution of a batch process in detail. It also facilitates
editing of batch recipes. Double-clicking on any of its bars brings up the dialog of the corresponding entity
(e.g., operation, procedure, recipe, etc.). The simulation calculations can then be redone, and the chart can
be updated by clicking on the refresh button of the chart ( ).

11
Furthermore, SuperPro can export its scheduling data to MS Project by selecting File Export to MS
Project XML File. Likewise, SuperPro can export its recipe data to SchedulePro by selecting File Export
to SchedulePro’s Recipe DB. SchedulePro is a resource management, production planning and
scheduling tool available from Intelligen. Please consult the Help facility for information on these two
exporting options (Help Index Exporting).

12
 RNA Synthesis

RNA Purification

Nano-
encapsulation

Figure 4: Equipment Occupancy Chart (EOC) for six batches. Buffer preparation tanks and filters have been omitted for the sake of clarity.

13
Figure 5: Operations Gantt Chart (part of one batch).

14
Material Balances

Table 1 presents overall process data such as batch size, annual production rate and number of batches
per year. This table was extracted from the RTF version of the Materials & Streams report, which can be
generated by selecting Reports Materials & Streams from the main menu bar of SuperPro. The format
of the report can be specified through the dialog that is displayed when you select Reports Options from
the main menu bar. Notice that the batch size is indicated as 96 g of mRNA, slightly lower than the number
mentioned in Process Description (101 g) because a batch failure rate of 5% was specified for this process.
Production failure rates can be defined by right-clicking on an empty space of the process flow diagram and
selecting Economic Evaluation Parameters Production Level. The total number of batches per year
is 334, which leads to a production volume of approximately 32 kg of mRNA per year. Considering that
each vaccine dose has 30 µg of mRNA (which is the actual dose of the Pfizer / BioNTech COVID-19 vaccine
[12]), and that the vaccine is filled into multi-dose vials with an overfill rate of 10%, the current manufacturing
process would be able to produce approximately 970 million vaccine doses per year.

Table 2, which is also extracted from the Materials & Streams report, displays the raw material requirements
in kg/year, kg/batch, and kg/kg of MP (“MP” stands for main product, which is mRNA in this case). It shows
that RNase-free water for injection (WFI), diafiltration and chromatography buffers are the dominant raw
materials in terms of quantity.

Table 1: Overall process data for mRNA vaccine manufacturing

OVERALL PROCESS DATA

Annual Operating Time 47.95 weeks

Unit Production Ref. Rate 32,041.40 g MP/year
Batch Size 95.93 g MP
Recipe Batch Time 63.86 h
Recipe Cycle Time 24.00 h
Number of Batches per Year 334.00
MP = Flow of Component 'mRNA' in Stream 'S-157'

15
Table 2: Material requirements for mRNA vaccine manufacturing

BULK MATERIALS (Entire Process)

Material kg/yr kg/batch kg/kg MP

AC Cleaning Bfr 228,720 684.79 7.14
AC Elution Bfr 169,991 508.96 5.31
AC Loading Bfr 652,319 1,953.05 20.36
AC Wash #2 Bfr 213,120 638.08 6.65
Cap Stock Sln 406 1.21 0.01
Citrate Bfr 10X 18,521 55.45 0.58
Citrate Bfr 1mM 258,780 774.79 8.08
Ethyl Alcohol 43,630 130.63 1.36
HIC Cleaning Bf 135,254 404.95 4.22
HIC Elut Bfr A 894,165 2,677.14 27.91
HIC Elut Bfr B 619,316 1,854.24 19.33
Inhibitor Stock 249 0.75 0.01
IVT Buffer 10x 1,043 3.12 0.03
Lipid Mix 896 2.68 0.03
NTP Mix 502 1.50 0.02
PBS 386,753 1,157.94 12.07
PBS 10X 27,058 81.01 0.84
PBS w/ sucrose 19,638 58.80 0.61
Plasmid 1 0.00 0.00
PPase Stock 199 0.60 0.01
Res Enz Bfr 10X 520 1.56 0.02
Restr Enz Stock 299 0.90 0.01
RNA Polym Stock 1,595 4.77 0.05
TFF San Buffer 55,738 166.88 1.74
TFF Storage Bfr 55,738 166.88 1.74
WFI 233,925 700.37 7.30
TOTAL 4,018,373 12,031.06 125.41

Cost Analysis

SuperPro Designer performs thorough cost analysis, estimating capital costs (CAPEX) as well as operating
costs (OPEX), and generates the following three pertinent reports (through the Reports menu): the
Economic Evaluation Report (EER), the Cash Flow Analysis Report (CFR), and the Itemized Cost Report
(ICR). Table 3 displays the Executive Summary of the Economic Evaluation Report.

The total capital investment for such a facility is approximately $111 million. This includes equipment
purchase and installation costs; other costs related to plant construction; startup and validation costs; and
the working capital required for this project. Plant construction costs such as the cost of buildings and piping
are estimated through multipliers in SuperPro Designer, and these were modified in this example to more
accurately represent the capital costs associated with a vaccine manufacturing facility. For more information

16
on capital cost estimations, please refer to the Industrial Enzymes example in the “…Examples / Bio-
Materials” folder or consult the Help tool (under the topic “Capital Investment Dialog: DFC Tab”).

Table 3 also displays the annual operating cost (AOC) and the unit production cost, which stand at $1.44
billion and $45/mg of mRNA, respectively. Considering the vaccine dose and overfill rate mentioned earlier,
this unit production cost translates to $1.49/dose. Further assuming a selling price of $151.5/mg of mRNA
(i.e., $5/dose) leads to a gross margin of 70% and a return on investment (ROI) of 2,317%. Such a high
ROI results from the low capital investment required by this project.

Table 3: Executive summary for mRNA vaccine manufacturing

EXECUTIVE SUMMARY (2021 prices)

Total Capital Investment 110,615,000 $

Capital Investment Charged to This Project 110,615,000 $
Operating Cost 1,443,294,000 $/yr
Revenues 4,854,271,000 $/yr
Batch Size 95.93 g MP
Cost Basis Annual Rate 32,041 g MP/yr
Unit Production Cost 45,044.66 $/g MP
Net Unit Production Cost 45,044.66 $/g MP
Unit Production Revenue 151,500.00 $/g MP
Gross Margin 70.27 %
Return On Investment 2,317.27 %
Payback Time 0.04 years
IRR (After Taxes) N/A
NPV (at 12.0% Interest) 19,294,918,000 $
MP = Flow of Component 'mRNA' in Stream 'S-157'

Table 4: Annual operating cost breakdown for mRNA vaccine manufacturing

ANNUAL OPERATING COST (2021 prices) - PROCESS SUMMARY

Cost Item $ %
Raw Materials 1,397,740,000 96.84
Labor-Dependent 5,531,000 0.38
Facility-Dependent 9,553,000 0.66
Laboratory/QC/QA 3,318,000 0.23
Consumables 26,880,000 1.86
Waste Treatment/Disposal 158,000 0.01
Utilities 113,000 0.01
TOTAL 1,443,294,000 100.00

Table 4 displays the breakdown of the operating costs, which was also extracted from the Economic
Evaluation Report. Note that raw materials account for almost the totality of the operating cost of this

17
process (approximately 97%). That is the case because mRNA vaccines constitute a brand new technology
that uses expensive (and often proprietary) raw materials such as special lipids and enzymes.

Table 5: Breakdown of material costs for mRNA vaccine manufacturing

MATERIALS COST - PROCESS SUMMARY

Unit Cost Annual Annual Cost

Bulk Material %
($) Amount ($)
AC Cleaning Bfr 28.62 228,720 kg 6,545,464 0.47
AC Elution Bfr 0.60 169,991 kg 102,352 0.01
AC Loading Bfr 0.95 652,319 kg 622,352 0.04
AC Wash #2 Bfr 0.88 213,120 kg 186,866 0.01
Cap Stock Sln 546.12 395,730 mL(STP) 216,117,151 15.46
Citrate Bfr 10X 3.39 18,521 kg 62,853 0.00
Citrate Bfr 1mM 0.50 258,780 kg 130,283 0.01
Ethyl Alcohol 20.00 53,985 L(STP) 1,079,700 0.08
HIC Cleaning Bf 0.91 135,254 kg 123,650 0.01
HIC Elut Bfr A 1.47 894,165 kg 1,310,583 0.09
HIC Elut Bfr B 0.88 619,316 kg 543,246 0.04
Inhibitor Stock 390.00 248,228 mL(STP) 96,808,784 6.93
IVT Buffer 10x 57.64 1,043 kg 60,096 0.00
Lipid Mix 490,717.00 896 kg 439,518,764 31.44
NTP Mix 4,740.28 502 kg 2,380,667 0.17
PBS 0.59 386,753 kg 228,528 0.02
PBS 10X 1.38 27,058 kg 37,261 0.00
PBS w/ sucrose 3.81 19,638 kg 74,827 0.01
Plasmid 50,000.00 501 g 25,050,000 1.79
PPase Stock 257.00 198,582 mL(STP) 51,035,605 3.65
Res Enz Bfr 10X 3.33 520 kg 1,731 0.00
Restr Enz Stock 265.00 297,873 mL(STP) 78,936,393 5.65
RNA Polym Stock 300.00 1,588,657 mL(STP) 476,597,089 34.10
TFF San Buffer 0.70 55,738 kg 39,091 0.00
TFF Storage Bfr 0.54 55,738 kg 30,113 0.00
WFI 0.50 233,925 kg 116,962 0.01
TOTAL 1,397,740,414 100.00

NOTE: Bulk material consumption amount includes material used as:

- Raw Material
- Cleaning Agent
- Heat Transfer Agent (if utilities are included in the operating cost)

This can be confirmed by inspecting Table 5, which presents a breakdown of raw material costs and was
also extracted from the Economic Evaluation Report. It clearly shows that most of the raw material costs

18
are due to process enzymes (especially RNA polymerase), lipids, the 5’-cap synthetic analog, and murine
RNase inhibitor.

Despite the very large raw material costs associated with this process, the results of the cost analysis
suggest that the production of mRNA vaccines can be highly profitable. It should be noted, however, that
many assumptions underlie this kind of analysis, such as the market demand and selling price of the mRNA
vaccine; the prices of raw materials; the yields of RNA synthesis, purification, encapsulation, the
requirement for royalties (in case the technology is licensed from another company), etc. As a result, the
actual economics of such an investment may be substantially better or worse than the current projection.

If the technology of mRNA vaccines matures in the future, the costs of the very expensive raw materials
will likely drop by more than an order of magnitude, drastically reducing the unit production cost and the
contribution of raw materials to the AOC. A useful exercise is to perform “what-if” analyses with SuperPro
to determine the impact of various changes (e.g., a higher in vitro transcription yield, a lower price for RNA
polymerase, etc.). This allows researchers to understand the potential risks and rewards of a project under
different sets of assumptions. What-if scenarios can be evaluated individually (by simply changing
parameter values manually and re-running the simulation to see the results), or they can be automated
through MS Excel. For information on how to drive SuperPro Designer through MS Excel and automate
sensitivity analysis, please consult the examples in the “…Examples \ COM” folder. Related information is
available in the Help facility of the tool, which can be accessed by selecting Help  COM Interface and
Library. Math optimization and Monte Carlo simulation can be performed in a similar manner.

References

[1] Messenger RNA n.d. https://en.wikipedia.org/wiki/Messenger_RNA (accessed October 29, 2021).

[2] Watson JD, Baker TA, Bell SP, Gann A, Levine M, Losick R. Mechanisms of Transcription. Mol.
Biol. Gene. 7th ed., New York, NY: Cold Spring Harbor Laboratory Press; 2013, p. 493–5.

[3] Kowalzik F, Schreiner D, Jensen C, Teschner D, Gehring S, Zepp F. mRNA-Based Vaccines.

Vaccines 2021;9:390. doi:10.3390/vaccines9040390.

[4] Park JW, Lagniton PNP, Liu Y, Xu R-H. mRNA vaccines for COVID-19: what, why and how. Int J
Biol Sci 2021;17:1446–60. doi:10.7150/ijbs.59233.

[5] Jackson NAC, Kester KE, Casimiro D, Gurunathan S, DeRosa F. The promise of mRNA vaccines:
a biotech and industrial perspective. Npj Vaccines 2020;5:3–8. doi:10.1038/s41541-020-0159-8.

[6] Rosa SS, Prazeres DMF, Azevedo AM, Marques MPC. mRNA vaccines manufacturing: Challenges
and bottlenecks. Vaccine 2021;39:2190–200. doi:10.1016/j.vaccine.2021.03.038.

19
[7] Pardi N, Hogan MJ, Porter FW, Weissman D. mRNA vaccines — a new era in vaccinology. Nat Rev
Drug Discov 2018;17:261–79. doi:10.1038/nrd.2017.243.

[8] Buschmann MD, Carrasco MJ, Alishetty S, Paige M, Alameh MG, Weissman D. Nanomaterial
Delivery Systems for mRNA Vaccines. Vaccines 2021;9:65. doi:10.3390/vaccines9010065.

[9] Kim J, Eygeris Y, Gupta M, Sahay G. Self-assembled mRNA vaccines. Adv Drug Deliv Rev
2021;170:83–112. doi:10.1016/j.addr.2020.12.014.

[10] Pilkington EH, Suys EJA, Trevaskis NL, Wheatley AK, Zukancic D, Algarni A, et al. From influenza
to COVID-19: Lipid nanoparticle mRNA vaccines at the frontiers of infectious diseases. Acta
Biomater 2021;131:16–40. doi:10.1016/j.actbio.2021.06.023.

[11] Ho W, Gao M, Li F, Li Z, Zhang X, Xu X. Next‐Generation Vaccines: Nanoparticle‐Mediated DNA

and mRNA Delivery. Adv Healthc Mater 2021;10:1–17. doi:10.1002/adhm.202001812.

[12] Schoenmaker L, Witzigmann D, Kulkarni JA, Verbeke R, Kersten G, Jiskoot W, et al. mRNA-lipid
nanoparticle COVID-19 vaccines: Structure and stability. Int J Pharm 2021;601:120586.
doi:10.1016/j.ijpharm.2021.120586.

[13] Borah P, Deb PK, Al-Shar’i NA, Dahabiyeh LA, Venugopala KN, Singh V, et al. Perspectives on
RNA Vaccine Candidates for COVID-19. Front Mol Biosci 2021;8:1–18.
doi:10.3389/fmolb.2021.635245.

[14] National Human Genome Research Institute. Understanding COVID-19 mRNA Vaccines n.d.
https://www.genome.gov/about-genomics/fact-sheets/Understanding-COVID-19-mRNA-Vaccines
(accessed October 29, 2021).

[15] Pardi N, Hogan MJ, Weissman D. Recent advances in mRNA vaccine technology. Curr Opin
Immunol 2020;65:14–20. doi:10.1016/j.coi.2020.01.008.

[16] Wolff JA, Malone RW, Williams P, Chong W, Acsadi G, Jani A, et al. Direct Gene Transfer into
Mouse Muscle in Vivo. Science (80- ) 1990;247:1465–8. doi:10.1126/science.1690918.

[17] Verbeke R, Lentacker I, De Smedt SC, Dewitte H. The dawn of mRNA vaccines: The COVID-19
case. J Control Release 2021;333:511–20. doi:10.1016/j.jconrel.2021.03.043.

[18] Gagnon P, Goričar B, Peršič Š, Černigoj U, Štrancar A. Two new capture options for improved
purification of large mRNA. Cell Gene Ther Insights 2020;6:1035–46. doi:10.18609/cgti.2020.114.

[19] Roces CB, Lou G, Jain N, Abraham S, Thomas A, Halbert GW, et al. Manufacturing Considerations
for the Development of Lipid Nanoparticles Using Microfluidics. Pharmaceutics 2020;12:1095.
doi:10.3390/pharmaceutics12111095.

20
[20] Huang J, Yuen D, Mintern JD, Johnston APR. Opportunities for innovation: Building on the success
of lipid nanoparticle vaccines. Curr Opin Colloid Interface Sci 2021;55:101468.
doi:10.1016/j.cocis.2021.101468.

[21] Thi TTH, Suys EJA, Lee JS, Nguyen DH, Park KD, Truong NP. Lipid-based nanoparticles in the
clinic and clinical trials: From cancer nanomedicine to COVID-19 vaccines. Vaccines 2021;9:1–29.
doi:10.3390/vaccines9040359.

[22] Aldosari BN, Alfagih IM, Almurshedi AS. Lipid Nanoparticles as Delivery Systems for RNA-Based
Vaccines. Pharmaceutics 2021;13:206. doi:10.3390/pharmaceutics13020206.

[23] File:Vaccines-09-00065-g001.webp n.d. https://commons.m.wikimedia.org/wiki/File:Vaccines-09-

00065-g001.webp (accessed October 29, 2021).

[24] Kis Z, Kontoravdi C, Shattock R, Shah N. Resources, production scales and time required for
producing RNA vaccines for the global pandemic demand. Vaccines 2021;9:1–14.
doi:10.3390/vaccines9010003.

[25] Kis Z, Shattock R, Shah N, Kontoravdi C. Emerging Technologies for Low‐Cost, Rapid Vaccine
Manufacture. Biotechnol J 2019;14:1–2. doi:10.1002/biot.201970055.

[26] Kartje ZJ, Janis HI, Mukhopadhyay S, Gagnon KT. Revisiting T7 RNA polymerase transcription in
vitro with the Broccoli RNA aptamer as a simplified real-time fluorescent reporter. J Biol Chem
2021;296:1–14. doi:10.1074/jbc.RA120.014553.

[27] Frugier M, Florentz C, Hosseini MW, Lehn J-M, Giegé R. Synthetic polyamines stimulate in vitro
transcription by T7 RNA polymerase. Nucleic Acids Res 1994;22:2784–90.
doi:10.1093/nar/22.14.2784.

[28] Schmidt A, Helgers H, Vetter FL, Juckers A, Strube J. Digital Twin of mRNA-Based SARS-COVID-
19 Vaccine Manufacturing towards Autonomous Operation for Improvements in Speed, Scale,
Robustness, Flexibility and Real-Time Release Testing. Processes 2021;9:748.
doi:10.3390/pr9050748.

[29] Fahr S, Thiel L, Sengoba C. Mobile on-Demand (MOD) mRNA Vaccine Production: A Design and
Optimal Location Study 2021.

[30] Shepherd SJ, Issadore D, Mitchell MJ. Microfluidic formulation of nanoparticles for biomedical
applications. Biomaterials 2021;274:120826. doi:10.1016/j.biomaterials.2021.120826.

[31] Forigua A, Kirsch RL, Willerth SM, Elvira KS. Recent advances in the design of microfluidic
technologies for the manufacture of drug releasing particles. J Control Release 2021;333:258–68.
doi:10.1016/j.jconrel.2021.03.019.

21
[32] Classification P. METHODS FOR PREPARING PARTICLES AND RELATED COMPOSITIONS. US
2018 / 0243230 A1, 2018.

[33] Crommelin DJA, Anchordoquy TJ, Volkin DB, Jiskoot W, Mastrobattista E. Addressing the Cold
Reality of mRNA Vaccine Stability. J Pharm Sci 2021;110:997–1001.
doi:10.1016/j.xphs.2020.12.006.

[34] Kis Z, Kontoravdi C, Dey AK, Shattock R, Shah N. Rapid development and deployment of high‐
volume vaccines for pandemic response. J Adv Manuf Process 2020;2:1–10.
doi:10.1002/amp2.10060.

[35] World Health Organization. Messenger RNA encoding the full-length SARS-CoV-2 spike
glycoprotein 2020.

22
View publication stats