mRNA-LNP manufacturing process: from IVT to sterile fill

mRNA-LNP is two processes in sequence: synthesis, then encapsulation

The first half makes the drug substance. A plasmid is cut into a linear template, and IVT synthesizes the mRNA from it with T7 RNA polymerase. Capping and a poly(A) tail give the mRNA its function, DNase removes the leftover template DNA, and purification removes dsRNA, short RNA, enzymes and residual DNA.

The second half formulates the product. LNP formation by rapid mixing encapsulates the mRNA. TFF (tangential flow filtration) removes organic solvent and sets the concentration and the buffer. Sterile filtration and filling put the product in vials, and the vials are stored frozen at ultra-low temperature. The step-by-step flow is on the mRNA-LNP process map.

Reading the process as two blocks of a different nature, synthesis and encapsulation, makes everything else easier to follow.

IVT decides yield and the impurity profile, dsRNA above all

mRNA is synthesized in a tube (in vitro), not in living cells. The starting point is plasmid DNA that carries the gene sequence of the target protein.

First, the circular plasmid is cut with a restriction enzyme to make it linear. Left circular, transcription does not stop and RNA that is too long is made. You cut at the position where transcription should end and make a linear template, and only then can IVT be controlled.

To the linear template you add T7 RNA polymerase, the four substrates (ATP, GTP, CTP and UTP) and magnesium, and mRNA is synthesized along the template. This is IVT. It gives a large amount of mRNA in a short time, but byproducts are unavoidable. The typical one is double-stranded RNA (dsRNA), which stimulates innate immunity strongly even in small amounts. The basic approach is therefore two layers: optimize the IVT reaction conditions to limit how much dsRNA forms, and remove what does form reliably in purification.

IVT decides yield and, at the same time, the impurity profile, dsRNA above all.

The 5' cap and 3' poly(A) tail make the mRNA work as a drug

Freshly synthesized mRNA does not work well in the body. Like natural mRNA, it needs a cap structure at the 5' end and a poly(A) tail at the 3' end before it is translated efficiently and resists degradation.

There are two ways to add the cap. In the co-transcriptional method, a cap analog is added to the IVT reaction and the cap goes on at the same time. In the enzymatic method, enzymes add the cap after transcription. The co-transcriptional method keeps the process simple, and the enzymatic method makes it easier to raise the fraction of mRNA that carries a cap. You choose by purpose and process design.

For the poly(A) tail, you can build the poly(A) sequence into the template DNA in advance, or add it after transcription with poly(A) polymerase. Tail length affects translation efficiency and stability, so how tightly the variation in length can be held is a point in quality control.

The 5' cap and the 3' poly(A) tail are essential elements for mRNA to function as a drug.

Purification decides the quality of the mRNA drug substance

The IVT reaction mixture holds more than the target mRNA. It also holds dsRNA, short RNA that broke off midway, the enzymes and nucleotides used in the reaction, and fragments of template DNA that DNase did not fully break down. The job of purification is to remove these and recover only high-purity mRNA.

DNase treatment first breaks down the template DNA. After that, several methods are combined to raise purity. The core is affinity purification with oligo-dT, which is complementary to the poly(A) tail and so selectively captures full-length mRNA that carries poly(A). Chromatography then separates short RNA and dsRNA, and ultrafiltration/diafiltration (TFF) removes nucleotides and salts while concentrating and exchanging the buffer.

The idea of taking out only the target at high purity is shared with antibody purification.

How far you can drive down residual DNA, dsRNA and short RNA decides the quality of the mRNA drug substance.

LNP quality is decided by how you mix

Purified mRNA is not taken up by cells as it is, and it breaks down quickly in blood. So the mRNA is wrapped in a lipid membrane and encapsulated in a nanoscale particle, the LNP. This is the core of the second half.

An LNP usually has four lipids. An ionizable lipid wraps the negatively charged mRNA and releases it inside the cell. A helper lipid (a phospholipid) stabilizes the membrane structure. Cholesterol keeps the particle rigid. A PEG lipid covers the particle surface, prevents aggregation and improves stability in blood.

In LNP formation, an organic phase in which the lipids are dissolved in ethanol and an acidic aqueous phase containing the mRNA are mixed rapidly, for example in a microfluidic device. At the moment of mixing the lipids self-assemble, and particles that have taken in the mRNA form all at once. Mixing speed, flow ratio and pH decide particle size and encapsulation efficiency. The key quality indicators are particle size (roughly tens of nanometers up to about 100 nm), uniformity of the size distribution, and encapsulation efficiency, which shows how much of the mRNA was enclosed. Size and distribution are measured and controlled by dynamic light scattering (DLS).

Particle size, distribution and encapsulation efficiency move with small differences in flow ratio or mixing conditions. At scale-up, keeping the mixing conditions equivalent is the key to carrying the quality over.

A microfluidic mixer used in lab formulation work and a turbulent mixer used in production can change the nature of the particles even with the same formulation. One paper held the formulation fixed, made mRNA-LNP with 10 kinds of mixers, and measured the difference. It is discussed in a paper review on mRNA-LNP mixing methods (in Japanese).

LNP formation is where reproducibility is tested, because particle quality is set by how you mix.

After mixing: ethanol removal, sterile filtration, fill and cold storage

Right after LNP formation, the liquid still holds the ethanol used for mixing and is still in the acidic buffer used for formation. Ultrafiltration/diafiltration on a TFF system removes the ethanol, replaces the acidic buffer with a buffer near neutral that suits dosing, and concentrates the product to the required level.

The final formulation is then sterile filtered through a sterile filter (generally 0.22 µm) and filled into containers under aseptic conditions. LNPs are particles, so you need to confirm in advance that filtration does not cause particle loss or clogging. Filling uses a vial filling machine while fill volume and sterility are controlled.

After filling, low-temperature storage is the premise. mRNA-LNP is sensitive to temperature, and some products must be stored frozen in an ultra-low freezer. The basic rule for quality is to keep the cold chain unbroken from shipping to dosing.

Solvent removal by TFF and an unbroken cold chain together hold up the quality of an mRNA-LNP product.

Eight questions to ask about an mRNA-LNP process

Use these questions with a developer, a CDMO or a supplier. Each follows one step and the quality attribute that belongs to it.

1. How are IVT conditions set to limit dsRNA formation, and where in the process is the dsRNA removed?
2. Is capping co-transcriptional or enzymatic, and what fraction of the mRNA carries a cap?
3. How tightly is the poly(A) tail length held, whether the tail is built into the template or added after transcription?
4. Which purification steps remove residual DNA, dsRNA and short RNA, and how far is each driven down in the drug substance?
5. What mixing method forms the LNP, and what particle size, size distribution and encapsulation efficiency does it give?
6. How are mixing conditions kept equivalent when the scale goes up or the mixer changes?
7. Has sterile filtration been checked for particle loss and clogging with this LNP?
8. What storage temperature does the product need, and how is the cold chain kept unbroken until dosing?