Small-molecule API process: synthesis to crystallization

Impurities in the starting material can travel all the way into the API

The process starts by preparing starting materials, reagents and solvents that meet their specifications. In small-molecule synthesis, the purity and impurity profile of the raw materials link directly to the quality of the final API. An impurity in a starting material can react and create a new impurity, and some of these cannot be removed by later steps and stay in the product.

That is why the choice of starting materials and their specifications draws regulators' attention. ICH Q11 sets out how to designate which materials count as starting materials in the development and manufacture of a drug substance, and how to control the process from there. Where you place the starting material changes the scope of GMP controls. The GMP requirements that apply to API manufacture itself are set out systematically in ICH Q7 (GMP for APIs; see also ICH Q7 explained (in Japanese)).

In small-molecule manufacture this chain can occur: an impurity in the starting material, changed by a reaction, remaining in the final API.

Tighten the specification on starting materials first, because it is the first step of impurity control in the final product.

Multistep reactions multiply yield loss and impurity carry-over

A small-molecule API is almost never finished in one reaction. Reagents are reacted with the starting material in sequence, and through intermediates the target skeleton is built up over several steps, sometimes a dozen or more. Bond formation, oxidation and reduction, protection and deprotection, and functional group transformation proceed along the designed route.

In each reaction, conditions such as temperature, time, reagent equivalents, solvent and pH decide yield and impurities. If a reaction does not go to completion, unreacted starting material remains. If it goes too far or side reactions occur, related substances (impurities structurally similar to the target) increase. So the progress of the reaction is tracked, for example on a chromatography system (in-process control), and the batch moves to the next step at the right moment.

The yield of each step also multiplies. Even with 90% yield per step, ten steps leave an overall yield of about 35%. Optimizing step by step moves the final cost and the quantity you can supply a great deal.

Batch synthesis charges the raw materials to a vessel at once. Continuous synthesis (flow chemistry) reacts them while they flow, and it has been spreading in recent years. The two differ in several ways. In batch, a large amount of an intermediate tends to stay in the vessel during a dangerous reaction, large vessels develop uneven temperature and mixing, and scaling up by building bigger vessels calls for attention to reproducibility. Batch works on general-purpose equipment, so it is flexible.

In flow, the reaction runs in narrow channels, so the holdup is small and dangerous reactions are easier to handle safely. Heat transfer and mixing are fast and the reaction is easy to control. You scale up by running longer. The price is dedicated equipment and process design.

In a multistep synthesis the yield and impurities of every step accumulate, and together they decide the quantity and quality of the final API.

Workup removes most impurities before crystallization, including mutagenic ones

When a reaction ends, the liquid holds not only the target but also leftover reagents, byproducts, inorganic salts and solvents. Workup is the step that removes these and takes out the target.

The typical method is extraction and phase separation, using an organic solvent that does not mix with water and an aqueous phase. The target moves into the organic phase, and water-soluble impurities and salts are driven into the aqueous phase, which gives a rough separation. The organic phase is then washed with water, acid or alkali to remove impurities whose properties are close to the target as well. For some intermediates, the product is taken out once as crystals here and purified.

Among impurities, ones that may be mutagenic have drawn particular attention in recent years. ICH M7 sets out the approach to assessing and controlling such impurities (often called genotoxic impurities), and it serves as a guide to how far workup and crystallization need to bring them down.

Workup removes most impurities by extraction and washing and hands over to crystallization.

Crystallization sets purity, polymorph and particle size

Crystallization is the core of small-molecule API purification. From a solution of the target, you create supersaturation by lowering the temperature, adding an antisolvent or concentrating the solvent, and only the target precipitates as crystals. Impurities stay in the mother liquor (the liquid left after crystals form), so crystallization purifies and solidifies at the same time. It is a powerful step.

Many factors have to be controlled. The type of solvent, the cooling rate, the timing of seed crystals and the strength of stirring decide not just purity but the quality of the crystals. The problem that arises most easily with small molecules is the crystal polymorph. The same molecule can pack into different crystal arrangements, and solubility, stability and ease of formulation change with them. Getting the desired polymorph stably is at the root of quality assurance.

Crystal particle size and shape are also set roughly in crystallization. Particle size affects how easy the later drying and filtration are, and the dissolution of the formulated product.

If an unintended polymorph mixes in, efficacy and stability can be affected. Fixing the crystallization conditions and obtaining the target polymorph reproducibly is the core of quality control here.

Fix the crystallization conditions so that the target polymorph and particle size come out the same every time.

Filtration, drying and milling turn crystals into a powder that meets specification

The crystals from crystallization are suspended in the mother liquor, so the solid and liquid are separated by filtration or centrifugation. With filters, Nutsche filters or centrifuges, you collect the crystals and drain the mother liquor, and if needed wash them with clean solvent to raise purity further.

The collected crystals (wet cake) contain solvent, so they are dried to bring residual solvent down to within specification. Residual solvents matter for safety, and ICH Q3C sets permitted amounts for each solvent. The requirement is to avoid highly toxic solvents (Class 1) as far as possible and to remove the solvents you do use sufficiently by drying. Drying is done in vacuum dryers and similar equipment, taking care that too much heat does not change the polymorph or decompose the product.

After drying, milling brings the particles to a uniform size. Particle size affects content uniformity and dissolution of the formulated product, so it is adjusted to the target particle size distribution, for example with a jet mill. For an API used in sterile products, aseptic filtration or aseptic handling may be built in near the final step, and some designs filter the dissolved API through a sterile filter before solidifying it.

Filtration, drying and milling are the final steps that bring the crystals to a powder that meets specification and is easy to handle.

Impurities fall into three ICH classes, each with its own guideline

Impurity control is at the center of quality control for small-molecule API. There are three broad kinds, and the ICH guidelines manage each of them systematically.

Organic impurities (related substances) are structurally similar to the target and come from side reactions, degradation or the starting material. They fall under ICH Q3A. Residual solvents are the solvents used in reaction, washing and crystallization, under ICH Q3C. Elemental impurities such as metals come from catalysts, reagents and equipment, under ICH Q3D.

For organic impurities, ICH Q3A, which covers new drug substances, sets out the thresholds for reporting, identification and qualification of safety. Reversed-phase HPLC on a chromatography system is widely used to identify and quantify related substances. Impurities on the drug product side are handled by ICH Q3B. In Japan, these principles are applied through the Japanese Pharmacopoeia (JP) and notifications from PMDA.

The impurities you manage are quite different from those of a biologic, but the idea of controlling process-derived impurities systematically is shared. Looking also at the very different mRNA-LNP process shows how the modalities differ.

Match each impurity to its class and to the step that creates it: related substances to Q3A, residual solvents to Q3C, elemental impurities to Q3D.

Seven checks along the route

These checks follow the order of the process. Use them when you review a route or a supplier.

1. Fix where the starting material sits and what its specification is, thinking through ICH Q11. The placement changes the scope of GMP controls.
2. Set in-process controls on every reaction with chromatography, and calculate the cumulative yield. At 90% per step over ten steps, you end up near 35%.
3. Assess mutagenic impurities against ICH M7, and decide where workup and crystallization have to bring them down.
4. Fix the crystallization conditions (solvent, cooling rate, timing of seeding, stirring) so that the target polymorph comes out reproducibly, and set the particle size target.
5. Dry at a temperature that does not shift the polymorph or decompose the product, and confirm residual solvents against ICH Q3C. Avoid Class 1 solvents as far as possible.
6. Mill to the target particle size distribution. For a sterile product, decide whether the dissolved API is sterile filtered before it is solidified.
7. Map related substances (Q3A), residual solvents (Q3C) and elemental impurities (Q3D) to the process steps where they arise.