Introduction
Hazardous chemistry can be a key element in the development of commercial scale processes to manufacture APIs. We focus on managing the risks of hazardous chemistry for developing safe and efficient processes on commercial scale. Based on an established pharmaceutical product – Oteseconazole/VT-1161 – we walk you through the project phases from development to scale-up and commercial production, showcasing a process safety methodology and tests. This example will focus on azide chemistry, involving in-situ generated hydrazoic acid as reactive species to introduce a tetrazole moiety. We show that optimized and robust large-scale commercial processes can be developed even though hazardous chemistry is involved or even thanks to the involvement of hazardous chemistry. (1, 2, 3).
Main contents
Oteseconazole is the active pharmaceutical ingredient (API) of VivjoaTM introduced to the market by Mycovia Pharmaceuticals Inc. in April 2022. VivjoaTM is an azole antifungal indicated to reduce the incidence of recurrent vulvovaginal candidiasis (RVCC) in females with a history of RVCC. RVCC (also known as chronic yeast infection) is a debilitating, chronic infectious condition that affects 138 million women worldwide each year.
In our internal process development methodology we are following a process safety protocol to evaluate the potential risk of the species and reaction mixtures. This protocol starts with a first theoretical assessment based on known critical structure elements. Oteseconazole (called VT-1161 during development) contains a tetrazole group (4, 5). Tetrazole moieties belong to the group of hazardous structure elements (figure 1) that lead to safety investigations with respect to thermal and mechanical stability as well as explosive and self-decomposing properties according to transport regulations described in the UN manual of tests and criteria (6).
The hazardous reaction step we are focusing on is the tetrazole synthesis. The tetrazole, substituted at N1, is generated on the last step (figure 2) of a multi-step synthesis of the API in a reaction of a primary amine with trimethyl orthoformate (TMOF), trimethylsilyl azide (TMSA), and sodium acetate in acetic acid as solvent.
The conversion comprises several chemical reactions as shown in figure 3. The primary amine of the precursor is converted with trimethyl orthoformate into the imide which reacts with trimethylsilyl azide (TMSA) or, more precisely, with hydrazoic acid released from TMSA and acetic acid (figure 5). TMSA is manufactured at Axplora from sodium azide and trimethylsilyl chloride at commercial scale. The liquid substance with a boiling point of 95°C is referred to as ‘liquid azide’ due to the ease of dosing compared to solid sodium azide. Even though the nitrogen content of TMSA is high (43%) the reagent is not explosive but classified as flammable liquid (class 3) and toxic substance (class 6.1) because it hydrolyses easily even with humidity in the lung to highly toxic hydrazoic acid(7). Inhalation must thus be absolutely avoided.
At this stage the practical part of our process safety protocol starts with a first determination of the energy content of the used reagents, reaction mixture, and generated product. Initially, we run microcalorimetry DSC (differential scanning calorimetry) experiments. The DSC experiments give an overview about critical temperatures to initiate decomposition reactions (runaways) and energy content. For VT-1161 a decomposition energy of 1375 J/g is shown. This high energy content initiates further tests for mechanical stability and explosive properties. For explosive properties the test series 1 in the UN manual for tests and criteria needs to be applied to have a full picture of the explosive properties and the transport classification.
The safety tests performed at the Leverkusen site of Axplora showed that VT-1161 does not exhibit mechanical instability or explosive properties (table 1).
The product is not mechanically sensitive as it neither had a positive result in the BAM (Bundesanstalt für Materialforschung und -prüfung / Federal Office for Materials Research and Testing) fall hammer test with a drop weight of 10 kg falling from 40 cm height onto the sample (equal to 40 J impact energy as upper limit for positive test result) to initiate decomposition/explosion, nor in the BAM friction apparatus test with 360 N force (upper limit for positive result) applied to the sample. The third test for explosive properties is the so-called Koenen-test, which simulates heating under partial confinement. The product is heated with gas flames in a metal cylinder with a small hole in the lid. A positive result for explosive properties is achieved, if the metal cylinder is shattered into at least three pieces at a hole diameter of 2 mm or bigger in at least one out of three trials.
VT-1161 gave a negative result even at 1 mm borehole diameter. Additional tests revealed VT-1161 to be not a flammable solid as it could not be ignited with a gas burner flame and that it did deflagrate rapidly in the time/pressure test. Relevant for transport classification is the UN gap test. A metal cylinder, placed on a metal plate is loaded with a defined volume of material at the bottom, followed by a PMMA spacer and a defined amount of explosive. The explosive is detonated with an igniter and if the metal cylinder and metal plate is damaged, we have a positive result. In the UN gap test, we obtained a negative result for VT-1161, as it did not rupture the tube and had no impact on the plate when a PMMA spacer was applied but could propagate a detonation without the spacer (figure 4). Overall, Oteseconazole could be classified for transport as no dangerous good.
Besides all the potential risks with the tetrazole bearing VT-1161, we have also a highly hazardous reaction itself, as the TMSA is reacting with acetic acid to form in-situ hydrazoic acid (7) as reactive species for the tetrazole formation (figure 5).
Neat hydrazoic acid (HN3) (7) is a low boiling liquid with a boiling point of 37°C. With an impact energy limit of 0.2 J it is extremely shock sensitive and behaves in this respect similar to the well-known explosive nitroglycerin. The detonation speed of 7000 m/s causes severe damage to equipment. An explosion can already be initiated in a condenser when a drop of hydrazoic acid falls from one coil to the next. Furthermore, heavy metals and especially copper form azide salts which are similarly sensitive. Copper free production equipment is therefore required for running azide chemistry. The gas phase concentration of hydrazoic acid above reaction mixtures must be monitored to avoid explosive concentrations. This can be done continuously by online Fourier transform infrared spectroscopy (FTIR) (8, 9).
The original process development by a former vendor who is not specialised in hazardous and especially azide chemistry relied on literature data(8,9) describing a limit of 11% v/v hydrazoic acid in the gas phase as explosive. This limit can be raised in the presence of volatile solvents like ethyl acetate. Of course, the gas phase concentration of low boiling hydrazoic acid increases with the temperature of the reaction mixture. At 65 °C critical levels of hydrazoic acid were found in the gas phase therefore the synthesis was performed at 50 °C with addition of a diluting solvent and a long TMSA dosing time. The drawback of this process is a very long reaction time of 48 h which is required to reach 99% conversion. Even so the residual imide causes an issue as it is hydrolysed to the corresponding formamide in a side reaction during the aqueous quench of the reaction mixture (see side reaction in figure 3). This by-product is purged during workup by a factor of three which is not sufficient to reach the specified limit of 0.15% formamide impurity in the API. The crude product had thus to be re-crystallised for another purge of two thirds of the impurity to meet the specification of Oteseconazole.
A potential way out of this problem was seen in a continuous flow process (10). But the tetrazole formation is so slow that even at 120 °C reaction temperature a 135 m long pipe reactor with 20 mm diameter would have been needed for the required productivity. At this high temperature a side reaction forming the formamide impurity becomes relevant so that the re-crystallisation would still be needed.
At this point, Mycovia appointed Axplora for a process development of the last step of their API synthesis. Axplora was chosen based on their long track record of experience with hazardous chemistry. All production equipment at the Leverkusen site is specifically designed to run azide reactions up to 12 m³ scale.
Regarding the safe process development from a thermal standpoint, we applied further calorimetric tests using accelerating rate calorimetry and reaction calorimetry. Those experiments were necessary to enable a safe upscaling and transfer to production scale. With those results we were able to evaluate the adiabatic temperature rise and exothermic temperature, which was needed to determine the TMR (time to maximum rate) or ADT24 (adiabatic decomposition temperature at which the reaction mass is stable for 24 h). This temperature was relevant as it set the upper limit for safe handling in production without having an unmanageable risk of runaways when a cooling failure happened. With the calorimetric data, we classified the reaction according to the Stoessel criticality system. A higher criticality class would lead to different scenarios such as process development to reduce criticality, higher level of safety installations (SIL classes), or remote production.
The new process developed used an over-pressure of nitrogen to dilute hydrazoic acid in the gas phase instead of the diluent solvent which slowed down the reaction.
Based on thermal data acquired during process development the temperature could be increased up to 75 °C so that overall 7 h reaction time was enough for reaching an in-process control limit of ≤0.4% residual imide. With this lower limit and the purge factor of three it was ensured that the product purity was above 99.9%, even without a re-crystallisation of the product. Besidesthe chemical yield could be increased from 67 to 86%. Process safety with respect to hydrazoic acid was ensured by online monitoring of its concentration in the gas phase at 3 m³ production scale. In addition, the batch time was decreased form more than two weeks to five days and the process mass intensity (calculated as mass of all chemicals and solvents going into the process by the output of product) was reduced by 55% from 176 to 79 kg/kg.
Overall, the process development lasted ten months from offer to first commercial pilot campaign.
Figure 1. critical structure elements.
Figure 2. Final step of the synthesis of Oteseconazole (VT-1161).
Figure 3. Conversion of the primary amine with TMOF and TMSA to the tetrazole VT-1161.
Figure 4. Outcome of the UN gap test for VT-1161 with (A) and without (B) PMMA spacer.
Figure 5. in situ generated hydrazoic acid as reactive species in the tetrazole formation.
Table 1. Results of safety tests for VT-1161.
References and Notes
- O. G. Bhusnure, R. B. Dongare, S. B. Gholve, P. S. Giram 2018, 12(3), 357.
- M. Blocher, P. Wallimann, Specialty Chemicals Magazine 2008, 28(4), 42.
- Y. Robin, Chimica Oggi 2008, 26(5), 56.
- T. L. Gilchrist, M. J. Alves, Organic Azides 2010, 167.
- D. S. Treitler, S. Leung, J. Org. Chem. 2022, 87(17), 11293.
- UN Recommendations on the Transport of Dangerous Goods – Manual of Tests and Criteria, 5th revised edition (“Orange Book”)
- G. W. Breton, P. J. Kropp, e-EROS Encyclopedia of Reagents for Organic Synthesis
- J. Wiss, C. Fleury, U. Onken, Org. Process Res. Dev. 2006, 10, 349.
- J. Wiss, C. Fleury, C. Heuberger, U. Onken, Org. Process Res. Dev. 2007, 11, 1096.
- M. Movsisyan, E. I. P. Delbeke, J. K. E. T. Berton, C. Battilocchio, S. V. Ley, C. V. Stevens, Chemical Society Reviews 2016, 45(18), 4892.
