Figure 1. Manufacturing of Roquinimex. Block diagram of
the final process step.
3 in a high yield. The last step, preparation of the carbox-
amide Roquinimex (4), was done via coupling of the
carboxylic acid 3 with N-methyl aniline, resulting in a high-
yielding product of high quality.
Figure 2. In vitro dissolution data of Roquinimex tablets. The
graph shows the batch specific mean value (%) of dissolved
Roquinimex for the formulated tablets (content 2.5, 5, and 10
mg) after 60 min (see ref 5).
The final manufacturing step can be seen as an operation
divided into three separate stages (Figure 1). The first stage
is the amidation of 3 with N-methyl aniline using 1,3-
dicyclohexylcarbodiimide (DCC) as a dehydrating coupling
agent. The crude Roquinimex, consisting of a mixture of
product and N,N′-dicyclohexyl urea (DCU), is then isolated
and dried. In the second stage, the purification procedure,
the crude Roquinimex is “titrated” in deionized water by
addition of a 2 M sodium hydroxide solution. Roquinimex
dissolves, and a plate filter removes the DCU. In the third
stage, the crystallization procedure, addition of hydrochloric
acid (5 M) at room temperature to the filtrated solution
induces precipitation of the product. The substance is isolated
by centrifugation and washed with deionized water before
drying.
Full-Scale Manufacturing. As the project proceeded
successfully, in terms of positive outcome of the clinical
trials, it was decided that Roquinimex should be produced
in full-scale batches and the process validated. The batches
F04-F06 were manufactured in the scale of 3 × 60 kg
analogous to previous pilot-batch protocols (entries 4-6,
Table 1). The outcome of the campaign was foreseen in terms
of yields, but the impurity profile of the substance was out
of expectation. Two of the batches contained significantly
higher levels of N-methyl aniline than previously observed
(700 and 600 ppm for batch F04 and F05, respectively)
although they were lower than the specified limit at that time
(<1000 ppm).4
Tablet Formulation. The full-scale batches F04-F06
were submitted to pharmaceutical formulation. When the
quality of the tablets was checked, it was realized that the
uniformity in content and in in vitro dissolution of the
manufactured tablets was unsatisfactory (Figure 2). As seen
in Figure 2, the dissolution capacity for the tablets containing
the batches F04 and F05 was significantly lower than for
the tablets containing batch F06, i.e. approximately 15-20%
less dissolved substance after 60 min. Instead, the obtained
average dissolution data for these latter tablets was in good
agreement with the dissolution previously observed for the
tablets produced from the batches P01-P03 (Figure 2). The
medium (0.1 M HCl) used in the tablet dissolution tests was
chosen to simulate the pH of the gastric fluid to give as true
a picture as possible of the bioavailability of the tablet in
vivo.5 For comparison, the dissolution capacities of the pure
Roquinimex substance from both pilot- and full-scale batches
were also investigated (Table 1).6 As seen in Table 1, again
the batches P01-P03 and full-scale batch F06 provided
substances that were substantially easier to dissolve. How-
ever, due to poor wetting (floating aggregates) and due to a
larger amount of substance to be weighed in, a test medium
with higher solubility of Roquinimex including Brij 35 as a
wetting agent had to be used.6 Consequently, the dissolution
data for the tablet versus substance could only be used for
qualitative comparison. Still, it was suspected that the reason
for the anomaly in dissolution capacity for the tablets was
directly linked to an apparent lack of control when crystal-
lizing the Roquinimex substance. Thus, the problem of not
being able to reliably administer the drug orally to the patient
would be an issue for the chemical process development
efforts.
Solid-State Properties of Roquinimex. It is well-known
that the physical properties of a drug substance may be
important for the pharmaceutical formulations containing the
substance.2 For instance, particle size distribution and poly-
morphism may have an influence on the dissolution rate and
the bioavailability of a solid oral formulation. As depicted
in stage 3 (Figure 1, vide supra), the produced Roquinimex
crystals were formed by precipitation from its sodium salt
in aqueous solution by adding hydrochloric acid. Thus, the
precipitation process took place at the interface between the
salt solution and the added acid. Because of the low solubility
of Roquinimex, 0.2 g/L of H2O below pH 4 at 25 °C,7 the
formation and growth of particles were very fast. The
presence of both amorphous structures and a large variety
(5) United States Pharmacopeia, paddle method. Rotation speed: 100 rpm.
Volume: 500 mL. Medium: 0.1 M HCl. Sample: 2.5, 5, or 10 mg
Roquinimex tablet. Temp: 37 °C.
(6) United States Pharmacopeia, paddle method. Rotation speed: 50 rpm.
Volume: 900 mL. Medium: 0.005 M Phosphate buffer pH 6.8 with 0.01%
Brij 35. Sample: ca. 100 mg Roquinimex. Temp: 37 °C.
(7) The solubility of Roquinimex (pKa ) 4.3) in water at 25 °C is strongly pH
dependent. The solubility ranges from 0.2 mg/mL at pH below 4 to 120
mg/mL at pH 6.9. Hansen, B. Roquinimex, the Solubility in Water at pH
3.8 to 6.9 and pKa; Pharmacia Document 21554F: Helsingborg, 1994.
(4) Later, it was realized that the FDA would not accept a limit higher than
100 ppm for the N-methyl aniline content. Consequently, future production
would have to take this 10-fold decrease of N-methyl aniline into account.
Vol. 8, No. 5, 2004 / Organic Process Research & Development
•
803