Removal of acrolein from active pharmaceutical ingredients using aldehyde scavengers

Article abstract of DOI:10.1021/op3000459

A variety of chemical compounds, intermediates, and reagents are used during the process of synthesizing active pharmaceutical ingredients (APIs). Some of these chemicals, intermediates, and reagents, as well as byproducts of synthetic processes, can have toxic properties and be present as impurities at low levels in the API or final drug formulation. If present at high concentrations, the toxic impurities could cause adverse health effects in humans. This paper describes a simple and rapid approach for selective removal of acrolein from APIs using iodixanol as a model API. Several scavengers were tested, and the resins which showed highest binding efficiency and selectivity were chosen for further evaluations. The kinetics of acrolein scavenging in the presence of the API iodixanol and the scavenging capacity of resins were demonstrated in this paper. The most complete scavenging is obtained with PS-NH₂ which removes 97.8% of acrolein without any substantial removal of the API during 20 min of reaction time.

Full text of DOI:10.1021/op3000459

Article
pubs.acs.org/OPRD
Removal of Acrolein from Active Pharmaceutical Ingredients Using
Aldehyde Scavengers
†
,‡
†
†
†
‡
,†
Rustem Kecili, David Nivhede, Johan Billing, Mats Leeman, Bo
̈
rje Sellergren, and Ecevit Yilmaz*
†
MIP Technologies AB, a Subsidiary of Biotage AB, Box 737, 22007 Lund, Sweden
INFU, Universitat Dortmund, Otto Hahn Strasse 6, 44221 Dortmund, Germany
‡
̈
ABSTRACT: A variety of chemical compounds, intermediates, and reagents are used during the process of synthesizing active
pharmaceutical ingredients (APIs). Some of these chemicals, intermediates, and reagents, as well as byproducts of synthetic
processes, can have toxic properties and be present as impurities at low levels in the API or final drug formulation. If present at
high concentrations, the toxic impurities could cause adverse health effects in humans. This paper describes a simple and rapid
approach for selective removal of acrolein from APIs using iodixanol as a model API. Several scavengers were tested, and the
resins which showed highest binding efficiency and selectivity were chosen for further evaluations. The kinetics of acrolein
scavenging in the presence of the API iodixanol and the scavenging capacity of resins were demonstrated in this paper. The most
complete scavenging is obtained with PS-NH which removes 97.8% of acrolein without any substantial removal of the API
2
during 20 min of reaction time.
INTRODUCTION
GTI is considered to be associated with an acceptable risk in
the absence of other data.
■
12
Pharmaceutical genotoxic impurities (GTIs) may induce
genetic mutations, chromosomal breaks, or chromosomal
rearrangements, and have the potential to cause cancer in
One of the potentially genotoxic impurities based on the
structural alerts is acrolein. Acrolein is an α,β-unsaturated
aldehyde that is used as a building block in the production of
1
−3
human.
Therefore, exposure to even low levels of such
15
pharmaceuticals and the highly reactive CC double bond
and CO carbonyl group moieties in the conjugated CC−
CO system are responsible for its electrophilicity. The
available toxicology studies for acrolein have recently been
summarized in a report by the U.S. Department of Health and
impurities present in the final active pharmaceutical ingredient
API) may be of significant toxicological concern.
4
−6
(
The
analysis of these impurities in APIs has received increased
7
−11
12
attention,
and guidelines were recently issued. During
production of active pharmaceutical ingredients (APIs),
reactive intermediates, catalyst, acids, or bases are often used.
These compounds or their derivatives can potentially end up at
trace levels in final pharmaceutical substances. The Viracept
16
Human Services. Acrolein irritates the gastrointestinal organs,
but despite the fact that it readily reacts with biological
nucleophiles, no clear carcinogenic effects related to oral intake
have been found, and a reasonable target limit in
pharmaceuticals may therefore be the ICHQ3A (R2)
qualification threshold of toxicological concern of 0.05% for
(
nelfinavir mesylate) contamination incident is an example of a
case that demonstrates potential dangers of genotoxic
impurities in a pharmaceutical compound. In June 2007, excess
levels of genotoxic sulfonate ester (ethyl mesylate) were
detected in the API of Viracept produced by Roche, and this
led to the global recall of Viracept from all European Union
17
APIs with a maximum daily dose (MDD) of 2 g/day.
Many APIs under development today are produced by
multistep processes requiring purification techniques such as
crystallization, chromatography, and other downstream pro-
cessing approaches. Such processes include several steps which
often result in loss of API and therefore increase the total cost
of the final product. The use of adsorbents and reactive
scavengers for the removal of undesired impurities from
1
3
(
EU) markets.
It is important for process chemists to explore possible
opportunities to avoid the use and generation of these
genotoxic materials in the manufacturing process. Since a
very large number of solutes are used or tested in drug
synthesis and development, no list of target solutes is available,
but rather a list of “structural alert functionalities” is used.
This list includes sulfonates (e.g., ethyl methane sulfonate),
18−23
8
,14
pharmaceutical compounds is a well-known approach;
another approach using membranes for removal of GTIs from
24
active pharmaceutical ingredients was recently reported.
alkylhalides, arylamines, epoxides, etc.
This paper presents the acrolein scavenging performance of
polystyrene and silica based aldehyde scavengers in organic
media in the presence of the API iodixanol. The kinetics of
genotoxic impurity removal and binding capacity of scavengers
is also demonstrated. It should be noted that iodixanol is used
The presence of genotoxic impurities in pharmaceutical
formulations is a serious issue for the pharmaceutical industry.
Therefore, the control of these genotoxic impurities (GTI)
during the development of pharmaceutical compounds is a
growing concern in the pharmaceutical industry. The European
Medicines Agency (EMEA) issued a guideline for GTI limits in
June 2006. According to the guideline, a threshold of
toxicological concern (TTC) value of 1.5 μg/day intake of
Received: February 22, 2012
Published: April 19, 2012
©
2012 American Chemical Society
1225
dx.doi.org/10.1021/op3000459 | Org. Process Res. Dev. 2012, 16, 1225−1229

Organic Process Research & Development
Article
as a model compound and we are not aware of any actual
contamination issues involving this API.
Figure 1 shows the structures of acrolein (GTI) and
that has a high degree of cross-linking and has been
functionalized with difunctional amino groups.
The following scavenging materials are based on silica: Si-
tosyl hydrazine is a silica-supported equivalent of p-
toluenesulfonyl hydrazine; Si-trisamine is a silica-bound
tris(2-aminoethyl)amine, and Si-triamine is a silica-bound
iodixanol (API).
25
tri(2-aminoethyl)amine.
In Figure 2, the schematic structures of the scavengers used
in this work are shown.
Figure 1. Structure of acrolein (GTI) and iodixanol (API).
RESULTS AND DISCUSSION
■
Analysis of Acrolein. The analysis of low-molecular weight
aldehydes, which are highly volatile and polar substances, is
difficult to accomplish by direct RP-HPLC methods. This is
first due to the fact that many such carbonyl compounds have
no chromophores and, thus, are not detectable by UV methods.
Due to the reactive carbonyl group present in those low-
molecular weight aldehydes and ketones, those can easily be
derivatized with, for example, 2,4-dinitrophenylhydrazine
(
DNPH) in order to impart a chromophore. In this particular
case, DNPH reacts in an acidic environment with the carbonyl
group of acrolein to form stable hydrazones (DNPH-carbonyls)
which can be separated by RP-HPLC on a C18 column and are
easily detected by UV. In Scheme 1, the derivatization reaction
of DNPH with an aldehyde or ketone is shown.
Scheme 1. Derivatization reaction of an aldehyde or ketone
Figure 2. Structures of scavengers used in this study.
A recent publication examined the potential use of such
22
resins to remove genotoxic impurities from APIs where the
successful removal of methyl sulfonate esters was reported.
However, related ethyl and isopropyl esters were only partially
removed. Nevertheless, the authors concluded that the use of
such resins showed some potential and suggested that this
could be extended to the other classes of genotoxic compounds.
The various scavenger resins that were evaluated in this study
were based either on cross-linked polystyrene−divinylbenzene
or on porous silica.
The majority of the polymeric scavenger resins are only
slightly cross-linked; i.e. the content of divinylbenzene is
around 1−2%. Only one candidate of the scavenger resins has a
higher degree of cross-linking. More specifically, the following
scavenger resins have a low degree of cross-linking: PS-amine
with aminomethyl groups as the scavenging moiety; PS-
trisamine modified with tris-(2-aminoethyl)amine groups as
multifunctional amine groups; PS-tosylhydrazine modified with
sulphonyl hydrazine groups on its surface as a reactive
scavenging functionality; MP-trisamine is a scavenger resin
We wanted to evaluate the impact of batch mode and flow-
through mode on the scavenging of acrolein in the presence of
iodixanol. This would fundamentally determine the way in
which the scavenger could potentially be used in a real process.
To begin with, the outcome of flow-through scavenging
experiments of acrolein in the presence of iodixanol using
modified polystyrene- and silica-based scavengers is summar-
ized in Figure 3. Trisamine-modified polystyrene showed 74%
acrolein removal with only very little loss of iodixanol.
However, trisamine-modified silica and triamine-modified silica
showed 71.4% and 66.8% acrolein removal, respectively, with
extensive removal of iodixanol. This nonspecific binding can be
expected to be adsorption to acidic silanol groups on the
surface of silica.
In the flow-through mode, the short interaction time
between analyte and resin reduces the effectiveness of
scavenging. For the next step, a batch-scavenging mode
recommended by the supplier of the scavengers was carried
1
226
dx.doi.org/10.1021/op3000459 | Org. Process Res. Dev. 2012, 16, 1225−1229

Organic Process Research & Development
Article
Figure 5. Results of kinetic study for acrolein scavenging. One
milliliter of a solution of 5 μg/mL acrolein and 2 mg/mL iodixanol in
EtOH was added to 75 mg of each scavenger. The solution was stirred,
and an aliquot of 100 μL of each mixture was taken at 0, 0.5, 1, 2, 5, 10,
and 20 min and derivatized by DNPH and assayed by HPLC.
Figure 3. Results of flow-through scavenging of acrolein in the
presence of iodixanol. A quantity of 75 mg of each scavenger was
packed in 1 mL SPE cartridges, and after loading of 1 mL of a solution
of 5 μg/mL acrolein and 2 mg/mL iodixanol in EtOH, 100 μL of each
of the collected aliquots of the sample was derivatized by DNPH and
assayed by HPLC.
of acrolein removal. However, if the process interaction time is
very limited, then PS-trisamine is the scavenger that shows the
fastest scavenging kineticsalready after 2 min ∼90% of
acrolein is removed.
Loading capacity is an important parameter for the resin
selection and for production process economics, and we have
therefore conducted loading capacity determinations of the
polymer-based resins (shown in Figure 6).
out to increase the interaction time and efficiency of scavenging
acrolein.
Batch-scavenging experiments of acrolein in the presence of
iodixanol using modified polystyrene- and silica-based scav-
engers are shown in Figure 4. Most of the scavengers showed
Figure 4. Results of batch scavenging of acrolein in the presence of
iodixanol. A quantity of 75 mg of each scavenger was placed in a
HPLC vial, and after the addition of 1 mL of a solution of 5 μg/mL
acrolein and 2 mg/mL iodixanol in EtOH, each mixture was shaken for
Figure 6. Results of loading capacity of scavengers. One milliliter of a
solution of 5 μg/mL acrolein and 2 mg/mL iodixanol in EtOH was
added to 25, 50, 75, 100, and 200 mg of scavengers. The solutions
stirred for 30 min, and aliquots of 100 μL were derivatized by DNPH
and assayed by HPLC.
30 min, and an aliquot of 100 μL was derivatized by DNPH and
assayed by HPLC.
extensive removal of acrolein. Amine-, multiple amine-, and
hydrazine-modified polystyrene scavengers led to higher than
9
0% acrolein removal with only very little loss of iodixanol,
As can be seen in Figure 6, at least 75 mg of scavenger is
needed for removal of a large amount of acrolein in the
solution. Increasing the amount of scavengers gives no further
improvement in acrolein removal. On the basis of the
assumption that 75 mg of scavenger affords close to complete
removal of acrolein, the capacity of the scavengers can be
whereas silica-based scavengers, on the other hand, resulted in
extensive losses of iodixanol. The less cross-linked polymer-
bound scavengers display a far better selectivity than the silica-
based scavengers and also the macro-porous highly cross-linked
scavengers. Therefore, only the polymer-bound scavengers
based on the less cross-linked polymers were used for further
evaluation. Apparently, under the conditions tested, iodixanol
exhibits a certain affinity for both the silanol surface and the
macro-porous surface of the polymer. Other APIs may behave
differently, and thus, other scavengers may display a more
advantageous selectivity profile than the currently chosen
scavengers.
estimated to be 64.8 μg/g for PS-NH , 60 μg/g for PS-TS-
2
NHNH , and 56 μg/g for PS-trisamine under these conditions.
2
Tosylhydrazine-modified reactive resins have been used for
removal of an aldehyde impurity in a previously reported
23
study, but it is important to consider that leaching of
hydrazine and hydrazine derivatives would be a concern with
this type of resin. The present work shows that PS-NH and
Next, the kinetics of the removal of acrolein using PS-NH2,
2
PS-trisamine resins are equally well suited for removal of
acrolein without this potential problem.
PS-TS-NHNH and PS-trisamine scavengers was studied in
order to determine the reaction duration required for the
2
If we take the model system above with 5 μg/mL acrolein,
then we can transfer the capacity to larger volumes. A 1 L API
solution with 5 mg/L acrolein (5 ppm) would require 75 g of
amine modified polystyrene scavenger to reduce the concen-
tration from 5 ppm down to 0.15 ppm. The initial
concentrations correspond to an impure API with 0.25% of
acrolein; to reduce this level to the ICH - Q3A limit of 0.05%
(80% reduction), ∼50 g of scavenger would be sufficient.
effective scavenging step.
As shown in Figure 5, PS-trisamine binds acrolein very
rapidly and removes over 80% within 2 min. After 20 min, no
scavenger exhibited any significant changes for acrolein binding.
The most effective scavenging is obtained with PS-NH which
2
removes up to 97.8% of acrolein and only 2.0% of iodixanol. If a
process can allow scavenging times of 20 min or more, then all
of the tested scavengers demonstrate a considerably high level
1
227
dx.doi.org/10.1021/op3000459 | Org. Process Res. Dev. 2012, 16, 1225−1229

Organic Process Research & Development
Article
EXPERIMENTAL SECTION
genotoxic compound, acrolein. We have found that, for our
model API process solution system, scavenging of acrolein was
seen to be quite fast and effective using both polymer- and
silica-based scavengers. However, less cross-linked polymer-
based scavengers display a more advantageous selectivity profile
than silica-based scavengers and scavengers with highly cross-
linked polymers; as we have observed, those scavengers
displayed an undesired high level of nonspecific binding to
the API.
In essence, we have described an easy and simple procedure
for effective removal of acrolein from contaminated API
solutions using iodixanol as a model compound. The most
effective and selective scavenging is obtained with PS-amine
which removes up to 97.8% of acrolein and only 2.0% of
iodixanol within 20 min using a batch-mode extraction
procedure. However, a comprehensive genotoxicological risk
assessment on aldehyde impurities in pharmaceutical for-
mulations should be carried out before using this or any other
type of cleanup, as impurities with the same structural alerts
may have similar modes of action and need to be considered
jointly.
■
Materials. Acrolein, 2,4-dinitrophenylhydrazine (DNPH),
formic acid, iodixanol, and also HPLC-grade solvents were
purchased from Sigma-Aldrich (Steinheim, Germany). Scav-
engers were from Biotage GB Limited (Cardiff, UK).
All distilled water used was purified using an ultrapure water
system from Elga (High Wycombe, UK)
Preparation of Stock Solutions. Stock solutions (1 mg/
mL) of acrolein (genotoxic impurity) and iodixanol (API) were
prepared in EtOH and stored in glass flasks in the freezer.
Derivatization of Acrolein with DNPH. To perform
derivatization of acrolein with DNPH, a procedure adapted
2
6
from literature was followed. Briefly, DNPH (50 mg) was
dissolved in 20 mL of acetonitrile and acidified with 0.4 mL of
formic acid. The DNPH solution (12 mM) was stable for 1
week when stored at 4 °C. Derivatization of acrolein was
performed by mixing 100 μL of the sample with 100 μL of the
derivatizing agent and incubating at room temperature for 1 h.
HPLC Analysis. HPLC experiments were carried out on a
Shimadzu LC10 AD equipped with a PDA detector (SPD-
M10A) and an autosampler (SIL-HT ). The column was
A
Supelco Ascentis Express C18 (2.7 μm, 50 mm × 4.6 mm).
Gradient elution was performed with ultrapure water (mobile
phase A) and acetonitrile (mobile phase B).
AUTHOR INFORMATION
■
Corresponding Author
The gradient was started with water/acetonitrile 60:40 (v/v),
then a linear gradient elution up to 55% acetonitrile within 2
min that was raised to 100% acetonitrile in another 2 min.
Then the gradient was decreased to 40% acetonitrile in 1 min.
The final composition was maintained for 3 min before
reequilibrating the column with the initial mobile phase (water/
acetonitrile 60:40 (v/v)). Flow rate was 0.6 mL/min, detection
wavelengths were 370 nm (for acrolein) and 254 nm (for
iodixanol), and the injection volume was 10 μL.
Flow-through Scavenging Procedure for Acrolein.
Flow-through scavenging experiments of acrolein in the
presence of iodixanol were carried out using a scavenger resin
packed in solid-phase extraction (SPE) cartridges. The resins
were manually packed as 75 mg amounts in 1 mL SPE
cartridges. The flow-through scavenging procedure for acrolein
and iodixanol is described below.
*
E-mail: ecevit.yilmaz@biotage.com. Telephone: +46 (0) 46
02 611. Fax: +46 (0) 46 163901.
1
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge the financial support of the European
Commission under FP7 - Marie Cure Actions, Contract
PITN-GA-2008-214226 [NEMOPUR] for the work presented
in this paper.
REFERENCES
■
(
1) Bolt, H. M.; Foth, H.; Hengstler, J. G.; Degen, G. H. Toxicol. Lett.
2
004, 151, 29.
(
2) Bouder, F. Expert Rev. Clin. Pharmacol. 2008, 1, 241.
Flow-through Scavenging Procedure.
(3) Muller, L.; Mauthe, R. J.; Riley, C. M.; Andino, M. M.; De
Antonis, D.; Beels, C.; De George, J.; De Knaep, A. G. M.; Ellison, D.;
Fagerland, J. A.; Frank, R.; Fritschel, B.; Galloway, S.; Harpur, E.;
Humfrey, C. D. N.; Jacks, A. S.; Jagota, N.; Mackinnon, J.; Mohan, G.;
Ness, D. K.; O’Donovan, M. R.; Smith, M. D.; Vudathala, G.; Yotti, L.
Regul. Toxicol. Pharmacol. 2006, 44, 198.
Conditioning
1
mL of MeOH - let equilibrate for 2 min
1
mL of EtOH - let equilibrate for 2 min
Loading
(
4) Kondo, K.; Watanabe, A.; Iwanaga, Y.; Abe, I.; Tanaka, H.;
Nagaoka, M. H.; Akiyama, H.; Maitani, T. Food Addit. Contam. 2006,
3, 1179.
5) Leblanc, B.; Charuel, C.; Ku, W.; Ogilvie, R. Int. J. Pharm. Med.
004, 18, 215.
6) Jacobson-Kram, D.; McGovern, T. Adv. Drug Delivery Rev. 2007,
9, 38.
1
mL of 5 μg/mL acrolein and 2 mg/mL iodixanol in
EtOH - let equilibrate for 5 min
2
(
2
(
5
(
(
Batch-Scavenging Procedure for Acrolein. Each scav-
enger (75 mg) was placed in HPLC vials; after the addition of 1
mL of a solution of 5 μg/mL acrolein and 2 mg/mL iodixanol
in EtOH, the mixtures were shaken for 30 min, derivatized by
DNPH, and assayed by HPLC-UV.
Capacity of Scavengers. Scavengers were put into HPLC
vials with varying amounts of resins (10, 25, 50, 75, 100, and
7) McGovern, T.; Jacobsen, K. D. Trends Anal. Chem. 2006, 25, 790.
8) Dobo, K. L.; Greene, N.; Cyr, M. O.; Caron, S.; Ku, W. W. Regul.
Toxicol. Pharmacol. 2006, 44, 282.
(
(10) Roy, A.; Nyarady, S.; Matchett, M. Chim. Oggi 2009, 27, 19.
(11) Pierson, D. A.; Olsen, B. A.; Robbins, D. K.; DeVries, K. M.;
Varie, D. L. Org. Process Res. Dev. 2009, 13, 285.
9) Bartos, D.; Gorog, S. Curr. Pharm. Anal. 2008, 4, 215.
̈
̈
2
00 mg) and 1 mL of a solution of 5 μg/mL acrolein and 2 mg/
mL iodixanol in EtOH was added to each vial, stirred for 30
min, derivatized by DNPH, and then assayed by HPLC-UV.
(
12) Guidelines on the Limits of Genotoxic Impurities, CPMP/SWP/
5
199/02; Committee for Medicinal Products for Human Use
CONCLUSIONS
■
(
CHMP), European Medicines Agency (EMEA): London, 2006
(13) Muller, L.; Singer, T. Toxicol. Lett. 2009, 190, 243.
(14) Robinson, D. I. Org. Process Res. Dev. 2010, 14, 946.
In the present study, we have evaluated a collection of
commercially available scavenging resins for the removal of the
1
228
dx.doi.org/10.1021/op3000459 | Org. Process Res. Dev. 2012, 16, 1225−1229

Organic Process Research & Development
Article
(
15) Kozekov, I. D.; Turesky, R. J.; Alas, G. R.; Harris, C. M.; Harris,
T. M.; Rizzo, C. J. Chem. Res. Toxicol. 2010, 23, 1701.
16) Toxicological Profile for Acrolein; U.S. Department of Health and
(
Human Services, Public Health Service Agency for Toxic Substances
and Disease Registry: Atlanta, GA, 2007; http://www.atsdr.cdc.gov/
toxprofiles/tp124.pdf.
(
17) , Impurities Testing Guideline: Impurities in New Drug Substances,
CPMP/ICH/2737/99; Committee for Medicinal Products for Human
Use (CHMP), European Medicines Agency (EMEA): London,
October, 2006; http://www.ema.europa.eu/docs/en_GB/document_
library/Scientific_guideline/2009/09/WC500002675.pdf.
(
18) LeVan, M. D., Ed. Fundamentals of Adsorption; Kluwer Academic
Publishers: Norwell, MA, 1996.
19) Rouqueol, F.; Rouquerol, J.; Sing, K. Adsorption by Powders
(
and Porous Solids: Principles. In Methodology and Applications;
Academic Press: New York, 1999.
(
20) Carpino, L. A.; Mansour, E. M.; Cheng, C. H.; Williams, J. R.;
MacDonald, R.; Knapczyk, J.; Carman, M. J. Org. Chem. 1983, 48, 661.
21) Carpino, L. A.; Mansour, E. M.; Knapczyk, J. J. Org. Chem. 1983,
8, 666.
22) Lee, C.; Helmy, R.; Strulson, C. Org. Process Res. Dev. 2010, 14
4), 1021.
23) Welch, C. J.; Biba, M.; Drahus, A.; Conlon, D. A.; Hsin Tung,
H.; Collins, P. J. Liq. Chromatogr. Rel. Technol. 2003, 26 (12), 1959.
24) Szekely, G.; Bandarra, J.; Heggie, W.; Sellergren, B.; Ferreira, F.
C. J. Membr. Sci. 2011, 381, 21.
(
4
(
(
(
(
́
(
(
25) www.biotage.com
26) Andreoli, R.; Manini, P.; Corradi, M.; Mutti, A.; Niessen, W. M.
Rapid Commun. Mass Spectrom. 2003, 17 (7), 637.
1
229
dx.doi.org/10.1021/op3000459 | Org. Process Res. Dev. 2012, 16, 1225−1229