Development of molecular adsorption processes for the ... · 1 Development of molecular adsorption processes for the removal of genotoxic impurities from active pharmaceutical ingredients

1

Development of molecular adsorption processes for the removal of genotoxic

impurities from active pharmaceutical ingredients

Mariana Duarte de Pina

Instituto Superior Técnico, Universidade de Lisboa

Avenida Rovisco Pais, 1049-001 Lisboa, Portugal

Abstract: Most of the drugs available in the market are synthesized using highly reactive

molecules. These molecules may be present in the final API as impurities, that may be genotoxic or

carcinogenic. The risk for patient’s health caused by these impurities has become an increasing

concern of pharmaceutical companies and regulatory authorities. A broad range of unrelated

chemicals from very different chemical families have been categorized as genotoxic. These

compounds have the ability to react with DNA, preventing its normal replication, resulting in an

associated carcinogenic risk. Although it is desirable to avoid the use of GTIs in the manufacture of

APIs, this is not always possible, since these compounds are synthetically useful. It is fundamental to

produce APIs with low GTI content, controlled below the Threshold of Toxicological Concern (TTC)

established by regulatory authorities (1,5 µg/day). So, it is necessary to find simple, robust and

economical routes to remove GTIs from APIs. During the development of this thesis, conventional

purification techniques (recrystallization, ionic exchange resins and adsorbents), as well as emergent

techniques (nanofiltration, molecularly imprinted polymers (MIPs)) were studied. The results achieved

suggest that recrystallization is not a cost-effective process. In that sense, it is necessary to find new

ways to increase its yield. Using ionic exchange resins and MIPs, it is possible to make

recrystallization a viable process for the pharmaceutical industry.

Keywords: Genotoxic impurity, purification, recrystallization, molecular imprinting

1. Introduction

Carcinogenesis includes three stages:

initiation, promotion and progression. Usually,

mutational events are involved in the initiation

stage; these events are usually corrected

almost immediately by DNA repairing

mechanisms. Yet, sometimes these

mechanisms fail to repair the DNA and the

mutated cells start to proliferate – promotion

state. Then, the cells undergo differentiation,

creating new genes. After differentiation, the

mutated cells transported by the bloodstream

invade healthy tissues; this process in known

as metastasis and occurs in the progression

stage1a,1b. Mutagenicity is the capacity to induce

transmissible genetic damage, including gene

mutations or chromosomal aberrations. The

term genotoxicity refers to all genetic damage,

including genetic alterations that may result in

mutations, which are not transmitted to

daughter cells2. Genotoxic compounds attack

the nucleophilic centers of the DNA, which can

lead to strand breaks. The nucleophilic centers

of DNA are the nitrogen and oxygen atoms of

pyrimidine and purine bases and the

phosphodiester backbone1a,2a,3. The

stereospecificity of the reaction depends on the

chemical nature of the genotoxic compound,

steric factors and nucleophilicity; the most

nucleophilic sites of the DNA bases are

endocyclic nitrogens; on the contrary, exocyclic

oxygens are the less nucleophilic4. Chemical

2

mutagens and carcinogens are metabolized by

a variety of enzymes; for instance, several

forms of human cytochrome P-450 are involved

in the oxidative metabolism of chemical

carcinogens.

1.1. Regulation

The International Conference on

Harmonisation of Technical Requirements for

Registration of Pharmaceutical for Human Use

(ICH) brings together the regulatory authorities

from Japan, Europe and United States; it

studies the scientific and technical aspects of

pharmaceutical product registration1a. ICH

guidelines address impurities in drug

substances (Q3A), degradants in drug product

(Q3B) and also, residual solvents in drug

substance (Q3C). However, these guidelines

fail to address a number of important issues,

such as the level of impurities in drugs during

development and control of GTIs5.

The European Medicines Agency (EMEA),

an agency for the evaluation of medicinal

products, published a limit guideline for

genotoxic impurities in new drug substances

that is only used to new applications for

manufacturing process changes5b,5d,6. This

guideline recommends the genotoxic impurities

division into those interacting directly with DNA

and those acting through other mechanisms.

The first group is of main concern, since there

is not enough evidence for a threshold-related

mechanism; they can damage DNA at any

concentration. In this case, the guideline

proposes the application of the ALARP principle

(As Low As Reasonably Practicable). This

principle is based on a balance between the

need to reduce the GTI concentration to the

lowest possible level and the possibility of

reducing it6. The guideline proposes the use of

a “threshold of toxicological concern” (TTC) for

genotoxic impurities; it refers to a threshold

exposure level of compounds that will not pose

a significant risk of carcinogenicity or other toxic

effects. The draft guideline proposes a TTC of

1,5 μg/day, which corresponds to a 10-5 lifetime

risk of cancer5b,5c; this risk is justified by the

anticipated health benefits for the patient in

taking the medicine5c,7.

The Pharmaceutical Research and

Manufacturing Association (PhRMA)

established a Genotoxic Impurity Task Force,

which developed a White Paper. It was

proposed that all identified or predicted

impurities should be classified into one of five

classes5a-c,7c,8.

1. Impurities known to be genotoxic

(mutagenic) and carcinogenic;

2. Impurities known to be genotoxic

(mutagenic) but with unknown

carcinogenic potential;

3. Impurities containing alerting

structures, unrelated to the structure of

the API, and of unknown genotoxic

(mutagenic) potential;

4. Impurities containing alerting

structures, which are related to the API;

5. Impurities with no alerting structures, or

where no sufficient evidence exists that

genotoxicity is absent.

The Center of Drug Evaluation and

Research (CDER) of the US FDA is developing

guidelines to address genotoxic impurities in

pharmaceutical products. Even though

genotoxic impurities should be avoided, it is

known that complete removal is not always

possible. When it is not possible to avoid

genotoxic impurities, they should be limited to a

level that does not represent a significant risk to

patients5b.

3

In 2013 the ICH M7 guideline was

published; it provides guidance on analysis of

structure activity relationships (SAR) for

genotoxicity.

1.2. Genotoxic compounds

A wide range of unrelated chemicals, with

very different structures and from very different

chemical families have been categorized as

genotoxic impurities. A genotoxic substance is

known as a genotoxin. Sources of genotoxic

impurities in the manufacture of APIs include

starting materials, reagents, intermediates, side

reactions and impurities. Functional groups may

also be responsible for starting materials and

intermediates’ genotoxicity. There is a small

group of APIs, such as chemotherapeutic

agents, for which genotoxic and carcinogenic

substances are acceptable; however on the

majority of cases, it is desirable to remove the

genotoxic impurities, which cannot always be

achieved.

From a chemical point of view, there are no

physical properties or chemical structural

elements that provide a definitive categorization

of genotoxic. There are some molecules whose

genotoxic effect is known, while others are

dangerous because they contain reactive

groups that may lead to genotoxicity; these

reactive groups are molecularly recognized and

are cataloged as “structural alerts”9. It was

estimated that about 20 to 25% of all

intermediates used in standard pharmaceutical

synthesis contain “structural alerts”10.

The model compounds selected for this

work were Mometasone furoate (Meta) as API,

4-dimethylaminopyridine (DMAP) and methyl p-

toluenesulfonate (MPTS) as GTIs. During the

synthesis of Meta, sulfonyl chlorides are used in

a DMAP base catalyzed sulfonylation reaction

(Figure 1).

Figure 1 - Model compounds.

1.3. GTI mitigation

The first strategy to ease GTIs in the

production of APIs is to avoid the use and

generation of GTIs by altering the synthetic

route. This can be achieved by using different

chemical synthesis to obtain the same API or

intermediate or by optimizing the existing

synthetic route. In many cases, reagents and

intermediates are reactive and synthetically

useful and cannot be avoided. In these cases, a

Quality by Design (QbD) approach can be

applied. This approach includes adjustment of

parameters such as pH, temperature, reaction

time and solvent matrix13.

1.4. API purification

The API synthesis includes some reaction

steps intercalated with purification steps; these

purification steps contribute to GTI removal,

even though they are not designed for that

purpose. For the specific removal of GTIs, the

selection of the purification method is

dependent on the chemicophysical properties of

the compound, such as reactivity, solubility,

volatility and ionisability of the GTI14. It is

necessary to guarantee that during the

purification steps, the API losses are not

significant; another scenario unacceptable at

industrial scale. It is important to select a

purification process highly selective to a specific

impurity, so the API losses are lower and the

removal efficiency of the impurity is higher.

4

Some of the conventional purification steps

include crystallization, precipitation, solvent

extraction, column chromatography, treatment

with activated carbon, resins as well as

distillation. The efficiency of the separation is

based on the differences in the properties of the

agents to be separated and/or their relative

affinities for a selective agent. During the last

decade some other techniques, such as

membrane separations or molecularly imprinted

polymers, have been developed15.

2. Materials and Methods

2.1. Materials

DMAP was purchased from Sigma-Aldrich,

MPTS was purchased from Acros Organic and

Meta was kindly provided by Hovione.

Amberlite resins were purchased from Sigma-

Aldrich, AG 50W-X2 resin was purchased from

BioRad and activated charcoal was purchased

from Merck.

2.2. Recrystallization

The recrystallization process was based on

the WO9800437 patent.

50 mL of a solution having 10000 ppm of

Meta and 1000 ppm of DMAP in DCM was

concentrated under reduced pressure to 5 mL.

10 mL of MeOH were added, the solution was

heated to 50ºC and the mixture was

concentrated to 5 mL. This procedure was

repeated twice and precipitation occurred. The

solution was cooled to 20ºC over 1 hour, cooled

further to 10ºC and agitated for 2 hours. The

Meta was filtered and washed 2 times with

MeOH cooled to 10ºC. 0,3 g of charcoal, 10 mL

of MeOH and 10 mL of DCM were added to the

wet cake. The API was dissolved at 50ºC

followed by filtration of the charcoal. The

filtration equipment was rinsed twice with 2 mL

of DCM. This solution was combined with the

API solution and concentrated under reduced

pressure to 5 mL. 10 mL of MeOH were added

and the solution was once again concentrated

under reduced pressure to 5 mL. The mixture

was cooled to 23ºC over 1 hour, cooled further

to 10ºC and agitated for 2 hours. The Meta was

filtered and washed 2 times with MeOH cooled

to 10ºC and dried in an oven at 70ºC for 24

hours.

This procedure was repeated twice with

cooling the solution to 4ºC in the crystallization

steps.

2.3. Resins screening

Ionic exchange resins (AG 50W-X2,

Amberlite CG400, IRA458, IRA68, IRC50, IRC

86, XAD16 and XAD7) and adsorption systems

(activated charcoal) may be used to purify the

washing solutions from the recrystallization

procedure. The influence of pH and

temperature in the adsorption process was

studied, as well as adsorption isotherms and

kinetics. Since the solvent in washing solutions

is MeOH and the resins are prepared to be

used in aqueous solutions, the following

procedures were made using aqueous solutions

of DMAP, then using a mixture of water and

MeOH (1:1) and finally MeOH.

20 mg of the scavenger resins were put in

contact with 4 mL of DMAP solutions with pH

values between 12 (DMAP completely

deprotonated) and 6 (DMAP completely

protonated); these solutions were stirred for 24

hours and then assayed by HPLC-UV. The

same was done to study the temperature

influence; in this case, the solutions were stirred

for 24 hours at 25ºC, 35ºC and 45ºC. To obtain

the adsorption isotherms, 20 mg of the selected

resins were completely dispersed in 4 mL of

DMAP solutions at concentration of 100, 250,

500, 750 and 1000 ppm. Then the solutions

5

were continuously agitated for 24 hours at a

desired temperature so that the adsorption

reached equilibrium. The resin was separated

from the solution and the residual content of

DMAP was determined by HPLC-UV. The

kinetic experiments were identical to the

isotherm experiments, while 1 mL of the sample

was taken at defined time intervals.

The same procedure was followed to study

MPTS adsorption onto resins. In this case only

two pH values were used (1,18 and 9,67) since

MPTS pKa value is -2,58.

2.4. MIPs synthesis

The polymers composition can be found in

Table 1. DMAP was used as template,

methacrylic acid (MAA) was used as monomer,

ethylene glycol dimethacrylate (EDGMA) was

used as cross linking agent and

azobisisobutyronitrile (AIBN) was used as

initiator. MAA was dissolved in DCM, which

works as porogen. The template was added to

the MAA solution and left for 5 minutes.

EGDMA and AIBN were added to the

polymerization solution, which was purged with

a flow of dry nitrogen. The polymerization tubes

were sealed and the polymerization occurred at

desired temperature. After the polymerization

was completed, the polymers were crushed

using a pestle and mortar. The template was

extracted in a Soxhlet-apparatus with a solution

of 0,1 M HCl in MeOH for 48 hours. The

remaining acid was washed out with MeOH

using a Soxhlet-apparatus for 24 hours. Then

the polymers were crushed again and sieved;

the fraction 38-63 µm was used to evaluate the

binding properties of the polymers. The

polymers were dried in the oven overnight at

40ºC. The same fraction was used for the

characterization of the scavengers. The non-

imprinted polymer was prepared in the same

way as described above, but without the

template molecule. To study the binding

properties of the different scavengers prepared,

25 mg and 50 mg of the polymer was added to

solutions of 100 and 1000 ppm DMAP in DCM.

These solutions were stirred for 24h at 60 rpm.

The scavengers were separated from the

solutions, and the DMAP concentration in

solution was determined by HPLC-UV.

Table 1 – Polymer composition and polymerization conditions.

Reaction Template

(mmol)

Monomer

(mmol)

Cross-linker

(mmol)

Initiator

(mmol) Polymerization conditions

MIP1 4 4 40 2% 65ºC for 24 hours

MIP2 4 4 40 2% 40ºC for 12 hours, then the

temperature was increased

(5ºC/30 minutes) to 65ºC for

an additional 3 hours

MIP3 1 4 10 1%

MIP4 1 4 20 1%

NIP4 0 4 20 1%

The isotherm adsorption was determined as

described above for resins. In this case, 75 mg

of the scavenger was added to 1,5 mL of DMAP

solutions at concentration of 20, 50, 150 and

250 ppm. To confirm that Meta was not

adsorbed by the polymers, 75 mg of MIP was

added to 1,5 mL of a solution of 1000 ppm

DMAP and 10000 ppm Meta in DCM. After the

solution was stirred at 60 rpm for 24 hours, it

was assayed by HPLC-UV.

6

3. Results and discussion

3.1. Recrystallization

GTI limits in APIs are calculated by the TTC

divided by the maximum daily dose (g/day)

giving the limit in ppm applied to the active

substance. Considering a 5 mg/day dose of

Meta the GTI is required to be controlled under

300 ppm (0,3 mgDMAP/gMeta). Having a post-

stream reaction containing 10.000 ppm of Meta

and 1000 ppm of DMAP (100 mgDMAP/gMeta), it is

necessary to remove more than 99,7% of the

GTI. The recrystallization shows a high API loss

without acceptable compensation in the API

purity achieved. This process is commonly used

to purify APIs in the pharmaceutical industry

due to the fact that it removes unwanted solvent

occlusions from the API and allows the control

of particle size. The largest fraction of API loss

was observed in the charcoal adsorption,

representing 53,32% (recrystallization at 10ºC)

and 78,86% (recrystallization at 4ºC) of the total

API loss over the three steps. When the

recrystallization procedure was performed at

10ºC, DMAP was almost all removed in the first

recrystallization. When these steps were

performed at 4ºC, DMAP removal could not be

assigned preferentially to any of the steps. This

may be due to the fact that at higher

temperatures, the crystallization occurs at a

slower rate. It was not possible to obtain a

solution with 0,3 mgDMAP/gMeta after

recrystallization. Meta was mostly lost during

charcoal adsorption and through mother liquors.

It is necessary to find a way to purify Meta lost

during this process.

Figure 2 – API lost and GTI removed during recrystallization procedures (in the left pictures the recrystallization

steps were made at 10ºC and in the right pictures at 4ºC)

0

100

200

300

400

500

600

AP

I (m

g)

Step

Recovered API

API loss in washing solutions

Another API losses

0

100

200

300

400

500

600

AP

I (m

g)

Step

Recovered API

API lost in washing solutions

Another API losses

0

10

20

30

40

50

GTI

(m

g)

Step

Another GTI losses

GTI removed in washing solutions

Recovered GTI

0

10

20

30

40

50

60

GTI

(m

g)

Step

Another GTI losses

GTI removed in washing solutions

Recovered GTI

7

3.2. MIPs

It was not possible to achieve a ratio of 0,3

mgDMAP/gMeta using recrystallization. Since most of

Meta is being lost during charcoal adsorption, this

adsorbent may be replaced with MIPs. MIPs are

highly specific for a given molecule.

The DMAP binding percentage was determined

for all the prepared MIPs.

Figure 3 – DMAP adsorbed by MIPs prepared.

The best results were obtained when using 50

mg of polymer in contact with a 100 ppm DMAP

solution.

The amount of template and cross-linker added

to the polymerization reaction affects the binding

properties. Increasing the amount of cross-linker,

the binding increases from 71% to 93%. This agent

fixes the functional groups of MAA around the

imprinted molecule and forms a highly cross-linked

rigid polymer. When the template is removed, the

polymer has some cavities complementary to the

target molecule. If the cross-linker content is low,

the cross-linking degree is smaller and,

consequently, the polymer cannot maintain a stable

cavity configuration. The template:monomer ratio

also affects the amount of GTI adsorbed. For a ratio

1:4, 93% of DMAP is removed whilst for a ratio 4:4

a lower amount of GTI is adsorbed (66%). This ratio

influences the number of binding sites available.

The polymerization conditions also influence the

binding properties. When the polymerization occurs

entirely at 65ºC the quantity of DMAP adsorbed is

lower (MIP1) than when a temperature gradient is

used. In the latter case, the polymeric chains were

formed conveniently, creating binding cavities for

molecular recognition. MIP4 revealed the best

performance for GTI removal. NIP4, which was

prepared like MIP4 but without the template,

removed 79% of DMAP. The Freundlich isotherm is

the best fit for the adsorption isotherm determined

for MIP4, which suggests a multilayer adsorption.

The charcoal adsorption stage is very critical, since

a great amount of API is lost during this procedure.

Since MIPs show good stability in DCM and are

effective in GTI removal, they can be used to

replace charcoal. Starting with 1,5 mL of a solution

of 100 ppm DMAP and 10000 ppm Meta and

adding 75 mg of MIP4, it is possible to remove 98%

while losing 9,65% of Meta.

Using data from the isotherm adsorption, it is

possible to conclude that MIPs allow to lower the

ratio from 74,24 mgDMAP/gMeta to 4,79 mgDMAP/gMeta.

3.3. Ionic exchange and adsorbent resins

During recrystallization at 10ºC, it is

possible to remove more than 97% of DMAP, but

there’s a great amount of Meta that is lost within

mother liquors from recrystallization. Ionic

exchange resins and adsorbents may be used to

remove DMAP from mother liquors.

3.3.1. DMAP in water

DMAP was successfully removed from the

solution using cationic exchange resins and

activated charcoal. The resins AG 50W-X2 (98%),

Amberlite IRC50 (93%), Amberlite IRC86 (98%)

and activated charcoal (91%) were capable of

removing DMAP from the solution. DMAP was more

efficiently removed when it was fully deprotonated

(higher pH values). The temperature did not

influence the adsorption process. After studying the

pH and temperature effect, the adsorption

isotherms were obtained. Freundlich isotherm was

0

100

MIP4 MIP3 MIP2 MIP1

% b

ind

ing

100 ppm 50 mg 100 ppm 25 mg1000 ppm 50 mg 1000 ppm 25 mg

8

the best model to describe the adsorption by

activated charcoal. This model is based on

multilayer adsorption on heterogeneous surfaces.

The ionic exchange resins follow the Sips model,

which is a combination of the Langmuir and

Freundlich isotherms. The kinetic studies were

performed with AG 50W-X2 resin. The adsorption

capacity of the resin increases rapidly with

increasing of time until the equilibrium. 5 minutes is

enough to reach equilibrium.

3.3.2. DMAP in water and MeOH (1:1)

When MeOH is added to the solution, the

binding capacity of the resins lowers, especially

with activated charcoal (60%). AG 50W-X2,

Amberlite IRC50 and Amberlite IRC86 were able to

remove 97%, 92% and 95% of DMAP, respectively.

Activated charcoal is able to adsorb MeOH, which

explains the decrease of the amount of DMAP

adsorbed by charcoal. It was also observed that the

amount of DMAP adsorbed when the pH of solution

is high (12,94) decreases abruptly. Possibly, ionic

species are formed and these species compete with

DMAP to bind to resins. The temperature influenced

the adsorption process, which may be due to the

fact that the temperature changes the equilibrium

constants. After determining the adsorption

isotherms, it was observed that the Sips model was

the best fit. The kinetic studies were performed with

AG 50W-X2. The adsorption capacity of the resin

increases rapidly with increasing of time until the

equilibrium. 1 minute is enough to reach

equilibrium.

3.3.3. DMAP in MeOH

When MeOH is the only solvent, the quantity of

DMAP adsorbed lowers. AG 50W-X2, Amberlite

IRC50 and Amberlite IRC86 removed 90%, 66%

and 64% of DMAP, respectively. Activated charcoal

only removed 17% of DMAP; as stated before, this

may be due to the fact that this adsorbent is able to

adsorb MeOH. Once again, it was observed that at

highest pH value (12,05), the DMAP adsorbed

decreases substantially, which suggests the

formation of ionic species competing with DMAP to

bind to the resins. It was also observed that

temperature does not affect the adsorption process.

The Freundlich isotherm is the best fit for the

adsorption isotherm obtained for AG 50W-X2 and

Amberlite IRC86, while Sips isotherm describes

more properly the adsorption of Amberlite IRC50.

Once again, the adsorption capacity of AG 50W-X2

increases rapidly over time, but in this case, 2 hours

were necessary to reach equilibrium.

Using AG 50W-X2, it is possible to purify the

mother liquor from recrystallization 1; a solution with

29,28 mgDMAP/gMeta was obtained. Since this ratio is

lower than 100, this solution could be fed again to

the recrystallization process.

3.4. OSN

This study was based on a theoretical model.

The data used was based on information available

from the membrane GMT-oNF-2, which shows a

good stability in DCM. This membrane retains Meta

effectively (99,1%), while DMAP can cross the

membrane easily (16,5%).

Figure 4 – GTI removal and API losses using OSN.

0,00

20,00

40,00

60,00

80,00

100,00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14% o

f G

TI r

emo

val a

nd

AP

I lo

sses

Dilution ratio

Remoção de GTI Perdas de API

20% API losses

95% GTI removal

9

OSN is very effective in the removal of DMAP.

GTI removals superior to 95% can be achieved at

the cost of only 2,69% Meta loss at diavolume 3. It

is possible to lower the ratio from 100 mgDMAP/gMeta

to 0,16 mgDMAP/gMeta at diavolume 6. However, the

higher the number of diavolumes, the higher the

API loss and solvent consumption and operation

time. It is possible to use a lower number of

diavolumes and then feed the mother liquor back to

the process and use MIPs to purify the API retained

in the membrane. For that purpose, it is possible to

dissolve the retained compounds in DCM and put

the solution in contact with MIP4 in order to remove

DMAP from the solution. Using this approach it is

possible to lower the ratio from 100 mgDMAP/gMeta to

0,33 mgDMAP/gMeta at diavolume 2 if an additional

step of MIPs adsorption is performed.

3.5. MPTS mitigation

3.5.1. MPTS in MeOH

The same studies as described above for

DMAP were performed with solutions of MPTS in

MeOH. Only Amberlite IRA68, whose functional

group is a tertiary amine, was able to remove MPTS

(96%). The amount of MPTS adsorbed increases

slightly with the increase of temperature, but it is not

significant. The adsorption isotherm may be

described by the Langmuir isotherm, which

describes the formation of monolayers. The

adsorption capacity of Amberlite IRA68 increases

slowly with time being necessary 24 hours to reach

equilibrium.

4. Conclusions

It was not possible to reduce the GTI in API

post reaction streams to levels below the

recommend TTC value using recrystallization. This

process shows a high API loss without

compensation in the API purity achieved, therefore

this process is not cost effective. Since this is the

purification process approved in the manufacture of

APIs, it is necessary to find alternatives to increase

the recrystallization yield. Ionic exchange resins

may be used to purify Meta lost in the mother

liquors, while MIPs are a good alternative to replace

charcoal in the adsorption step.

OSN is another purification process that can be

used instead of recrystallization. OSN requires 6

diavolumes to remove 99,85% of the GTI with

acceptable API losses (5,28%). Adding a step of

MIPs adsorption, only two diavolumes are required

to obtain a solution with 0,33 mgDMAP/gMeta.

MPTS can be removed from solution using ionic

exchange resins and adsorbents. Amberlite IRA68

is very efficient in the removal of MPTS.

5. Acknowledgements

FCT – Fundação para a Ciência e Tecnologia for

funding through the project PTDC/QEQ-

PRS/2757/2012, “Removal of Genotoxic Impurities

from Active Pharmaceutical Ingredients”, and

Hovione for supply of API used. To IST, FFUL and

FCT-UNL team members that participated in this

project.

6. References

1. (a) Székely, G. S., Miriam; Ferreira, Frederico C.; Gil, Marco; Heggie, William, Sources of genotoxic impurities in drug synthesis: a systematic review. p 115. To be submitted; (b) Woo, Y.-t. L., David Y., (Q)Sar Analysis Of Genotoxic And Nongenotoxic Carcinogens: A State-Of-The-Art Overview. In Cancer Risk Assessment: Chemical Carcinogenesis, Hazard Evaluation, and Risk Quantification, Wiley, Ed. 2010; p 517. 2. (a) Benigni, R.; Bossa, C., Chemical reviews 2011, 111 (4), 2507-36; (b) Zeiger, E., Environmental and Molecular Mutagenesis 2004, 44 (5), 363-371.

10

3. Pullman, B. P., Alberte, Nucleophilicity of DNA. Relation to Chemical Carcinogenesis. In Carcinogenesis: Fundamental Mechanisms and Environmental Effects, Springer Netherlands: 1980; pp 55-66. 4. In Carcinogenesis: Fundamental Mechanisms and Environmental Effects, Thirteenth Jerusalem Symposium on Quantum Chemistry and Biochemistry, Jerusalem, Israel, Pullman, B. T. o., Paul O. P.; Gelboin, Harry, Ed. Springer Netherlands: Jerusalem, Israel, 1980. 5. (a) Müller, L.; Mauthe, R. J.; Riley, C. M.; Andino, M. M.; Antonis, D. D.; Beels, C.; DeGeorge, 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., Regulatory Toxicology and Pharmacology 2006, 44 (3), 198-211; (b) Jacobson-Kram, D.; McGovern, T., Advanced drug delivery reviews 2007, 59 (1), 38-42; (c) Robinson, D. I., Organic Process Research & Development 2010, 14 (4), 946-959; (d) Jacobson-Kram, D., Genotoxic and Carcinogenic Impurities in Drug Substances and Products: Recommended Approaches. Services, U. S. D. o. H. a. H., Ed. 2008; (e) Snyder, R. D.; Smith, M. D., Drug Discovery Today 2005, 10 (16), 1119-1124. 6. Giordani, A.; Kobel, W.; Gally, H. U., European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences 2011, 43 (1-2), 1-15. 7. (a) (CHMP), C. f. H. M. P., The potential risks of carcinogens, mutagens and substances toxic to reproduction when these substances are used as excipients of medicinal products for human use. In EMEA/CHMP/SWP/146166/2007, Agency, E. M., Ed. 2008; (b) Medley, C. D.; Smith, J. E.; Wigman, L. S.; Chetwyn, N. P., Analytical and bioanalytical chemistry 2012, 404 (8), 2233-9; (c) Pierson, D. A.; Olsen, B. A.; Robbins, D. K.; DeVries, K. M.; Varie, D. L., Organic Process Research & Development 2008, 13 (2), 285-291. 8. Dobo, K. L.; Greene, N.; Cyr, M. O.; Caron, S.; Ku, W. W., Regulatory toxicology and pharmacology : RTP 2006, 44 (3), 282-93. 9. Sawatari, K.; Nakanishi, Y.; Matsushima, T., Industrial health 2001, 39 (4), 341-5. 10. Benigni, R. B., Cecilia, Current Computer-Aided Drug Design 2006, 2 (2), 169-176. 11. (a) Neumann, H. G., Journal of cancer research and clinical oncology 1986, 112 (2), 100-6; (b) Meier, C. 12. (a) Elder, D.; Facchine, K. L.; Levy, J. N.; Parsons, R.; Ridge, D.; Semo, L.; Teasdale, A., Organic Process Research & Development 2012, 16 (11), 1707-1710; (b) Teasdale, A.; Delaney, E. J.; Eyley, S. C.; Jacq, K.; Taylor-Worth, K.; Lipczynski, A.; Hoffmann, W.; Reif, V.; Elder, D. P.; Facchine, K. L.; Golec, S.; Schulte Oestrich, R.;

Sandra, P.; David, F., Organic Process Research & Development 2010, 14 (4), 999-1007. 13. (a) Raman, N. V.; Prasad, A. V.; Ratnakar Reddy, K., Journal of pharmaceutical and biomedical analysis 2011, 55 (4), 662-7; (b) Vogt, F. G.; Kord, A. S., Journal of Pharmaceutical Sciences 2011, 100 (3), 797-812; (c) Cimarosti, Z.; Bravo, F.; Stonestreet, P.; Tinazzi, F.; Vecchi, O.; Camurri, G., Organic Process Research & Development 2009, 14 (4), 993-998. 14. Teasdale, A.; Elder, D.; Chang, S.-J.; Wang, S.; Thompson, R.; Benz, N.; Sanchez Flores, I. H., Organic Process Research & Development 2013, 17 (2), 221-230. 15. (a) Rohani, S.; Horne, S.; Murthy, K., Organic Process Research & Development 2005, 9 (6), 858-872; (b) Shekunov, B. Y.; York, P., Journal of Crystal Growth 2000, 211 (1–4), 122-136; (c) Kim, S.; Wei, C.; Kiang, S., Organic Process Research & Development 2003, 7 (6), 997-1001; (d) Yang, Y.; Tjia, R., Computers & Chemical Engineering 2010, 34 (7), 1030-1035; (e) Lee, C.; Helmy, R.; Strulson, C.; Plewa, J.; Kolodziej, E.; Antonucci, V.; Mao, B.; Welch, C. J.; Ge, Z.; Al-Sayah, M. A., Organic Process Research & Development 2010, 14 (4), 1021-1026; (f) Iverlund, N.; Parsons, A.; Muller, F., Chemical Engineering Research and Design 2009, 87 (6), 852-858; (g) Kecili, R.; Billing, J.; Nivhede, D.; Sellergren, B.; Rees, A.; Yilmaz, E., Journal of chromatography. A 2014, 1339, 65-72; (h) Székely, G.; Bandarra, J.; Heggie, W.; Sellergren, B.; Ferreira, F. C., Journal of Membrane Science 2011, 381 (1-2), 21-33; (i) Piletsky, S. A.; Alcock, S.; Turner, A. P. F., Trends in Biotechnology 2001, 19 (1), 9-12.