Journal of Pharmaceutical and Biomedical Analysis 120 (2016) 106–111

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journal homepage: www.elsevier.com/locate/jpba

A critical evaluation of Amicon Ultra centrifugal filters for separating

proteins, drugs and nanoparticles in biosamples
Elin Johnsen a , Ole Kristian Brandtzaeg a , Tore Vehus a,b , Hanne Roberg-Larsen a ,
Vanya Bogoeva c , Ornela Ademi a , Jon Hildahl d , Elsa Lundanes a , Steven Ray Wilson a,∗
a
Department of Chemistry, University of Oslo, Post Box 1033, Blindern, NO-0315 Oslo, Norway
b
Department of Engineering Sciences, University of Agder, Jon Lilletunsvei 9, 4891 Grimstad, Norway
c
Department of Molecular Biology of the Cell Cycle, Institute of Molecular Biology, Bulgarian Academy of Sciences, Acad. George Bonchev Str, bl. 21, 1113
Sofia, Bulgaria
d
Department of Biosciences, University of Oslo, Blindernveien 31, 0371 Oslo, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Amicon® Ultra centrifugal filters were critically evaluated for various sample preparations, namely (a)
Received 11 September 2015 proteome fractionation, (b) sample cleanup prior to liquid chromatography mass spectrometry (LC–MS)
Received in revised form 7 December 2015 measurement of small molecules in cell lysate, and (c) separating drug-loaded nanoparticles and released
Accepted 9 December 2015
drugs for accurate release profiling in biological samples. (a) Filters of supposedly differing molar mass
Available online 12 December 2015
(MM ) selectivity (10, 30, 50 and 100K) were combined to attempt fractionation of samples of various
complexity and concentration. However, the products had surprisingly similar MM retentate/filtrate
Keywords:
profiles, and the filters were unsuited for proteome fractionation. (b) Centrifugal filtration was the
Centrifugal filters
Proteins
only clean-up procedure in a FDA-guideline validated LC–MS method for determining anti-tuberculosis
Fractionation agents rifampicin and thioridazine in macrophage cell lysate. An additional organic solvent washing step
Rifampicin (drug/protein-binding disruption) was required for satisfactory recovery. (c) The centrifugation filters are
Thioridazine well suited for separating drugs and nanoparticles in simple aqueous solutions, but significantly less so
Nanoparticles for biological samples, as common drug–protein binding disruptors can dissolve NPs or be incompatible
with LC–MS instrumentation.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction gel pieces following electrophoresis. The various Amicon® Ultra

filter types – 10, 30, 50 and 100K – were expected to have filtra-
Centrifugal filters are used for e.g., desalting, protein enrichment tion properties related to molar mass (molecular weight) (as these
and deproteinization, and are familiar to many life scientists. How- and other related filters are often referred to as molecular weight
ever, they can have recovery/selectivity issues [1–4] and can differ “cut-off” units, e.g. [6,7,8]). Protein fractionation was attempted by
in performance after product re-design [2,5]. Also, performance can combining 10, 30, 50 and 100K Amicon® Ultra filters for whole
vary between brands [4]. In light of such factors, there are relatively blood, plasma, serum, and a commercial protein ladder standard
few papers that highlight the traits, limitations and optimizations, mixture.
of these ostensibly “no-brainer” products. Experiences with cur- (b) For small molecule determination, centrifugal filters have
rent Amicon® Ultra centrifugal filters are here collected regarding e.g., been used as a clean-up step following liquid–liquid extrac-
(a) protein separation, (b) small molecule determination in cell tion [9] or protein precipitation [2]. Surprisingly, few papers have
lysates and (c) studying release of drugs from nanoparticles (NPs) described the use of centrifugal filters as a stand–alone sample
in biological samples. preparation step prior to mass spectrometric analysis [10]. It was
(a) Centrifugal filters were tested for crude fractionation of blood hypothesized that Amicon® Ultra centrifugal filters could be used
proteins (e.g., in connection with MS-based proteomics), as a sim- as a single preparation procedure for accurate and precise LC-
ple alternative to size exclusion chromatography (SEC) or cutting ESI–MS/MS measurements of rifampicin (RIF) and thioriazine (TZ)
in limited amounts of cell lysate. Such a method is needed to
assess synergistic effects/uptake when combining an RNA poly-
∗ Corresponding author. merase inhibitor (RIF) and efflux pump inhibitor (TZ) in targeting
E-mail address: [email protected] (S.R. Wilson). drug-resistant tuberculosis [11].

http://dx.doi.org/10.1016/j.jpba.2015.12.010
0731-7085/© 2015 Elsevier B.V. All rights reserved.
 E. Johnsen et al. / Journal of Pharmaceutical and Biomedical Analysis 120 (2016) 106–111 107

(c) It was desired to assess whether the method in (b) could be 2.2. Sample clean-up prior to LC-ESI–MS/MS measurement of
modified for measuring drug release from poly(lactic-co-glycolic small molecules in cell lysate: rifampicin and thioridazine
acid) (PLGA) NPs [12,13] in biological samples. NP drug delivery
allows sustained release of therapeutics during week-scale parti- 2.2.1. Reagents
cle degradation [12]. The slow release of NPs in vivo is associated All reagents used were of analytical grade. See SM B for specifi-
with lowered toxicity. Specific cell types can be targeted [12], which cations and suppliers.
also has toxicological benefits [14]. Isolating released drug prior to
measurement avoids contamination from still-loaded NPs. Filter-
2.2.2. Stock solutions
trapping NPs could be a practical option to e.g., ultracentrifugation
1 mg/mL stock solutions of TZ, RIF, TZ internal standard (IS) or
instrumentation (as standard centrifugation will not “spin down”
RIF IS were made by dissolving appropriate amounts in 300 ␮M
NPs of e.g., nanometer- to low micrometer sizes).
ascorbic acid, 0.1% FA in ACN/HPLC water (40/60, v/v). All stock
solutions were stored at −20 ◦ C. These stock solutions were used
2. Materials and methods to prepare analyte- and internal standard solutions (see SM B).

For all experiments, Amicon® Ultra centrifugation filters (Merck

2.2.3. Sample clean up using centrifugation filters
Millipore, Billericia, MA, USA) of with 10K, 30K, 50K or 100K cut-off
Amicon® Ultra 10K centrifugation filters were pre-washed with
was used,
400 ␮L 600 ␮M ascorbic acid, 0.1% FA in HPLC water. Cell lysate
(100 ␮L, 100 000 cells, see SM-B for preparation) was transferred
2.1. Proteome fractionation to a 10K Amicon centrifugal filter and centrifuged at 14,000 × g for
10 min. To obtain an acceptable recovery of the analytes, 100 ␮L
2.1.1. Reagents of LC–MS mobile phase solvent was added to the filter before an
All reagents used were of analytical grade. See Supplementary additional centrifugation at 14,000 × g for 10 min was performed.
materials, Section A (SM A) for specifications and suppliers. The filtrate was collected and 0.5 ␮L was injected directly onto the
LC–MS system (described below).

2.1.2. Centrifugal filtration and 0.45 m filtration

Spin filtration was executed as recommended by the manufac- 2.2.4. LC–MS analysis
turer manual supplied with the product with the exception of the The analytes were determined using an Ultimate 3000 UHPLC
combined filtration. The chosen centrifugation time was 15 min at system coupled to a TSQ Vantage triple quadrupole-MS (Thermo
14,000 × g. The combinations were 30 and 10, 50 and 30, and 100 Scientific, Waltham, MA) equipped with a heated electrospray
and 50K filters. ionization (HESI) source operated in positive mode (prelimi-
The 100 and 50K filter combination will be explained as exam- nary experiments were performed using a 2695 Separations
ple. Aliquots of 500 ␮L whole blood, serum or plasma was applied Modules + Micromass ZQ single quadropole LC–MS from Waters
to the 100K filter and centrifuged at 14,000 × g for 15 min. Further, (Milford, MA)). Chromatographic separation was achieved using a
the filtrate was applied to a 50K filter and centrifuged at 14,000 × g Hypersil GOLD column from Thermo (50 × 1 mm ID, 1.9 ␮m parti-
for 15 min. The resulting supernatant was collected by reversing cles) with a flow rate of 0.2 mL/min. The MP was 300 ␮M ascorbic
the filter to a new vial and collected for further analysis. acid in 0.1% FA, 60% HPLC grade H2 O and 40% ACN (isocratic sepa-
ration). Ascorbic acid was added to prevent oxidation of rifampicin
during ionization. The column temperature was set to 40 ◦ C and all
2.1.3. Investigation of ionic, hydrophobic and protein–cellulose compounds eluted within 4 min. For more details and experimen-
interactions tal parameters together with complete validation procedures, see
For experimental procedure on filter pre-treatment, see SM A. SM B.

2.1.4. Gel electrophoresis 2.3. Separating nanoparticles, drug and proteins

Samples were boiled in 1X SDS sample loading buffer (1% (w/v)
SDS, 1 mM DTT, 20% (v/v) glycerol, 20 mM Tris–HCl pH 7.0, 0.01% 2.3.1. Reagents
(w/v) bromophenol blue) at 70 ◦ C for 15 min. Prepared samples All reagents used were of analytical grade. See SM C for specifi-
were loaded onto NuPAGE® Novex® 12% Bis–Tris protein gels and cations and suppliers.
separated for 1 h at 200 V in 1X MOPS running buffer. After separa-
tion, the gels were fixated overnight (o/n) in 10% (v/v) acetic acid,
25% (v/v) isopropanol, 65% (v/v) water. The gels were then sub- 2.3.2. Materials and sample preparation
sequently soaked in Coomassie Blue Staining solution (10% (v/v) PLGA NPs were loaded with RIF (25%) or TZ (27%) and synthe-
acetic acid, 0.005% (w/v) Coomassie Brilliant Blue and 90% (v/v) sized as described in Refs. [12,14]. Degradation of NP particles in
water) for 1–2 h. The gels were destained in 40% (v/v) MeOH, 10% 10–40% ACN/methanol was assessed by comparing NP release in
(v/v) acetic acid and 50% (v/v) water. Imaging was done with a cell lysate with that of lysate spiked with non-encapsuled drug
benchtop scanner. Silver staining was done with ProteoSilverTM (using the methodology for measuring RIF and TZ described above,
Silver Stain Kit according to manufacturers’ protocol. 0.01 mg particles per mL, 360◦ sample rotation during mixing, sam-
ples collected over the course of 24 h) and by inspection of particle
pellet size after centrifugation (observed in microcentrifugal tubes,
2.1.5. Characterization of filters using scanning electron not centrifugal filters). Degradation of RIF in acidic environment
microscopy was assessed using the Waters LC–MS system described above.
The centrifugal filters (10, 30, 50 and 100K) were characterized The effect of aqueous ZnSO4 heptahydrate (10/90, w/v) on ESI–MS
by scanning electron microscope (SEM). A section of the filter was signal/stability was assessed by comparing analyte signal-to-noise
placed on a carbon tape. SEM images were obtained in high vacuum ratios for solutions (cell lysate and pure water) processed with and
using a FEI Quanta 200 FEG-ESEM (FEI, Hillsboro, OR, USA). without the salt present during the centrifugal filtration procedure.
108 E. Johnsen et al. / Journal of Pharmaceutical and Biomedical Analysis 120 (2016) 106–111

Fig. 2. Gel electrophoresis of marker proteins. (a) GE with Coomassie Blue staining
of retentate from 8 marker proteins when applied to 10K (lane 1), 30K (lane 2), 50K
(lane 3) and 100K (lane 4) centrifugal filters. (b) GE with Coomassie Blue staining
of filtrate from standard ladder when applied to 100K centrifugal filters that have
been pre-treated with 0.1 M NaCl (lane 5), 10% ACN (lane 6) and 5% BSA in 1x PBS
(lane 7) before sample filtration. The lane to the left is in a–e the protein standard
ladder. (For interpretation of the references to color in this figure legend, the reader
is referred to the web version of this article.)

that attempted “heart-cutting” (e.g., 30K + 10K = 10K < protein frac-
tion <30K) resulted in low protein recovery (Fig. 1c, lane 9–11, Silver
stain), and an absence of fractioning selectivity.
Plasma was diluted 10, 100 and 1000-fold prior to spin filtration
to assess the effect of sample concentration on centrifugal filter
selectivity. Fig. 1d, lane 12–14 (10, 100, 1000 × dilution, respec-
tively) shows that the distribution of proteins (exemplified with
100K filter) was similar as for undiluted samples. Hence, applying
lower amounts of protein did not improve selectivity.
It was questioned whether the lack of selectivity was related to
sample complexity and therefore assessed the performance of the
Fig. 1. Gel electrophoresis of filtered plasma. (a) GE with Coomassie Blue staining of filters using a very simple commercial kit of 8 marker proteins,
retentate from human plasma when applied to 10K (lane 1), 30K (lane 2), 50K (lane 3)
commonly used for SDS-PAGE MM calibration. The marker pro-
and 100K (lane 4) centrifugal filters. (b) GE with silver staining of filtrate from human
plasma when applied to 10K (lane 5), 30K (lane 6), 50K (lane 7) and 100K (lane tein standard sample (denaturated form) was applied to 10, 30, 50
8) centrifugal filters. (c) GE with silver staining of heart-cuts from human plasma and 100K filters, respectively. Again, there was no clear difference
when applied to the combination of centrifugal filters 30 + 10K (lane 9), 50 + 30K between the filters in protein selectivity (Fig. 2a, lane 1–4).
(lane 10), 100 + 50K (lane 11). (d) GE with Coomassie Blue staining of filtrate from Also investigated was whether conditioning/saturating the fil-
human plasma when diluted 1:10 (lane 12), 1:100 (lane 13) and 1:1000 (lane 14)
and applied to 100K centrifugal filters. (For interpretation of the references to color
ters with salts, organic solvent and human serum albumin (HAS)
in this figure legend, the reader is referred to the web version of this article.) would have an impact on filter selectivity. This was not the case,
indicating an absence of hydrophobic or ionic interactions between
the Amicon® Ultra filters and proteins Fig. 2b, lane 5–7).
3. Results
3.2. Sample clean-up prior to LC-ESI–MS/MS measurement of
3.1. Proteome fractionation small molecules in cell lysate: rifampicin and thioridazine

Unprocessed blood (i.e., whole blood), serum and plasma were Amicon® Ultra centrifugation filtration of aqueous standard
subjected to filtration with 10, 30, 50 and 100K Amicon® Ultra solutions allowed rifampicin (RIF, 822.9 Da) and thioridazine (TZ,
centrifugal filters. SDS-PAGE analysis of samples collected fol- 370.6 Da) to elute with high recovery (90–95%). When applying
lowing centrifugation showed little difference/pattern in filter cell lysate, however, a subsequent “sample washing” step with
performance regarding proteins; most of the filters are able to 100 ␮L solvent containing 40% ACN was needed to obtain satisfac-
concentrate proteins in general, seemingly with no particular pref- tory recovery (58–65%). Centrifugation with 10K filters for 10 min
erence regarding MM (exemplified with plasma; Fig. 1a, lane 1–4). at 14 000 × g provided best repeatability. To remove glycerol from
This was in accordance with protein levels measured using UV- the filter and to avoid potential oxidation of rifampicin during fil-
spectroscopy (results not shown). However, there was a tendency tration, a pre-wash with 400 ␮L 600 ␮M ascorbic acid, 0.1% FA in
that the 100K filters allowed somewhat more proteins to elute HPLC water was included prior to sample application.
through the filter (Fig. 1b, lane 8) compared to 10, 30 and 50K For the subsequent LC–MS determination, isocratic mobile
(Fig. 1b, lane 5–7, Silver stain), but without an obvious “cut-off” phase conditions were used, so no column reconditioning was
function. Serum and plasma could be filtered, but whole blood needed between injections. RIF, TZ and internal standards were
(expectedly) plugged the filters, also after pre-filtration through chromatographed within 4 min. Using selective reaction monitor-
0.45 ␮m filters. Given the similar performance of the filters (i.e., ing (SRM), the limits of detection were 0.8 ng/mL (TZ) and 4 ng/mL
equally high enrichment for all protein sizes), it was unsurprising (RIF) in cell lysate (100 000 cells). The limits of quantification (LOQ,
 E. Johnsen et al. / Journal of Pharmaceutical and Biomedical Analysis 120 (2016) 106–111 109

Fig. 3. Validation data for measuring thioridazine and rifampicin in cell lysate. Within-day and between-day repeatability, linear range, recovery etc. Level L = 5 ng/mL TZ,
50 ng/mL RIF (LOQ), level M = 25 ng/mL TZ, 250 ng/mL RIF, level H = 75 ng/mL TZ, 750 ng/mL RIF. All solutions contained 50 ng/mL TZ IS and 500 ng/mL RIF IS. See Section 2.2.2
and 2.2.3 for more details regarding the solutions/samples employed.

Fig. 4. LC–MS of rifampicin and thioridazine. Representative chromatograms (spiked cell lysate, level M) and SRM transitions for measuring thioridazine and rifampicin and
internal standards. See Section 2.2.4 for more details on LC–MS conditions.
110 E. Johnsen et al. / Journal of Pharmaceutical and Biomedical Analysis 120 (2016) 106–111

defined as concentrations providing a S/N ≈ 10) were 5 ng/mL (RIF) after sample preparation was ≈100%). More extensive pre-washing
and 50 ng/mL (TZ). The mobile phase contained 300 ␮M ascor- can be more crucial prior to for NMR analysis [20], as glycerol
bic acid [14] to avoid oxidation of RIF in the electrospray source proton signals can overlap with the other compounds [20]. The
[15]. Ascorbic acid (600 ␮M) was also added during cell lysis, min- post-loading filter wash with organic solvent enabled satisfactory
imizing oxidation of the labile RIF during sample preparation. The recovery of small molecule analytes from biological samples. ACN
method was validated following FDA guidelines and had satisfac- (used as part of the washing solution) has previously been asso-
tory within- and between day precision (1–18% RSD). Figs. 3 and 4 ciated with less effective deproteinization with centrifugal filters
show the validation data and chromatograms, respectively. Carry- [9]. However, such symptoms were not observed regarding the
over was absent in the concentration ranges investigated. More described method (e.g., substantial retention time shifts due to
details on SRM conditions and additional data are presented in protein adsorbance on stationary phase) with the small washing
Supplementary materials, SM B. volume applied (100 ␮L on to the 0.5 mL capacity filter). Precision
was satisfactory (<20% RSD = within FDA guideline requirements),
and would likely be even better if labeled internal standards were
3.3. Separating nanoparticles, drug and proteins with
used (available from e.g., Alsachim or Santa Cruz Biotechnology)
centrifugation filters is problematic
instead of the chemical analogs (rifapentine and trifluoperazine)
used here. However, these analogs were chromatographically sep-
The authors [14] and others e.g. [16,17] have applied centrifuga-
arable and hence also suited for simpler, non-MS based systems
tion filters for separating PLGA NPs and drugs for release profiling,
(e.g., RIF is measured with fluorescence spectroscopy detection in
studied in simple aqueous solutions. In the present study it was
other applications of the authors).
investigated whether an approach as that described in (b) could
Contemporary methodologies can measure TZ in e.g., low ng/mL
be expanded/modified to separate drugs from proteins and NPs in
levels in biofluids e.g., blood and plasma [21], which is comparable
biological samples. Unsurprisingly, the organic solvent wash step
to that obtained in this work. However, to the authors’ knowl-
(40% ACN) in (b) dissolved the filter-trapped NPs so the validated
edge, virtually no methods have been reported for measuring TZ
method for RIF and TZ determination was not applicable for NP
in limited cell lysates. RIF is also measured in biofluids such as
release studies. However, also neutral solutions with less amounts
plasma and dried blood spots (e.g., Refs. [22–24]), but only limited
of ACN and methanol (e.g., 10%) caused a partial NP dissolvation
numbers of recent papers describe LC–MS methods for measuring
as well. Aqueous ZnSO4 , which is an alternative to organic solvents
RIF in cell lysates (earlier papers typically describe methods with
for protein precipitation [18], was also assessed for drug–protein
significantly lower performance due to low detection limits and
binding disruption but was associated with a >10-fold increase MS
low resolving chromatography). For instance, Oswald et al. mea-
noise and a very poor repeatability. Washing with TCA [18] was not
sured RIF and other related compounds in broncho-alveolar cells
considered due to the instability of RIF in strong acids (SM B).
[22]. This method and that described in this paper are comparable
regarding precision and analysis times; however, the Oswald et al.
4. Discussion method features a (arguably) more laborious liquid–liquid extrac-
tion procedure, intended for larger samples (≈20 times more cells
Surprisingly, the molecular size selectivity of the centrifugation per sample).
filters investigated was minimal for proteins, as all variants had Although the centrifugal filters were found to be well suited
very similar performance. It has previously been shown that some for separating NPs and drugs in simple solutions [14], they are
centrifugal filter variants can have issues with protein selectivity less applicable for separating drugs, e.g., RIF and TZ, from NPs and
[3,4]. However, the membrane chemistry and geometry of filters drug binding proteins. A key challenge is that efficient drug/protein
similar to the Amicon® Ultra filters have previously been described binding disruptors [18] such as organic solvents (ACN, methanol)
as suited for e.g., isolating small proteins (>30 kDa), based on eval- dissolves PLGA particles, highly acidic once (e.g., TCA) can degrade
uations with another brand [4]; in other words, centrifugal filters RIF and salts (ZnSO4 ) can generate noise in the MS. Additional
from various vendors (even with similar membrane chemistry and steps may reduce mass spectra noise when employing salts for
geometry) are by no means “one and the same product”. binding disruption, e.g., desalting with solid phase extraction, but
To examine why the centrifugal filters allowed proteins above such a string of preparation steps is arguably more extensive
their presumed MM cut-off values to pass the filters, the filters were than ultracentrifugation-based approaches. Of course, other small
characterized using SEM (the filters were observed dry, a require- molecules than those studied here may be more compatible with
ment for SEM). Dry centrifugal filter membranes contain pores NP/protein separation with centrifugation filters. NPs may also be
of increasing sizes, depending on their assigned cut-off/selectivity carriers of proteins [17]. According to the findings from part (a), one
value (10–100K) (see SM A). However, large pores in the ␮m-range should be cautious to attempt separating proteins and NPs using the
were observed for all the filters; for comparison, HSA [66 kDa] has Amicon filters; Even a 100K filter retains a considerable amount of
a Stokes–Einstein radius of 3.5 nm [19]. Such large pores would lower mass proteins, and this can dramatically affect the recovery
allow even extremely large molecules to pass, if the pore sizes of a protein release assay. Nonetheless, a 3K Amicon filter has been
had remained intact when wetted. It was therefore speculated that employed to remove PLGA particles from a 20 kDa protein [17].
when solvent is applied, the filter pores shrink to nm-scale but However, the recovery of this approach was not reported.
somewhat independent of the initial pore sizes observed in SEM. In summary, Amicon centrifugation filters were found to
Described herein is a validated method (following FDA guide- be: suited for isolating small molecules from proteins prior to
lines) for determining rifampicin and thioriazine in cell lysate using LC-ESI–MS/MS; unsuited for separating proteins into fractions;
centrifugal filters and LC–MS/MS, in connection with studying syn- unsuited for separating small molecules from NPs and proteins
ergetic properties of the two drugs for addressing drug resistance in when using common drug/protein-binding disruptors.
tuberculosis [11]. Centrifugal filtration was the only sample prepa-
ration procedure prior to LC–MS determination of the two basic
drugs in cell lysate. Protein precipitation prior to filtration [2] was Acknowledgments
not required for the cell lysate matrix. By including a single pre-
washing step (400 ␮L), traces of glycerol filter preservative did This work was funded by the Norwegian Research Council
not interfere with the LC–MS analysis (recovery of analyte spiked (FRIPRO 143686) (JH) and an EEA mobility grant (VB).
 E. Johnsen et al. / Journal of Pharmaceutical and Biomedical Analysis 120 (2016) 106–111 111

Appendix A. Supplementary data [12] R. Kalluru, F. Fenaroli, D. Westmoreland, L. Ulanova, A. Maleki, N. Roos, M.P.
Madsen, G. Koster, W. Egge-Jacobsen, S. Wilson,
Poly(lactide-co-glycolide)-rifampicin nanoparticles efficiently clear
Supplementary data associated with this article can be found, in Mycobacterium bovis BCG infection in macrophages and remain
the online version, at http://dx.doi.org/10.1016/j.jpba.2015.12.010. membrane-bound in phago-lysosomes, J. Cell Sci. 126 (2013) 3043–3054.
[13] X. Xu, C.R. Sabanayagam, D.A. Harrington, M.C. Farach-Carson, X. Jia, A
hydrogel-based tumor model for the evaluation of nanoparticle-based cancer
References therapeutics, Biomaterials 35 (2014) 3319–3330.
[14] C.B. Vibe, F. Fenaroli, D. Pires, S.R. Wilson, V. Bogoeva, R. Kalluru, M. Speth, E.
[1] C.A. Daykin, P.J. Foxall, S.C. Connor, J.C. Lindon, J.K. Nicholson, The comparison Anes, G. Griffiths, J. Hildahl, Thioridazine in PLGA nanoparticles reduces
of plasma deproteinization methods for the detection of toxicity and improves rifampicin therapy against mycobacterial infection in
low-molecular-weight metabolites by 1H nuclear magnetic resonance zebrafish, Nanotox (2015) 1–9.
spectroscopy, Anal. Biochem. 304 (2002) 220–230. [15] K. Morand, G. Talbo, M. Mann, Oxidation of peptides during electrospray
[2] L. Pont, F. Benavente, J. Barbosa, V. Sanz-Nebot, An update for human blood ionization, Rapid Commun. Mass Spectrom. 7 (1993) 738–743.
plasma pretreatment for optimized recovery of low-molecular-mass peptides [16] S. Wohlfart, A.S. Khalansky, S. Gelperina, O. Maksimenko, C. Bernreuther, M.
prior to CE–MS and SPE–CE–MS, J. Sep. Sci. 36 (2013) 3896–3902. Glatzel, J. Kreuter, Efficient chemotherapy of rat glioblastoma using
[3] H.M. Georgiou, G.E. Rice, M.S. Baker, Proteomic analysis of human plasma: doxorubicin-loaded PLGA nanoparticles with different stabilizers, PLoS One 6
failure of centrifugal ultrafiltration to remove albumin and other high (2011) e19121.
molecular weight proteins, Proteomics 1 (2001) 1503–1506. [17] N. Jimenez, J. Galan, A. Vallet, M. Egea, M. Garcia, Methyl trypsin loaded poly
[4] D.W. Greening, R.J. Simpson, Low-molecular weight plasma proteome (D,l-lactide-coglycolide) nanoparticles for contact lens care, J. Pharm. Sci. 99
analysis using centrifugal ultrafiltration, in: Serum/Plasma Proteomics, (2010) 1414–1426.
Springer, 2011, pp. 109–124. [18] C. Polson, P. Sarkar, B. Incledon, V. Raguvaran, R. Grant, Optimization of
[5] D. Pellerin, H. Gagnon, J. Dubé, F. Corbin, Amicon-adapted enhanced FASP: an protein precipitation based upon effectiveness of protein removal and
in-solution digestion-based alternative sample preparation method to FASP, ionization effect in liquid chromatography–tandem mass spectrometry, J.
F1000 Research 4 (2015). Chromatogr. B 785 (2003) 263–275.
[6] A. Beloqui, M.Á. Solinís, A.R. Gascón, A. del Pozo-Rodríguez, A. des Rieux, V. [19] F. Yuan, M. Dellian, D. Fukumura, M. Leunig, D.A. Berk, V.P. Torchilin, R.K. Jain,
Préat, Mechanism of transport of saquinavir-loaded nanostructured lipid Vascular permeability in a human tumor xenograft: molecular size
carriers across the intestinal barrier, J. Controll. Release 166 (2013) 115–123. dependence and cutoff size, Cancer Res. 55 (1995) 3752–3756.
[7] Z. Deng, S. Yan, H. Hu, Z. Duan, L. Yin, S. Liao, Y. Sun, D. Yin, G. Li, Proteomic [20] S. Tiziani, A.-H. Emwas, A. Lodi, C. Ludwig, C.M. Bunce, M.R. Viant, U.L.
profile of carbonylated proteins in rat liver: discovering possible mechanisms Günther, Optimized metabolite extraction from blood serum for 1H nuclear
for tetracycline-induced steatosis, Proteomics 15 (2015) 148–159. magnetic resonance spectroscopy, Anal. Biochem. 377 (2008) 16–23.
[8] J.J. Dalluge, J.L. McCurtain, A.J. Gilbertsen, K.A. Kalstabakken, B.J. Williams, [21] D. Montenarh, M. Hopf, H.H. Maurer, P. Schmidt, A.H. Ewald, Development
Determination of agmatine using isotope dilution UPLC–tandem mass and validation of a multi-analyte LC–MS/MS approach for quantification of
spectrometry: application to the characterization of the arginine neuroleptics in whole blood, plasma, and serum, Drug Test Anal. (2015).
decarboxylase pathway in Pseudomonas aeruginosa, Anal. Bioanal. Chem. [22] S. Oswald, J. Peters, M. Venner, W. Siegmund, LC–MS/MS method for the
(2015) 1–7. simultaneous determination of clarithromycin rifampicin and their main
[9] T. Reyns, M. Cherlet, S. De Baere, P. De Backer, S. Croubels, Rapid method for metabolites in horse plasma epithelial lining fluid and broncho-alveolar cells,
the quantification of amoxicillin and its major metabolites in pig tissues by J. Pharm. Biomed. Anal. 55 (2011) 194–201.
liquid chromatography-tandem mass spectrometry with emphasis on [23] D. Vu, R. Koster, A. Wessels, B. Greijdanus, J. Alffenaar, D. Uges,
stability issues, J. Chromatogr. B 861 (2008) 108–116. Troubleshooting carry-over of LC–MS/MS method for rifampicin,
[10] M. Sergi, C. Montesano, S. Napoletano, D. Pizzoni, C. Manetti, F. Colistro, R. clarithromycin and metabolites in human plasma, J. Chromatogr. B 917
Curini, D. Compagnone, Analysis of bile acids profile in human serum by (2013) 1–4.
ultrafiltration clean-up and LC–MS/MS, Chromatographia 75 (2012) 479–489. [24] D. Vu, R. Koster, M. Bolhuis, B. Greijdanus, R. Altena, D. Nguyen, J. Brouwers, D.
[11] L. Amaral, M. Viveiros, Why thioridazine in combination with antibiotics Uges, J. Alffenaar, Simultaneous determination of rifampicin, clarithromycin
cures extensively drug-resistant Mycobacterium tuberculosis infections, Int. J. and their metabolites in dried blood spots using LC–MS/MS, Talanta 121
Antimicrob. Agents 39 (2012) 376–380. (2014) 9–17.