A fluorescent probe for the quantification of heparin in clinical samples with minimal matrix interference

Article abstract of DOI:10.1039/b917287d

A red-fluorescent perylene diimide probe allows sensitive quantification of heparin, a widely used antithrombotic drug, in plasma and serum samples with minimal interference from matrix components.

Full text of DOI:10.1039/b917287d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
A fluorescent probe for the quantification of heparin in clinical samples
with minimal matrix interferencew
a
b
a
Helga Szelke,a Sarah Schu
¨
bel, Job Harenberg and Roland Kramer*
¨
Received (in Cambridge, UK) 21st August 2009, Accepted 30th November 2009
First published as an Advance Article on the web 11th January 2010
DOI: 10.1039/b917287d
A
red-fluorescent perylene diimide probe allows sensitive
samples. Such probes were pioneered by Anslyn7 and also
described by Chang8 and Zhang and Zhu.9 Heparin binding in
solution is accompanied by a decrease of UV7 or increase of
visible fluorescence.8,9 Problems related to matrix interference
have not yet been fully overcome by the currently available
probes. Compared with buffered aqueous solutions, higher
concentrations of heparin are required to trigger
a comparable fluorescence response in plasma or serum
containing samples.7,8 In two of the above-mentioned assays,
the volume of serum added to the detection solutions had to be
kept rather low (2 vol%7 or 0.2 vol%9), so that heparin levels
above the clinically relevant range might be required in the
original serum sample. Probe–heparin interaction can be
affected by competitive binding of either heparin or the probe
(or both). Matrix interference is generally of great concern in
serum and plasma diagnostics and drug monitoring10 since
it may lead to inaccurate and poorly reproducible data.
Calibration considers general matrix effects, but specific
interferences involve the risk that the response of the probe
depends on medium composition, considering that about
30 more abundant (i.e. c Z 1 mM) plasma proteins vary
2–30 fold in their individual concentration from person
to person, or depending on the physiological state.11 For
laboratory and point-of-care monitoring of heparin and for
pharmacokinetic studies,12 a probe that is less sensitive to
medium interference would be more desirable.
quantification of heparin, a widely used antithrombotic drug,
in plasma and serum samples with minimal interference from
matrix components.
Heparins are major clinical anticoagulants, with about half a
billion doses applied annually.1 Heparin is a linear, poly-
disperse polysaccharide that consists predominantly (470%)
of a trisulfated disaccharide repeating unit (Scheme 1) and
has a high negative charge density. Unfractionated heparin
(UFH, mean molecular weight between 13 000 and 15 000),
isolated from natural sources, has been clinically applied since
the 1930s. By many indications, there has been a trend2
towards the use of fractionated low molecular weight heparins
(LMWH, mean molecular weight between 3000 and 5000),
manufactured by partial depolymerisation of UFH. LMWHs
have lower side effects and a more favourable pharmaco-
kinetics. It is often critical to maintain heparin levels that
are sufficient to prevent thrombosis but on the other hand
avoid risks of bleeding. Monitoring of heparins is recom-
mended2 for a variety of patient subgroups, in anticoagulation
reversal during acute bleeding and routinely for therapeutic
UFH application. Despite their recognized limitations,2b,3
measurement of activated partial thrombine time APTT
(the ability to delay clotting) and anti-Xa assay (inhibition of
a specific blood coagulation factor), remain the currently
accepted practice for laboratory monitoring of UFH and
LMWH, respectively, in patients’ blood. There are no
established methods for the reliable and convenient point-of-
care monitoring of LMWH.4 These limitations have
stimulated interest in alternative detection schemes,5 with a
particular emphasis on methods that would facilitate the
detection of heparin by simple means in clinical samples.
Detection of heparin by molecular indicators goes back to
toluidine blue staining, and interest in simple, colorimetric or
fluorimetric assays has been greatly renewed in recent years.6
Very few small-molecule probes, however, are capable of the
sensitive fluorescence detection of heparin in clinically relevant
plasma (cell-free blood) or serum (plasma without clotting factors)
Encouraged by
a study on polyammonium-modified
calixarenes which bind heparin very strongly,13 we have
recently linked multiple ammonium groups to a fluorescent
ruthenium(II)-tris(2,20-bipyridine) complex in order to strengthen
electrostatic probe–target interaction.14 However, the problem
of altered emission response to heparin in serum samples
persisted. A number of studies, including those in our own
laboratory, on the targeting of polyanionic DNA by cationic
a Universitat Heidelberg, Anorganisch-Chemsiches Institut,
¨
Im Neuenheimer Feld 270, 69120 Heidelberg, Germany.
E-mail: Roland.Kraemer@urz.uni-heidelberg.de;
Fax: +49 62 21-54 85 99; Tel: +49 62 21-54 84 38
b Universitat Heidelberg, Medizinische Fakultat Mannheim, Institut
¨
¨
fur Experimentelle und Klinische Pharmakologie und Toxikologie,
¨
Maybachstr. 14, 68169 Mannheim, Germany.
E-mail: job.harenberg@medma.uni-heidelberg.de;
Fax: +49 621 383-9622; Tel: +49 62 383-9621
w Electronic supplementary information (ESI) available: Details of
synthesis and characterization of 1. See DOI: 10.1039/b917287d
Scheme 1 Structure of 1 (above, major 1,7-isomerz) and of the major
repeating disaccharide unit of heparin (below).
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 1667–1669 | 1667

View Article Online
perylene diimide or naphthalene diimide derivatives,15 often
with modulation of the chromophores optical properties,
provided a promising starting point for the design of
alternative probes for heparin sensing.
Here we describe the novel probe 1, a modified perylene-
3,4:9,10-tetracarboxylic acid diimide which minimises
medium-dependent interferences in heparin quantification
and is at the same time sensitive enough to cover the clinically
relevant concentration range.
1 was prepared in five reaction steps from commercially
available compounds and isolated in the form of its ammo-
nium salt (H81)Cl8 as a water soluble violet powder.w
Protonation of the amino groups at neutral pH is expected
since the pKa values of the protonated parent amines are
8.0–10.9 for spermine and 10.2 for 3-aminopropanol at I =
0.1 M.16 As suggested for other water soluble perylene
diimides,17 strong fluorescence of 1 (broad emission at
lmax = 615 nm; F = 0.23 on excitation at 485 nm in 1 mM
HCl, relative to rhodamine 6G in ethanol) is in accordance
with a reduced tendency, due to good solubilisation and
electrostatic repulsion, to form nonfluorescent aggregates,
held together by hydrophobic interactions between the
perylene cores. The O-substituents at the perylene skeleton
shift the fluorescence to longer wavelength compared to the
unsubstituted analog (lmax = 548 nm). A red fluorescence
signal improves the signal-to-background ratio in serum and
plasma samples, which on excitation at 485 nm have a
significant autofluorescence in the range 490–560 nm.
At longer wavelength, autofluorescence is reduced (to 25%,
11% and 3% at 615, 650 and 700 nm, respectively).
Fig. 1 Fluorescence intensity (arbitrary values) of solutions contain-
ing 1 (1 mM) and varying LMWH levels in different media: m Control
(buffered H2O), & 20 vol% human blood serum, ꢂ 20 vol% human
blood plasma. T = 20 1C, pH 7.0, buffer 60 mM 3-(N-morpholino)
propanesulfonic acid,
I E 0.2 M (200 mM sodium tosylate).
Excitation at 485 nm, emission 4665 nm. All data points are averages
of three measurements. The top LMWH scale indicates actual LMWH
concentrations (using
a molar mass 665.5 of the disaccharide
tetrasodium salt in Scheme 1) in assay solutions. *The bottom scale
relates to the LMWH activity IU/mL which a corresponding clinical
plasma or serum (or a control) sample would have prior to addition to
the assay solution. Actual LMWH activity levels in the assay solutions
are five times lower due to dilution. Inset: red fluorescent (left cuvette)
and heparin-quenched (right cuvette) solutions of 1.
Apparently, 1 associates with heparin in a highly selective
manner, with minimal interference by serum or plasma
components.
To study the response of 1 to heparin we focused on more
widely applied LMWH, using a commercial LMWH sodium
salt with average molecular weight 5000 (similar results
were obtained with UFH but are not included in this
communication). Fluorescence was recorded with a portable
fluorimeter, equipped with a 485 nm excitation filter and a
665 nm cut off emission filter. The latter filter offers a good
signal-to-background ratio since serum/plasma autofluores-
cence is very low at 4665 nm. Fluorescence of 1 in buffered
solution decreases approximately linearly (Fig. 1) with
increasing heparin concentration (given as the concentration
of the disaccharide in Scheme 1) with a calculated ‘‘loading’’ of
one dye molecule/4.6 sugar moieties. We interpret fluorescence
quenching as a consequence of the aggregation of 1 at the
polyanionic heparin template. Metal-coordination induced
aggregation was recently described as a signalling mechanism
of a Hg2+ selective perylene diimide probe.18 Response of 1 to
LMWH in samples containing 20 vol% serum or plasma
correlates well with the data of the control solution. Tosylate
anion is added to counteract a more significant decrease of
fluorescence in plasma/serum; presumably, the amphiphilic
anion associates with 1 and disrupts weak interactions with
plasma/serum components leading to fluorescence quenching.
Note that, although heparin-free solutions have a somewhat
lower fluorescence in the plasma/serum medium, response of
the probe to heparin is comparable in buffer, serum and
plasma. Equilibration is fast (within seconds) in all media.
Binding is reversible since fluorescence is restored on
addition of protamine, a strongly heparin-binding polypeptide.
In a clinical context, LMWH is quantified by its anti-Xa
activity in IU mLꢁ1 of a patients plasma sample. For the
LMWH sample used in this study, we determined the anti-Xa
activity 140 IU mgꢁ1. The bottom scale of Fig. 1 relates to the
heparin activity level in the plasma or serum sample that
would be added to the assay solution in a routine diagnostic
setting. At 1 mM 1, the probe covers both the generally
recommended19 therapeutic (0.5–1.0 IU mLꢁ1) and prophyl-
actic (0.2–0.4 IU mLꢁ1) range of LMWH in plasma. If
necessary, concentration of
1 can be adjusted for a
more precise monitoring at both higher or lower levels. For
example, when the concentration of 1 is reduced to 0.5 mM,
a 50% and 20% fluorescence decrease is triggered by
0.25 IU mLꢁ1 and 0.1 IU mLꢁ1, respectively, at a signal-to-
background ratio 44 (data not shown).
In a preliminary evaluation of inter-individual variations,
we analysed the response of 1 to heparin-spiked serum samples
of five randomly selected healthy people. For clarity, data in
Fig. 2 are limited to the fluorescence intensity readout for
selected heparin levels. The coefficient of variation CV =
(standard deviation/mean ꢂ 100) for the five patients is 8%
at 0.25 and 0.5 IU mLꢁ1. Note that in a validation protocol of
anti-Xa assay,19 the established clinical method for LMWH
quantification, the CV even for within-run precision
(same sample measured repeatedly) is, depending on instru-
mentation, in the range 5–25% for comparable heparin levels.
In conclusion, we designed a fluorescent perylene diimide
probe 1 for the detection of heparin with high sensitivity and
ꢀc
This journal is The Royal Society of Chemistry 2010
1668 | Chem. Commun., 2010, 46, 1667–1669

View Article Online
2 (a) J. Harenberg, Semin. Thromb. Hemostasis, 2008, 34, 779;
(b) E. Gray, B. Mulloy and T. W. Barrowcliffe, Thromb. Haemost.,
2008, 99, 807.
3 T. Baglin, T. W. Barrowcliffe, A. Cohen and M. Greaves,
Br. J. Haematol., 2006, 133, 19.
4 S. A. Spinler, A. K. Wittkowsky, E. A. Nutescu and M. A. Smythe,
Ann. Pharmacother., 2005, 39, 1275.
5 R. Jelinek and S. Kolusheva, Chem. Rev., 2004, 104, 5987.
6 (a) Zh. Zhong and E. V. Anslyn, J. Am. Chem. Soc., 2002, 124,
9014; (b) W. Sun, H. Bandmann and T. Schrader, Chem.–Eur. J.,
2007, 13, 7701; (c) J. C. Sauceda, R. M. Duke and M. Nitz,
ChemBioChem, 2007, 8, 391; (d) K.-Y. Pu and B. Liu, Macro-
molecules, 2008, 41, 6636; (e) T. Briza, Z. KejikI. Cisarova,
J. Kralova, P. Martasek and V. Kral, Chem. Commun., 2008,
1901; (f) R. B. C. Jagt, R. F. Gomez-Biagi and M. Nitz, Angew.
Chem., Int. Ed., 2009, 48, 1995.
7 A. T. Wright, Z. L. Zhong and E. V. Anslyn, Angew. Chem., Int.
Ed., 2005, 44, 5679.
8 S. Wang and Y.-T. Chang, Chem. Commun., 2008, 1173.
9 M. Wang, D. Zhang, G. Zhang and D. Zhu, Chem. Commun.,
2008, 4469.
10 (a) B. Maibohn, Pharmacokinetics and Pharmacodynamics of
Biotech Drugs, Wiley-VCH, 2006, p. 159; (b) S. Rochat, Z. Grote
and K. Severin, Org. Biomol. Chem., 2009, 7, 1147.
11 N. L. Anderson and N. G. Anderson, Mol. Cell. Proteomics, 2002,
1, 845.
12 F. Zhang, P. Sun, E. Munoz, L. Chi, S. Sakai, T. Toida, H. Zhang,
S. Mousa and R. J. Linhardt, Anal. Biochem., 2006, 353, 284.
13 T. Mecca, G. M. L. Consoli, C. Geraci, R. La Spina and
F. Cunsolo, Org. Biomol. Chem., 2006, 4, 3763.
Fig. 2 Fluorescence intensity (arbitrary values) of aqueous solutions
containing 1 (1 mM) after addition of 20 vol% LMWH-spiked (0, 0.25,
0.5, 1 and 2 U/mL) human blood serum of five different individuals.
Error bars indicate the inter-individual standard deviation. pH 7.0,
60 mM 3-(N-morpholino) propanesulfonic acid; T = 20 1C, I E 0.2 M
(200 mM sodium tosylate). Excitation at 485 nm, emission 4665 nm.
selectivity in clinical samples. While the alkoxy-substituents at
the perylene core shift fluorescence to wavelengths 4600 nm
to provide a high signal-to-background ratio, the multiple
ammonium groups attached to the N-imide sites favour
selective association with the polyanionic heparin target. 1 is
a rare example of a small-molecule fluorescent probe that
quantifies a molecular target with minimal interference from
highly competitive serum and plasma components. Ease of
application combined with matrix-independent response make
this heparin probe attractive for both routine laboratory and
fast point-of-care testing as well as high-throughput pharma-
cokinetic studies, ideally without the need for patient-specific
calibration.
14 H. Szelke, J. Harenberg and R. Kramer, Thromb. Haemost., 2009,
102, 859.
15 Selected examples: (a) O. Y. Fedoroff, M. Salazar, H. Han,
V. V. Chemeris, S. M. Kerwin and L. H. Hurley, Biochemistry,
1998, 37, 12367; (b) J. T. Kern, P. W. Thomas and S. M. Kerwin,
Biochemistry, 2002, 41, 11379; (c) L. Rossetti, M. Franceschin,
A. Bianco, G. Ortaggi and M. Savino, Bioorg. Med. Chem. Lett.,
2002, 12, 2527; (d) A. Mokhir and R. Kramer, Bioconjugate Chem.,
2003, 14, 877; (e) A. Mokhir, H. Wolf and R. Kramer, J. Am.
Chem. Soc., 2004, 126, 6208; (f) S. Krauß, M. Lysetska and
F. Wurthner, Lett. Org. Chem., 2005, 2, 349.
The authors thank Christina Giese for fluorescence assay
measurements of patients serum/plasma samples.
Notes and references
16 (a) D. Templeton and B. Sarkar, Can. J. Chem., 1985, 63, 3122;
(b) J. Bunting and D. Stefanidis, J. Am. Chem. Soc., 1990, 112, 779.
17 Ch. Kohl, T. Weil, J. Qu and K. Muellen, Chem.–Eur. J., 2004, 10,
5297.
18 Y. Che, X. Yang and L. Zang, Chem. Commun., 2008, 1413.
19 L. Houbouyan, B. Boutiere, G. Contant, M. D. Dautzenberg,
P. Fievet, G. Potrn, A. Vassault and Y. Gourmelin, Clin. Chem.,
1996, 42, 1223.
z 1 is isolated as a 2 : 1 mixture of two isomers with respect to the
position (1,7 or 1,6) of the –OR substituents. Only the major 1,7-
isomer is shown in Fig. 1. Small quantities of the isomers could be
isolated by HPLC and were shown to have both very similar fluorescence
emission profiles and response to LMWH in serum samples.
1 N. Mackman, Nature, 2008, 451, 914; S. J. Williams and
G. J. Davies, Trends Biotechnol., 2001, 19, 356.
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 1667–1669 | 1669