Anal. Chem.

2008, 80, 2717-2727

Glucose and Lactate Biosensors for Scanning

Electrochemical Microscopy Imaging of Single Live
Cells
Madalina Ciobanu,† Dale E. Taylor Jr.,‡ Jeremy P. Wilburn,† and David E. Cliffel*,†

Department of Chemistry, Vanderbilt University, Nashville, Tennessee 37235, and Department of Chemistry and Life
Science, United States Military Academy, West Point, New York 10996

We have developed glucose and lactate ultramicroelec- Of particular interest are the enzyme-based sensors because
trode (UME) biosensors based on glucose oxidase and they offer the selectivity toward a single analyte optimized by
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

lactate oxidase (with enzymes immobilized onto Pt UMEs natural evolution. Enzymes achieve molecular recognition of
by either electropolymerization or casting) for scanning substrates (i.e., analytes of interest) based on structural comple-
electrochemical microscopy (SECM) and have determined mentarity, leaving little room for error.5 Enzymes catalyze with
Downloaded via UNIV DEL VALLE on September 23, 2024 at 23:00:23 (UTC).

their sensitivity to glucose and lactate, respectively. The high specificity chemical reactions in biological systems according
results of our evaluations reveal different advantages for to eq 1:
sensors constructed by each method: improved sensitiv-
enzyme
ity and shorter manufacturing time for hand-casting, and substrate + cofactor 98 products (1)
increased reproducibility for electropolymerization. We
have acquired amperometric approach curves (ACs) for
each type of manufactured biosensor UME, and these ACs Enzymes have been successfully employed in the miniaturiza-
can be used as a means of positioning the UME above a tion of sensor designs, including one of the most important
substrate at a known distance. We have used the glucose sensors for the health industry that emerged decades ago: the
biosensor UMEs to record profiles of glucose uptake glucose sensor which is based on glucose oxidase (GOx), whether
above individual fibroblasts. Likewise, we have employed it is intended for in vivo monitoring of glucose levels6 or for a
the lactate biosensor UMEs for recording the lactate diabetic’s regular glucose meter for use at home. Besides GOx,
production above single cancer cells with the SECM. We there have been numerous other enzymes used for the develop-
also show that oxygen respiration profiles for single cancer ment of miniaturized amperometric sensors, such as sarcosine
cells do not mimic cell topography, but are rather more oxidase,7 galactose oxidase,7 hexokinase,8 choline oxidase,7 lactate
convoluted, with a higher respiration activity observed at oxidase (LOx),9 alcohol oxidase,7 horseradish peroxidase,10 cho-
the points where the cell touches the Petri dish. These lesterol oxidase,11,12 etc.
UME biosensors, along with the application of others In addition to their importance for diabetes patients and in vivo
already described in the literature, could prove to be studies, enzyme-based amperometric sensors are irreplaceable
powerful tools for mapping metabolic analytes, such as tools for the noninvasive study of the metabolism at the cellular
glucose, lactate, and oxygen, in single cancer cells. level. A simplified view of the metabolic processes at the single-
cell level is presented in Figure 1, indicating that information about
The past decade has witnessed a growth in the application of the way a cell performs glycolysis could be acquired by using
biosensors to micrometer and submicrometer level investigations
(4) Marzouk, S. A. M.; Buck, R. P.; Dunlap, L. A.; Johnson, T. A.; Cascio, W. E.
in a wide variety of disciplines. Rapid development, both in
Anal. Biochem. 2002, 308 (1), 52-60.
miniaturization techniques and in understanding of biological (5) Garrett, R. H.; Grisham, C. M. Biochemistry, 2nd ed.; Saunders College
processes, has accelerated the expansion of biosensors in clinical Publishing: Fort Worth, TX, 1999.
(6) Chen, T.; Schmidtke, D. W.; Heller, A. Clin. Chem. Lab. Med. 2002, 40
applications and in areas such as biology, neurobiology, pharma-
(8), 786-789.
cology, and tissue engineering.1,2 In this respect, researchers have (7) Strike, D. J.; de Rooij, N. F.; Koudelka-Hep, M. Biosens. Bioelectron. 1995,
been able to acquire real-time quantitative measurements through 10 (1/2), 61-66.
(8) Kueng, A.; Kranz, C.; Mizaikoff, B. Biosens. Bioelectron. 2005, 21 (2), 346-
the application of microsensors on living cells, both in vitro and
353.
in vivo.2-4 (9) Ito, N.; Matsumoto, T.; Fujiwara, H.; Matsumoto, Y.; Kayashima, S.; Arai,
T.; Kikuchi, M.; Karube, I. Anal. Chim. Acta 1995, 312 (3), 323-328.
* To whom correspondence should be addressed. E-mail: d.cliffel@ (10) Horrocks, B. R.; Schmidtke, D.; Heller, A.; Bard, A. J. Anal. Chem. 1993,
vanderbilt.edu. 65 (24), 3605-3614.
† Vanderbilt University.
(11) Devadoss, A.; Burgess, J. D. J. Am. Chem. Soc. 2004, 126 (33), 10214-
‡ United States Military Academy, West Point.
10215.
(1) Suzuki, H. Electroanalysis 2000, 12 (9), 703-715. (12) Jiang, D.; Devadoss, A.; Palencsar, M. S.; Fang, D.; White, N. M.; Kelley, T.
(2) Lindner, E.; Buck, R. P. Anal. Chem. 2000, 72 (9), 336A-345A. J.; Smith, J. D.; Burgess, J. D. J. Am. Chem. Soc. 2007, 129 (37), 11352-
(3) Yao, T.; Yano, T.; Nanjyo, Y.; Nishino, H. Anal. Sci. 2003, 19 (1), 61-65. 11353.

10.1021/ac7021184 CCC: $40.75 © 2008 American Chemical Society Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2717
Published on Web 03/18/2008
Figure 1. Cancer cell metabolic pathways suitable for electrochemical probing by SECM. Cancer cells have an altered metabolism compared
to normal cells, with glycolysis being one of the affected pathways. TCA ) tricarboxylic acid cycle.

enzyme-based sensors for glucose and lactate. We have already The glucose and lactate amperometric sensors are based on a
employed GOx- and LOx-based sensors for probing cellular Pt electrode that is in contact with a polymer containing the
metabolism in a multianalyte microphysiometer (MAMP).13-16 Our specific enzyme (Figure 2); glucose and lactate are measured
laboratory has modified the Cytosensor microphysiometer (orig- indirectly at the electrode surface by amperometric oxidation of
inally developed by Molecular Devices for the measurement of hydrogen peroxide. The hydrogen peroxide is formed in aqueous
acidification rates) into the MAMP to allow for simultaneously environments during the reaction catalyzed by the enzyme
recording four analytes: glucose and oxygen uptake, lactate entrapped in the matrix (e.g., GOx in Figure 2A or LOx in Figure
production, and extracellular acidification rates for large numbers 2B) involving the analyte of interest (i.e., glucose for Figure 2A
(>105) of cells. The MAMP proves a useful tool in studies such or lactate for Figure 2B) in aerobic conditions, as described by
as the response of various cell types to chemical agents and toxins, eqs 2 and 3:
with ramifications in biodefense studies,15 and cancer cellular
GOx
metabolism relating genetic mutations to different metabolic D-glucose + oxygen 98 D-gluconolactone +
phenotypes, with implications in mathematical modeling of hydrogen peroxide (2)
cancer.17,18 The MAMP yields metabolic rates (e.g., lactate
LOx
production rate) for a large number of cells but has no ability to L-lactate + oxygen 98 pyruvate +
assess cell to cell variability. This paper is not a study of the hydrogen peroxide (3)
variability of metabolic rates from one single cell to another, but
rather it presents the technical means that could be employed
for addressing this problem by opening up the range of analytes Horrocks et al.10 have used an enzyme (horseradish peroxidase)
for single-cell electrochemistry. ultramicroelectrode (UME) in conjunction with the scanning
electrochemical microscopy (SECM) for the measurement of H2O2
(13) Eklund, S. E.; Cliffel, D. E.; Kozlov, E.; Prokop, A.; Wikswo, J.; Baudenbacher, at surfaces. They do not present any amperometric approach
F. Anal. Chim. Acta 2003, 496 (1-2), 93-101.
(14) Eklund, S. E.; Kozlov, E.; Taylor, D. E.; Baudenbacher, F.; Cliffel, D. E. In curves (ACs), but rather conductance-distance curves. The
Methods in Molecular Biology; Rosenthal, S. J., Wright, D. W., Eds.; Humana conductance measurements have the advantage of being insensi-
Press, Inc.: Totowa, NJ, 2005; Vol. 303, pp 209-223. tive to the enzyme-polymer layer present on the tip; however,
(15) Eklund, S. E.; Snider, R. M.; Wikswo, J.; Baudenbacher, F.; Prokop, A.; Cliffel,
D. E. J. Electroanal. Chem. 2006, 587 (2), 333-339. they have to be performed in solutions of very low conductivity
(16) Eklund, S. E.; Taylor, D.; Kozlov, E.; Prokop, A.; Cliffel, D. E. Anal. Chem. (1 mM KCl) which makes them unsuitable for physiological
2004, 76 (3), 519-527. applications, such as approaching and imaging in biological buffers
(17) Anderson, A. R. A. Math. Med. Biol. 2005, 22 (2), 163-186.
(18) Anderson, A. R. A.; Weaver, A. M.; Cummings, P. T.; Quaranta, V. Cell 2006, (e.g., RPMI). Kueng et al.8 have also reported the use of an
127 (5), 905-915. enzyme UME for SECM; they have assembled a Pt dual microdisk
2718 Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
Figure 2. Enzyme-based UME biosensor. An enzyme layer (GOx (A) or LOx (B)) is deposited onto a 25 µm Pt UME by either
electropolymerization or casting. The tip is rastered across the cell adhered to a Petri dish: (A) in the case of GOx it senses extracellular
glucose depletion due to cellular uptake; (B) in the case of LOx it senses lactate production by the cell. The enzyme film breaks down the
analyte, and the resulting H2O2 is oxidized at the underlying Pt UME. The arrows entering the cell symbolize glucose uptake (A), while the
arrows exiting the single cell denote lactate production (B).

electrode with one 5 µm Pt for electrode positioning via ACs and activities across different types of cells, examining the differences
the other 5 µm Pt modified through electropolymerization with between nonmetastatic and metastatic human breast cells. They
both GOx and hexokinase for indirect imaging of ATP transport have identified series of redox mediators that can cross the cell
through a porous polycarbonate membrane. However, they have membrane, investigated their distribution between cytosol and
not used the enzyme sensor of this dual UME for either nucleus, and found them unable to penetrate the nuclear envelope.
approaching the surface or for imaging of live organisms. Acquir- Matsue and co-workers36-38 studied single-cell topography and
ing ACs using the electrochemical process that takes place at an paid particular attention to respiration studies by means of SECM
enzyme-modified tip would remove the need in this case for for cells such as protoplasts and PC12. Liebetrau et al.31 have used
manufacturing the more complicated dual UME. Kueng et al.19 SECM to monitor the real-time morphological changes in PC12
have also used an enzyme (GOx)-coated tip integrated in an cells that are used as model neurons. Kurulugama et al.30 have
AFM-SECM setup for imaging glucose transport through a used a constant distance SECM for a study of PC12 cells where
porous membrane; in this case they bypassed the need for ACs they showed that constant impedance imaging yields superior
by using the AFM tip. topographic images of single cells compared to more common
SECM20-23 can be developed as a useful technique for the study constant height SECM. Hengstenberg et al.28 recorded topo-
of cell metabolic fluxes (as an important component of metabo- graphic images of PC12 cells by SECM, and studied catecholamine
lomics), since it can map electrochemical activity across the entire release from PC12 cells, using SECM as a positioning tool for an
surface of a single cell and can record dynamic changes. Although amperometric sensor in the vicinity of the cells. Isik and Schuh-
SECM has been used over the years for investigating biological mann29 have detected the release of nitric oxide from single
systems,23,24 only recently has attention been given to SECM for
T-HUVEC cells by also using a constant distance SECM. Bard
single-cell studies.25-38 Mirkin and co-workers26,27,32-34 have inves-
and co-workers have used SECM to study the metabolism of
tigated single cells by SECM in order to determine various redox
menadione to thiodione in single hepatoblastomas35 and have
(19) Kueng, A.; Kranz, C.; Lugstein, A.; Bertagnolli, E.; Mizaikoff, B. Angew.
Chem., Int. Ed. 2005, 44 (22), 3419-3422. (28) Hengstenberg, A.; Blochl, A.; Dietzel, I. D.; Schuhmann, W. Angew. Chem.,
(20) Bard, A. J.; Fan, F. R. F.; Kwak, J.; Lev, O. Anal. Chem. 1989, 61 (2), 132- Int. Ed. 2001, 40 (5), 905-908.
138. (29) Isik, S.; Schuhmann, W. Angew. Chem., Int. Ed. 2006, 45 (44), 7451-7454.
(21) Bard, A. J.; Fan, F. R. F.; Mirkin, M. V. In Electroanalytical Chemistry; (30) Kurulugama, R. T.; Wipf, D. O.; Takacs, S. A.; Pongmayteegul, S.; Garris,
Bard, A. J., Ed.; Marcel Dekker, Inc.: New York, 1994; Vol. 18, pp 243- P. A.; Baur, J. E. Anal. Chem. 2005, 77 (4), 1111-1117.
373. (31) Liebetrau, J. M.; Miller, H. M.; Baur, J. E.; Takacs, S. A.; Anupunpisit, V.;
(22) Bard, A. J.; Fan, F. R. F.; Pierce, D. T.; Unwin, P. R.; Wipf, D. O.; Zhou, F. Garris, P. A.; Wipf, D. O. Anal. Chem. 2003, 75, 563-571.
Science 1991, 254 (5028), 68-74. (32) Liu, B.; Cheng, W.; Rotenberg, S. A.; Mirkin, M. V. J. Electroanal. Chem.
(23) Bard, A. J.; Mirkin, M. V. Scanning Electrochemical Microscopy; Marcel 2001, 500, 590-597.
Dekker, Inc.: New York, 2001. (33) Liu, B.; Rotenberg, S. A.; Mirkin, M. V. Proc. Natl. Acad. Sci. U.S.A. 2000,
(24) Shiku, H.; Ohya, H.; Matsue, T. In Encyclopedia of Electrochemistry; Bard, 97 (18), 9855-9860.
A. J., Stratman, M., Wilson, G. S., Eds.; Wiley-VCH: Weinheim, Germany, (34) Liu, B.; Rotenberg, S. A.; Mirkin, M. V. Anal. Chem. 2002, 74, 6340-6348.
2002; Vol. 9, pp 257-275. (35) Mauzeroll, J.; Bard, A. J.; Owhadian, O.; Monks, T. J. Proc. Natl. Acad. Sci.
(25) Bard, A. J.; Li, X.; Zhan, W. Biosens. Bioelectron. 2006, 22 (4), 461-472. U.S.A. 2004, 101 (51), 17582-17587.
(26) Cai, C.; Liu, B.; Mirkin, M. V.; Frank, H. A.; Rusling, J. F. Anal. Chem. 2002, (36) Nishizawa, M.; Takoh, K.; Matsue, T. Langmuir 2002, 18 (9), 3645-3649.
74, 114-119. (37) Takii, Y.; Takoh, K.; Nishizawa, M.; Matsue, T. Electrochim. Acta 2003,
(27) Feng, W.; Rotenberg, S. A.; Mirkin, M. V. Anal. Chem. 2003, 75, 4148- 48, 3381-3385.
4154. (38) Yasukawa, T.; Kaya, T.; Matsue, T. Electroanalysis 2000, 12 (9), 653-659.

Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2719

probed the activity of HeLa cells using the oxidation of a ferrocene of glycolytic analytes (e.g., pH, lactate, oxygen, glucose) an area
derivative.25 of great interest. The intracellular pH for both cancerous and
Our choice of using the constant height SECM versus the normal cells is approximately 7.4,45 and the extracellular pH for
constant distance SECM for the study of single-cell metabolism cancerous growths is acidic, with values as low as 6.7.46 This
was predicated on using a chemically modified tip (i.e., Pt UME increased acidity in tumors is a result of the Warburg effect,44
coated with an enzyme-containing polymer); the modified tips can i.e., enhanced glycolysis, which under hypoxic conditions leads
be integrated easier with the typical constant height SECM mode to an increased lactate production that in turn lowers the
of operation. The major ways of achieving the constant distance extracellular pH (Figure 1). Walenta et al.47 reported mean lactate
mode in SECM are all relatively limited for the distance signal concentrations in head and neck tumors of ∼12 µmol/g, as
possibilities, due to their intrinsic design. One possibility for the compared to similar normal tissue lactate concentrations of ∼5
constant distance mode is the use of the impedance SECM,39 µmol/g. Acidity profiles32 were previously studied by SECM for
which is based on acquiring data based on high-frequency single living cells; however, the recording of these pH profiles
capacitance measurements. The capacitance of the chemically presents certain technical difficulties, such as the use of an
modified tips would be dramatically affected by the presence of additional potentiostat in conjunction with the SECM, as the pH
the polymer layer entrapping the enzymes on the Pt UME, sensor (Sb-based in Liu et al.’s paper)32 is by definition potentio-
compared to bare metal tips. Another way of achieving the metric. Using the lactate production to characterize cancer cells
constant distance mode for SECM measurements is by using presents its own advantages, such as the use of an amperometric
shear force.40 Hengstenberg et al. have shown that enzyme-filled sensor for lactate sensing, the ability to record lactate profiles in
capillaries can be used as tips for scanning using a shear force highly buffered solutions (as most cell culture media are) as
based constant distance SECM mode;41 however, these tips are opposed to a pH potentiometric sensor that can be used solely in
not amperometric and are suited to generation-collection modes, a modified low buffered medium. Lactate production profiles could
not imaging. The application of a polymer film to the tip for shear also provide information that could lead to a deconvolution of the
force based SECM applications appears technically challenging pH profiles of cancerous cells.
because the mass loading of the enzyme film will reduce the Cell respiration37 was also previously studied by SECM for
resonant Q of the oscillating tip in addition to the hydrodynamic single living cells; however, to our knowledge, there have been
loss in Q from a large volume of solution above the cells. Another no reported glucose or lactate spatio-temporal profiles for single
mode of operation for the SECM would be the AFM-SECM.42 live cells. We present in this paper the manufacturing and
Although Kueng et al. reported a GOx-coated AFM tip,19 the use characterization of glucose and lactate UME biosensors for SECM
of AFM-SECM for scanning live cells has yet to be developed of single live cells, and demonstrate that the GOx-based sensors
into a practicable method, and the difficulty of making each tip can be used for the positioning of the UME in the vicinity of the
by FIB (focused ion beam) micromachining is considerable. The cell, as well as for the imaging of the cell, and also show that the
integration of polymer film-modified tips for use in conjunction LOx-based sensors can be used for recording of lactate release
with a constant distance SECM for the study of live cell profiles from single cells.
metabolism will be the subject of future studies. With the typical
UME sizes of 5-25 µm used in our work, the higher resolution EXPERIMENTAL SECTION
possible by constant distance control would not be significant, Materials.Thematerialswerepurchasedasfollows: MgCl2‚6H2O,
justifying our use of constant height imaging. NaH2PO4‚H2O, HCl, CaCl2, MgSO4, NaCl, KCl, KH2PO4, NaHCO3,
A single cell is the basic structural and functional unit of all L-lactic acid, sodium acetate trihydrate, 35 mm Petri dishes, and

organisms. Although the MAMP provides information about the 50 mM phosphate buffer pH 7.0 from Fisher Scientific; KPF6 from
metabolism of cells, it averages signals over large populations Aldrich Chemical; glutaric dialdehyde as 25% solution in water
(>105 cells) and cannot distinguish differences between individual (GDA) and 2-aminophenol (OAP) from Acros; Na2HPO4‚7H2O,
cells. One cannot develop a deep understanding of metabolomics bovine serum albumin (BSA), RPMI 1640, D-glucose, GOx, type
unless the processes are explored in the context of the individual II-S from Aspergillus niger, and LOx from Pediococcus species from
cell level, wherein multiple biochemical processes, distributed Sigma; H2SO4 and NaOH from EM Science; ethanol (absolute
throughout the cell and various cell compartments, are integrated proof) from Aaper; ferrocenylmethyltrimethylammonium iodide
into a functional system. Differences in the average rate of from Strem Chemicals; N2 from Gibbs Welding Supply; conductive
glycolysis between large numbers of cancerous and healthy cells silver epoxy from Epoxy Technology; white sealant Hysol Epoxi-
have been demonstrated,43,44 making spatio-temporal mapping Patch from Dexter Corporation; 35 mm glass bottom Petri dishes
from Bioscience Tools. All materials were used as received unless
(39) Wipf, D. O.; Bard, A. J. Anal. Chem. 1992, 64 (13), 1362-1367. otherwise specified.
(40) Ludwig, M.; Kranz, C.; Schuhmann, W.; Gaub, H. E. Rev. Sci. Instrum. 1995,
66 (4), 2857-2860. (45) Griffiths, J. R.; McIntyre, D. J. O.; Howe, F. A.; Stubbs, M. In The Tumour
(41) Hengstenberg, A.; Kranz, C.; Schuhmann, W. Chem.sEur. J. 2000, 6 (9), Microenvironment: Causes and Consequences of Hypoxia and Acidity; Goode,
1547-1554. J. A., Chadwick, D. J., Eds.; Novartis Foundation Symposium, Vol. 240; John
(42) Macpherson, J. V.; Unwin, P. R. Anal. Chem. 2001, 73 (3), 550-557. Wiley & Sons, Ltd.: Chichester, 2001; pp 46-62, discussion pp 62-67.
(43) Semenza, G. L.; Artemov, D.; Bedi, A.; Bhujwalla, Z.; Chiles, K.; Feldser, (46) Bhujwalla, Z. M.; Artemov, D.; Aboagye, E.; Ackerstaff, E.; Gillies, R. J.;
D.; Laughner, E.; Ravi, R.; Simons, J.; Taghavi, P.; Zhong, H. In The Tumour Natarajan, K.; Solaiyappan, M. In The Tumour Microenvironment: Causes
Microenvironment: Causes and Consequences of Hypoxia and Acidity; Goode, and Consequences of Hypoxia and Acidity; Goode, J. A., Chadwick, D. J., Eds.;
J. A., Chadwick, D. J., Eds.; Novartis Foundation Symposium, Vol. 240; John Novartis Foundation Symposium, Vol. 240; John Wiley & Sons, Ltd.:
Wiley & Sons, Ltd.: Chichester, 2001; pp 251-260, discussion pp 260- Chichester, 2001; pp 23-38, discussion pp 38-45.
264. (47) Walenta, S.; Salameh, A.; Lyng, H.; Evensen, J. F.; Mitze, M.; Rofstad, E.
(44) Warburg, O. The Metabolism of Tumors; Constable: London, 1930. K.; Mueller-Klieser, W. Am. J. Pathology 1997, 150 (2), 409-415.

2720 Analytical Chemistry, Vol. 80, No. 8, April 15, 2008

 ASTM type I (18 MΩ) analytical grade deionized water (DW) after the use of the biosensor UMEs in conjunction with the
was obtained with a Solution 2000 water purification system from SECM.
Solution Consultants. All solutions were filtered prior to use with SECM Approach Curves of Biosensor UMEs. ACs were
0.2 µm syringe filters from Fisher. Ferrocenylmethyltrimethylam- acquired for GOx and LOx UMEs in a three-electrode system:
monium hexafluorophosphate (FcTMA) was prepared according WEsbiosensor 25 µm UME; REsAg/AgCl, 3 M KCl; CTRsPt
to the method of Forouzan et al.48 mesh. The tip was approached to the substrate (35 mm Petri dish)
All reported potentials were measured against a commercial and allowed to touch it. Glucose solutions of 1.0-5.0 mM and
Ag/AgCl, 3 M KCl reference electrode (RE, model CHI111, CH lactate solutions of 0.5 mM were used in the approach experi-
Instruments). ments.
SECM Testing Prior to Converting UMEs into Biosensors. Cells. Fibroblasts (mouse fibroblast CRL-10225) were cultured
All SECM measurements were conducted with a CHI900 instru- in high glucose (4.5 g/L) DMEM (Dulbecco’s modified Eagle’s
ment from CH Instruments equipped with an adjustable stage for medium) with 4 mM L-glutamine, supplemented with 10% fetal
tilt correction. The electrochemical cell was in a typical four- bovine serum and 1.0 mM sodium pyruvate. MCF10CA1a cells50
electrode configuration for testing the UMEs: UME tip electrode were cultured in DMEM/F12 medium, supplemented with cholera
(working electrode, WE), Pt wire counter electrode (CTR), Ag/ toxin (0.1 µg/mL), insulin (10 µg/mL), hydrocortisone (0.5 µg/
AgCl, 3 M KCl (RE), and substrate electrode (SE), which was a mL), epidermal growth factor (20 ng/mL), and horse serum (5%).
2 mm Pt disk (CHI303). All UMEs were manufactured with 5 or PC12 cells (CRL-1721) were cultured in Ham’s F12K medium with
25 µm Pt wire from Goodfellow according to Bard et al.,21 polished 2 mM L-glutamine adjusted to contain 1.5 g/L sodium bicarbonate
with 0.05 µm alumina from Buehler, and sonicated in ethanol and (82.5%), horse serum (15%), and fetal bovine serum (2.5%); PC12
water. UMEs were electrochemically cleaned for 5 min in 0.5 M cells were plated on dishes treated with human collagen (type
H2SO4 using a CHI660a potentiostat and were tested with 1 mM IV).
FcTMA in 100 mM KCl for performance. The Pt SE was polished SECM Imaging. The SECM was used in conjunction with an
and cleaned in an identical manner to the UMEs. Only UMEs inverted microscope (VistaVision, VWR) equipped with a stage
capable of achieving IT ) 8 (e.g., 800% increase in current over and objective warmer (TCII, 2Ch micro temperature controller,
steady-state conditions far from the SE)23 during approach to the Bioscience Tools). The inverted microscope allowed for visually
Pt SE were used in biosensor construction. The UMEs were also monitoring the morphology of the cells to be scanned. The SECM
characterized by cyclic voltammetry (CV) and consistently yielded tip holder was extended above the microscope objective with the
the expected sigmoidal CV. The UMEs used had RG ratios21 of aid of a light but sturdy aluminum piece. For positioning, the
=5 for the 5 µm Pt tips (7 µm actual measured diameter) and =2 UMEs were approached to the dish and then withdrawn a known
for the 25 µm Pt tips. distance. The data were recorded in a three-electrode system:
Enzyme-Based UMEs. (a) Manufacturing. The enzyme- WEsUME; REsAg/AgCl, 3 M KCl; CTRsPt mesh, in the
containing polymer layer was deposited onto 25 µm Pt UMEs via constant height mode.
two distinct methods: (i) Electropolymerizationsthe enzyme was (a) Glucose Profiles. The 25 µm UME GOx biosensors were
deposited on the electrode in a similar manner as described by positioned near the cell to be imaged, and then the biosensor tip
Zhang et al.49 The enzyme (10 mg/mL for GOx and 2 mg/mL for was rastered across the surface above the cell while the tip was
LOx) and the OAP (0.8 mg/mL) were dissolved in 50 mM acetate maintained at E ) +0.6 V versus Ag/AgCl, 3 M KCl. The buffer
buffer, pH 5.9, the solution was degassed (N2), and the elec- used was Hank’s balanced salt solution (HBSS), pH 7.4, 5 mM
trodeposition occurred under stirring at E ) +0.65 V versus Ag/ glucose, and no phenol red.
AgCl, 3 M KCl for 1 h (CHI660a workstation equipped with a (b) Lactate Profiles. The 25 µm UME LOx biosensors were
Faraday cage); the electrodes were allowed to dry for another positioned near the cell to be imaged, and then the biosensor tip
hour and were then briefly rinsed with DW and stored at 4 °C in was rastered across the surface above the cell while the tip was
50 mM phosphate buffer until use. (ii) Castingsthe enzyme (8 maintained at E ) +0.6 V versus Ag/AgCl, 3 M KCl. The buffer
mg/mL for GOx and 10 mg/mL for LOx), GDA (14 µL/mL used was RPMI, pH 7.4; there was no lactate added/present in
buffer), and BSA (62.5 mg/mL) were all dissolved in 50 mM the solution. There are possible interfering analytes secreted by
phosphate buffer, pH 7.0;16 the enzyme solution was then hand- the cells (e.g., peroxide, dopamine, etc.) that would be active at
cast onto the surface of the UME by briefly touching the tip of the potential set for the UME biosensor. However, the types of
the UME to a tiny droplet of enzyme solution; the UMEs were sensors we are using are enzyme-based, and the polymer films
air-dried and stored at room temperature until further use. are highly selective, as the polymer helps to block interferences.
(b) Calibrations. The performance of the biosensor UMEs We have extensively tested the use of the sensors in the presence
was tested using a CHI660a potentiostat with a three-electrode of live cells in the MAMP and found no analyte detection problems
configuration: WEsenzyme UME; REsAg/AgCl, 3 M KCl; due to cross interferences at this potential.14-16
CTRsPt mesh. Current versus time curves were recorded under (c) Oxygen Profiles. The buffer used was HBSS, pH 7.4, with
moderate stirring at E ) +0.6 V versus Ag/AgCl, 3 M KCl (for 5 mM glucose and 1 mM FcTMA. The Pt 7 µm UME was used
both GOx and LOx). Standard additions were made from stock for recording topographical images first, with the tip potential E
solutions of 500 mM glucose and 100 mM lactate. To ensure the ) +0.6 V versus Ag/AgCl, 3 M KCl, ensuring the oxidation of
validity of the results, the calibrations were performed before and the FcTMA. We chose FcTMA as mediator for topography
(50) Santner, S. J.; Dawson, P. J.; Tait, L.; Soule, H. D.; Eliason, J.; Mohamed, A.
(48) Forouzan, F.; Bard, A. J.; Mirkin, M. V. Isr. J. Chem. 1997, 37, 155-163. N.; Wolman, S. R.; Heppner, G. H.; Miller, F. R. Breast Cancer Res. Treat.
(49) Zhang, Z.; Liu, H.; Deng, J. Anal. Chem. 1996, 68 (9), 1632-1638. 2001, 65 (2), 101-110.

Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2721

imaging of single cells based on previous work in our lab showing
that this particular redox mediator does not cross bilayer lipid
membranes (BLMs),51 i.e., it yields a negative feedback when the
UME approaches the BLM. Once the SECM topography informa-
tion was acquired, the tip potential was switched to E ) -0.5 V
versus Ag/AgCl, 3 M KCl, ensuring the reduction of the oxygen
present in the solution, thus yielding oxygen uptake profiles.

RESULTS AND DISCUSSION

We have manufactured GOx enzyme UME biosensors by
employing two methods: electropolymerization and hand-casting.
We have tested the properties of the enzyme films deposited onto
the Pt UMEs by doing amperometric calibrations prior to any
SECM applications. Figure 3A displays a typical example of a
calibration toward the analyte of interest (i.e., glucose) of an
electropolymerized GOx UME biosensor. The sensor displays
linear sensitivity to glucose for concentrations of up to 18 mM,
and there is no response to blank additions (buffer, no glucose).
In order to find the “zero distance point” on the AC, the sensors
were allowed to approach until they physically touched the
substrate surface. In order to determine whether the procedure
adversely affected performance, after the SECM experiments the
sensors were tested again for sensitivity with respect to their
analyte. Figure 3B shows an example of amperometric calibration
for a GOx UME sensor after use in conjunction with SECM for
recording ACs. The sensor UME continues to display linear
sensitivity to glucose for concentrations of up to 18 mM, and there
is no response to the blank additions. Comparatively, the cast GOx
UME biosensors display a narrower linear sensitivity range, for
glucose concentration of up to 14 mM. Figure 3C presents the
stability of an electropolymerized GOx UME sensor in time. The
data points in Figure 3C have been obtained from amperometric
calibration curves similar to the ones presented in Figure 3, parts
A and B. Figure 3C displays one measurement (data point) for
each day for each concentration represented on the graph. The
electrode was stored in the refrigerator in 50 mM phosphate buffer
pH 7 between runs. There is little difference between the Figure 3. Amperometric calibrations of electropolymerized GOx 25
µm UME biosensors, E ) +0.6 V vs Ag/AgCl, 3 M KCl; additions up
calibrations done over the course of 12 days, and this minute to 18 mM glucose in 50 mM phosphate buffer, pH 7: (A) GOx UME
variation is not important in the context of the fact that the sensors exhibit sensitivity to glucose prior to SECM experiments; (B)
electrodes should always be calibrated the day of the measure- the sensors preserve the sensitivity to their analyte after being
ments. The electropolymerized GOx sensors (if stored at 4 °C in employed in extensive SECM experiments; (C) the electropolymerized
buffer) preserve their sensitivity toward glucose for periods of GOx biosensors are stable in time when stored in 50 mM phosphate
buffer, pH 7 at 4 °C for periods of at least 2 weeks; the same sensor
time for up to 1 month. In contrast, the cast 25 µm GOx sensors shows almost identical sensitivity after 1, 10, and 12 days from the
(stored at room temperature, no buffer) start to lose sensitivity electropolymerization date. The data presented in this figure are for
to glucose after 2-3 days. Different storage conditions have been three distinct electropolymerized GOx sensors (one sensor for the
explored for all types of sensors presented in this paper, and the data in (A), one sensor for the data in (B), and another sensor for the
optimum ones for each sensor type are the only ones discussed data in (C)).
here (dry storage at room temperature for hand-cast biosensors
and 4 °C in buffer for electropolymerized biosensors). One The lactate sensor production is complicated by the relatively
advantage that the cast GOx sensors have over the similar lower stability of LOx as compared to GOx. Figure 4 compares
electropolymerized ones is the more facile manufacturing proce- the behavior of electropolymerized and cast LOx UME sensors.
dure, which renders the films applied on the 25 µm Pt UME as For both types of LOx-based sensors, the calibration plots seem
“disposable”. The overall success rate for production of sensors to exhibit two distinct linear ranges: 0.1-0.3 mM that covers the
(determined by whether sensors exhibited sensitivity toward their likely physiological range for single-cell release, and 0.4-0.9/1.0
analyte the same day as manufacturing) was ∼70% for the mM lactate, which may be useful for future studies (e.g., scanning
electropolymerization process and ∼90% for the casting method across surfaces with patterned LOx, in the presence of lactate in
(we have tested at least 20 electrodes for each method). solution). Horrocks et al. 10 have observed the same type of dual
(51) Wilburn, J. P.; Wright, D. W.; Cliffel, D. E. Analyst 2006, 131 (2), 311- range when they calibrated a horseradish peroxidase-modified
316. UME with respect to H2O2 sensitivity; they have proposed that
2722 Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
Figure 4. Amperometric calibrations of (A) electropolymerized and (B) cast LOx 25 µm UME biosensors, E ) +0.6 V vs Ag/AgCl, 3 M KCl;
additions up to 1 mM lactate in 50 mM phosphate buffer, pH 7. The sensors were calibrated for concentrations of up to 1 mM lactate and display
two distinct linear regions: below 0.3 mM and above 0.4 mM, with the expected cellular lactate release range below 0.3 mM. The sensors retain
lactate sensitivity upon being used in SECM experiments. Double-reciprocal plot representations of these amperometric calibrations: (C) the
electropolymerized LOx sensor for which the corresponding data is shown in (A); (D) cast LOx sensor for which the corresponding data is
shown in (B)sbefore use in SECM.

this decay in sensitivity was due to the redox process in parts of previously suggested.10 Since the concentration of lactate produc-
the polymer in poor contact with the UME element (carbon fiber tion by cells is very low, the concentrations of produced lactate
in their case). The electropolymerized LOx sensors generally lose are contained in the lower regions (below 0.3 mM lactate), which
sensitivity to lactate above 0.9 mM lactate (Figure 4A). However, are linear under the Michaelis-Menten behavior, and thus the
these UME sensors are assembled with the help of oxidases, and enzyme saturation with the substrate does not hinder the detection
as such they should obey the enzyme kinetics laws. For enzymes for our purposes.
following the Michaelis-Menten behavior, one should look at the Just as for the GOx UME biosensors, we have tested the
Lineweaver-Burk double-reciprocal plot, where the inverse of the properties of the LOx sensors before and after the use in
reaction rate is directly proportional to the inverse of the substrate conjunction with the SECM, to validate the SECM results. Figure
concentration5 (e.g., lactate in our case). Since the reaction rate 4B displays examples of calibration plots for a cast LOx UME
is directly proportional to the current,52 we can plot the inverse biosensor in the lower linear concentration range before and after
of the current versus the inverse of the lactate concentration, for recording ACs (where the tip was allowed to touch the Petri dish
example, and if the plot is linear, then the enzyme follows the substrate in order to find the “zero distance point”), showing that
Michaelis-Menten kinetics. Figure 4C displays the data from sensitivity toward lactate is retained. Although the electropoly-
Figure 4A in the form of the Lineweaver-Burk double-reciprocal merized LOx sensors are more reproducible, they have a lower
plot, and Figure 4D does the same thing for the data from Figure life time (∼24 h, stored at 4 °C in buffer) compared with the cast
4B (before SECM). Both plots appear linear, indicating that the LOx sensors (2 days, room temperature). The manufacturing
presence of two distinct regions in Figure 4, parts A and B, are success rate is overall lower for the LOx sensors than for the GOx
due to a saturation of the LOx enzyme with the lactate substrate, sensors: ∼20% for electropolymerized LOx and ∼60% for cast LOx
meaning that the enzymes follow the Michaelis-Menten kinetics. sensors (we have tested at least 20 electrodes for each method).
The R2 values for the Lineweaver-Burk double-reciprocal plots Just as in the case of the GOx sensors, the cast LOx sensors are
in both cases (Figure 4, parts C and D) are higher than the R2 easier to manufacture than the electropolymerized ones, and the
values for the linear ranges corresponding to the higher concen- cast LOx films can be regarded as “disposable”.
trations (>0.4 mM, Figure 4, parts A and B), indicating that these
films should be regarded as obeying the Michaelis-Menten (52) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and
behavior, rather than having two distinct linear regions as Applications, 2nd ed.; John Wiley & Sons, Inc.: New York, 2001.

Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2723

 Figure 6. Glucose uptake profile for a single fibroblast. E ) +0.6
V vs Ag/AgCl, 3 M KCl; 5 mM glucose in HBSS, pH 7.4; scanning
speed 15 µm/s; tip, GOx cast film on 25 µm UME biosensor; d ) 10
µm from Petri dish; data have been corrected for background tilt in
Origin Pro 7.5. The GOx cast 25 µm UME tip was positioned at 10
µm from the Petri dish by using approach curves (ACs).

Figure 7. Lactate production from a single cancer cell MCF10CA1a.

E ) +0.6 V vs Ag/AgCl, 3 M KCl; no lactate added in RPMI media,
pH 7.4; scan speed 3 µm/s; tip, LOx cast film on 25 µm UME
biosensor; red line, lactate production current associated with a single
MCF10CA1a cancer cell; blue line, in the same Petri dish, same
electrode, scan across the surface when no cancer cells are present
in the path of the electrode; data have been corrected for background
tilt in Origin Pro 7.5.

Figure 5. Approach curves (ACs) for 25 µm UME enzyme-based

biosensors; E ) +0.6 V vs Ag/AgCl, 3 M KCl; in 50 mM phosphate For the sensor no. 2 in Figure 5A the current increases upon the
buffer, pH 7: (A) GOx electropolymerized sensors no. 1 (approach tip touching the substrate; this is probably due to a slight
speed 1.5 µm/s) and no. 2 (approach speed 3 µm/s) yielded similar compression of the polymer layer which in turn could cause a
curves in 5 mM glucose, indicating reproducibility; (B) GOx cast
sudden increase in the concentration of electroactive species at
sensor (5 mM glucose, approach speed 3 µm/s); (C) LOx cast (0.5
mM lactate, approach speed 15 µm/s). In all plots the vertical dashed the Pt sensing element. Parts B and C of Figure 5 display ACs
line represents the point where the UME tip touches the Petri dish, for cast electrodes: GOx and LOx cast UME biosensors, respec-
coming from the right side (solution side). tively. For the case of GOx-based sensors the ACs can be used
as means of positioning the tip at a known distance from the
substrate when imaging live cells, for example (since the solution
As expected, for both types of sensors discussed in this paper, contains glucose needed for cell viability). In the case of metabolic
the electropolymerized ones were more reproducible than the cast lactate imaging, there is insufficient lactate in the extracellular
ones. During the casting procedure it is difficult to control the solution (e.g., produced by a single cell, since the solution has
exact amount of enzyme solution that gets deposited onto the Pt no added lactate) to allow an AC to be performed, although there
tip and/or the casting solution homogeneity (for electrodeposition might be other applications where the lactate ACs might find use
the solution is stirred). However, individual sensor calibrations (e.g., study of LOx kinetics). Upon conducting an extensive review
must be performed prior to use, making this less important. Figure of the literature we were unable to find any published studies
5A presents ACs for two distinct electropolymerized GOx UME featuring ACs generated by enzyme electrodes; therefore, Figure
sensors manufactured a few months apart (both tested the second 5 presents the first amperometric ACs recorded with enzyme-
day after manufacturing, approaching to plastic Petri dishes) and coated UMEs. The ACs for the enzyme-coated UMEs do not obey
illustrates overall reproducibility of the electropolymerization the classic SECM feedback theory, due to the presence of the
manufacturing process. We can compare the ACs since we have polymer layer deposited on the Pt surface. The attempts of fitting
the “zero separation distance” for each curve, and currents have the ACs to the simple kinetically limited (and/or mass transfer
been normalized with respect to their steady-state current values. limited) feedback modes for SECM did not succeed either,
2724 Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
Figure 8. Topography (A and C) and oxygen uptake profile (B and D) of a single PC12 cell. E ) +0.6 V vs Ag/AgCl, 3 M KCl for topography
and E ) -0.5 V vs Ag/AgCl, 3 M KCl for respiration; 1 mM FcTMA in Ringer’s solution pH 7.4; tip, Pt 7 µm UME and scanning speed 30 µm/s
for both images; data have been corrected for background tilt in Origin Pro 7.5. The tip was positioned by ACs to the insulating Petri dish
surface: d = 10 µm from the Petri dish for topography, and d = 8 µm from Petri dish for respiration. Panels A and B display the 3D contour plots
for topography and respiration, respectively. Panels C and D are the same images from panels A and B with “crosshairs” for easier comparison
of the topography and respiration profiles: horizontal crosshair corresponds to the scan presented above the image; vertical crosshair corresponds
to the scan presented on the right side of the image. All distances are in micrometers.

pointing to the fact that the processes occurring at the tip are of the substrate37 is that the tip can be damaged in the process.
complex, and this opens the door to future studies for unfolding By using the ACs in conjunction with the optical monitoring of
the intricacies of enzyme-coupled electrode kinetics for these UME the moving tip, we minimize the opportunities for damaging the
sensors. sensor. Upon establishing physical contact between the tip and
We have used the ACs for the GOx UME sensors to position the substrate, the tip is retracted the desired distance using the
the tip at a certain distance (10 µm) from the glass bottom Petri z-axis inchworm of the SECM setup, and consequently the tip-
dish on which fibroblasts were cultured at low density (2000 cells substrate separation distance that we report for the cell imaging
in the entire dish) to permit the imaging of a single fibroblast is as accurate as the movement of the z-axis inchworm. Although
without interference from neighboring cells. We have used both it is possible that the polymer layer would slightly adhere to the
the visual monitoring of the enzyme UME approaching the surface surface of the dish and the geometry of the enzyme film would
of the Petri dish with the help of the inverted microscope and be slightly disturbed in the retraction movement, our calibration
the ACs to find the point where the tip touches the substrate. experiments (Figures 3 and 4) clearly indicate that the analyte
The problem with only physically lowering the tip onto the surface sensitivity of the UME biosensor is not altered upon recording
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2725
ACs and gently touching the surface of the substrate. Upon al.37 reported that the acquired respiration image identically
positioning the UME biosensor at the desired distance above the matched the expected topography (profile of the oxygen diffusion
Petri dish, the tip is rastered across the live cell while monitoring hindrance atop the cell, i.e., the oxygen distribution around the
its movement through the optical microscope. Figure 6 shows cell), and they were not able to show distinctive oxygen uptake
the results of such an experiment, revealing a glucose uptake profiles. While we have acquired images (taken at d ≈ 10 µm
profile above a single fibroblast from a background solution from the dish surface, not shown) that resemble the previously
containing 5 mM glucose. The current is the lowest on the top of published cell respiration images as cell topography, when the
the cell (as expected) and the uptake profile is uniform. The tip is scanned closer than 10 µm to the surface, it is possible to
glucose uptake profile above a single fibroblast looks similar to obtain images that allow differentiation in respiratory activity as
the topography profiles of single fibroblasts. It is possible that a function of spatial location. Since the diffusion of the oxygen is
the glucose uptake observed here is a combination of both extremely rapid in aqueous solutions, it is important to approach
diffusion/distribution profile of glucose around the fibroblast and closer to the cell so that the image obtained (E ) -0.5 versus
nutrient uptake by the cell. Consequently, for the study of cell Ag/AgCl, 3 M KCl) is not solely the result of the cell topography
metabolism one should acquire glucose uptake images in conjunc- but includes oxygen uptake concentration profiles. We have
tion with measuring the cellular production of an analyte, such recorded a respiration profile at d = 8 µm from the dish surface,
as lactate. and the respiration image (3D contour plot, Figure 8B) raster
One of the hallmarks of cancer cells is a lowered extracellular shows the lowest respiration activity in the middle of the cell
pH which is in part due to increased anaerobic respiration and surface, and the highest respiration activity at the edge of the
cell, where the cell makes contact with the Petri dish. Although
lactate production by the cell (Figure 1). We have picked for this
this image does certainly have a component that comes from the
study an invasive cell line (MCF10CA1a) that forms tumors and
distribution of the oxygen around the cell, it is obvious that we
metastasis in mice.50 Figure 7 represents lactate production by a
were able to record the oxygen depletion by the cell. Also, one
single MCF10CA1a cell, as sensed by a cast LOx 25 µm UME
could argue that the reason we are not observing uniform oxygen
tip. The onset and falloff of current during the scan was observed
diffusion across the cell membrane is due to the higher distribu-
(via the inverted microscope) to coincide with the period when
tion of mitochondria (where cellular respiratory activity occurs)
the UME sensor was physically above the cell. The topography
at the points where the cell makes contact with the Petri dish
of the MCF10CA1a cells is hemispherical; however, because the
substrate. Parts C and D of Figure 8 are the corresponding plots
lactate is produced by the cell (as opposed to being consumed,
for topography (Figure 8A) and respiration (Figure 8B), respec-
as in the case of glucose uptake), an increase in the lactate
tively, and we have inserted crosshairs to better illustrate portions
concentration above the cell is a consequence of cellular activity,
of the 2D profiles. Figure 8C, which uses an external redox
and it cannot be mistaken for changes in cellular topography. The mediator and so is independent of physiological activity, shows a
overlaid background curve was generated by repeating the scan, clear dependence between the current profile and topography for
at the same tip-substrate separation, in a region of the dish where a roughly spheroid cell. Figure 8D, however, displays a more
no cells were present. We chose the control experiment of convoluted oxygen profile that cannot be explained solely by
scanning above the bare surface of the Petri dish in order to prove topography and, therefore, is indicative of physiological activity
that unless there is a source for the lactate (i.e., MCF10CA1a cell (i.e., respiration) and/or transport properties of the cell or
in our case), no current increase will be observed. membrane itself.
Cancer cells present enhanced aerobic glycolysis,44 and thus
oxygen is an important analyte (Figure 1) that has to be accounted CONCLUSIONS
for when the metabolic profiling of cancerous cells is sought. We
We have manufactured enzyme UMEs for SECM use in the
have recorded topographical images of PC12 cells in a medium
detection of lactate and glucose. We have calibrated these sensors
containing FcTMA, with the potential of the 7 µm Pt UME tip set with respect to their analyte, and we have shown for the first time
at E ) +0.6 V versus Ag/AgCl, 3 M KCl; at this potential, the that enzyme-based UMEs can be used for positioning the SECM
FcTMA present in solution gets oxidized at the Pt tip and a tip via amperometric ACs with respect to the substrate to be
topography-type profile is observed, corresponding to the amount scanned. By employing the GOx- and LOx-based UME biosensors,
of FcTMA available for immediate diffusion to the tip, i.e., we have obtained glucose uptake and lactate release profiles for
corresponding to the distance between the tip and the imaged single cancer cells as a proof of concept for the application of
object (PC12 cell in this case). An example of PC12 topography SECM toward the study of basic metabolic processes that are
can be observed in Figure 8A, where the 3D contour plot for the affected by disease in cancerous cells. We have also shown that
FcTMA oxidation profile around the cell is presented. The shape respiration studies performed at closer tip-substrate separations
of the cell appears uniform, as a hemispherical object. Upon (<10 µm) reveal a nonuniform distribution of oxygen that cannot
obtaining information about the cell surface, we have switched be explained by the cell topography alone; these oxygen uptake
the potential of the same Pt tip to E ) -0.5 versus Ag/AgCl, 3 M profiles could give a more complete metabolic picture of the single
KCl, to ensure the reduction of the oxygen present in solution at cancer cells. Further experimentation, however, is required to
the Pt UME, and we have scanned the same area and obtained a definitively assign this convoluted oxygen uptake distribution to
respiration profile of the same single PC12 cell as seen in Figure a difference in respiratory activities within the cell itself. The ability
8B. It is very interesting to observe that the oxygen profile we to look at single-cell lactate production and glucose and oxygen
recorded across the cell does not mirror the topography. Takii et consumption, in addition to previously published acidification rates
2726 Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
profiles, will provide for a direct comparison of single-cell data to 01-8064; the Vanderbilt Integrative Cancer Biology Center; and
our multicell data from MAMP experiments. the Vanderbilt Institute for Integrative Biosystems Research and
Education.
ACKNOWLEDGMENT
We thank Evgheni Kozlov and Ales Prokop (fibroblasts and
PC12), Julie Maier and Alissa Weaver (MCF10CA1a), and Kris- Received for review October 15, 2007. Accepted February
topher Kahlig and Aurelio Galli (PC12) for cell culture and plating. 04, 2008.
We acknowledge the following funding sources: NIAID U01
AI061223; NCI U54 CA113007; DARPA/ONR contract N66001- AC7021184

Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2727