Letter

Cite This: Nano Lett. 2019, 19, 6346−6351 pubs.acs.org/NanoLett

Wearable Sweat Band for Noninvasive Levodopa Monitoring

Li-Chia Tai,†,‡,§ Tiffany S. Liaw,†,‡,§ Yuanjing Lin,†,‡,§,⊥| Hnin Y. Y. Nyein,†,‡,§ Mallika Bariya,†,‡,§
Wenbo Ji,†,‡,§ Mark Hettick,†,‡,§ Chunsong Zhao,†,‡,§ Jiangqi Zhao,†,‡,§ Lei Hou,†,‡,§ Zhen Yuan,†,‡,§
Zhiyong Fan,⊥| and Ali Javey*,†,‡,§
†
Department of Electrical Engineering and Computer Sciences and ‡Berkeley Sensor and Actuator Center, University of California,
Berkeley, California 94720, United States
§
Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States
⊥|
Department of Electronic and Computer Engineering, Hong Kong University of Science and Technology, Clear Water Bay,
Kowloon, Hong Kong SAR, China
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

*
S Supporting Information
Downloaded via UNIV OF MISSOURI-ST. LOUIS on October 22, 2019 at 10:20:39 (UTC).

ABSTRACT: Levodopa is the standard medication clinically

prescribed to patients afflicted with Parkinson’s disease. In
particular, the monitoring and optimization of levodopa
dosage are critical to mitigate the onset of undesired
fluctuations in the patients’ physical and emotional conditions
such as speech function, motor behavior, and mood stability.
The traditional approach to optimize levodopa dosage
involves evaluating the subjects’ motor function, which has
many shortcomings due to its subjective and limited
quantifiable nature. Here, we present a wearable sweat band
on a nanodendritic platform that quantitatively monitors levodopa dynamics in the body. Both stationary iontophoretic
induction and physical exercise are utilized as our methods of sweat extraction. The sweat band measures real-time
pharmacokinetic profiles of levodopa to track the dynamic response of the drug metabolism. We demonstrated the sweat band’s
functionalities on multiple subjects with implications toward the systematic administering of levodopa and routine management
of Parkinson’s disease.
KEYWORDS: Wearable sweat sensor, Parkinson’s disease, levodopa detection, noninvasive drug monitoring, dosage optimization

L evodopa is the medication administered to treat patients

with Parkinson’s disease.1−7 Despite its success, an
individual’s responses to levodopa can vary due to factors
levodopa excretes through sweat with its concentration in
sweat exhibiting a potential correlation with that in human
plasma.19−22 Additionally, under standard levodopa dosage,
such as dietary intake, age, gender, and drug administration sweat levodopa level shows up in the micromolar ranges,17−19
history.8−12 These variations can lead to unfavorable making it amenable for reliable detection with current
fluctuations of the subject’s motor and cognitive functions if technologies.23,24 For these reasons, sweat presents an ideal
the levodopa dosage is not tailored toward the individuals.12−14 means for nonobstructive monitoring of levodopa for dosage
Therefore, levodopa monitoring is an essential part of the optimization. Sweat sensing patches have been employed for
treatment for Parkinson’s disease. The gold standard for drug tests in athletic doping control, drug abuse investigation,
optimizing levodopa dosage involves an assessment of the and forensic inspection.25,26 Recently, sweat sensors for drugs
motor function of a Parkinson’s disease patient.9 This method and their related biomolecules have also been demonstrated via
requires clinicians to evaluate the subject’s motor functions,
optical and electrochemical techniques.27−32 In particular, the
leading to difficulty of point-of-care testing and ambiguity in
electrochemical approach represents an attractive method
drug dosage. To address this challenge, monitoring blood
owing to its advantages for electronic integration, economical
levodopa concentration stands out as a viable solution.15−18
However, blood-based detection is hampered by its need for cost, sensitivity, and selectivity.29−36
invasive sampling and separate analytical tools, making it In this work, we expand the strength of the electrochemical
inappropriate for the long-term and frequent measurements sensor through integrated surface innovations at the physical
that are necessary due to the dynamic nature of drug and chemical levels. The incorporation of gold dendritic
metabolism. nanostructures onto the electrodes remarkably enhanced the
Taking this into consideration, human sweat is an alternative
to blood due to its accessibility through noninvasive Received: June 18, 2019
procedures and its abundance of biomolecules. Like other Revised: July 26, 2019
drug molecules that undergo xenobiotic metabolism pathway, Published: August 5, 2019

© 2019 American Chemical Society 6346 DOI: 10.1021/acs.nanolett.9b02478

Nano Lett. 2019, 19, 6346−6351
Nano Letters Letter

Figure 1. Schematic of the s-band and drug sensing mechanism. (a) Optical image of the s-band worn on a subject’s wrist. (b) Sensing mechanism
of the levodopa sensor. WE, RE, and CE are working electrode, reference electrode, and counter electrode, respectively. (c) Cross-section view of
the gold electrodes on a flexible sensor patch. (d) Scanning electron microscope image of the gold dendritic structures. (e) Real-time sweat
levodopa monitoring using the s-band after levodopa intake.

Figure 2. Characterization of the functionalized working electrode in PBS. (a) CV of levodopa (LD) dissolved in PBS and (b) zoom-in view for
smaller concentration range. (c) Amperometric response of levodopa dissolved in PBS. Inset shows the calibration curve.

detection limit of the levodopa sensor down to about 1 μM in evaporated Au/Cr conductive layer.38,39 Further, thionin
sweat solution, a couple-times improvement compared to acetate salts are deposited via cyclic voltammetry (CV)
previous work tested in human fluid.37 Moreover, the cross- method conformally on the as-synthesized dendritic gold
linking mechanisms with glutaraldehyde provide us chemically structures, and glutaraldehyde/tyrosinase is drop-casted onto
robust enzymatic structures to achieve sensor stability for long- the working electrode. Glutaraldehyde serves as the cross-
term and continuous usage.37−40 This approach effectively linker to immobilize the tyrosinase enzyme that facilitates the
connects existing sensor enhancement technologies into a electrochemical oxidation of levodopa.37,40 It is also worth
consolidated platform for prolonged sensor operation. Our mentioning that the dendritic gold plays a pivotal role to
solutions effectively address the challenges for drug detection, provide adequate interface for enzyme loading and molecular
which are attributed to the generally low concentration of contact to achieve an improved sensing performance. The final
drugs in human sweat and the long time scale of drug modification step for the working electrode involves the drop-
metabolism. casting of Nafion, which enhances the long-term stability and
The design of the wearable sensor packaged into a sweat antifouling features of the electrochemical sensor.41 The
band (s-band) is illustrated in Figure 1a. The sensor is prospective application of the sensor is illustrated in Figure
fabricated on a polyethylene terephthalate (PET) substrate, 1e, which shows the application of iontophoresis for
and it employs a standard three-electrode configuration with a noninvasive and stationary stimulation of sweat to monitor
functionalized levodopa sensing electrode as working elec- levodopa levels after a subject consumes the drug. The s-band
trode, a Ag/AgCl top layer as reference electrode, and a Au top enables continuous monitoring of levodopa, which allows for
layer as counter electrode. As shown in Figure 1b, during personalized optimization of levodopa dosage.
amperometric measurement, levodopa excreted in sweat can be The functionalized levodopa sensing electrode was charac-
oxidized via tyrosinase enzyme to dopaquinone.37 This process terized with CV scanning, which indicated the oxidation peak
generates a Faradaic current that can be further calibrated into of levodopa. Figure 2a shows the CV curves using the
its corresponding sweat levodopa concentration. The cross- functionalized electrode in phosphate-buffered saline (PBS)
section schematic of the electrodes is shown in Figure 1c, and with different concentrations of levodopa. The oxidation and
Figure 1d shows a scanning electron microscope image of the reduction peaks for levodopa are around 0.34 and 0.30 V
gold dendrites to validate the successful growth and density of (method of peak identification is illustrated in Figure S1),
the nanostructures. To achieve the enhancement on sensitivity, respectively, which are consistent with the literature.23 Figure
gold nanodendrites with largely increased surface area are 2b shows a CV in the proximity of levodopa’s oxidation peak in
synthesized via overpotential deposition approach on the the physiologically relevant concentration range.19 The
6347 DOI: 10.1021/acs.nanolett.9b02478
Nano Lett. 2019, 19, 6346−6351
Nano Letters Letter

Figure 3. Characterization of the levodopa sensor in sweat solution. (a) CV of levodopa dissolved in sweat. (b) Amperometric response of
levodopa dissolved in sweat and the (c) corresponding calibration curve. (d) Interference study of the levodopa sensor after the addition of
levodopa (LD), uric acid (UA), glucose (G), and ascorbic acid (AA).

Figure 4. Levodopa monitoring via iontophoresis-induced sweat. Sweat levodopa concentration is monitored continuously after fava beans
consumption and subsequent applications of iontophoresis (a) without and (b) with prior dietary consumption. The horizontal axis indicates the
time elapsed after the subject consumes 450 g of fava beans.

functionalized electrode responds to levodopa with high concentration (e.g., 10 μM) over a period of operation (e.g., 30
sensitivity in the micromolar range. This is a remarkable min). The sensor drift is approximately 18 nA as shown in
enhancement of more than two orders of magnitude compared Figure S3. The error of concentration measurement due to
to the response of a bare Au electrode (Figure S2). Figure 2c drift is estimated to be 1.2 μM. The limit of detection is 1.25
shows the amperometric response of levodopa at the oxidation μM as shown in Figure 3b. The signal-to-noise ratio (SNR) is
potential (0.34 V). The inset displays the calibration curve the ratio of the square of the amplitude of the signal to the
with a sensitivity of 15 nA/μM, which is on par with the best background noise. On the basis of Figure S3b, the
levodopa sensors reported.23 This is notable considering the amperometric response signal is found to be 225 nA. During
simplicity of electrode functionalization and the electrode’s the measurement, the noise level is observed to be
excellent stability for long-term usage. approximately 50 nA, which results in a SNR of 20. The
The sensors were further characterized in sweat solution to selectivity of the sensor is essential in the presence of other
demonstrate the s-band’s practical applications for noninvasive common sweat biomolecules. Therefore, the amperometric
monitoring. Figure 3a shows a CV in the proximity of responses of the addition of levodopa and its potential
levodopa’s oxidation peak, which is identified to be 0.25 V. interferents such as uric acid (20 μM), glucose (166 μM),
The amperometric response of levodopa was similarly tested at and ascorbic acid (16 μM) are recorded in Figure 3d. The
the oxidation potential, shown in Figure 3b. The correspond- concentrations are chosen to be in their physiologically
ing calibration curve in sweat is displayed in Figure 3c, and the relevant ranges.29−32 The result shows that the interference
sensitivity is found to be 1.7 nA/μM. This value is different on the levodopa sensor performance is within an error range of
from that in PBS and is expected due to biofouling activity.29 0.35 μM.
For the purpose of the study, we define drift as the maximum To explore the viability of using the s-band to track the
change in the signal of the amperometric response of a fixed metabolism of levodopa in human subjects, sweat was
6348 DOI: 10.1021/acs.nanolett.9b02478
Nano Lett. 2019, 19, 6346−6351
Nano Letters Letter

extracted from a healthy volunteer through iontophoresis and

after Vicia faba (fava bean) consumption. Fava beans are
levodopa-containing legumes consumed for culinary and
medicinal purposes.18 The use of fava beans allows for
extensive testing of the s-band’s functionalities on non-
vulnerable, healthy subjects.
Figure 4a shows the time progression panel of fava beans
consumption and iontophoresis application as well as
continuous sensor readings of the sweat levodopa concen-
tration. The result demonstrates that the s-band can
continuously capture the sweat levodopa trend that resembles
the blood levodopa profile observed.15−18 During iontopho-
retic stimulation, sweating begins after iontophoresis and lasts
for about 20 to 45 min. Therefore, two consecutive
applications of iontophoresis are performed at −20 and 25
min, with respect to the time of fava beans intake, to cover a
broad time range that captures levodopa’s metabolic trend.
The subjects consume 450 g of fava beans after 12 h of fasting,
and their sweat levodopa concentration is monitored after each
iontophoresis. The observed sweat levodopa level versus time
shows an increasing trend up to 6.6 μM at about 47 min, after
which the concentration begins to decrease. The levodopa
concentration decreases to 3.3 μM at about 74 min, so the half-
life of the decay is found to be 27 min. The time scales of
levodopa’s half-life and its time of peak concentration are
expected based on previous observations.15−18 In Figure 4b,
the subject first consumes a 426 g sandwich (see Methods)
followed by 450 g of fava beans. Three consecutive
applications of iontophoresis are performed at −20, 25, and
80 min, with reference to the time of fava beans consumption. Figure 5. Levodopa monitoring via exercise induced sweat. (a)
The result shows a slight delay of 13 min in the Cycling and sweat analysis. Examples of sweat levodopa concen-
pharmacokinetic peak time compared to that in Figure 4a. trations for three different subjects after they consume 450 g of fava
beans. (b) Averaged time of peak levodopa concentration for three
This finding is expected as dietary intake can affect the
different subjects across multiple exercise trials.
pharmacokinetic profile of levodopa in human secretory
systems.11 The data in Figures 4a and b verifiy the s-band’s
capability of continuous levodopa measurement. of levodopa, the standard medication prescribed to Parkinson’s
While sweating caused by iontophoresis can be limited in its disease patients. The levodopa s-band integrates various
duration, sweating generated through exercise can typically last material innovations and enables us to gain fundamental
longer and allow us to capture a more complete picture of the insights into the pharmacokinetic behavior of levodopa
drug’s pharmacokinetics. Figure 5 explores the possibility of noninvasively. Dendritic growth, enzyme immobilization, and
exercise as a means to extend the sweating period. Figure 5a stabilizing film are seamlessly incorporated to improve the
shows the time progression panel of fava beans intake and an robustness and stability of the electrochemical sensor. We have
image of a subject engaging in ergometer cycling. The also demonstrated the application of the s-band for prolonged,
representative pharmacokinetic profiles of sweat levodopa for continuous, and noninvasive drug monitoring in human
three different subjects are included. In each trial, the subject subjects after fava beans intake. Through analyzing sweat
consumes 450 g of fava beans and exercises on a stationary generated via iontophoresis and physical activities, the
ergometer. Each subject exercises for multiple trials, and the metabolism of the drug can be tracked in real-time to allow
cumulative result is shown in Figure 5b, with the averaged time for dosage optimization. Future directions include investigating
of peak concentration for subject 1, 2, and 3 being 44 ± 20 pharmacodynamics between drugs, lengthening iontophoresis
min, 42 ± 26 min, and 67 ± 14 min. It is worth noting that sweating duration, and improving electrode lifetime upon
Figure 4a and Subject 1 of Figure 5a correspond to the same repeated use. We envision that the wearable s-band can be
subject. The results from iontophoresis and exercise sweat leveraged to study the intrinsically complex drug profiles,
show similar time of peak concentration (47 min versus 50 optimize drug dosages to regulate Parkinsonian behaviors in
min). The sets of on-body studies are compared to the case patients, and integrate with drug delivery systems. This
where no fava beans are consumed, which demonstrates almost platform serves as a pathway toward drug management for
zero concentration of levodopa (Figure S4). The on-body increasingly personalized, point-of-care medicine for the future.
experiments demonstrate the novelties and feasibilities of the Methods. Sensor Fabrication. The flexible electrodes were
wearable s-band for noninvasive and continuous monitoring of fabricated on PET substrates via photolithography and
levodopa’s dynamic metabolic rate. We envision that this evaporation. The electrodes were patterned through photo-
sensor platform can enable clinical understanding of xenobiotic lithography with positive photoresist (Shipley Microposit
metabolisms and dosage optimizations. S1818) and electron-beam evaporation of Cr (30 nm) and
In summary, we demonstrated the performance of a Au (50 nm). Afterward, lift-off in acetone solution was
wearable sweat band for monitoring the metabolic behavior performed. Au nanodendrites were grown on top of the
6349 DOI: 10.1021/acs.nanolett.9b02478
Nano Lett. 2019, 19, 6346−6351
Nano Letters Letter

electrodes with a Gamry Electrochemical Potentiostat (signal deviation of all the exercise trials for the same subject. One trial
type, square wave; signal frequency, 50 Hz; amplitude, 1 V; was performed for subject 3, and the uncertainty was estimated
DC offset, −1 V; cycles, 6000) and chloroauric acid solution from the slope (concentration/time) around the proximity of
(mixture of 50 mM AuCl3 and 50 mM HCl). The precursor the pharmacokinetic peak and the sensor drift (uncertainty in
concentration will affect the morphology and length of the concentration).
nanostructures, while increasing the deposition time will Food Intake. Fava beans were purchased from a local
normally increase the length of the nanostructures.42 The community market (Berkeley Bowl). Two-hundred-fifty grams
6000 cycles (120 s) of Au deposition were chosen because the of fava beans consumption is equivalent to about 125 mg of
resulting functionalized electrode showed the largest current levodopa intake.18 The sandwich was a 12-in. Italian B. M. T.
change upon the addition of 10 μM of levodopa, as shown in sandwich (Subway). The fava beans’ seeds were taken out from
Figure S3. The electrodes were immediately cleaned with the bean pods and cooked in boiling water for 20 min prior to
deionized water and left at room temperature for 2 h for consumption.
drying. On top of the Au nanodendrites, 0.25 mM of thionin
acetate salt (Sigma-Aldrich) was deposited electrochemically
with CV (initial potential, −0.6 V; final potential, 0.1 V; scan
■
*
ASSOCIATED CONTENT
S Supporting Information
rate, 0.1 V/s; segment, 40). The electrodes were then left at The Supporting Information is available free of charge on the
room temperature for 2 h. Subsequently, a mixture of enzyme ACS Publications website at DOI: 10.1021/acs.nano-
and cross-linker solution was drop-casted on top of the lett.9b02478.
electrodes (2.5 μL). The solution was prepared by mixing
tyrosinase (Sigma-Aldrich. One mg) with 2% glutaraldehyde Method of determining peak potential; CV of levodopa
(Sigma-Aldrich. 0.866 μL) in PBS (66.6 μL). The electrodes dissolved in PBS tested with bare Au working electrode;
were left at room temperature for 12 h. Finally, Nafion 117 long-term amperometric test for functionalized working
(Sigma-Aldrich. One μL) was drop-casted on top of the electrode; levodopa sensor response with iontophoresis-
electrodes, and the electrodes were left at room temperature induced sweat without fava beans consumption (PDF)
for 2 h. For the reference electrodes, Ag/AgCl paste was
painted on top of the Au electrodes and left at room
temperature for 12 h.
■ AUTHOR INFORMATION
Corresponding Author
Sensor Characterization and Calibration. The function- *E-mail: [email protected].
alized electrodes were characterized electrochemically using
ORCID
CHI 1230C potentiostat (CH Instrument) in CV and
amperometric measurements. The CV and amperometric Li-Chia Tai: 0000-0001-7042-5109
responses corresponding to different concentrations of Yuanjing Lin: 0000-0002-8568-1786
levodopa were subsequently evaluated. The interference test Hnin Y. Y. Nyein: 0000-0002-5692-6182
involved the addition of various biomolecules at the Mallika Bariya: 0000-0002-3416-8157
physiologically relevant concentrations. In Figure 2, the Wenbo Ji: 0000-0002-7913-361X
functionalized working electrodes were characterized with Zhiyong Fan: 0000-0002-5397-0129
commercial Ag/AgCl reference electrode in PBS. In Figure 3, Ali Javey: 0000-0001-7214-7931
on-body functionalized electrode arrays, including function- Author Contributions
alized working electrodes, Ag/AgCl pasted reference electro-
L.-C.T. and A.J. designed the experiments. L.-C.T., T.S.L., Y.L.,
des, and Au counter electrodes (all on PET substrate), were
and A.J. contributed to data collection, analysis, and
characterized in sweat solutions collected from the exercise
interpretation. L.-C.T., T.S.L., Y.L., H.Y.Y.N., M.B., W.J.,
trials. M.H., and C.Z. contributed to sensor fabrication and
Iontophoresis Sweat Analysis. Iontophoresis was con- preparation. L.-C.T. and A.J. wrote the paper, and all authors
ducted by gently mounting pilocarpine hydrogel (ELIT- provided feedback.
echGroup SS-023 Pilogel) on a subject’s wrist for 5 min at 1
mA DC current (ELITechGroup Model 3700 Webster Sweat Notes
The authors declare no competing financial interest.

■
Inducer). During in situ evaluation, electrode arrays were
placed conformal to the skin and connected to the CHI 1230C
potentiostat. The on-body sweat analysis of human subjects ACKNOWLEDGMENTS
was approved by the institutional review board (CPHS 2016− This work was supported by the National Science Foundation
06−8853) at the University of California, Berkeley. The on- (NSF) Nanomanufacturing Systems for Mobile Computing
body sweat analyses were processed with MATLAB’s Hampel and Mobile Energy Technologies (NASCENT) and the
and Smooth functions for noise reduction. Berkeley Sensor & Actuator Center (BSAC). The sensor
Exercise Sweat Analysis. The subjects engaged in stationary fabrication was performed in the Electronic Materials (E-
cycling on an ergometer (Gold’s Gym 290C Upright Cycle MAT) laboratory funded by the Director, Office of Science,
Trainer) at a biking power of 100 W. The electrode arrays Office of Basic Energy Sciences, Materials Sciences and
were calibrated and tested with the same method as that in the Engineering Division of the U.S. Department of Energy
Iontophoresis Sweat Analysis section. The continuous data are under Contract No. DE-AC02-05CH11231.
plotted when the sensor starts to respond in sweat solutions.
This corresponds to the time when we observe obvious
sweating on the subject and sufficient accumulation of sweat
■ REFERENCES
(1) Piccini, P.; Brooks, D. J.; Björklund, A.; Gunn, R. N.; Grasby, P.
around the sensor. The uncertainty for the time of peak M.; Rimoldi, O.; Brundin, P.; Hagell, P.; Rehncrona, S.; Widner, H.;
concentration in Figure 5b was defined as the standard Lindvall, O. Nat. Neurosci. 1999, 2, 1137−1140.

6350 DOI: 10.1021/acs.nanolett.9b02478

Nano Lett. 2019, 19, 6346−6351
Nano Letters Letter

(2) Kravitz, A. V.; Freeze, B. S.; Parker, P. R. L.; Kay, K.; Thwin, M. Lien, D.-H.; Ota, H.; Cho, G.; Javey, A. Adv. Mater. 2018, 30,
T.; Deisseroth, K.; Kreitzer, A. C. Nature 2010, 466, 622−626. 1707442.
(3) Garnett, E. S.; Firnau, G.; Nahmias, C. Nature 1983, 305, 137− (30) Gao, W.; Emaminejad, S.; Nyein, H. Y. Y.; Challa, S.; Chen, K.;
138. Peck, A.; Fahad, H. M.; Ota, H.; Shiraki, H.; Kiriya, D.; Lien, D.-H.;
(4) Kortekaas, R.; Leenders, K. L.; van Oostrom, J. C. H.; Vaalburg, Brooks, G. A.; Davis, R. W.; Javey, A. Nature 2016, 529, 509−514.
W.; Bart, J.; Willemsen, A. T. M.; Hendrikse, N. H. Ann. Neurol. 2005, (31) Emaminejad, S.; Gao, W.; Wu, E.; Davies, Z. A.; Nyein, H. Y.
57, 176−179. Y.; Challa, S.; Ryan, S. P.; Fahad, H. M.; Chen, K.; Shahpar, Z.;
(5) van Kessel, S. P.; Frye, A. K.; El-Gendy, A. O.; Castejon, M.; Talebi, S.; Milla, C.; Javey, A.; Davis, R. W. Proc. Natl. Acad. Sci. U. S.
Keshavarzian, A.; van Dijk, G.; El Aidy, S. Nat. Commun. 2019, 10, A. 2017, 114, 4625−4630.
310. (32) Kim, J.; Jeerapan, I.; Imani, S.; Cho, T. N.; Bandodkar, A.;
(6) Frankel, J. P.; Kempster, P. A.; Bovingdon, M.; Webster, R.; Cinti, S.; Mercier, P. P.; Wang, J. ACS Sens. 2016, 1, 1011−1019.
Lees, A. J.; Stern, G. M. J. Neurol., Neurosurg. Psychiatry 1989, 52, (33) Krishnan, S. R.; Ray, T. R.; Ayer, A. B.; Ma, Y.; Gutruf, P.; Lee,
1063−1067. K.; Lee, J. Y.; Wei, C.; Feng, X.; Ng, B.; Abecassis, Z. A.; Murthy, N.;
(7) Cotzias, G. C.; Van Woert, M. H.; Schiffer, L. M. N. N. Engl. J. Stankiewicz, I.; Freudman, J.; Stillman, J.; Kim, N.; Young, G.;
Med. 1967, 276, 374−379. Goudeseune, C.; Ciraldo, J.; Tate, M.; Huang, Y.; Potts, M.; Rogers, J.
(8) Deleu, D.; Jacob, P.; Chand, P.; Sarre, S.; Colwell, A. Neurology A. Sci. Transl. Med. 2018, 10, No. eaat8437.
2006, 67, 897−899. (34) Takei, K.; Takahashi, T.; Ho, J. C.; Ko, H.; Gillies, A. G.; Leu,
(9) Haaxma, C. A.; Bloem, B. R.; Borm, G. F.; Oyen, W. J. G.; P. W.; Fearing, R. S.; Javey, A. Nat. Mater. 2010, 9, 821−826.
Leenders, K. L.; Eshuis, S.; Booij, J.; Dluzen, D. E.; Horstink, M. W. I. (35) Ota, H.; Chen, K.; Lin, Y.; Kiriya, D.; Shiraki, H.; Yu, Z.; Ha,
T.-J.; Javey, A. Nat. Commun. 2014, 5, 5032.
M. J. Neurol., Neurosurg. Psychiatry 2007, 78, 819−824.
(36) Boutry, C. M.; Beker, L.; Kaizawa, Y.; Vassos, C.; Tran, H.;
(10) Sossi, V.; de la Fuente-Fernández, R.; Schulzer, M.; Adams, J.;
Hinckley, A. C.; Pfattner, R.; Niu, S.; Li, J.; Claverie, J.; Wang, Z.;
Stoessl, J. Brain 2006, 129, 1050−1058.
Chang, J.; Fox, P. M.; Bao, Z. Nat. Biomed. Eng. 2019, 3, 47−57.
(11) Crevoisier, C.; Zerr, P.; Calvi-Gries, F.; Nilsen, T. Eur. J. Pharm.
(37) Brunetti, B.; Valdés-Ramírez, G.; Litvan, I.; Wang, J.
Biopharm. 2003, 55, 71−76. Electrochem. Commun. 2014, 48, 28−31.
(12) Svenningsson, P.; Rosenblad, C.; af Edholm Arvidsson, K.; (38) Lin, Y.; Gao, Y.; Fan, Z. Adv. Mater. 2017, 29, 1701736.
Wictorin, K.; Keywood, C.; Shankar, B.; Lowe, D. A.; Björklund, A.; (39) Wang, S.; Wu, Y.; Gu, Y.; Li, T.; Luo, H.; Li, L.-H.; Bai, Y.; Li,
Widner, H. Brain 2015, 138, 963−973. L.; Liu, L.; Cao, Y.; Ding, H.; Zhang, T. Anal. Chem. 2017, 89,
(13) Bézard, E.; Ferry, S.; Mach, U.; Stark, H.; Leriche, L.; Boraud, 10224−10231.
T.; Gross, C.; Sokoloff, P. Nat. Med. 2003, 9, 762−767. (40) Dempsey, E.; Diamond, D.; Collier, A. Biosens. Bioelectron.
(14) Picconi, B.; Centonze, D.; Håkansson, K.; Bernardi, G.; 2004, 20, 367−377.
Greengard, P.; Fisone, G.; Cenci, M. A.; Calabresi, P. Nat. Neurosci. (41) Qi, L.; Thomas, E.; White, S. H.; Smith, S. K.; Lee, C. A.;
2003, 6, 501−506. Wilson, L. R.; Sombers, L. A. Anal. Chem. 2016, 88, 8129−8136.
(15) Senek, M.; Aquilonius, S.-M.; Askmark, H.; Bergquist, F.; (42) Ye, W.; Yan, J.; Ye, Q.; Zhou, F. J. Phys. Chem. C 2010, 114,
Constantinescu, R.; Ericsson, A.; Lycke, S.; Medvedev, A.; Memedi, 15617−15624.
M.; Ohlsson, F.; Spira, J.; Westin, J.; Nyholm, D. Eur. J. Clin.
Pharmacol. 2017, 73, 563−571.
(16) Harder, S.; Baas, H.; Rietbrock, S. Clin. Pharmacokinet. 1995,
29, 243−256.
(17) Nyholm, D.; Lewander, T.; Gomes-Trolin, C.; Bäckström, T.;
Panagiotidis, G.; Ehrnebo, M.; Nyström, C.; Aquilonius, S.-M. Clin.
Neuropharmacol. 2012, 35, 111−117.
(18) Rabey, J. M.; Vered, Y.; Shabtai, H.; Graff, E.; Korczyn, A. D. J.
Neurol., Neurosurg. Psychiatry 1992, 55, 725−727.
(19) Tsunoda, M.; Hirayama, M.; Tsuda, T.; Ohno, K. Clin. Chim.
Acta 2015, 442, 52−55.
(20) Kovacs, E. M. R.; Stegen, J. H. C. H.; Brouns, F. J. Appl. Physiol.
1998, 85, 709−715.
(21) Birkett, D. J.; Miners, J. O. Br. J. Clin. Pharmacol. 1991, 31,
405−408.
(22) Kintz, P.; Henrich, A.; Cirimele, V.; Ludes, B. J. Chromatogr.,
Biomed. Appl. 1998, 705, 357−361.
(23) Movlaee, K.; Beitollahi, H.; Ganjali, M. R.; Norouzi, P.
Microchim. Acta 2017, 184, 3281−3289.
(24) Yue, H. Y.; Zhang, H.; Huang, S.; Lin, X. Y.; Gao, X.; Chang, J.;
Yao, L. H.; Guo, E. J. Biosens. Bioelectron. 2017, 89, 592−597.
(25) Thieme, D.; Rautenberg, C.; Grosse, J.; Schoenfelder, M. Drug
Test. Anal. 2013, 5, 819−825.
(26) Cone, E. J.; Hillsgrove, M. J.; Jenkins, A. J.; Keenan, R. M.;
Darwin, W. D. J. Anal. Toxicol. 1994, 18, 298−305.
(27) Koh, A.; Kang, D.; Xue, Y.; Lee, S.; Pielak, R. M.; Kim, J.;
Hwang, T.; Min, S.; Banks, A.; Bastien, P.; Manco, M. C.; Wang, L.;
Ammann, K. R.; Jang, K.-I.; Won, P.; Han, S.; Ghaffari, R.; Paik, U.;
Slepian, M. J.; Balooch, G.; Huang, Y.; Rogers, J. A. Sci. Transl. Med.
2016, 8, 366ra165.
(28) Matzeu, G.; Fay, C.; Vaillant, A.; Coyle, S.; Diamond, D. IEEE
Trans. Biomed. Eng. 2016, 63, 1672−1680.
(29) Tai, L.-C.; Gao, W.; Chao, M.; Bariya, M.; Ngo, Q. P.; Shahpar,
Z.; Nyein, H. Y. Y.; Park, H.; Sun, J.; Jung, Y.; Wu, E.; Fahad, H. M.;

6351 DOI: 10.1021/acs.nanolett.9b02478

Nano Lett. 2019, 19, 6346−6351