micromachines

Article Interfacing Digital Microfluidics with Ambient Using SU-8 as Layer

Gowtham Sathyanarayanan , Markus Haapala and Tiina Sikanen * Drug Research Program, Faculty of Pharmacy, Division of Pharmaceutical Chemistry and Technology, University of Helsinki, Viikinkaari 5E, 00790 Helsinki, Finland; gowtham.sathyanarayanan@helsinki.fi (G.S.); markus.haapala@helsinki.fi (M.H.) * Correspondence: tiina.sikanen@helsinki.fi; Tel.: +358-2941-59173

 Received: 20 November 2018; Accepted: 6 December 2018; Published: 8 December 2018 

Abstract: This work describes the interfacing of -on-dielectric based digital microfluidic (DMF) sample preparation devices with ambient mass spectrometry (MS) via desorption atmospheric pressure photoionization (DAPPI). The DMF droplet manipulation technique was adopted to facilitate drug distribution and metabolism assays in droplet scale, while ambient mass spectrometry (MS) was exploited for the analysis of dried samples directly on the surface of the DMF device. Although ambient MS is well-established for bio- and forensic analyses directly on surfaces, its interfacing with DMF is scarce and requires careful optimization of the surface-sensitive processes, such as sample precipitation and the subsequent desorption/ionization. These technical challenges were addressed and resolved in this study by making use of the high mechanical, thermal, and chemical stability of SU-8. In our design, SU-8 served as the dielectric layer for DMF as well as the substrate material for DAPPI-MS. The feasibility of SU-8 based DMF devices for DAPPI-MS was demonstrated in the analysis of selected pharmaceuticals following on-chip liquid-liquid extraction or an enzymatic dealkylation reaction. The lower limits of detection were in the range of 1–10 pmol per droplet (0.25–1.0 µg/mL) for all pharmaceuticals tested.

Keywords: digital microfluidics; SU-8; desorption/ionization; ambient mass spectrometry; liquid-liquid extraction; drug distribution; drug metabolism; cytochrome P450; microreactor; desorption atmospheric pressure photoionization

1. Introduction Digital microfluidics (DMF), based on electrowetting-on-dielectric, is an established technology for performing automated chemical and biochemical assays in droplet scale [1]. In DMF, droplets of reagents (in the range of nL–µL) are manipulated by applying a series of to an array of electrodes coated with an insulating (dielectric) layer, and further with an additional hydrophobic (e.g., a fluoropolymer) layer to reduce droplet sticking [2]. Application of to individual electrodes results in charge accumulation on the , thus allowing precise control of droplet movement (incl. merging, mixing, and splitting) based on changes in the surface’s electrowetting properties [3]. A variety of materials have been introduced as the dielectric layer in DMF [4], of which parylene C is by far the most widely used [5]. Although chemical vapor deposition (CVD) of parylene C results in a homogenous and defect-free layer, CVD is time-consuming and requires relatively expensive instrumentation [6]. Therefore, spin-coatable and UV curable photoresists, such as the epoxy polymer SU-8, appear as appealing alternatives to rapid and low-cost DMF microdevice fabrication, facilitating chip manufacturing even under standard laboratory conditions [7]. Owing to its good mechanical rigidity and wide thickness range, SU-8 has been the material of choice for a variety of laboratory-on-a-chip applications requiring high aspect ratio structures or ultimately sharp features,

Micromachines 2018, 9, 649; doi:10.3390/mi9120649 www.mdpi.com/journal/micromachines Micromachines 2018, 9, 649 2 of 15 such as on-chip electrospray emitters for mass spectrometric (MS) detection [8–10]. With a view to DMF applications, the strong adhesion of SU-8 [11] facilitates fabrication of high quality dielectric layers by spin-coating. On the other hand, the high internal film stresses characteristic to SU-8 [11], do not play a major role in DMF, as the thickness of the dielectric layer is typically in the range of few micrometers only. Although the relatively high autofluorescence of SU-8 [9] may not favor its use in optically detected DMF assays, the possibility of lithographic patterning of the dielectric layer provides a convenient approach to incorporate, e.g., integrated metal elements (underneath the dielectric layer) into the active assay design. For example, to facilitate voltammetric analysis of the sample droplet, on-chip electrodes patterned below the dielectric have been exposed via lithographic patterning of SU-8 [12]. The good thermal stability of SU-8 has also allowed its use as the dielectric layer for high temperature polymerase chain reaction on a thermal DMF device [13]. In addition to its favorable mechanical and thermal properties, the good chemical stability of SU-8 [14] favor its use over other dielectric materials under harsh experimental conditions, e.g., when the dielectric needs to tolerate high (>100 ◦C) temperatures in the presence of organic solvents. An experimentally challenged approach like this is, for example, ambient mass spectrometry (MS), which would facilitate direct analysis of sample components from the surfaces of the DMF bottom plate. Ambient MS is an analytical concept increasingly used for mapping the spatial distribution of metabolic biomarkers directly from the surfaces of biological substrates, such as human or animal tissue [15] or plant leaves [16]. The technology also allows for forensic and drug analysis directly, without additional sample preparation, from bodily fluids, fabrics, tablets, and fruit, for example [17,18]. A large variety of desorption/ionization methods have been developed to facilitate ambient MS, of which desorption (DESI) and direct analysis in real time (DART) are probably the most widely used techniques in biochemical and forensic analyses, respectively. By facilitating non-targeted analysis of the assay components and reaction products in situ, ambient MS also appears as a logical, supplementary tool for monitoring of DMF-based assays in situ. However, the combination has not seen widespread exploitation among the variety of optically and electrochemically detected DMF assays, likely because it requires controlled evaporation of the sample solution prior to detection. Instead, interfacing of DMF with MS has mainly been realized via electrospray ionization (ESI) using integrated off-chip or on-chip emitters [19,20] or by transferring the sample solution to off-chip liquid chromatography (LC) MS analysis [21]. A few prior reports also exist on matrix-assisted laser desorption ionization (MALDI) MS from the DMF bottom plate [22,23]. All of the above-mentioned concepts, however, have their own limitations. For example, integration of an ESI emitter with the DMF device is technically demanding, same as the transfer of the very small volume droplet for analysis on a macro LC-MS instrument. MALDI-MS on the other hand requires substantial amount of manual sample preparation prior to analysis and is mostly feasible for large biomolecules only. In this work, we examined the critical steps related to interfacing of DMF based biochemical assays, more precisely drug distribution (liquid-liquid extraction) and metabolism assays, with ambient MS. First of all, we chose to use desorption atmospheric photoionization (DAPPI) [24] for desorption/ionization on DMF devices. Even if DESI is the most used technique for ambient MS based bioanalysis in general, it suffers from pronounced signal suppression (loss of sensitivity) in the presence of high salt concentrations, characteristic to biochemical assays typically conducted by DMF. In terms of drug analysis, DAPPI not only provides better salt tolerance, but also enables more efficient ionization of non-polar compounds (such as most of the drugs) over DESI [25,26]. Here, we show that by exposing SU-8 spots to the assay design, to allow for controlled sample precipitation on a hydrophilic surface, the efficiency of ambient MS analysis is significantly improved over that performed on the hydrophobic (fluoropolymer) layer. Moreover, the high chemical, thermal, and mechanical stability of SU-8 are in a key role in enabling desorption/ionization directly from the SU-8 surface. Micromachines 2018, 9, 649 3 of 15

2. Materials and Methods

2.1. Materials and Reagents Pentaerythritol tetrakis (3-mercaptopropionate), >95% (tetra-thiol), 1,3,5-Triallyl-1,3,5-triazine- 2,4,6(1H,3H,5H)-trione (tri-ene), 98%, β-nicotinamide adenine dinucleotide 20-phosphate reduced tetrasodium salt hydrate (NADPH), biotin-PEG4-alkyne, rhodamine B, magnesium chloride, paracetamol, naproxen, carbamazepine, phenacetin, testosterone, octanol, ethylene glycol, chlorobenzene, 28–30% ammonium hydroxide, 2-Amino-2-(hydroxymethyl)-1,3-propanediol (Tris), phosohate buffer saline (PBS) and chloroform were purchased from Sigma-Aldrich (Steinheim, Germany). Potassium dihydrogen phosphate and trifluoroacetic acid (TFA) were from Riedel-de Haen (Seelze, Germany). Dipotassium hydrogen phosphate was from Amresco (Solon, OH, USA). Methanol was from Honeywell Specialty Chemicals GmbH (Seelze, Germany) and Irgacure® TPO-L (photoinitiator) was kindly donated by BASF (Ludwigshafen, Germany). Alexa Fluor® 488 streptavidin conjugate (Alexa-SA) was purchased from Thermo Fisher Scientific (Waltham, MA, USA). The lipids 1,2-dioleoyl-sn-glycero-3-phophoethanolamine, 1,2-dioleoyl-3-trimethylammonium-propane (chloride salt), 1,2-dioleoyl-sn-glycero-3-phophoethanolamine-N-(Cap biotinyl) (sodium salt), and 1,2-dioleoyl- sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (ammonium salt) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL, USA). Recombinant cytochrome P450 1A2 (rCYP1A2) (1000 pmol·mL−1) was purchased from Corning (Wiesbaden, Germany). Water was purified with a Milli-Q water purification system, (Merck Millipore, Molsheim, France). All solvents were of analytical or HPLC grade. SU-8 5 was purchased from Microchem Corp., (Westborough, MA, USA). Fluoropolymer Cytonix FluoroPel PFC 1604V and PFC 110 fluoro-solvent were purchased from Cytonix LLC (Beltsville, MD, USA). AZ 351B resist developer was purchased from MicroChemicals GmbH (Ulm, Germany). Ormocomp was purchased from Microresist technology (Berlin, Germany).

2.2. Digital Microfluidic Chip Fabrication The top- and bottom plates of DMF devices were fabricated and assembled as described previously [7]. Briefly, an array of 15 × 6 driving electrodes and 8 reservoir electrodes for sample loading were lithographically patterned on a Cr coated glass plate (Telic, Valencia, CA, USA) which served as the bottom plate of the DMF device (Figure1a). The masked UV exposure was performed on a Dymax 5000-EC Series UV flood lamp (Dymax corporation, Torrington, CT, USA; nominal intensity 225 mW/cm2) followed by resist development and Cr etching under non-clean room conditions. After electrode patterning, a 8-µm-thick SU-8 layer was spin-coated (1600 rpm, 30 s) onto the electrodes, soft baked according to the standard protocol (2 min at 65 ◦C and 2 min at 95 ◦C) and flood exposed for 15 s using the Dymax lamp. After post-exposure bake (2 min at 65 ◦C and 2 min at 95 ◦C), a 1-µm-thick fluoropolymer layer was spin-coated (1% solution, 1000 rpm, 30 s) on SU-8 and baked at 150 ◦C for 30 min. Hydrophilic spots were made by covering the bare SU-8 layer with circular pieces of dicing tape, cut with a biopsy puncher (I.D. 1 mm), prior to fluoropolymer coating. After fluoropolymer coating, the tape was removed and the SU-8 spots were washed with methanol-water (1:1) three times and dried under nitrogen stream before use. Commercial coated glass plates (Structure Probe, Inc., West Chester, PA, USA), spin-coated with fluoropolymer in the same way, were used as the top plates of the DMF devices. Droplet actuation on the DMF devices was conducted with the help of an open-source DropBot automation system [27]. For heating of the DMF based drug metabolism assays to physiological temperature (37 ◦C), ink-jet printed thin-film silver microheaters were used. The microheaters (R = 30.8 ± 0.4 Ω, n = 5) were printed with a Brother MFC-J5910DW printer out of a commercial silver ink (AGIC-AN01 Silver Nano Ink, AgIC Inc., Tokyo, Japan) attached to the DMF bottom plate and resistively heated as described earlier [7]. Micromachines 2018, 9, 649 4 of 15 Micromachines 2018, 9, x FOR PEER REVIEW 4 of 15

FigureFigure 1.1. (a(a) )Schematic Schematic view view of ofthe the fabrication fabrication process process of digital of digital microfluidic microfluidic (DMF) (DMF) bottom bottom plate plateincorporating incorporating hydrophilic hydrophilic spots, spots, (b) (ba) aphotograph photograph of of DMF DMF-desorption‐desorption atmospheric pressurepressure photoionizationphotoionization (DAPPI) interface, interface, and and (c) ( cschematic) schematic representation representation of the of process the process flow of flow DMF of‐ DMF-DAPPI-massDAPPI‐mass spectrometry spectrometry (MS) (MS) analysis. analysis. For details For details of the of heated the heated nebulizer nebulizer chip chip fabrication fabrication and anduse, use,see references see references [24,28], [24, 28respectively.], respectively. 2.3. Mass Spectrometry 2.3. Mass Spectrometry For DAPPI-MS analysis, the DMF devices were connected to a Bruker micrOTOF™ mass For DAPPI‐MS analysis, the DMF devices were connected to a Bruker micrOTOF™ mass spectrometer (Bruker Daltonics, Bremen, Germany). The DAPPI-MS setup includes a microfabricated spectrometer (Bruker Daltonics, Bremen, Germany). The DAPPI‐MS setup includes a microfabricated heated nebulizer chip for vaporizing the organic solvent jet needed for desorption, and a krypton heated nebulizer chip for vaporizing the organic solvent jet needed for desorption, and a krypton filled vacuum UV lamp (PKR 106; Heraeus Noblelight GmbH, Hanau, Germany) for photoionization filled vacuum UV lamp (PKR 106; Heraeus Noblelight GmbH, Hanau, Germany) for photoionization (Figure1b). Schematic views of the DAPPI-MS setup and the heated nebulizer chip are given in the (Figure 1b). Schematic views of the DAPPI‐MS setup and the heated nebulizer chip are given in the Supplementary Materials (Figure S1). The heated nebulizer chip fabrication as well as optimization of supplementary material (Figure S1). The heated nebulizer chip fabrication as well as optimization of the key operation parameters related to DAPPI-MS analysis are described elsewhere [24,26]. Briefly, the key operation parameters related to DAPPI‐MS analysis are described elsewhere [24,26]. Briefly, in DAPPI-MS, the desorption/ionization process begins with thermal desorption (vaporization) of in DAPPI‐MS, the desorption/ionization process begins with thermal desorption (vaporization) of solid sample into the gas phase with help of a heated jet of vaporized organic solvent (Figure1c). solid sample into the gas phase with help of a heated jet of vaporized organic solvent (Figure 1c). The The gas phase solvent molecules are ionized with the help of photons (10 eV), after which they induce gas phase solvent molecules are ionized with the help of photons (10 eV), after which they induce ionization of the sample components via gas phase charge exchange or proton transfer reactions. ionization of the sample components via gas phase charge exchange or proton transfer reactions. As As the result, the sample components are detected on the MS as radical cations (M+•) or protonated the result, the sample components are detected on the MS as radical cations (M+•) or protonated ions ions [M + H]+, respectively, depending on their proton affinity and ionization energy. In this work, [M + H]+, respectively, depending on their proton affinity and ionization energy. In this work, sample sample desorption/ionization from the DMF bottom plate was performed using chlorobenzene (250 ◦C, desorption/ionization from the DMF bottom plate was performed using chlorobenzene (250 °C, 10 10 µL/min) as the organic solvent, vaporized with help of the nebulizer chip using heating power of μL/min) as the organic solvent, vaporized with help of the nebulizer chip using heating power of 4 W. In addition, nitrogen (180 μL/min) was used as an assisting (nebulizer) gas. The nebulizer chip

Micromachines 2018, 9, 649 5 of 15

MicromachinesMicromachinesMicromachines 2018 2018 2018, 9,, 9x,, 9FORx, FORx FOR PEER PEER PEER REVIEW REVIEW REVIEW 5 of5 of5 15 of 15 15 4 W. In addition, nitrogen (180 µL/min) was used as an assisting (nebulizer) gas. The nebulizer chip waswaswas fixed fixed fixed fixed at at an at an anangle angle angle of of 45° 45of 45°◦ 45° withwith with with respect respect respect respect to to the to the the sample sample sample (DMF) (DMF) (DMF) plate plate plate (Figure (Figure (Figure S1). S1). S1). On On On the the the MS, MS, MS, the the the drying drying drying drying gasgasgas temperature temperature temperature was was was set set set to to 300to 300 300 ◦°C C°C °Cand and and the the the flow flow flow flow rate rate rate to to 4to 4 L/min. 4L/min. L/min. AfterAfterAfter performing performing performing the the the desired desired desired sample sample sample manipulations manipulations manipulations on on onthe the the DMF DMF DMF device, device, device, the the the droplets droplets droplets were were were transferredtransferredtransferred onto onto onto the the the hydrophilichydro hydrophilic hydrophilicphilic spots spots spots by by byDMF DMF DMF actuation. actuation. actuation. The The The top top top top plate plate plate was was was then then then removed removed removed and and and the the the dropletsdropletsdroplets werewere were were letlet let dry letdry dry dry prior prior prior prior to to DAPPI-MSto DAPPIto DAPPI DAPPI‐MS‐MS analysis‐MS analysis analysis analysis (Figure (Figure (Figure (Figure1c). 1c). The 1c). 1c). The sample The The sample sample sample spots spots werespots spots were then were were alignedthen then then aligned aligned withaligned withrespectwithwith respect respect torespect the to heatedto theto the the heated heated solvent heated solvent solvent jet,solvent the jet, jet, UV jet, the the lamp,the UV UV UV lamp, andlamp, lamp, the and and MSand the the inlet the MS MS MS withinlet inlet inlet helpwith with with ofhelp help an help of xyz-positioning of anof an anxyz xyz xyz‐positioning‐positioning‐positioning stage (Figurestagestagestage (Figure(F (Figure1igure b,c).(Figure The1b,c). 1b,c). 1b,c). distance The The The distance distance betweendistance between between thebetween sample the the the sample sample spot sample and spot spot spot the and and MS and the inletthe the MS MS wasMS inlet inlet inlet ca. was was 2 was mm ca. ca. ca.2 and 2mm 2mm themm and and UV and the the lamp the UV UV UV waslamplamplamp fixed was waswas ca. fixed 10 fixedfixed mm ca. aboveca. ca. 10 10 10 the mm mmmm sample above aboveabove spot. the thethe The sample sample DAPPI-MSsample spot. spot.spot. method The TheThe DAPPI qualificationDAPPIDAPPI‐MS‐MS‐MS method (determinationmethodmethod qualification qualificationqualification of the (determinationlower(determination(determination limits of detection)of of theof the the lower lower lower was limits limits conducted limits of of detection)of detection) detection) with help was was was of conducted threeconducted conducted pharmaceuticals, with with with help help help of of three includingof three three pharmaceuticals, pharmaceuticals, pharmaceuticals, paracetamol, includingnaproxen,includingincluding andparacetamol, paracetamol, paracetamol, carbamazepine, naproxen, naproxen, naproxen, using and testosteroneand and carbamazepine,carbam carbamazepine, carbamazepine,azepine, as an internal using using using qualifiertestosterone testosterone testosterone (Table as 1as ).asan Foran aninternal internal optimization internal qualifier qualifier qualifier of (Tablethe(Table(Table MS 1). parameters, 1). 1).For For For optimization optimization optimization dummy of glass of theof the the plates MS MS MS parameters, withoutparameters, parameters, the dummy electrodesdummy dummy glass glass andglass plates patternedplates plates without without without with the SU-8the the electrodes electrodes (hydrophilicelectrodes and and and patternedspots)patternedpatterned and with fluoropolymer with with SU SU SU‐8‐ 8‐(hydrophilic 8(hydrophilic (hydrophilic in the same spots) spots) manner, spots) and and and were fluoropolymer fluoropolymer fluoropolymer used as substrates in in inthe the the onto same same same which manner, manner, manner, the sample were were were dropletsused used used as as as substrates(1substratesµsubstratesL) were onto manuallyonto onto which which which dispensed. the the the sample sample sample The droplets MSdroplets droplets data (1 was (1 μ(1 μL) μ analyzedL) L)were were were manually withmanually manually Compass dispensed. dispensed. dispensed. DataAnalysis The The The MS MS 4.0MS dadata softwaredata tadata was was was analyzed(Brukeranalyzedanalyzed Daltonics, with with with Compass Compass Compass Bremen, DataAnalysis DataAnalysis Germany). DataAnalysis For 4.0 4.0 4.0 datasoftware software software analysis (Bruker (Bruker (Bruker (peak Daltonics, areaDaltonics, Daltonics, integration), Bremen, Bremen, Bremen, extraction Germany). Germany). Germany). window For For For data data ofdata analysis±analysis0.05analysism /(peak z(peak was(peak area used.area area integration), integration), integration), extraction extraction extraction window window window of of ±0.05of ±0.05 ±0.05 m m/ zm/ zwas/ zwas was used. used. used.

TableTableTable 1.1. 1. The1. TheThe The chemical chemicalchemical chemical structures, structures, structures, structures, monoisotopic monoisotopicmonoisotopic monoisotopic masses massesmasses masses and andand and chemical chemical chemical chemical properties properties properties properties of of of ofthe the themodel model model pharmaceuticalspharmaceuticalspharmaceuticals used.used. used. used. MolecularMolecular Molecular Molecular propertiesproperties properties properties from from from from Chemicalize Chemicalize Chemicalize Chemicalize ( https://chemicalize.com/(https://chemicalize.com/ (https://chemicalize.com/ (https://chemicalize.com/)) developed) developed developed) developed byby ChemAxonby ChemAxon ChemAxon Ltd. Ltd. Ltd. (Budapest, (Budapest, (Budapest, Hungary). Hungary). Hungary).

CarbamazepineCarbamazepineCarbamazepine TestosteroneTestosteroneTestosteroneTestosterone NaproxenNaproxenNaproxenNaproxen Carbamazepine ParacetamolParacetamolParacetamol Naproxen

CompoundCompoundCompoundCompound

MolecularMolecularMolecular Molecular CC8HC8H98NOH9NO9NO2 2 2 CC14HC14H1414HO14O143 O3 3 CC15HC15H1512HN12N122ON2O 2O CC19HC19H1928HO28O282 O2 2 CC8HH9NONO2 CC14HH14OO3 CC15H12NN2OOC C19H28OO2 FormulaFormulaFormulaFormula 8 9 2 14 14 3 15 12 2 19 28 2 ObservedObservedObserved Ion Ion Ion 152.071152.071152.071 230.095230.095230.095 237.102237.102237.102 289.216289.216289.216 Observed Ion 152.071 230.095 237.102 289.216 TypeTypeType ([M([M([M + +H] +H] +H]) +) +) (M(M+•(M)+• )+• ) ([M([M([M + +H] +H] +H]) +) +) ([M([M([M + +H] +H] +H]) +) +) Type ([M([M + +H] H]) +) (M(M+)• ) ([M([M ++ H]+)) ([M([M + H] +)) MonoisotopicMonoisotopicMonoisotopic Monoisotopic 151.063151.063151.063 230.094230.094230.094 236.095236.095236.095 288.209288.209288.209 MassMassMass (g/mol) (g/mol) (g/mol) 151.063 230.094 236.095 288.209 Mass (g/mol) pKpKap (acidic)Ka (acidic)a (acidic) 9.469.469.46 4.194.194.19 15.9615.9615.96 n.n.d.n.d.d.n.d. LogDpKLogDLogDa (acidic) (pH (pH (pH 2) 2) 2) 0.99.460.9 0.9 2.98 4.192.982.98 15.962.762.762.76 LogD (pH 2) 0.9 2.98 2.76 (internal(internal(internaln.d. qualifier) qualifier) qualifier) LogD (pH 12) −1.15 −0.54 2.76 (internal qualifier) LogDLogDLogD (pH (pH 12) 2) 12) −1.15−1.150.9 −0.54−2.980.54 2.762.76 (internal qualifier) LogD (pH 12) −1.15 −0.54 2.76 2.4.2.4.2.4. On On On‐Chip‐Chip‐Chip Drug Drug Drug Distribution Distribution Distribution Assays Assays Assays 2.4. On-Chip Drug Distribution Assays TheTheThe on on ‐onchip‐chip‐chip drug drug drug distribution distribution distribution assays assays assays were were were based based based on on onthree three three‐phase‐phase‐phase liquid liquid liquid‐liquid‐liquid‐liquid extraction extraction extraction (LLE) (LLE) (LLE) protocolprotocolprotocolThe as on-chipas asdescribed described described drug in distributionin inFigure Figure Figure 2a. 2a. assays 2a.For For For on wereon ‐onchip‐chip‐chip based LLE, LLE, LLE, onthe the the three-phasesample sample sample solution solution solution liquid-liquid (containing (containing (containing extraction three three three pharmaceuticals(LLE)pharmaceuticalspharmaceuticals protocol as and describedand and testosterone testosterone testosterone in Figure as as asan 2an a. aninternal internal For internal on-chip qualifier, qualifier, qualifier, LLE, each each the each 45 sample 45 μ45 μg/mL) μg/mL)g/mL) solution was was was acidified acidified (containing acidified with with with three 0.1% 0.1% 0.1% TFApharmaceuticalsTFATFAH (p(pH (pH (pH 2), 2), 2),after after andafter which testosteronewhich which the the the aqueous asaqueous aqueous an internal sample sample sample qualifier, droplet droplet droplet each (donor) (donor) 45(donor)µg/mL) was was was wasencapsulated encapsulated encapsulated acidified withwith with with 0.1% octanol octanol octanol TFA (acceptor).(pH(acceptor).(acceptor). 2), after The The which The first first first theextraction extraction extraction aqueous (water (water sample (water → → droplet→octanol octanol octanol (donor)distribution) distribution) distribution) was encapsulated was was was carried carried carried for with for for 30 30 octanol 30min min min to to (acceptor).allowto allow allow the the the extractionTheextractionextraction first extraction between between between the (water the the donor donor donor→ (volume octanol(volume (volume equal distribution) equal equal to to twoto two two electrodes) waselectrodes) electrodes) carried and and forand the the 30 the acceptor minacceptor acceptor to allow(volume (volume (volume the equal extractionequal equal to to sixto six six electrodes).betweenelectrodes).electrodes). the The donor The The process process process (volume (step (step (step equal 2 2in 2in Figurein to Figure Figure two 2a2a) electrodes) 2a)) 2a) was was was then then then and repeated repeated repeated the acceptor three three three times (volume times times by by bybringing equal bringing bringing to a six afresh afresh electrodes). fresh droplet droplet droplet ofTheof ofsample processsample sample solution (stepsolution solution 2 in (donor) Figure(donor) (donor) 2intoa) into wasinto contact contact thencontact repeated with with with the the threethe same same same times octanol octanol octanol by bringing droplet droplet droplet a (acceptor). fresh (acceptor). (acceptor). droplet The The ofThe samplesecond second second extractionsolutionextractionextraction (donor) (octanol (octanol (octanol into → → contact water)→ water) water) withwas was was carried the carried carried same out out octanol out by by bybringing bringing bringing droplet a a (acceptor).droplet adroplet droplet of of ammoniumof The ammonium ammonium second hydroxide extraction hydroxide hydroxide (octanol(pH (pH (pH 12) 12) 12) (acceptor,→(acceptor,(acceptor,water) was1 1electrode 1electrode carriedelectrode volume) out volume) volume) by bringing inside inside inside the a the droplet the octanol octanol octanol of (donor). ammonium (donor). (donor). The The The hydroxideextraextraction extraction extractionction (pHtime time time 12) was was (acceptor,was again again again 30 30 1 min,30 electrode min, min, after after after whichvolume)whichwhich the the insidethe aqueous aqueous aqueous the droplet octanol droplet droplet was (donor). was was separated separated separated The extraction from from from octanol, octanol, octanol, time brought was brought brought again onto onto onto 30 the min, the the hydrophilic hydrophilic after hydrophilic which SU SU the SU‐8‐ 8spot aqueous‐ 8spot spot and and and alloweddropletallowedallowed was to to evaporateto separatedevaporate evaporate prior from prior prior octanol,to to DAPPIto DAPPI DAPPI brought‐MS‐MS‐MS analysis. analysis. onto analysis. the During hydrophilicDuring During both both both SU-8extractions, extractions, extractions, spot and the the allowed the octanol octanol octanol to phase evaporatephase phase was was was stationarypriorstationarystationary to DAPPI-MS and and and only only only analysis.the the the aqueous aqueous aqueous During droplets droplets droplets both extractions,were were were brought brought brought the in octanolin andin and and out out phase out of of theof the was the octanoloctan octanol stationaryoctanolol droplet. droplet. droplet. and This onlyThis This was thewas was facilitatedaqueousfacilitatedfacilitated droplets by by bythe the the werefact fact fact that broughtthat that octanol octanol octanol in and was was was out immobile immobile ofimmobile the octanol at at frequenciesat frequencies frequencies droplet. Thisused used used wasfor for for actuation facilitated actuation actuation of by of theof thethe the aqueous fact aqueous aqueous that phase.phase.phase. The The The driving driving driving potential potential potential used used used was was was 100–140 100–140 100–140 V Vrms Vrmsz rmsat at 10at 10 kH10kHz kHz kHz and and and 300 300 300 Hz Hz Hz for for for the the the aqueous aqueous aqueous phase phase phase and and and octanol,octanol,octanol, respectively. respectively. respectively.

Micromachines 2018, 9, 649 6 of 15 octanol was immobile at frequencies used for actuation of the aqueous phase. The driving potential usedMicromachines was 100–140 2018, 9, x V FORrms atPEER 10 REVIEW kHz and 300 Hz for the aqueous phase and octanol, respectively. 6 of 15

Figure 2. Schematic views of the main steps of (a) the on-chip drug distribution assays and (b) the Figure 2. Schematic views of the main steps of (a) the on‐chip drug distribution assays and (b) the on‐ on-chip drug metabolism assays. chip drug metabolism assays. 2.5. Preparation of Immobilized Cytochrome P450 Reactors 2.5. Preparation of Immobilized Cytochrome P450 Reactors For conducting drug metabolism reactions on the DMF devices, porous polymer monoliths functionalizedFor conducting with immobilized drug metabolism recombinant reactions cytochrome on the P450DMF (rCYP) devices, enzymes porous were polymer prepared monoliths similar tofunctionalized [7]. Briefly, the with porous immobilized polymer monoliths recombinant were preparedcytochrome out ofP450 thiol-ene (rCYP) polymer enzymes and were functionalized prepared sequentiallysimilar to [7]. with Briefly, biotin, the fluorescent porous polymer labeled streptavidinmonoliths were (Alexa-SA, prepared ex/em out 495/519of thiol‐ene nm), polymer biotinylated and fusiogenicfunctionalized liposomes sequentially (b-FL), with and rCYP.biotin, To fluorescent prepare the labeled porous streptavidin polymer monoliths, (Alexa‐SA, a 0.5 ex/em g mixture 495/519 of tetra-thiolnm), biotinylated and tri-ene fusiogenic monomers liposomes (50 mol% (b‐FL), excess and thiol) rCYP. was To stirredprepare along the porous with 5 polymerµL of 10% monoliths, Irgacure TPO-La 0.5 g inmixture methanol of tetra and‐thiol 2 g of and methanol tri‐ene monomers (porogen). (50 While mol% stirring, excess 2 thiol)µL of was the stirred emulsion along was with on a5 hydrophilicμL of 10% Irgacure spot (2 mmTPO dia.)‐L in patternedmethanol and onto 2 the g of bottom methanol plate (porogen). and cured While under stirring, UV (1.5 2 min μL of flood the exposure)emulsion was with on the a top-platehydrophilic attached. spot (2 Aftermm dia.) curing, patterned the monolith onto the plugs bottom were plate thoroughly and cured washed under withUV (1.5 a few min mL flood of MilliQ exposure) water. with The the monoliths top‐plate were attached. then functionalized After curing, withthe monolith biotin following plugs were the thoroughly washed with a few mL of MilliQ water. The monoliths were then functionalized with previously published protocol [29]. Briefly, 2 µL of 1 mM biotin-PEG4-alkyne in polyethylene glycol containingbiotin following 1% (v /thev) Irgacurepreviously TPO-L published was dispensed protocol [29]. onto Briefly, the monolith 2 μL of and 1 mM allowed biotin to‐PEG crosslink4‐alkyne with in thepolyethylene surface thiols glycol under containing UV for 1 1% min, (v after/v) Irgacure which the TPO monolith‐L was was dispensed sequentially onto washedthe monolith with a fewand mLallowed of methanol to crosslink and phosphate-buffered with the surface thiols saline under (PBS). UV Next, for 2 µ1L min, of 0.5 after mg/mL which Alexa-SA the monolith in PBS waswas pipettedsequentially on the washed monoliths with a and few incubated mL of methanol for 15 min and atphosphate room temperature.‐buffered saline After (PBS). washing Next, with 2 μ fewL of mL0.5 ofmg/mL PBS, 2.5 AlexaµL of‐SA b-FLs in PBS was was added pipetted on to the on monoliths, the monoliths incubated and atincubated room temperature for 15 min for at 30 room min, followedtemperature. by PBS After wash. washing Finally, with 2.5 µfewL of mL the of recombinant PBS, 2.5 μ CYP1A2L of b‐FLs isoform was added stock solution on to the was monoliths, pipetted onincubated the monoliths at room and temperature allowed to for fuse 30 with min, the followed b-FLs at by 37 PBS◦C forwash. 30 min. Finally, The 2.5 b-FLs μL of were the prepared recombinant and CYP1A2 isoform stock solution was pipetted on the monoliths and allowed to fuse with the b‐FLs at 37 °C for 30 min. The b‐FLs were prepared and fused with rCYP similar to the previous work [30]. After the fusion, the monoliths were washed with PBS and kept wet until use.

Micromachines 2018, 9, 649 7 of 15 fused with rCYP similar to the previous work [30]. After the fusion, the monoliths were washed with PBS and kept wet until use.

2.6. On-Chip Drug Metabolism Assays The on-chip drug metabolism reactions were performed on the monolith-based rCYP1A2 reactors using phenacetin O-deethylation as the model reaction (Figure2b). Before initiating the CYP reaction, the storage buffer (PBS) was replaced by DMF actuation with the reaction solution containing 100 µM phenacetin as the model substrate and 0.5 mM NADPH as the co-substrate in 15 mM Tris buffer (pH 7.5) containing 0.5 mM MgCl2. The reaction was initiated by applying heating power of 0.1 W via the printed microheater to reach the physiological temperature (37 ◦C). The initial reaction volume was ca. 3 µL (equal to area of four DMF electrodes) and the evaporation due to heating was compensated by adding 1.5 µL of fresh buffer to the reaction solution after 30 min (terminal point of the reaction). After the reaction, the droplet was separated from the monolith and brought onto the hydrophilic SU-8 spot. Next, the top plate was removed and the reaction solution was let dry prior to DAPPI-MS analysis. The specificity of the DAPPI-MS identification of phenacetin metabolites was confirmed with help of blank incubation matrix, i.e., control reactions performed in the absence of phenacetin.

3. Results and Discussion

3.1. Material Considerations The feasibility of SU-8 for low-cost chip fabrication by spin-coating and photolithographic patterning has raised considerable interest in its use as the dielectric material in DMF devices. SU-8 not only provides mechanically strong films with high thermal and chemical stability, but it also has notably high dielectric strength (4.4 MV/cm [31]), similar to that of parylene C (the most used dielectric in DMF). In this work, the DMF driving voltages typically ranged between 80 and 120 Vrms at 10 kHz (for actuation of aqueous solutions). Under these conditions, each chip could be used for several subsequent droplet manipulations and for long term (here, 30 min) continuous actuation, such as mixing of the drug distribution assays. In addition to high dielectric strength, the good uniformity and the high crosslinking degree of the cured SU-8 layer (film) are also in a critical role in terms of the robustness of SU-8 based DMF devices. Namely, when compared with another commercial negative tone photoresist, OrmoComp®, showing equally high dielectric strength [32], we observed that, while SU-8 based devices remained undamaged, the OrmoComp® based devices suffered from frequent dielectric breakdown between the electrodes. As the result of the electric discharge, discoloration of the dielectric OrmoComp layer was often visible even by eye as illustrated in Figure S2 (see brownish color at the edges of electrodes). The dielectric breakdown typically occurred in the proximity of the edges of the metal patterns, e.g., between the active electrode and its adjacent electrode (grounded), which were the most challenging spots of our assay design in terms of the layer quality. Even if both materials were applied (by spin-coating) and cured in the same way, the good processability of SU-8 was shown to result in superior performance over OrmoComp in electrowetting. Furthermore, SU-8 tolerated well the harsh desorption/ionization conditions characteristic to DAPPI-MS. Namely, in DAPPI-MS, sample desorption is induced by a heated (>200 ◦C) jet of vaporized organic solvent (Figure3a), which sets strict requirements for the stability of the substrate material. However, the heated solvent jet did not cause any visible defects or degradation of the SU-8 surface within the normal desorption time (1–2 min) as evidenced by the lack of interfering background ions in the MS spectra (Figure3b). Although the fluoropolymer surface gave equally low background interference (see Figure S3) as that of SU-8, the sample precipitation on the fluoropolymer posed substantial challenges in terms of repeatability as explicated below. Micromachines 2018, 9, 649 8 of 15 Micromachines 2018, 9, x FOR PEER REVIEW 8 of 15

FigureFigure 3.3. ((aa)) Schematic Schematic illustration illustration of of the the process process of ( ofa) (desorption/ionizationa) desorption/ionization of samples of samples from fromDMF DMFchip surface, chip surface, and integration and integration of peak of area peak of areaextracted of extracted ion chromatogram ion chromatogram (EIC) in (EIC)DAPPI in‐MS DAPPI-MS analysis. analysis.(b) Mass ( bspectrum) Mass spectrum of paracetamol of paracetamol (66 pmol), (66 pmol),naproxen naproxen (44 pmol), (44 pmol), carbamazepine carbamazepine (42 pmol), (42 pmol), and andtestosterone testosterone (70 (70pmol) pmol) dissolved dissolved in inmethanol methanol-water‐water 1:1, 1:1, deposited deposited onto onto the the hydrophilic hydrophilic SU SU-8‐8 spot, andand analyzed analyzed by by DAPPI-MS. DAPPI‐MS. ( c(c)) Photograph Photograph of of rhodamine rhodamine B B precipitation precipitation pattern pattern on on hydrophilic hydrophilic SU-8 SU‐ spots8 spots and and on on hydrophobic hydrophobic fluoropolymer fluoropolymer in in two two different different solvents: solvents: water water (top) (top) and and methanol-water methanol‐water 1:11:1 (bottom). (bottom). ((dd,,ee)) EICsEICs ofof testosterone testosterone (internal(internal qualifier,qualifier, 70 70 pmol pmol in in methanol-water methanol‐water 1:1) 1:1) deposited deposited andand analyzedanalyzed fromfrom parallelparallel spotsspots (n(n == 6)6) byby DAPPI-MS:DAPPI‐MS: ((dd)) onon hydrophilichydrophilic SU-8SU‐8 spots,spots, andand ((ee)) onon hydrophobichydrophobic fluoropolymer. fluoropolymer.

Micromachines 2018, 9, 649 9 of 15

One of the critical factors with a view to successful ambient MS analysis on artificial (not tissue or plant derived) surfaces is the uniformity of the sample precipitation pattern. Besides the compound’s solubility (to the evaporated solvent), the surface’s wetting properties play a key role in facilitating controlled precipitation of the sample solution [33]. In case of aqueous sample solutions, the high water of the fluoropolymer coating (top most layer of the DMF devices) results in a somewhat dense precipitation pattern because of decreasing droplet size upon evaporation (Figure3c). This sets limitations on the feasibility of the hydrophobic coating for desorption/ionization, because very small precipitate complicates the alignment of the heated solvent jet to the sample spot and thus results in reduced repeatability. Instead by exposing hydrophilic SU-8 spots into the DMF chip design, we were able to control the droplet size upon evaporation with help of , which resulted in a uniform and less dense precipitation pattern regardless of the solvent composition (Figure3c). In this study, we mainly compared the sample precipitation from fully aqueous solvent (needed for drug distribution and metabolism assays) and from water-methanol 1:1 solution (used for method validation). No significant difference was observed in the precipitation patterns between these two solvents. Instead, desorption/ionization efficiency and repeatability were shown to be much better on top of SU-8 (Figure3d) than on the fluoropolymer (Figure3e). This was likewise because of the homogenous precipitate and the fact that the edges of the SU-8 spots facilitated more precise visual alignment of the heated vapor jet to the sample spot. The hydrophilic spots also allowed us to ‘anchor’ the droplets to the bottom plate, when removing the top plate of the DMF device to allow for evaporation prior to ambient MS analysis. On the basis of these qualitative comparisons, SU-8 was concluded well feasible for fabrication of DMF microdevices even under non-cleanroom conditions as well as for multiple repeated use (of the hydrophilic SU-8 spots) in DAPPI-MS.

3.2. Mass Spectrometry Method Qualification The lower limits of detection (LLODs) of the three model compounds selected were determined with help of a series of calibration samples ranging between 0.1 and 10 µg/mL (each compound, in methanol-water 1:1) and were 1 µg/mL (7 pmol) for paracetamol, 0.25 µg/mL (1 pmol) for naproxen, and 0.5 µg/mL (2 pmol) for carbamazepine. Of these, paracetamol and carbamazepine (as well as the internal qualifier, testosterone) were detected as protonated [M + H]+ ions and naproxen as a radical M+• cation. The LLODs were calculated based on five replicates spots at each concentration. LLOD was defined as the lowest concentration that gave detectable signal (integrated peak) from all five spots. As such, the method was concluded feasible for, e.g., detection of drugs in bodily fluids, such as plasma or urine, which typically feature drug concentrations in the range of few µg/mL. However, it should be noted that DAPPI-MS is inherently not a quantitative analysis method, but best applicable for qualitative monitoring of in vivo administered drugs or identification of the in vitro reaction products (as will be shown in the next chapters). Minor variation in the ion current intensity between parallel spots as in Figure3e is also inherent to DAPPI-MS. For comparison, the repeatability of paracetamol (66 pmol) signal intensity on SU-8 and on polymethylmetacrylate (a common standard substrate for DAPPI-MS [26]) were 35% (RSD, n = 5 spots) and 23% (RSD, n = 5 spots), respectively. To evaluate the reuse possibility of the DMF chips, i.e., the cross-contamination risk between subsequent samples applied onto the same hydrophilic SU-8 spot, we also compared the background spectra and ion intensities between fresh (unused) and cleaned (reused) hydrophilic spots. Cleaning of the spots between experiments was carried out by rinsing the spots manually in methanol and water, which is our standard procedure for cleaning of the DMF chips between subsequent uses. As illustrated in Figure S4 in the Supplementary Materials, there were no traces of standards on the spots after washing and the intensities of the recorded signals were within the normal run-to-run variation before and after washing (Figure S4). Micromachines 2018, 9, x FOR PEER REVIEW 10 of 15

Micromachines 2018, 9, 649 10 of 15 3.3. On‐Chip DAPPI‐MS Analysis of Drug Distribution Assays Octanol‐water distribution, i.e., logD coefficient, is a routine tool for assessing the solubility of a drug3.3. On-Chip compound. DAPPI-MS Solubility Analysis is not of Drugonly Distributiona critical factor Assays to distinguish between continuation and terminationOctanol-water of the development distribution, of i.e., a new logD drug coefficient, candidate, is a routinebut it also tool serves for assessing as the basis the solubility for sample of purificationa drug compound. possibilities Solubility by liquid is not‐liquid only aextraction critical factor (LLE). to distinguishIn this work, between we demonstrated continuation andthe feasibilitytermination of the of the DMF development‐DAPPI‐MS ofcombination a new drug for candidate, preliminary but examination it also serves of asa drug’s the basis solubility for sample and distributionpurification possibilitiesproperties by by conducting liquid-liquid a extractionthree‐phase (LLE). (aqueous In this → work, organic we demonstrated → aqueous) theLLE feasibility on chip andof the monitoring DMF-DAPPI-MS the distribution combination forof preliminarythree model examination compounds of a drug’s(naproxen, solubility paracetamol, and distribution and carbamazepine)properties by conducting between athe three-phase organic solvent (aqueous and→ aqueousorganic phases.→ aqueous) The LLE onwas chip performed and monitoring in two stepsthe distribution to extract the of selected three model pharmaceuticals compounds first (naproxen, from aqueous paracetamol, (acidic) and to organic carbamazepine) (octanol) solution between andthe organicthen from solvent the andorganic aqueous solvent phases. back The to an LLE aqueous was performed (alkaline) in twosolution. steps Since to extract octanol the selectedcan be actuatedpharmaceuticals only at lower first from frequencies aqueous (300 (acidic) Hz) in to DMF organic device, (octanol) it was solution possible and to selectively then from actuate the organic the aqueoussolvent back droplets to an only aqueous by using (alkaline) higher solution. frequency Since (10 octanol kHz) canto bring be actuated them inside only at and lower outside frequencies of the larger,(300 Hz) immobile in DMF octanol device, droplet it was possible(Supplementary to selectively Video actuateS1). By theencapsulating aqueous droplets the aqueous only bydroplets using insidehigher octanol, frequency the (10assay kHz) was to also bring insensitive them inside to evaporation and outside effects of the owing larger, to immobile the very low octanol evaporation droplet rate(Supplementary of octanol. Video S1). By encapsulating the aqueous droplets inside octanol, the assay was also insensitiveAll selected to evaporation test compounds effects possessed owing to the acidic very nature low evaporation (pKa ranging rate between of octanol. 4.19 and 15.96, Table 1) andAll were selected thus test in compoundsneutral form possessed in the first acidic aqueous nature (pK donora ranging solution between (0.1% 4.19 TFA, and pH 15.96, 2), Tablewhich1) facilitatedand were thustheir inextraction neutral form to the in organic the first acceptor aqueous solution donor solution (octanol). (0.1% However, TFA, pH owing 2), which to their facilitated varying watertheir extractionsolubility and to the acidity organic (Table acceptor 1), the solution compounds’ (octanol). distribution However, from owing the organic to their phase varying and water back tosolubility the alkaline and aqueous acidity (Table phase1 ),varied the compounds’ according to distribution these parameters. from the Naproxen organic was phase the and most back acidic to compoundthe alkaline (pKa aqueous 4.19) phasewith moderate varied according lipophilicity to these (logD parameters. 2.98) at pH Naproxen 2 and low was lipophilicity the most (logD acidic −compound0.54) at pH (pKa 12, and 4.19) thus, with it moderatewas efficiently lipophilicity extracted (logD from 2.98) aqueous at pH to 2 andorganic low phase lipophilicity (in a neutral (logD form)−0.54) and at pH further 12, and from thus, organic it was to efficiently aqueous phase extracted (in an from ionized aqueous form). to organic Paracetamol phase on (in the a neutral other hand form) wasand more further water from soluble organic (logD to aqueous 0.91) at phase pH 2, (in which an ionized reduced form). its Paracetamolextraction efficiency on the other from hand the first was aqueousmore water donor soluble solution (logD to 0.91) the atoctanol pH 2, phase which compared reduced its with extraction naproxen efficiency as illustrated from the in first Figure aqueous 4a. Nevertheless,donor solution abundant to the octanol naproxen phase comparedand paracetamol with naproxen peaks as were illustrated detected in Figure by DAPPI4a. Nevertheless,‐MS after evaporationabundant naproxen of the final and paracetamolacceptor solution peaks were(Figure detected 4b). Instead by DAPPI-MS carbamazepine after evaporation as the least of the water final‐ solubleacceptor (logD solution 2.76) (Figure and only4b). Insteadweakly carbamazepine acidic (pKa 15.96) as the compound least water-soluble was not (logDback‐extracted 2.76) and from only octanolweakly to acidic the alkaline (pKa 15.96) aqueous compound phase and was was not thus back-extracted not detected from at all octanol in the tofinal the acceptor alkaline solution. aqueous Applyingphase and mixing was thus (by DMF not detected actuation at of all the in octanol the final phase acceptor at 300 solution. Hz) during Applying LLE did mixing not much (by affect DMF theactuation extraction of the yields octanol in phaseany of at the 300 cases Hz) during (data not LLE shown), did not muchbecause affect the the distribution extraction processes yields in any only of occurthe cases at the (data water not‐octanol shown), interface. because the distribution processes only occur at the water-octanol interface.

Figure 4. Cont.

Micromachines 2018, 9, 649 11 of 15 Micromachines 2018, 9, x FOR PEER REVIEW 11 of 15

FigureFigure 4. ((aa)) Comparison Comparison of of the the repeatability repeatability (n (n = = 3 3 LLE LLE experiments) experiments) of of intensities intensities of of extracted paracetamolparacetamol (initial (initial amount amount 300 300 pmol) pmol) and and naproxen naproxen (initial (initial amount amount 200 200 pmol) pmol) after after LLE LLE and and evaporationevaporation of of the the final final aqueous acceptor solution. ( (bb)) Mass Mass spectra spectra of of the the final final acceptor solution solution featuringfeaturing abundant abundant paracetamol paracetamol and and naproxen naproxen ions. ions. ( (cc)) Illustration Illustration of of monolith monolith functionalization functionalization and and phenacetinphenacetin-O-deethylation‐O‐deethylation reaction reaction and and ( (dd)) scanning scanning electron electron microscope microscope image image of of porous porous thiol thiol-ene‐ene monolith.monolith. ( (ee)) Comparison Comparison of of paracetamol paracetamol signal signal intensities intensities in in pure pure solvent solvent (methanol (methanol-water‐water 1:1) 1:1) vs. vs. 6060 mM mM Tris buffer (n = 5 samples each). ( (ff)) EICs EICs of of paracetamol (red (red line) line) produced via via CYP1A2 mediatedmediated phenacetin phenacetin (black (black line) line) O O-dealkylation‐dealkylation analyzed analyzed from from the the dried dried precipitate precipitate of of the the enzymatic enzymatic reactionreaction solution solution after after the the on on-chip‐chip drug metabolism assays.

3.4. On-Chip DAPPI-MS Analysis of Drug Metabolism Assays 3.4. On‐Chip DAPPI‐MS Analysis of Drug Metabolism Assays Drug metabolism assays are typically conducted at a physiological pH (7–7.5) in the presence Drug metabolism assays are typically conducted at a physiological pH (7–7.5) in the presence of of a relatively high salt concentration to preserve the enzymatic activity. The most commonly used a relatively high salt concentration to preserve the enzymatic activity. The most commonly used electrolytes in CYP enzyme assays include phosphate and Tris buffers. However, phosphate is electrolytes in CYP enzyme assays include phosphate and Tris buffers. However, phosphate is inherently not feasible for MS analysis, because of its nonvolatile nature. Upon sample evaporation, inherently not feasible for MS analysis, because of its nonvolatile nature. Upon sample evaporation, the use of a phosphate buffer also results in a very dense precipitate, which would impair ambient the use of a phosphate buffer also results in a very dense precipitate, which would impair ambient MS analysis independent of the chosen desorption/ionization method. Therefore, we chose to MS analysis independent of the chosen desorption/ionization method. Therefore, we chose to conduct the on-chip CYP assays by using the more volatile Tris (base) as the buffer component conduct the on‐chip CYP assays by using the more volatile Tris (base) as the buffer component and and determined its impact on the ionization of paracetamol (the model metabolite produced via determined its impact on the ionization of paracetamol (the model metabolite produced via CYP1A2 CYP1A2 induced phenacetin O-dealkylation). Although some variation between the average (n = 5 induced phenacetin O‐dealkylation). Although some variation between the average (n = 5 samples) paracetamol intensities were observed between pure solvent (methanol‐water 1:1) and the Tris buffer,

Micromachines 2018, 9, 649 12 of 15 samples) paracetamol intensities were observed between pure solvent (methanol-water 1:1) and the Tris buffer, these were within the normal spot-to-spot variation of DAPPI-MS (Figure4e). However, in Tris buffer, the protonated paracetamol ion [M + H]+ = 152.071 was dominated by the protonated Tris adduct ion [M + Tris + H]+ = 273.199. Nevertheless, the sensitivity of DAPPI-MS analysis directly from the DMF bottom plate was shown to be good enough to facilitate detection of the trace-level amounts (few pmol) of drug metabolites produced in the droplet based CYP reactions in situ, as illustrated in Figure4f. As seen in Figure4f, no metabolite was observed in blank control assay (with no substrate) whereas paracetamol Tris adduct ion peak was detected in all positive control assays.

4. Conclusions MS analysis of droplet-scale (~µL) biochemical reactions is often compromised by the sample loss associated with off-chip handling of small volumes, which may be required for, e.g., purification of sample from biological buffers or transfer of the sample onto a conventional LC-MS separation system. Ambient MS is therefore an appealing alternative to facile and rapid detection of droplet-based assays directly on the surfaces of DMF devices. Besides the fact that excessive sample purification can be omitted, the combination also benefits from the small sample volume, which allows rapid evaporation of the sample by simply removing the top plate of the DMF device prior to desorption/ionization. This work describes the feasibility of DAPPI-MS for instant analysis of sample components directly from the surface of a DMF based sample preparation device. With help of DAPPI-MS, the sample components can be identified in situ without need for integration of additional detector elements (e.g., ESI emitters) or transfer of the sample droplets for off-chip analysis. The successful combination of DMF and DAPPI-MS was promoted by the favorable mechanical, thermal, and chemical properties of SU-8, which served a bifunctional role, as dielectric material for electrowetting and as substrate material for desorption/ionization. Owing to their high dielectric and mechanical strength, SU-8 dielectric layers showed robust performance in DMF even upon prolonged continuous use. On the other hand, the hydrophilic nature of SU-8 (spots exposed to the assay design) allowed us to control the size and uniformity of the sample precipitate upon evaporation of sample droplet on top of the SU-8 spot. As the result, we were able to detect as low as pmol amounts of drugs per droplet (corresponding to less than µg/mL sample concentrations) by the DMF-DAPPI-MS setup, suggesting that the method is readily feasible for detection of, e.g., in vivo administered drug concentrations in bodily fluids. Most importantly, the good chemical and thermal stability of SU-8 facilitated multiple desorption/ionization cycles (repeated use) of each SU-8 spot without any visible damage of the substrate surface. In all, the straightforward integration of DMF with DAPPI-MS increases the flexibility of the analysis of DMF assays by facilitating identification of the sample components in situ directly on the DMF bottom plate. Moreover, the chip fabrication protocol developed herein can be carried out under standard laboratory conditions, which significantly reduces the fabrication costs and makes rapid prototyping of the DMF chips much easier compared with conventional parylene C based devices. In all, these technical advances may pave the way for adaptation of the droplet technology in preclinical drug discovery and development. Even if quantitation of the produced drug metabolites or drug distribution ratios is not feasible because of the qualitative nature of the DAPPI-MS analysis, the possibility to categorize new drug candidates according to their solubility using the droplet scale liquid-liquid extraction experiments could allow for a cost-efficient tool to distinguish between poorly and highly soluble drugs (a key parameter for continuation vs. termination of the development of new drug candidates). Alternatively, at a later stage of drug development, preliminary identification of the most abundant drug metabolites isoenzyme-specifically could allow for more efficient design of experiments and thus result in substantial savings in the analysis costs. Owing to the inherent possibility of DMF for parallelism of multiple individual reactions (each with different CYP isoenzyme), the metabolic profile of a new drug candidate could be preliminarily examined with help of the droplet reactions at much lower cost and reagent consumption compared with the conventional liquid Micromachines 2018, 9, 649 13 of 15 chromatography mass spectrometry protocols. Thereby, the developed DMF-DAPPI-MS setup could allow for competitive advantage to rapid, preliminary characterization of new drug candidates with a view to their metabolism and disposition before more in-depth studies.

Supplementary Materials: The following are available online at http://www.mdpi.com/2072-666X/9/12/649/ s1, Figure S1. Schematic illustrations of (a) DAPPI-MS analysis of samples precipitated onto the DMF bottom plate, and (b) the key functions of the heated nebulizer chip; Figure S2: Comparison of SU-8 and Ormocomp dielectric layers before and after use (in DMF, driving voltages (100–140 Vrms at 10 kHz); Figure S3: Extracted ion chromatograms corresponding to protonated testosterone ion ([M + H]+ = 289.216) measured from eight parallel SU-8 spots following (a) first application of 70 pmol testosterone in methanol-water (1:1) per spot (fresh plate), (b) washing of the spots with methanol and water (cleaned plate), and (c) re-application of 70 pmol testosterone (cleaned and reused plate); Figure S4: Extracted ion chromatograms corresponding to protonated testosterone ion ([M + H]+ = 289.216) measured from eight parallel SU-8 spots following (a) first application of 70 pmol testosterone in methanol-water (1:1) per spot (fresh plate), (b) washing of the spots with methanol and water (cleaned plate), and (c) re-application of 70 pmol testosterone (cleaned and reused plate); Video S1: Digital microfluidic actuation of an aqueous droplet into an octanol droplet in drug distribution assay. Author Contributions: Conceptualization and Methodology, G.S., M.H. and T.S.; Formal Analysis, G.S. and M.H.; Data Curation, G.S.; Writing—Original Draft Preparation, G.S.; Writing—Review and Editing, M.H. and T.S.; Visualization, G.S.; Supervision, Project Administration and Resources, T.S.; Funding Acquisition, T.S. and G.S. Funding: This research was funded by the EUROPEAN RESEARCH COUNCIL (ERC) under the European Union’s Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement number 311705 (CUMTAS), the ACADEMY OF FINLAND (grant numbers 304400 and 309608), the UNIVERSITY OF HELSINKI (UH) RESEARCH FOUNDATION and the CENTRE FOR INTERNATIONAL MOBILITY (CIMO), Finland. Acknowledgments: Anu Vaikkinen (Faculty of Pharmacy, UH) and the Electron Microscopy Unit (Institute of Biotechnology, UH) are thanked for help with the MS instrumentation and for the access to SEM, respectively. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Choi, K.; Ng, A.H.C.; Fobel, R.; Wheeler, A.R. Digital Microfluidics. Annu. Rev. Anal. Chem. 2012, 5, 413–440. [CrossRef][PubMed] 2. Pollack, M.G.; Shenderov, A.D.; Fair, R.B. Electrowetting-based actuation of droplets for integrated microfluidics. Lab Chip 2002, 2, 96–101. [CrossRef][PubMed] 3. Cho, S.K.; Moon, H.; Kim, C.-J. Creating, transporting, cutting, and merging liquid droplets by electrowetting-based actuation for digital microfluidic circuits. J. Microelectromech. Syst. 2003, 12, 70–80. 4. Liu, H.; Dharmatilleke, S.; Maurya, D.K.; Tay, A.A.O. Dielectric materials for electrowetting-on-dielectric actuation. Microsyst. Technol. 2009, 16, 449–460. [CrossRef] 5. Kahouli, A.; Sylvestre, A.; Ortega, L.; Jomni, F.; Yangui, B.; Maillard, M.; Berge, B.; Robert, J.-C.; Legrand, J. Structural and dielectric study of parylene C thin films. Appl. Phys. Lett. 2009, 94, 152901. [CrossRef] 6. Choy, K.L. Chemical vapour deposition of coatings. Prog. Mater. Sci. 2003, 48, 57–170. [CrossRef] 7. Sathyanarayanan, G.; Haapala, M.; Kiiski, I.; Sikanen, T. Digital microfluidic immobilized cytochrome P450 reactors with integrated inkjet-printed microheaters for droplet-based drug metabolism research. Anal. Bioanal. Chem. 2018, 410, 6677–6687. [CrossRef] 8. Arscott, S. SU-8 as a material for lab-on-a-chip-based mass spectrometry. Lab Chip 2014, 14, 3668–3689. [CrossRef] 9. Sikanen, T.; Heikkilä, L.; Tuomikoski, S.; Ketola, R.A.; Kostiainen, R.; Franssila, S.; Kotiaho, T. Performance of SU-8 Microchips as Separation Devices and Comparison with Glass Microchips. Anal. Chem. 2007, 79, 6255–6263. [CrossRef] 10. Nissilä, T.; Sainiemi, L.; Franssila, S.; Ketola, R.A. Fully polymeric integrated microreactor/electrospray ionization chip for on-chip digestion and mass spectrometric analysis. Sens. Actuators B Chem. 2009, 143, 414–420. [CrossRef] 11. Becker, H.; Gärtner, C. Polymer microfabrication technologies for microfluidic systems. Anal. Bioanal. Chem. 2008, 390, 89–111. [CrossRef][PubMed] 12. Dryden, M.D.M.; Rackus, D.D.G.; Shamsi, M.H.; Wheeler, A.R. Integrated Digital Microfluidic Platform for Voltammetric Analysis. Anal. Chem. 2013, 85, 8809–8816. [CrossRef][PubMed] Micromachines 2018, 9, 649 14 of 15

13. Chen, T.; Jia, Y.; Dong, C.; Gao, J.; Mak, P.-I.; Martins, R.P. Sub-7-second genotyping of single-nucleotide polymorphism by high-resolution melting curve analysis on a thermal digital microfluidic device. Lab Chip 2016, 16, 743–752. [CrossRef][PubMed] 14. Lorenz, H.; Despont, M.; Fahrni, N.; LaBianca, N.; Renaud, P.; Vettiger, P. SU-8: A low-cost negative resist for MEMS. J. Micromech. Microeng. 1997, 7, 121. [CrossRef] 15. Eberlin, L.S.; Ferreira, C.R.; Dill, A.L.; Ifa, D.R.; Cheng, L.; Cooks, R.G. Non-Destructive, Histologically Compatible Tissue Imaging by Desorption Electrospray Ionization Mass Spectrometry. Chembiochem. Eur. J. Chem. Biol. 2011, 12, 2129–2132. [CrossRef][PubMed] 16. Jackson, A.U.; Tata, A.; Wu, C.; Perry, R.H.; Haas, G.; West, L.; Graham Cooks, R. Direct analysis of Stevia leaves for diterpene glycosides by desorption electrospray ionization mass spectrometry. Analyst 2009, 134, 867–874. [CrossRef] 17. Morelato, M.; Beavis, A.; Kirkbride, P.; Roux, C. Forensic applications of desorption electrospray ionisation mass spectrometry (DESI-MS). Forensic Sci. Int. 2013, 226, 10–21. [CrossRef] 18. Pavlovich, M.J.; Musselman, B.; Hall, A.B. Direct analysis in real time—Mass spectrometry (DART-MS) in forensic and security applications. Mass Spectrom. Rev. 2018, 37, 171–187. [CrossRef] 19. Shih, S.C.C.; Yang, H.; Jebrail, M.J.; Fobel, R.; McIntosh, N.; Al-Dirbashi, O.Y.; Chakraborty, P.; Wheeler, A.R. Dried Blood Spot Analysis by Digital Microfluidics Coupled to Nanoelectrospray Ionization Mass Spectrometry. Anal. Chem. 2012, 84, 3731–3738. [CrossRef] 20. Kirby, A.E.; Wheeler, A.R. Microfluidic origami: A new device format for in-line reaction monitoring by nanoelectrospray ionization mass spectrometry. Lab Chip 2013, 13, 2533–2540. [CrossRef] 21. Choi, K.; Boyacı, E.; Kim, J.; Seale, B.; Barrera-Arbelaez, L.; Pawliszyn, J.; Wheeler, A.R. A digital microfluidic interface between solid-phase microextraction and liquid chromatography–mass spectrometry. J. Chromatogr. A 2016, 1444, 1–7. [CrossRef][PubMed] 22. Moon, H.; Wheeler, A.R.; Garrell, R.L.; Loo, J.A.; Kim, C.-J. An integrated digital microfluidic chip for multiplexed proteomic sample preparation and analysis by MALDI-MS. Lab Chip 2006, 6, 1213–1219. [CrossRef][PubMed] 23. Nichols, K.P.; Gardeniers, J.G.E. A Digital Microfluidic System for the Investigation of Pre-Steady-State Enzyme Kinetics Using Rapid Quenching with MALDI-TOF Mass Spectrometry. Anal. Chem. 2007, 79, 8699–8704. [CrossRef][PubMed] 24. Haapala, M.; Pól, J.; Saarela, V.; Arvola, V.; Kotiaho, T.; Ketola, R.A.; Franssila, S.; Kauppila, T.J.; Kostiainen, R. Desorption Atmospheric Pressure Photoionization. Anal. Chem. 2007, 79, 7867–7872. [CrossRef][PubMed] 25. Suni, N.M.; Lindfors, P.; Laine, O.; Östman, P.; Ojanperä, I.; Kotiaho, T.; Kauppila, T.J.; Kostiainen, R. Matrix effect in the analysis of drugs of abuse from urine with desorption atmospheric pressure photoionization-mass spectrometry (DAPPI-MS) and desorption electrospray ionization-mass spectrometry (DESI-MS). Anal. Chim. Acta 2011, 699, 73–80. [CrossRef][PubMed] 26. Luosujärvi, L.; Arvola, V.; Haapala, M.; Pól, J.; Saarela, V.; Franssila, S.; Kotiaho, T.; Kostiainen, R.; Kauppila, T.J. Desorption and Ionization Mechanisms in Desorption Atmospheric Pressure Photoionization. Anal. Chem. 2008, 80, 7460–7466. [CrossRef] 27. Fobel, R.; Fobel, C.; Wheeler, A.R. DropBot: An open-source digital microfluidic control system with precise control of electrostatic driving force and instantaneous drop velocity measurement. Appl. Phys. Lett. 2013, 102, 193513. [CrossRef] 28. Saarela, V.; Haapala, M.; Kostiainen, R.; Kotiaho, T.; Franssila, S. Glass microfabricated nebulizer chip for mass spectrometry. Lab Chip 2007, 7, 644–646. [CrossRef] 29. Feidenhans’l, N.A.; Lafleur, J.P.; Jensen, T.G.; Kutter, J.P. Surface functionalized thiol-ene waveguides for fluorescence biosensing in microfluidic devices. Electrophoresis 2014, 35, 282–288. [CrossRef] 30. Kiiski, I.; Pihlaja, T.; Urvas, L.; Wiedmer, S.; Jokinen, V.; Sikanen, T. Overcoming the Pitfalls of Cytochrome P450 Immobilization Through the Use of Fusogenic Liposomes. Adv. Biosyst. 2018, accepted. [CrossRef] 31. Melai, J.; Salm, C.; Smits, S.; Visschers, J.; Schmitz, J. The electrical conduction and dielectric strength of SU-8. J. Micromech. Microeng. 2009, 19, 065012. [CrossRef] Micromachines 2018, 9, 649 15 of 15

32. Fraunhofer Insititute for Silicate Research ISC. Available online: https://www.isc.fraunhofer.de/content/ dam/isc/de/documents/Publikationen/Dielectric_ORMOCERs_for_system_in_package_electronics.pdf (accessed on 19 November 2018). 33. Li, Y.-F.; Sheng, Y.-J.; Tsao, H.-K. Evaporation Stains: Suppressing the Coffee-Ring Effect by Contact Angle Hysteresis. Langmuir 2013, 29, 7802–7811. [CrossRef][PubMed]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).