ARTICLE IN PRESS

0969-8043/$ - se

doi:10.1016/j.ap

�CorrespondE-mail addr

Applied Radiation and Isotopes 64 (2006) 333–336

www.elsevier.com/locate/apradiso

Microfluidic technology for PET radiochemistry

J.M. Gilliesa,�, C. Prenanta,b, G.N. Chimona,b, G.J. Smethursta, B.A. Dekkera, J. Zweita,b

aCancer Research-UK/University of Manchester Radiochemical Targeting and Imaging Group, Paterson Institute for Cancer Research,

Manchester, M20 4BX, UKbSchool of Chemical Engineering and Analytical Sciences, University of Manchester, P.O. Box 88, Manchester, M60 1QD, UK

Received 8 December 2004; received in revised form 30 August 2005; accepted 30 August 2005

Abstract

This paper describes the first application of a microfabricated reaction system to positron emission tomography (PET) radiochemistry.

We have applied microfluidic technology to synthesise PET radiopharmaceuticals using 18F and 124I as labels for fluorodeoxyglucose

(FDG) and Annexin-V, respectively. These reactions involved established methods of nucleophilic substitution on a mannose triflate

precursor and direct iodination of the protein using iodogen as an oxidant. This has demonstrated a proof of principle of using

microfluidic technology to radiochemical reactions involving low and high molecular weight compounds. Using microfluidic reactions,

[18F]FDG was synthesised with a 50% incorporation of the available F-18 radioactivity in a very short time of 4 s. The radiolabelling

efficiency of 124I Annexin-V was 40% after 1min reaction time. Chromatographic analysis showed that such reaction yields are

comparable to conventional methods, but in a much shorter time. The yields can be further improved with more optimisation of the

microfluidic device itself and its fluid mixing profiles. This demonstrates the potential for this technology to have an impact on rapid and

simpler radiopharmaceutical synthesis using short and medium half-life radionuclides.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: PET radiochemistry; Radiosynthesis; Microfabrication; Microfluidics

1. Introduction

Positron emission tomography (PET) allows the study ofin vivo biochemistry and biology underlying disease andtherapeutic intervention (Gambhir, 2002, Massoud andGambhir, 2003, Reader and Zweit, 2001). This uniquecapability allows a rational assessment of, for example,anti-cancer drug development in early clinical trials.Current practice involves the manipulation of macroscopicquantities of material in the synthesis of various radio-pharmaceuticals. Nanotechnology, the miniaturisation ofmacroscale processes and devices, offers distinct advan-tages to PET radiochemistry. In particular, intrinsicreduction in resources and logistics is required for PETradiochemical preparations. Microfluidic technologies arecapable of controlling and transferring tiny quantities ofliquids which allow chemical and biochemical assays to be

e front matter r 2005 Elsevier Ltd. All rights reserved.

radiso.2005.08.009

ing author. Tel.: +44 161 446 3150; fax: +44 161 446 3109.

ess: jgillies@picr.man.ac.uk (J.M. Gillies).

integrated and carried out on a small scale. In the firstplace, radiochemical reactions on-chip can be easilyshielded and will not require the space and resourcesrequired for conventional hot cell synthesis. Secondly, itprovides a scope for an integrated total system (synthesis,separation and analysis). Thirdly, due to the efficient andrapid mixing in miniaturised reactors (Regenfuss et al.,1985), the speed of radiochemical synthesis and purifica-tion can be accelerated. Finally, the photolithographicfabrication of the microfabricated device allows themanufacture of complex, yet relatively inexpensive anddisposable devices (Mitchell, 2001; Ramsey, 1999). Wedemonstrate the rapid radiosynthesis of the PET metabolictracer 2-[18F]fluorodeoxyglucose (2-[18F]FDG) and theradioiodination of the apoptosis marker, [124I]Annexin-V(Dekker et al., 2005a, 2005b; Keen et al., 2005) using asimple microreactor. Radiolabelling using the microreactordemonstrated considerable improvements in speed ofreaction while using reduced reagent volumes and con-centrations in comparison to conventional radiosynthesis.

ARTICLE IN PRESSJ.M. Gillies et al. / Applied Radiation and Isotopes 64 (2006) 333–336334

Here, we show the first application of a microfabricatedreaction system to PET radiochemistry, we term ‘‘micro-fluidic PET’’.

2. Experimental

The design and fabrication of a simple microfluidicreactor, to generate adequate mixing and transfer ofreactants, is shown in Fig. 1 (Stuernstrom and Roeraade,1998; Lin et al. 2001; Tsai and Lin, 2001). The microreactorwas constructed from three layers of thermally bondedsoda-lime glass plates (15� 15� 1mm) using standardphotolithographic techniques. The microreactor disc(Fig. 1) was 10mm in diameter and 0.1mm deep. Themiddle and bottom plates were etched using 50% HFsolution to form two 10mm diameter� 100 mm deep discs,in which the reagents mixing would take place (Fig. 1). Thisgave a total internal volume of 16.0 mL. The top plate hadthree 1mm holes drilled to form the inlets. The centralmixing disc was connected via 0.1mm channels to the threeinlets. The inlets were connected, via fused silica capillaries,to external reagent reservoirs (PEEK HPLC loops, 200 mL)linked to a nitrogen gas manifold (51.4mLmin�1 flowrate).This generated a back pressure of 6.0� 104Nm�2 whichwas enough to drive the contents of the three reagentreservoirs through the microreactor at a flowrate of250 mLmin�1. The middle and bottom plates were thenaligned and a hole (1mm diameter hole� 5mm deep) wasdrilled horizontally into the interface between the middleand bottom plates, to act as placement for the outlet.

3. Discussion

To demonstrate ‘‘proof of principle’’, we have investi-gated the radioiodination of small and large molecularweight molecules using the microfluidic device. Thesereactions involved the direct radioiodination of theapoptosis marker Annexin-V, and the radiofluorinationof the PET radiotracer 2-[18F]FDG.

Top Plate

Fitting Inlets

Middle Plate

Vortex Mixer (0.2 mm diameter)

15 mm

3 mm

Bottom Plate

Fitting Outlet

Etched mixing discs(10 mm diameter,100µm deep)

Fig. 1. Three-dimensional diagram representing the construction of the

microreactor incorporating a three tier system of inlets, reactor and outlet.

Detection and imaging of apoptosis involves the highlyspecific binding of Annexin-V to phosphatidylserine (PS)that appears on the extracellular membrane of cellsundergoing apoptosis (Tokita et al., 2003). Direct methodsfor the radiolabelling of Annexin-V have been developedusing 124I-iodine (Dekker et al., 2005a, 2005b; Keen et al.,2005). This method was used in conjunction with themicrofabricated device. In the radioiodination of Annexin-V described here, reservoir 1 contained unlabelled Annex-in-V (10 mL in 200 mL PBS at pH 7.0). Reservoir 2contained 124I, 5 mL (20MBq) in 200 mL PBS at pH 7.0and reservoir 3 contained the oxidising agent iodogen(40 mg in 200 mL acetonitrile). The reaction on the micro-fabricated device was carried out by flushing all three of thereagent reservoirs (under a stream of nitrogen) through themicrofabricated device at the same time, at a total flowrateof 250 mLmin�1. The reaction product generated from thiscontinuous flow was sampled and analysed over a range oftime points (0–100 s) during the course of the reaction, untilall the reagent reservoirs were completely empty. Theunpurified reaction mixture appearing at the outlet of thechip was analysed by radioTLC at various time points(10–120 s) (Fig. 2) with a mobile phase 5% trichloroaceticacid and analysed using an Instant Imager electronicautoradiography system (Packard, USA). The labellingefficiency refers to the proportion of radioactivity that isincorporated into the compound as a fraction of the totalradioactivity used. After 2min, a 4075% labellingefficiency was obtained for both the microfluidic andconventional reactions (Fig. 2).

Fig. 2. On-chip radiolabelling of [124I]Annexin-V. Labelling efficiency

assessed over the first 120 s of the reaction. Samples were analysed using

radioTLC. This graph shows that both the conventional and microfluidic

reactions are almost instantaneous and reach the maximum labelling

efficiencies within 120 s. Labelling efficiency was determined as percentage

of radioactivity incorporated/total radioactivity used.

ARTICLE IN PRESS

Fig. 3. Schematic of the experimental set-up for the production of 2-

[18F]FDG. FDG was synthesised according to the method of Hamacher et

al. Fluorine-18 was produced from proton bombardment of [18O]H2O

target. In this reaction, two microfabricated devices were linked in

sequence using lengths of fused silica capillary connecting the outlet of

device 1 to inlet of device 2. Reagent reservoirs 1 and 2 connected to device

1 were primed with the following solutions, respectively, [18F]KF/K2.2.2./

K2CO3 in 200mL DMF and mannose triflate (25mg in 200mL DMF).

Inlet 3 of both devices were blocked off. A third reagent reservoir was

primed with sodium methanolate in MeOH and connected to inlet 2 of

device 2. The total volume of reactants passed through the chip within 6 s

of the start of reaction and were collected and analysed by radioTLC

(silica gel 60F254: 90% acetonitrile).

Fig. 4. RadioTLC analysis of the products generated from the multichip

synthesis of 2-[18F]FDG.

J.M. Gillies et al. / Applied Radiation and Isotopes 64 (2006) 333–336 335

The second PET tracer selected for production using themicrofabricated device was 2-[18F]FDG. The currentmethod of 2-[18F]FDG synthesis is based on the approachdeveloped by Hamacher et al. (1986). A modification of theHamacher reaction was carried out using two microfabri-cated devices linked together in sequence (Fig. 3). The firstmicrofabricated device was designed to carry out the[18F]fluorination of the protected mannose triflate pre-cursor. This was followed by deprotection on the secondmicrofabricated device using sodium methanolate inmethanol.

Device 1 produced a reaction between 18F(500MBq)/KF/Kryptofix.2.2.2/K2CO3 in N,N-dimethylformamide(DMF) and mannose triflate in DMF. This resulted inthe production of the 2-[18F]fluoro-tetra-O-acetyl-mannose(Fig. 3).

The 2-[18F]fluoro-tetra-O-acetyl-mannose was thenpumped onto the second chip where it mixed with asolution of 10% sodium methanolate in methanol resultingin the production of 2-[18F]FDG. Fifty percent of theavailable F-18 radioactivity was incorporated as 2-[18F]FDG within 6 s. This was determined by analysing afraction (1 mL) of the total volume collected at the outlet ofthe device. Fig. 4 shows a radiochromatogram of thisunpurified 2-[18F]FDG fraction. The amount of 2-[18F]FDG incorporated F-18 radioactivity was 450%,with approximately 20–30% of the radioactivity associated

with the unhydrolysed 2-[18F]fluoro-tetra-O-acetyl-man-nose and a further 10–20% was unreacted [18F]F�. Theunreacted [18F]F� is present due to the incomplete reactionwithin device 1 and 2-[18F]fluoro-tetra-O-acetyl-mannose ispresent on the radioTLC from the incomplete hydrolysis ofproducts produced in device 2. The radiochemical puritythat could be expected using this technology will exceed the495% specification stipulated by the European Pharma-copia.. As mentioned earlier, one advantage of microfluidicPET radiochemistry is the feasibility of an integratedsystem incorporating synthesis, purification and analysis.In this context, a miniaturised radioHPLC can beincorporated into the system for on-line analysis. For thepurpose of demonstrating proof of principle, the amount of18F used was only 500MBq. Since the method is based oncontinuous flow of reactants, much higher amounts (tens ofGBq) of 18F can be utilised in the reaction.Taken together, these preliminary results demonstrate

the feasibility of microfluidic radiochemistry, and thepreliminary data presented here indicate similar reactionyields of the two methods. Further studies based onmodelling and experimental validation are necessary foroptimisation of the device.

Acknowledgements

This work is funded by Cancer Research UK. Thanks toDr. Graham Cowling for proof reading the manuscript.

References

Dekker, B.A., Keen, H.G., Shaw, D., Disley, L., Hastings, D., Hadfield,

J., Reader, A., Allan, D., Julyan, P., Watson, A., Zweit, J., 2005a.

ARTICLE IN PRESSJ.M. Gillies et al. / Applied Radiation and Isotopes 64 (2006) 333–336336

Functional comparison of annexin V analogues labeled indirectly and

directly with iodine-124. Nucl. Med. Biol. 32, 403–413.

Dekker, B.A., Keen, H.G., Lyons, S., Disley, L., Hastings, H., Reader,

A.J., Ottewell, P., Watson, A., Zweit, J., 2005b. MBP–annexin V

radiolabeled directly with iodine-124 can be used to image apoptosis in

vivo using PET. Nucl. Med. Biol. 32, 241–252.

Gambhir, S.S., 2002. Molecular imaging of cancer with positron emission

tomography. Nat. Rev. Cancer 2, 683.

Hamacher, K., Coenen, H.H., Stocklin, G., 1986. Efficient stereospecific

synthesis of no-carrier-added 2-[18F]-fluoro-2-deoxy-D-glucose using

aminopolyether supported nucleophilic substitution. J. Nucl. Med. 27,

235–238.

Keen, H.G., Dekker, B.A., Disley, L., Hastings, H., Lyons, S., Reader,

A.J., Ottewell, P., Watson, A., Zweit, J., 2005. Imaging

apoptosis in vivo using 124I-annexin V and PET. Nucl. Med. Biol.

32, 395–402.

Lin, C.H., Lee, G.B., Lin, Y.H., Chang, G.L., 2001. A fast prototyping

process for fabrication of microfluidic systems on soda-lime glass. J.

Micromech. Microeng. 11, 726.

Massoud, T.F., Gambhir, S.S., 2003. Molecular imaging in living subjects:

seeing fundamental biological processes in a new light. Genes Dev. 17,

545.

Mitchell, P., 2001. Microfluidics—downsizing large-scale biology. Nature

Biotech. 19, 717.

Ramsey, J.M., 1999. The burgeoning power of the shrinking laboratory.

Nature Biotech. 17, 1061.

Reader, A.J., Zweit, J., 2001. Developments in whole-body molecular

imaging of live subjects. Trends Pharmacol. Sci. 22, 604.

Regenfuss, P., Clegg, R.M., Fulwyler, M.J., Barrantes, F.J., Jovin, T.M.,

1985. Mixing liquids in microseconds. Rev. Sci. Instrum. 56, 283.

Stuernstrom, M., Roeraade, J., 1998. Method for fabrication of

microfluidic systems in glass. J. Micromech. Microeng. 8, 33.

Tokita, N., Hasegawa, S., Maruyama, K., Izumi, T., Blankenberg, F.G.,

Tait, J.F., Strauss, H.W., Nishimura, T., 2003. 99mTc–Hynic-annexin V

imaging to evaluate inflammation and apoptosis in rats with auto-

immune myocarditis. Eur. J. Nucl. Med. Mol. Imaging 30, 232–238.

Tsai, J.H., Lin, L., 2001. Micro-to-macro fluidic interconnectors with an

integrated polymer sealant. J. Micromech. Microeng. 11, 577.