- Email: [email protected]
Membrane microfilter device for selective capture, electrolysis and genomic analysis of human circulating tumor cells
Journal of Chromatography A, 1162 (2007) 154–161
Membrane microfilter device for selective capture, electrolysis and genomic analysis of human circulating tumor cells Siyang Zheng a,∗,1 , Henry Lin b,1 , Jing-Quan Liu a,c , Marija Balic b , Ram Datar b , Richard J. Cote b , Yu-Chong Tai a a b
Department of Electrical Engineering, M/C 136-93, California Institute of Technology, Pasadena, CA 91125, USA Department of Pathology, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033, USA c Key Laboratory for Thin Film and Microfabrication of Ministry of Education, Research Institute of Micro/ Nanometer Science and Technology, Shanghai Jiao Tong University, Shanghai 200030, China Available online 29 May 2007
Abstract This paper presents development of a parylene membrane microfilter device for single stage capture and electrolysis of circulating tumor cells (CTCs) in human blood, and the potential of this device to allow genomic analysis. The presence and number of CTCs in blood has recently been demonstrated to provide significant prognostic information for patients with metastatic breast cancer. While finding as few as five CTCs in about 7.5 mL of blood (i.e., 1010 blood cells in) is clinically significant, detection of CTCs is currently difficult and time consuming. CTC enrichment is performed by either gradient centrifugation of CTC based on their buoyant density or magnetic separation of epithelial CTC, both of which are laborious procedures with variable efficiency, and CTC identification is typically done by trained pathologists through visual observation of stained cytokeratin-positive epithelial CTC. These processes may take hours, if not days. Work presented here provides a micro-electro-mechanical system (MEMS)-based option to make this process simpler, faster, better and cheaper. We exploited the size difference between CTCs and human blood cells to achieve the CTC capture on filter with ∼90% recovery within 10 min, which is superior to current approaches. Following capture, we facilitated polymerase chain reaction (PCR)-based genomic analysis by performing on-membrane electrolysis with embedded electrodes reaching each of the individual 16,000 filtering pores. The biggest advantage for this on-membrane in situ cell lysis is the high efficiency since cells are immobilized, allowing their direct contact with electrodes. As a proof-of-principle, we show beta actin gene PCR, the same technology can be easily extended to real time PCR for CTC-specific transcript to allow molecular identification of CTC and their further characterization. © 2007 Elsevier B.V. All rights reserved. Keywords: Circulating tumor cells; MEMS; Membrane filter; Parylene; Capture; Electrical lysis
1. Introduction Disseminated tumor cells (DTCs) can travel to distant organs through the hematogenous and/or lymphatic system; moreover, DTCs are prognostically critical, associated with clinical stage, disease recurrence, tumor metastasis and patient survival following therapy [1–4]. Circulating tumor cells (CTCs), in particular, are tumor cells flowing in the blood stream with the possibility of extravasation at distant organ sites [5,6]. A recent study with CTCs in metastatic breast cancer patients shows that those with more than five CTCs per 7.5 mL of blood have a much lower survival rate than patients with fewer cells [6].
∗ 1
Corresponding author. Tel.: +1 626 395 2227; fax: +1 626 584 9104. E-mail address: [email protected] (S. Zheng). The first two authors contributed equally.
0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.05.064
Having a technology that can reliably capture CTCs may allow for detecting the earliest signs of tumor metastasis, which is the cause of death for 90% of cancer patients, as well as enabling earlier therapeutic intervention which can lead to improved treatment outcome [7–9]. Several technologies are available for CTC isolation from whole human blood. Traditionally, density gradient centrifugation is employed to enrich the mononucleocyte (MNCs) fraction, which includes CTCs due to their similar buoyant density. The washed MNC fraction cells are cytospun onto glass slides followed by immunohistochemical staining for epithelial marker cytokeratin to detect CTC [10–12]. A trained pathologist examines each slide for the presence of CTCs, which is both time consuming and subjective. Moreover, density gradient centrifugation has a maximum recovery rate of ∼70%. An alternative technology based on predominant existence on CTC surface of epithelial marker proteins such as epithelial membrane antigen
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
(EMA) allows for their enrichment via immunomagnetic separation [11,13–15]. This method is expensive owing to cost of antibodies conjugated to magnetic beads, labor intensive and subject to a large range of yield (9–90%) [11] due to variable expression of surface markers. Isolation of CTC based on cell size using polycarbonate filters has been demonstrated to be an efficient, inexpensive and userfriendly way for enrichment of CTCs [16–20] by exploiting the fact that circulating epithelial tumor cells are significantly larger than the surrounding blood cells [11,19]. These polycarbonate filters are fabricated with track etching [21], which results in random placement of pores with relatively low density, and track etching often results in fusion of two or more pores, reflected as lower CTC capture efficiency of 50–60% [11,22]. There are two main challenges in building a microfabricated system for CTC isolation. First is the requirement of high efficiency of CTC recovery and their effective separation away from blood cells. The detection sensitivity required is high, with ability to capture as few as one CTC in 7.5 mL of whole blood, which contains about 10 billion blood cells. Secondly, the sample volume required to be processed is in the milliliter range, while microdevices are normally used to process nanoliter or even femtoliter volumes of sample. Such a challenge is further exacerbated when dilution of blood is required. Filters made with microfabrication technologies have several advantages for CTCs capture. Unlike the track-etched polycarbonate filters, the size, geometry and density of the pores can be precisely controlled. With batch fabrication, this technology can be very cost effective, which makes it suitable to develop a device for routine test in the clinics. The filter with uniformly spaced pores of identical diameter itself can afford maximal parallel processing capability, which reduces processing time and filter clogging due to back-pressure. In the present study, we demonstrate employment of a novel parylene membrane and CTC recovery rates higher or comparable to existing technologies. We used parylene-C to make the CTC capture filter. Several distinct properties make it one of the best candidates for this application. First, as the highest USP class IV biocompatible polymer for implementation, bio fouling is expected to be minimal for parylene-C. This strong but flexible material has excellent mechanical properties. It has a Young’s modulus of 4 GPa and high malleability that can withstand up to 200% elongation. It also has desirable electrical properties with low dielectric constant and high resistivity, which make it a good isolating material for electronics. Unlike the opaque polycarbonate filters, parylene is transparent in UV and visible range, which enables staining and observation of captured CTCs directly onmembrane without transferring captured CTCs to glass slides; this translates into minimal cell loss. Finally, we have established processing technologies to fabricate the filters. With room temperature, conformal and pinhole-free deposition, high quality parylene-C film can be routinely obtained. Metal deposition and oxygen plasma etching in reactive ion etching (RIE) system makes it possible to be integrated. Cell lysis in microdevices has been demonstrated based on various principles. Chemical lysis was achieved by mixing with
155
lysis buffer [23–25] or local hydroxide electro-generation [26]. Flowing cells through nano-structured barbs [27] or spinning them with beads can affect mechanical cell lysis [28]. Finally bacterial, yeast and mammalian cells have been electrolyzed with either DC or AC signals on microchips [29–34]. Electrolysis has the advantages of not requiring additional chemicals or mixing, no additional moving structures need to be introduced, and the electrodes are prefabricated. Micro-electrical cell lysis devices also have the advantage of lower applied electrical voltage relative to macroscopic electrical cell lysis instruments, which minimizes electrode damage and water hydrolysis of target cells, while also reducing operator-risk. But even for microdevices, the working potential is much higher than the voltage threshold for water hydrolysis (∼1 V), so rapid alternating signal is normally preferred to minimize gas bubble formation inside device and the extreme pH conditions close to electrodes. Unlike previous micro-cell electrolysis devices, in our devices, the cells were lysed in situ on the membrane instead of lysing them inside physically distinct fluidic channels or chambers, which typically can result in biomaterial loss; our approach has the potential of improving the lysis efficiency due to minimal cell movement and requires lower working voltage. 2. Experimental 2.1. Device design and fabrication We have two designs for membrane filters without integrated electrodes. Both designs are 1 cm by 1 cm square sheets with effective filter area of 0.6 cm by 0.6 cm. For design I, each pore is consisted of 10 m diameter circular holes with center to center distance between adjacent pores of 20 m. Design II uses oval shaped pores that were formed by etching rectangular masks of 14 m by 8 m with 12 m edge to edge distance. To fabricate the membrane filter without integrated electrodes, photoresist AZ1518 (Clariant, Somerville, NJ, USA) was spin-coated on silicon wafer. Parylene-C (Specialty Coating Systems, Indianapolis, IN, USA) was conformally deposited to 10 m thick and patterned with oxygen plasma in RIE where ˚ ˚ could be used either AZ9260 (Clariant) or Cr/Au (100 A/2000 A) as mask layer. Finally the whole film was released in acetone or photoresist stripper ST22 (ATMI, San Jose, CA, USA) at 80 ◦ C overnight. Fig. 1 shows fabricated filters. Membrane filters with integrated electrodes were designed to have 11 m diameter circular pores. The smallest distance between the pair of electrodes for each pore is 8 m. The center to center distance between adjacent pores was increased to 40 m to enable electrodes routing to reach individual pores. All anodes were tied together on one side of the filter. All cathodes were on the other side of the filter (Fig. 2). For membrane filters with integrated electrodes, a selfaligned process was used to produce electrodes with sharp edges precisely defined around each pore (Fig. 3). A 7.3 m thick parylene-C was conformally deposited on Si wafer and doublelayered photoresist compose of liftoff photoresist LOR3B (Microchem, Newton, MA, USA) as the first layer and AZ1518 as the second layer was deposited on the wafer front side.
156
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
Fig. 1. Fabricated membrane filters without electrodes. (A) Released membrane, (B) close-up view on one circular pore membrane filter and (C) close-up view on one oval pore membrane filter.
˚ ˚ was e-beam evaporated on wafer front Cr/Au (200 A/2500 A) side. Metal was then liftoff by dissolving photoresist in photoresist stripper ST22 at 80 ◦ C for 2 h with ultrasonication. The pore positions on the metal layer were patterned by wet etching (Transene, Danvers, MA, USA). Parylene pores were RIE etched with same mask. Finally the photoresist was striped and the whole film was released in deionized (DI) water. 2.2. Membrane filter assembly
Fig. 2. Device design for filters with integrated electrodes for cell electrical lysis. Left illustration is an overview of one filter device with individual electrodes wired up on top and bottom sides to form pads. Right picture shows a close-up view of the configuration of individual pores and electrodes reaching them.
Individual membrane filter was cut from the whole film with scissors and sandwiched between two pieces of polydimethylsiloxane (PDMS) with wells to form a chamber, which is then clamped between acrylic jig from the top and polyether ether ketone (PEEK) jig from the bottom for a sealed system (Fig. 4). The fluidic accesses to the filter were provided with two syringes with needles penetrating the PDMS pieces. Sample and washing buffer were applied to the filter from top syringe and collected in the bottom one.
Fig. 3. Fabrication process for filter with integrated electrodes for cell electrolysis.
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
157
osmium tetroxide at pH 7.4 for 60 min. Following fixation, each sample is washed in DI water for 5 min and dehydrated with 25, 50, 75% and 2× 100% ethanol in DI water for 5 min each. After the second 100% ethanol dehydration step, the samples were transfer to hexamethyldisilazane for 5 min and air-dried. 2.6. Electrical lysis of cells on-membrane filter After tumor cells were captured on the membrane filter, the filter was separated from the assembly and laid on a glass slide. Two mini clamps were connected to the pair of electrodes located on opposite sides of the filter to provide electrical signal for cell lysis. Ten kilohertz, 5 V peak-to-peak (Vpp), 30% duty cycle square wave signal from HP 33120A waveform generator was applied to each membrane for cell lysis and the voltage across electrodes was monitored with an oscilloscope during the process. 2.7. PCR after cell lysis
Fig. 4. Device assembly. (A) Illustration of the parylene membrane filter sandwiched between PDMS pieces and clamped by the jig and (B) a picture of the assembled device.
2.3. Sample preparation Cultured LNCaP cells derived from human metastatic prostatic adenocarcinoma were stained with Mayer’s modified hematoxylin (Ventana Medical Systems, Tucson, AZ, USA). The stained cells were serially diluted in Dulbecco’s phosphate buffered saline (DPBS; Mediatech, Herndon, VA, USA) to the desired concentration for device testing. For recovery testing, the cells were diluted to between 50 and 500 cells/mL to minimize dilution and counting errors. For detection limit testing, the cells were diluted to <10 cells/mL. Tumor cells were counted manually on hemacytometers (Hausser Scientific, Horsham, PA, USA). Human blood samples were collected from healthy donors in collection tube with EDTA to prevent coagulation and used within 24 h. 2.4. Cell filtration through membrane filter Cell suspension was loaded into a syringe, which is connected to the PDMS top chamber directly. Another empty syringe was connected to the bottom PDMS chamber. Sample was dispensed to traverse through the filter manually by pushing the plunger of the top syringe. The flow-through was collected by the bottom syringe. Further rinses with DPBS were performed in a similar way. 2.5. Sample preparation for scanning electron microscopy (SEM) imaging Filters containing cells were first fixed with formalin followed with secondary osmium fixation by submerging in a solution containing 0.1 M sodium cacodylate, 0.1 M sucrose and 1%
After cell electrolysis, DNA was eluted from the membrane with 30 L of PBS buffer and dispense into 3 L × 10 L aliquots subjecting to (1) DNA purification with QIAquick kit (Qiagen, Hilden, Germany), (2) alcohol precipitation and (3) without purification. Each aliquot was PCR amplified for actin with the following primer sequence: CTC CTT AAT GTC ACG CAC GAT, GTG GGG CGC CCC AGG CAC CA, giving a product size of 541 base pairs (bp). 3. Results and discussion 3.1. Capture of hematoxylin pre-stained tumor cells in PBS Although the diameter of erythrocytes (RBCs) is in the range of 5–9 m [35], its deformability enables it to traverse capillaries of 4 m diameter [36]. Early study on filtration of normal human red blood cells (RBCs) concludes that their transmission can achieve 100% for pore diameter larger than 3.3 m [37]. The average diameter of LNCaP cell line was measured to be 17 ± 1.5 m, which is larger than most of the blood cells. The size difference between CTCs and blood cells enable us to use the membrane filters for CTCs capture. The typical dimensions of blood cells are 5–9 m for erythrocytes, 10–15 m for granulocytes, 7–18 m for lymphocytes and 12–20 m for monocytes [38,39]. It is observed that very large leukocytes and erythrocyte agglomerates (rouleaux of erythrocytes) can be captured on the microfilters. Subsequent immunostaining can discriminate the CTCs from other cells on the filter. Because the number of these large particles in blood is minimal compared with erythrocytes, the enrichment of the microfilters can be at least 6 log10 . To evaluate the recovery rate of the membrane filters, LNCaP cells were stained with hematoxylin, which stains cell nucleuses as dark blue. Cells were serially diluted to desired concentration and dispensed through the membrane filter followed by rinsing twice with DPBS. Table 1 summarize the recovery rates for both circular (87.3 ± 7.0%) and oval designs (89.1 ± 7.0%). To characterize the sensitivity of this system,
158
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
Table 1 Recovery tests for circular pore design and oval pore design Test number
1 2 3 4 5
Circular pore
Oval pore
Cells recovered
Cells in flow-through
Cells recovered
Cells in flow-through
346 400 344 327 339
2 1 0 1 0
79 84 74 78 68
0 0 0 0 0
The expected cell number for circular pore design was 402 ± 56. The expected cell number for oval pore design was 86 ± 15. Table 2 Capture limit tests for circular pore design and oval pore design Test number
1 2 3 4 5
Circular pore
Oval pore
Cells recovered
Cells in flow-through
Cells recovered
Cells in flow-through
4 3 3 3 3
0 0 0 0 0
7 7 6 9 7
0 0 0 0 0
The expected cell number for circular pore design was 4 ± 1. The expected cell number for oval pore design was 8 ± 2.
recovery rates of samples containing <10 cells/mL were measure and demonstrated in Table 2. Using the circular pore design, we were able to recover at least three cells when four cells were expected and at least six cells were recovered when expected cell count of eight cells were applied with the oval pores. For each experiment, the expected cell numbers were determined by manual count using hemacytometer and the intersample variation is 10–20%. Judging from the results above, we did not experience a significant difference in performance between the oval and circular pore designs; therefore, we have based our filters with integrated electrodes using the circular design.
SEM pictures were taken for LNCaP cells isolated onmembrane filer (Fig. 5). Compared with commercially available polycarbonate filters (Fig. 5A), parylene filters (Fig. 5B and C) are more dense and without fused pores. The SEM fixation procedure (Fig. 5D) preserved the cell shape better than those without fixation (Fig. 5C). 3.2. Capture of hematoxylin pre-stained tumor cells spiked in whole human blood To mimic the real clinical samples obtained from cancer patients, the device performance was also tested by spiking
Fig. 5. SEM pictures. (A) Commercial membrane filter with sparse and occasionally fused pores, (B) microfabricated parylene membrane filter, (C) parylene membrane filter with cells captured without SEM fixation treatment and (D) parylene membrane filter with cells captured after SEM fixation procedure.
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
159
Fig. 7. FEMLAB simulation showing electrical field strength around pore region. The applied voltage is 3 V.
Fig. 6. Capture of blood spiked tumor cells (LNCaP) on parylene membrane filters. (A) Circular pore design and (B) oval pore design.
known numbers of LNCaP cells in peripheral blood from healthy donors (Fig. 6). Judging from the figure, the cell (blue circle) appears to be the same size as the filter pores; however, hematoxylin stains only the cell nucleus, which does not represent the entire cell. Increased back-pressure up to 3.45 kPa was experienced while applying the cells through the filters as compared with the previous experiments where LNCaP cells were spiked in PBS only. Although there is noticeable pressure build up, the parylene membrane filters can withstand the pressure without visible damages. Similar to experiments without blood, circular filter design resulted in 89.0 ± 9.5% recovery from blood as shown in Table 3. 3.3. Electrical lysis of tumor cells on-membrane filter after capture Assuming that transmembrane breakdown potential for a cell is 1 V [40], electrical field strength of 0.3 × 105 V/m will be required for electrical lysis of cell with 20 m diameter. The electrical field of the filter was simulated in two dimensions Table 3 Recovery for tumor cells spiked in 1 mL whole blood with circular pore design Test number
1
2
3
4
5
Tumor cells recovered Tumor cells in flow-through
33 0
35 0
42 0
39 0
34 0
The expected cell number was 41 ± 8.
(2D) with FEMLAB (Comsol Multiphysics, Stockholm, Sweden). Electrical field strength in the range of 105 –106 V/m was found on the edges of the electrodes, when 3 V of electrical potential was applied across the electrodes (Fig. 7). However, in our fabricated devices, the ideal sharp edge assumed in the simulation cannot be achieved; therefore, the actual electrical field strength is expected to be lower than the simulation. At 4 Vpp applied electrical potential, hematoxylin prestained cells barely showed signs of lysis compared with cells before electrical lysis (Fig. 8A and B). At 6 Vpp, the lysing efficiency reached over 90% with cells becoming more transparent and yellowish as the dye leaks out due to disintegration of the cell plasma and nucleus membrane (Fig. 8C). Reloaded fresh LNCaP cells were visually different from the lysed cells (Fig. 8D). Thanks to the small distance between electrodes and close contact with captured cells on-membrane, the applied voltage was much lower than the working potential of bulk electroporation instruments and even lower compared with other microdevices for cell electrical lysis [29–34]. The whole electrolysis process was monitored with oscilloscope (Fig. 9). During lysis, the voltage drop across the filter holes shows the capacitive feature. After the lysis, the voltage returns back to the applied waveform. The electrical lysis was normally performed in a couple of seconds. Prolonged lysis over 10 s could cause electrode loss. 3.4. PCR study of tumor cell lysate To confirm the electrical lysis and demonstrate our ability to perform single stage capture and electrical lysis for genomic analysis, PCR was performed on tumor cell lysate. As shown in Fig. 10, purified lysate by either alcohol precipitation or QIAquick purification resulted in a positive band at 541 bp on the electrophoresis gel; however, untreated lysate failed to be amplified by the PCR reaction.
160
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
Fig. 8. Electrolysis of captured tumor cells with 10 kHz and 30% duty cycle square wave at different peak-to-peak voltage for 20 s. (A) Before lysis, (B) after 4 Vpp is applied, (C) after 6 Vpp is applied and (D) fresh tumor cells (as pointed) are reloaded.
Fig. 10. PCR of electrolysis elute on-membrane filter. Lane 1, 2, 7: various negative controls; lane 8: positive control with 104 cells loaded; lane 3: electrolysis elute without purification; lane 5: electrolysis elute purified by alcohol precipitation; lane 6: electrolysis elute purified by QIAquick kit. Lane 4: 100 bp DNA ladder.
4. Conclusions and perspectives
Fig. 9. Electrical potential waveform change during electrolysis. (A) The applied electrical signal and (B) voltage potential of the membrane during cell lysis.
We demonstrated an efficient, reliable and low-cost technology to capture and electrically lyse the captured CTCs from human whole blood on a single platform. With microfabricated membrane filter, tumor cell isolation can be achieved with recovery rate of about 90% and fast processing time within 10 min. The excellent mechanical, electrical, optical and biocompatible properties of parylene-C made it a good candidate for the membrane filter. Highly efficient on-chip electrical lysis after cell capture was confirmed optically with microscope and chemically by DNA PCR amplification to prove the feasibility of down stream genomic analysis. To further optimize the system, the effect of pore size, shape and membrane thickness on the recovery rate will be studied. Because of the special hydrodynamic properties of whole blood,
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
it will be interesting to investigate the relationship between pressure and flow rate of the filtration process. Maintaining the lowest possible pressure is important to keep the shear stress on cells low, thus preventing cell damage. Clinical testing for this technology is underway. Multiple filters with various pore sizes stacked together will be useful in clinical testing where sizes of CTCs might vary. Finally, integrating other microfluidic components on the membrane filter and making it a lab-on-a-chip device for characterization of single cells will be a very powerful tool for cancer research. Acknowledgment The funding of the project was provided by NIH 1R21 CA123027-01. References [1] M. Cristofanilli, G.T. Budd, M.J. Ellis, A. Stopeck, J. Matera, M.C. Miller, J.M. Reuben, G.V. Doyle, W.J. Allard, L.W.M.M. Terstappen, D.F. Hayes, New Engl. J. Med. 351 (2004) 781. [2] M. Cristofanilli, D.F. Hayes, G.T. Budd, M.J. Ellis, A. Stopeck, J.M. Reuben, G.V. Doyle, J. Matera, W.J. Allard, M.C. Miller, H.A. Fritsche, G.N. Hortobagyi, L.W.M.M. Terstappen, J. Clin. Oncol. 23 (2005) 1420. [3] D.F. Hayes, M. Cristofanilli, G.T. Budd, M.J. Ellis, A. Stopeck, M.C. Miller, J. Matera, W.J. Allard, G.V. Doyle, L.W.W.M. Terstappen, Clin. Cancer Res. 12 (2006) 4218. [4] G.T. Budd, M. Cristofanilli, M.J. Ellis, A. Stopeck, E. Borden, M.C. Miller, J. Matera, M. Repollet, G.V. Doyle, L. Terstappen, D.F. Hayes, Clin. Cancer Res. 12 (2006) 6403. [5] G.P. Gupta, J. Massague, Cell 127 (2006) 679. [6] D. Hanahan, R.A. Weinberg, Cell 100 (2000) 57. [7] M. Cristofanilli, J. Mendelsohn, P. Natl. Acad. Sci. USA 103 (2006) 17073. [8] L. Dirix, P. Van Dam, P. Vermeulen, Curr. Opin. Oncol. 17 (2005) 551. [9] S. Meng, D. Tripathy, S. Shete, R. Ashfaq, H. Saboorian, B. Haley, E. Frenkel, D. Euhus, M. Leitch, C. Osborne, E. Clifford, S. Perkins, P. Beitsch, A. Khan, L. Morrison, D. Herlyn, L.W.M.M. Terstappen, N. Lane, J. Wang, J. Uhr, P. Natl. Acad. Sci. USA 103 (2006) 17361. [10] M.K. Baker, K. Mikhitarian, W. Osta, K. Callahan, R. Hoda, F. Brescia, R. Kneuper-Hall, M. Mitas, D.J. Cole, W.E. Gillanders, Clin. Cancer Res. 9 (2003) 4865. [11] O. Lara, X.D. Tong, M. Zborowski, J.J. Chalmers, Exp. Hematol. 32 (2004) 891. [12] K. Pantel, R.H. Brakenhoff, Nat. Rev. Cancer 4 (2004) 448. [13] A. Benez, A. Geiselhart, R. Handgretinger, U. Schiebel, G. Fierlbeck, J. Clin. Lab. Anal. 13 (1999) 229.
161
[14] J.E. Hardingham, D. Kotasek, B. Farmer, R.N. Butler, J.X. Mi, R.E. Sage, A. Dobrovic, Cancer Res. 53 (1993) 3455. [15] E. Racila, D. Euhus, A.J. Weiss, C. Rao, J. McConnell, L. Terstappen, J.W. Uhr, P. Natl. Acad. Sci. USA 95 (1998) 4589. [16] P. Pinzani, B. Salvadori, L. Simi, S. Bianchi, V. Distante, L. Cataliotti, M. Pazzagli, C. Orlando, Hum. Pathol. 37 (2006) 711. [17] G. Vona, C. Beroud, A. Benachi, A. Quenette, J.P. Bonnefont, S. Romana, A. Munnich, M. Vekemans, Y. Dumez, B. Lacour, P. Paterlini-Brechot, Am. J. Pathol. 160 (2002) 51. [18] G. Vona, L. Estepa, C. Beroud, D. Damotte, F. Capron, B. Nalpas, A. Mineur, D. Franco, B. Lacour, S. Pol, C. Brechot, P. Paterlini-Brechot, Hepatology 39 (2004) 792. [19] G. Vona, A. Sabile, M. Louha, V. Sitruk, S. Romana, K. Schutze, F. Capron, D. Franco, M. Pazzagli, M. Vekemans, B. Lacour, C. Brechot, P. PaterliniBrechot, Am. J. Pathol. 156 (2000) 57. [20] H.J. Kahn, A. Presta, L.Y. Yang, J. Blondal, M. Trudeau, L. Lickley, C. Holloway, D.R. McCready, D. Maclean, A. Marks, Breast Cancer Res. Treat. 86 (2004) 237. [21] F. Rl, H.W. Alter, R.M. Walker, S.C. Furman, P.B. Price, Science 178 (1972) 255. [22] P. Rostagno, J.L. Moll, J.C. Bisconte, C. Caldani, Anticancer Res. 17 (1997) 2481. [23] J. El-Ali, S. Gaudet, A. Gunther, P.K. Sorger, K.F. Jensen, Anal. Chem. 77 (2005) 3629. [24] D. Irimia, R.G. Tompkins, M. Toner, Anal. Chem. 76 (2004) 6137. [25] E.A. Schilling, A.E. Kamholz, P. Yager, Anal. Chem. 74 (2002) 1798. [26] D. Di Carlo, C. Ionescu-Zanetti, Y. Zhang, P. Hung, L.P. Lee, Lab. Chip 5 (2005) 171. [27] D. Di Carlo, K.H. Jeong, L.P. Lee, Lab. Chip 3 (2003) 287. [28] J. Kim, S.H. Jang, G.Y. Jia, J.V. Zoval, N.A. Da Silva, M.J. Madou, Lab. Chip 4 (2004) 516. [29] J. Cheng, E.L. Sheldon, L. Wu, M.J. Heller, J.P. O’Connell, Anal. Chem. 70 (1998) 2321. [30] J. Cheng, E.L. Sheldon, L. Wu, A. Uribe, L.O. Gerrue, J. Carrino, M.J. Heller, J.P. O’Connell, Nat. Biotechnol. 16 (1998) 541. [31] S.W. Lee, Y.C. Tai, Sens. Actuat. A-Phys. 73 (1999) 74. [32] F.T. Han, Y. Wang, C.E. Sims, M. Bachman, R.S. Chang, G.P. Li, N.L. Allbritton, Anal. Chem. 75 (2003) 3688. [33] H. Lu, M.A. Schmidt, K.F. Jensen, Lab. Chip 5 (2005) 23. [34] Q. Ramadan, V. Samper, D. Poenar, Z. Liang, C. Yu, T.M. Lim, Sens. Actuat. B-Chem. 113 (2006) 944. [35] S.B. McKenzie, Clinical Laboratory Hematology, Prentice Hall, Upper Saddle River, NJ, 2004. [36] D. Braasch, Physiol. Rev. 51 (1971) 679. [37] S. Chien, S. Usami, Rj. Dellenba, Mi. Gregerse, Am. J. Physiol. 219 (1970) 136. [38] C.E. Lentner, Geigy Scientific Tables, Ciba-Geigy, Basel, Switzerland, 1981. [39] H.M. Shapiro, E.R. Schildkraut, R. Curbelo, C.W. Laird, R.B. Turner, T. Hirschfeld, J. Histochem. Cytochem. 24 (1976) 396. [40] T.Y. Tsong, Biophys. J. 60 (1991) 297.
Membrane microfilter device for selective capture, electrolysis and genomic analysis of human circulating tumor cells Siyang Zheng a,∗,1 , Henry Lin b,1 , Jing-Quan Liu a,c , Marija Balic b , Ram Datar b , Richard J. Cote b , Yu-Chong Tai a a b
Department of Electrical Engineering, M/C 136-93, California Institute of Technology, Pasadena, CA 91125, USA Department of Pathology, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033, USA c Key Laboratory for Thin Film and Microfabrication of Ministry of Education, Research Institute of Micro/ Nanometer Science and Technology, Shanghai Jiao Tong University, Shanghai 200030, China Available online 29 May 2007
Abstract This paper presents development of a parylene membrane microfilter device for single stage capture and electrolysis of circulating tumor cells (CTCs) in human blood, and the potential of this device to allow genomic analysis. The presence and number of CTCs in blood has recently been demonstrated to provide significant prognostic information for patients with metastatic breast cancer. While finding as few as five CTCs in about 7.5 mL of blood (i.e., 1010 blood cells in) is clinically significant, detection of CTCs is currently difficult and time consuming. CTC enrichment is performed by either gradient centrifugation of CTC based on their buoyant density or magnetic separation of epithelial CTC, both of which are laborious procedures with variable efficiency, and CTC identification is typically done by trained pathologists through visual observation of stained cytokeratin-positive epithelial CTC. These processes may take hours, if not days. Work presented here provides a micro-electro-mechanical system (MEMS)-based option to make this process simpler, faster, better and cheaper. We exploited the size difference between CTCs and human blood cells to achieve the CTC capture on filter with ∼90% recovery within 10 min, which is superior to current approaches. Following capture, we facilitated polymerase chain reaction (PCR)-based genomic analysis by performing on-membrane electrolysis with embedded electrodes reaching each of the individual 16,000 filtering pores. The biggest advantage for this on-membrane in situ cell lysis is the high efficiency since cells are immobilized, allowing their direct contact with electrodes. As a proof-of-principle, we show beta actin gene PCR, the same technology can be easily extended to real time PCR for CTC-specific transcript to allow molecular identification of CTC and their further characterization. © 2007 Elsevier B.V. All rights reserved. Keywords: Circulating tumor cells; MEMS; Membrane filter; Parylene; Capture; Electrical lysis
1. Introduction Disseminated tumor cells (DTCs) can travel to distant organs through the hematogenous and/or lymphatic system; moreover, DTCs are prognostically critical, associated with clinical stage, disease recurrence, tumor metastasis and patient survival following therapy [1–4]. Circulating tumor cells (CTCs), in particular, are tumor cells flowing in the blood stream with the possibility of extravasation at distant organ sites [5,6]. A recent study with CTCs in metastatic breast cancer patients shows that those with more than five CTCs per 7.5 mL of blood have a much lower survival rate than patients with fewer cells [6].
∗ 1
Corresponding author. Tel.: +1 626 395 2227; fax: +1 626 584 9104. E-mail address: [email protected] (S. Zheng). The first two authors contributed equally.
0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.05.064
Having a technology that can reliably capture CTCs may allow for detecting the earliest signs of tumor metastasis, which is the cause of death for 90% of cancer patients, as well as enabling earlier therapeutic intervention which can lead to improved treatment outcome [7–9]. Several technologies are available for CTC isolation from whole human blood. Traditionally, density gradient centrifugation is employed to enrich the mononucleocyte (MNCs) fraction, which includes CTCs due to their similar buoyant density. The washed MNC fraction cells are cytospun onto glass slides followed by immunohistochemical staining for epithelial marker cytokeratin to detect CTC [10–12]. A trained pathologist examines each slide for the presence of CTCs, which is both time consuming and subjective. Moreover, density gradient centrifugation has a maximum recovery rate of ∼70%. An alternative technology based on predominant existence on CTC surface of epithelial marker proteins such as epithelial membrane antigen
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
(EMA) allows for their enrichment via immunomagnetic separation [11,13–15]. This method is expensive owing to cost of antibodies conjugated to magnetic beads, labor intensive and subject to a large range of yield (9–90%) [11] due to variable expression of surface markers. Isolation of CTC based on cell size using polycarbonate filters has been demonstrated to be an efficient, inexpensive and userfriendly way for enrichment of CTCs [16–20] by exploiting the fact that circulating epithelial tumor cells are significantly larger than the surrounding blood cells [11,19]. These polycarbonate filters are fabricated with track etching [21], which results in random placement of pores with relatively low density, and track etching often results in fusion of two or more pores, reflected as lower CTC capture efficiency of 50–60% [11,22]. There are two main challenges in building a microfabricated system for CTC isolation. First is the requirement of high efficiency of CTC recovery and their effective separation away from blood cells. The detection sensitivity required is high, with ability to capture as few as one CTC in 7.5 mL of whole blood, which contains about 10 billion blood cells. Secondly, the sample volume required to be processed is in the milliliter range, while microdevices are normally used to process nanoliter or even femtoliter volumes of sample. Such a challenge is further exacerbated when dilution of blood is required. Filters made with microfabrication technologies have several advantages for CTCs capture. Unlike the track-etched polycarbonate filters, the size, geometry and density of the pores can be precisely controlled. With batch fabrication, this technology can be very cost effective, which makes it suitable to develop a device for routine test in the clinics. The filter with uniformly spaced pores of identical diameter itself can afford maximal parallel processing capability, which reduces processing time and filter clogging due to back-pressure. In the present study, we demonstrate employment of a novel parylene membrane and CTC recovery rates higher or comparable to existing technologies. We used parylene-C to make the CTC capture filter. Several distinct properties make it one of the best candidates for this application. First, as the highest USP class IV biocompatible polymer for implementation, bio fouling is expected to be minimal for parylene-C. This strong but flexible material has excellent mechanical properties. It has a Young’s modulus of 4 GPa and high malleability that can withstand up to 200% elongation. It also has desirable electrical properties with low dielectric constant and high resistivity, which make it a good isolating material for electronics. Unlike the opaque polycarbonate filters, parylene is transparent in UV and visible range, which enables staining and observation of captured CTCs directly onmembrane without transferring captured CTCs to glass slides; this translates into minimal cell loss. Finally, we have established processing technologies to fabricate the filters. With room temperature, conformal and pinhole-free deposition, high quality parylene-C film can be routinely obtained. Metal deposition and oxygen plasma etching in reactive ion etching (RIE) system makes it possible to be integrated. Cell lysis in microdevices has been demonstrated based on various principles. Chemical lysis was achieved by mixing with
155
lysis buffer [23–25] or local hydroxide electro-generation [26]. Flowing cells through nano-structured barbs [27] or spinning them with beads can affect mechanical cell lysis [28]. Finally bacterial, yeast and mammalian cells have been electrolyzed with either DC or AC signals on microchips [29–34]. Electrolysis has the advantages of not requiring additional chemicals or mixing, no additional moving structures need to be introduced, and the electrodes are prefabricated. Micro-electrical cell lysis devices also have the advantage of lower applied electrical voltage relative to macroscopic electrical cell lysis instruments, which minimizes electrode damage and water hydrolysis of target cells, while also reducing operator-risk. But even for microdevices, the working potential is much higher than the voltage threshold for water hydrolysis (∼1 V), so rapid alternating signal is normally preferred to minimize gas bubble formation inside device and the extreme pH conditions close to electrodes. Unlike previous micro-cell electrolysis devices, in our devices, the cells were lysed in situ on the membrane instead of lysing them inside physically distinct fluidic channels or chambers, which typically can result in biomaterial loss; our approach has the potential of improving the lysis efficiency due to minimal cell movement and requires lower working voltage. 2. Experimental 2.1. Device design and fabrication We have two designs for membrane filters without integrated electrodes. Both designs are 1 cm by 1 cm square sheets with effective filter area of 0.6 cm by 0.6 cm. For design I, each pore is consisted of 10 m diameter circular holes with center to center distance between adjacent pores of 20 m. Design II uses oval shaped pores that were formed by etching rectangular masks of 14 m by 8 m with 12 m edge to edge distance. To fabricate the membrane filter without integrated electrodes, photoresist AZ1518 (Clariant, Somerville, NJ, USA) was spin-coated on silicon wafer. Parylene-C (Specialty Coating Systems, Indianapolis, IN, USA) was conformally deposited to 10 m thick and patterned with oxygen plasma in RIE where ˚ ˚ could be used either AZ9260 (Clariant) or Cr/Au (100 A/2000 A) as mask layer. Finally the whole film was released in acetone or photoresist stripper ST22 (ATMI, San Jose, CA, USA) at 80 ◦ C overnight. Fig. 1 shows fabricated filters. Membrane filters with integrated electrodes were designed to have 11 m diameter circular pores. The smallest distance between the pair of electrodes for each pore is 8 m. The center to center distance between adjacent pores was increased to 40 m to enable electrodes routing to reach individual pores. All anodes were tied together on one side of the filter. All cathodes were on the other side of the filter (Fig. 2). For membrane filters with integrated electrodes, a selfaligned process was used to produce electrodes with sharp edges precisely defined around each pore (Fig. 3). A 7.3 m thick parylene-C was conformally deposited on Si wafer and doublelayered photoresist compose of liftoff photoresist LOR3B (Microchem, Newton, MA, USA) as the first layer and AZ1518 as the second layer was deposited on the wafer front side.
156
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
Fig. 1. Fabricated membrane filters without electrodes. (A) Released membrane, (B) close-up view on one circular pore membrane filter and (C) close-up view on one oval pore membrane filter.
˚ ˚ was e-beam evaporated on wafer front Cr/Au (200 A/2500 A) side. Metal was then liftoff by dissolving photoresist in photoresist stripper ST22 at 80 ◦ C for 2 h with ultrasonication. The pore positions on the metal layer were patterned by wet etching (Transene, Danvers, MA, USA). Parylene pores were RIE etched with same mask. Finally the photoresist was striped and the whole film was released in deionized (DI) water. 2.2. Membrane filter assembly
Fig. 2. Device design for filters with integrated electrodes for cell electrical lysis. Left illustration is an overview of one filter device with individual electrodes wired up on top and bottom sides to form pads. Right picture shows a close-up view of the configuration of individual pores and electrodes reaching them.
Individual membrane filter was cut from the whole film with scissors and sandwiched between two pieces of polydimethylsiloxane (PDMS) with wells to form a chamber, which is then clamped between acrylic jig from the top and polyether ether ketone (PEEK) jig from the bottom for a sealed system (Fig. 4). The fluidic accesses to the filter were provided with two syringes with needles penetrating the PDMS pieces. Sample and washing buffer were applied to the filter from top syringe and collected in the bottom one.
Fig. 3. Fabrication process for filter with integrated electrodes for cell electrolysis.
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
157
osmium tetroxide at pH 7.4 for 60 min. Following fixation, each sample is washed in DI water for 5 min and dehydrated with 25, 50, 75% and 2× 100% ethanol in DI water for 5 min each. After the second 100% ethanol dehydration step, the samples were transfer to hexamethyldisilazane for 5 min and air-dried. 2.6. Electrical lysis of cells on-membrane filter After tumor cells were captured on the membrane filter, the filter was separated from the assembly and laid on a glass slide. Two mini clamps were connected to the pair of electrodes located on opposite sides of the filter to provide electrical signal for cell lysis. Ten kilohertz, 5 V peak-to-peak (Vpp), 30% duty cycle square wave signal from HP 33120A waveform generator was applied to each membrane for cell lysis and the voltage across electrodes was monitored with an oscilloscope during the process. 2.7. PCR after cell lysis
Fig. 4. Device assembly. (A) Illustration of the parylene membrane filter sandwiched between PDMS pieces and clamped by the jig and (B) a picture of the assembled device.
2.3. Sample preparation Cultured LNCaP cells derived from human metastatic prostatic adenocarcinoma were stained with Mayer’s modified hematoxylin (Ventana Medical Systems, Tucson, AZ, USA). The stained cells were serially diluted in Dulbecco’s phosphate buffered saline (DPBS; Mediatech, Herndon, VA, USA) to the desired concentration for device testing. For recovery testing, the cells were diluted to between 50 and 500 cells/mL to minimize dilution and counting errors. For detection limit testing, the cells were diluted to <10 cells/mL. Tumor cells were counted manually on hemacytometers (Hausser Scientific, Horsham, PA, USA). Human blood samples were collected from healthy donors in collection tube with EDTA to prevent coagulation and used within 24 h. 2.4. Cell filtration through membrane filter Cell suspension was loaded into a syringe, which is connected to the PDMS top chamber directly. Another empty syringe was connected to the bottom PDMS chamber. Sample was dispensed to traverse through the filter manually by pushing the plunger of the top syringe. The flow-through was collected by the bottom syringe. Further rinses with DPBS were performed in a similar way. 2.5. Sample preparation for scanning electron microscopy (SEM) imaging Filters containing cells were first fixed with formalin followed with secondary osmium fixation by submerging in a solution containing 0.1 M sodium cacodylate, 0.1 M sucrose and 1%
After cell electrolysis, DNA was eluted from the membrane with 30 L of PBS buffer and dispense into 3 L × 10 L aliquots subjecting to (1) DNA purification with QIAquick kit (Qiagen, Hilden, Germany), (2) alcohol precipitation and (3) without purification. Each aliquot was PCR amplified for actin with the following primer sequence: CTC CTT AAT GTC ACG CAC GAT, GTG GGG CGC CCC AGG CAC CA, giving a product size of 541 base pairs (bp). 3. Results and discussion 3.1. Capture of hematoxylin pre-stained tumor cells in PBS Although the diameter of erythrocytes (RBCs) is in the range of 5–9 m [35], its deformability enables it to traverse capillaries of 4 m diameter [36]. Early study on filtration of normal human red blood cells (RBCs) concludes that their transmission can achieve 100% for pore diameter larger than 3.3 m [37]. The average diameter of LNCaP cell line was measured to be 17 ± 1.5 m, which is larger than most of the blood cells. The size difference between CTCs and blood cells enable us to use the membrane filters for CTCs capture. The typical dimensions of blood cells are 5–9 m for erythrocytes, 10–15 m for granulocytes, 7–18 m for lymphocytes and 12–20 m for monocytes [38,39]. It is observed that very large leukocytes and erythrocyte agglomerates (rouleaux of erythrocytes) can be captured on the microfilters. Subsequent immunostaining can discriminate the CTCs from other cells on the filter. Because the number of these large particles in blood is minimal compared with erythrocytes, the enrichment of the microfilters can be at least 6 log10 . To evaluate the recovery rate of the membrane filters, LNCaP cells were stained with hematoxylin, which stains cell nucleuses as dark blue. Cells were serially diluted to desired concentration and dispensed through the membrane filter followed by rinsing twice with DPBS. Table 1 summarize the recovery rates for both circular (87.3 ± 7.0%) and oval designs (89.1 ± 7.0%). To characterize the sensitivity of this system,
158
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
Table 1 Recovery tests for circular pore design and oval pore design Test number
1 2 3 4 5
Circular pore
Oval pore
Cells recovered
Cells in flow-through
Cells recovered
Cells in flow-through
346 400 344 327 339
2 1 0 1 0
79 84 74 78 68
0 0 0 0 0
The expected cell number for circular pore design was 402 ± 56. The expected cell number for oval pore design was 86 ± 15. Table 2 Capture limit tests for circular pore design and oval pore design Test number
1 2 3 4 5
Circular pore
Oval pore
Cells recovered
Cells in flow-through
Cells recovered
Cells in flow-through
4 3 3 3 3
0 0 0 0 0
7 7 6 9 7
0 0 0 0 0
The expected cell number for circular pore design was 4 ± 1. The expected cell number for oval pore design was 8 ± 2.
recovery rates of samples containing <10 cells/mL were measure and demonstrated in Table 2. Using the circular pore design, we were able to recover at least three cells when four cells were expected and at least six cells were recovered when expected cell count of eight cells were applied with the oval pores. For each experiment, the expected cell numbers were determined by manual count using hemacytometer and the intersample variation is 10–20%. Judging from the results above, we did not experience a significant difference in performance between the oval and circular pore designs; therefore, we have based our filters with integrated electrodes using the circular design.
SEM pictures were taken for LNCaP cells isolated onmembrane filer (Fig. 5). Compared with commercially available polycarbonate filters (Fig. 5A), parylene filters (Fig. 5B and C) are more dense and without fused pores. The SEM fixation procedure (Fig. 5D) preserved the cell shape better than those without fixation (Fig. 5C). 3.2. Capture of hematoxylin pre-stained tumor cells spiked in whole human blood To mimic the real clinical samples obtained from cancer patients, the device performance was also tested by spiking
Fig. 5. SEM pictures. (A) Commercial membrane filter with sparse and occasionally fused pores, (B) microfabricated parylene membrane filter, (C) parylene membrane filter with cells captured without SEM fixation treatment and (D) parylene membrane filter with cells captured after SEM fixation procedure.
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
159
Fig. 7. FEMLAB simulation showing electrical field strength around pore region. The applied voltage is 3 V.
Fig. 6. Capture of blood spiked tumor cells (LNCaP) on parylene membrane filters. (A) Circular pore design and (B) oval pore design.
known numbers of LNCaP cells in peripheral blood from healthy donors (Fig. 6). Judging from the figure, the cell (blue circle) appears to be the same size as the filter pores; however, hematoxylin stains only the cell nucleus, which does not represent the entire cell. Increased back-pressure up to 3.45 kPa was experienced while applying the cells through the filters as compared with the previous experiments where LNCaP cells were spiked in PBS only. Although there is noticeable pressure build up, the parylene membrane filters can withstand the pressure without visible damages. Similar to experiments without blood, circular filter design resulted in 89.0 ± 9.5% recovery from blood as shown in Table 3. 3.3. Electrical lysis of tumor cells on-membrane filter after capture Assuming that transmembrane breakdown potential for a cell is 1 V [40], electrical field strength of 0.3 × 105 V/m will be required for electrical lysis of cell with 20 m diameter. The electrical field of the filter was simulated in two dimensions Table 3 Recovery for tumor cells spiked in 1 mL whole blood with circular pore design Test number
1
2
3
4
5
Tumor cells recovered Tumor cells in flow-through
33 0
35 0
42 0
39 0
34 0
The expected cell number was 41 ± 8.
(2D) with FEMLAB (Comsol Multiphysics, Stockholm, Sweden). Electrical field strength in the range of 105 –106 V/m was found on the edges of the electrodes, when 3 V of electrical potential was applied across the electrodes (Fig. 7). However, in our fabricated devices, the ideal sharp edge assumed in the simulation cannot be achieved; therefore, the actual electrical field strength is expected to be lower than the simulation. At 4 Vpp applied electrical potential, hematoxylin prestained cells barely showed signs of lysis compared with cells before electrical lysis (Fig. 8A and B). At 6 Vpp, the lysing efficiency reached over 90% with cells becoming more transparent and yellowish as the dye leaks out due to disintegration of the cell plasma and nucleus membrane (Fig. 8C). Reloaded fresh LNCaP cells were visually different from the lysed cells (Fig. 8D). Thanks to the small distance between electrodes and close contact with captured cells on-membrane, the applied voltage was much lower than the working potential of bulk electroporation instruments and even lower compared with other microdevices for cell electrical lysis [29–34]. The whole electrolysis process was monitored with oscilloscope (Fig. 9). During lysis, the voltage drop across the filter holes shows the capacitive feature. After the lysis, the voltage returns back to the applied waveform. The electrical lysis was normally performed in a couple of seconds. Prolonged lysis over 10 s could cause electrode loss. 3.4. PCR study of tumor cell lysate To confirm the electrical lysis and demonstrate our ability to perform single stage capture and electrical lysis for genomic analysis, PCR was performed on tumor cell lysate. As shown in Fig. 10, purified lysate by either alcohol precipitation or QIAquick purification resulted in a positive band at 541 bp on the electrophoresis gel; however, untreated lysate failed to be amplified by the PCR reaction.
160
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
Fig. 8. Electrolysis of captured tumor cells with 10 kHz and 30% duty cycle square wave at different peak-to-peak voltage for 20 s. (A) Before lysis, (B) after 4 Vpp is applied, (C) after 6 Vpp is applied and (D) fresh tumor cells (as pointed) are reloaded.
Fig. 10. PCR of electrolysis elute on-membrane filter. Lane 1, 2, 7: various negative controls; lane 8: positive control with 104 cells loaded; lane 3: electrolysis elute without purification; lane 5: electrolysis elute purified by alcohol precipitation; lane 6: electrolysis elute purified by QIAquick kit. Lane 4: 100 bp DNA ladder.
4. Conclusions and perspectives
Fig. 9. Electrical potential waveform change during electrolysis. (A) The applied electrical signal and (B) voltage potential of the membrane during cell lysis.
We demonstrated an efficient, reliable and low-cost technology to capture and electrically lyse the captured CTCs from human whole blood on a single platform. With microfabricated membrane filter, tumor cell isolation can be achieved with recovery rate of about 90% and fast processing time within 10 min. The excellent mechanical, electrical, optical and biocompatible properties of parylene-C made it a good candidate for the membrane filter. Highly efficient on-chip electrical lysis after cell capture was confirmed optically with microscope and chemically by DNA PCR amplification to prove the feasibility of down stream genomic analysis. To further optimize the system, the effect of pore size, shape and membrane thickness on the recovery rate will be studied. Because of the special hydrodynamic properties of whole blood,
S. Zheng et al. / J. Chromatogr. A 1162 (2007) 154–161
it will be interesting to investigate the relationship between pressure and flow rate of the filtration process. Maintaining the lowest possible pressure is important to keep the shear stress on cells low, thus preventing cell damage. Clinical testing for this technology is underway. Multiple filters with various pore sizes stacked together will be useful in clinical testing where sizes of CTCs might vary. Finally, integrating other microfluidic components on the membrane filter and making it a lab-on-a-chip device for characterization of single cells will be a very powerful tool for cancer research. Acknowledgment The funding of the project was provided by NIH 1R21 CA123027-01. References [1] M. Cristofanilli, G.T. Budd, M.J. Ellis, A. Stopeck, J. Matera, M.C. Miller, J.M. Reuben, G.V. Doyle, W.J. Allard, L.W.M.M. Terstappen, D.F. Hayes, New Engl. J. Med. 351 (2004) 781. [2] M. Cristofanilli, D.F. Hayes, G.T. Budd, M.J. Ellis, A. Stopeck, J.M. Reuben, G.V. Doyle, J. Matera, W.J. Allard, M.C. Miller, H.A. Fritsche, G.N. Hortobagyi, L.W.M.M. Terstappen, J. Clin. Oncol. 23 (2005) 1420. [3] D.F. Hayes, M. Cristofanilli, G.T. Budd, M.J. Ellis, A. Stopeck, M.C. Miller, J. Matera, W.J. Allard, G.V. Doyle, L.W.W.M. Terstappen, Clin. Cancer Res. 12 (2006) 4218. [4] G.T. Budd, M. Cristofanilli, M.J. Ellis, A. Stopeck, E. Borden, M.C. Miller, J. Matera, M. Repollet, G.V. Doyle, L. Terstappen, D.F. Hayes, Clin. Cancer Res. 12 (2006) 6403. [5] G.P. Gupta, J. Massague, Cell 127 (2006) 679. [6] D. Hanahan, R.A. Weinberg, Cell 100 (2000) 57. [7] M. Cristofanilli, J. Mendelsohn, P. Natl. Acad. Sci. USA 103 (2006) 17073. [8] L. Dirix, P. Van Dam, P. Vermeulen, Curr. Opin. Oncol. 17 (2005) 551. [9] S. Meng, D. Tripathy, S. Shete, R. Ashfaq, H. Saboorian, B. Haley, E. Frenkel, D. Euhus, M. Leitch, C. Osborne, E. Clifford, S. Perkins, P. Beitsch, A. Khan, L. Morrison, D. Herlyn, L.W.M.M. Terstappen, N. Lane, J. Wang, J. Uhr, P. Natl. Acad. Sci. USA 103 (2006) 17361. [10] M.K. Baker, K. Mikhitarian, W. Osta, K. Callahan, R. Hoda, F. Brescia, R. Kneuper-Hall, M. Mitas, D.J. Cole, W.E. Gillanders, Clin. Cancer Res. 9 (2003) 4865. [11] O. Lara, X.D. Tong, M. Zborowski, J.J. Chalmers, Exp. Hematol. 32 (2004) 891. [12] K. Pantel, R.H. Brakenhoff, Nat. Rev. Cancer 4 (2004) 448. [13] A. Benez, A. Geiselhart, R. Handgretinger, U. Schiebel, G. Fierlbeck, J. Clin. Lab. Anal. 13 (1999) 229.
161
[14] J.E. Hardingham, D. Kotasek, B. Farmer, R.N. Butler, J.X. Mi, R.E. Sage, A. Dobrovic, Cancer Res. 53 (1993) 3455. [15] E. Racila, D. Euhus, A.J. Weiss, C. Rao, J. McConnell, L. Terstappen, J.W. Uhr, P. Natl. Acad. Sci. USA 95 (1998) 4589. [16] P. Pinzani, B. Salvadori, L. Simi, S. Bianchi, V. Distante, L. Cataliotti, M. Pazzagli, C. Orlando, Hum. Pathol. 37 (2006) 711. [17] G. Vona, C. Beroud, A. Benachi, A. Quenette, J.P. Bonnefont, S. Romana, A. Munnich, M. Vekemans, Y. Dumez, B. Lacour, P. Paterlini-Brechot, Am. J. Pathol. 160 (2002) 51. [18] G. Vona, L. Estepa, C. Beroud, D. Damotte, F. Capron, B. Nalpas, A. Mineur, D. Franco, B. Lacour, S. Pol, C. Brechot, P. Paterlini-Brechot, Hepatology 39 (2004) 792. [19] G. Vona, A. Sabile, M. Louha, V. Sitruk, S. Romana, K. Schutze, F. Capron, D. Franco, M. Pazzagli, M. Vekemans, B. Lacour, C. Brechot, P. PaterliniBrechot, Am. J. Pathol. 156 (2000) 57. [20] H.J. Kahn, A. Presta, L.Y. Yang, J. Blondal, M. Trudeau, L. Lickley, C. Holloway, D.R. McCready, D. Maclean, A. Marks, Breast Cancer Res. Treat. 86 (2004) 237. [21] F. Rl, H.W. Alter, R.M. Walker, S.C. Furman, P.B. Price, Science 178 (1972) 255. [22] P. Rostagno, J.L. Moll, J.C. Bisconte, C. Caldani, Anticancer Res. 17 (1997) 2481. [23] J. El-Ali, S. Gaudet, A. Gunther, P.K. Sorger, K.F. Jensen, Anal. Chem. 77 (2005) 3629. [24] D. Irimia, R.G. Tompkins, M. Toner, Anal. Chem. 76 (2004) 6137. [25] E.A. Schilling, A.E. Kamholz, P. Yager, Anal. Chem. 74 (2002) 1798. [26] D. Di Carlo, C. Ionescu-Zanetti, Y. Zhang, P. Hung, L.P. Lee, Lab. Chip 5 (2005) 171. [27] D. Di Carlo, K.H. Jeong, L.P. Lee, Lab. Chip 3 (2003) 287. [28] J. Kim, S.H. Jang, G.Y. Jia, J.V. Zoval, N.A. Da Silva, M.J. Madou, Lab. Chip 4 (2004) 516. [29] J. Cheng, E.L. Sheldon, L. Wu, M.J. Heller, J.P. O’Connell, Anal. Chem. 70 (1998) 2321. [30] J. Cheng, E.L. Sheldon, L. Wu, A. Uribe, L.O. Gerrue, J. Carrino, M.J. Heller, J.P. O’Connell, Nat. Biotechnol. 16 (1998) 541. [31] S.W. Lee, Y.C. Tai, Sens. Actuat. A-Phys. 73 (1999) 74. [32] F.T. Han, Y. Wang, C.E. Sims, M. Bachman, R.S. Chang, G.P. Li, N.L. Allbritton, Anal. Chem. 75 (2003) 3688. [33] H. Lu, M.A. Schmidt, K.F. Jensen, Lab. Chip 5 (2005) 23. [34] Q. Ramadan, V. Samper, D. Poenar, Z. Liang, C. Yu, T.M. Lim, Sens. Actuat. B-Chem. 113 (2006) 944. [35] S.B. McKenzie, Clinical Laboratory Hematology, Prentice Hall, Upper Saddle River, NJ, 2004. [36] D. Braasch, Physiol. Rev. 51 (1971) 679. [37] S. Chien, S. Usami, Rj. Dellenba, Mi. Gregerse, Am. J. Physiol. 219 (1970) 136. [38] C.E. Lentner, Geigy Scientific Tables, Ciba-Geigy, Basel, Switzerland, 1981. [39] H.M. Shapiro, E.R. Schildkraut, R. Curbelo, C.W. Laird, R.B. Turner, T. Hirschfeld, J. Histochem. Cytochem. 24 (1976) 396. [40] T.Y. Tsong, Biophys. J. 60 (1991) 297.