Integrated microfluidic bioprocessor for single-cell gene expression analysis

Bioprocessor Fabrication.

The fabrication protocol is similar to that used in previous nucleic acid amplification microdevices (22). Briefly, to form the pneumatic manifold wafer, valve seats and actuation channels were photolithographically defined and etched to a depth of 38 μm on a 0.5-mm-thick borofloat 100-mm glass wafer. Valve actuation access holes were drilled, and the manifold was diced into reusable 9-mm × 6-cm strips. Removable polydimethylsiloxane (PDMS) elastomer valves were formed by activating both sides of the 254-μm PDMS membrane with a UV ozone cleaner for 1.5 min to improve PDMS-glass bonding and then sandwiching the membrane between the manifold and the bonded channel wafers (21).

The reactor/channel wafer was fabricated on a 0.5-mm-thick borofloat glass wafer. Fluidic channels for pumping were photolithographically defined on the front side and etched to a depth of 38 μm. Reaction, hold, and capture chambers along with separation channels were photolithographically defined on the back side and etched to a depth of 20 μm. Electrophoresis reservoirs, resistance temperature detection (RTD) access holes, and valve via holes were diamond drilled. To form the RTD wafer, a 0.5-mm-thick borofloat glass wafer sputter deposited with 200 Å of Ti and 2,000 Å of Pt (Ti/Pt; UHV Sputtering) was photolithographically patterned and etched with 90 °C aqua regia to form the 30-μm-wide RTD elements and 300-μm-wide leads. The drilled reactor/channel wafer was aligned and thermally bonded to the RTD wafer by using a programmable vacuum furnace at 655 °C for 6 h.

To form the removable modular heater, a 0.5-mm-thick borofloat glass wafer was sputter-deposited with 2,200 Å of Ti/Pt. Heater leads were formed by electroplating 6 μm of gold onto photolithographically defined areas. Ti/Pt serpentine resistive heater elements connecting the gold leads were formed by anisotropically etching photolithographically exposed Ti/Pt in an ion mill.

Jurkat Cell Preparation.

T lymphocyte Jurkat cells were cultured in 50-mL flasks (Nalge–Nunc International) for 48 h in 10 mL of medium (RPMI medium 1640; Invitrogen) containing 1% penicillin/streptomycin (1% P/S; Invitrogen) and 25 μM Ac₄ManNAz resulting in the display of N-azidoacetylsialic acids on the cell surface glycans (24). For the first 24 h of growth, the cells were cultured with 10% FBS (JR Scientific). The Jurkat cells were washed and incubated in serum-depleted medium containing 25 μM Ac₄ManNAz for an additional 24 h for synchronization. Fresh DNA-functionalized Jurkat cells were prepared 1 h before the analysis. Cells were washed twice with 5 mL of PBS (Ambion) containing 1% FBS and reacted with 125 μM phosphine-modified ssDNA (5′-phos-GTA ACG ATC CAG CTG TCA CT-3′) in 1% FBS/PBS for 1 h at 37 °C (20). The cells were then rinsed 3 times with 5 mL of 1% FBS/PBS solution before introduction into the microfluidic device.

siRNA Treatment.

For gene-silencing studies, 150,000 Jurkat cells were electroporated with 2.5 μg of double-stranded GAPDH siRNA (sense, 5′-GGU CAU CCA UGA CAA CUU UdTdT-3′; Ambion). Cells were suspended in 75 μL of siPORT electroporation buffer (AM1629; Ambion) and a single pulse is performed in a 1-mm cuvette (Bio-Rad) for 250 μs at 250 V. Cells were then grown and prepared in the same manner as described in the Jurkat cell preparation section above. For negative control studies, cells were electroporated with 150 pmol of Cy3-labeled siRNA that does not bind to mRNA.

RT-PCR Mixture.

Multiplex RNA RT-PCR was performed on GAPDH and 18S rRNA transcripts directly from Jurkat cells. A 25-μL RT reaction mixture comprises a Cell-to-cDNA II kit [4 units of Moloney murine leukemia virus (Mo-MLV) reverse transcriptase, 0.4 unit of RNase inhibitor, 0.1 μM dNTPs, 1× RT buffer (Ambion)], 0.08 unit of platinum Taq polymerase (Invitrogen), along with 800 nM forward and reverse primers for the GAPDH gene and 20 nM forward and reverse primer for the 18S rRNA target. The GAPDH forward (5′-AGG GCT GCT TTT AAC TCT GG-3′) and reverse (5′-FAM-TTG ATT TTG GAG GGA TCT CG-3′) primers generate a 200-bp amplicon. The 18S rRNA forward (5′-CGG CTA CCA CAT CCA AGG AAG-3) and reverse (5′-FAM-CGC TCC CAA GAT CCA ACT AC-3′) primers generate a 247-bp amplicon. Controls without RT and without template were performed on the microdevice by removing the Mo-MLV RT and Jurkat cells from the reaction mixture, respectively.

Matrix Synthesis.

A DNA affinity capture gel is synthesized by copolymerizing LPA with 2 5′-acrydite-modified capture oligonucleotides. The affinity capture matrix is synthesized at 4 °C by sparging a 2-mL solution containing 6% wt/vol acrylamide, 1× TTE, and 40 nmol of the 2 acrydite-modified oligonucleotides (IDT) for 2 h with argon followed by the addition of 0.015% wt/vol ammonium persulfate (APS; Fisher Scientific) and tetramethylethylenediamine (TEMED; Fisher Scientific). The affinity capture matrix contains capture probes for GAPDH (5′-Acry-ATC CCA TCA CCA TCT TCC AG-3′, T_M = 54.2; 50 mM monovalent salt, 20 μM) and 18S rRNA (5′-Acry-GCA GCC GCG GTA ATT CCA GC-3′, T_M = 61.9; 50 mM monovalent salt, 20 μM). The GAPDH capture oligonucleotide is complementary to a 20-base sequence in the 200-bp amplicon, 23 bases from the 5′ FAM-labeled terminus. The 18S rRNA capture oligonucleotide is complementary to a 20-base sequence in the 247-bp amplicon, 60 bases from the 5′ FAM-labeled terminus.

Reactor Preparation.

The glass surface is derivatized with polydimethylacrylamide (PDMA) by using a modified Hjerten coating protocol to prevent nonspecific cell adhesion (32). First, the reactors glass surface is deprotonated by incubating with 1 M NaOH for 1 h. The NaOH solution is replaced with a 0.6% (vol/vol) (γ-methacryloxypropyl)trimethoxysilane solution (γ, Sigma) in 3.5 pH H₂O. The bifunctional γ-solution prepares the glass surface for acrylamide polymer nucleation. During γ-solution incubation, 250 μL of dimethylacrylamide is dissolved in 4.75 mL of H₂O and sparged with Ar for 1 h. After Ar sparging, 100 μL of isopropyl alcohol (IPA), 20 μL of TEMED, and 25 μL of 10% (vol/vol) APS were sequentially added to the acrylamide solution to form linear PDMA. The γ-solution is removed from the channel, and PDMA solution incubates in the channel for 1 h. The channel is then rinsed and dried with acetonitrile.

Next, the photolithographically defined 25-μm × 25-μm gold pad in the center of the reaction chamber is functionalized with ssDNA by incubating for 1 h with Tris(2-carboxyethyl)phosphine (TCEP, 200 μM; Invitrogen) deprotected thiol-DNA (5′-thiol-AGT GAC AGC TGG ATC GTT AC-3′, 20 μM). The chamber is then rinsed and dried to remove unbound DNA.

Gel Loading Sequence.

The device is prepared for affinity capture and separation by treating the separation channels, hold chambers, and capture chambers with a dynamic coating diluted in methanol for 1 min (1:1; DEH-100; The Gel Company) to suppress electroosmotic flow. Multiplex affinity capture matrix (20 μM, 6%, yellow) is loaded from each cathode (C) reservoir up to the separation channel cross at room temperature. Separation matrix (red) is then loaded from the central anode (A) past the capture chamber to the sample load cross. With all of the valves opened, the rest of the system is hydrated by adding 3 μL of RNase-free water at the sample ports and applying a vacuum at the waste (W) reservoirs. The microdevice is then placed on a 44 °C temperature-controlled stage. An additional 2 μL of water is flushed through the system to remove thermally expanded gel from the sample load cross.

Bioprocessor Operation.

As depicted in Fig. 3, the microdevice operation begins by preparing the reaction chamber for cell capture. Cell modified to contain ssDNA on their surface were suspended in the reaction mixture and drawn into the reaction by vacuum. When a cell flows over the gold cell capture pad, DNA hybridization occurs between the ssDNA on the surface of the cell and the complementary ssDNA immobilized on the gold pad. To maximize capture efficiency, DNA-mediated cell capture was performed for 15 min. The uncaptured cells were washed out of the system and removed at the waste port, and the valves were closed for thermal cycling. During this study, all 4 reactors were used in parallel. This enabled simultaneous analysis from 4 individual cells. After capture, the morphology of each cell is recorded to ensure that it has remained viable. The total data collection time for the single-cell studies was 1 month.

Freeze–thaw lysis and 1-step RT-PCR thermal cycling were performed in a single 200-nL reactor. A piece of dry ice was placed over the reaction chambers of all 4 reactors for 30 s to freeze–thaw lyse the captured Jurkat cell. Freeze–thaw lysis was used in this 1-step RT-PCR to prevent early unwanted activation of the Hot Start Taq polymerase, to prevent denaturation of the reverse transcriptase enzyme, and to minimize RNA degradation by RNases (33, 34). Next, a linear 15-min cDNA synthesis from the cells' RNA was performed at 42 °C by using primers complementary to the RNA transcripts of interest (GAPDH and 18S rRNA). After cDNA synthesis, the Mo-MLV RT was denatured, and the platinum Taq polymerase was activated at 95 °C for 60 s followed by 30 cycles of PCR at 95 °C for 5 s, 47 °C for 20 s, and 72 °C for 25 s. Because of the rapid heating and cooling rates (>15 °C s⁻¹), each cycle of PCR is completed in 50 s, and the total reaction time is 46 min.

After thermal cycling, affinity capture, purification and concentration of the products of interest were performed. The reactor contents were pumped into the hold chamber by using a 5-step pump cycle. A 350-ms actuation was used with each step, resulting in a 30-nL stroke volume. A 23-s delay was used between each pump cycle to allow sufficient time for the analyte to migrate into the capture region and to prevent analyte accumulation in the hold chamber. A constant 100-V/cm field between the waste (W) and cathode (C) reservoirs electrophoretically drives the analyte toward the capture chamber. Analytes complementary to the capture probe were hybridized at the entrance of the capture chamber, creating a sample plug. The electric field between the waste and cathode was maintained until residual PCR reactants (excess primer, salts, and buffer) were washed into the cathode reservoir thus resulting in a purified amplicon sample plug. Thirty pump cycles were used resulting in a total capture and wash time of 12.2 min. After the capture process was completed, the temperature of the entire device was raised to 80 °C to thermally release the captured DNA fragments thermally from the affinity capture gel, and the sample was separated with a field of 150 V/cm between the cathode and the anode. The electrophoretically separated FAM-labeled products from all 4 lanes were detected by using laser-induced fluorescence with the Berkeley rotary confocal scanner (27). The entire capture and release process was performed on a temperature-controlled stage on the scanner to prevent thermal gradients.