cDNA synthesis and analysis in microfluidic droplets

(1)

Master Thesis in Nanobiotechnology

cDNA synthesis and analysis in microfluidic

droplets

Lovisa Söderberg

Royal institute of Technology (KTH)

Division of Nanobiotechnology

Examiner: Helene Andersson-‐Svahn

Supervisor: Håkan Jönsson

(2)

Abstract

In this master thesis a technique is developed to perform thousands of parallel cDNA synthesis reactions compartmentalized in picoliter reaction vessels. The approach of using droplet microfluidics will enable a high throughput technique for gene expression analysis of isolated single cells. High throughput is required to perform gene expression analysis on a large number of cells, which is needed to be able to identify heterogeneity within a cell population.

Significant cDNA yields were achieved from picoliter samples of extracted RNA at concentrations corresponding to the RNA content of one cell per droplet. cDNA

syntheses were also performed from whole lung cancer cells, lysed with a temperature step, with added extracted RNA. This approach gave a higher cDNA yield then samples only containing extracted RNA in the microfluidic droplets, which indicate that the approach will work also from whole cells in microfluidic droplets. To perform the cDNA synthesis in microfluidic droplets the reverse transcription enzyme was mixed with RNA sample before encapsulation in droplets. The droplets were incubated for cDNA

(3)

1. Introduction

1.1 Project background

Miniaturization of analysis and reactions to microscale have been more and more applied in biology during recent years due to advantages such as; smaller sample sizes, less reagents needed, faster reaction times, and better process control compared to reactions in macroscale (Lindström et al, 2011). Droplet microfluidics is a high

throughput technique (>1000reactions/second) used to miniaturize reaction volumes and generate droplets that contains all the components needed for a reaction. Each individual droplet functions as a microreactor divided from each other by an oil barrier. (Taly et al, 2007)

To analyze gene expression, the most common way is to study transcribed mRNA. This since the gene corresponding to a protein often includes exons of non-‐coding areas, which need to be removed by splicing to get the final sequence (Black, 2003). RNA is single stranded and easily degraded a more efficient and reproducible technique, then to study the RNA content directly, is to do a reverse transcription of RNA to cDNA. cDNA is more stable and can be amplified in a PCR reaction. This has long been a standard procedure to study gene expression in bulk samples and this gives a good image of the average state in a population of cells. (Nolan et al, 2006)

However the gene expression have been shown to vary within a cell population (Elowitz et al 2002). Heterogeneity in gene expression between cells leads to differences in phenotype, variation in cell division and expression of receptors, and efficiency of drugs towards the cell (Spudich et al, 1976)(Raser et al, 2005). It is therefore interesting to be able to study gene expression on a single-‐cell level.

(5)

To look at the gene expression of a single cell and still get a complete image of the expression in a larger population a lot of cells has to be analyzed so that the distribution of subpopulations can be determined and so that wrong conclusions are not drawn because only rare cell are studied that does not correspond to the average cell

population (Lindström et al, 2010). Also to be able to sort out a rare cell type in a cell sample requires that many cells can be discarded in an effective manner. An extreme example would be circulating tumor cells in a blood sample where only one or a few tumor cells will be present per 109_{blood
cells
(Nagrath
et
al,
2007).
Therefore
a
high-‐} throughput analysis technique is needed so that many reactions can be performed in parallel under a short period of time.

1.2 Aim

The aim of the project is to perform a reverse transcription and synthesize cDNA from single cells. To be able to separate single cells in individual reaction chambers the platform of droplet microfluidics will be applied. The droplet microfluidic approach gives a sensitive analysis under high-‐throughput conditions where a single cell can be encapsulated in a 30µm droplet that will function as a reaction vessel. The cDNA syntheses will initially by performed from extracted RNA, from a lung cancer cell line. When this approach gives a significant yield of cDNA in droplets the project will move on to its next stage, which will be to synthesis cDNA directly from whole lung cancer cells. The produced cDNA will be analyzed in qPCR.

(6)

2. Theoretical background

2.1 Microfluidics

In microfluidics micro-‐channel-‐systems are constructed on chips with patterns in micro-‐ scale. External pumps are used to actuate pressure driven flows inside the channels. These systems are used to perform experiments with small reagent volumes, precise control and the possibility of parallel reactions. Laminar flows are achieved inside the channels, which gives a controlled flow without turbulence. (Beebe et al, 2002) The reduction in sample volumes allows for more reactions to be performed when only small amounts of sample can be obtained or is expensive. Reduction in regent volumes also reduces the cost per reaction. And by doing the reactions on a micro-‐scale instead of macro-‐scale the diffusion distances becomes shorter which increase the reaction rates (Ahmed et al, 2006).

2.1.1 Chip manufacturing

The microfluidic channel system are constructed on chips and are often made out of PDMS (poly(dimethylsiloxane)). PDMS is a polymeric material with many advantages compared to other materials used in microfluidic chips, such as glass and silicon. PDMS is cheaper, less fragile and the devices are faster to manufacture. (Duffy et. al, 1998) Other properties of PDMS are that it is; optically transparent, nontoxic for mammalian cells, it can be irreversibly sealed after oxygen plasma treatment to, for example glass. PDMS is an elastomer which means that it can be shaped after a non planar pattern and after curing it will release from the structure without damaging the mold or change the formed shape. This property makes it possible to produced micro-‐sized pattern in a fast and reproducible manner for prototyping. (McDonald et. al, 2000)

(7)

procedure a master is formed with the negative pattern of the wanted channel system. This master will be used as a mold for the PDMS devices and can be used multiple times. (McDonald et al, 2000)

The PDMS molding is performed by mixing the polymer and cross-‐linking agent, which is then pored onto the silcon wafer and cured in an oven. After curing, the PDMS

structure can be peeled of the wafer. (Fiorini et al, 2005) To get complete channels, the PDMS device is sealed by binding it to a glass surface. The PDMS device and a glass slide are exposed to oxygen plasma and afterwards they are pressed together. This will form an irreversible waterproof bound. (McDonald et al, 2000)

2.2 Droplet Microfluidics

In droplet microfluidics two liquid phases are utilized to form an emulsion, oil droplets in water or more common water droplets in oil. The droplets can be varied in size from several nanoliters down to femtoliters. The droplets can be generated at a frequency of 1-‐10kHz which makes it possible perform many reactions in small volumes at a very short time. (Therberge et al, 2010)

In chemical reactions the droplet platform minimizes the volume of reagents needed. Compartmentalization to small reaction volumes also gives shorter diffusion distances and faster mass transfer. Reactions in micro volumes have been proven to increase reaction rates (Ahmed et al, 2006).

The phase inside the droplets are called the dispersed phase and the continuous phase is the liquid outside the droplets. The droplet dynamics in a two-‐phase device is

determined by the dimensionless capillary number, Ca which depends on the viscosity, η, and velocity of the disperse phase, ν, and the interfacial tension, γ, between the continuous and the discrete phase, in the following relation: Ca = ην/γ. When the Ca number increases, droplets are generated. (The et al, 2007)

(8)

phase travels through a narrow part of the channel. The sheer stress between the continuous and disperse phase will elongate the flow of disperse phase into a thin line and the reduction of surface tension by lowering the surface to volume ratio will make the disperse phase break up into droplets (Thorsen et al, 2001). The droplet size is determined by the nozzle width and increasing the flow rate of the continuous phase can generate a decreased droplet size. (Anna et al, 2003)

Figure 1: a) Droplet generation in a flow-‐focusing device. The water phase is in the middle channel and the oil phase in the two side channels. The water phase will break up into droplets after the channel crossing. b) Generated monodisperse droplets stabilized with a surfactant to prevent coalescence.

With these techniques droplets with high monodispersity are formed. Reports have been made with generated droplets with a coefficient of variance less then 2% (Nisisako et al, 2006). This property is required for the droplet microfluidic platform to be used in high-‐throughput analysis. To be able to have a high degree of parallelization the droplets need to be the same size so that all chemical regents will be at the same concentration in each droplet. It is important for chemical reactions and biological applications that all chemical reagents inside the droplet are at the same concentration so that each droplet can be seen as a single experiment to be compared to the rest of the droplet population.

2.2.1 Oils and Surfactants

(9)

with tightly packed emulsion without coalescing with neighboring droplets. (Baret, 2012)

Some oils that are used in droplet microfluidics, as the continuous phase, are

fluorocarbon oils such as FC-‐40, FC-‐77, hydrocarbon oils and mineral oils. In this project a hydrofluoroether oil, HFE was used which is a fluorocarbon oil. (3M) Different oils and surfactants are combined, depending on the molecules inside the droplets to avoid interaction with the droplet interface and leakage into the continuous phase.

The surfactant used in this project is a block copolymer consisting of a PEG and two PFPE chains. The PFPE part is soluble in the HFE oil and the PEG is soluble in the water phase and prevents interaction with biological molecules inside the droplet and the interface. This surfactant stabilizes droplets almost directly after generation and while they are closely packed. Cell proliferation is possible inside the droplets, which shows a good biocompatibility of the droplets. (Holtze et al. 2008)

2.2.2 Droplet manipulation

In order to use this platform for applications with high parallelization the droplets need to be individually controlled and manipulated. For manipulation of droplets there are two types of modules, passive and active. A passive manipulation is controlled by the channel geometry and affects all droplets. Active manipulations are controlled by for example electrodes or valves, that often require external power, and individual droplets can be selected for the manipulation while others are not affected. (The et al, 2007) Splitting of droplets is accomplished by introducing a T-‐junction on the chip. Droplets in a compressed state will split into two daughter droplets when the channel divides in a T-‐ junction. (Links et al, 2004)

(10)

liquid can be added to the droplet. When the droplet moves out of the electric field the droplets becomes stable again. The volume injected depends on the injection pressure and flow rate of the continuous phase. (Abate et. al, 2010)

Droplets can also be actively sorted to select specific droplets out of a larger population. In one technique dielectrophoretic forces are applied on droplets that will pull them in different directions and go into selected channels. The dielectrophoretic force is applied by an electric field generated by electrodes on the chip. For this technique the droplets does not need to be charged. (Ahn et. al, 2006)

By incorporating a fluorescent microscope in the experimental setup and adding a fluorescent molecule inside the droplets one can do on-‐chip detection. Quantification of the molecules binding to the fluorescent molecules is an often-‐used approach for

analysis. (Therberge et. al, 2010) Very sensitive detections have been made down to a few proteins expressed on a cell surface encapsulated in a droplet using enzymatic amplification (Joensson et. al, 2009) or a single DNA molecule (Srisa-‐Art et al, 2009). Each of these operations can be seen as a module in the workflow and a separate experimental step. So by placing these modules after each other an entire experiment can be built on a single chip or be divided over several different chips. The emulsion can also be collected and stored off chip before being reinjected on chip. (Kintses et. al, 2010)

2.2.3 Encapsulation of biological molecules in droplets

(11)

For single cell analysis the encapsulation of a single cell in each droplets enables a linkage between genotype and phenotype. Even after the cell is lysed the cell content will be contained in an individual reaction vessel. (Taly et al, 2007) The interest in individually studying a large number of single cells at a time comes from the heterogeneity between cells in the same population (Lindström et al, 2010). By

gathering information from each individual cell, instead of measuring an average value for a large population of cells, a better understanding of the function of a cell can be achieved and smaller subpopulation can be detected.

2.3 Droplets as a reaction vessel

In previous work in droplets microfluidics several applications have been developed. In this project I will mainly focus on applications where the generated droplets are used as reaction vessel.

Directed evolution is a selection of nucleotides or protein with some wanted properties generated by induced mutations. Agresti et al shows the possibility of droplet based directed evolution that reduces both the analysis time and cost. In this study a mutant of horseradish peroxidase enzyme was selected that has a 10 times higher activity then the initial population. A library was created that were transformed into yeast cells and then encapsulated in droplets. Droplets sorting with fluorescent detection were used to select the best mutants with the highest activity. A screen of a 107_{variants
take
only
a
few} hours effectively with this technique and the reactions require only a fraction of the reagents needed for an equivalent screen with traditional techniques. (Agresti et al, 2010)

(12)

exposed to thermo cycling as in a traditional PCR. At specific locations a fluorescent measurement is performed to get a real-‐time quantification of the amplification process for each droplet. (Kiss, 2008). It is also possible to simplify the reaction by performing an isotherm DNA amplification (Mazutis et al, 2009). Beer et al, 2008, has performed reverse transcription-‐PCR in droplets where RNA is first transcribed to DNA before amplification. Instruments that perform PCR in microfluidic droplets are commercially available by companies such as Bio-‐Rad and RainDance Technologies. (Baker, 2012) Mary et al performed a study of cell-‐to-‐cell gene expression variations in droplets by performing a reverse transcription PCR, RT-‐PCR. Droplets were immobilized on chip containing 0-‐3 cells and RT-‐PCR reaction mixture and exposed to thermal cycling. In the study 30 droplets were analyzed containing MDCK cells and it was possible to detect differences in gene expression in the different cells. (Mary et al, 2011) However by immobilizing the droplets on chip the ability to perform high-‐throughput analysis is lost. Detection of DNA in droplets without PCR has also been performed down to a single molecule per droplet. By binding several fluorescent dye molecule to a DNA molecule a single molecule could be detected. (Srisa-‐Art et al, 2009)

2.4 qPCR

To quantify cDNA one of the most sensitive techniques is quantitative polymerase-‐ chain-‐reactions, qPCR. The difference between a traditional PCR that amplifies DNA fragments is that the qPCR reaction mixture contains a fluorescent dye that binds to double stranded DNA. After each thermo cycle a plate read is performed that reports the increase in fluorescent molecules that have bound to amplified DNA. Each reactions target molecule quantity is determined by at which cycle the fluorescent signal increases above a specific threshold, Ct. The higher quantity of target molecules in the starting reaction the lower the Ct value. All Ct values are then compared to a series of standard samples with know target molecule concentration. (Nolan et al, 2006)

2.5 Cell lysis technique

A big challenge that is always present in droplet microfluidics is the fact that the

(13)

droplet approach once a reagent is added it cannot be removed. In this project it has affected the choice of cell lysis technique. To be able to perform a reverse transcription from whole cells requires that the cell membrane is permeable so that the reverse transcription enzyme can reach its target.

Traditional chemical lysis agents at high concentration are difficult to use since some inhibits the reverse transcription in later steps (Bengtsson, 2008). Electric fields can also be used to lyse cells (Lee et al, 1999) but this requires that electrodes be introduced on the chip. That would give a more expensive and complicated chip manufacturing. Cell lysis induced by elevated temperatures is a technique that has previously been used by Liu et al, 2004, to lyse cell on-‐chip before PCR. A temperature lysis leaves no chemical traces behind and no extra components are needed on chip, which makes it the choice for this application.

2.6 cDNA synthesis

The reverse transcription enzyme was first discovered by Baltimore, 1970. The enzyme was able to synthesis DNA from an RNA template. This technique can be used to study the expressed genes in cells by studying the already transcribed RNA.

When performing cDNA synthesis in bulk the RNA content from all cells, in the sample, are mixed when the cells are lysed. By encapsulating the cells in microfluidic droplets before cell lysis the RNA content from different cells will be kept separate, see figure 2.

(14)

3. Material and Methods

3.1 Cell cultivation

The A549 adherent lung cancer cell were incubated in DMEM (Dulbecco's Modified Eagle's Medium D6429, Sigma) cell medium containing 10% FBS (fetal bovine serum, Sigma) and 1% antibiotics (Antibiotic, Antimycotic solution, Sigma). The cells were incubated in petri dishes at 37°C.

The cultures were split approximately every 60-‐80 hours. The cells were first washed with PBS and detached from the petri dish using 1000µl 1xTrypsine. The cells were then collected in a 15 ml centrifuge tube and diluted in culture medium to inactivate the trypsin. The cell culture was centrifuge to a pellet at 160 rcf for 3 min. The supernatant containing trypsin was removed and replaced with 1000µl fresh cell culture medium. The cell culture was transferred to a new petri dish with warm cell culture medium and incubated at 37°C.

3.2 RNA extraction

To extract RNA from the cultivated cell the RNeasy Midi columns kit for RNA extraction from Qiagen were used. The cells were prepared from their cultivation in petri dishes. The cells were Trypsinated to detach from the surface and then collected in a 15ml centrifuge tube and centrifuged for 3 min at 160 rcf. The supernatant was removed and the cells were washed with PBS.

RNA was extracted at two different occasions. The first time a starting material of 5.1 x 106_{cells
were
used
and
the
second
time
1.65
x
10}7_cells.

To start the RNA extraction the cells were lysed with 2,0 ml of RLT buffer containing 0,1% β-‐Mercaptoethanol. The cell lysate was homogenized by vortex for 10 s and then passed through a 20-‐gauge needle fitted to an RNase free syringe 10 times.

2 ml of 70% ethanol were added to the homogenized lysate. The sample was transferred to the RNeasy Midi column placed in a 15 ml centrifuge tube and centrifuged for 5 min at 3500 rcf. After each centrifugation step the flow-‐through from the column was

(15)

ml RPE buffer and centrifuged for 5min to dry the column to avoid any residuals of ethanol that could interfere with the results.

To elute the RNA from the column 150 µl of RNase free water were added and the tube was centrifuged for 5min at 3500 rcf. This step was then repeated with another 150 µl RNase free water.

3.2.1 RNA Quality control

The quantity of RNA was first checked in a NanoDrop 1000 instrument. The purity, degradation and quantity of the extracted RNA were then determined in an Agilent Bioanalyzer 2100 (Agilent technologies, Germany) using a 6000 nano RNA chip, following the protocol of the manufacturer.

3.3 cDNA synthesis in bulk

After extraction of RNA from A549 lung cancer cells cDNA was synthesized. To test the protocol and to have a reference yield, for the synthesis in droplets, the synthesis were first conducted in bulk using the SuperScript III First-‐Strand Synthesis system for RT-‐ PCR (Invitrogen). The protocol was tested both with random hexamers and Oligo(dT)20 as primers and with different amounts of RNA as starting material diluted with DEPC-‐ treated water. Controls without primers and SuperScript III RT were tested.

8 µl of RNA sample were mixed with 1µl of dNTPs and 1 µl of random hexamers or Oligo(dT)20. The mixture was incubated at 65°C for 5 min and then placed on ice. The cDNA synthesis mix was prepared with 2 volumes 5X RT buffer, 4 volumes 25 mM

MgCl2, 2 volumes 0,1 M DTT, 1 volume RNaseOUT (40 U/µl) and 1 volume of SuperScript III RT (200 U/µl). 10 µl of cDNA synthesis mix were added to each sample. The samples were then mixed and briefly centrifuged before incubation. The samples with random hexamers were incubated for 10 min at 25°C and the 50 min at 50°C. The samples with Oligo(dT)20 were incubated for 50 min at 50°C. All reactions were terminated at 85°C for 5 min and then placed on ice before being stored at -‐20°C.

3.4 cDNA synthesis in droplets

(16)

10mM dNTP mix. Nuclease free water was added to a total volume of 20 µl. The mixture was incubated at 65°C for 5 min and then placed on ice for at least 1 min. The synthesis mix consisted of 4 µl 5X RT buffer, 8 µl 25 mM MgCl2, 4 µl 0.1 M DTT, 2µl RNaseOUT (40 U/µl) and 2 µl SuperScript III RT (200 U/µl). The synthesis mixture where also placed on ice before use.

The droplets experimental setup includes an inverted microscope, Olympus IX51, a Nemesys pump system for four syringes and a camera operated by computer software. Two droplet setups were used during the project, see figure 3a. In the first setup, Mixed, the RNA sample and the synthesis mix were mixed before introduced to the chip. In the second setup, T-‐junc, the RNA sample and the synthesis mix were mixed in a T-‐junction in the tubing before entering the chip. When the T-‐junction setup is used only the droplets containing both RNA sample and synthesis mixture will have a possible cDNA synthesis. It is therefore important to have synchronized flows, see figure 3b, with a large effective volume where both RNA sample and synthesis mix is present in each droplet.

The oil phase used was HFE oil (3M) with 0,5% surfactant (Raindance Technologies). The emulsion was collected in a syringe with HFE oil with 0,5% surfactant. The water phase inlet; RNA sample and synthesis mix, were connected with syringes filled with HFE oil without surfactant. The RNA sample and the synthesis mixture were loaded using a negative flow on the syringe pumps so that the samples were loaded in the tubing connecting the syringe to the chip. Long enough tubings were used so that all of the sample could fit inside the tubing without entering the syringe. In the T-‐junc setup, see figure 3, the RNA sample and synthesis mix were loaded into the tubings before being fitted to a T-‐junction that were then connected to the chips water inlet.

(17)

Figure 3: a) Experimental droplet setup, microchannels on chip in black and tubings connected to syringes in blue. The top setup, Mixed, shows the setup where the RNA sample and the synthesis is mixed on before hand. The bottom setup, T-‐junc, shows the setup where the RNA sample and synthesis mix is mixed in a T-‐junction before entering the chip. b) Possible mixing of the flows A and B in a T-‐junction. With a synchronized flow the maximum effective volume is achieved with both regent A and B present in each droplet. With an

unsynchronized flow the effective volume is decreased.

The chip was first filled with oil from the oil syringe. The collect syringe also had a low flow to avoid that any debris went into the collect. After the chip was filled, the pumps connected to the water phase syringes were turned on. When the water phase reached the chip the flows were adjusted as follows: oil syringe: 1000 µl/h, water phase syringe: 50 µl/h each, or when only one syringe were used for the water phase: 100 µl/h, and the collect syringe: 200 µl/h. When the droplet generation became stable, the collection of emulsion started by applying a stepwise negative flow on the collect syringe until the droplets went into the collect, usually around -‐800 µl/h.

After droplets were generated the collect syringe was placed in the oven and incubated at 50°C for 50 min and then placed on ice. The emulsion was stored at -‐20°C. Before qPCR the sample were thawed and the emulsion was broken with a droplet destabilizer (RainDance Technologies).

3.5 qPCR

For quantification of the cDNA product quantitative chain reaction, qPCR, was used. b)

(18)

3.5.1 Selection of primer sequences

The primer pairs for the qPCR were provided by Afshin Ahmadian, KTH. The selected primer pairs are from three housekeeping genes (HKG). The HKG are a good choice for this application since they are expressed in all cells. The HKGs selected were, Homo sapiens beta-‐2-‐microglobulin (B2M), Homo sapiens hydroxymethylbilane synthase (HMBS) Homo sapiens hypoxanthine phosphoribosyltransferase 1 (HPRT1). B2M encodes for a serum protein that is expressed on almost all nucleated cells. HMBS encodes for a protein in the heme biosynthetic pathway that is a member of the

hydroxymethylbilane synthase super family. HPRT1 encodes for a transferase that has an important role in the production of purine nucleotides. The B2M primer pair gave the shortest amplified sequences, amplicon, of 293 bp. (NCBI, 2012) Shorter fragments often gives better results in the qPCR so this primer pair will initially be used.

Tabell 1: Primers used in qPCR from three different housekeeping genes; B2M, HMBS. HPRT1.

Primer Gene Forward primer Reverse primer

1 B2M CAGCGTACTCCAAAGATTCAG CATGTCTCGATCCCACTTAAC

2 HMBS CAGTTTGAAATCATTGCTATGTC AAGCCGGGTGTTGAGGTTTC

3 HPRT1 AGTGATGATGAACCAGGTTATG GACCATCTTTGGATTATACTG

3.5.2 qPCR program optimization

For qPCR Maxima SYBR Green/Fluorescein qPCR Master Mix (Fermentas) was used. A qPCR program was optimized for the B2M primer pair as follows; 95°C for 10 min, 95°C for 15 s, 59°C for 30s and 72°C for 30 s. These three steps were repeated 60 times. The reactions were terminated at 95°C for 10 s. After each run a melt curve analysis was performed from 65°C to 95°C with an increase of 0,5°C every 5 s. This gave the melting temperature of the product in each well to ensure purity of the product.

The total volume in each well was 20µl with 10µl Maxima SYBR Green/Fluorescein qPCR Master Mix (2X), 0,5 µl each of forward and reverse primer to a total

(19)

In each qPCR run a series of samples with known cDNA concentration were added to construct a standard curve to which all samples were normalized. 9 different

concentrations were used in construction of the standard curve from 108_{molecules/µl} to 100_{molecules/µl
with
a
dilution
of
10
times
between
each
step.
Each
concentration} was repeated in two wells.

3.5.3 Gel electrophoresis to control purity

Gel electrophoresis was used to control that only one length of DNA fragments where amplified in the qPCR reaction. In the electrophoresis 1% Agarose gel stained with etidium bromide was used. In each run a GeneRuler 100 bp ladder plus and MassRuler loading dye x6 from Fermentas were used to be able to distinguish the length of the fragments and to be sure that the sample did not travel out of the chip. The

electrophoresis was run for 1,5 h at a voltage of 3 V/cm.

3.6 Cell lysis

To determine at which temperature cells were lysed two culture plates with

approximately 2x106_{adherent
lung
cancer
cells
were
used.
The
cell
culture
medium
was} aspirated and 2 ml of 0,5 mM Calcein AM Green (Invitrogen) added as a live stain and incubated for 37° C. The cells were trypsinized as for a cell passage and transferred to a 15 ml Falcon tube and centrifuged to a pellet. The pellet was washed with 1xPBS and then diluted in 3,5 ml 1xPBS divided in seven eppendorf tubes. The positive control was kept at room temperature and the negative control was placed in 70% ethanol in -‐20° C for 2 h. The other samples were incubated for 10 min at 55° C, 60° C, 65° C, 70° C and one sample were incubated for 50 min at 50° C. After incubation each sample were centrifuged and the supernatant removed. 0.5 ml of 3 µM Propidium Iodide (Molecular Probes, Invitrogen) were added and the samples were incubated for 15 min at room temperature. The cells were then studied in fluorescent microscope and the live and dead cells were counted using a counting chamber.

3.7 cDNA synthesis from cells

To see if the cDNA synthesis could be performed from whole cells, six samples were

(20)

where the cDNA synthesis enzyme was added before cell lysis and one where the

enzyme was added after cell lysis. One sample was diluted in 1xPBS and the enzyme was added before cell lysis. Three other samples were prepared in the same way but 3,2 µg RNA was added to each sample.

For controls a sample with only RNA and one sample with nuclease-‐free water was included in the cDNA synthesis. In the qPCR; controls with only lysed cells and controls with cDNA diluted in lysed cells were added.

Reactions with different cell concentrations were also performed. The highest concentration included approximately 160 000 cells, corresponding to a cell

concentration of 8000 cells/µl. For each sample the cell concentration was diluted five times and five samples were made.

The cDNA syntheses mix contained 2 volumes 10xRT buffer, 4 volumes 25 mM MgCl2, 2 volumes 0,1 M DTP, 1 volume RNaseOUT and 1 volume SuperScript III RT (200 U/µl) (Invitrogen). 10 µl cDNA synthesis mix was added to each sample.

(21)

4. Results

4.1 Extraction of RNA from A549 lung cancer cell line

RNA was extracted from A549 lung cancer cells on two different occasions. After each extraction the quality and degradation of the RNA were controlled in an Agilent Bioanalyzer 2100. The degradation was evaluated by calculating the RNA integrity number, RIN, which is performed by the Agilent software. The calculation is based on the amplitude ratio of the 18s and the 28s peak, that is given by the ribosomal RNA, and the area under these peaks compared to the area under the base line and also the flatness of the baseline (Schroeder et al, 2006).

For the first RNA extraction a good starting material with little degradation was

obtained, see figure 4. For this run I did not get a quantity of RNA or RIN number but by comparison to other graphs I evaluated the RIN to be between 9 and 10. I also used a NanoDrop 1000 instrument to get the quantitative measurement of about 400 ng RNA/µl.

Figure 4: Quality control of RNA from the first RNA extraction in Agilent Bioanalyzer. RNA concentration measured to 400 ng RNA/µl. The two distinct peaks show that the RNA has a good quality without degradation.

(22)

Figure 5: Quality control of RNA from the second RNA extraction in Agilent Bioanalyzer. RNA concentration measured to 720 ng RNA/µl in a 5x dilution with RIN is determined to 10.

4.2 qPCR serial dilution

Quantification of the amount of produced cDNA was done with qPCR. In each run a serial dilution with known concentrations of PCR fragments, of the specific gene, were

included to calculate a standard curve. The standard curve showed a linear behavior with only small deviations between the replicates.

Three replicates were made of each sample in the qPCR.

4.2.1 Gel electrophoresis

For the first qPCR runs the purity of the product was controlled by gel electrophoresis. This gives the length of the amplified fragments. When only one band is visible on the gel it indicates that only fragments of the same length has been amplified. In figure 6 the visible band corresponds to the length given by the B2M primers used. Since only one band is visible for each lane a pure product is determined.

(23)

4.3 cDNA synthesis in bulk from RNA

From the experiment in bulk with different amount of RNA as starting material with Oligo (dT)20 as primers an increasing amount of cDNA were produced with increasing amount of RNA as staring material see figure 7. For the samples with random hexamers as primers the same relation could be seen for the 320 ng and 32 ng RNA as starting material but the highest amount of RNA gave a lower yield. Overall the samples with Oligo (dT)20 as primers gave a 10 times higher yield then the samples with random hexamers. In the controls with no primers for the cDNA synthesis there was a cDNA product but a drastic increase in product could be seen when primers were used. Also in the control with RNA where no cDNA synthesis had been performed there were DNA fragments, but again a significant difference can be seen between this control and the samples where cDNA synthesis had been performed.

Figure 7: cDNA synthesis with either random hexamers or Oligo (dT)20 as primers. The reactions also have

different amount of RNA as starting material 3,2 µg, 320 ng or 32 ng. Controls without primers in cDNA synthesis and a RNA sample where no cDNA synthesis has been performed. Oligo (dT)20 primers gives a

higher product then random hexamers. Overall a higher product yield is seen with more RNA as starting material.

To try to explain the lower yield with the highest amounts of RNA with random

hexamers the experiment was repeated. The reactions with the same color on the bars in figure 8 were performed at the same time. The samples with the 3,2 µg RNA gave a varied amount of cDNA between the replicates but they all gave a lower yield then expected. The limiting step in these reactions is not the amount of RNA. Most likely is that the reactions are limited either by the availability of enzyme or primers in the

1,00E+00 1,00E+01 1,00E+02 1,00E+03 1,00E+04 1,00E+05 1,00E+06 1,00E+07 1,00E+08

Hex 3,2µg Hex 320ng Hex 32ng Oligo 3,2µg Oligo 320ng Oligo 32ng No primer RNA sample

cDNA synthesis with different primers and

different amount of RNA

(24)

reaction mixture. This will be further discussed in the discussion part of the report. The reactions with lower amount of RNA gave a more even cDNA yield.

Figure 8: cDNA synthesis with random hexamers as primers and different amount of RNA as starting material, 3,2 µg, 320 ng or 32 ng. Bars with the same color are reactions performed at the same occasion. Almost the same amount of cDNA is synthesized with 3,2 µg RNA and 320 ng RNA indicates that the reactions are not limited by the amount of RNA for higher RNA concentrations. 32 ng RNA as starting materials gives a lower yield then the two others.

4.4 cDNA synthesis in droplets from RNA

In the first droplet experiment the RNA sample and the cDNA synthesis mix were mixed in a tube before being introduced to the chip, (Mixed). And in the second droplet

experiment the RNA sample and the cDNA synthesis were combined in a T-‐junction right before entering the chip, (T-‐junc), see figure 3a for experimental setup. There are slightly more cDNA produced in the samples conducted in droplets compared to the average value of the 3,2 µg RNA with random hexamers as primers which were also used in the droplet samples, see figure 9.

A higher cDNA yield can also be seen in Mixed, the droplet experiment where the RNA sample and the cDNA synthesis mix is mixed before it is introduced on the chip then the sample where the mixing is done in a T-‐junction, T-‐junc, see figure 3a. This is probably due to that the two flows will not be perfectly synchronized in the setup with the T-‐ junction. Some droplets will only contain RNA samples and some will only contain cDNA synthesis mix. All experiments in microfluidic droplets gives a comparable or slightly higher cDNA yield than the average yield in bulk with the same RNA concentration.

1 10 100 1000 10000 100000 1000000

Hex 3,2µg Hex 320ng Hex 32ng

cDNA synthesis with random hexamers

(25)

Figure 9: cDNA synthesis in droplets compared with bulk samples and controls. The reactions performed in droplets with random hexamers as primers gives a lower yield then the reactions performed in bulk with Oligo (dT)20 as primers. The reactions in droplets gives a cDNA yield comparable or slightly higher than the

average yield in bulk. Also a significant cDNA product compared to the RNA sample where no cDNA synthesis has been performed.

4.4.1 Reproducibility of droplet experiment

The cDNA synthesis in droplets, when the mixing of the cDNA synthesis mix and the RNA sample was conducted on the chip in a T-‐junction, see figure 3a, were repeated two more times to control the reproducibility of the experiment. All experiments gave a significant product compared to the negative control where no cDNA synthesis was conducted, see figure 10. There is some variation in the cDNA yield from the different experiments. 1 10 100 1000 10000 100000 1000000 10000000

Drop Mixed Drop T-‐junc Hex 3,2µg in

bulk RNA sample

cDNA synthesis in droplets

(26)

Figure 10: All three experiments have been performed in the same way in droplets. Some difference in the cDNA produced can be seen between the samples. The controls are the average value for the reaction in bulk and an RNA sample where no cDNA synthesis has been performed.

4.5 Cell lysis

When a good cDNA yield could be seen in droplets from extracted RNA the next goal would be to do a cDNA synthesis from whole cells.

To lyse cells a temperature step was introduced. To see at which temperature the cells would lyse. The cells were stained with live and dead stain green and red fluorescent respectively. The cells were stained with live stain before the temperature step so cells with only green fluorescent were considered to be alive. Cells with both red and green fluorescent were considered dead and, this since the dead stain, Propidium Iodide, will only be able to pass the cell membrane when the cell is dead and then have a more permeable cell membrane.

Temperature (°C) Incubation time (min)

Number of cells Number of dead cells

Percent of dead cells 25 10 234 0 0% 55 10 139 30 21.6% 60 10 70 39 55.7% 65 10 51 51 100% 70 10 109 109 100% 50 ∼ 120 64 59 92.2% -‐ 20 ∼ 120 98 97 99.0% 1 10 100 1000 10000 100000 1000000

Drop T-‐junc Drop 1 Drop 2 Hex 3,2µg RNA sample

Reproducibility of drop experiment

(27)

Table 2: Percentage of cells lysed at different temperature and incubation time. The numbers of dead cells are calculated compared to the total number of cells using a fluorescent microscope. Cells incubated at 65°C and 70°C for 10 min were all lysed.

Cells that had been incubated for 10 min at 65°C were all seen to be dead, see table 2, so in the following experiments this procedure will be used to lyse cells.

4.6 cDNA synthesis from whole cells

To test this setup the cDNA synthesis was performed with approximately 150 000 cells (7500 cells/µl) in each sample instead of the extracted RNA. The cells were diluted in either PBS or MilliQ water. None of the samples gave a response in the qPCR

quantification.

To see if the cDNA synthesis was inhibited by the presence of cells, 3,2 µg extracted RNA was added to each cell samples. In this setup different concentrations of cell were used where the highest concentration was 6000 cells/µl this was then diluted 5 times to 1200 cells/µl, 240 cells/µl and 48 cells/µl.

For the highest concentration of cells there was no response in the qPCR but for the samples with lower cell concentration there was no significant inhibition of the cDNA synthesis, see figure 11.

To see if the qPCR was inhibited by the presence of cells a cDNA sample was prepared and divided into different samples with different cell concentrations. The highest concentration was 6000 cells/µl this was then diluted 5 times to 1200 cells/µl and 240 cells/µl.

Again for the highest concentration there was no response in the qPCR and for the middle cell concentration some inhibitory effect could be seen in comparisons to the cDNA sample without any cells. For the sample with low cell concentration no difference could be seen, see figure 11.

cDNA synthesis and analysis in microfluidic droplets

Master Thesis in Nanobiotechnology