In Situ RNA Quality Control

(1)

In Situ RNA Quality Control

A spatial heat map of RNA integrity with single cell resolution

DEGREE PROJECT IN MEDICAL BIOTECHNOLOGY BY KONSTANTIN CARLBERG

Supervisor: Professor Joakim Lundeberg, Royal Institute of Technology

Co-‐supervisor: PhD student Linda Kvastad, Royal Institute of Technology

(2)

Abstract

The quality of RNA is of great importance in gene expression studies. It is mostly measured using the RNA integrity number (RIN). Lately it has been shown that samples with low RIN and different fragmentation patterns could affect quality of sequencing data. For such low RIN samples a new approach has been developed by Illumina called the DV200 metric, which is the percentage of fragments >200 nucleotides. For samples with low RIN, DV200 has proved to be a better method to predict if good quality data from RNA sequencing can be generated. However, neither RIN nor DV200 provide spatial information on the RNA integrity. Thus, tissues with areas of heterogeneous RNA integrity, where regions of good quality RNA sequencing data could be generated from are missed. We have designed a method to spatially evaluate the RNA integrity in tissue, which we named in situ RNA QC. The method uses three probes with three different fluorophores, each bound to three specific cDNA regions synthesized from the high abundant and well conserved 18S rRNA.

With the help of in-‐house technology from the Spatial Transcriptomics (ST) group at SciLifeLab, we enable creation of heat maps over the RNA integrity to show spatial fragmentation patterns of RNA in tissue. This could reveal the regional quality of transcripts in situ, which is crucial knowledge when selecting samples for further RNA sequencing.

The assay has been tried using different tissue fixation methods in order to show a proof of concept that formalin gives shorter cDNA fragments than acetone. The generated heat-‐map provides a visual overview of RNA integrity in situ; hence this method could be used to select samples for sequencing by evaluating the spatial quality of RNA. For instance from fresh frozen and formalin fixated paraffin embedded (FFPE) tissue (biobanks contain large number of long-‐

term storage FFPE samples). With this assay we will be able to determine which samples are suitable for sequencing.

(3)

Contents

ABSTRACT ... 1

1 INTRODUCTION ... 3

1.1 Problem statement ... 3

1.2 Aim of study ... 6

2 BACKGROUND ... 8

2.1 Fixation agents ... 10

2.2 CodeLink slides ... 10

2.3 Probes with fluorophores ... 11

2.4 Full-‐length cDNA synthesis ... 11

2.5 Removal of rRNA strand ... 12

2.6 Hybridization ... 12

3 MATERIALS AND METHODS ... 12

3.1 Primer design ... 12

3.2 Evaluation of probes using qPCR ... 14

3.3 Printing of activated slides ... 14

3.4 RNA QC ... 15

3.5 Scanning and microscope use ... 17

3.6 Total RNA control ... 17

3.7 qPCR on released material ... 18

4 RESULTS ... 18

4.1 Primers ... 18

4.2 qPCR results ... 19

4.3 qPCR after release ... 21

4.4 Bright field images ... 21

4.5 Quality control assay ... 23

4.6 Total RNA experiments ... 27

5 DISCUSSION ... 29

5.1 Spatial heat map ... 29

5.2 Alternative methods ... 31

5.3 Further tests ... 32

6 REFERENCES ... 33

7 APPENDIX ... 36

7.1 Sequences ... 36

7.2 Melt curve results ... 41

7.3 Tables over scanning intensities ... 43

ACKNOWLEDGEMENTS ... 44

(4)

1 Introduction 1.1 Problem statement

Ribosomal RNA is the most abundant RNA species in human cells and corresponds to >80 % of the total RNA content [1]. Ribosomal RNA (rRNA) is a collection of RNA components of the ribosomes essential for the protein synthesis during the translation where messenger RNA (mRNA) is decoded into polypeptides. One of these rRNA, 18S ribosomal RNA (18S rRNA) is an rRNA subunit with a highly conserved sequence, frequently used in phylogenetic studies [2].

According to the NCBI database human 18S rRNA is 1869 bases in length and has evident secondary structure characteristics [3]. Various conformations of 18S rRNA have been advocated with some conformations favored depending on the surrounding environmental conditions [4]. One of the suggested conformations presented by Apollo Chemistry in Georgia Institute of Technology can be seen in Figure 1 [5].

Figure 1. Representation of one conformation of human 18S rRNA with the clearly

folded nature, presented by Apollo Chemistry of Georgia Institute of Technology [5].

(5)

Due to the high abundance of 18S rRNA compared to mRNA, 18S rRNA has been used as an indicator of the overall RNA status in cells and tissue, for instance the overall degradation of RNA [6]. The single stranded nature of RNA molecules and the highly reactive carbon (C2) hydroxyl group on the ribose unit makes RNA more sensitive to degradation than DNA. Some RNA-‐mediated enzymatic events act on the C2 hydroxyl group, for instance RNase A, an enzyme in the RNase family which are nucleases that catalyzes the degradation of RNA [7]. RNase H, another enzyme in the RNase family degrades RNA hybridized to DNA by breaking the 3’-‐

phosphodiester bond on the RNA strand [7]. RNases are very ubiquitous and hard to eliminate i.e. RNA is vulnerable to chemical degradation [7]. RNA is in general more sensitive to heat degradation than DNA. Single stranded RNA, depending on environmental conditions for instance type of buffer, might decay from temperatures >65°C, whilst double stranded RNA forming secondary structures heat denatures at around 70-‐75°C [8-‐9]. Fragmented RNA eventually leads to loss of function and a way to measure RNA integrity is to calculate the RNA Integrity Number (RIN) [6]. Degraded mRNA is a major concern in genomic research since the loss of the full RNA sequences leads to difficulties in obtaining information on gene isoforms (splice variants) and Single Nucleotide Polymorphism (SNPs), thus the knowledge of RIN is valuable for gene expression studies [10]. The RNA quality has traditionally been calculated by evaluating the 28S rRNA to 18S rRNA ratio on agarose gel. Nowadays the RIN algorithm is based on results derived from an electropherogram. The RNA molecules are stained with a nucleotide binding dye and are exposed to an electrical current and migrate through the gel with different speed depending on size. The electropherogram measures the fluorescent signal from the dyed RNA over time, giving different electropherogram patterns of the fragmentation ratio with RIN scores from 1 to 10, where 1 is completely degraded RNA while 10 is completely intact RNA. Agilent 2100 bioanalyzer is an instrument that uses this electropherogram technology that is built on a machine-‐learning algorithm. Thus, it solves the drawbacks with the traditional subjective interpretation of the RNA quality and offers a quantitative non-‐subjective score [6].

The RNA quality is crucial in single cell studies but there is a certain loss of information when using RIN to determine RNA integrity. The importance of the RNA quality for single cell studies has been mentioned in articles, like Pietersen’s et al. article on Human postmortem brain tissue using Laser Capture Microdissection (LCM) [11]. Similar methods have been discussed in other articles [12].

Because most of the studies on RNA are solution based the spatial information of expression in tissue is lost during the sample preparation. Concerning RNA quality, spatial information of tissue may show if certain cells and regions of the tissue are more prone to be affected by RNA degradation than other. At SciLifeLab (Stockholm, Sweden) Professor Joakim Lundeberg’s group in collaboration with Professor Jonas Frisén of Karolinska Institute have developed a method that enables the determination of gene expression spatially on tissue sections. This method is called Spatial Transcriptomics (ST) and is a technology that allows the integration of gene expression information with tissue morphology. The method provides a better understanding of the diversity within tissue and between cells. Spatial Transcriptomics is a sequencing-‐based method which uses poly-‐T probes on an Arrayit^® slide to capture mRNA. The probes are printed like dots covering the active surface of the slide with features that makes it possible to trace the origin of each sequenced transcript from the tissue with the aim to reach

(6)

single cell resolution. The mRNA is vertically diffused to its corresponding featured dot in order to recreate a spatial genetic expression profile of the tissue. In articles to come this groundbreaking technology will be presented in detail. The possibility to study multiple genes simultaneously and trace the origin of the transcript and the level of expression on tissue will make ST a complement to the Human Protein Atlas (HPA). This technology will allow for studies on differences between mRNA expression and protein expression, since not all mRNA translates into protein. It will also provide information on diseased tissue, some interesting features on different areas of tissue and further the information on development of diseases at cellular level.

In medical research, before further investigations, tissues are fixated. Fixation is a step where the histological sections are treated to prevent decay of the tissue. The fixation is most often performed with various chemical detergents but heat and immersion are also utilized methods.

The most widely used fixation method is formaldehyde fixation (commercial name formalin) where tissues are fixated in 10% Neutral Buffered Formalin (NBF), that is approximately 4%

formaldehyde in Phosphate Buffered Saline (PBS) [13]. Formalin is a beneficial fixation method for keeping tissue morphology. Formalin makes the tissue more rigid by creating crosslinks between tissue, proteins and nucleic acids. This however makes DNA, RNA and protein studies on tissue more challenging [14]. These challenges can be exemplified when performing full-‐

length complementary DNA (cDNA) synthesis from RNA. The reverse transcriptase (RT) enzyme that acts on the RNA strand during cDNA synthesis is affected by steric hindrance from the protein-‐RNA crosslinks. This fact complicates the detection of full-‐length rRNA using formalin fixation and furthermore the measurement of RNA integrity. Other fixation reagents denature the proteins in the tissue by disrupting the hydrophobic interactions that keeps the tertiary structures of the proteins. This further leads to aggregation (precipitation) of the proteins, the solubility of the proteins decreases and the morphology of the tissue is preserved.

Those reagents are usually alcohols or ketones like methanol and acetone. Acetone also has a permeabilizing effect on the tissue [13-‐14]. With these differences between the ketone based methods compared to formalin, ketone based methods may have advantages studying full-‐

length RNA sequences on tissue as the nucleic acids are exempt from protein cross-‐linking.

The study of cancer tissue has been highly important using spatial methods. The heterogenous nature of cancer makes array based technologies like the study of spatial expression of genes a research method for cancer. These array-‐based methods provide information on the spatial expression of genes between different cancer tissue and different cells inside the same cancer tissue may show unique features. Not only is heterogenousity revealed between cancer cells, but also knowledge of the activity of surrounding cells. The same methods can be applied on other heterogenous diseases like Reumatoid arthritis in order to gain essential information on the disease. Some of these spatial methods use barcoding, microfluidics and different tracebacking algorithms in order to gain the spatial information, as in Rahul Satija’s study about spatial reconstruction of single-‐cell gene expression data [15]. Other studies have evaluated spatial gene expression and the field is a hot topic presently [16-‐18].

The RNA integrity number may be the same for a vast variety of RNA integrity patterns among low RIN samples. Thus RIN may not be the ultimate method to evaluate RNA quality of tissue.

An Illumina^® protocol using another metric system for RNA quality called DV200 has been

(7)

suggested. This method uses the percentage of RNA fragments >200 nucleotides to determine the RNA quality [19].

As the genetic studies on tissue improves, the demand for RNA quality control methods for the tissue increases rapidly. The drawbacks with total RNA studies signify the importance of spatial RNA integrity measurement methods. Fresh frozen tissue and Formalin-‐Fixed Paraffin-‐

Embedded (FFPE) tissue may show different patterns of degradation around the surface compared to the interior of the tissue. Some spatial integrity patterns may arise from different fixation methods and the biopsy (the sampling of tissue). Spatial information of the RNA quality could provide information on which areas of the tissue biopsies beneficial RNA sequencing results could lead to. If cells adjacent to tumor cells have intact RNA while other parts of the tissue have cells with low RNA integrity, sequencing of the tissue could still hypothetically provide informative sequencing data about the tumor cells.

New methods to obtain the spatial information of RNA integrity could in other words act as a quality control for the biopsy, for the fixation method and for the research of cancerous or other disease affected tissue as well as for tissues stored in biobanks prior conducting further gene expression studies.

1.2 Aim of study

This thesis aims to create a quality control assay for evaluating RNA integrity on tissue.

The quality control assay will be based on the study of 18s rRNA degradation in tissue by creating a heat-‐map using fluorescently dyed oligonucleotides (probes). The fluorescently dyed probes will bind on specific locations on the captured and full-‐length cDNA-‐synthesized 18s rRNA from tissue. The information gained from the spatial heat-‐map will provide information about the degree of fragmentation of the tissue. For instance which tissues actually provide satisfying RNA sequencing information or if different fixation methods affect the overall quality of the tissue.

The purpose is to capture 18S rRNA from sectioned tissue on an activated glass slide, synthesize cDNA from RNA and use three differently fluorescently dyed probes binding to three different regions of the cDNA sequence, one close to the 5’-‐end, one in the middle and one in the 3’-‐end and by that evaluate the integrity of RNA based on the binding of the probes. The fluorophores used for the study are Fluorescein isothiocyanate (FITC) with emission maximum at 519 nm, Cyanine 3 (Cy3) with emission maximum at 523 nm and Cyanine 5 (Cy5) with emission maximum at 635 nm [20]. A schematic representation of the workflow can be seen in Figure 2.

(8)

Figure 2. Representation of the workflow for the in situ RNA QC. 1. Cryosectioning: Tissue sections and capture probes specific to a region close to the 3’ site of the 18s rRNA. In reality the tissue and the probes are stringently stuck on the glass surface. 2. Permeabilization: This step allows diffusion through the cells. RNA and the capture probes diffuse through the cell membrane and can interact. The RNA diffuses vertically and attaches to the probe right under it. 3. cDNA synthesis: after the rRNA is attached to the capture probes, cDNA synthesis is executed. 4.

Tissue removal: tissue is removed. 5. rRNA removal: the rRNA is removed from the cDNA strand using RNase H that degrades the rRNA strand. 6. Hybridization of detection probes: Three different fluorescently dyed probes bind to specific regions of the cDNA. Each fluorophore has its specific site on the RNA. The yellow circles represent FITC, the green Cy3 and the red Cy5. The excess probes are washed away. And thereafter imaging of the heat-‐map is performed.

The probe bound to FITC were closest to the 3’-‐site on the rRNA molecule; the probe bound to Cy3 were in the middle region of the rRNA molecule and the probe bound to Cy5 were closest to the 5’-‐site of the rRNA molecule. The hypothesis is that FITC would be visible all over the tissue while Cy5 would be visible only in the intact regions and should therefore not be hard to distinguish from the rest of the tissue while Cy3 would be intermediately visible.

For studying fresh frozen tissue a long probe panel (long-‐range) is used where the probes are spread over the whole cDNA molecule synthesized from the 18S rRNA. For FFPE tissue a short probe panel (short-‐range) is used with probes binding a smaller span of the cDNA molecule synthesized from the 18S rRNA, see Figure 3 for a representation on the probe panels. The short probe panel was not experimentally tested on tissue in this thesis.

②

①

③

④

⑤ ⑥

(9)

Figure 3. Representation of the long panel and the short panel. Note that the

probes with the same name has identical sequences but have different fluorophores depending on the panel.

Furthermore this is a visual sequencing-‐free method that evades time consuming and demanding library preparation. Some methods implement algorithms for quality control of the RNA after the sequencing step. The in situ RNA quality control assay circumvents data processing and data analysis steps. The results are easy to interpret, offering the user a desirable method to study RNA quality spatially on tissue in order to decide whether the tissue is suitable for RNA sequencing or not.

2 Background

Below in Figure 4 is a representation of the workflow and the experimental setup of the quality control assay. The quality control assay was performed on mouse brain tissue.

(10)

Figure 4. Flowchart over the method. The optimization of the method is visualized.

Each step represents critical points in the assay where optimization was executed.

The reason for the Hematoxylin and Eosin staining (H&E) being detoured is because it was performed on different tissue sections than the ones that the rest of the assay was performed on. The H&E staining is an optimal step, allowing tissue morphology with spatial manifestation of RNA integrity.

Sectioning

Formalin Methanol

Acetone

H&E staining

Permeabilization

RNA removal

cDNA synthesis

Tissue removal

Hybridization

Scanning

Fixation

(11)

2.1 Fixation agents

In a proof of concept study of spatial RNA integrity, comparison between formalin fixation, acetone fixation and methanol fixation was executed. The formalin fixation should in theory not generate signal from fluorophores bound far away from the capture site on the rRNA because of the protein-‐RNA crosslinks. Methanol could act as a comparison towards acetone in order to evaluate which of the method that is superior. There are a vast number of other fixation detergents that can be evaluated including combination of methods, for instance mixing of methanol and acetone.

2.2 CodeLink slides

The RNA capture step requires probes complementary to the RNA sequence bound to a 2-‐

dimensional transparent surface to enable scanning on microscopes and laser scanners.

CodeLink® provides activated glass slides with hydrophilic polymers containing N-‐

hydroxysuccinimide ester groups that covalently bind amine-‐modified DNA, see Figure 5 [21].

The covalent amine-‐ester bond between the DNA probes needs a 5’-‐amine 6 carbon chain (C6) modification in order to bind the activated slide. The activated surface disrupts at temperature above 55°C during exposures over 24 hours [22]. Due to that it is necessary to keep temperatures below 55°C throughout most of the experiment.

Figure 5. Representation of the chemistry on the surface of the activated

CodeLink slides. The picture is taken from ParaBioSys of Harvard Medical School [21].

(12)

2.3 Probes with fluorophores

The use of fluorescently dyed probes in order to evaluate integrity of RNA enables the visualization of degradation on a 2-‐dimensional surface. The probes were designed with around 50 % Guanine-‐Cytosine (GC) content since that increases the chances for the probes to bind sufficiently enough. The melting temperature Tm, where 50% of the oligonucleotides are hybridized, will be below 50°C for the probes. The amine-‐ester do not disrupt on the chip and the Tm for the capture probe will have its optimum at 42°C, the temperature the cDNA synthesis is set.

The ideal probe lengths for reducing the risk of unspecific binding are between 18-‐22 nucleotides with optimum at 20. The free energy (∆G) is an important factor to consider when designing probes. If the free energy of the probe has a negative number some reactions like folding can occur spontaneously. Too high negative numbers on the free energy of the probes may result in undesirable hairpin structures, therefore a good rule of thumb when designing hybridization probes is to exclude probes that have secondary structures at free energies with greater negative numbers than -‐6 kcal/mol [23].

2.4 Full-‐length cDNA synthesis

Mouse 18S rRNA is 1870 nucleotides with similar secondary structure characteristics of the human sequence. Since 18S rRNA is highly conserved the quality control method for evaluating the RNA integration supposedly works for both species. The sequence similarity between human 18S rRNA (NCBI reference NR_003286.2) and mouse 18S rRNA (NCBI reference NR_003278.3) is 99% with 1857/1872 matches and 7/1872 gaps [3, 24]. The sequences can be seen in appendix 1, 2 and 3.

In order for the cDNA synthesis to cover the whole span of the sequence the reverse transcriptase needs to give rise to product sizes of at least 1870 bases (1.87 kb). Thermo Fisher Scientific claims that SuperScript® III Reverse Transcriptase has the ability to transcribe sequences up to 12.3 kb. It also has strand displacement activity, which means that the enzyme has the ability to loosen up secondary structures during the transcription [25-‐26].

The cDNA synthesis can be enhanced in different ways for instance by adding additional amount of nucleotides. Also adding Betaine to the reaction, which reduce secondary structure formation in GC-‐rich regions and thereby decreases its melting temperature and enhances the cDNA synthesis. Adding extra MgCl2 act as a catalyst to the enzyme and additional Dimethyl Sulfoxide (DMSO) to decrease thermostability and inhibit secondary structures will also improve the cDNA synthesis [27-‐28].

(13)

2.5 Removal of rRNA strand

In order to leave room for the fluorescently dyed probes to bind to the cDNA, the RNA strand needs to be removed. Removal of RNA can occur in different ways for instance by denaturation, but the DNA/RNA hybrid denatures in temperatures above 95°C [23]. This is not the optimal way of removing RNA since the amine-‐ester bind on the CodeLink® will disrupt and a lot of signal will get lost. RNase H cleaves and degrades the DNA/RNA hybrid [29]. The enzyme leaves some un-‐cleaved gaps on the molecule, which is a drawback. In order to remove as much RNA as possible the RNase H treatment can be executed for longer durations. Short rRNA leftovers on the cDNA strand may dehybridize easier since the Tm is lower for shorter fragments.

The removal of rRNA is a critical step for the assay to work and the RNA strand needs to be completely removed to allow proper hybridization of the fluorescently dyed probes for the assay to work optimally.

2.6 Hybridization

The hybridization temperature is usually set between 5-‐10°C below Tm. Too low temperatures will result in unspecific bindings and higher temperatures than Tm will result in the melting of the DNA hybrids [23]. The optimal hybridization solution depends on the probe structure and the target [30]. It can be calculated using databases as the OligoAnalyzer 3.1 from Integrated DNA Technologies^® (IDT) [31-‐32].

3 Materials and methods 3.1 Primer design

3.1.1 Capture probes

Capture probes were designed to bind in the area close to the 3’-‐end of 18S rRNA.

The assay was constructed to work for both mouse and human tissue. The sequences used for finding optimal primers had the NCBI Reference Sequence number NR_003286.2 (human 18S rRNA) and NR_003278.3 (mouse 18S rRNA) [3, 24]. The primer3web database (version 4.0.0) was used for finding primers [33-‐35].

The primer conditions were a sequence length of between 18-‐23 nucleotides with 20 as optimal length, a Tm between 38-‐50°C with optimal temperature at 42°C, and a content of guanine and cytosine (GC-‐content) of 30-‐60 % with an optimum at 50 %. The first five bases of the 3’-‐ends of the primers were set to include two of either guanine or cytosine or one of each. The primers were then controlled in the mfold web server for determination of secondary structures [36-‐37]. The free energy of the formed secondary structures was set to not have any larger negative number than -‐6 kcal/mol.

Primer dimers (self-‐complementarity of the primers) were also evaluated using the Oligo Calc: Oligonucleotide Properties Calculator [38-‐39]. Primers that gave rise to primer dimers in the databases were excluded.

Each probe was checked in the Basic Local Alignment Search Tool BLAST^® database for complementarity, in order to see if any unspecific binding could occur. Probes that had off-‐

target bindings were excluded.

(14)

3.1.2 Hybridization probes

Hybridization probes complementary to cDNA from 18S rRNA were designed.

The purpose of the hybridization probes was to bind the cDNA synthesized from rRNA.

Four hybridization probes were designed for the long-‐range experiment and for the short-‐

range experiment. The conditions set in primer3web, mfold and Calc: Oligonucleotide Properties Calculator was identical to the ones for the capture probes.

The probes were set as left primers (forward primers) from the original 18s rRNA sequence.

The probes were named Probe 1, Probe 2, Probe 3 and Probe 4. The probes were ordered in a way so that the capture probe designed as a right primer (reverse primer) was closest to the 3’-‐end of the 18s rRNA, thereafter the Probe 1 designed as a left primer second closest to the 3’-‐end, Probe 2 third, Probe 3 fourth and Probe 4 fifth. Thus Probe 1 was set to bind close to the 5’-‐end of the cDNA, Probe 2 and Probe 3 in the middle region and Probe 4 close to the 3’-‐end. Probe 1, 2 and 3 was used for the short-‐range probe panel and Probe 2, 3 and 4 for the long-‐range probe panel that was experimentally tested. Thus the short-‐range probe setup had no probes close to the 3’-‐end of the cDNA. For the capture probes and each probe help primers were designed. For the capture probes the help primers act as forward primers and for Probe 1, Probe 2, Probe 3 and Probe 4 help primers acts as a reverse primers, see Figure 6.

From a pool of probes a panel of one capture probe, named Capture probe, and one of each;

Probe 1, Probe 2 and Probe 3 and Probe 4 were selected for the in situ RNA QC procedure.

Each probe was checked in the BLAST^® database for complementarity, in order to see if any unspecific binding could occur.

Figure 6. Representation of the theory behind the qPCR runs on the

selected probes with the example of Probe 1 and Help primer 1 seen in the

(15)

picture. Probe 2 and Probe 3 bind to areas to the right of Probe 1 on the cDNA strand. NOTE: the picture is not in scale.

3.2 Evaluation of probes using qPCR

From the pool of several candidate probes, evaluation of probes for further studies using qPCR was performed. The evaluation of probes was performed on human total RNA. After the evaluation probes with 5’-‐attached fluorophores were ordered from IDT^®.

3.2.1 Reverse transcription

For each reaction a mixture was prepared containing a final concentration of 1x 5xSuper script III First strand buffer (250 mM Tris-‐HCl, 375 mM, KCl, 15 mM MgCl2), 5 mM DTT, 1 M Betaine, 6 µM MgCl2, 0.25 µM Capture primer, 4x1 mM dNTPs, 0.2 mg/ml BSA, 10 % DMSO, 20 U/µl SuperScript® III Reverse Transcriptase, 2 U/µl RNaseOUT™ Recombinant Ribonuclease Inhibitor. All reaction were carried out at a total volume of 20µl. Total RNA was added to each PCR tube to a final amount of 10 ng in each tube. The reverse transcription was run at 42°C for 50 minutes, and thereafter kept in 10°C. The samples were stored in -‐20°C.

3.2.2 qPCR

A reaction mixture of 1x iQTM SYBR® Green supermix (2x concentration), 500 nM reverse primers (hybridization probes), 500 nM forward primers (help primers) were prepared.

Thereafter the cDNA was added to a final amount of 1 ng in each well. The qPCR reaction was done using a white-‐coated 96 well PCR plate, and the total volume used in each well was 20 µl. The qPCR program used started with 3 minutes preheating at 95°C, 10 seconds of denaturation at 95°C, 30 seconds of annealing in 55°C, 30 seconds of extension in 72°C. The denaturation, annealing and extension steps were run in 50 cycles, thereafter a melt curve between 55-‐95°C was run.

3.3 Printing of activated slides

One Capture probe was selected for printing on a CodeLink® Activated Slide. An amine C6 modified version of the Capture probe was ordered from IDT. Printing was done using Arrayit® SuperMask™ 16 Substrate Slides and holders. Each well of the mask was printed with a mix containing 20 µM of the Capture probe and 1x of 2x Printing buffer (containing 100mM Sodium Phosphate, 0.12 % Sarkosyl). Each well was treated with the printing mix for one minute before it was removed by pipetting. Once the Printing mix was removed from each well the masks were dried in room temperature for 5 hours.

After drying the slides were put in a rack and placed in a humidity chamber over night.

(16)

The next day the slides were treated in a chamber with 50°C pre-‐warmed blocking solution (containing 0.1 % Tris, 50 mM ethanolamine) for 30 minutes with 100 rpm shake and protected from light. Thereafter the slides were quickly washed in deionized water. After washing with deionized water the slides were treated in a chamber with 50°C prewarmed washing solution (containing 4x SSC, 0.1 % SDS) for 30 minutes with 100 rpm shake. After washing the slides, they were quickly washed in deionized water and spun dry.

3.4 RNA QC

3.4.1 Tissue sectioning

Mouse brain tissues embedded in optimal cutting temperature compound (OCT) were sectioned using a CryoStar NX50 Cryostat. The sectioning conditions set in the cryostat where -‐20°C for the cutting blade and -‐20°C for the tissue holder. The tissues were sectioned in 10 µm sections. After sectioning the tissues were attached to the active wells of the activated slides. After the attachment of the tissues they were further adhered to the slides by heating to 37°C for 1 minute.

3.4.2 Fixation

After sectioning, fixation was done by treating each tissue with different fixation methods.

Acetone fixation was done using 100 % pre-‐cooled (-‐20°C) acetone for 10 minutes in -‐20°C.

Thereafter the acetone was removed by pipetting and the slides were washed in 1xPBS and incubated in 37°C until the slides were dry. The methanol fixation was carried out exactly the same way as the acetone fixation. The concentration of methanol used was 100%.

For the formalin fixation 4% formalin diluted in 1xPBS was used for 10 minutes in room temperature. Thereafter the slides were washed in 1xPBS and incubated in 37°C until the slides were dry.

3.4.3 Hematoxylin and Eosin staining

The staining was performed by initially adding 100% Isopropanol (2-‐propanol) to cover the whole tissue. After the Isopropanol evaporated a Mayer’s Hematoxylin solution was added to each tissue and incubated at room temperature for 7 minutes. After staining with Hematoxylin the slides were washed in nuclease free water and thereafter a Dako Bluing Buffer was added to the tissue and incubated at room temperature for 2 minutes. The slides were washed in nuclease free water after the Bluing buffer treatment. Subsequently the tissue was treated with 1:20 Eosin (diluted in Tris-‐acrylamide) for 10 seconds after being washed in nuclease free water. The slides were dried in room temperature and then incubated in 37°C for 5 minutes.

After the Hematoxylin and Eosin staining the slides were scanned in a Zeiss Imager.Z2 with the Vslide and Metafer software from MetaSystems. The slides were treated with 85% glycerol before the scan and the settings in the microscope were set at LED intensity 10 with white panel, a 20x intensity bright field classifier, camera gain 1 and integration time 0.0015 seconds.

(17)

After the bright field microscopy the slides were washed from the glycerol in nuclease free water, and after that washed in 80% ethanol. Thereafter the slides were washed in room temperature and after that incubated in 37°C for 1 minute.

3.4.4 Permeabilization

Permeabilization was performed by treating each well containing tissue with 0.1 M pepsin in 0.1 % HCl. It was performed in 37°C for 5 minutes, thereafter the permeabilization mixture was removed from each well by pipetting and the slides were washed by pipetting 0.1xSSC in and out of each well.

3.4.5 cDNA synthesis

A cDNA synthesis mixture was prepared containing 1x 5xSuper script III First strand buffer (250 mM Tris-‐HCl, 375 mM, KCl, 15 mM MgCl2), 5 mM DTT, 50 ng/µl Actinomycin D (500 ng/µl in 10 % DMSO), 1 M Betaine, 6 µM MgCl2, 0.2 mg/ml BSA, 10 % DMSO, 4x1 mM dNTPs.

The cDNA synthesis mixture was preheated to 42°C before the addition of SuperScript® III Reverse Transcriptase with a final concentration of 20 U/µl and RNaseOUT™ Recombinant Ribonuclease Inhibitor final concentration of 2 U/µl.

Thereafter the cDNA synthesis mixture was added to each well with tissue and incubated in 42°C over night.

After the incubation the cDNA synthesis mixture was removed from each well by pipetting and the wells were washed by pipetting 0.1xSSC in and out of each well.

3.4.6 Tissue removal

Tissue removal was done using 40 mAU/ml QIAGEN® Proteinase K in Proteinase K digest buffer (PKD buffer). The tissues were treated with the tissue removal mixture for 1 hour in 56°C using 300 rpm interval mix (6 seconds mixing 3 seconds rest). After incubation the tissue removal mixture was removed from each well by pipetting and the wells were washed by pipetting 0.1xSSC in and out of each well. The slides were checked for absence of tissue before proceeding to the denaturation step.

3.4.7 rRNA removal

The rRNA was removed from the cDNA using a reaction mixture containing 1x 5xSuper script III First strand buffer (250 mM Tris-‐HCl, 375 mM, KCl, 15 mM MgCl2), 0.2 mg/ml BSA, 40 U/ml RNase H.

The reaction was executed at 37°C for 1 hour. After incubation the tissue removal mixture was removed from each well by pipetting and the wells were washed by pipetting 0.1xSSC in and out on each well.

3.4.8 Hybridization of probes

Hybridization of fluorescently dyed probes was done using a mixture containing 1x 2xHybridization buffer (20mM Tris-‐HCl, 2mM EDTA 100mM NaCl), 0.5 µM of each fluorescently dyed probes. The hybridization solution was preheated to 42°C before being added to each well. The hybridization was executed at 42°C for 30 minutes protected from light. The hybridization mix was removed by pipetting and thereafter washing of the plates was performed.

(18)

3.4.9 Plate Wash

The slides were washed in 42°C heated washing solution 1 (containing 2xSSC and 0.1% SDS) for 10 minutes with 300 rpm shake. The slides were thereafter washed in room temperature with washing solution 2 (0.2xSSC) for 1 minute with 300 rpm shake. Lastly the slides where washed in room temperature in washing solution 3 (0.1xSSC) for 1 minute with 300 rpm shake, and thereafter the slides where spun dry.

3.5 Scanning and microscope use 3.5.1 Control scan

The slides were scanned with an Agilent G2565CA DNA Microarray Scanner. Both red (for Cy5) and green (for Cy3) lasers were used and the resolution was set to 5 µm and the intensity for both lasers was set to 50 %.

3.5.2 Microscope scan

After the control scan SlowFade® Gold Antifade Reagent was added to cover all active wells.

Thereafter a glass cover was put on top of the antifade reagent and the slides were scanned in a Zeiss Imager.Z2 with the Vslide and Metafer software from MetaSystems. The laser source was an X-‐cite® exacte with the intensity set to 100. The classifier used in Metafer was a three fluorophore classier that scanned for FITC, Cy3, and Cy5. The camera gain was set to 6 and the integration time was set to 0.02.

3.6 Total RNA control

3.6.1 Fragmentation of total RNA

An amount of 9 µg of total RNA was used for fragmentation with the NEBNext® Magnesium RNA Fragmentation Module Protocol. The fragmentation mix contained 2x RNA Fragmentation Buffer (10X) together with the total RNA. The sample volume was 20 µl and the fragmentation was executed at 94°C. To gain a length of 1000 nucleotides for the fragments the total RNA was heated for 1 minute until stopped with x2.2 10X RNA Fragmentation Stop Solution. To gain a length of 300 nucleotides for the fragments the total RNA was heated for 3 minutes until stopped with x2.2 10X RNA Fragmentation Stop Solution.

After the fragmentation, the fragmented total RNA samples were purified using the RNeasy®

MinElute® Cleanup Kit from QIAGEN.

3.6.2 Purification of fragmented RNA

The RNeasy® MinElute® Cleanup Kit protocol was followed using the starting volume of 100 μl [40]. Thereafter the samples were analyzed for fragment sizes in a 2100 Bioanalyzer Instrument from Agilent Technologies.

3.6.3 Bioanalyzer run

A Bioanalyzer run was performed on the fragmented samples using the Agilent RNA 6000 Nano Kit Guide [41].

3.6.4 RNA QC

The RNA QC protocol was run on total RNA in a similar fashion as for the tissue. After the Bioanalyzer run the samples were treated as in 3.4.5 cDNA synthesis to 3.5.2 Microscope scan, excluding the 3.4.6 Tissue removal step.

(19)

3.7 qPCR on released material 3.7.1 PolyU-‐Capture probe

A poly-‐uracil modified capture probe with 5 uracils in the 5’-‐site of the probe was designed.

The poly-‐U capture probe had the amine C6 modification as well in order to be printed on a CodeLink slide. The probe was printed as in 3.3 Printing of activated slides. Thereafter the quality control assay was executed from 3.4.1 Tissue sectioning to 3.4.6 Tissue removal. After the tissue removal step a release step was executed were the captured material from the surface of the CodeLink slide were released. The release was executed using a mix diluted in nuclease free water of 100 U/ml USER™ Enzyme and 1XCutSmart® Buffer. The mix was added to the wells on the slide and a one-‐hour treatment in 300 rpm shake and 37°C was performed.

After the treatment the liquid from the wells was pipetted out and a qPCR as in step 3.2.2 was performed on the released material.

4 Results 4.1 Primers

The sequence of the cDNA from 18s rRNA from human and mouse with the positions of each probe can be seen in appendix. The information about the sequence similarity between the sequences from mouse and human can be seen in appendix. The panels of chosen probes can be seen in Table 1 and 2. For the qPCR the capture probe was used as a primer for the reverse transcriptase of the total RNA. Each probe without its associated fluorophore and with a help probe was used for the qPCR.

Table 1. The panel of long-‐range probes with positions counted from the 3’ end on human 18S rRNA

Probe name Fluorophore Sequence 5’-‐3’ Tm rRNA position 3’-‐5’

Capture probe None TTTACTTCCTCTAGATAGTC 47.92 45

Probe 2 Help primer 2 Probe 3 Help primer 3 Probe 4 Help primer 4

FITC^* None Cy3^* None Cy5^* None

GAGATTGAGCAATAACAG AATCAACGCAAGCTTATGAC GTAGTTCCGACCATAAAC GTGTTGAGTCAAATTAAG GGTGACTCTAGATAACCT CGAAAGAGTCCTGTATTG

47.77 54.69 49.75 45.93 48.95 49.77

396 192 801 615 1581 1325

*The primers did not have the associated fluorophore during the qPCR experiment

Table 2. The panel of short-‐range probes with positions counted from the 3’ end on human 18S rRNA

Probe name Fluorophore Sequence 5’-‐3’ Tm rRNA position 3’-‐5’

Capture probe None TTTACTTCCTCTAGATAGTC 47.92 45

Probe 1 Help primer 1 Probe 2 Help primer 2 Probe 3 Help primer 3

FITC^* None Cy3^* None Cy5^* None

GAGGAATTCCCAGTAAGT TCCTCTAGATAGTCAAGTTC GAGATTGAGCAATAACAG AATCAACGCAAGCTTATGAC GTAGTTCCGACCATAAAC GTGTTGAGTCAAATTAAG

48.62 45.74 47.77 54.69 49.75 45.93

233 51 396 192 801 615

*The primers did not have the associated fluorophore during the qPCR experiment