Optimisation of high-quality total ribonucleic acid isolation from cartilaginous tissues for real-time polymerase chain reaction analysis

Tissue samples

Tissue samples from the AC, M, intervertebral disc AF and intervertebral disc NP were obtained from skeletally mature three- to five-year-old Dutch female dairy goats (n = 3) at autopsy. The research protocol was approved by the scientific board and the animal ethics committee of the VU University Medical Center, and is in accordance with national guidelines and regulations. After collection, samples were cut in half, immediately frozen in liquid nitrogen and stored at -80°C until further use. Tissue samples of less than 100 mg were used in this study. Additionally, a few samples were stored in RNA Later Solution (Life Technologies, Bleiswijk, The Netherlands) to test for optimal storage conditions of the tissues. Figure 1 describes the flow of the performed experiments (arrows) as well as the comparisons made in order to evaluate the different RNA extraction and homogenisation methods (numbers).

Fig. 1

Overview of the experimental series. Arrows show the flow of the experiment: tissue samples were collected from goats, homogenised using two different methods, and ribonucleic acid (RNA) was isolated from each homogenate by four different techniques. RNA yield and purity was determined by NanoDrop. Real-time polymerase chain reaction (PCR) was used to determine the quality and quantity of mRNA transcripts for the gene expression of type II collagen and aggrecan. Numbers in boxes show the different comparisons between the different RNA isolation kits.(1: RNA yield and purity compared for the four different kits.2: Comparison of the two homogenisation methods, i.e. MagNA Lyser and Freezer Mill. 3: Gene expression analysis used to compare the different RNA isolation kits.) (18S rRNA, 18S ribosomal RNA; UBC, Ubiquitin C; YWHAZ, Tyrosine 3-monooxygenase; RIN, RNA integrity number.)

Tissue homogenisation

Each half of the obtained tissue samples was homogenised by either the MagNA Lyser or the Freezer Mill. The MagNA Lyser is a homogeniser that disrupts, and simultaneously homogenises tissue by rapid agitation of a 2 ml screw tube containing the tissue, 1.4 mm ceramic beads and a lysis buffer. Although tissue disruption and tissue homogenisation are two different processes, in this manuscript we will only use the term homogenisation for both processes. Tissues to be homogenised by the MagNA Lyser were divided into four parts and transferred into pre-cooled tubes, filled with ceramic beads, and the lysis buffers derived from the different RNA isolation kits which were added. Homogenisation occurred in four runs of 40 seconds each at 6500 rpm. In between the runs, samples were cooled at 4°C for two minutes. In the case that the samples were very viscous, they were placed on a roller mixer for approximately ten minutes. The Freezer Mill is an impact grinder that uses a steel impactor which moves back and forth between the metal ends of a vial by strong magnetic fields in liquid nitrogen. For the Freezer Mill, frozen samples were added to a pre-cooled grinding container. Tissue was ground in liquid nitrogen in three runs of two minutes each, at a frequency of 10 Hz. A cooling step of two minutes between each run was applied to achieve optimal brittleness. The homogenised tissue powder was divided into four groups and suspended in the lysis buffer provided by the different RNA isolation kits.

RNA isolation

After tissue homogenisation, samples were subjected to RNA isolation using either TRIzol (Life Technologies), the RNeasy Lipid Tissue kit, the RNeasy Fibrous Tissue kit, or the Aurum Total RNA Fatty and Fibrous Tissue kit. Kits are further referred to as TRIzol, Lipid, Fibrous and Aurum, respectively. The homogenised tissue was separated into two equal parts and RNA was isolated with the same kit but independently, i.e. n = 2 per RNA isolation method per homogenisation method performed for cartilaginous material obtained from three different goats. RNA isolation occurred according to the manufacturer’s protocol. For TRIzol isolations, the proposed additional steps for RNA isolation from tissues rich in proteoglycans were tested: high salt precipitation; extra centrifuging step; and precipitation using lithium chloride (LiCl, 8M) instead of propanol. Minor modifications and suggested improvements for the RNeasy Fibrous kit include using 20 μl instead of 10 µl of proteinase K for NP tissues as pilot studies showed increased RNA quality and quantity. Furthermore, 70% ethanol (EtOH), isopropanol and phenol:chloroform:isoamyl alcohol were tried, in addition to 100% EtOH. Both the Lipid and Fibrous kits were used in centrifuge form, whereas the Aurum kit was used in vacuum form. DNase I treatment (Qiagen) was performed according to the manufacturer’s protocol for all kits except TRIzol. Additionally, RNA clean-up sets, MinElute (Qiagen) and RNA clean-up kit (Invitek, Glasgow, United Kingdom) were used after RNA isolation via TRIzol.

RNA analysis

The quantity and quality of the isolated RNA were determined by NanoDrop ND-1000 (NanoDrop Technologies, Inc. Wilmington, New Jersey). The RNA concentration was determined by measuring the absorbance at 260 nm, and purity was assessed by determining the A260/A280 ratio, which should be between 1.8 and 2.0, and the A260/A230 ratio. Additionally, RNA integrity of specific samples was verified by calculating the RNA integrity number (RIN) from the 28S/18S ratios, using the Agilent RNA 6000 Pico Chips (Agilent Technologies, Amstelveen, The Netherlands) according to the manufacturer’s instructions.

Complementary DNA synthesis

First strand complementary (c)DNA synthesis was performed using 100 ng total RNA as input, in a 25 μl reverse transcription reaction mix consisting of 5 U Transcriptor Reverse Transcriptase, 0.08 U random primers, 10 U Protector RNase Inhibitor, and 1 mM of each dNTP in 1 × Transcriptor reverse transcription (RT) buffer (all Roche Diagnostics). Reactions were incubated at 42°C for 45 minutes, followed by an inactivation of the transcriptase enzyme at 80°C for five minutes.

Real-time PCR

Real-time PCR reactions were performed using the SYBR Green reaction kit according to the manufacturer’s protocol (Roche Diagnostics) in a LightCycler 480 System (Roche Diagnostics). The LightCycler reactions were prepared in 20 μl total volume with 7 μl PCR-H₂O, 0.5 μl forward primer (0.2 μM), 0.5 μl reverse primer (0.2 μM), 10 μl LightCycler Mastermix (LightCycler 480 SYBR Green I Master; Roche Diagnostics), to which 2 μl of ten times diluted cDNA was added as a PCR template. Controls in the real-time PCR reaction included RT reactions without the reverse transcriptase (control for DNA carryover) and RT reactions without template (control for reagent contamination). Primers (Life Technologies) used for real-time PCR are listed in Table I. Primers were designed from sequences available in the data banks, based on homology in conserved domains comparing human, mouse, rat, dog and/or cow. Primers were previously used by Vonk et al.^18,19 The amplified PCR fragment extended over at least one exon border, except for 18S. The 18S ribosomal RNA (18S rRNA), Ubiquitin C (UBC) and Tyrosine 3-monooxygenase (YWHAZ) were used as housekeeping genes. The stability of the housekeeping genes was checked by the geNorm software.²⁰ Gene expression levels were normalised for the normalisation factor calculated with the equation 18S*UBC*YWHAZ3. Expression of the target genes, type II collagen (Col2A1) and aggrecan (ACAN), was quantified with the fit point method in the LightCycler software, crossing points from a standard curve of six serial dilutions ranging from 2 pg to 10 fg, of DNA, derived from each gene, were assessed. Relative gene expression is shown as the ratio between absolute expression of the gene of interest and the absolute normalisation factor for housekeeping gene expression of the same sample. PCR efficiency (E) was automatically calculated using the fit point method, as provided by LightCycler (E = 10−1/slope). Gene expression data were used only if the PCR efficiency was calculated as being between 1.85 and 2.0 (Table I).

Statistical analysis

The normality of the data for the different groups was checked and a Wilcoxon rank-sum test was used to determine significant differences in gene expression. SPSS Statistics software version 20 was used (IBM Corp., Armonk, New York). To assess RNA quality and be able to compare the different homogenisation methods and RNA isolation kits, the variance in gene expression of duplicate samples was calculated, assuming that the fold difference for perfect duplicates is one (Fig. 1; orange blocks 2 and 3). After tissue homogenisation, samples were split and processed independently, i.e. n = 2 per RNA isolation method per homogenisation method performed for cartilaginous material obtained from three different goats.

RNA Isolation

Table II shows the average weight of the tissue samples, RNA yield and quality (A260/A280 and A260/A230 ratio) determined for four different cartilaginous tissues, homogenised by either the MagNA Lyser (Magna) or Freezer Mill (Freezer). RNA extraction was executed using four different kits (TRIzol (TRIzol), RNeasy Lipid Tissue kit (Lipid), RNeasy Fibrous Tissue kit (Fibrous), and Aurum Total RNA Fatty and Fibrous Tissue kit (Aurum)). RNA isolation with all four kits for all cartilaginous tissues resulted in quantifiable amounts of RNA. Yields obtained with the Lipid and Fibrous kits were lower for all tissues and homogenisation methods, compared with TRIzol and the Aurum kit. RNA yields obtained with the Aurum kit show high standard deviations (sd) in comparison with the other three kits, especially for the yields of the Aurum samples homogenised by the MagNA Lyser. For all four tissues, higher A260/ A280 ratios were found for the Lipid and Fibrous kits compared with TRIzol and Aurum RNA isolations, suggesting that although the RNA yield for these kits is lower, the obtained RNA is less contaminated. The high sds and low purity of the RNA isolated with the Aurum kit indicate that the kit has low reproducibility for RNA isolation of cartilaginous tissues. The additional steps as suggested by the manufacturer to isolate pure RNA from material containing a very high level of proteoglycans for TRIzol and the Fibrous kit, did all not improve RNA quality (data not shown). In addition, the modifications prior to isolation, i.e., tissue storage in an RNA Later Solution, and RNA clean-up kits after isolation, also did not improve the quality of the obtained RNA (data not shown). Moreover, storage of the tissue samples in the RNA Later solution deteriorated the homogenisation process as the tissues became very rigid.

As protein contaminations and solvent carryovers could interfere with the UV photometric quantification of nucleic acids, RNA concentration and integrity of randomly selected samples was also determined with the Agilent NanoChip in parallel with the NanoDrop. The mean RIN value for all tissues was 4.13 (sd 0.64). For all tissues, homogenisation with the MagNA Lyser resulted in better RIN values compared with the Freezer Mill (mean 4.45, sd 0.57 and 3.80, sd 0.48, respectively). Clear peaks in both the 18S and 28S bands were observed for all four tissues (see supplementary material).

Gene expression analysis to validate RNA quality

The high concentration of proteoglycans in the cartilaginous tissues not only causes problems in the RNA isolation process but can also inhibit real-time PCR reactions, therefore quantitative real-time PCR analysis was performed to validate the quality and purity of the isolated RNA by analysing collagen type II (COL2A1) and aggrecan (ACAN) gene expression. For NP, AF and AC, expression of COL2A1 and ACAN was measured after RNA isolation with both the Lipid and the Fibrous kits, whereas for some samples isolated by TRIzol or the Aurum kit no gene expression could be observed (Table III). For meniscus tissue, only the Lipid kit resulted in quantifiable gene expression of all samples. RNA quality was validated by comparing the variance in gene expression between duplicate samples, assuming that the fold difference for perfect duplicates is one. Tissue samples were split into two after homogenisation; RNA isolation and gene expression analysis were performed in a similar manner but independently for each half. Since the expression of COL2A1 was highly variable between all four cartilaginous tissues, we only present the data of ACAN expression.

Comparison of tissue homogenisation methods

By determining the fold differences in gene expression between duplicate samples, we compared the two different homogenisation methods for each isolation kit. For the Aurum RNA isolation kit, large variations in gene expression between duplicate samples, up to > 1500-fold differences, were observed. We decided to exclude results obtained with the Aurum kit from further analysis due to these large variations, the relative high number of samples where gene expression could not be determined (Table III), and the abnormal RNA yields and A260/A280 ratios (Table II).

Large differences between duplicates were observed for the samples isolated with TRIzol, regardless of homogenisation method, especially for meniscus and AF tissue as can be observed for aggrecan expression in Figure 2. Only for articular cartilage did RNA isolation with TRIzol result in similar gene expression for the duplicate samples. For the Lipid and Fibrous kits, homogenisation method does not influence gene expression analysis as no significant differences could be detected between duplicate samples (Fig. 2a).

Fig.

Graph showing the comparison of tissue homogenisation methods. The variances in aggrecan gene expression between two duplicate samples for the four cartilaginous tissues isolated with the same kit homogenised by either the MagNA Lyser (dark orange) or Freezer Mill (second bar from left): a) for the RNeasy Lipid (left hand bars × 2) and Fibrous (right hand bars × 2) kits; and b) for TRIzol. Data are presented as mean and standard error. Note the linear scale for graph (a) and the logarithmic scale for graph (b). As the aggrecan gene expression of NP tissue after RNA isolation with TRIzol could only be determined for the duplicates of the samples of one goat for both homogenisation methods, no error bars are shown. (AC, articular cartilage; M, meniscus; AF, annulus fibrosus; NP, nucleus pulposus.)

Comparison and validation of the different RNA isolation kits

Since no differences between the two homogenisation methods for the Lipid and Fibrous kits were observed, data from samples isolated by either one of these kits but homogenised by different methods were pooled. In order to compare these two kits with the frequently used RNA isolation method TRIzol, we also pooled the data from the two homogenisation methods, despite the observed differences (Fig. 2b).

Reproducible gene expression was found for all tissues isolated with the Lipid and Fibrous kits and no differences between the two kits were observed (Fig. 3a), whereas large variations were observed for the TRIzol RNA isolations (Fig. 3b). Although for the meniscus no significant difference between the Lipid and Fibrous kits was found, Figure 3b shows less difference between duplicates for the Lipid kit.

Fig.

Comparison of ribonucleic acid (RNA) isolation kits analysed by real-time polymerase chain reaction. Results of RNA isolation with the different kits and for the different cartilaginous tissues are compared by analysing the differences in aggrecan gene expression between duplicate samples: a) for the RNeasy Lipid (dark orange) and Fibrous (light orange) kits; and b) for TRIzol. Tissues homogenised with the same method are pooled. Data are presented as mean and standard error of the mean. Note the linear scale for graph (a) and the logarithmic scale for graph (b). (AC, articular cartilage; M, meniscus; AF, annulus fibrosus; NP, nucleus pulposus.)