Isolation of RNA from small human articular cartilage specimens allows quantification of mRNA expression levels in local articular cartilage defects

ELSEVIER

Journal of Orthopaedic Research

Journal of Orthopaedic Research 19 (2001) 478481

www.elsevier.nl/locate/orthrea

Isolation of RNA from small human articular cartilage specimens allows quantification of mRNA expression levels in local articular cartilage defects Angelika Gehrsitz, Louise A. McKenna, Stefan Soder, Thomas Kirchner, Thomas Aigner * Curiilage Research Group, Institute of Pathology, Uniorrsity qf Erlungen-Niirnberg, Krunkenhausstrasse 8-10, D-91054Erlangm, Germany

Received 31 March 2000; accepted 6 September 2000

Abstract

Human adult cartilage is an inherently difficult tissue from which to isolate RNA. The RNA isolation techniques described so far have generally only been successfully applied to the isolation of RNA from larger amounts of cartilage. However, it is important to be able to analyse focal cartilage lesions in order to understand the local processes in the cartilage degeneration process. Therefore, we have developed a protocol for isolating RNA directly from as little as 10 mg wet weight of cartilage followed by quantitative PCR analysis. We were able to analyse the expression levels of several genes in parallel including aggrecan and type I1 collagen. 0 2001 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved.

Introduction The isolation of RN A from human articular cartilage has been a long-standing problem due to a number of factors. Cartilage has a low cell content (1-2% total mass) and a highly cross-linked extracellular matrix containing a high concentration of proteoglycan (PG) which tends to co-purify with the R N A as they too are large and highly negatively charged macromolecules [4]. Therefore, much of the published data on human cartilage is based on RNA obtained from chondrocyte cultures. However, in vitro chondrocytes are known to significantly change their phenotype and hence, mRNA expression levels measured in vitro hardly reflect the in vivo situation. Alternatively, people have used animal tissue such as bovine, rabbit, and porcine cartilage [1,7,9]. However, cartilage lesions in animals are not always a good model for direct comparison to human osteoarthritic cartilage degeneration. So far, all reported protocols for isolating R N A from human articular cartilage used larger amounts of tissue [5,8]. This, however, does not allow the analysis of local defects, which is needed for investigating focal events

*Corresponding author. Tel.: +49-9131-8522857; fax: +49-91318534745. E-muil address: thomas.aigner(&atho.imed.uni-erlangen.de (T. Aigner).

characteristic of osteoarthritic cartilage degeneration. Thus, the potential for analysing small cartilage areas is needed. In our laboratory we have used in situ hybridisation in order to compare focal gene expression pattern in normal and early osteoarthritic cartilage [2,3]. However, in situ hybridisation is not quantitative and thus, differences in detected mRNA expression levels are often difficult to interpret. Our interest was, therefore, to develop a protocol which would allow quantitative gene expression analysis of small cartilage pieces (1040 mm3).

Materials and methods Cartilage samples

~

Tissue grading

All cartilage samples were obtained at autopsy, within 48 h of death. Full thickness cartilage plugs of 4 mm of diameter were taken from weight bearing areas of the femoral condyles. Cartilage tissues taken for the study underwent an initial macroscopic inspection of all joint surfaces and the synovial membrane. “Normal” joints did not show any overt cartilage lesions and pathological changes of the synovial membrane (mean age 48 yr). Cases with superficial cartilage lesions in the weight bearing femoral condyle, as visualized by macroscopic surface fissuring and tissue softening, were pre-classifed as showing early degenerative changes (mean age 58 yr). In none of these cases was synovial effusion or discolouration of the synovial fluid encountered. None of the cases had a known history of a metabolic disease or medication known to influence the metabolism of chondrocytes. Cartilage plugs were cut in half and one half was immediately frozen in liquid nitrogen and stored at -80°C until required for RNA

0736-0266/01/$ - see front matter 0 2001 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved PII: S 0 7 3 6 - 0 2 6 6 ( 0 0 ) 9 0 0 2 8 - 7

isolation. The other half was embedded in paraffin and histological sectioning and haematoxylin and eosin-staining was performed. P G concentration was evaluated histochemically by toluidine blue staining. All tissues were graded according to Mankin et al. [6]. Normal cartilage samples showed Mankin’s grade 0-1 and early degenerative cartilage Mankin’s grade 3.

at 13,000 rpm at room temperature. The clear lysate was applied to an RNeasy mini-column (Qiagen, Germany) and R N A was isolated according to the manufacturer’s protocol with an “on-column” DNAse digestion step. Tiryi?iun PCR

Iwlation oj. RNA

R N A was isolated from 1 0 4 0 mg (wet weight) of frozen articular cartilage following micro-dismembranation. Plugs were finely ground in the pre-frozen cups of a micro-dismembrator (Micro-dismembrator S, Braun Biotech International, Germany) using 7 cycles of 1 min at 2200 rpm. Milled cartilage was resuspended in 1 ml of Qiagen lysis buffer and a clear lysate was produced after centrifugation for 10 min

a.

1 Mankin’s I

Agg/

I

COL2/

1

DNA-

1

b. normal articular cartilage

-

6

-

-

5

-

I

First strand synthesis was performed using the First Strand Syathesis Kit from Boehringer (Boehringer Mannheim, Germany), exactly according to the manufacturer’s instructions using a third of the RNA obtained from one isolation. TAQMAN PCR was used to detect collagen type I1 (COL2). aggrecan and glyceraldehyde 3-phosphate dehydrogenase ( G A P D H ) mRNA in human articular cartilage RNA samples. The primers ( M W G Biotech, Gemany) and TAQMAN probes (Eurogentech, Belgium) were designed using PRIMER EXPRESS T M software (Perkin Elmer). A separate master-mix was made up for each of the primer pairs and contained a final coiicentration of 200 pM NTPs, 450 n M Roxbuffer and 100 nM TAQMAN probe. For COL2 the final reaction mix contained 10 p1 cDNA (corresponding t o 5’i.b of the total cDNA synthesized), 0.5 U polymerase (Eurogentech. Belgium) 50 nM forward primer (FP) (CAACACTGCCAACGTCCAGAT), 300 nM reverse primer ( R P ) (CTGCTTCGTCCAGATAGGCAAT), 100 n M COL2-probe (ACCTTCCTACGCCTGCTGTCCACG), and 5.5 mM MgCl?. For aggrecan the final reaction mix contained 10 111 cDNA, 0.5 U DNA-polymerase (Eurogentech. Belgium), 50 nM F P (ACTTCCGCTGGTCAGATGGA). 900 nM R P (TCTCGTGCCAGATCATCACC), and 100 nM aggrecan-probe (CCATGCAATTTGAGAACTGGCGCC),and 6.0 mM MgCI?. For GAPDH the final reaction mix contained 10 1.11 cDNA, 0.5 U polymerase, 300 n M FP (GAAGGTGAAGGTCGGAGTC), 50 nni R P (GAAGATGGTGATGGGATTTC), 100 nM GAPDH-probe (CAAGCTTCCCGTTCTCAGCC), and 5.5 mM MgCI?. The reaction was carried out in an ABI Prism 7700 Sequence Detector (Perkin Elmer. Germany) for 4 min at 95°C and 40 I [I5 s at 95°C and 1 min at 6 0 T ] in triplicate. Results were captured using SDS software from Perkin Elmer and evaluated using Microsoft Excel. Calculation of copy numbers was essentially performed by comparing detection values with values of the accompanying standard curve as suggested by the Perkin Elmer.

I

Results R N A isolation AggrecanlWPDH

COLZIWPDH

Fig. I . ( a ) Quantitative determination of m R N A expression levels for aggrecan and COL1 using TAQMAN PCR technology. The values were normalised against GAPDH in normal ( N 1 4 ; Mankin’s grade 0 I ) and early degenerative joint lesions ( D I 4 ; Mankin’s grade 3); ( * ) ratio genomic DNA:cDNA-molecules; ( b ) shows mean m R N A expression levels for aggrecan ( n = 4 ) and COL2 (17=3)normalised and early degenerative joint lesions ( W ) . against GAPDH in normal (0)

Cartilage plugs of 1 0 4 0 mg wet weight could be rapidly powdered in the micro-dismembrator. The frozen cartilage powder was then allowed to come to room temperature in lysis buffer and a final vortexing step prior to centrifugation of the cartilage powder ensured maximum extraction efficiency of the RNA. Due to the very limited amount of RNA to be expected in 1 0 4 0 mg

Table 1 Validation experiment showing results of TAQMAN analysis on cartilage half-plugs (from normal articular cartilage) which were quarlerrd and proceeded with separatelv (A and B). mRNA copy numbers for aggrecan and GAPDH were determined by TAQMQN PCR

C 1A C 1B C 2A C 2B C 3A C 3B C 4A C 4B

Aggrecan

GAPDH

Agg/GAPDH

674 3479 505 450 605 326 394 245

896 4764 I33 138 227 109

0.75 0.73 3.51 3.76 2.67 7.98 3.44 3.97

114

62

180

ri. Gi4rsitz et ul. I JoumuI of’ Ortlioptredic Research I 9 (2001i 4 7 8 3 8 1

of adult articular cartilage tissue, electrophoretic analysis of the isolated RNA of small cartilage plugs was not feasible. We had previously isolated R N A from larger amounts of cartilage using this method and RNA could be visualised on agarose gels. The two characteristic bands for 18s and 38s ribosomal RNA were seen as discrete bands without any signs of degradation or DNA contamination (not shown). DNA contamination, which can affect quantitative PCR analysis, was additionally excluded by TAQMAN analysis (Fig. l ( a ) last column).

Four plugs from different areas of one joint of a norinal donor (not included in Fig I ) were taken and split in half. R N A was isolated from each half independently and at a different time. The R N A was reverse transcribed and used in TAQMAN analysis. As shown in Table 1, reproducible aggrecan/GAPDH ratios were obtained, although regional differences were encountcred. Two similar experiments, including only two plugs each, were additionally perfomed and they gave comparable results (not shown). Qumztijicrrtion of mRN.4 expression lwels j b r GAPDH, uggrecun, und COL2

Quantitative PCR analysis of gene expression levels was performed on four cases of normal and four cases of early degenerative cartilage samples using the TAQMAN-technology from Perkin Elmer (Fig. 1). This enabled us to carry out sensitive, quantitative gene expression analysis for several genes from as little as 1 0 mg of cartilage tissue. Quantification of RNA without reverse transcription demonstrated that the isolated RNA was largely DNA free. GAPDH was used as a house-keeping gene for normalization of the gene expression levels. In our samples, we specifically analysed aggrecan and COL:! mRNA expression as the main markers of anabolic activity of articular chondrocytes. Results are shown in Fig. I .

The possibility to carry out quantitative PCR using TAQM AN technology has significant advantages over many of the techniques otherwise used to examine gene expression patterns in cartilage. In situ hybridisation, for example, is not a quantitative technique. Also mRNA expression profiling using cDNA-array technology requires confirmation of differences in expression levels by other quantifying techniques. Northern hybridisation requires rather large amounts of RNA. As we have shown in this paper, it is possible to quantitatively analyse the expression profile of genes of interest in parallel in biopsy size cartilage specimens with high sensitivity. The high degree of sensitivity is exemplified by the TAQMAN detection collagen type I1 expression in normal articular cartilage, which had not been detected by in situ hybridisation [3,3].However, one significant drawback of any method of isolation of RNA directly from full-depth articular cartilage specimens is the missing zonal resolution of the technique, which is so far only possible by in situ hybridization. This might be one explanation for the missing differences in gene expression levels for aggrecan and COL2 in between normal and early degenerative cartilage lesions. Alternatively, early cartilage degeneration might not show significant differences in mRNA expression levels for aggrecan and COL2 and the interindividual variability might be higher than differences in very early phases of the degeneration process. This, however, requires an exstensive study based on much more samples than used by us for this study. The high interindividual differences themselves, as well as the predominance of aggrecan versus COL2 expression at least in normal articular cartilage fits very well with previously reported data [3] and more recent data obtained in our laboratory using cDNA-array techniques. Overall, our protocol provides great potential for further elucidating gene expression pattern in local lesions on a more quantitative pattern and thus to promote our understanding of the disease process despite the fact that cartilage has an extremely low cell population. This will allow the analysis of focal processes in cartilage degeneration, which is needed in order to understand the complex disease processes.

Discussion This paper describes a reliable protocol, which incorporates the Qiagen RNEasy procedure with an “oncolumn’’ DNAse-digestion, to isolate DNA-free RNA from very small samples of adult human articular cartilage (10-40 mg wet weight). Purified RNA is suitable for quantitative analysis of gene expression using TAQMAN PCR to analyse mRNA expression levels from human normal and early degenerative cartilage plugs.

Acknowledgements The authors would like to thank Ms Michaela Schafer, Ms Sabine Stegner, Ms Freya Boggasch and Ms Pia Gebhard for excellent technical assistance. This work was supported by the German Ministry of Research (BMBF Grant 01GG9834 as well as the Interdisciplinary Center for Clinical Research, University of Erlangen-Niirnberg (IZKF - Grant D4)).

A. Gehrsitz et al. I Jnurnul o j Ortholiucdic Rcwurch 19 (2001I 4 7 8 4 x 1

References [I] Adams ME, Huang DQ, Yao LY, Sandell LJ. Extraction and isolation of mRNA from adult articular cartilage. Anal Biochem 1992:202:89-95. [2] Aigner T, Ston H, Weseloh G, Zeiler G, von der Mark K. Activation of collagen type I1 expression in osteoarthritic and rheumatoid cartilage. Virchows Arch B 1992;62:337F-lS. [ 3 ] Aigner T, Vornehm SI, Zeiler G, Dudhia J, von der Mark K, Bayliss MT. Suppression of cartilage matrix gene expression in upper zone chondrocytes of osteoarthritic cartilage. Arthritis Rheum 1997;10:562-9. [a] Carney SL, Muir H. The structure and function of cartilage proteoglycans. Physiol Rev 1988;68:858-910. [S] Cs-Szabo G, Melching LI, Roughley PJ, Glant TT. Changes in messenger RNA and protein levels of proteoglycans and link

48 1

protein in human osteoarthritic cartilage samples. Arthritis Rheum 1997;40:103745. [6] Mankin HJ, Dorfniaii H, Lippiello L, Zarins A . Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. J Bone and J Surg 1971;53-A:523-37. [7] Matyas JR, Adams ME, Huang D, Sandell LJ. Discoordinate gene expression of aggrecan and type I1 collagen in experimental osteoarthritis. Arthritis Rheum 1995;38:120-5. [8] helching LT, Cs-Szabo G, Roughley PJ. Analysis of proteoglycan messages in human articular cartilage by a compitative PCR technique. Matrix Biol 1997;16:1-11. [9] Smale G, Sasse J. RNA isolation from cartilage using density gradient centrifugation in cesium trifluoroacetate: An RNA preparation tequnique effective of high proteoglycan content. Anal Biochem 1992;703:351-6.