Skip to content


Arthritis Research & Therapy

Open Access

Differential gene expression and protein expression levels of MMP and TIMP molecules in response to glucocorticoid treatment in arthritic patients

  • T Andersson1, 5,
  • E af Klint1,
  • S Strobel2,
  • U Andersson3,
  • A Stark4,
  • R Gelinas2,
  • J Lundeberg5,
  • L Klareskog1 and
  • A-K Ulfgren1
Arthritis Res Ther20046(Suppl 1):53

Received: 16 January 2004

Published: 24 February 2004


GlucocorticoidImmunohistochemical StainingGene Expression DataProtein Expression LevelDifferential Gene Expression


In this study we attempted to map the molecular events that underlie the clinical effects of glucocorticoid (GC) treatment in patients.


DNA arrays and immunohistochemical staining was used for comparison of patient material isolated before and after intra-articular glucocorticoid (GC) treatment. Gene expression data analysis was performed in the GeneSpring™ software.


Hierarchical clustering of gene expression data could distinguish samples taken before from those taken after treatment to a surprisingly high degree, in spite of anticipated individual and experimental variations. Patterns were further analyzed by identification of genes with both statistically significant differential expression (P < 0.05) using the cross gene error model implemented in GeneSpring™, and greater than twofold differential expression. Twelve genes satisfying both of these criteria were found, including matrix metalloproteinase (MMP)-1 and MMP-3, for which mRNA expression was downregulated, and tissue inhibitor of MMP (TIMP)-1 and TIMP-4, for which mRNA expression was upregulated by GC treatment. Immunohistochemical staining for several MMP and TIMP molecules confirmed data for MMP-1 and MMP-3 on a protein level, whereas the result for TIMP-1 and other TIMPs revealed reduced protein expression of these molecules.


Our data demonstrate that DNA arrays may be used for evaluating the molecular effects of novel as well as established therapies, but also that data should be interpreted with caution concerning their ability to predict therapeutic effects at the protein level.

Authors’ Affiliations

Rheumatology Unit, Department of Medicine, Karolinska Hospital, Sweden
Celltech Group PLC, Seattle, USA
Department of Woman and Child Health, Karolinska Hospital, Sweden
Department of Orthopedic Surgery, Karolinska Hospital, Sweden
Department of Biotechnology, Royal Institute of Technology (KTH), Stockholm, Sweden


© BioMed Central Ltd 2004