Blood biomarker panel studies aim to predict knee osteoarthritis, progression

AMSTERDAM – Results of separate U.S. and European studies hold promise that predictive blood tests for the diagnosis and progression of knee osteoarthritis could soon be developed.

Researchers at Duke University in Durham, N.C., and at the Hospital Universitario de A Coruña, Spain, reported their studies’ findings at the World Congress on Osteoarthritis sponsored by the Osteoarthritis Research Society International.

Courtesy National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS)

Clinical trials are often statistically underpowered to measure progression of knee osteoarthritis, noted Dr. Virginia B. Kraus of Duke University, because of inefficient methods for identifying patients who would progress.

“This is because, with the exception of generalized OA, traditional risk factors are very poor predictors, and that includes age, gender, BMI [body mass index], knee pain, and baseline joint space width,” she explained. Radiographic changes and changes in Kellgren-Lawrence grade are often slow to appear and may only occur in a few subjects at a time. Imaging biomarkers have been identified but are a “rather expensive way of predicting progression,” Dr. Kraus said. There have been some associations between OA progression and biomarkers in the urine and serum, she said, but still not at a level that could be put to clinical use.

“We hypothesized that we could improve the utility of OA biomarkers by using a nonbiased, nontargeted [mass spectrometry] approach,” Dr. Kraus said, explaining that this allowed them to create a Multiple Reaction Monitoring (MRM) panel of 146 serum peptides that were based on analyses of serum, synovial fluid, and urine samples from knee OA radiographic progressors and nonprogressors. The MRM panel was used first to identify candidate biomarkers and then confirm their differential expression patterns, followed by a process of verifying and validating the selected serum biomarkers.

The main aim of the study was to look for serum biomarkers in data from the Prediction of Osteoarthritis Progression (POP) and Genetics of Generalized Osteoarthritis (GOGO) patient cohorts that would be highly predictive for knee OA progression, with a secondary aim to look at potential early diagnostic biomarkers. The investigators used data from 83 patients with symptomatic knee progression to search for prognostic biomarkers and 126 with symptoms but without knee progression to find diagnostic biomarkers. The mean age of patients was 64 years, and 82% were female. The mean BMI was 28 kg/m².

As expected, clinical predictors alone – age, gender, and BMI – had low predictive power for OA progression, with an area under the curve (AUC) of about 0.7. However, these clinical parameters in conjunction with a panel of 10 markers identified via MRM improved the AUC to 0.9 in the verification set and 0.8 in the validation set. For diagnosis, the AUC for clinical covariates alone was 0.8, but combined with a set of 19 MRM-identified markers, the AUC was 1.0 for the verification set and 0.9 for the validation set.

Dr. Kraus said that the serum peptide biomarkers identified from the MRM panel predicted that OA progression pathways would involve aggregation of cells, complement activation, cell movement of leukocytes, and cellular infiltration of leukocytes. The pathways predicted to be involved in a diagnosis of OA included skeletal function, cell movement, formation of filaments, cellular protrusions, and hemostasis.

“Results from this work in progress suggest that a select set of biomarkers from nontargeted analyses could significantly improve prediction of knee OA progression,” Dr. Kraus concluded. They also suggest a set of biomarkers could be used for diagnosis because of their distinction from those for progression, with one exception. Future work will look at whether the diagnostic panel could be used in patients with nonradiographic OA to potentially diagnose and perhaps help treat OA early before structural damage is apparent, she said.

The proteomics group at the Hospital Universitario de A Coruña collaborated with researchers at the KTH-Royal Institute of Technology in Stockholm and the University of Oxford (England) to also identify a serum protein biomarker panel that might prove useful for the diagnosis of OA.

Their study compared serum samples taken from 461 patients with knee OA with those taken from 412 patients with rheumatoid arthritis and 159 nonarthritic individuals. Serum samples from each of these groups were divided into a screening (n = 593) and a verification set (n = 439). The data were adjusted for variations in participant’s age, BMI, and gender.

In the screening phase, the investigators used an antibody suspension bead array composed of 174 antibodies to examine 78 protein targets to try to identify the best serum protein biomarker candidates. They replicated the findings in an adapted array of 79 antibodies and 34 protein targets and then used this array in the verification set of serum samples.

There were three proteins that could potentially differentiate patients with OA from controls and perhaps be used to diagnose OA. These were: complement 3, which is involved in inflammation; inter-alpha-trypsin inhibitor heavy chain I, which is involved in the stability of the extracellular matrix; and S100 calcium-binding protein A6, which has a role in chondrocyte differentiation and thus bone remodeling.

“Compared to rheumatoid arthritis patients, we observed that the level of complement 3 and inter-alpha-trypsin inhibitor heavy chain I are significantly increased in osteoarthritis patients,” said the presenting study author Lucía Maria Lourido Salas, PhD. “This suggests the potential of these two proteins as specific proteins for osteoarthritis,” she proposed. A further 28 proteins were found that might also help to differentiate OA from RA; 16 of these were increased and 12 decreased in OA patients. Together these 30 proteins could potentially help identify a specific serum protein signature of OA, Dr. Salas said.

Dr. Kraus and Dr. Salas did not report having any financial disclosures. Dr. Kraus’ research is funded by the Duke Biomarker Factory.

AMSTERDAM – Results of separate U.S. and European studies hold promise that predictive blood tests for the diagnosis and progression of knee osteoarthritis could soon be developed.

Researchers at Duke University in Durham, N.C., and at the Hospital Universitario de A Coruña, Spain, reported their studies’ findings at the World Congress on Osteoarthritis sponsored by the Osteoarthritis Research Society International.

Courtesy National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS)

Clinical trials are often statistically underpowered to measure progression of knee osteoarthritis, noted Dr. Virginia B. Kraus of Duke University, because of inefficient methods for identifying patients who would progress.

“This is because, with the exception of generalized OA, traditional risk factors are very poor predictors, and that includes age, gender, BMI [body mass index], knee pain, and baseline joint space width,” she explained. Radiographic changes and changes in Kellgren-Lawrence grade are often slow to appear and may only occur in a few subjects at a time. Imaging biomarkers have been identified but are a “rather expensive way of predicting progression,” Dr. Kraus said. There have been some associations between OA progression and biomarkers in the urine and serum, she said, but still not at a level that could be put to clinical use.

“We hypothesized that we could improve the utility of OA biomarkers by using a nonbiased, nontargeted [mass spectrometry] approach,” Dr. Kraus said, explaining that this allowed them to create a Multiple Reaction Monitoring (MRM) panel of 146 serum peptides that were based on analyses of serum, synovial fluid, and urine samples from knee OA radiographic progressors and nonprogressors. The MRM panel was used first to identify candidate biomarkers and then confirm their differential expression patterns, followed by a process of verifying and validating the selected serum biomarkers.

The main aim of the study was to look for serum biomarkers in data from the Prediction of Osteoarthritis Progression (POP) and Genetics of Generalized Osteoarthritis (GOGO) patient cohorts that would be highly predictive for knee OA progression, with a secondary aim to look at potential early diagnostic biomarkers. The investigators used data from 83 patients with symptomatic knee progression to search for prognostic biomarkers and 126 with symptoms but without knee progression to find diagnostic biomarkers. The mean age of patients was 64 years, and 82% were female. The mean BMI was 28 kg/m².

As expected, clinical predictors alone – age, gender, and BMI – had low predictive power for OA progression, with an area under the curve (AUC) of about 0.7. However, these clinical parameters in conjunction with a panel of 10 markers identified via MRM improved the AUC to 0.9 in the verification set and 0.8 in the validation set. For diagnosis, the AUC for clinical covariates alone was 0.8, but combined with a set of 19 MRM-identified markers, the AUC was 1.0 for the verification set and 0.9 for the validation set.

Dr. Kraus said that the serum peptide biomarkers identified from the MRM panel predicted that OA progression pathways would involve aggregation of cells, complement activation, cell movement of leukocytes, and cellular infiltration of leukocytes. The pathways predicted to be involved in a diagnosis of OA included skeletal function, cell movement, formation of filaments, cellular protrusions, and hemostasis.

“Results from this work in progress suggest that a select set of biomarkers from nontargeted analyses could significantly improve prediction of knee OA progression,” Dr. Kraus concluded. They also suggest a set of biomarkers could be used for diagnosis because of their distinction from those for progression, with one exception. Future work will look at whether the diagnostic panel could be used in patients with nonradiographic OA to potentially diagnose and perhaps help treat OA early before structural damage is apparent, she said.

The proteomics group at the Hospital Universitario de A Coruña collaborated with researchers at the KTH-Royal Institute of Technology in Stockholm and the University of Oxford (England) to also identify a serum protein biomarker panel that might prove useful for the diagnosis of OA.

Their study compared serum samples taken from 461 patients with knee OA with those taken from 412 patients with rheumatoid arthritis and 159 nonarthritic individuals. Serum samples from each of these groups were divided into a screening (n = 593) and a verification set (n = 439). The data were adjusted for variations in participant’s age, BMI, and gender.

In the screening phase, the investigators used an antibody suspension bead array composed of 174 antibodies to examine 78 protein targets to try to identify the best serum protein biomarker candidates. They replicated the findings in an adapted array of 79 antibodies and 34 protein targets and then used this array in the verification set of serum samples.

There were three proteins that could potentially differentiate patients with OA from controls and perhaps be used to diagnose OA. These were: complement 3, which is involved in inflammation; inter-alpha-trypsin inhibitor heavy chain I, which is involved in the stability of the extracellular matrix; and S100 calcium-binding protein A6, which has a role in chondrocyte differentiation and thus bone remodeling.

“Compared to rheumatoid arthritis patients, we observed that the level of complement 3 and inter-alpha-trypsin inhibitor heavy chain I are significantly increased in osteoarthritis patients,” said the presenting study author Lucía Maria Lourido Salas, PhD. “This suggests the potential of these two proteins as specific proteins for osteoarthritis,” she proposed. A further 28 proteins were found that might also help to differentiate OA from RA; 16 of these were increased and 12 decreased in OA patients. Together these 30 proteins could potentially help identify a specific serum protein signature of OA, Dr. Salas said.

Dr. Kraus and Dr. Salas did not report having any financial disclosures. Dr. Kraus’ research is funded by the Duke Biomarker Factory.

AMSTERDAM – Results of separate U.S. and European studies hold promise that predictive blood tests for the diagnosis and progression of knee osteoarthritis could soon be developed.

Researchers at Duke University in Durham, N.C., and at the Hospital Universitario de A Coruña, Spain, reported their studies’ findings at the World Congress on Osteoarthritis sponsored by the Osteoarthritis Research Society International.

Courtesy National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS)

Clinical trials are often statistically underpowered to measure progression of knee osteoarthritis, noted Dr. Virginia B. Kraus of Duke University, because of inefficient methods for identifying patients who would progress.

“This is because, with the exception of generalized OA, traditional risk factors are very poor predictors, and that includes age, gender, BMI [body mass index], knee pain, and baseline joint space width,” she explained. Radiographic changes and changes in Kellgren-Lawrence grade are often slow to appear and may only occur in a few subjects at a time. Imaging biomarkers have been identified but are a “rather expensive way of predicting progression,” Dr. Kraus said. There have been some associations between OA progression and biomarkers in the urine and serum, she said, but still not at a level that could be put to clinical use.

“We hypothesized that we could improve the utility of OA biomarkers by using a nonbiased, nontargeted [mass spectrometry] approach,” Dr. Kraus said, explaining that this allowed them to create a Multiple Reaction Monitoring (MRM) panel of 146 serum peptides that were based on analyses of serum, synovial fluid, and urine samples from knee OA radiographic progressors and nonprogressors. The MRM panel was used first to identify candidate biomarkers and then confirm their differential expression patterns, followed by a process of verifying and validating the selected serum biomarkers.

The main aim of the study was to look for serum biomarkers in data from the Prediction of Osteoarthritis Progression (POP) and Genetics of Generalized Osteoarthritis (GOGO) patient cohorts that would be highly predictive for knee OA progression, with a secondary aim to look at potential early diagnostic biomarkers. The investigators used data from 83 patients with symptomatic knee progression to search for prognostic biomarkers and 126 with symptoms but without knee progression to find diagnostic biomarkers. The mean age of patients was 64 years, and 82% were female. The mean BMI was 28 kg/m².

As expected, clinical predictors alone – age, gender, and BMI – had low predictive power for OA progression, with an area under the curve (AUC) of about 0.7. However, these clinical parameters in conjunction with a panel of 10 markers identified via MRM improved the AUC to 0.9 in the verification set and 0.8 in the validation set. For diagnosis, the AUC for clinical covariates alone was 0.8, but combined with a set of 19 MRM-identified markers, the AUC was 1.0 for the verification set and 0.9 for the validation set.

Dr. Kraus said that the serum peptide biomarkers identified from the MRM panel predicted that OA progression pathways would involve aggregation of cells, complement activation, cell movement of leukocytes, and cellular infiltration of leukocytes. The pathways predicted to be involved in a diagnosis of OA included skeletal function, cell movement, formation of filaments, cellular protrusions, and hemostasis.

“Results from this work in progress suggest that a select set of biomarkers from nontargeted analyses could significantly improve prediction of knee OA progression,” Dr. Kraus concluded. They also suggest a set of biomarkers could be used for diagnosis because of their distinction from those for progression, with one exception. Future work will look at whether the diagnostic panel could be used in patients with nonradiographic OA to potentially diagnose and perhaps help treat OA early before structural damage is apparent, she said.

The proteomics group at the Hospital Universitario de A Coruña collaborated with researchers at the KTH-Royal Institute of Technology in Stockholm and the University of Oxford (England) to also identify a serum protein biomarker panel that might prove useful for the diagnosis of OA.

Their study compared serum samples taken from 461 patients with knee OA with those taken from 412 patients with rheumatoid arthritis and 159 nonarthritic individuals. Serum samples from each of these groups were divided into a screening (n = 593) and a verification set (n = 439). The data were adjusted for variations in participant’s age, BMI, and gender.

In the screening phase, the investigators used an antibody suspension bead array composed of 174 antibodies to examine 78 protein targets to try to identify the best serum protein biomarker candidates. They replicated the findings in an adapted array of 79 antibodies and 34 protein targets and then used this array in the verification set of serum samples.

There were three proteins that could potentially differentiate patients with OA from controls and perhaps be used to diagnose OA. These were: complement 3, which is involved in inflammation; inter-alpha-trypsin inhibitor heavy chain I, which is involved in the stability of the extracellular matrix; and S100 calcium-binding protein A6, which has a role in chondrocyte differentiation and thus bone remodeling.

“Compared to rheumatoid arthritis patients, we observed that the level of complement 3 and inter-alpha-trypsin inhibitor heavy chain I are significantly increased in osteoarthritis patients,” said the presenting study author Lucía Maria Lourido Salas, PhD. “This suggests the potential of these two proteins as specific proteins for osteoarthritis,” she proposed. A further 28 proteins were found that might also help to differentiate OA from RA; 16 of these were increased and 12 decreased in OA patients. Together these 30 proteins could potentially help identify a specific serum protein signature of OA, Dr. Salas said.

Dr. Kraus and Dr. Salas did not report having any financial disclosures. Dr. Kraus’ research is funded by the Duke Biomarker Factory.