POS0231 GENETIC BIOMARKERS, SNP GENES AND mtDNA HAPLOGROUPS, PREDICT OSTEOARTHRITIS STRUCTURAL PROGRESSORS THROUGH THE USE OF SUPERVISED MACHINE LEARNING

Background

Knee osteoarthritis is the most prevalent chronic musculoskeletal debilitating disease. Current treatments are only symptomatic and to improve this, we need a robust prediction model to stratify patients at an early stage according to the risk of joint structure disease progression. Some genetic factors, including single nucleotide polymorphism (SNP) genes and mitochondrial (mt)DNA haplogroups/clusters, have been linked to this disease.

Objectives

For the first time, we aim to determine, by using machine learning, whether some SNP genes and mtDNA haplogroups/clusters alone or combined could predict early knee osteoarthritis structural progressors.

Methods

Participants (901) were first classified for the probability of being structural progressors. Genotyping included SNP genes TP63, FTO, GNL3, DUS4L, GDF5, SUPT3H, MCF2L, TGFA, mtDNA haplogroups H, J, T, Uk, others, and clusters HV, TJ, KU, C-others. They were considered for prediction with major risk factors of osteoarthritis, namely, age and body mass index (BMI). Seven supervised machine learning methodologies were evaluated. The support vector machine was used to generate gender-based models. The best input combination was assessed using sensitivity and synergy analyses. Validation was performed using 10-fold cross-validation as well as an external cohort (TASOAC).

Results

From 277 models, two were defined. Both used age and BMI in addition for the first one of the SNP genes TP63, DUS4L, GDF5, FTO with an accuracy of 85.0%; the second profits from the association of mtDNA haplogroups and SNP genes FTO and SUPT3H with 82.5% accuracy. The highest impact was associated with the haplogroup H, the presence of CT alleles for rs8044769 at FTO, and the absence of AA for rs10948172 at SUPT3H. Validation accuracy with the cross-validation (about 95%) and the external cohort (90.5%, 85.7%, respectively) was excellent for both models.

Conclusion

This study introduces a novel source of decision support in precision medicine in which, for the first time, two models were developed consisting of i) age, BMI, TP63, DUS4L, GDF5, FTO and ii) the optimum one as it has one less variable: age, BMI, mtDNA haplogroup, FTO, SUPT3H. Such a framework is translational and would be of benefit to patients at risk of structural progressive knee osteoarthritis.

Acknowledgements

The authors would like to thank the Osteoarthritis Initiative (OAI) participants and Coordinating Center for their work in generating the clinical and radiological data of the OAI cohort and for making them publicly available. The OAI is a public-private partnership comprised of five contracts (N01-AR-2-2258; N01-AR-2-2259; N01-AR-2-2260; N01-AR-2-2261; N01-AR-2-2262) funded by the National Institutes of Health, a branch of the Department of Health and Human Services, and conducted by the OAI Study Investigators. Private funding partners include Merck Research Laboratories; Novartis Pharmaceuticals Corporation, GlaxoSmithKline; and Pfizer, Inc. Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health. This manuscript was prepared using an OAI public use data set and does not necessarily reflect the opinions or views of the OAI investigators, the NIH, or the private funding partners. None of the authors are part of the OAI investigator team. Moreover, the authors are also grateful to the TASOAC participants.

A special thanks to ArthroLab Inc. for having provided the MRI data used for classifying structural progressors for each individual.

Disclosure of Interests

Hossein Bonakdari: None declared, Jean-Pierre Pelletier Shareholder of: ArthroLab Inc., Grant/research support from: Work supported in part by the Osteoarthritis Research Unit of the University of Montreal Hospital Research Centre and the Chair in Osteoarthritis from the University of Montreal., Francisco J. Blanco: None declared, Ignacio Rego-Perez: None declared, Alejandro Durán-Sotuela: None declared, Dawn Aitken: None declared, Graeme Jones: None declared, Flavia Cicuttini: None declared, Afshin Jamshidi Grant/research support from: Received a bursary from the Canada First Research Excellence Fund through the TransMedTech Institute in Canada., François Abram Employee of: was an employee of ArthroLab Inc., Johanne Martel-Pelletier Shareholder of: ArthroLab Inc., Grant/research support from: Work supported in part by the Osteoarthritis Research Unit of the University of Montreal Hospital Research Centre and the Chair in Osteoarthritis from the University of Montreal.

POS0347 SPECIFIC MITO-NUCLEAR INTERACTIONS AND mt16519C VARIANT AS PREDICTIVE BIOMARKERS FOR THE RAPIDLY PROGRESSIVE OSTEOARTHRITIS OF THE KNEE. DATA FROM THE OSTEOARTHRITIS INITIATIVE

Background:

The early identification of patients with rapid progressive osteoarthritis (RPOA) could allow the implementation of prevention strategies and their inclusion in clinical trials. Polymorphisms in nuclear and mitochondrial DNA (mtDNA) have been associated with OA. Preliminary analyses by our group showed nuclear single nucleotide polymorphism (nSNP) rs12107036 of TP63 as a potential risk factor for RPOA of the knee.

Objectives:

i) To analyze interactions between mtDNA haplogroups and rs12107036 ii) To apply Next Generation Sequencing (NGS) to discover novel mitochondrial variants to construct predictive models of RPOA of the knee.

Methods:

1102 Caucasian subjects from the OAI were classified as follows: i) Rapid progressors (N=255), baseline KL grade 0-1 or 2 in at least one knee, that increases up to KL≥ 3 or 4 respectively during 48-month follow-up. ii) Non-rapid progressors (N=847), with the same baseline characteristics as rapid progressors, but with slower or no evolution over time.

mtDNA haplogroups and rs12107036 were assigned by mini-sequencing techniques. Novel mtDNA variants were studied by NGS. Statistical analyses included chi-square tests and generalized estimating equations. Relative excess risk due to interaction (RERI) and attributable proportion (AP) were evaluated for the additive interaction between mtDNA clusters and nSNP rs12107036. A nomogram for the estimation of the risk of RPOA was also developed. Analyses were performed using SPSS Statistics v24 and epi.R package included in R software v3.6.3.

Results:

Chi-square analyses revealed an increased risk of RPOA in patients with the allele G of rs12107036 and mtDNA cluster UK (OR 2,013; p=0,001). An excess of 70,3% of RERI between nSNP rs12107036 and mtDNA clusters was detected, indicating that 47,1% (AP) of the risk is attributable to this interaction, therefore harboring both genetic factors increase the risk of RPOA up to 4,7 times compared to harboring just one. mtDNA sequencing revealed the variant mt16519 overrepresented in rapid-progressors (OR 1,620; p=0,002).

Table 1 shows the predictive model for the risk of RPOA. The interaction between the allele G of rs12107036 and mtDNA cluster KU (OR 1,727; p=0,036), in addition to the variant mt16519C (OR 1,690; p=0,003), showed a significant association with the RPOA phenotype regardless of age, BMI, contralateral knee OA, previous injury and WOMAC pain. Image 1 displays the nomogram for predicting risk of RPOA; as an example, a 70 year old male, with a BMI of 28, WOMAC pain score of 10, contralateral OA and presence of both mito-nuclear interaction and mt16519C, has a risk of RPOA of 0,7.

Conclusion:

mtDNA genetic variants are useful, not only as modulators of the influence of specific nuclear polymorphisms on the risk of developing RPOA, but also as candidate genetic biomarkers of this phenotype.

Table 1.

Predictive model for the risk of RPOA phenotype

Variablep-valueORmin 95% CIMAX 95% CIClinical and genetic variablesAge<0,001#1,0561,0381,074Female0,1431,2600,9251,718BMI<0,001#1,0651,0301,101Contralateral OA<0,001#1,9271,4132,626Previous Injury<0,001#1,7701,2932,422WOMAC pain0,001#1,0971,0391,159rs12107036 G0,1721,2260,9151,643mt16519 C0,003#1,6901,2022,375mtDNA Clusters$Others0,8030,9210,4821,760TJ0,4821,2090,7122,052UK0,1360,6980,4351,120HVReferencers12107036 G*mtDNA ClusterG * Others0,5020,7890,3951,576G * TJ0,1580,6470,3531,185G * UK0,036#1,7271,0362,881G * HVReference

$mtDNA Clusters: haplogroups with a common phylogenetic origin BMI: Body Mass Index; WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index; OR: Odds Ratio; CI: confidence interval; #: statistical significance declared at P ≤ 0.05, in bold.

Image 1.

Nomogran for the estimation of the risk of RPOA phenotype. Circles represent the values for the example. Clusters: haplogroups with a common phylogenetic origin BMI: Body Mass Index; WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index.

Disclosure of Interests:

None declared

THU0012 IMPACT OF RS12107036 POLYMORPHISM OF TP63 ON THE RISK OF RAPID PROGRESSIVE OSTEOARTHRITIS OF THE KNEE. DATA FROM THE OSTEOARTHRITIS INITIATIVE

Background:

There is a need to identify patients with the rapid progressive phenotype of Osteoarthritis (RPOA) to include them in clinical trials and to implement prevention strategies. During the last years, nuclear single nucleotide polymorphisms (SNPs) were associated with susceptibility and progression of the disease, but not with the rapid progression phenotype.

Objectives:

Analyze the influence of previously knee OA-associated nuclear SNPs on the risk of RPOA in patients of the OAI.

Methods:

Caucasian patients from the OAI were selected and assigned into three different groups (N=252/group) based on the following criteria:

A.rapid progressors; baseline KL grade 0-1 in at least one knee and increase up to KL≥ 3 during a 48-month period; or baseline KL grade 2 in at least one knee and increase up to KL 4 or total knee replacement during the follow-up.

B.no-rapid progressors; baseline KL grade 0-1 in at least one knee and increase up to KL 2 during 48-month period; or baseline KL grade 2 in at least one knee and increase up to KL 3 during the follow-up.

C.no-progressors; KL grade 0-2 at baseline in at least one knee and bilaterally stable during 48-month period.

Groups were re-categorized into two groups: non-progressors and progressors (pooling A and B). Nuclear SNPs were previously assigned by mini-sequencing techniques. Preliminary chi-square analyses and binary and multinomial logistic regression models adjusted by gender, age, body mass index (BMI), contralateral OA, previous injury in target knee and WOMAC pain, were performed with IBM SPSS Statistics v24.

Results:

We analyzed the effect of 7 SNPs that had been strongly associated with knee OA susceptibility in different GWAS studies: rs11177, rs4730250, rs11842874, rs12107036, rs8044769, rs10948172 and rs143383. Chi-square analyses only showed differences in the frequency distribution of rs12107036 between groups (p=0,028), being the GG genotype over-represented in the rapid progressors group and the AA genotype in the non-progressors group (Figure 1).

The binary logistic regression showed that G allele was significantly over-represented in the (pooled) progressors group when compared with non-progressors (p=0,008) (Table 1). And the multinomial logistic regression showed that, in addition to age and previous injury in target knee, the GG genotype (p=0,032) emerged as a potential risk factor for the RPOA when compared with non-rapid progressors (Table 2).

Table 1.

Binary regression model comparing progressors pool vs. no-progress

Variablesp-valueORC.I. 95%Min.Max.Age0,2171,0120,9931,030Sex (Female)0,000#2,0491,4782,842BMI0,000#1,0851,0441,127Contralateral OA (Yes)0,044#1,4001,0091,942Previous Injury (Yes)0,002#1,7231,2232,429WOMAC pain0,003#1,1021,0331,177rs12107036 G (Yes)0,008#1,6821,1482,463

CI: confidence interval; OR: Odd Ratio; #: statistical significance declared at P ≤ 0.05

Table 2.

Multinomial regression model comparing rapid vs. no-rapid progressors.

Variablesp-valueORC.I. 95%Min.Max.Age0,000#1,0641,0411,088Sex (Female)0,4980,8750,5951,287BMI0,0961,0340,9941,077Contralateral OA (Yes)0,7921,0520,7191,539Previous Injury (Yes)0,028#1,5231,0472,216WOMAC pain0,0911,0550,9921,123rs12107036 GG (Yes)0,032#1,5741,0392,382

CI: confidence interval; OR: Odd Ratio; #: statistical significance declared at P ≤ 0.05

Conclusion:

The G allele of the nuclear SNP rs12107036 of TP63 gen increases the risk of knee OA progression. Depending on the number of risk allele copies the level of progression varies, being the GG genotype a risk factor for the RPOA of the knee. The assignment of this nuclear polymorphism could be useful as complementary genetic biomarker for the early identification of this phenotype.

Disclosure of Interests:

Alejandro Durán-Sotuela: None declared, Mercedes Fernandez-Moreno: None declared, Maria Eugenia Vazquez Mosquera: None declared, Paula Ramos-Louro: None declared, Andrea Dalmao-Fernandez: None declared, Sara Relaño-Fernandez: None declared, Natividad Oreiro: None declared, Francisco J. Blanco Grant/research support from: Sanofi-Aventis, Lilly, Bristol MS, Amgen, Pfizer, Abbvie, TRB Chemedica International, Glaxo SmithKline, Archigen Biotech Limited, Novartis, Nichi-iko pharmaceutical Co, Genentech, Jannsen Research & Development, UCB Biopharma, Centrexion Theurapeutics, Celgene, Roche, Regeneron Pharmaceuticals Inc, Biohope, Corbus Pharmaceutical, Tedec Meiji Pharma, Kiniksa Pharmaceuticals, Ltd, Gilead Sciences Inc, Consultant of: Lilly, Bristol MS, Pfizer, Ignacio Rego-Perez: None declared

FRI0387 A PROGNOSTIC MODEL OF PRE-RADIOGRAPHIC KNEE OSTEOARTHRITIS: DATA FROM THE OSTEOARTHRITIS INITIATIVE

Background:

The improvement of the existing diagnostic methods to detect pre-radiographic knee OA (KOA) may facilitate the development of preventive strategies. It has been postulated that combining biochemical with clinical markers, may increase the prognostic power to detect who is at high risk for developing KOA.

Objectives:

To validate and qualify the ability of 6 proteins with biomarker potential to generate a prognostic model of knee OA prediction through the combination of validated OA biomarkers and clinical markers.

Methods:

In thevalidation phase(Figure 1), 749 sera at the baseline visit belonging to participants from the Osteoarthritis Initiative (OAI) Cohort were randomly selected to blindly quantify 6 biomarkers using in-house custom sandwich microarrays built using the xMAP technology. Among these, only 540 participants have a Kellgren and Lawrence (KL) grade = 0-1 at the beginning of the OAI study in at least one knee. After a follow-up period of 96 months, 209 participants developed KOA in at least one knee (KL ≥ 2) and were classified as incident group, whereas 331 did not developed the disease (KL = 0-1) and were classified as not-incident group. Statistical differences between the outcome groups were assessed by non-parametric Mann-Whitney U tests. In thequalification phase(n=540), univariate regression analyses were carried out to investigate whether the individual biomarkers were associated with the risk of KOA development. A clinical prognostic model was defined by stepwise regression analysis using clinical non-radiographic variables significantly associated with the OA incidence. The utility of the potential biomarkers, alone or in combination, was evaluated by comparing the Area Under the Curve (AUC) of the clinical prognostic model with the biomarkers plus clinical prognostic models. In addition, sen

Figure 1.

Study desing

Results:

The incident group showed significant higher serum concentrations at the baseline visit (p < 0.05) for all the potential biomarkers analyzed in this study. Moreover, 5 of them were also significantly associated with the future appearance of radiographic KOA, yielding Odds Ratios (OR) ≥ 10 per 10 µg/ml increase. Among all the possible combinations, the inclusion of 2 biomarkers to the clinical prognostic model showed a significant improvement of the predictive capacity (AUCs = 0.78 vs 0.82, p= 0,044) with 65% (95% Confidence Interval (95%CI): 60-70%) specificity and 88% (95%CI: 81-91%) sensitivity. Variables included in the regression model and all metrics comparing the biomarkers plus clinical prognostic model with the clinical prognostic model are shown in Figure 2A. The ROC curves of the biomarkers-only model, clinical prognostic model and biomarkers plus clinical prognostic model are represented in Figure 2B.

Figure 2.

(A) Multivariate regression analysis (B) ROC curves of the models

Conclusion:

We have generated a prognostic model for the prediction of KOA by combining biomarkers and clinical variables, which showed a putative utility in the clinical setting by improving the predictive capacity of a clinical prognostic model to identify patients at a higher risk to develop radiographic KOA.

Disclosure of Interests:

Maria Camacho Encina: None declared, Vanesa Balboa-Barreiro: None declared, Ignacio Rego-Perez: None declared, Rocío Paz González: None declared, Valentina Calamia: None declared, Lucía Lourido: None declared, Cristina Ruiz-Romero: None declared, Francisco J. Blanco Grant/research support from: Sanofi-Aventis, Lilly, Bristol MS, Amgen, Pfizer, Abbvie, TRB Chemedica International, Glaxo SmithKline, Archigen Biotech Limited, Novartis, Nichi-iko pharmaceutical Co, Genentech, Jannsen Research & Development, UCB Biopharma, Centrexion Theurapeutics, Celgene, Roche, Regeneron Pharmaceuticals Inc, Biohope, Corbus Pharmaceutical, Tedec Meiji Pharma, Kiniksa Pharmaceuticals, Ltd, Gilead Sciences Inc, Consultant of: Lilly, Bristol MS, Pfizer

THU0011 ANALYSIS OF METABOLIC STATUS IN CYBRIDS REVEALED IMPAIRED METABOLIC FLEXIBILITY IN OA PROCESS

Background:

There are several metabolic pathways involved in cell metabolism, including glycolysis, tricarboxylic acid (TCA) cycle and fatty acid (FA) oxidation. Metabolic flexibility has previously described as the ability to respond or adapt to changes in metabolic demand; assessed by the ability to switch from fat to carbohydrate oxidation. In the last years there is a growing interest to assess the influence of metabolic flexibility, as a mechanism to explain how lipids can accumulate in the tissue. During OA, it has been established a relationship between mitochondrial dysfunction and cellular damage due to impairments in mitochondrial function and metabolic flexibility. Several studies have suggested that fatty acids may play an important role in OA development and progression.

Objectives:

The aim of this work was to examine the differences in glucose and fatty acid metabolism, with special focus on metabolic flexibility, in cybrids from healthy (N) or OA donors.

Methods:

Cybrids were developed using 143B.TK-Rho-0 cell line (nuclear donor) and platelets (mitochondrial donors) from healthy (N) and OA donors. Glucose and FA metabolism were measured using D-[14C(U)]glucose and [1-14C]oleic acid respectively. Metabolic flexibility was evaluated by co-culturing with glucose and oleic acid acutely by using inhibitors against glucose and FA oxidation, 20µM UK5099 and 10µM etomoxir, respectively. Incorporation of FA into lipid droplet (LD) was evaluated by thin layer chromatography and LD were stained by LD540 and analyzed by confocal microscope and flow cytometry. Mitochondrial dynamics was measured by real-time PCR method. Percentage of mitochondrial Anion Superoxide (O2-) production was evaluated incubating cells with MitoSox® using Flow Cytometer. Appropriate statistical analyses were performed with GraphPad Prism v6.

Results:

There were no changes in basal glucose metabolism between cybrids. N cybrids had higher acid-soluble metabolites, reflecting incomplete FA β-oxidation than OA cybrids. Comparing glucose and FA metabolism showed that both types of cybrids preferred to oxidize glucose. Co-culturing with glucose and Oleic acid, increased total cellular uptake and oxidation of glucose in N compared to basal condition (Figure-1) and in this condition the OA cybrids showed an increase in mitochondrial O2-production. Inhibition of FA oxidation by etomoxir increased complete glucose oxidation of N cybrids but not in OA cybrids that had a preference to oxidize oleic acid compared to basal condition. Gene expression of mitofusin-2 (MFN2) was higher in N than OA cybrids under inhibiting conditions. Combine these data indicate that N cybrids are more metabolically flexible and have better adaptative response than OA. Cybrids presented different lipid distribution patterns. Lipid droplet (LD) formation increased in both groups incubated in presence of FA. Furthermore, N cybrids showed less LD formation than OA.

Conclusion:

The results indicated that cybrids from OA patients had reduced metabolic flexibility compared to N cybrids. These results enhance our understanding of the mitochondria metabolism in OA, suggesting a mitochondrial dysfunction and impairment of metabolic flexibility during the OA process.

Disclosure of Interests:

Andrea Dalmao-Fernandez: None declared, Jenny Lund: None declared, Tamara Hermida Gómez: None declared, Maria Eugenia Vazquez Mosquera: None declared, Ignacio Rego-Perez: None declared, Francisco J. Blanco Grant/research support from: Sanofi-Aventis, Lilly, Bristol MS, Amgen, Pfizer, Abbvie, TRB Chemedica International, Glaxo SmithKline, Archigen Biotech Limited, Novartis, Nichi-iko pharmaceutical Co, Genentech, Jannsen Research & Development, UCB Biopharma, Centrexion Theurapeutics, Celgene, Roche, Regeneron Pharmaceuticals Inc, Biohope, Corbus Pharmaceutical, Tedec Meiji Pharma, Kiniksa Pharmaceuticals, Ltd, Gilead Sciences Inc, Consultant of: Lilly, Bristol MS, Pfizer, Mercedes Fernandez-Moreno: None declared

Insights into the genetic architecture of osteoarthritis from stage 1 of the arcOGEN study

Objectives

The genetic aetiology of osteoarthritis has not yet been elucidated. To enable a well-powered genome-wide association study (GWAS) for osteoarthritis, the authors have formed the arcOGEN Consortium, a UK-wide collaborative effort aiming to scan genome-wide over 7500 osteoarthritis cases in a two-stage genome-wide association scan. Here the authors report the findings of the stage 1 interim analysis.

Methods

The authors have performed a genome-wide association scan for knee and hip osteoarthritis in 3177 cases and 4894 population-based controls from the UK. Replication of promising signals was carried out in silico in five further scans (44 449 individuals), and de novo in 14 534 independent samples, all of European descent.

Results

None of the association signals the authors identified reach genome-wide levels of statistical significance, therefore stressing the need for corroboration in sample sets of a larger size. Application of analytical approaches to examine the allelic architecture of disease to the stage 1 genome-wide association scan data suggests that osteoarthritis is a highly polygenic disease with multiple risk variants conferring small effects.

Conclusions

Identifying loci conferring susceptibility to osteoarthritis will require large-scale sample sizes and well-defined phenotypes to minimise heterogeneity.