Next Article in Journal
Investigation of the Effect of Curcumin on Protein Targets in NAFLD Using Bioinformatic Analysis
Next Article in Special Issue
The Controversial Role of HCY and Vitamin B Deficiency in Cardiovascular Diseases
Previous Article in Journal
Stop the Pop: A Mixed-Methods Study Examining Children’s Physical and Emotional Responses during Three Days of Sugary Drink Cessation
Previous Article in Special Issue
A Consistency Model for Identifying the Effects of n-3 and n-6 Fatty Acids on Lipoproteins in Dialysis Patients
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Nutrients
Volume 14
Issue 7
10.3390/nu14071327
nutrients-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Causal Associations of Circulating Lipids with Osteoarthritis: A Bidirectional Mendelian Randomization Study

by
Hongen Meng
,
Li Jiang
,
Zijun Song
and
Fudi Wang
*
The Fourth Affiliated Hospital, School of Public Health, Institute of Translational Medicine, Zhejiang University School of Medicine, Hangzhou 310058, China
*
Author to whom correspondence should be addressed.
Nutrients 2022, 14(7), 1327; https://doi.org/10.3390/nu14071327
Submission received: 1 March 2022 / Revised: 17 March 2022 / Accepted: 18 March 2022 / Published: 22 March 2022
(This article belongs to the Special Issue Lipid Metabolism and Relevance to Chronic Disease)
Review Reports Versions Notes

Abstract

:
Osteoarthritis (OA) imposes an increasing social burden due to global activity limitations, especially among the aged. Links between circulating lipids and OA have been reported; however, confounding data from observational studies have hindered causal conclusions. We used Mendelian randomization (MR) approach to evaluate the genetic causal effects of circulating apolipoproteins and lipoprotein lipids on OA risk. Genetic instruments at the genome-wide significance level (p < 5 × 10−8) were selected from genome-wide association studies (n = 393,193–441,016 individuals). Summary-level OA data were obtained from the UK Biobank (39,427 cases, 378,169 controls). Bidirectional two-sample Mendelian randomization (MR) analyses used MR-Egger, weighted median, and MR-PRESSO for sensitivity analysis. Genetic predisposition to 1-SD increments of Apolipoprotein B (APOB), and low-density lipoprotein (LDL) was associated with a decreased risk of knee or hip OA (KHOA) (odds ratio (OR) = 0.925, 95% confidence interval (95% CI): 0.881–0.972, p = 0.002; OR = 0.898, 95% CI: 0.843–0.957, p = 0.001) and hip OA (HOA) (OR = 0.894; 95% CI: 0.832–0.961, p = 0.002; OR = 0.870 95% CI: 0.797–0.949, p = 0.002). Genetically predicted APOB showed an association with knee OA (KOA) (OR per SD increase, 0.930, 95% CI: 0.876–0.987, p = 0.016). The OR of KOA was 0.899 (95% CI: 0.835–0.968, p = 0.005) for a 1-SD increase in LDL. Apolipoprotein A1, high-density lipoprotein, and triglycerides showed no association. Inverse MR showed no causal effect of KOA, HOA, or KHOA on these serum lipids. Distinct protective genetic-influence patterns were observed for APOB and LDL on OA, offering new insights into relationships between lipids and OA risk and a better understanding of OA etiology.

1. Introduction

Osteoarthritis (OA) is the main cause of activity limitation in the elderly, especially in older women [1]. It is a chronic degenerative joint disease (occurring preferentially in the hands, knees, hips, and feet) or a wear-and-tear disorder. Globally, more than 500 million people suffer from this disease, accounting for 7% of the world population [2], and this is expected to increase to 15–20% of the population by 2050, according to the United Nations [3]. Until now, no effective treatments have been available for OA apart from end-stage surgeries, such as total joint arthroplasty. However, the medical costs of surgeries are increasing [4], and OA takes a considerable personal, economic, and societal toll. Therefore, clarifying the causes and early-stage management of this tricky disease is very important.
OA is a complicated disease that is caused by many genetic, environmental, and systemic factors, including aging, gender, race, inflammation, and metabolism [5,6]. The concept of metabolic OA has recently been put forward, with a crucial role for metabolic disturbances and interactions proposed for the onset and development of OA [6,7]. Gkretsi et al. were the first to report a correlation between abnormal high-density cholesterol (HDL) metabolism and OA pathogenesis through epidemiological studies that revealed a link between increased cholesterol levels and the risk of knee, hip, and/or generalized OA [8]. An association between serum cholesterol levels and OA was subsequently reported with hand OA in women [9] and with knee OA [10]. Nevertheless, many studies have presented contradictory results [11,12], including the lack of a significant correlation between serum cholesterol and radiographic knee OA in an early study from Gothenburg [11], and no association between plasma low density cholesterol (LDL) or low HDL and the radiographic or symptomatic OA risks in the Multicenter Osteoarthritis Study (MOST) cohort [12]. By contrast, two recent large meta-analyses studies have shown a strong relationship between dyslipidemia and OA [13,14]. This unclear relationship between lipids and OA necessitates further clarification of the causal effects of serum lipids on OA.
Mendelian randomization (MR) is a new strategy for investigating causation between different traits based on Mendel’s laws of inheritance. MR uses genetic predisposition to avoid biases, such as sample size and short follow-up duration, in observational studies. At present, only one study has explored the effects of obesity-related risk factors, including body mass index (BMI), bone mineral density, serum HDL and LDL, triglyceride (TG) levels, type 2 diabetes, systolic blood pressure (BP), and C-reactive protein levels, on OA incidence, and the authors claimed no evidence of causality for HDL, LDL, and TG on knee OA (KOA), hip OA (HOA), or hand OA [15]. In the same year, another related study examined the effects of cardiometabolic risk factors (LDL, HDL, TG, BMI, fasting plasma glucose, and BP) on OA and reported protective effects of LDL on OA but no strong evidence for any effects of HDL and TG on OA [16]. Another study also reported protective effects of high LDL on OA risk [17]. However, the genetic instrument variables (IVs) used were all obtained from selected genome-wide association studies (GWASs) conducted in 2013 or even before.
In the present study, we conducted a bidirectional two-sample MR study and focused on serum lipoproteins and lipids, including Apolipoprotein A1 (APOA1), Apolipoprotein B (APOB), HDL, LDL, and TG. Our study was based on the latest databases and explored the correlation between these lipid factors and the risk of OA (KOA, HOA, and OA of the knee or hip (KHOA)).

2. Materials and Methods

2.1. Study Design and Data Sources

The study design overview and the assumptions of the MR study are shown in Supplementary Figure S1. The genetic instruments for the exposures were obtained from published GWASs [18]. Data for OA were obtained from the UK Biobank [19]. Detailed information about the data sources used is displayed in Table 1. All studies had been approved by a relevant ethical review board, and all participants had given informed consent.

2.2. Genetic Instrument Selection

Single nucleotide polymorphisms (SNPs) associated with APOA1, APOB, HDL, LDL, and TG at the genome-wide level were obtained from large GWAS’ [18]. The mean age of participants was 56.9 y (range 39–73 y) and 54.2% of them were women. Detailed information about the GWASs used is presented in Table 1. SNPs without linkage disequilibrium (defined by r2 < 0.001 and clump distance >1 Mb) were used as IVs. All selected SNPs for APOA1, APOB, HDL, LDL, and TG are listed in Supplementary Dataset S1.

2.3. Data Sources for OA

Summary-level data for the associations of exposure-associated SNPs with OA were available from the UK Biobank study [19], with up to 39,427 cases and 378,169 controls of European ancestry. The OA cases were from Arthritis Research UK Osteoarthritis Genetics (arcOGEN, which is a collection of unrelated UK-based individuals of European ancestry with knee and/or hip osteoarthritis from the arcOGEN Consortium) and were ascertained based either on clinical evidence of disease to a level requiring joint replacement or on radiographic evidence of disease (Kellgren–Lawrence grade ≥ 2). The controls were from the United Kingdom Household Longitudinal Study (UKHLS), which is a longitudinal panel survey of 40,000 UK households (England, Scotland, Wales, and Northern Ireland) representative of the UK population. Detailed information regarding the SNPs of the 5 lipids used as IVs to conduct this MR analysis is displayed in Supplementary Dataset S2.
All studies included in the GWASs had been approved by relevant ethical review committees, and participants provided written informed consent. The current study only used summary-level data that was publicly available. Thus, no additional ethical review was required for this study.

2.4. Statistical Analysis

We used the multiplicative random-effects inverse variance weighted (IVW) model as the main statistical method. Three sensitivity analyses, namely the weighted median [20], MR-Egger [21], and MR-PRESSO [22], were performed to examine the consistency of associations and to detect possible pleiotropy. The weighted median method provides consistent causal estimates when more than 50% of the weight comes from valid instruments [20]. The MR-Egger regression can detect pleiotropy by its intercept and provide estimates after the correction for pleiotropy; however, it compromises statistical power [21]. The MR-PRESSO method can detect outliers and provide causal estimates after the removal of identified outliers [22]. The distortion test embedded in the MR-PRESSO analysis can distinguish the difference between the estimates before and after outlier removal [22]. Cochrane’s Q value was used to assess the heterogeneity among SNP estimates in each analysis. The horizontal heterogeneity effect was examined by the IVW test and the MR-Egger regression. A leave-one-out sensitivity analysis was also performed to monitor whether significant associations were dominated by a single SNP. Associations with a p-value < 0.01 (0.05/5 exposures) were deemed significant associations, and associations with a p-value > 0.01 and <0.05 were regarded as suggestive associations. Statistical powers were calculated using the mRnd power calculation tool, assuming the true causal OR was 1.10 [23]. We performed the Mendelian randomization analysis using the Two-Sample MR package, an R package, and a GWAS summary data library developed as a platform for performing Mendelian randomization tests and sensitivity analyses [24,25]. All the statistical analyses were performed using R (version 4.1.1) software (TwoSampleMR [24], MR-PRESSO [22], and MendelianRandomization [26] packages).

3. Results

Genetic liability due to APOB showed an association with a decreased risk of KHOA and HOA and a suggestive association with a decreased risk of KOA (p = 0.002, p = 0.002, and p = 0.016, respectively) (Table 2). These associations remained significant in the other sensitivity analyses. For a 1 SD increment of APOB, the odds ratio (OR) and 95% confidence intervals (CI) for KHOA, KOA, and HOA were OR = 0.925 (95% CI: 0.881–0.972), OR = 0.930 (95% CI: 0.876–0.987), and OR = 0.894 (95% CI: 0.832–0.961), respectively.
The genetic predisposition to higher LDL was associated with a decreased risk of KHOA (OR = 0.898, 95% CI: 0.843–0.957, p = 0.001) (Table 3). The associations remained consistent in the sensitivity analyses using the weighted median and MR-Egger methods (Table 3). Five outliers were detected in the MR-PRESSO analyses; however, the associations persisted after the removal of these outliers (Table 3). Only limited data supported associations of genetically predicted LDL with KOA (OR = 0.899, 95% CI: 0.835–0.968, p = 0.005) or HOA (OR = 0.870, 95% CI: 0.797–0.949, p = 0.002) (Table 3). For KOA, the protective effect persisted in the sensitivity analyses in the MR-Egger methods and MR-PRESSO analyses after the removal of five outliers, but not in the weighted median methods. For HOA, the negative association was borderline in the weighted median and MR-Egger methods but became significant in MR-PRESSO analyses after the removal of five outliers. No associations were detected for APOA1, HDL, and TG with OA in the primary analysis or in the sensitivity analyses of each data source (Supplementary Tables S1–S3).
The intercept in the MR-Egger regression indicated no pleiotropy in any analysis (all p values > 0.05) (Table 2 and Table 3, Supplementary Tables S1–S3). No clear violations of the MR assumptions were indicated after generating forest, funnel, and scatter plots (Supplementary Figures S2–S4).
An inverse MR of OA on serum lipids was similarly conducted, except the Bonferroni-adjusted p < 0.05/3 = 0.017 was used. No causal effect was detected for KOA, HOA, or KHOA on serum lipids (APOA1, APOB, HDL, LDL, and TG) using 7 IVs of KOA, 23 IVs of HOA, 22 IVs of KHOA as exposures and serum lipids as outcomes, through multiple MR sensitivity analysis (Supplementary Table S4).

4. Discussion

The multiple MR sensitive analyses led to the conclusion that circulating APOB was negatively associated with the risk of KHOA and HOA, and that serum LDL was negatively associated with the risk of KHOA, KOA, and HOA. No other genetic relationships were detected between other lipids and the risk of KOA, HOA, or KHOA, nor was there a genetic relationship between KOA, HOA, or KHOA and these five lipids.
This study is the first one to demonstrate a protective effect of APOB on KOA, HOA, and KHOA, despite previous demonstrations of no or positive correlations with OA risks in a number of studies. For example, Zhang et al. found no significant differences in the APOB levels in the serum and synovial fluid of primary KOA patients or healthy individuals [27]. In a mouse model of feeding a cholesterol-rich diet, APOB accumulation in synovial macrophages did not affect the thickening of the synovial lining or induce cartilage damage [28]. Similarly, the mutations in APOB (rs693 and rs1042031) increased the risk of steroid-induced osteonecrosis of the femoral head [29]. However, an osteochondrosis study of horses determined that APOB and APOB-3G (similar to Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G) from leukocytes were relatively under-expressed in the osteochondrosis-affected group, indicating an inverse correlation between APOB and osteochondrosis which culminated in OA [30]. As the apolipoprotein of LDL and the ligand for the LDL receptor, the protective effects of APOB on OA might be mediated through LDL, which was negatively associated with the risk of KHOA, KOA, and HOA in this study.
LDL has also shown an association with OA. One comparative study showed that the mean level of LDL was significantly (p < 0.01) higher in KOA [31], in agreement with another study of primary KOA [22]. A case-control study showed that LDL was independently associated with the radiographic severity of KHOA [7]. An experimental study showed that high LDL participated in the progression of KOA through synovial inflammation and ectopic bone formation [32]. Many other studies have supported an association between the oxidation of low-density lipoprotein (ox-LDL) and OA [33,34,35]. However, de Seny et al. found that LDL could decrease APOA1 levels and serum amyloid A protein-induced joint inflammation in human primary chondrocytes and fibroblast-like synoviocytes [36]. A recent MR study using the Malmö Diet and Cancer Study cohort [17] also claimed that LDL levels show a genetically inverse relationship to OA risks, which supports our present findings. Whether the simultaneous elevation of APOB and LDL exerts a protective effect is unclear and further studies are needed.
Some epidemiological and experimental studies have also concluded that no relationship exists between OA and either APOA1, HDL, or TG. One previous study reported no significant difference in serum APOA1 levels between OA patients and healthy controls [37]. Schwager et al. also found no significant association between HDL and OA based on the Multicenter Osteoarthritis Study (MOST) cohort [12]. Genetic causality between HDL and TG with OA was also absent in a UKBB based MR study [15]. The use of a statin (a serum-lipid lowering drug) had no impact on KOA progression [38], implying no effect of HDL and TG on OA.
The underlying mechanisms of the protective effect of APOB and LDL on OA remain to be explored through additional research. APOB in the synovial fluid appears to be important in the response to the TNFα-NFκB inflammatory signal through alpha-enolase (ENO1), independent of the LDL receptor [1]. Much research is now focused on the effect of ox-LDL on OA. The lectin-like oxidized low density lipoprotein receptor 1 (ox-LDL/LOX1) system might be involved in the progression of OA through vascular endothelial growth factor (VEGF)/ peroxisome proliferators-activated receptor γ (PPAR-γ) [35], monocyte chemoattractant protein 1 (MCP-1) [33], IL-1β [39], and ROS [40,41]. A recent study reported that feeding a serum LDL-increasing western diet to a mouse model of KOA did not play a role in the Nox isoform 2 (Nox2)-mediated ROS production or the IL-1β-mediated pro-inflammatory effect [42,43], indicating that LDL might not be deleterious for OA. Consistent with this, a high-fat diet intake in healthy and OA rabbits did not change the expression of pro-inflammatory and catabolic (IL-1β, IL-6, matrix metallopeptidase 13 (MMP-13), MCP1, or cyclooxygenase-2 (COX-2)) genes or proteins, nor did it modify the cartilage structure. Further, in cultured human OA articular chondrocytes, ox-LDL did not affect chondrocyte viability in the presence or absence of IL-1β and TNFα pro-inflammatory stimuli and even decreased the induction of IL-1β, IL-6, MCP-1, MMP-13, iNOS, and COX-2 gene expression increased MMP-13 and COX-2 protein levels [44]. Ascone et al. also showed that the protective effects of LDL on OA arose due to attenuation of the onset of inflammation and cartilage destruction through Fc gamma Receptors (FcγRs) [45]. Further studies are warranted to identify the potential mechanisms by which APOB and LDL protect against the occurrence and progression of OA.
To the best of our knowledge, this study is the largest genetic correlation study of the effects of serum APOA1, APOB, HDL, LDL, and TG levels on KOA, HOA, and KHOA using MR analyses based on a bidirectional two-sample MR approach. The datasets employed were the most recent and largest, and both came from UKBB; therefore, the cohorts share the same European descent, and this reduces the bias caused by population stratification. The use of strict SNP selection terms, with LD < 0.001 and p < 0.01, and choice of SNPs at locations ≥1 MB distant from each other provided further increases in precision and statistical power. In addition to using MR-Egger, weighted median, IVW, and MR-PRESSO (raw, corrected outliers) methods in the MR sensitivity analysis, we also employed an inverse MR sensitivity analysis of OA on lipids to minimize the bias and provide a strong causal result. A few studies have previously shown a correlation between APOB or LDL and OA, but we have further analyzed the genetically correct effect to rule out false-positive results.
Despite all these efforts, the present study still has some limitations. One is that only European descent participants were included, but African Americans have a high incidence of both KOA and HOA [46,47]. A second limitation is that only APOA1, APOB, HDL, LDL, and TG were examined. More genetic IVs of other lipid components should be taken into account to draw more accurate conclusions. Another limitation is that, even with the use of the weighted median method, bias caused by heterogeneity cannot be fully avoided. Moreover, although our reverse MR analysis found no significant causal evidence, the effect of OA on blood lipids requires further investigation based on independent GWAS and large prospective studies. Nonetheless, our study offers new insights into the relationships between lipids and the risk of OA and provides a better understanding of OA etiology.

5. Conclusions

A genetic protective effect was found for APOB and LDL on OA based on bidirectional MR analysis. Specifically, increased APOB was negatively related to the risks of KOA, HOA, and KHOA, and elevated LDL decreased the risks of KHOA.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu14071327/s1. Supplementary Text S1. Mendelian randomization analysis of osteoarthritis (OA) on lipids; Supplementary Table S1 Associations of genetically predicted APOA1 with OA risks in MR analyses; Supplementary Table S2 Associations of genetically predicted HDL with OA risks in MR analyses; Supplementary Table S3 Associations of genetically predicted TG with OA risks in MR analyses; Supplementary Table S4 Associations of genetically predicted OA with lipids in MR analyses; Supplementary Figure S1. Diagram of Mendelian randomization framework in the current paper; Supplementary Figure S2. Forest plots of variant specific inverse variance estimates for causal association between lipids and OA; Supplementary Figure S3. Funnel plots of causal association between lipids and OA; Supplementary Figure S4. Leave-one-out plots to assess if a single variant is driving the association between lipids and OA.

Author Contributions

The authors’ responsibilities were as follows—H.M., L.J., Z.S. and F.W.: designed research; H.M.: conducted research with assistance from L.J. and Z.S.; H.M. and Z.S.: performed data analysis; H.M., L.J. and F.W.: wrote the manuscript and had primary responsibility for final content. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by grants from the National Natural Science Foundation of China (31930057 (F.W.)), the National Key R&D Program of China (2018YFA0507802 (F.W.)), and the National Natural Science Foundation of China (32000820 (L.J.)), and the China Postdoctoral Science Foundation (2019M662028 (L.J.)).

Institutional Review Board Statement

Appropriate patient consent and ethics approval were obtained in the original studies from which the data for this work were obtained [18,19].

Informed Consent Statement

Not applicable.

Data Availability Statement

All data are available in the submitted manuscript or related sources described in the manuscript. Supplementary datasets were hosted on the Figshare website and cited as Supplementary. Data described in the manuscript, code book, and analytic code will be made available upon request pending.

Acknowledgments

The authors thank all the investigators of the arcOGEN Consortium and UK Biobank for providing the data publicly.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Katz, J.N.; Arant, K.R.; Loeser, R.F. Diagnosis and Treatment of Hip and Knee Osteoarthritis: A Review. JAMA 2021, 325, 568–578. [Google Scholar] [CrossRef] [PubMed]
  2. Hunter, D.J.; March, L.; Chew, M. Osteoarthritis in 2020 and beyond: A Lancet Commission. Lancet 2020, 396, 1711–1712. [Google Scholar] [CrossRef]
  3. Papathanasiou, I.; Anastasopoulou, L.; Tsezou, A. Cholesterol metabolism related genes in osteoarthritis. Bone 2021, 152, 116076. [Google Scholar] [CrossRef] [PubMed]
  4. Delanois, R.E.; Mistry, J.B.; Gwam, C.U.; Mohamed, N.S.; Choksi, U.S.; Mont, M.A. Current Epidemiology of Revision Total Knee Arthroplasty in the United States. J. Arthroplast. 2017, 32, 2663–2668. [Google Scholar] [CrossRef] [PubMed]
  5. Hunter, D.J.; Bierma-Zeinstra, S. Osteoarthritis. Lancet 2019, 393, 1745–1759. [Google Scholar] [CrossRef]
  6. Bijlsma, J.W.; Berenbaum, F.; Lafeber, F.P. Osteoarthritis: An update with relevance for clinical practice. Lancet 2011, 377, 2115–2126. [Google Scholar] [CrossRef]
  7. Tootsi, K.; Märtson, A.; Kals, J.; Paapstel, K.; Zilmer, M. Metabolic factors and oxidative stress in osteoarthritis: A case-control study. Scand. J. Clin. Lab. Investig. 2017, 77, 520–526. [Google Scholar] [CrossRef] [PubMed]
  8. Gkretsi, V.; Simopoulou, T.; Tsezou, A. Lipid metabolism and osteoarthritis: Lessons from atherosclerosis. Prog. Lipid Res. 2011, 50, 133–140. [Google Scholar] [CrossRef]
  9. Kellgren, J.H. Osteoarthrosis in patients and populations. Br. Med. J. 1961, 2, 1–6. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  10. Hart, D.J.; Doyle, D.V.; Spector, T.D. Association between metabolic factors and knee osteoarthritis in women: The Chingford Study. J. Rheumatol. 1995, 22, 1118–1123. [Google Scholar] [PubMed]
  11. Bagge, E.; Bjelle, A.; Edén, S.; Svanborg, A. Factors associated with radiographic osteoarthritis: Results from the population study 70-year-old people in Göteborg. J. Rheumatol. 1991, 18, 1218–1222. [Google Scholar]
  12. Schwager, J.L.; Nevitt, M.C.; Torner, J.; Lewis, C.E.; Matthan, N.R.; Wang, N.; Sun, X.; Lichtenstein, A.H.; Felson, D.; Multicenter Osteoarthritis Study Group. Association of Serum Low-Density Lipoprotein, High-Density Lipoprotein, and Total Cholesterol With Development of Knee Osteoarthritis. Arthritis Care Res. 2022, 74, 274–280. [Google Scholar] [CrossRef] [PubMed]
  13. Liu, S.Y.; Zhu, W.T.; Chen, B.W.; Chen, Y.H.; Ni, G.X. Bidirectional association between metabolic syndrome and osteoarthritis: A meta-analysis of observational studies. Diabetol. Metab. Syndr. 2020, 12, 38. [Google Scholar] [CrossRef]
  14. Xiong, J.; Long, J.; Chen, X.; Li, Y.; Song, H. Dyslipidemia Might Be Associated with an Increased Risk of Osteoarthritis. Biomed Res. Int. 2020, 2020, 3105248. [Google Scholar] [CrossRef]
  15. Funck-Brentano, T.; Nethander, M.; Movérare-Skrtic, S.; Richette, P.; Ohlsson, C. Causal Factors for Knee, Hip, and Hand Osteoarthritis: A Mendelian Randomization Study in the UK Biobank. Arthritis Rheumatol. 2019, 71, 1634–1641. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  16. Gill, D.; Karhunen, V.; Malik, R.; Dichgans, M.; Sofat, N. Cardiometabolic traits mediating the effect of education on osteoarthritis risk: A Mendelian randomization study. Osteoarthr. Cartil. 2021, 29, 365–371. [Google Scholar] [CrossRef] [PubMed]
  17. Hindy, G.; Åkesson, K.E.; Melander, O.; Aragam, K.G.; Haas, M.E.; Nilsson, P.M.; Kadam, U.T.; Orho-Melander, M. Cardiometabolic Polygenic Risk Scores and Osteoarthritis Outcomes: A Mendelian Randomization Study Using Data From the Malmö Diet and Cancer Study and the UK Biobank. Arthritis Rheumatol. 2019, 71, 925–934. [Google Scholar] [CrossRef] [PubMed]
  18. Richardson, T.G.; Sanderson, E.; Palmer, T.M.; Ala-Korpela, M.; Ference, B.A.; Davey Smith, G.D.; Holmes, M.V. Evaluating the relationship between circulating lipoprotein lipids and apolipoproteins with risk of coronary heart disease: A multivariable Mendelian randomisation analysis. PLoS Med. 2020, 17, e1003062. [Google Scholar] [CrossRef] [Green Version]
  19. Tachmazidou, I.; Hatzikotoulas, K.; Southam, L.; Esparza-Gordillo, J.; Haberland, V.; Zheng, J.; Johnson, T.; Koprulu, M.; Zengini, E.; Steinberg, J.; et al. Identification of new therapeutic targets for osteoarthritis through genome-wide analyses of UK Biobank data. Nat. Genet. 2019, 51, 230–236. [Google Scholar] [CrossRef] [PubMed]
  20. Bowden, J.; Davey Smith, G.; Haycock, P.C.; Burgess, S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet. Epidemiol. 2016, 40, 304–314. [Google Scholar] [CrossRef] [Green Version]
  21. Bowden, J.; Davey Smith, G.; Burgess, S. Mendelian randomization with invalid instruments: Effect estimation and bias detection through Egger regression. Int. J. Epidemiol. 2015, 44, 512–525. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  22. Verbanck, M.; Chen, C.Y.; Neale, B.; Do, R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat. Genet. 2018, 50, 693–698. [Google Scholar] [CrossRef]
  23. Brion, M.J.; Shakhbazov, K.; Visscher, P.M. Calculating statistical power in Mendelian randomization studies. Int. J. Epidemiol. 2013, 42, 1497–1501. [Google Scholar] [CrossRef] [Green Version]
  24. Hartwig, F.P.; Davies, N.M.; Hemani, G.; Davey Smith, G. Two-sample Mendelian randomization: Avoiding the downsides of a powerful, widely applicable but potentially fallible technique. Int. J. Epidemiol. 2016, 45, 1717–1726. [Google Scholar] [CrossRef] [Green Version]
  25. Hemani, G.; Zheng, J.; Elsworth, B.; Wade, K.H.; Haberland, V.; Baird, D.; Laurin, C.; Burgess, S.; Bowden, J.; Langdon, R.; et al. The MR-Base platform supports systematic causal inference across the human phenome. Elife 2018, 7, e34408. [Google Scholar] [CrossRef]
  26. Yavorska, O.O.; Burgess, S. MendelianRandomization: An R package for performing Mendelian randomization analyses using summarized data. Int. J. Epidemiol. 2017, 46, 1734–1739. [Google Scholar] [CrossRef]
  27. Zhang, K.; Ji, Y.; Dai, H.; Khan, A.A.; Zhou, Y.; Chen, R.; Jiang, Y.; Gui, J. High-Density Lipoprotein Cholesterol and Apolipoprotein A1 in Synovial Fluid: Potential Predictors of Disease Severity of Primary Knee Osteoarthritis. Cartilage 2021, 13, 1465S–1473S. [Google Scholar] [CrossRef]
  28. Miyanishi, K.; Yamamoto, T.; Irisa, T.; Noguchi, Y.; Sugioka, Y.; Iwamoto, Y. Increased level of apolipoprotein B/apolipoprotein A1 ratio as a potential risk for osteonecrosis. Ann. Rheum. Dis. 1999, 58, 514–516. [Google Scholar] [CrossRef] [PubMed]
  29. Chen, X.; Zhang, L.; Liang, D.; Li, J.; Liu, F.; Ma, H. Lipid Transporter Activity-Related Genetic Polymorphisms Are Associated with Steroid-Induced Osteonecrosis of the Femoral Head: An Updated Meta-Analysis Based on the GRADE Guidelines. Front. Physiol. 2018, 9, 1684. [Google Scholar] [CrossRef] [PubMed]
  30. Serteyn, D.; Piquemal, D.; Vanderheyden, L.; Lejeune, J.P.; Verwilghen, D.; Sandersen, C. Gene expression profiling from leukocytes of horses affected by osteochondrosis. J. Orthop. Res. 2010, 28, 965–970. [Google Scholar] [CrossRef]
  31. Mishra, R.; Singh, A.; Chandra, V.; Negi, M.P.; Tripathy, B.C.; Prakash, J.; Gupta, V. A comparative analysis of serological parameters and oxidative stress in osteoarthritis and rheumatoid arthritis. Rheumatol. Int. 2012, 32, 2377–2382. [Google Scholar] [CrossRef] [PubMed]
  32. De Munter, W.; Bosch, M.H.V.D.; Slöetjes, A.W.; Croce, K.J.; Vogl, T.; Roth, J.; Koenders, M.I.; van de Loo, F.A.; Berg, W.B.V.D.; van der Kraan, P.M.; et al. High LDL levels lead to increased synovial inflammation and accelerated ectopic bone formation during experimental osteoarthritis. Osteoarthr. Cartil. 2016, 24, 844–855. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  33. Akagi, M.; Ueda, A.; Teramura, T.; Kanata, S.; Sawamura, T.; Hamanishi, C. Oxidized LDL binding to LOX-1 enhances MCP-1 expression in cultured human articular chondrocytes. Osteoarthr. Cartil. 2009, 17, 271–275. [Google Scholar] [CrossRef] [Green Version]
  34. Ertürk, C.; Altay, M.A.; Bilge, A.; Çelik, H. Is there a relationship between serum ox-LDL, oxidative stress, and PON1 in knee osteoarthritis? Clin. Rheumatol. 2017, 36, 2775–2780. [Google Scholar] [CrossRef]
  35. Kanata, S.; Akagi, M.; Nishimura, S.; Hayakawa, S.; Yoshida, K.; Sawamura, T.; Munakata, H.; Hamanishi, C. Oxidized LDL binding to LOX-1 upregulates VEGF expression in cultured bovine chondrocytes through activation of PPAR-gamma. Biochem. Biophys. Res. Commun. 2006, 348, 1003–1010. [Google Scholar] [CrossRef] [PubMed]
  36. De Seny, D.; Cobraiville, G.; Charlier, E.; Neuville, S.; Lutteri, L.; Le Goff, C.; Malaise, D.; Malaise, O.; Chapelle, J.-P.; Relic, B.; et al. Apolipoprotein-A1 as a damage-associated molecular patterns protein in osteoarthritis: Ex vivo and in vitro pro-inflammatory properties. PLoS ONE 2015, 10, e0122904. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  37. Oliviero, F.; Sfriso, P.; Baldo, G.; Dayer, J.-M.; Giunco, S.; Scanu, A.; Bernardi, D.; Ramonda, R.; Plebani, M.; Punzi, L. Apolipoprotein A-I and cholesterol in synovial fluid of patients with rheumatoid arthritis, psoriatic arthritis and osteoarthritis. Clin. Exp. Rheumatol. 2009, 27, 79–83. [Google Scholar]
  38. Riddle, D.L.; Moxley, G.; Dumenci, L. Associations between statin use and changes in pain, function and structural progression: A longitudinal study of persons with knee osteoarthritis. Ann. Rheum. Dis. 2013, 72, 196–203. [Google Scholar] [CrossRef] [PubMed]
  39. Nakagawa, T.; Yasuda, T.; Hoshikawa, H.; Shimizu, M.; Kakinuma, T.; Chen, M.; Masaki, T.; Nakamura, T.; Sawamura, T. LOX-1 expressed in cultured rat chondrocytes mediates oxidized LDL-induced cell death-possible role of dephosphorylation of Akt. Biochem. Biophys. Res. Commun. 2002, 299, 91–97. [Google Scholar] [CrossRef]
  40. Nishimura, S.; Akagi, M.; Yoshida, K.; Hayakawa, S.; Sawamura, T.; Munakata, H.; Hamanishi, C. Oxidized low-density lipoprotein (ox-LDL) binding to lectin-like ox-LDL receptor-1 (LOX-1) in cultured bovine articular chondrocytes increases production of intracellular reactive oxygen species (ROS) resulting in the activation of NF-kappa B. Osteoarthr. Cartil. 2004, 12, 568–576. [Google Scholar] [CrossRef] [Green Version]
  41. Akagi, M.; Nishimura, S.; Yoshida, K.; Kakinuma, T.; Sawamura, T.; Munakata, H.; Hamanishi, C. Cyclic tensile stretch load and oxidized low density lipoprotein synergistically induce lectin-like oxidized ldl receptor-1 in cultured bovine chondrocytes, resulting in decreased cell viability and proteoglycan synthesis. J. Orthop. Res. 2006, 24, 1782–1790. [Google Scholar] [CrossRef] [PubMed]
  42. Kruisbergen, N.N.L.; Di Ceglie, I.; van Gemert, Y.; Walgreen, B.; Helsen, M.M.A.; Slöetjes, A.W.; Koenders, M.I.; van de Loo, F.A.J.; Roth, J.; Vogl, T.; et al. Nox2 Deficiency Reduces Cartilage Damage and Ectopic Bone Formation in an Experimental Model for Osteoarthritis. Antioxidants 2021, 10, 1660. [Google Scholar] [CrossRef] [PubMed]
  43. Kruisbergen, N.N.L.; van Gemert, Y.; Walgreen, B.; Helsen, M.M.A.; Slöetjes, A.W.; Koenders, M.I.; van de Loo, F.A.J.; Roth, J.; Vogl, T.; van der Kraan, P.M.; et al. A single dose of anti-IL-1β antibodies prevents Western diet-induced immune activation during early stage collagenase-induced osteoarthritis, but does not ameliorate end-stage pathology. Osteoarthr. Cartil. 2021, 29, 1462–1473. [Google Scholar] [CrossRef] [PubMed]
  44. Villalvilla, A.; Larrañaga-Vera, A.; Lamuedra, A.; Pérez-Baos, S.; López-Reyes, A.G.; Herrero-Beaumont, G.; Largo, R. Modulation of the Inflammatory Process by Hypercholesterolemia in Osteoarthritis. Front. Med. 2020, 7, 566250. [Google Scholar] [CrossRef] [PubMed]
  45. Ascone, G.; Di Ceglie, I.; Bosch, M.H.J.V.D.; Kruisbergen, N.N.L.; Walgreen, B.; Sloetjes, A.W.; Lindhout, E.; Joosten, L.A.B.; Loo, F.A.J.V.D.; Koenders, I.M.; et al. High LDL-C levels attenuate onset of inflammation and cartilage destruction in antigen-induced arthritis. Clin. Exp. Rheumatol. 2019, 37, 983–993. [Google Scholar] [PubMed]
  46. Allen, K.D.; Golightly, Y.M. State of the evidence. Curr. Opin. Rheumatol. 2015, 27, 276–283. [Google Scholar] [CrossRef] [PubMed]
  47. Vina, E.R.; Kwoh, C.K. Epidemiology of osteoarthritis: Literature update. Curr. Opin. Rheumatol. 2018, 30, 160–167. [Google Scholar] [CrossRef]
Table
Table 1. Characteristics of UK Biobank datasets.
Table 1. Characteristics of UK Biobank datasets.
Exposures ConsortiumNo. SNPsSample Size AdjustmentsPopulation
APOA1UK Biobank299393,193 Age, sex, and genotyping chip arrayEuropean
APOBUK Biobank198439,214
HDLUK Biobank362403,943
LDLUK Biobank177440,546
TGUK Biobank313441,016
Main Outcomes DatasetNo. CasesControlTotalAdjustmentsPopulation
KHOAUK Biobank39,427378,169417,596Age, sex, genotyping chip array, and 10 genetic principal componentsEuropean
KOAUK Biobank24,955378,169403,124
HOAUK Biobank15,704378,169393,873
Abbreviations: APOA1, Apolipoprotein A1; APOB, Apolipoprotein B; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; TG, triglycerides; SNP, single-nucleotide polymorphism; KHOA, Osteoarthritis of knee or hip; KOA, Knee OA; HOA, Hip OA.
Table
Table 2. Associations of genetically predicted APOB with OA risks in MR analyses.
Table 2. Associations of genetically predicted APOB with OA risks in MR analyses.
Main OutcomeMethodNo. of SNPsOR (95% CI)P for AssociationP for Heterogeneity TestP for MR-Egger InterceptP for MR-PRESSO Global TestStatistical
Power
KHOAIVW1880.925 (0.881–0.972)0.002 8.23 × 10−120.09 1.00
MR Egger1880.889 (0.832–0.950)0.001 2.21 × 10−11
Weighted median1880.900 (0.848–0.956)0.001
MR-PRESSO (raw,3outliers)1850.927 (0.924–0.930)0.001 <1 × 10−4
KOAIVW1880.930 (0.876–0.987)0.016 3.81 × 10−110.19 1.00
MR Egger1880.896 (0.827–0.972)0.009 5.79 × 10−11
Weighted median1880.892 (0.824–0.966)0.005
MR-PRESSO (raw,2outliers)1860.927 (0.923–0.930)0.006 <1 × 10−4
HOAIVW1880.894 (0.832–0.961)0.002 5.05 × 10−90.44 0.96
MR Egger1880.871 (0.789–0.961)0.0064.79 × 10−9
Weighted median1880.873 (0.800–0.953)0.002
MR-PRESSO (raw,4outliers)1840.881 (0.877–0.885)<0.001 <1 × 10−4
Abbreviations: APOB, Apolipoprotein B; OA, Osteoarthritis; KHOA, Osteoarthritis of the knee or hip; KOA, Knee OA; HOA, Hip OA; IVW, multiplicative random-effects inverse variance-weighted; SNP, single-nucleotide polymorphism; OR, odds ratio; CI, confidence interval.
Table
Table 3. Associations of genetically predicted LDL with OA risks in MR analyses.
Table 3. Associations of genetically predicted LDL with OA risks in MR analyses.
Main OutcomeMethodNo. of SNPsOR (95% CI)P for AssociationP for Heterogeneity TestP for MR-Egger InterceptP for MR-PRESSO Global TestStatistical
Power
KHOAIVW1620.898 (0.843–0.957)0.001 3.24 × 10−180.18 1.00
MR Egger1620.856 (0.778–0.941)0.002 6.80 × 10−18
Weighted median1620.867 (0.810–0.927)<0.001
MR-PRESSO (raw, 5outliers)1570.901 (0.897–0.905)<0.001 <1 × 10−4
KOAIVW1630.899 (0.835–0.968)0.005 2.99 × 10−140.25 0.97
MR Egger1630.857 (0.767–0.957)0.007 4.22 × 10−14
Weighted median1630.934 (0.854–1.021)0.135
MR-PRESSO (raw, 5outliers)1580.900 (0.896–0.905)0.002 <1 × 10−4
HOAIVW1630.870 (0.797–0.949)0.002 1.07 × 10−90.56 0.88
MR Egger1630.845 (0.741–0.963)0.012 9.06 × 10−10
Weighted median1630.891 (0.798–0.994)0.039
MR-PRESSO (raw, 5outliers)1580.863 (0.858–0.868)<0.001 <1 × 10−4
Abbreviations: LDL, low-density lipoprotein cholesterol; OA, Osteoarthritis; KHOA, Osteoarthritis of the knee or hip; KOA, Knee OA; HOA, Hip OA; IVW, multiplicative random-effects inverse variance-weighted; SNP, single-nucleotide polymorphism; OR, odds ratio; CI, confidence interval.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Meng, H.; Jiang, L.; Song, Z.; Wang, F. Causal Associations of Circulating Lipids with Osteoarthritis: A Bidirectional Mendelian Randomization Study. Nutrients 2022, 14, 1327. https://doi.org/10.3390/nu14071327

AMA Style

Meng H, Jiang L, Song Z, Wang F. Causal Associations of Circulating Lipids with Osteoarthritis: A Bidirectional Mendelian Randomization Study. Nutrients. 2022; 14(7):1327. https://doi.org/10.3390/nu14071327

Chicago/Turabian Style

Meng, Hongen, Li Jiang, Zijun Song, and Fudi Wang. 2022. "Causal Associations of Circulating Lipids with Osteoarthritis: A Bidirectional Mendelian Randomization Study" Nutrients 14, no. 7: 1327. https://doi.org/10.3390/nu14071327

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop