Rheumatoid arthritis (RA) is an autoimmune and inflammatory disorder that leads to cartilage and bone deterioration. This inflammatory activity causes extra-articular manifestations, including the acceleration of the atherosclerotic process. However, the exact causes of this accelerated process are under investigation. In this study, we compared the advanced lipid profile between patients with RA, patients with metabolic disorders, and controls. We also explored how microRNAs previously associated with subclinical atherosclerosis in RA are linked to these lipid subfractions in RA.
MethodsThe study included 219 patients with RA, 82 with metabolic disorders and 64 controls. Clinical evaluations were performed, and blood samples were collected. Quantification of microRNAs (Let7a, 24, 96, 103, 125a, 125b, 132, 146, 191, 223, 425, 451) and measurement of the advanced lipid profile using nuclear magnetic resonance (NMR) were carried out. Kruskal–Wallis tests and multivariate linear models were applied.
ResultsPatients with RA exhibited elevated total, large, medium, and small VLDL particles compared to controls. Their LDL subfractions were decreased compared to patients with metabolic disorders, with differences with controls. Patients with RA had fewer and smaller HDL particles than both groups. MicroRNA-125a was associated with VLDL subfractions and small LDL particles. Other microRNAs (96, 132, 191, 451) showed associations with certain LDL subfractions.
ConclusionsIn patients with RA, elevated levels of VLDL particles have been observed, while LDL levels remain similar to controls. The notable association of microRNA-125a with the metabolism of both VLDL and LDL in RA patients suggests its involvement in lipid regulation. This could point to microRNA-125a as a promising therapeutic target to address the increased cardiovascular risks of RA.
La artritis reumatoide (AR) es un trastorno autoinmunitario e inflamatorio que provoca el deterioro de cartílagos y huesos. Esta actividad inflamatoria causa manifestaciones extraarticulares, como la aceleración del proceso aterosclerótico. Sin embargo, las causas exactas de este proceso acelerado aún se están investigando. En este estudio, comparamos el perfil lipídico avanzado entre pacientes con AR, pacientes con trastornos metabólicos y controles. También exploramos cómo micro-ARN previamente asociados con aterosclerosis subclínica en AR están asociados a estas subfracciones lipídicas en AR.
MétodosEl estudio incluyó a 219 pacientes con AR, 82 con trastornos metabólicos y 64 controles. Se realizaron evaluaciones clínicas y se recogieron muestras de sangre. Se cuantificaron micro-ARN (Let7a, 24, 96, 103, 125a, 125b, 132, 146, 191, 223, 425, 451) y se midió el perfil lipídico avanzado utilizando resonancia magnética nuclear. Se aplicaron pruebas de Kruskal-Wallis y modelos lineales multivariados.
ResultadosLos pacientes con AR exhibieron partículas VLDL elevadas (totales, grandes, medianas y pequeñas) en comparación con los controles. Sus subfracciones LDL eran menores comparadas con las de los pacientes metabólicos, sin diferencias respecto a los controles. Los pacientes con AR tenían menos partículas de HDL y de menor tamaño en comparación con los otros grupos. El micro-ARN-125a se asoció con las subfracciones VLDL y partículas pequeñas de LDL. Otros micro-ARN (96, 132, 191, 451) también mostraron asociaciones con subfracciones LDL.
ConclusionesEn pacientes con AR se observaron niveles elevados de partículas VLDL, mientras que los niveles de LDL permanecen similares a los controles. La notable asociación del micro-ARN-125a con el metabolismo tanto de VLDL como de LDL en pacientes con AR evidencia su participación en la regulación lipídica. Esto podría señalar al micro-ARN-125a como un objetivo terapéutico prometedor para abordar los riesgos cardiovasculares aumentados de la AR.
Rheumatoid arthritis (RA) is a chronic, autoimmune, inflammatory disorder that causes inflammation in articulations and joints. Without treatment, it can result in persistent pain, deterioration of cartilage and bone, and worsening disability. Affecting about 1% of the population, RA is the most common chronic inflammatory condition and occurs more frequently in women.1 The inflammation caused by RA extends beyond the joints, impacting other areas of the body and causing pulmonary, hematological, and ocular affectations. Moreover, the inflammatory process also affects the vascular system, accelerating the atherosclerotic process and consequently increasing the risk of cardiovascular disease.2,3 However, the relationship between inflammation and its impact on the atherosclerotic process in RA remains a subject of ongoing study. Therefore, new biomarkers are needed to better comprehend the underlying pathways of the inflammatory process of RA and its association with cardiovascular disease.
In this context, various microRNAs have been linked to the inflammatory pathophysiology of RA as well as to atherosclerosis and cardiovascular disease associated with this condition.4,5 MicroRNAs are small non-coding RNA molecules, 21–25 nucleotides in length that regulate gene expression by repressing translation or directly cleaving RNA. Their ability to target multiple RNAs allows them to play key roles in numerous biological processes, such as cellular differentiation, immunity, atherosclerosis, and inflammation.6 In this direction, both our group and others have identified several microRNAs as potential predictors of cardiovascular disease in patients with RA. Specifically, we observed a strong relationship between microRNA 24, 146, and Let7a expression levels and carotid plaque presence in male patients with RA.7 Other microRNAs such as 125b, 132, 96, 425, or 451, were also associated with carotid intima-media thickness or arterial stiffness in patients with RA.7,8 Other findings also indicate that microRNAs 24, 146, and Let7a, both individually and collectively, showed strong associations with the inflammatory parameters of the disease.9 Taken together, these findings suggest that microRNAs may play a key role in connecting inflammation and atherosclerosis in RA. However, it is essential to acknowledge that further comprehensive and longitudinal studies are required to establish definitive causal links and to fully elucidate the underlying mechanisms.
Apart from microRNAs, other metabolic biomarkers, such as the different lipoprotein subfractions measured by proton nuclear magnetic resonance (1H-NMR), have been shown to be altered in patients with RA. In this direction, it has been observed that patients with RA show increased number of large HDL particles compared to healthy subjects, which are suspected to be less protective against atherosclerosis and positively associated with inflammation.10,11 Regarding LDL particle size, it has also been observed that patients with RA show increased small and dense LDL particles, with pro-atherogenic properties.12,13 These characteristics indicate that patients with RA exhibit a distinctive atherogenic dyslipidemia pattern, which may partly explain their increased cardiovascular risk.14 However, the relationship between advanced lipoprotein profile and cardiovascular risk-associated microRNAs has not been previously studied.
In this work, we have described the lipoprotein profile of patients with RA, measured with 1H-NMR, and we have compared it with controls and patients with other metabolic disorders such as type 2 diabetes mellitus (T2DM), obesity, hypertension, or dyslipidemia. Moreover, for the first time, we have explored the associations of these molecules with the studied microRNAs (Let7a, 24, 96, 103, 125a, 125b, 132, 146, 191, 223, 425, 451) in order to explore how these microRNAs might regulate lipoprotein concentrations in the context of RA. Understanding the relationship between the expression of the studied microRNAs and the advanced lipoprotein profile might open a door for future therapeutic targets to reduce the cardiovascular risk exhibited by patients with RA.
Material and methodsPatients and clinical variablesThe RA cohort in this current research has been previously characterized.7,15 Individuals who randomly visited University Hospital Sant Joan de Reus through external appointments and met the diagnostic criteria for RA outlined by the American College of Rheumatology in 1987 were selected by our rheumatology team. Those who were over 80 years old or under 18 years old, individuals with acute concurrent illnesses, and those whose disease diagnosis had changed were excluded from the study. Recruitment of patients occurred between November 2011 and January 2015. A total of 219 patients aged between 18 and 80 years were enrolled in the investigation, and blood samples were collected on the same day as their medical appointments. Clinical evaluation of the patients has been previously described (Supplementary Data S1). As a measure of disease activity and inflammation, disease activity score (DAS28) was calculated according to the erythrocyte sedimentation rate (ESR).
In addition, we incorporated 82 patients who voluntarily took part in the study and were receiving care at the Vascular Medicine and Metabolism Unit of our hospital for issues related to lipid metabolism disruptions and related conditions, including T2DM, obesity, hypertension, or dyslipidemia. Diagnosis of T2DM, obesity, hypertension, or dyslipidemia was based on established clinical criteria. Furthermore, we enlisted 64 control subjects who were drawn from the hospital's personnel pool as individuals free from RA, T2DM, obesity, dyslipidemia, or any other chronic and debilitating condition.
The study was approved by The Clinical Research Ethics Committee of Hospital Sant Joan de Reus (patients with RA: 11-04-28/4proj5, controls and patients with metabolic disorders: CEIm: 222/2020) and all the participants gave written informed consent. We executed the investigation in accordance with our institution's guidelines and the Helsinki Declaration.
Laboratory measurementsWe obtained blood samples from each individual included in the study with a previous fasting period of no less than 12h, using EDTA used as the anticoagulant. Plasma was separated from the whole blood through centrifugation at 3000rpm for a duration of 10min, and these plasma samples were subsequently preserved at −80°C for further analysis. Analytical assessments were carried out using enzymatic and standard methods. These assessments encompassed the measurement of rheumatoid factor (RF) (Rheumatoid Factors—II, Roche, Germany), anti-CCP (Bioplex 2200 kit, BioRad, USA), and inflammatory markers (ESR (Automatic system VesMATIC cube 80, Diesse DiagnosticaSenese S.pA., Italy), C-reactive protein (CRP) (latex particle enhanced immunoturbidimetric assay, CRP4 reagent, Roche, Germany)), and fibrinogen (Human recombinant (HemosIL Recombiplastin 2G Instrumentation Laboratory) tromboplastin, Instrumentation Laboratory SpA, USA) using conventional techniques.
Plasma microRNA expressionWe analyzed a panel of microRNAs, including Let7a, 24, 96, 103, 125a, 125b, 132, 146, 191, 223, 425, and 451, using separate plasma samples from 219 individuals with RA. Before RNA extraction, a hemolysis assessment was performed on 200μl plasma aliquots. Hemolysis was detected and excluded through spectrophotometric analysis at λ=414nm, which corresponds to the absorption peak for oxyhemoglobin contamination. RNA extraction was performed using the commercial miRCURY RNA Isolation Kit (Exiqon, Denmark). Candidate miRNAs were measured by qPCR using commercial miRCURY LNA Universal RT microRNA PCR, ExiLENT SYBR Green master mix Kit (Exiqon, Denmark), along with commercial primers specific for each miRNA (hsa-miR LNA™ PCR primer set, UniRT). Detailed procedures for plasma microRNA extraction are provided in Supplemental Data S2. For normalization, microRNA-16-5p was chosen as the reference. The relative expression of each microRNA was calculated using the ΔCt method, where ΔCt=Ct of the candidate microRNA−Ct of microRNA-16-5p. A higher ΔCt value for a given microRNA indicated a lower expression level of that microRNA. The cycle threshold (Ct) for each sample and miRNA was obtained with SDS v2.3 software (Applied Biosystems, USA).
2D-1H-NMR lipoprotein profileThe lipoprotein profiles were analyzed in plasma samples using 1H-NMR, following established protocols.16 200μl of serum were mixed with 50μl of deuterated water and 300μl of a 50mM phosphate buffer solution at pH 7.4. The 1H-NMR spectra were recorded at 305.95K using a Bruker Avance III 600 spectrometer operating at a proton frequency of 600.20MHz (14.1T). Regarding the lipoprotein profile, this method allowed us to determine the particle sizes and numbers of nine subtypes of lipoproteins, including large, medium, and small very low-density lipoprotein (VLDL), LDL, and high-density lipoprotein (HDL), as well as the mean diameter of each lipoprotein class (VLDLz, LDLz, HDLz), as previously reported.16 The particle concentrations and diffusion coefficients were obtained from the measured amplitudes and attenuation of their spectroscopically distinct lipid methyl group NMR signals using the 2D diffusion-ordered NMR spectroscopy pulse. Each lipoprotein subtype was represented by a Lorentzian function, with the area of each function reflecting the lipid concentration. Particle numbers for each lipoprotein subtype were calculated by dividing the lipid volume by the particle volume of a given class. Lipid volumes were determined using standard conversion factors to transform concentration units into volume units. Notably, the variation coefficients for particle number ranged between 2% and 4%, while those for particle size were below 0.3%.
Statistical analysisMean and standard deviation (SD) were reported for normally distributed variables, while median and interquartile range (IQR) were provided for variables with nonnormal distributions. For categorical variables, both percentages and absolute counts were presented. To compare differences between normal, nonnormal, and categorical variables, we employed t-tests, Mann–Whitney U tests, and Chi-squared tests, respectively. Kruskal–Wallis test were used to assess differences in the lipid profile among patients with RA, those with metabolic disorders, and controls. When significant differences were observed Bonferroni post hoc tests were performed. To explore associations between microRNAs and various lipoprotein subfractions, multivariate linear models were applied. These models were adjusted for age, sex, body mass index (BMI), RA treatments and lipid-lowering therapies, and were analyzed both in the overall cohort and separately by sex. For each model, R-squared (R2), change in R-squared (ΔR2), and Akaike Information Criterion (AIC) values were provided. R2 indicates the proportion of variability explained by the model, while ΔR2 shows the additional variance explained with the inclusion of different microRNAs. AIC was used to assess model quality, with lower AIC values indicating better model fit. All statistical analyses were conducted using R Studio, version 4.0.1. A p-value of <0.05 was considered statistically significant.
ResultsGeneral characteristics of the cohortsTable 1 presents the general characteristics of the 219 patients with RA included in the study, both overall and stratified by sex. The cohort presented a mean age of 57.59±12.11, with 65.29% of the patients being female. The mean disease duration and DAS28 score were 7 (2.50–13) and 3.48 (2.67–4.37), respectively. Furthermore, 75.79% of the patients were positive for RF, 73.97% were positive for anti-CCP and 55.25% presented erosions. Regarding RA treatments, 73.97% of the patients received conventional synthetic disease-modifying antirheumatic drugs (csDMARDS), while 58.44% received nonsteroidal anti-inflammatory drugs (NSAIDs), 21% received biological drugs, 52.51% received corticosteroids (mean dose=3.11mg). Moreover, 17.35% of the patients with RA were on lipid-lowering therapies. Male patients exhibited increased waist circumference, diastolic blood pressure (DBP) and a higher prevalence of hypertension. Conversely, women presented higher levels of HDL cholesterol, DAS28 and health assessment questionary (HAQ) punctuation. Additionally, Supplementary Table 1 shows the general characteristics of the RA cohort, control subjects and patients with metabolic disorders. Patients with RA and metabolic disorders were older, had decreased HDL levels, and exhibited increased triglyceride levels compared to controls. They also showed a higher prevalence of T2DM, hypertension, and dyslipidemia, along with elevated inflammatory variable values. Furthermore, RA patients had higher CRP and ESR levels compared to patients with metabolic disorders.
General characteristics of the overall cohort of RA patients and stratified by sex.
| RA (n=219) | Female (n=143) | Male (n=76) | p | |
|---|---|---|---|---|
| Characteristics of the groups | ||||
| Sex – female (%, n) | 65.29%, 143 | |||
| Age (mean in years, SD) | 57.59 (12.11) | 57.4 (12.31) | 57.96 (11.81) | 0.74 |
| Body mass index (median in kg/m2, IQR) | 26.89 (23.54–30.84) | 26.54 (22.88–31.47) | 27.84 (25.47–30.58) | 0.14 |
| Waist circumference (mean in cm, SD) | 92.18 (14.78) | 88.27 (14.85) | 100 (11.58) | <0.001 |
| SBP (median in mmHg, IQR) | 135 (120–150) | 133 (120–146.5) | 137.5 (124.5–155) | 0.06 |
| DBP (median in mmHg, IQR) | 80 (72–89) | 80 (71.50–87) | 82.50 (75–90) | 0.03 |
| LDL cholesterol (median in mg/dL, IQR) | 114 (99–134.5) | 114 (95.5–134.5) | 115 (100.8–134.2) | 0.49 |
| HDL cholesterol (median in mg/dL, IQR) | 66 (53–76) | 69 (61–80) | 54 (43–66) | <0.001 |
| Triglycerides (median in mg/dL, IQR) | 93 (69–128) | 88 (65–126.5) | 95 (74–128.5) | 0.31 |
| Glucose (median in mg/dL, IQR) | 89 (82–98.5) | 89 (81.50–97) | 92 (83–101.25) | 0.14 |
| Current smoker (%, n) | 27%, 59 | 27.27%, 39 | 26.31%, 20 | 1 |
| Hypertension (%, n) | 57.99%, 127 | 51.74%, 74 | 69.74%, 53 | 0.02 |
| Diabetes mellitus (%, n) | 11.41%, 25 | 11.11%, 16 | 11.84%, 9 | 1 |
| Dyslipidemia (%, n) | 40.63%, 89 | 39.86%, 57 | 42.10%, 32 | 0.86 |
| Disease features | ||||
| Disease duration (median in years, IQR) | 7 (2.50–13) | 8 (3–13.50) | 6 (2–11) | 0.09 |
| DAS28 (median, IQR) | 3.48 (2.67–4.37) | 3.59 (2.82–4.69) | 3.02 (2.45–3.76) | <0.001 |
| HAQ (median, IQR) | 0.38 (0–0.88) | 0.50 (0.13–0.88) | 0 (0–0.375) | <0.001 |
| Rheumatoid factor+ (%, n) | 75.79%, 166 | 74.13%, 106 | 78.94%, 60 | 0.53 |
| Anti-CCP+ (%, n) | 73.97%, 162 | 74.12%, 106 | 73.68%, 56 | 1 |
| Erosions (%, n) | 55.25%, 121 | 57.34%, 82 | 51.31%, 39 | 0.48 |
| ESR (median in mm/h, IQR) | 31 (19–50) | 31 (19–53) | 28.50 (19.75–47.25) | 0.25 |
| CRP (median in mg/dL, IQR) | 0.5 (0.2–0.95) | 0.4 (0.2–0.9) | 0.5 (0.2–1) | 0.28 |
| Fibrinogen (mean in mg/dL, SD) | 444.1 (95.11) | 440 (95.56) | 459.5 (96.12) | 0.30 |
| Treatments (%, n) | ||||
| DMARDsc (%, n) | 73.97%, 162 | 62.57%, 102 | 78.94%, 60 | 0.29 |
| Biological agent (%, n) | 21%, 46 | 23.77%, 34 | 15.79%, 12 | 0.23 |
| NSAIDs (%, n) | 58.44%, 128 | 58.74%, 84 | 57.89%, 44 | 1 |
| Corticosteroids (%, n) (mean dose: 3.11mg) | 52.51%, 115 | 53.15%, 76 | 51.31%, 39 | 0.91 |
| Lipid-lowering therapies (%, n) | 17.35%, 38 | 16.78%, 24 | 19.74%, 15 | 0.58 |
n=number of individuals, SBP=systolic blood pressure, DBP=diastolic blood pressure, LDL=low density lipoprotein, HDL=high density lipoprotein, HAQ=health assessment questionnaire index, ACPA=citrullinated anti-cyclic peptide antibodies, ESR=erythrocyte sedimentation rate, CRP=C-reactive protein, DAS28=disease activity score, DMARDs=disease-modifying antirheumatic drugs, NSAIDs=non-steroidal anti-inflammatory drugs, cIMT=carotid intima media thickness, PWV=pulse wave velocity, CV=cardiovascular, p=p value.
We conducted an analysis of H-NMR lipoprotein profiles in patients with RA, patients with metabolic disorders, and healthy subjects (Table 2). Our findings revealed significant differences among these groups when post hoc analyses were performed. Specifically, RA patients showed higher values in all VLDL particle measurements, including total VLDL particle number (p<0.001), large VLDL particles (p=0.004), medium VLDL particles (p<0.001) and small VLDL particles (p=0.002) when compared to healthy subjects. In terms of LDL profile, RA patients demonstrated lower values in total LDL particle number (p<0.001), large LDL particles (p=0.03), and medium LDL particle number (p=0.01) when compared to patients with metabolic disorders. No significant differences were observed in the advanced LDL profile when comparing healthy subjects and subjects with RA. In addition, patients with RA showed decreased levels of total HDL particle number (p=0.03) compared to controls and decreased number of small HDL particle number (p=0.01) compared to patients with metabolic disorders. RA patients also showed decreased HDL diameter (p=0.005) compared to patients with metabolic disorders. Sex-stratified analyses were performed in every cohort and are shown in Supplementary Table 2. Regarding RA patients, we found that men exhibited significant increased VLDL particles (p=0.02), including large (p=0.01), medium (p=0.04), and small VLDLs (p=0.03). Male RA patients also had a higher number of small LDL particles (p=0.02) and a reduced LDL particle diameter (p=0.002). In contrast, women with RA showed an increased total number of HDL particles (p<0.001), including large (p=0.04), medium (p<0.001), and small HDLs (p=0.003). In patients with metabolic disorders, women exhibited a significantly higher number of medium HDL particles (p=0.02), with no significant differences observed in other lipid subfractions. Finally, no differences were found between men and women in any lipid subfraction in the control group.
Comparison of the lipid parameters between patients with RA, controls and patients with metabolic disorders, presented as medians and IQRs.
| RA (n=219) | Controls (n=64) | Met. dis. (n=82) | p-Value | |
|---|---|---|---|---|
| Total VLDL particles (nmol/L) | 39.62 (30.75–54.40)a | 30.85 (24.97–38.70)c | 35.65 (26.39–55.27) | 0.003 |
| Large VLDL (nmol/L) | 1.089 (0.84–1.34)a | 0.86 (0.66–1.12) | 0.99 (0.72–1.35) | 0.004 |
| Medium VLDL (nmol/L) | 4.62 (3.59–6.04)a | 3.78 (2.72–4.61)c | 4.37 (3.43–5.78) | <0.001 |
| Small VLDL (nmol/L) | 33.89 (25.96–46.46)a | 26.34 (21.49–33.38)c | 30.44 (22.02–48.04) | 0.002 |
| Total LDL particles (nmol/L) | 1195.4 (1087.4–1328.3)b | 1215.3 (1090.6–1354.9)c | 1292 (1180.5–1434.7) | <0.001 |
| Large LDL (nmol/L) | 203.2 (104.8–218.2) | 212 (196–229.1)c | 208.8 (192.1–237.4) | 0.03 |
| Medium LDL (nmol/L) | 354.3 (300.4–410.1)b | 377.2 (318.2–423.9) | 388.2 (332.4–457.4) | 0.01 |
| Small LDL (nmol/L) | 648.7 (584.2–707.7)b | 631.4 (578.4–671.9)c | 679.5 (619.8–774.6) | <0.001 |
| Total HDL particles (μmol/L) | 30.41 (27.14–33.82)a | 32.11 (29.43–34.48) | 31.09 (28.40–34.01) | 0.03 |
| Large HDL (μmol/L) | 0.30 (0.28–0.33) | 0.31 (0.28–0.33) | 0.30 (0.28–0.33) | 0.8 |
| Medium HDL (μmol/L) | 11.37 (10.34–13.04) | 11.99 (10.57–13.59) | 10.99 (10.04–13.09) | 0.3 |
| Small HDL (μmol/L) | 18.47 (15.79–21.08)b | 19.62 (17.79–21.76) | 19.26 (17.51–22.89) | 0.01 |
| VLDLz (nm) | 42.24 (42.11–42.37) | 42.28 (42.17–42.40) | 42.28 (42.14–42.40) | 0.27 |
| LDLz (nm) | 21.16 (21–21.30)a | 21.24 (21.12–21.39)c | 21.19 (21.06–21.35) | 0.03 |
| HDLz (nm) | 8.30 (8.25–8.36)b | 8.30 (8.26–8.33) | 8.28 (8.24–8.32) | 0.005 |
Table 3 shows the different β coefficients obtained from the multivariate linear models of the microRNAs statistically associated with the different particles of the lipoprotein profile of the patients with RA. These models were adjusted for age, sex, BMI, RA treatments and lipid-lowering medications. First, in the overall cohort, we showed that microRNA-125a (β=−0.19, p=0.005) was significantly associated with total VLDL particle number. Moreover, microRNA-125a was also associated with large (β=−0.19, p=0.006), medium (β=−0.21, p=0.004) and small VLDL particles (β=−0.19, p=0.006). In addition, microRNA-451 was associated with VLDL diameter (β=−0.14, p=0.04). Regarding LDL and HDL profile, we also found a statistical association between microRNA-125a and small LDL particles (β=−0.16, p=0.03) and microRNA-132 and large HDL particles (β=0.16, p=0.03).
Study of the associations between the different microRNAs and the lipid parameters analyzed by NMR in patients with RA. Multivariate lineal models, adjusted for age, sex, BMI, RA treatments and lipid-lowering medication were computed.
| Overall cohort | |||||
|---|---|---|---|---|---|
| β | p | R2 | ΔR2 | AIC | |
| Total VLDL particles | |||||
| Initial model | 11.64 | 613.39 | |||
| MicroRNA-125a | −0.19 | 0.005 | 13.99 | 2.35 | 541.95 |
| Large VLDL | |||||
| Initial model | 12.96 | 621.79 | |||
| MicroRNA -125a | −0.19 | 0.006 | 15.31 | 2.35 | 543.16 |
| Medium VLDL | |||||
| Initial model | 10.22 | 626.75 | |||
| MicroRNA -125a | −0.21 | 0.004 | 12.51 | 2.29 | 553.92 |
| Small VLDL | |||||
| Initial model | 11.49 | 613.75 | |||
| MicroRNA -125a | −0.19 | 0.006 | 13.75 | 541.68 | |
| VLDLz | |||||
| Initial model | 5.81 | 627.37 | |||
| MicroRNA -451 | −0.14 | 0.04 | 7.24 | 624.01 | |
| Small LDL | |||||
| Initial model | 6.08 | 626.89 | |||
| MicroRNA -125a | −0.16 | 0.03 | 8.88 | 2.80 | 562.48 |
| Large HDL | |||||
| Initial model | 5.40 | 626.33 | |||
| MicroRNA -132 | 0.16 | 0.03 | 8.18 | 2.78 | 523.08 |
| Men patients | |||||
|---|---|---|---|---|---|
| β | p | R2 | ΔR2 | AIC | |
| Large LDL | |||||
| Initial model | 3.66 | 229.84 | |||
| MicroRNA -191 | 0.25 | 0.04 | 13.01 | 9.35 | 190.69 |
| MicroRNA -96 | 0.28 | 0.04 | 14.66 | 11 | 180.87 |
| Medium LDL | |||||
| Initial model | 2.32 | 230.88 | |||
| MicroRNA -96 | 0.30 | 0.04 | 11.30 | 8.98 | 186.09 |
| Total HDL particles | |||||
| Initial model | 10.35 | 224.37 | |||
| MicroRNA -Let7a | 0.26 | 0.04 | 16.12 | 5.77 | 221.31 |
| Large HDL | |||||
| Initial model | 10.11 | 222.57 | |||
| MicroRNA -125b | 0.24 | 0.04 | 22.61 | 12.50 | 179.29 |
| Female patients | |||||
|---|---|---|---|---|---|
| β | p | R2 | ΔR2 | AIC | |
| Total VLDL particles | |||||
| Initial model | 14.75 | 399.99 | |||
| MicroRNA -125a | −0.25 | 0.003 | 19.56 | 35.62 | 364.37 |
| Large VLDL | |||||
| Initial model | 15.54 | 412.93 | |||
| MicroRNA -125a | −0.21 | 0.01 | 18.64 | 3.10 | 375.97 |
| Medium VLDL | |||||
| Initial model | 12.45 | 413.99 | |||
| MicroRNA -125a | −0.27 | 0.001 | 18.33 | 5.88 | 368.76 |
| Small VLDL | |||||
| Initial model | 14.63 | 400.20 | |||
| MicroRNA -125a | −0.24 | 0.004 | 19.80 | 5.17 | 364.84 |
| Total LDL particles | |||||
| Initial model | 3.05 | 418.38 | |||
| MicroRNA -125a | −0.20 | 0.03 | 7.46 | 4.41 | 382.34 |
| Small LDL | |||||
| Initial model | 7.22 | 412.10 | |||
| MicroRNA -125a | −0.23 | 0.008 | 13.30 | 6.08 | 371.65 |
| Medium HDL | |||||
| Initial model | 8.58 | 409.99 | |||
| MicroRNA -125a | 0.18 | 0.04 | 11.71 | 3.13 | 376.35 |
β=beta coefficient; p=p-value, AIC=Akaike information criteria.
When sex-stratified analyses were performed, we observed in male patients that microRNAs 191 (β=0.25, p=0.04) and 96 (β=0.28, p=0.04) were associated with large LDL particles. We also observed that microRNA-96 was associated with medium LDLs (β=0.30, p=0.04). Finally, we also found that microRNA-Let7a was associated with total number of HDL particles (β=0.26, p=0.04) and microRNAs-125b with large HDLs (β=0.24, p=0.04). Regarding female patients, we observed that microRNA-125a was associated with the total VLDL particle number (β=−0.25, p=0.003) as well as with large (β=−0.21, p=0.01), medium (β=−0.27, p=0.001) and small VLDL particles (β=−0.24, p=0.004). MicroRNA-125a was also associated with total LDL particle number (β=−0.20, p=0.03) and small LDL particles (β=−0.23, p=0.008). Finally, we also found an association between microRNA-125a and medium HDL particles (β=0.19, p=0.03).
Incorporating the different microRNAs into the initial models improved the overall model qualities, leading to an increase in the explained variability (R2) and a decrease in the AIC in every model. Table 3 presents a comprehensive summary of the adjusted models, encompassing both the overall cohort and the sex-stratified subgroups.
DiscussionIn the present study, we conducted a comparison of the advanced lipoprotein profile, measured with H-NMR, among a cohort of patients with RA, patients with metabolic disorders (T2DM, obesity, hypertension, or dyslipidemia) and healthy subjects. Furthermore, we explored the associations between microRNAs previously linked to cardiovascular disease and the advanced lipoprotein profile in patients with RA.
Regarding the comparison of the advanced lipid profile between patients with RA, those with metabolic disorders, and healthy controls, we observed that RA patients had significantly higher levels of VLDL particles, both in total number and across most subfractions, compared to controls. Although their VLDL levels were also higher than those in patients with metabolic disorders, this difference did not reach statistical significance. It is well-established that RA patients have elevated levels of VLDL, triglyceride-rich lipoproteins that are pro-atherogenic and serve as an important, independent predictor of cardiovascular disease.17 Therefore, it is plausible to think that the increase of TG, probably due to the chronic inflammatory state of RA, contributes to the increased risk of cardiovascular disease that patients with RA show.18 On the other hand, patients with RA exhibited lower levels of LDL particles compared to individuals with metabolic disorders. Moreover, we observed a noticeable trend indicating that patients with RA exhibited a higher concentration of small LDL particles in comparison to healthy subjects, though this difference did not reach statistical significance. It is known that patients with RA show more oxidized and small LDL particles, likely due to the reduction of HDL cholesterol, which are contributors of the accelerated atherosclerotic process that patients with RA show.19 However, despite this observed trend, the differences in our study were not statistically significant, possibly due to limited statistical power. We also observed that patients with RA showed decreased total and small HDL particle number, and increased HDL diameter compared to patients with metabolic disorders. Some studies suggest that patients with RA show increased small HDL particles compared to subjects without this condition, which are generally considered less effective at carrying out their cholesterol-transporting functions. Inflammation is probably the main cause of the HDL dysfunctionality exhibited by patients with RA, which might potentially contribute to the increased cardiovascular risk.20 Interestingly, in sex-stratified analyses of RA, men exhibited a more atherogenic profile than women, characterized by higher levels of VLDL particles and small LDL particles, along with lower HDL levels. This profile may partially explain their elevated cardiovascular disease risk compared to women.
Moreover, we investigated the associations between microRNAs previously associated with cardiovascular disease and inflammation and the advanced lipoprotein profile, yielding several noteworthy findings. Specifically, our observations indicate that microRNA 451 is associated with VLDL size. As previously mentioned, we have linked this microRNA to subclinical atherosclerosis in RA,8 while VLDL particles are known to exhibit inflammatory activity and, consequently, are associated with cardiovascular disease.21 However, this study represents the first exploration of these specific associations in this context. We also found associations between microRNA-125a and the total particle number as well as with small, medium, and large VLDLs. MicroRNA-125a has been previously considered as a potential biomarker of RA and insulin resistance, often associated with hypertriglyceridemia.22,23 A previous study has reported the presence of microRNA-125a in all VLDL fractions of healthy subjects24; however, this is the first instance of this association in RA patients. Dysregulation of microRNA-125a may be influential in altering VLDL metabolism, which is linked to both inflammation and cardiovascular disease.25 Hence, microRNA-125a emerges as a promising therapeutic target for RA. Regarding LDL particles, microRNA-125a also demonstrated an association with small LDLs in both the overall cohort and among women. Furthermore, it exhibited an association with the total LDL particle number in women. While microRNA-125a has been previously detected in LDL particles, its presence in RA patients is novel. Moreover, evidence suggests that microRNA-125a may regulate small and oxidized LDLs, aligning with previous results.26 Finally, our study revealed fewer associations with HDL subfractions. In this context, we also observed associations between mcroRNA-125a with several HDL subfractions. Lipoproteins can serve as carriers for microRNAs,27 which could add another layer of complexity as microRNAs encapsulated within lipoproteins might be transported to distant sites, potentially influencing gene expression and cellular function in a variety of tissues.
Our study does have several limitations. Firstly, we cannot establish causality in any of the associations we found due to the cross-sectional design of our study. Secondly, our cohort of patients with metabolic disorders, while exhibiting a similar inflammatory pattern, is relatively homogeneous. Lastly, the associations we discovered do not fully elucidate the metabolic interactions between the studied microRNAs and the lipoprotein particles identified. Consequently, conducting experimental and in vitro studies is imperative to gain a deeper understanding of the relationships between these parameters.
In conclusion, our study provides a comprehensive comparison of advanced lipoprotein profile among three distinct cohorts: patients with RA, healthy subjects, and individuals with metabolic disorders. Notably, we have identified significant differences among these groups. Additionally, we have uncovered intriguing associations between this advanced profile and several microRNAs. Of particular significance is our observation regarding the pivotal role of microRNA-125a in VLDL metabolism, positioning it as a promising therapeutic target for mitigating inflammation and cardiovascular risk. Furthermore, these results contribute to a more profound understanding of lipid metabolism alterations in RA patients and highlight potential therapeutic microRNAs, such as targeting microRNA-125a.
Informed consent statementInformed consent was obtained from all subjects involved in the study.
FundingThis study has been funded by Instituto de Salud Carlos III (ISCIII) through the project “FIS PI20/00443” and co-funded by the European Union.
Conflict of interestThere is no conflict of interest to disclose.
We would like to thank all the patients for their essential collaboration.
The followings are the supplementary data to this article:
