Genetic association of lipids and lipid lowering drug target genes with sarcoidosis

Abstract

To determine the potential causal association between serum lipid levels and sarcoidosis, and to investigate the potential impact of lipid lowering agents on sarcoidosis. Two-sample Mendelian randomization (TSMR) was used to investigate the association between lipid levels (including LDL-c, HDL-c, TG, and TC) and sarcoidosis risk. In addition, we used Mendelian drug target randomization (DMR) to analyze the relationship between drug targets for lowering LDL-c levels (HMGCR, PCSK9, and NPC1L1) and drug targets for lowering TG levels (LPL and APOC3) and the risk of sarcoidosis. According to the TSMR analysis, a positive correlation was observed between the serum LDL-c concentration and sarcoidosis incidence (n = 153 SNPs, OR = 1.232, 95% CI = 1.018–1.491; p = 0.031). Similarly, serum TG concentration was found to be positively associated with sarcoidosis (n = 52 SNPs, OR = 1.287, 95% CI = 1.024–1.617; p = 0.03). The DMR results demonstrated a positive correlation between PCSK9-mediated serum LDL-c levels and sarcoidosis (n = 35 SNPs, OR = 1.681, 95% CI = 1.220–2.315; p = 0.001). Similarly, serum TG levels mediated by LPL were positively associated with sarcoidosis (n = 28 SNPs, OR = 1.569, 95% CI = 1.223–2.012; p = 0.0003). This study suggested that elevated serum TG and LDL-c levels may increase the risk of sarcoidosis. PCSK9-mediated reduction of LDL-C levels (simulating the effects of PCSK9 inhibitors) and LPL-mediated reduction of TG levels (simulating the effects of LPL-related lipid lowering drugs) can decrease the risk of developing sarcoidosis.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-75322-3.

Introduction

Sarcoidosis is a multisystem disease of unknown origin characterized by noncaseous necrotizing epithelioid granuloma¹. It often presents with bilateral hilar lymph node enlargement, lung infiltration, and ocular and skin lesions². The highest incidence is 5–40 cases / 100,000 in Northern Europe³, followed by the Americas, and the incidence is low in Asia, Australia and Africa⁴. However, due to the lack of effective treatment and the difficulty of clinical diagnosis, the prognosis is poor when the lung fibrosis is extensive, and the overall case fatality rate is 1 − 5%.

An observational study revealed that the serum samples from patients with sarcoidosis exhibited similar characteristics, including significant decreases in total cholesterol and HDL-c. This finding suggested a functional role for mononuclear phagocytes in regulating serum lipoprotein cholesterol levels⁵. Changes in lipid metabolism, including reduced HDL-c levels and decreased apolipoprotein A1 concentrations, induce endothelial cell damage in serous membranes, bronchi, and pulmonary capillaries in patients with sarcoidosis. Furthermore, dyslipidemia has been associated with increased oxidative stress and decreased overall antioxidant protection, with the precise mechanisms and directionality of these associations still being actively investigated⁶. This suggests that lipid abnormalities may contribute to an increased risk of sarcoidosis, though the exact nature of this relationship remains to be fully elucidated. Lipids are major participants and facilitators of physiological processes in the lung⁷, and a large number of studies have shown that changes in lipids and their metabolism can lead to chronic lung diseases and progression such as pulmonary fibrosis. A meta-analysis showed that statins reduced mortality and CRP levels in patients with COPD, and there is evidence suggesting that their anti-inflammatory and immunomodulatory effects may contribute to maintaining a more stable or normal pH balance in the respiratory system of these patients⁸. In 2018, a clinical study in which lipid parameters were measured in 77 patients with pulmonary nodules and 139 normal subjects showed that triglyceride and LDL-c levels were significantly elevated in patients with pulmonary nodules, and that blood lipid parameters were associated with an increased risk of sarcoidosis⁹. A decrease in high-density lipoprotein cholesterol (HDL-c) was observed in untreated sarcoidosis patients. Multiple regression analysis of a clinical trial comparing 40 patients not treated with prednisone to 22 patients treated with prednisone revealed that, among the treated group, HDL-cholesterol (HDL-c) and apolipoprotein A1 (ApoA1) levels were found to be independently and significantly associated with steroid therapy, suggesting potential differences in lipid profiles as a result of glucocorticoid (GC) exposure¹⁰. A cross-sectional study in China showed that high LDL-c was more likely to be associated with sarcoidosis¹¹. PCSK9 and HMGCR are target proteins of drugs used to lower LDL-c levels by regulating lipid metabolism¹². LPL and APOC 3 are target proteins for drugs that lower TG levels¹³. However, the impact of the use of these drugs on sarcoidosis remains to be further studied.

Clinical studies have limitations, such as confounders, survival bias, and reverse causality, which can cause errors or contradictions in observational results¹⁴. However, these problems can be avoided by using Mendelian randomization (MR). The TSMR was used to study the correlation between exposure factors and outcome factors. Its advantage is that it can eliminate the interference of confounding factors. Mendelian randomization of drug targets (DMR) can provide important information for drugs, such as lipid lowering drugs, including predicting efficacy and revealing target-mediated adverse reactions¹⁵. In this study, we used the TSMR to analyze the causal relationship between sarcoidosis and the four serum lipid levels. In the DMR analysis, we evaluated the impact of various lipid lowering drugs on sarcoidosis incidence.

Materials and methods

Study design

The flow chart of the study is shown in Fig. 1. We conducted two MR analyses. On the one hand, TSMR analysis was performed to evaluate the effect of different lipids (e.g., TG, TC, LDL-c, and HDL-c) on the risk of sarcoidosis. On the other hand, we performed drug-targeted MR (DMR) analyses to assess the causal effect of variants in TG drug target genes (LPL and APOC3) on sarcoidosis risk. The effects of HMGCR and PCSK9, which target lipid lowering drugs to reduce LDL-c levels, on sarcoidosis was also analyzed.

Fig. 1.

The design scheme of this study.

GWAS data source

The statistical data for sarcoidosis were obtained from the 217,758 European descent GWAS meta-analysis (https://gwas.mrcieu.ac.uk/). From this meta-analysis we obtained 2046 patients with sarcoidosis and 215,712 controls. All the serum lipid data were obtained from the freely available IEU database (https://gwas.mrcieu.ac.uk/). In this study, after several screenings, we ultimately obtained a highly representative dataset. All the data involved in this study can be found in the supplementary materials (supplementary Table 1). In accordance with the original research ethics approval of the GWAS, this study exempted patients from informed consent and other ethical requirements.

Instrumental variable selection

In the TSMR analysis, first, we obtained single nucleotide polymorphisms (SNPs) for sarcoidosis from the FinnGen database. We subsequently screened the SNPs as follows: (1) genome-wide significance threshold P < 5 × 10⁻⁸; (2) linkage disequilibrium R² value < 0.001; (3) distance between adjacent SNPs less than 10 Mb; (4) palindromic instrumental variables (IVs) with exclusion of intermediate allelic frequencies; (5) threshold F-statistic greater than 10¹⁶.

Finally, we identified SNPs that are strongly associated with sarcoidosis. We used the selected SNPs as IVs for TSMR analysis. In drug target MR analysis, the association between eQTLs and GWAS-based instrument variables was confirmed by drug target MR analysis using eQTLs within a range of 100 kb on either side of the encoded gene. The screening criteria for IVs were the same as those above.

MR analyses

Inverse variance weighting (IVW) was used as the main analysis method for the TSMR¹⁷. Horizontal pleiotropy was determined by the MR Egger method, and the heterogeneity and directed pleiotropy were estimated by Cochran’s Q statistics and the MR‒Egger test¹⁷. The MR-PRESSO method removes all “outliers” from the original SNP dataset to improve the reliability of the results¹⁸. We used these methods to analyze the causal relationship between serum lipid levels and sarcoidosis incidence. In the DMR analysis, IVW was used as the main analysis method to analyze the influence of various lipid lowering drugs on sarcoidosis incidence. Cochran’s Q test was used to test for heterogeneity (p < 0.05). MR‒Egger regression analysis and MR-PRESSO analysis were utilized to evaluate horizontal pleiotropy (p > 0.05). We used these methods to analyze the causal relationship between lipid lowering drugs and sarcoidosis.

Results

Instrumental variable

In the TSMR analysis, we used the SNPs that were screened in Sect. 2.3 and strongly associated with sarcoidosis. We also used the harmonize function to modify the lipid profile and sarcoidosis results (Supplemental Tables 2–5).

Causal effects of lipids on sarcoidosis

The number of SNPs with a causal relationship between lipid markers and sarcoidosis was as follows: TG, 52; TC, 81; LDL-c, 153; and HDL-c, 314 (Fig. 2; Supplementary Tables 2–5). In the TSMR analysis, we primarily used the IVW analysis method, and employed the MR‒Egger and weighted median methods as auxiliary methods. This study provides a visual representation of the dispersion of individual study estimates around the population MR Estimates through a visualized funnel plot, allowing us to identify any asymmetries or outliers that might suggest level pleiotropy. Before performing the IVW analysis, we enhance the stability of the results by utilizing the MR-PRESSO analysis, which is designed to eliminate outliers. According to the TSMR analysis, the serum TG and LDL-c levels were significantly associated with sarcoidosis. Increased serum TG levels were associated with an increased risk of sarcoidosis (OR = 1.287, 95% CI = 1.024–1.617; p = 0.03) (Supplementary Figs. 4–6). Increased serum LDL-c levels were associated with an increased risk of sarcoidosis (OR = 1.232, 95% CI = 1.018–1.491; p = 0.031) (Supplementary Figs. 1–3). However, the results of this study did not reveal any correlation between the serum TC or HDL-c level and sarcoidosis (Fig. 2). As a positive control, we selected CHD patients and verified the presence of the abovementioned lipid-related SNPs under these conditions. The P values for the serum LDL-C, HDL-c, TG, and TC were 6.35E-29, 1.38E-14, 1.61E-08, and 2.76E-32, respectively. (Supplementary Figs. 40–51; Supplementary Table 9)

Fig. 2.

Results of the TSMR analysis of the associations between LDL-c, HDL-c, TG, and TC and sarcoidosis risk were obtained using the IVW, MR Egger, and weighted median methods. The lipid levels mentioned above were validated using CHD as a positive control. Number of SNPs.

Causal relationship between lipid lowering drug targets and sarcoidosis

DMR analysis revealed that PCSK9-mediated serum changes in LDL-c levels were positively associated with sarcoidosis risk (OR = 1.681, 95% CI = 1.220–2.315; p = 0.001) (Fig. 3, Supplementary Figs. 7–9). However, our results did not reveal any association between HMGCR and NPC1L1-mediated changes in the serum LDL-c concentration and sarcoidosis risk. We utilized CHD as a positive control and validated the aforementioned drug-related SNPs in patients with CHD. The P values for PCSK9, HMGCR, and NPC1L1 were 3.57E-24, 6.79E-09, and 0.0002, respectively (Supplementary Figs. 16–24; Supplementary Table 8). We subsequently conducted an analysis of LPL-mediated changes in serum TG levels using DMR and observed a positive correlation with sarcoidosis risk (OR = 1.569, 95% CI = 1.223–2.012; p = 0.0003) (Supplementary Figs. 10–12). However, our results did not reveal any significant association between APOC3 or ANGPTL3-mediated changes in the serum TG concentration and sarcoidosis risk. Similarly, CHD was also utilized as a positive control, and the aforementioned drug-related SNPs were validated and evaluated in patients with CHD. The P values for LPL, APOC3, and ANGPTL3 were 3.51E-26, 0.0001, and 0.038, respectively (Supplementary Figs. 25–33; Supplementary Table 8). Subsequently, we used DMR analysis to investigate the association between CETP-mediated changes in serum HDL-c levels and sarcoidosis. The results indicated a positive correlation (OR = 1.289, 95% CI = 1. 157–1.435; p = 4.02E-06) (Supplementary Figs. 13–15; Supplementary Table 7). However, our analysis did not find any association between APOB-mediated serum TC levels and sarcoidosis risk. Finally, CHD was utilized as a positive control, and the drug-related SNPs associated with TG and HDL-c were validated and analyzed in patients with CHD. The findings revealed that the P-values for CETP and APOB were 3.03E-12 and 1.32E-08, respectively (Supplementary Figs. 34–39; Supplementary Table 8).

Fig. 3.

DMR analysis of the eight drug targets. DMR validation analysis of the above drug targets and CHD incidence. Number of SNPs.

Discussion

Sarcoidosis, a highly prevalent granulomatous disease of unknown cause¹⁹, is usually common in the lungs but can affect almost all organs²⁰. Lipid molecules are involved in triggering and resolving inflammation in lung diseases, and abnormal lipid metabolic pathways are involved in the pathogenesis of many lung diseases²¹. Therefore, in clinical practice, we may be able to predict the occurrence, development and prognosis of lung diseases through observing the levels of various lipids. Recent studies have shown that abnormal regulation of lipid metabolism plays an important role in idiopathic pulmonary fibrosis²². Studies have found that a lack of PPAR-γ levels can lead to sarcoidosis²³. Endogenous PPAR-gamma ligands are fatty acid derivatives, and their deficiency leads to lung inflammation. Since PPAR-γ has anti-inflammatory properties, the relationship between PPAR-γ and sarcoidosis development and inflammatory status can be essentially identified as a suitable therapeutic target. Synthesis of PPAR-gamma receptor agonists or PPAR-gamma ligands may be an effective way to treat sarcoidosis in the future²⁴. Abnormal fatty acids, cholesterol and other lipids affect the regenerative function of alveolar epithelial cells and promote the transformation of fibroblasts into myofibroblasts. Drugs that lower blood lipid levels, such as statins, have been shown to delay the progression of idiopathic pulmonary fibrosis and reduce the impairment of lung function²⁵. However, the effect of these drugs on the risk or progression of sarcoidosis, especially in the context of genetic predisposition, remains unclear. Therefore, we performed a single-factor MR analysis to explore the causal effect of changes in the serum levels of various lipid factors on sarcoidosis risk. In addition, we performed drug target MR analyses to determine the association of lipid lowering drug -targeting genes with sarcoidosis.

Our TSMR analysis suggested that high serum TG levels and high serum LDL-c levels increase the risk of sarcoidosis. These findings suggest that lowering serum TG levels and serum LDL-c levels may reduce the risk of sarcoidosis. However, the results of this study did not reveal any correlation between serum HDL-c or TC levels and sarcoidosis. In a study by Jelena Veki et al., with 77 patients, adverse lipoprotein subprofiles in the serum of sarcoidosis patients continued to change over the course of the disease, with an increase in the serum LDL II and III subclasses (p < 0.001)²⁶. In an observational study of 133 patients with newly diagnosed sarcoidosis and 51 patients with untreated risk-like arthritis, Arzu Cennet Izek et al. reported that metabolic syndrome (MetS) was more common in patients with sarcoidosis (OR = 5.3; 95% CI = 2.4–11.5; p < 0.001), and the serum TG levels in female sarcoidosis patients were significantly greater than those in the control group²⁷. Our data are consistent with these two results. Previous data have shown that elevated serum LDL-c and TG levels are significantly associated with sarcoidosis. In vivo and in vitro experiments by Xian Guang Shi et al. revealed that serum LDL-c levels were elevated in a mouse model of idiopathic pulmonary fibrosis (IPF), and that LDLR levels were restored by combination treatment with atorvastatin and alirocumab. It inhibited bleomycin-induced LDL rise, apoptosis, and fibroblast-like cell accumulation, and alleviated IPF in mice²⁸. An elevated triglyceride-glucose index (TYG) is a risk marker for chronic obstructive pulmonary disease (COPD)²⁹, and our findings suggest that serum TG and LDL-c levels are positively associated with sarcoidosis risk.

Previous studies, have not established that lipid lowering drugs are linked to sarcoidosis, and this area remains unexplored. Studies have shown the potential role of statins in idiopathic pulmonary fibrosis, but the association between statin use and sarcoidosis has not been established. A clinical trial showed the advantage of statins in treating patients with chronic obstructive pulmonary disease³⁰. PCSK9 inhibitors reduce serum LDL-c levels, and a randomized controlled trial showed that PCSK9 inhibitors improve survival in patients with pneumonia with high inflammatory intensity³¹. Our study investigated the association between sarcoidosis and serum LDL-C levels modulated by PCSK9-mediated mechanisms (simulating the effects of PCSK9 inhibitors) and by HMGCR-mediated mechanisms (simulating the effects of HMGCR inhibitors, commonly known as statins). PCSK9 inhibitors or HMGCR lipid lowering drugs can reduce the serum LDL-c concentration. In this DMR analysis, we analyzed the causal relationship between sarcoidosis and reduced serum LDL-c levels of HMGCR and PCSK9 gene-associated SNPs. IVW analysis revealed that PCSK9-mediated changes in the serum LDL-c concentration were positively associated with sarcoidosis risk. However, our findings did not support an association between the serum LDL-c levels modulated by HMGCR and sarcoidosis. Given the complexity of genetic associations and the potential for horizontal pleiotropy, we acknowledge the need to consider alternative explanations for the observed differences in the associations between PCSK9- and HMGCR-mediated changes in serum LDL-c levels and sarcoidosis risk. Horizontal pleiotropy, where a genetic variant influences multiple traits through distinct biological pathways, could underlie the disparate findings between these two SNPs. Therefore, it is possible that the effect of PCSK9 on sarcoidosis risk is mediated through a pathway distinct from that of HMGCR, or that there are other, yet unidentified, factors modulating the relationship between LDL-c levels and sarcoidosis. In addition, this study analyzed the association between sarcoidosis and the SNPS associated with the LPL and APOC3 genes, which reduce the serum TG concentration. IVW-MR was used to analyze LPL-mediated changes in serum TG levels and the risk of sarcoidosis, and the results showed a positive association. However, there was no significant association between APOC3-mediated changes in the serum TG concentration and sarcoidosis risk. Our analysis suggests a potential causal relationship between alterations in serum LDL-c levels mediated by PCSK9-targeted therapies and alterations in serum TG concentrations mediated by LPL-related mechanisms, and the risk of developing sarcoidosis.

Our study has the following advantages. First, this study is the first to establish causal relationships between serum TG levels, and serum LDL-c levels, and sarcoidosis using GWAS data. The TG and LDL-c levels may potentially serve as test parameters for predicting sarcoidosis risk and could be tested in future studies for their discriminative ability in this regard. Second, through Mendelian randomization analysis of genetic data, we found evidence suggesting that reducing LDL-C levels mediated by PCSK9-targeted mechanisms and reducing TG levels mediated by LPL-related mechanisms may decrease the risk of developing sarcoidosis. Third, we use the following methods to verify the accuracy of the MR data: (1) the F statistic was calculated to remove weakly correlated IVs; (2) CHD was used as a positive control group to verify the reliability of the SNPs; and (3) the IVW method was used to verify the accuracy of the results obtained by MR analysis of the drug targets. Fourth, we used genetic tools instead of drug exposure for drug target MR analysis, minimizing confounding bias and reverse causation.

However, there are several limitations to our study. First, despite our use of various analytical methods to exclude outliers and verify horizontal pleiotropy, we cannot fully eliminate the possibility of confounding bias in this study. On the other hand, the sarcoidosis database we used this time was limited to individuals in continental Europe. Therefore, we cannot easily use these data to explain the risk of sarcoidosis in other countries or populations on other continents.

Conclusion

Our analysis indicated that both elevated serum TG levels and increased serum LDL-C levels augment the risk of sarcoidosis. Moreover, PCSK9-mediated reduction of LDL-C levels (simulating the effects of PCSK9 inhibitors) and LPL-mediated reduction of TG levels (simulating the effects of LPL-related lipid lowering drugs) can decrease the risk of developing sarcoidosis. Identifying elevated blood lipid levels during the treatment of sarcoidosis patients may be useful. Our research team plans to conduct randomized controlled trials on lipid lowering medications in the future.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

MR

Mendelian randomization

TSMR

Two-sample Mendelian randomization

DMR

drug target Mendelian randomization

IVW

Instrumental variable weighted regression

SNPs

Single nucleotide polymorphisms

IVs

Instrumental variables

Author contributions

WT wrote the manuscript and performed the quality assessment. WT designed the project and performed the statistical analysis. ZL and XT contributed to the revision of the manuscript and reviewed the results. Conceptualization: WT and XT. Methodology: ZL. Software: ZL. Validation: WT. Formal analysis: WT. Investigation: WT. Resources: WT. Data curation: WT. Writing-original draft preparation: WT. Writing-review and editing: WT. Visualization: WT. Supervision: YL and GT. Funding acquisition: YL and GT. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by the National Natural Science Foundation of China (82004306). The Hunan Province Traditional Chinese Medicine “Shennong Talent” project; the General Research Project of Health Commission of Hunan Province (D202303027798); the Hunan Province Key Research and Development Project (2023SK2057); and the Hunan Nature Foundation (2022JJ40238).

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The authors declare no competing interests.