Exploring the Causal Relationship Between Immune Cells and Idiopathic Pulmonary Fibrosis: A Mendelian Randomization Analysis

ABSTRACT

Background

Idiopathic pulmonary fibrosis (IPF) is a progressive and irreversible interstitial lung disease with a complex pathogenesis involving multiple immune cells. This study investigates the relationship between immune cells and IPF using Mendelian randomization (MR) analysis.

Methods

A two‐sample MR analysis was performed using genome‐wide association studies (GWAS) and immune cell databases by R software. We analyzed data from 1028 European individuals with IPF, focusing on 731 immune traits. The primary method of analysis was inverse variance weighting (IVW), supplemented with sensitivity analyses, including MR‐Egger regression and MR‐PRESSO, to detect and correct for pleiotropy.

Results

The MR analysis identified six immune panels and 23 immune traits significantly associated with IPF, including five traits that increase and 18 traits that decrease IPF risk. Notable traits increasing IPF risk included switched memory B‐cells (OR = 1.27, p = 0.0029) and IgD‐ CD38dim B‐cells (OR = 1.08, p = 0.0449). Traits associated with a reduced IPF risk included central memory CD4+ T‐cells (%CD4+, OR = 0.96, p = 0.0489), CD20 on naive‐mature B‐cells (OR = 0.94, p = 0.0499), and CD33br HLA‐DR+ absolute count (AC) (OR = 0.93, p = 0.0489). There was no significant causal relationship between IPF disease and some immune traits (p > 0.05).

Conclusion

This study suggests a potential causal link between specific immune cell traits and the development of IPF, providing new insights into the disease's immunological mechanisms. Future research should focus on validating these findings in larger, more diverse populations to inform drug development and therapeutic strategies.

Immunophenotypes of immune cells associated with idiopathic pulmonary fibrosis analyzed by the IVW method.

1. Introduction

Idiopathic pulmonary fibrosis (IPF) is a progressive and irreversible interstitial lung disease (ILD) marked by abnormal accumulation of extracellular matrix (ECM) within the lung parenchyma [1]. The pathogenesis of IPF is complex, involving multiple profibrotic cytokines, immune cells, and signaling pathways. Nintedanib and pirfenidone, first‐line drugs for IPF, target fibrotic processes and have shown efficacy in slowing lung function decline, although palliative care remains essential throughout the disease's progression. Lung transplantation, the only life‐saving intervention for advanced‐stage IPF, is associated with significant complications, with approximately 80% of patients experiencing drug toxicity.

The exposure of the lung to external and internal environments makes it susceptible to various factors. Oxidative stress, abnormal cellular signaling, elevated telomerase activity, inflammatory factors, and immune cells contribute to IPF pathogenesis. Oxidative stress, characterized by an imbalance in the redox environment, leads to excessive production of reactive oxygen species and a defective antioxidant system, damaging DNA, proteins, and lipids, thereby promoting pulmonary fibrosis. Epithelial cells, fibroblasts, macrophages, and T lymphocytes produce profibrotic factors through pathways such as MAPK, NF‐kB, and TGF‐β, stimulating fibroblast proliferation and collagen production [2].

Immune cells play a crucial role in lung health by recognizing and removing abnormal or damaged cells. However, overactivation or abnormal differentiation of immune cells can lead to autoimmune diseases like IPF and multiple sclerosis (MS). For instance, T cell overactivation in IPF patients can cause lung tissue damage and fibrosis [3]. Additionally, the number of regulatory T cells (Tregs) with anti‐inflammatory effects is reduced in IPF patients, exacerbating the inflammatory response and fibrosis. Modulating immune cells, using immunosuppressants or modulators to suppress overactivated immune responses or stimulate anti‐inflammatory immune cells, could be an effective strategy for treating IPF [4].

Epidemiological surveys indicate that the survival period of IPF is 2.5–3.5 years [5], with 10%–15% of patients experiencing lung failure within a short period [6]. In Europe, the incidence rate is approximately 25% per year and tends to increase with time. Significant risk factors for IPF include male sex, with approximately 85% of patients aged > 70 years [7], and environmental factors such as smoking, occupational exposure, and air pollution. IPF is also hereditary in 5%–20% of cases, potentially related to gene mutations, protein denaturation, and immune function. Early‐stage IPF involves inflammatory factors and immune cell involvement in the inflammatory response and disease healing [8]. IPF is associated with the innate and adaptive immune systems, including mutations in genes related to innate immune function and the regulation of inflammatory mediators by alveolar macrophages [9].

The two‐sample Mendelian randomization (MR) method screens more neglected immune cells and observes immune phenotypes in greater detail. Advancements in technology have deepened our understanding of the genetic characteristics of immune cells, providing more theories for developing new immune therapies and mechanisms of action. This study investigates the relationship between immune cells and IPF using MR analysis, aiming to elucidate the immunological mechanisms underlying IPF and inform future drug development.

2. Methods and Materials

2.1. IPF GWAS Data Sources

The Genome‐Wide Association Studies (GWAS) data used in this study were obtained from the IEU OpenGWAS project (https://gwas.mrcieu.ac.uk/), specifically from the dataset with GWAS ID “finn‐b‐IPF.” This dataset includes summary statistics from 1028 European men and women, encompassing 16,380,413 single‐nucleotide polymorphisms (SNPs). The quality control measures included replicating results and using independent cohort data to ensure robustness. Instrumental variables (IVs) were selected based on a stringent p‐value threshold (p < 5e‐8) to filter SNPs significantly associated with the exposure factors. Linkage disequilibrium (LD) pruning was performed with a correlation coefficient threshold (r ² < 0.001) and a distance threshold of 1000 kb to remove correlated SNPs.

2.2. Immunity‐Wide GWAS Data Sources

Data on 731 immune traits were downloaded from the GWAS Catalog (http://ftp.ebi.ac.uk/pub/databases/gwas/summary_statistics/), encompassing datasets GCST90001391‐GCST90002121. The traits included Relative Count (RC), Absolute Count (AC), Median Fluorescence Intensities (MFI), and Morphological Parameter (MP) measurements, with data from immune cell types such as cDCs, myeloid cells, TBNK cells, Tregs, and B cells.

2.3. Evaluation of Data Quality and Representativeness

The quality and representativeness of the data were evaluated using several criteria. SNPs were selected based on their genome‐wide significance (p < 5e‐8) to ensure strong instrument relevance. We also performed LD pruning to remove SNPs in high linkage disequilibrium, which reduces bias due to correlated variants. To mitigate potential confounding, we used bidirectional MR analysis and conducted sensitivity analyses, including MR‐Egger regression and MR‐PRESSO global test, to detect and correct for pleiotropy.

2.4. Proxy Parameter Description

The proxy parameter was set to FALSE initially, indicating no proxy SNPs were used for the outcome variable. When set to TRUE, parameters such as rsq, align_alleles, and palindrome were specified, and a minor allele frequency (maf_threshold) of 0.01 was set to ensure the validity of the proxy SNPs used.

2.5. Statistical Analysis

MR relies on three core assumptions: The relevance of the genetic variants as instruments, the independence of these variants from confounders, and the absence of a direct effect of the variants on the outcome. We assessed these assumptions through multiple statistical tests. The F‐statistic was used to assess instrument strength (F < 10 indicates a weak instrument). MR‐Egger regression was utilized to detect directional pleiotropy, with the intercept providing a test for pleiotropy. MR‐PRESSO global test further assessed pleiotropy and corrected for outliers.

For our main analysis, we used the inverse variance weighting (IVW) method, which combines the effects of multiple genetic variants to estimate the causal effect. The IVW method offers robustness against missing data, outliers, and errors, making it suitable for complex traits. Sensitivity analyses, including leave‐one‐out tests, were conducted to evaluate the robustness of our findings and to identify any influential SNPs.

2.6. Selection of Instrumental Variables (IVs)

Using R software (version 4.2.3), we batch‐analyzed immune traits and extracted SNPs with p < 1e‐5 and r ² < 0.001. The stringent threshold (p < 1 × 10⁻⁵) helps minimize false positives and reduce the potential impact of pleiotropy. We harmonized exposure and outcome datasets to ensure consistency in allele frequencies and effect sizes. Scatterplots, forest plots, and funnel plots were generated to visualize the data, and multiple comparison corrections were applied to account for the testing of multiple traits.

2.7. Heterogeneity and Sensitivity Analyses

Heterogeneity was assessed using Cochran's Q statistic and funnel plots, while sensitivity analyses included MR‐Egger, weighted median estimator, and MR‐PRESSO. These analyses ensured the robustness and reliability of our results by accounting for potential pleiotropy and heterogeneity among SNPs.

3. Results

3.1. Impact of Immunophenotype on IPF

The analysis identified six immune panels and twenty‐three immune traits significantly associated with IPF, including the B‐cell panel, myeloid cell panel, monocyte panel, maturation stages of T‐cell panel, TBNK panel, and Treg panel. The odds ratio (OR) in case–control studies reflects the relationship between disease and exposure, with OR > 1 indicating an increased risk and OR < 1 indicating a decreased risk. A p‐value of < 0.05 denotes statistical significance between immune traits and IPF (Figure 1 and Table S1). Table 1 demonstrates the significant immunophenotypes that were found to increase the risk of IPF. The most significant factor identified with increased risk of IPF was the percentage of Sw mem % B‐cell. With an OR of 1.27 and a 95% CI of 1.08 to 1.48, this factor indicates a 27% increase in the odds of the outcome for each unit increase in the percentage of these cells (p = 0.0029). Conversely, Table 2 shows the significant immunophenotypes associated with a reduced risk of IPF. The strongest factor with decreased risk of IPF were CD25 on activated Tregs, CD28 on resting Tregs, and TD CD4+ absolute count (AC), each with an OR of 0.80. These factors indicate a 20% reduction in the odds of the outcome for each unit increase in their respective measurements. The Detailed MR results and scatter plots on immunophenotyping versus IPF are available in Figures [Link], [Link], [Link].

FIGURE 1.

Immunophenotypes of immune cells associated with idiopathic pulmonary fibrosis analyzed by the IVW method.

TABLE 1.

Increased immunophenotyping of IPF disease.

Trait	Odds ratio (OR)	95% confidence interval (CI)	p
Sw mem %B‐cell	1.27	1.08 ~ 1.48	0.0029
IgD‐ CD38dim %B‐cell	1.08	1.00 ~ 1.17	0.0449
CD25 on CD4 Treg	1.07	1.00 to 1.14	0.0359
CD45RA+ CD28‐ CD8br %T‐cell	1.00	1.00 ~ 1.00	0.0329
CD45RA+ CD28‐ CD8br AC	1.00	1.00 ~ 1.00	0.0441

TABLE 2.

Decreased immunophenotyping of IPF disease.

Trait	Odds ratio (OR)	95% confidence interval (CI)	p
Central memory CD4+ T‐cells (%CD4+)	0.96	0.93 ~ 1.00	0.0489
CD20 on naive‐mature B‐cells	0.94	0.89 ~ 1.00	0.0499
CD33br HLA‐DR+ absolute count (AC)	0.93	0.86 ~ 1.00	0.0489
CD19 on IgD+ CD38‐ naive B‐cells	0.94	0.88 ~ 1.00	0.0367
CD33 on CD33dim HLA‐DR+ CD11b + cells	0.93	0.86 ~ 1.00	0.0369
CD20 on IgD+ B‐cells	0.93	0.86 ~ 0.99	0.0283
HVEM on TD CD4+ T‐cells	0.92	0.85 ~ 1.00	0.0461
CD33br HLA‐DR+ CD14dim absolute count (AC)	0.92	0.87 ~ 0.98	0.0057
CD28 on activated Tregs	0.90	0.81 ~ 1.00	0.0410
CD127‐ CD8br %CD8br	0.89	0.79 ~ 1.00	0.0440
CD66b on CD66b++ myeloid cells	0.89	0.82 ~ 0.96	0.0050
CD127 on CD45RA‐ CD4 (not Tregs)	0.89	0.79 ~ 0.99	0.0364
CD16‐CD56 on NK cells	0.87	0.78 ~ 0.97	0.0128
CD16+ monocyte %monocyte	0.86	0.75 ~ 1.00	0.0488
CD24 on transitional cells	0.81	0.69 ~ 0.96	0.0131
CD28 on resting Tregs	0.80	0.66 ~ 0.98	0.0333
CD25 on activated Tregs	0.80	0.67 ~ 0.97	0.0199
TD CD4+ absolute count (AC)	0.80	0.65 ~ 0.98	0.0314

3.2. Impact of IPF Disease on Immunophenotypes

Reverse Mendelian randomization analysis did not reveal a significant causal association between IPF disease and immune phenotypes (Figure 2 and Table S2). Specifically, the ORs for the examined traits, such as HVEM on TD CD4+ (OR: 1.05, 95% CI: 0.99–1.11, p = 0.9328), CD28 on resting Treg (OR: 1.04, 95% CI: 0.95–1.13, p = 0.4056), and CD127 on CD45RA‐ CD4 not Treg (OR: 1.04, 95% CI: 0.99–1.08, p = 0.9981) were close to 1.00, indicating no strong effect in either direction. Additionally, the confidence intervals for these ORs spanned 1.00, and p‐values were above conventional significance thresholds, further emphasizing the lack of statistical significance. This pattern was consistent across all measured traits, such as CD25 on activated Tregs (OR: 1.00, 95% CI: 0.97–1.04, p = 0.8230) and Sw mem % B cell (OR: 1.00, 95% CI: 0.97–1.04, p = 0.9928).

FIGURE 2.

IVW method for analyzing phenotypes associated with immune cells in idiopathic pulmonary fibrosis.

3.3. Sensitivity Analysis

Sensitivity analyses, including heterogeneity and multiple comparison corrections, were conducted to validate the robustness of our findings in Tables S3 and S4. Leave‐one‐out analyses identified multiple SNPs that influenced the causality assessment. The F‐statistic for instrumental variables (IVs) ranged from 0.266 to 0.960, indicating no weak instrument bias in this MR study. We employed gene multiplicity tests, heterogeneity tests, and leave‐one‐out methods to ensure the reliability of our results. Additionally, the MR‐PRESSO global test and MR‐Egger regression were used to detect and correct for pleiotropy, enhancing the robustness of our causal inferences. MR‐PRESSO detects outliers by comparing observed vs. expected residual sum of squares under the null hypothesis of no pleiotropy, while MR‐Egger regression tests directional pleiotropy via its intercept term. However, MR‐Egger assumes the InSIDE condition (instrument strength independent of direct effects), which, if violated, may bias estimates. MR‐PRESSO is more robust to violations but requires sufficient outlier SNPs for correction.

4. Discussion

This study investigated the causal relationship between 731 immune cell traits and IPF by analyzing data from GWAS databases and using MR analysis. Our results demonstrate that certain immune cell traits are significantly associated with the risk of developing IPF, indicating potential therapeutic targets. According to our findings, six immune panels (B‐cells, cDCs, myeloid cells, monocytes, maturation stages of T‐cells, TBNK, and Tregs) and twenty‐three immune traits significantly affect IPF. The immune traits Switched memory B‐cells (Sw mem % B‐cell), IgD‐ CD38dim B‐cells, CD25 on CD4 Treg cells, CD45RA+ CD28‐ CD8br T‐cells, and CD45RA+ CD28‐ CD8br AC were found to be the main factors increasing IPF risk.

Switched memory B‐cells originate from the germinal center and include preswitched IgM+ cells and switched IgG, IgA, and IgE immune cells. They enhance the body's immune response during viral infections, and in the case of severe acute respiratory syndrome coronavirus‐2 (SARS‐CoV‐2), they increase the risk of IPF [10]. Switched memory B‐cell subpopulations remain positively correlated with the development of pneumonia. IgD‐ CD38dim B‐cells are mainly found in the spleen and lymph nodes. These relatively naive immune cells, which differentiate from the B‐cell subset, can induce B‐cell proliferation and promote IPF through relevant immune cell interactions [11]. The CD38dim B‐cell percentage can exacerbate SARS‐CoV‐2 infection in children [12]. Literature suggests that B cells, especially the naive CD38dim B‐cell population, are closely related to the pathogenesis of IPF, increasing Bruton's tyrosine kinase (BTK) levels, promoting substrate level phosphorylation, and thereby promoting the development of IPF [12].

Our results found that CD25 expression on CD4 Tregs showed an increasing risk of IPF. CD25 expression on CD4 Tregs, a subpopulation of natural regulatory T cells distributed in peripheral blood and splenic tissues, can inhibit T‐cell proliferation and induce autoimmune diseases through antigenic nonspecific mechanisms [13]. CD4 and CD25 proteins can increase the expression of the transcription factor Foxp3 [14]. CD45RA+ CD28‐ CD8br T cells secrete TGF‐β and cytokines IL‐10, IL‐2, and IL‐6 [15], promoting the proliferation of human lung fibroblasts via the TGF‐β pathway through IL‐6/STAT3/Smad3 trans‐signaling [16]. These cells are a subpopulation of CD8 cells in which T cells re‐express CD45RA, categorized as terminally differentiated effector memory cells (TEMRA) [17]. These cells are strongly associated with the severity of IPF post‐injury [18]. CD28‐ expression is an indicator of impaired telomere function in T cells, potentially producing a senescent‐associated secretory phenotype (SASP), promoting systemic inflammation [19]. A retrospective study of immune cell markers in the peripheral blood of 65‐year‐old patients with IPF found that CD45RA was significantly elevated in pretreatment patients but not after nintedanib treatment [20].

CD20 is an IgD nanocluster organization distributed on the B‐cell membrane, restoring B‐cell surface protein expression. Anti‐CD20 antibody rituximab treatment is used for B‐cell tumors and related autoimmune diseases [21]. Inhibition of CD20 B cells in bleomycin‐induced pulmonary fibrosis involves the removal of mature B cells through ablative therapy [22]. CD28 on activated Tregs, dependent on the T‐cell receptor (TCR) and the CD28 signaling pathway, effectively inhibits effector T cell proliferation. CD28 T cells promote lung structural remodeling in IPF by inhibiting alveolar type II epithelial cells or regulating surfactant protein C production [23]. CD127‐ CD8br T cells, a regulatory B‐cell (Breg) subtype, produce IL‐10 and TGF‐β. These cells are potential antifibrotic targets in IPF [24], with CD127, the α‐subunit of the IL‐7 receptor, distributed in monocytes and promoting inflammation in various diseases [25].

Mechanistically, switched memory B‐cells may promote fibrosis by secreting autoantibodies that activate lung fibroblasts via BTK signaling, enhancing collagen deposition [12]. Similarly, CD25+ Tregs could impair immune homeostasis, permitting unchecked TGF‐β and IL‐6 secretion from CD8+ TEMRA cells [16, 20]. These cytokines drive fibroblast‐to‐myofibroblast differentiation through STAT3/Smad3 trans‐signaling, a pathway corroborated by elevated TGF‐β levels in IPF patient sera. The role of CD20 on B‐cells aligns with preclinical models where B‐cell depletion via anti‐CD20 therapy reduced ECM accumulation in bleomycin‐induced fibrosis, suggesting a direct link between B‐cell activity and fibrotic progression [22].

Reverse Mendelian randomization analysis did not reveal a significant causal association between IPF disease and immune phenotypes. This may be due to several factors, including a weak causal link between genetic variants and immune phenotypes or IPF, insufficient sample size, and confounding by uncontrolled variables. These non‐significant results do not negate the existence of a causal relationship but suggest the need for further research with optimized study designs and larger sample sizes. Moreover, IPF is a relatively rare disease, leading to fewer genome‐wide significant variants that can be used as robust instruments, which in turn limits statistical power. Additionally, it is possible that any effect of IPF on immune traits is modest or context‐dependent, requiring larger multi‐ethnic datasets or refined phenotypic definitions to detect.

These findings are consistent with clinical studies showing increased levels of inflammatory cytokines (e.g., IL‐6, TNF‐α) in IPF patients, as well as animal models demonstrating that inhibition of key immune pathways can slow fibrotic progression [26, 27, 28, 29, 30]. By integrating our genetic evidence with these clinical and experimental observations, we further deepen the biological plausibility that immune cell dysregulation contributes to IPF.

The study has certain limitations. First, the reverse immune cell Mendelian randomization study did not identify a significant causal relationship between IPF disease and immune phenotype, potentially due to a weak causal link and limited sample size. This outcome does not disprove the existence of a causal connection but warrants further in‐depth investigation. Second, the GWAS database used predominantly includes European populations, which may limit the generalizability of our results to other ethnic groups. Future studies should aim to include more diverse populations to enhance the applicability of the findings. Lastly, while our MR approach reduces confounding by using genetic variants as instrumental variables, horizontal pleiotropy remains a critical source of bias. Although we employed MR‐PRESSO and MR‐Egger regression to detect and account for pleiotropy, these tools cannot completely eliminate the bias. In addition, the relatively smaller sample size in some GWAS datasets, especially for IPF, may limit the statistical power to detect weaker causal effects. Future studies leveraging larger consortia or meta‐analyses and incorporating diverse ethnic populations are needed to validate and extend our findings.

Future research should focus on validating these findings in larger, more diverse populations to confirm their generalizability. Additionally, in‐depth investigation into the mechanisms by which these immune traits influence IPF pathogenesis could lead to the development of novel therapeutic strategies. By targeting these specific immune traits, it may be possible to develop treatments that more effectively manage or even prevent IPF, improving patient outcomes and quality of life. Furthermore, the criterion of a stringent p‐value threshold may exclude borderline phenotypes that could still be biologically relevant. As such, future studies could adopt a two‐stage approach, initially applying a more inclusive threshold followed by replication analyses to ensure broader coverage of potentially significant immune traits. Finally, differences in genetic architecture and immune response across ethnic groups may yield divergent associations. Therefore, future research should include multi‐ethnic cohorts or trans‐ethnic meta‐analyses to validate whether the observed causal relationships persist in diverse populations.

In conclusion, this study identified specific immune cell traits that significantly influence the risk of developing IPF, suggesting these traits as potential therapeutic targets. The findings highlight the importance of immune modulation in IPF management and provide a foundation for future research to develop targeted treatments.

5. Visualization of Sensitivity Analysis

The results were analyzed using separate exclusion analyses, as described above, which demonstrated significant causal associations between SNPs, immune cells, and lung disease (Figure S1). Additionally, the symmetry of the funnel plots indicated the absence of heterogeneity's influence on the results (Figures S2 and S3).

Author Contributions

P.G. and X.L. designed the experimental process. P.G., X.C., and Y.L. carried out experimental operations. P.G. analyzed the data and wrote the manuscript. X.L. reviewed the content and grammar of the manuscript. All the authors have read and approved the final manuscript.

Ethics Statement

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1. Sensitivity‐analysis of causal effects of immune cells on IPF.

Figure S2. MRanalysis of immune cells in IPF funnel plots.

Figure S3. MRanalysis of the scatterplot of SNP effects on IPF.

Table S1. IVW results of causal effects of immune cells on IPF.

Table S2. IVW results of causal effects of IPF on immune cells.

Table S3. Causal effects of immune cells on IPF.

Table S4. Causal effects of IPF on immune cells.

Data Availability Statement

Genome‐Wide Association Studies data source: https://gwas.mrcieu.ac.uk/. GWAS publicly available database: http://ftp.ebi.ac.uk/pub/databases/gwas/summary_statistics/. Operating system(s): R for Windows GUI front‐end. Programming language: R. Other requirements: R 4.2.3. The data are all included in this manuscript.