Mendelian randomization analysis reveals the combined effects of epigenetics and telomere biology in hematologic cancers

Abstract

Background

Telomere shortening and epigenetic modifications are key factors in aging and hematologic diseases. This study investigates the relationship of telomere length and epigenetic age acceleration (EAA) with hematologic cancers, blood cells, and biochemical markers through the epigenetic clocks.

Methods

This study primarily utilizes genome-wide association studies of populations of European descent as instrumental variables, exploring the causal relationships between exposures and outcomes through a bidirectional two-sample Mendelian randomization (MR) approach. MR techniques include inverse variance weighted (IVW), MR Egger, and weighted median modes. Heterogeneity and pleiotropy in MR are assessed using Cochran's Q test and the MR Egger intercept, with the robustness of the conclusions further validated by multivariable MR (MVMR).

Results

Our research shows that longer telomere lengths significantly increase the risk of multiple myeloma, leukemia, and lymphoma (OR > 1, P < 0.05) and establish a causal relationship between telomere length and red blood cell indices such as RBC (OR = 1.121, P_IVW = 0.034), MCH (OR = 0.801, P_IVW = 2.046e-06), MCV (OR = 0.801, P_IVW = 0.001), and MCHC (OR = 0.813, P_IVW = 0.002). Additionally, MVMR analysis revealed an association between DNA methylation PhenoAge acceleration and alkaline phosphatase (OR = 1.026, P_IVW = 0.007).

Conclusion

The study clarifies the relationships between telomere length, EAA, and hematological malignancies, further emphasizing the prognostic significance of telomere length and EAA. This deepens our understanding of the pathogenesis of hematological diseases, which can inform risk assessment and therapeutic strategies.

Introduction

During the aging process, profound transformations occur in cellular, subcellular, and nuclear levels, with telomere attrition identified as a key molecular hallmark [1]. Telomeres are composed of repetitive DNA sequences and are associated with protein complexes such as shelterin, playing a crucial role in maintaining chromosome stability and integrity [2]. Telomerase, primarily regulated by the TERT gene encoding human telomerase reverse transcriptase and the TERC gene encoding template RNA, governs the maintenance of telomere length. Epigenetic modifications of human telomeres play a relevant role in telomere function and cellular proliferation [3]. Telomeres and subtelomeric regions are rich in epigenetic markers, which regulate telomere length through DNA methylation and post-translational histone modifications [4]. Unlike other chromosomal regions, telomeres are specialized structures that cap the ends of linear chromosomes, preventing them from being recognized as DNA breaks. Epigenetic modifications in telomeres are crucial for regulating the access of telomerase and other DNA repair proteins to telomere ends. In most somatic cells, telomere length progressively shortens during cell division, and when a critical length is reached, cells enter replicative senescence and can no longer divide. The TERT gene is abnormally overexpressed in 80%–90% of human malignancies, enabling tumor cells to maintain telomere length, thereby avoiding replicative aging and apoptosis. Extensive research has demonstrated a significant relationship between telomere length and the risk of blood diseases [5, 6].

Previous studies have established a robust correlation between DNA methylation at specific CpG sites and chronological age, known as the epigenetic clocks. Predictions from these epigenetic clocks often diverge from actual age, leading to what is known as epigenetic age acceleration (EAA) [7, 8]. These clocks, particularly as measured through various DNA methylation age indices like estimated granulocyte proportions (Gran) [9], Hannum [10], Horvath [11], GrimAge [12], and PhenoAge [13], have been utilized as a biomarker of aging, correlated with a healthy lifespan. Compared to traditional and emerging biomarkers, such as telomere length and omics-based indicators, the epigenetic clocks has proven more effective in predicting chronological age and mortality [13].

First-generation epigenetic clocks, such as HannumAge [10] and Intrinsic HorvathAge [14], which utilize DNA methylation levels at CpG sites, correlate with chronological age. HannumAge, indicative of 'extrinsic' epigenetic age acceleration (EEAA), serves as a biomarker for the aging immune system, capturing elements of immune decline including age-related changes in blood cell counts linked to lifestyle and health. Conversely, HorvathAge is determined by calculating the difference between epigenetic age and chronological age. This metric provides an overall assessment of epigenetic age across various tissues and cell types, capturing the overall biological aging process [14]. IEAA specifically measures the "pure" epigenetic aging effects in blood that are not influenced by differences in blood cell counts. This parameter focuses on intrinsic cellular aging, independent of changes in external cellular composition, thus providing a more refined assessment of the intrinsic aging process [11]. More recently, 'second-generation' epigenetic clocks like PhenoAge [13] and GrimAge [12] have been developed to predict age-related morbidity and mortality. PhenoAge integrates CpG data correlated with mortality and clinical biomarkers, while GrimAge includes metrics related to smoking pack-years and seven plasma proteins, namely cystatin C, leptin, tissue inhibitor of metalloproteinases 1, adrenomedullin, β-2-microglobulin, growth differentiation factor 15, and plasminogen activator inhibitor-1 (PAI-1). These second-generation clocks include a broader range of CpG sites and clinical biomarkers that are relevant to mortality and disease.

Although changes in telomere length do not directly affect the methylation status of CpG sites in epigenetic clocks, both telomere dynamics and epigenetic modifications are influenced by similar biological processes (such as oxidative stress and cellular senescence). Therefore, studying the association between the two helps us understand how these biological processes jointly affect cellular aging and health. Observational studies suggest that telomere length and EAA are potential risk factors for certain cancers; however, inherent study limitations and external variables complicate the establishment of direct causal relationships. This article investigates the possible mechanisms that associate telomere length and EAA with various hematologic malignancies, circulating blood cells, and biochemical markers. This research utilizes bidirectional two-sample Mendelian randomization (MR) and multivariable Mendelian randomization (MVMR) to navigate the complexities of these associations. Utilizing genetic variants as natural tools, this method enables a more definitive assessment of the effects of susceptibility between exposures and outcomes. By integrating genetic data with multiple clinical indicators, we aim to elucidate the causal connections between telomere length, EAA, and hematologic malignancies, thereby providing essential insights for future preventive and therapeutic strategies.

Materials and methods

Ethical considerations

While this research does not entail the participation of human or animal subjects, the team remains steadfast in preserving the integrity and confidentiality of the historical documents under examination. All pertinent literature and data will be sequestered securely, with stringent measures in place to prohibit the disclosure of sensitive information pertaining to individuals or organizations. A commitment to transparency will be unflaggingly upheld throughout the research lifecycle, encompassing methodology, data management, and analytical outcomes [8].

Research design and data source

This study utilizes a bidirectional two-sample MR approach to analyze telomere length, incorporating GWAS summary data authored by Codd et al [15] (2021,PMID:34,611,362), which encompasses 20,134,421 SNPs identified via advanced sequencing technologies. For EAA, we sourced summarized genetic association estimates from recent GWAS meta-analyses of biological aging, including mitochondrial DNA copy number [16] (2022,PMID:35,023,831), DNA methylation-estimated granulocyte proportions [9] (2021,PMID:34,187,551), DNA methylation GrimAge acceleration [9] (2021,PMID:34,187,551), DNA methylation PhenoAge acceleration [9] (2021,PMID:34,187,551), DNA methylation Hannum age acceleration [9] (2021,PMID:34,187,551), DNA methylation-estimated PAI-1 levels [9] (2021,PMID:34,187,551), and intrinsic epigenetic age acceleration [9] (2021,PMID:34,187,551) (Table 1). Initial data for outcome variables on hematologic malignancies, such as multiple myeloma, leukemia, lymphoma, and their subtypes, were obtained from relevant databases and literature (Table 1). All participants were of European descent and provided informed consent. Biochemical indicators analyzed included lactate dehydrogenase [17] (2019,PMID:29,403,010), serum alkaline phosphatase levels [18, 19] (2021,PMID:34,226,706); (2021,PMID:34,594,039), and circulating blood cell metrics such as red blood cell count [18] (2021,PMID:34,226,706), hemoglobin concentration [20](2020,PMID:32,888,493), hematocrit [20] (2020,PMID:32,888,493), mean corpuscular hemoglobin [19] (2020,PMID:34,594,039), mean corpuscular volume [18] (2021,PMID:34,226,706), mean corpuscular hemoglobin concentration [18] (2021,PMID:34,226,706), red cell distribution width [19] (2020,PMID:32,888,493), platelet count [20] (2020,PMID:32,888,493), and mean platelet volume [20] (2020,PMID:32,888,493). The Blood Cell Consortium provided the counts for white blood cells, neutrophils, lymphocytes, monocytes, basophils, and eosinophils [21] (Table 1). Compared to the one-sample MR method, the two-sample MR approach offers enhanced efficiency and robustness by leveraging data from multiple independent GWAS studies. This research strictly adheres to the three foundational principles of the two-sample MR framework: (1) instrumental variables (IVs) are strongly associated with the exposure; (2) IVs are not associated with any confounders that might influence the exposure–outcome relationship; and (3) IVs affect the outcome exclusively through their impact on the exposure.

Table 1 Information related to the data of exposure and outcomes

Instrumental variables (IVs) selection

To undertake a MR analysis, adherence to three cardinal assumptions is imperative: relevance, independence, and exclusion of confounding [23]. Consequently, the IVs slated for subsequent examination must be meticulously vetted. SNPs demonstrating a robust correlation with telomere length and EAA were initially selected, and those manifesting F-values less than 10 were subsequently excised to mitigate the influence of weak instrument bias and enhance result fidelity [24]. We utilized Phenoscanner V2 to investigate the relationships between specific SNPs and various phenotypes. Phenoscanner V2 integrates data from multiple GWAS studies and allows users to search for associations with specific SNPs, genes, or phenotypes (Fig. 1).

Fig. 1

Mechanism diagram

Data harmonization and attrition control

In this study, we harmonized GWAS datasets. We selected telomere length and EAA as exposures, and hematologic malignancies, circulating blood cells, and biochemical markers as outcomes. This was done to ensure consistency among effect alleles and to exclude palindromic SNPs with moderate allele frequencies. By leveraging summary statistics from an extensive consortium, we minimized dataset attrition bias and performed sensitivity analyses on SNPs with incomplete data. Adhering to Mendelian randomization principles, the study presupposes the random segregation of alleles during meiosis, thus bolstering the credibility of our causal inferences.

Multivariable Mendelian randomization

We conducted a MVMR analysis to clarify the relationship between EAA and serum alkaline phosphatase levels. Utilizing an extension of the inverse variance weighted (IVW) method, termed MVMR-IVW, we selected either random or fixed effects based on heterogeneity, as outlined in the two-sample MR approach. Our analysis revealed a significant correlation between DNA methylation indicators of aging—PhenoAge acceleration, IEAA, GrimAge acceleration—and serum alkaline phosphatase levels. Subsequently, we conducted an additional MVMR analysis, treating these DNA methylation markers as exposures to further investigate their individual impacts on serum alkaline phosphatase levels, accounting for underlying genetic correlations.

Statistical analysis

We employed various statistical techniques to explore the genetic association between exposures and outcomes (exposures: telomere length and EAA, outcomes: hematologic malignancies, circulating blood cells, and biochemical markers). These included IVW, MR Egger, weighted median, simple mode, and weighted mode approaches. The IVW method, relying on the validity of all selected SNPs, served as the primary analytical tool. To ensure the reliability of the results, we conducted comprehensive diagnostic assessments. We also performed multiple hypothesis testing, using a P-value threshold of less than 1.838 × 10^–4, after Bonferroni correction, to indicate the presence of a direct causal association. Additionally, we considered P-values between 1.838 × 10^–4 and 0.05 as indicative of a potential risk predictor. These assessments encompassed Cochrane's Q test and funnel plots for detecting heterogeneity, MR Egger intercept for assessing pleiotropy, MR-PRESSO for scrutinizing outliers, and leave-one-out analysis for confirming sensitivity. All statistical analyses were conducted using R version 4.3.3, and results were considered significant at P < 0.05.

Results

Selection of instrumental variables

Following a meticulous screening process, we identified SNPs that demonstrated a strong correlation with exposure data (with P-values < 5 × 10^-8 and F-values > 10) and independence (with r² < 0.001 within a 10,000 kb physical window). It is important to note that the use of Phenoscanner V2 to identify SNPs correlated with both outcomes and potential confounding variables did not lead to any exclusions at this stage. After aligning the exposure and outcome datasets, a carefully selected group of SNPs was designated for further MR analysis, as detailed in Additional files 1, 2. This selection process specifically excluded palindromic SNPs with medium allele frequencies to ensure the robustness of the MR analysis.

This study aims to elucidate the potential causal relationships between telomere length, EAA, and a range of hematologic malignancies, circulating blood cells, and biochemical markers. Employing bidirectional two-sample MR methods, we rigorously investigated these associations using a variety of MR techniques. The diversity of these methodologies provides a comprehensive evaluation of how variations in telomere length and EAA affect the risk of developing various types of hematologic malignancies, circulating blood cells, and biochemical markers.

Forward Mendelian randomization analysis

This study utilized the random effects IVW method as the primary analytical tool to explore the genetic connections between telomere length, EAA and hematologic malignancies, circulating blood cells, and biochemical markers (Table 2 and Additional file 4). The IVW method confirms that elongated telomere length is an independent risk factor for various hematologic malignancies. Specifically, there is a direct causal relationship between telomere length and lymphoid leukemia (P = 1.449e-08, OR [95%CI] = 1.002 [1.002,1.003]), and leukemia (P = 2.351e-06, OR [95%CI] = 1.003 [1.002,1.004]). Additionally, there are potential causal relationships between telomere length and multiple myeloma (P = 0.005, OR [95%CI] = 1.001 [1.000,1.002]), malignant lymphoma (P = 2.733e-03, OR [95%CI] = 1.403 [1.124–1.750]), Hodgkin lymphoma (P = 0.012, OR [95%CI] = 2.036 [1.166,3.554]), non-follicular lymphoma (P = 0.037, OR [95%CI] = 1.378 [1.019,1.862]), and other unspecified types of non-Hodgkin lymphoma (P = 0.006, OR [95%CI] = 1.982 [1.219,3.223]). Elongated telomere length is negatively correlated with several indices, including a direct causal relationship between telomere length and mean corpuscular hemoglobin (P = 2.046e-06, OR [95%CI] = 0.801 [0.731,0.878]), and potential causal relationships with mean corpuscular volume (P = 1.079e-03, OR [95%CI] = 0.801 [0.701,0.915]), mean corpuscular hemoglobin concentration (P = 2.307e-03, OR [95%CI] = 0.813 [0.712,0.929]), and eosinophil cell count (P = 7.353e-03, OR [95%CI] = 0.895 [0.825,0.971]). Conversely, extended telomere length positively correlated with neutrophil cell count (P = 2.627e-03, OR [95%CI] = 1.089 [1.030,1.151]) and platelet count (P = 9.552e-03, OR [95%CI] = 1.127 [1.030,1.233]). For every unit increase in mitochondrial DNA copy number, the risk of developing myeloid leukemia increases by 1.001 times (P = 0.044, OR [95%CI] = 1.001 [1.000,1.002]). The increase in mitochondrial DNA copy number may affect the occurrence of leukemia by influencing these processes. Additionally, we have observed a negative correlation between mitochondrial DNA copy number and serum alkaline phosphatase levels(P = 0.026, OR [95%CI] = 0.760 [0.598,0.967]). DNA methylation GrimAge acceleration positively correlated with serum alkaline phosphatase levels (P = 0.004, OR [95%CI] = 1.033 [1.010,1.056]) and negatively with hematocrit (P = 0.025, OR [95%CI] = 0.945 [0.900,0.993]). DNA methylation PhenoAge acceleration was positively correlated with several markers including serum alkaline phosphatase levels (P = 0.021, OR [95%CI] = 1.007 [1.001,1.013]; P = 0.043, OR [95%CI] = 1.005 [1.000,1.010]), and negatively with white blood cell count (P = 0.026, OR [95%CI] = 0.982 [0.967,0.998]), neutrophil cell count (P = 0.036, OR [95%CI] = 0.984 [0.969,0.999]), and leukemia (P = 0.049, OR [95%CI] = 1.000 [0.999,1.000]). Hannum age acceleration is a measure of biological aging that reflects epigenetic changes in DNA methylation patterns. We observed a negative correlation between DNA methylation-based Hannum age acceleration and the risk of BCR/ABL-positive chronic myeloid leukemia (CML). This finding suggests that a lower rate of DNA methylation Hannum age acceleration is associated with an increased risk of developing CML(P = 0.007, OR [95%CI] = 0.539 [0.343,0.846]). DNA methylation-estimated plasminogen activator inhibitor-1 levels negatively correlated with mean platelet volume(P = 0.009, OR [95%CI] = 1.000 [1.000,1.000]) and chronic myeloid leukemia, BCR/ABL-positive(P = 0.029, OR [95%CI] = 1.001 [1.000,1.002]). Intrinsic epigenetic age acceleration was positively correlated with serum alkaline phosphatase levels (P = 0.036, OR [95%CI] = 1.008 [1.001,1.015])(Fig. 2).

Table 2 Association of telomere length, EAA, and hematologic malignancies, circulating blood cells, biochemical markers using heterogeneity, pleiotropy, weighted median, and MR Egger analysis

Fig. 2

Causal effects of telomere length and EAA on hematologic malignancies, circulating blood cells, and biochemical markers (P-value < 0.05)

To evaluate the heterogeneity of individual causal effects, Cochran's Q test was employed. A p-value > 0.05 indicates homogeneity, while a p-value < 0.05 suggests heterogeneity, thereby necessitating the use of random effects IVW MR analysis(Table 2). Funnel plots demonstrated the symmetry of the selected SNPs (Supplement 3 S2, S5, S8,S11). Both the MR-Intercept and MR-PRESSO global tests excluded horizontal heterogeneity (Table 2 and Supplement 4), with MR-PRESSO additionally confirming the absence of outliers. Leave-one-out sensitivity analysis reinforced the consistency and reliability of the MR findings, even with the exclusion of individual SNPs (Supplement 3 S3,S6,S9,S12). Subsequent analyses employing MR Egger, weighted median, simple mode, and weighted mode methods corroborated the genetic causal relationships (Table 2, Supplement 3 S1,S4,S7,S10 and Supplement 4).

Reverse Mendelian randomization analysis

This study primarily utilized the IVW method to analyze genetic connections. The focus was on exploring the causal relationships between hematologic malignancies, circulating blood cells, biochemical markers, and both telomere length and EAA (Table 3). Our study utilized the IVW method to explore the genetic associations between hematological parameters and telomere length. We found that mean corpuscular hemoglobin (P = 0.006, OR [95% CI] = 0.960 [0.933, 0.989]), mean corpuscular volume (P = 0.002, OR [95% CI] = 0.963 [0.940, 0.986]), and mean corpuscular hemoglobin concentration (P = 0.001, OR [95% CI] = 0.959 [0.936, 0.983]) are inversely correlated with telomere length. This suggests that these parameters may serve a protective role in telomere maintenance. Furthermore, our results indicate that red blood cell count is an independent risk factor for telomere length (P = 0.001, OR [95% CI] = 1.053 [1.020, 1.086]), highlighting the complexity of the relationship between hematological parameters and telomere biology. Our study has demonstrated a significant positive correlation between neutrophil cell count and DNA methylation PhenoAge acceleration(P = 0.019, OR [95% CI] = 1.352 [1.051, 1.739]), suggesting that higher levels of neutrophils may contribute to an accelerated biological aging process. This finding indicates that neutrophils, which are a crucial component of the immune system and typically involved in inflammatory responses, may play a role in the aging process. Inflammation is recognized as a key driver of aging; hence, an increase in neutrophil count could indicate heightened inflammatory activity, which in turn could affect the rate of PhenoAge acceleration. Mean platelet volume is a risk factor for DNA methylation-estimated plasminogen activator inhibitor-1 levels (P = 0.015, OR [95% CI] = 1.313 [1.053, 1.637]) (Fig. 3 and Supplement 3  S13-S14).

Table 3 Association of hematologic malignancies, circulating blood cells, and biochemical markers with telomere length and EAA using heterogeneity, pleiotropy, weighted median, and MR Egger analysis

Fig. 3

Causal effects of hematologic malignancies, circulating blood cells, and biochemical markers on telomere length and EAA (P-value < 0.05)

Multivariable Mendelian randomization analysis

We applied multivariable Mendelian randomization analysis to evaluate the potential causal relationships between EAA and serum alkaline phosphatase levels. Compared to the results from the two-sample MR, the associations between intrinsic epigenetic age acceleration, DNA methylation GrimAge acceleration, and serum alkaline phosphatase levels were no longer significant (P > 0.05). Moreover, there is strong evidence that DNA methylation PhenoAge acceleration has a direct potential causal impact on serum alkaline phosphatase levels (P = 0.007, OR [95% CI] = 1.026 [1.007, 1.045]) (Table 4).

Table 4 Multivariable Mendelian randomization analysis of the potential causal effects of EAA on serum alkaline phosphatase levels

Discussion

This study employed a comprehensive GWAS dataset and bidirectional two-sample MR methods to clarify the causal relationships between telomere length, EAA and 16 types of hematologic malignancies, along with circulating blood cells, and biochemical markers. The findings reveal significant scientific implications: extended telomere lengths correlate with increased risks of various hematologic malignancies such as multiple myeloma, leukemia, lymphocytic leukemia, malignant lymphoma, Hodgkin's lymphoma, non-follicular lymphoma, and unspecified types of non-Hodgkin's lymphoma. Increasing research supports the association between telomere length and malignant blood diseases, emphasizing its potential as a biomarker for early detection and disease evaluation. Moreover, telomere length influences several blood cell indicators (RBC, MCH, MCV, MVHC, neutrophil cell count, eosinophil cell count, and platelet count). Additionally, this research explored the potential causal links between EAA and hematologic malignancies, as well as the influence on circulating blood cells and biochemical markers. Our findings indicate that increased EAA might be linked to a higher incidence of myeloid leukemia and possibly to serum alkaline phosphatase levels, which we confirmed through multivariable Mendelian analysis.

Several studies have established genetic risk scores (GRS) utilizing SNPs associated with telomere length, demonstrating a significant correlation with lymphoma risk [25, 26]. Moreover, findings from the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort study suggest that longer telomeres are associated with an increased risk of B cell lymphomas [27]. In prospective research on Hodgkin's lymphoma, pre-treatment telomere shortening was strongly linked to the development of secondary tumors [28, 29]. Clinical investigations have pinpointed genetically determined telomere length as a critical marker for multiple myeloma (MM) risk, substantially influencing patient survival rates [30, 31]. The regulation of telomerase activity in multiple myeloma (MM) stem cells, which controls clonal growth, is crucial. Studies have confirmed the role of telomere length as an independent prognostic marker for various leukemias, including chronic lymphocytic leukemia (CLL) 32,33,34], childhood acute lymphoblastic leukemia [35], adult T cell leukemia, and acute myeloid leukemia (AML) 36,37,38]. The mechanisms underlying the connection between telomere length and hematologic malignancies include potential chromosomal anomalies, DNA damage, and genomic instability due to telomere shortening, which could drive cancer cell development [39]. Telomeres, serving as protective chromosome caps, progressively shorten with each cell division, influencing cancer susceptibility. Additionally, the clonal expansion of mutated hematopoietic stem and progenitor cells (HSPCs), termed clonal hematopoiesis (CH) [40], and myeloid malignancies often act as genetic drivers of CH. This involves the expansion of specific blood cell lineages, which is correlated with aging and adverse health outcomes. Extended leukocyte telomere length is recognized as a causative risk factor for CH [41]. Observations suggest a bidirectional causal relationship between leukocyte telomere length (LTL) and clonal hematopoiesis of indeterminate potential (CHIP), thereby enhancing the understanding of the link between malignancies and telomere length [40]. Several models, such as the heterogeneous multihit theory and the biphasic effects of TERT promoter mutations, have been proposed to explain tumorigenesis processes [42]. An extended LTL might facilitate the acquisition of CHIP, heightening its mutagenic potential.

Bidirectional Mendelian randomization has confirmed a causal relationship between telomere length and various red blood cell indices (RBC, MCH, MCV, MCHC). In hematopoietic stem cells (HSCs), longer telomeres correlate with improved genomic integrity maintenance, which may prevent premature aging and enhance cellular proliferation and division capabilities. This facilitates the increased production and differentiation of erythroid progenitor cells, leading to higher red blood cell output. Optimal telomere length also minimizes replication-induced errors and premature aging, resulting in functionally intact and morphologically standard RBCs. However, extended telomeres that accelerate RBC maturation may yield smaller cells (lower MCV) and have less hemoglobin content per cell (lower MCH). This subsequently decreases the MCHC. This effect may stem from the accelerated transition of red blood cells through hematopoietic stages, potentially altering their growth and hemoglobin accumulation. Insights into telomere biology’s impact on RBC indices could elucidate the widespread occurrence of specific anemia types in older adults.

Numerous studies underscore the critical role of epigenetic regulation in hematologic cancer development. An correlation between epigenetic clocks and various forms of leukemia—including myeloid and chronic myeloid leukemia—has been established. Maegawa's research revealed that methylation changes, occurring as a function of age in normal hematopoiesis, accelerate during the transition to acute myeloid leukemia (AML) from myelodysplastic syndromes (MDS) [43]. Similarly, studies identified a significant relationship between epigenetic age acceleration and relapse timing in patients with chronic lymphocytic leukemia [44]. Furthermore, the aging of the hematopoietic system and the stem cell niche influences age-related phenotypes of hematopoietic stem cells, such as leukemia and lymphoma [45].

Analysis of EAA and biochemical markers has revealed correlations between several epigenetic indicators and lactate dehydrogenase. These indicators include the DNA methylation GrimAge acceleration, DNA methylation PhenoAge acceleration, and intrinsic epigenetic age acceleration. Such correlations suggest a potential causal relationship between epigenetic changes and increased serum alkaline phosphatase levels. Research shows that accelerated epigenetic aging might reflect chronic inflammation, which might influence ALP levels through multiple mechanisms. In liver diseases like non-alcoholic steatohepatitis, accelerated aging linked to these epigenetic clocks has been documented. Patients with high rates of age acceleration have been found to display elevated lactate dehydrogenase levels in liver biopsies [46]. Thus, the link between accelerated epigenetic aging and higher ALP levels could signify inflammation's role in aging and associated diseases. Serum alkaline phosphatase, crucial for bone formation and metabolism, may be impacted by accelerated epigenetic aging, potentially triggering pathways involved in bone remodeling and repair. This relationship might also illuminate the connection between liver health and the aging process, as the liver plays a significant role in methylation metabolism, with minor functional changes potentially affecting broader biological processes.

This study presents several advantages. Primarily, the GWAS is an effective tool for analyzing complex diseases, capable of identifying specific genes or gene clusters, thus overcoming the limitations inherent to single-gene association studies. This pioneering research explores the associations between telomere length, EAA, and GWAS data across 16 hematologic malignancies, circulating blood cells, and biochemical markers, thereby enhancing our understanding of these causal relationships. Moreover, the implementation of two-sample MR and multivariate MR methods successfully mitigates typical observational study limitations such as reverse causation, confounding variables, and biases. Rigorous screening and selection of instrumental variables have substantially increased analytical precision. A comprehensive suite of statistical tests assessed sensitivity, multicollinearity, and heterogeneity. Bidirectional Mendelian randomization further corroborated the causal relationships. Ultimately, our results underscore the capacity of epigenetics to predict and clarify the connections between biological aging and clinical biochemical markers. Future research should investigate the impact of these indicators across diverse populations and disease conditions, aiming to refine prevention and treatment strategies for age-related diseases.

Despite these advantages, there are still some limitations. First, the study population mainly consists of individuals of European descent, which may introduce ethnocentric bias. Although measures were taken to control for multicollinearity through MR-Intercept and MR-PRESSO global testing, the existing GWAS meta-analysis data prevents stratification based on geographical region, ethnicity, or age cohort. Therefore, the conclusions drawn from MR may lack generalizability in diverse populations. Second, whole blood samples may not accurately reflect the biological status of specific tissues or cell types. Therefore, future research should explore the associations between EAA and telomere length in different cell types and consider how these factors jointly affect disease risk and progression.

Conclusion

We have established the causal relationships between telomere length and EAA with hematologic malignancies, circulating blood cells, and biochemical markers. We have demonstrated that extended telomere length is an independent risk factor for various hematologic malignancies. Additionally, we have identified a causal relationship between telomere length and multiple red blood cell indices, including RBC, MCH, MCV, and MCHC. These findings further underscore the critical role of epigenetic regulation in the development of hematologic cancers and highlight the significance of serum alkaline phosphatase. Consequently, our research underscores the potential clinical implications of telomere length and EAA, which presents promising opportunities for clinical applications.

Data Availability Statement

No datasets were generated or analyzed during the current study.

Competing interests

The authors declare no competing interests.