A Mendelian Randomization Study Found Causal Linkage Between Telomere Attrition And Chronic Kidney Disease

Contact: Audrey Hu Whatsapp/hp: 0086 13880143964 Email: audrey.hu@wecistanche.com

Sehoon Park, Soojin Lee, Yaerim Kim, Semin Cho, Kwangsoo Kim, Yong Chul Kim, Seung Seok Han, Hajeong Lee, Jung Pyo Lee, Kwon Wook Joo, Chun Soo Lim, Yon Su Kim, and Dong Ki Kim

1 Department of Biomedical Sciences, Seoul National University College of Medicine, Seoul, Korea;

2 Department of Internal Medicine, Armed Forces Capital Hospital, Gyeonggi-do, Korea;

3 Division of Nephrology, Department of Internal Medicine, Uijeongbu Eulji University Medical Center, Gyeonggi-do, Korea;

4 Department of Internal Medicine, Keimyung University School of Medicine, Daegu, Korea;

5 Department of Internal Medicine, Seoul National University Hospital, Seoul, Korea;

6 Transdisciplinary Department of Medicine and Advanced Technology, Seoul National University Hospital, Seoul, Korea;

7 Kidney Research Institute, Seoul National University, Seoul, Korea;

8 Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea; and

9 Department of internal medicine, Seoul National University Boramae Medical Center, Seoul, Korea

Chronic kidney disease (CKD)is highly prevalent in the elderly population. However, it is rarely investigated whether kidney function is causally linked to biological aging itself. In this Mendelian randomization study, genetic instruments for telomere attrition were applied to a CKDGen genome-wide association study results for 41,395 cases of CKD among 480,698 individuals as summary-level Mendelian randomization. A replicative analysis was performed by polygenic score analysis using independent United Kingdom Biobank data for 8,118 cases of CKD among 321,024 white individuals of British ancestry. Reverse-direction Mendelian randomization analysis was performed utilizing genetic instruments for log-estimated glomerular filtration rate change with Z-standardized telomere length outcome data for 326,075 participants in the UK Biobank. Genetic predisposition toward telomere attrition (one Zscore decrease in length) was found to be a causative factor for a higher CKD risk[Odds Ratio 1.20 (95%confidence interval 1.08-1.33)], as supported by pleiotropyrobust Mendelian randomization sensitivity analyses implemented using the CKDGen data. Based on United Kingdom Biobank data, the polygenic score for telomere attrition was significantly associated with a higher risk of CKD [1.20(1.04-1.39)]. In reverse-direction Mendelian randomization, the genetically predicted kidney function decrease was significantly associated with a higher degree of telomere attrition [beta 0.039 (0.009-0.069)]. Thus, our study supports the causal linkage between telomere attrition and kidney function impairment.

KEYWORDS: aging; chronic kidney disease; Mendelian randomization; telomere

Cistanche improve kidney function

Chronic kidney disease (CKD), leading comorbidity worldwide, was estimated to have affected nearly 7 million individuals in 2017.1 As CKD is related to a large socioeconomic burden and a high risk of mortality, the identification of causative factors for CKD is important for elucidating biological pathways that are potential therapeutic targets.

As CKD(Chronic kidney disease) prevalence is high in the elderly population, there have been debates about whether a direct effect of aging, independent of coexisting comorbidities, on the risk of CKD exists.2,3 Such a biological aging process may be assessed by telomeres,4 of which are repeated nucleotide sequences of TTAGGG located at the ends of chromosomes that serve to protect DNA and secure chromosomal stability. Telomeres shorten with aging, and the degree of telomere loss varies among individuals. Furthermore, accelerated shortening of telomeres, a biomarker of cellular aging, is a causative factor for degenerative disorders, such as coronary artery disease and Alzheimer's disease.5–7 However, whether telomere attrition has a causal effect on the risk of CKD has yet to be confirmed. In addition, whether kidney function impairment is a causative factor for accelerated telomere attrition should be investigated to reveal the consequences of CKD on the biological aging process. Nevertheless, telomere length is a difficult variable to measure on large scale, and conventional observational findings are inevitably affected by reverse causation or unmeasured confounding effects.8,9

Mendelian randomization (MR) is a widely adopted analytic method to reveal causal effects from various genetically predictable exposures on complex diseases,10 and MR has been implemented in the field of nephrology.11–13 As instrumented genetic information is fixed before birth, genetically predicted exposures in MR are minimally affected by reverse causation or confounding effects. Indeed, the causal effects of telomere attrition on various degenerative diseases have been investigated through the MR approach via the implementation of genetically predicted leukocyte telomere lengths.5–7,14

In this study, we performed a 2-sample MR analysis to test the causal linkage between telomere attrition and CKD using 2 population-scale independent genetic databases. To assess whether there is a direct effect of biological aging on CKD risks, we hypothesized that a shorter telomere length would be a causal factor increasing the risk of CKD. Furthermore, we investigated reverse-direction causal effects of kidney function impairment on telomere attrition.

CISTANCHE BENEFIT: PREVENT KIDNEY FAILURE

METHODS

Ethical considerations

The Institutional Review Boards of Seoul National University Hospital approved the current study (Institutional Review Board No, E-2102-007-1192). UK Biobank approved the study and investigation of the database (approval No.53799). The requirement for informed consent was waived because we studied anonymous data in the public domain.

Study setting

This study involved a 2-sample bidirectional MR analysis(Figure 1). First, causal estimates based on telomere attrition on the risk of CKD were tested by implementing genetic instruments for telomere length. A summary-level MR analysis was performed with data from CKDGen Consortium,15, and replicative individual-level MR was performed via polygenic score (PGS)analysis using UK Biobank data.6

To investigate reverse-direction causal estimates from kidney function impairment on telomere attrition, the genetic instruments used included lead single-nucleotide polymorphisms(SNPs)associated with log-transformed estimated glomerular filtration rate (eGFR) values identified by the CKDGen study. Outcome summary statistics for telomere attrition were developed from the telomere length data provided by UK Biobank.

MR assumptions

MR analysis requires 3 assumptions to demonstrate causal effects.10 The relevance assumption is that the instrument should be strongly associated with the exposure phenotype, and as we implemented genetic variants identified by genome-wide association study (GWAS), this was considered to be attained. The independence assumption is that the instrument should not be associated with confounders and we performed pleiotropy-robust MR sensitivity analysis, which provides causal estimates with the relaxation of this assumption. We also conducted MR sensitivity analysis by excluding SNPs with potential associations with confounders. The exclusion-restriction assumption is that the causal effects should be through the exposure trait. Although this assumption is untestable, we performed median-based MR analysis, which can relax this assumption for some of the instruments, serving as a sensitivity analysis.

CISTANCHE BENEFIT: IMPROVE KIDNEY FUNCTION SO TO IMPROVE SEXUAL FUNCTION

Genetic instruments for telomere length

A genetic instrument for telomere length was introduced in a previous GWAS meta-analysis for leukocyte telomere length5 The recently updated GWAS meta-analysis for leukocyte telomere length, which is the largest to date, included EPIC-InterAct (N = 19,799), EPIC-CVD (N= 11,915), and ENGAGE consortium (N = 46.898).

In this study, we included 52 independent SNPs identified with a false discovery rate < 0.05(P<1.03 x 10-5)and associated with leukocyte telomere length and their joint effect sizes as genetic instruments, as in a previous study.'Among them,3 monoallelic and 3 non-overlapping SNPs associated with outcome data were disregarded. and the instrumented SNPs explained 2.65% of the variance in leukocyte telomere length(Supplementary Table S1). As we tested the causal effects of telomere attrition, all GWAS effect sizes were aligned toward a"shortening" of telomeres.

CISTANCHE BENEFIT: IMPROVE KIDNEY FUNCTION SO TO IMPROVE SEXUAL FUNCTION

To attain the independence assumption, we performed an additional GWAS adjusted for age， sex， age × sex， age²， and the first 10 genetic principal components within the individual-level UK Bio-bank data and determined whether the chosen SNPshad genome-wide significant (P = 5 x 10-)or stronger association with a confounder than that with leukocyte telomere length. The potential confounders included hypertension, diabetes mellitus, cholesterol-lowering medications or blood lipid profiles, obesity, or smoking (previous or current; Supplementary Methods); however, none of the associations reached a genome-wide significance level or an association stronger than that with telomere length(Supplementary Table S1). We further screened Phenoscanner to identify potential pleiotropic effects. Among the 16 SNPs associated with phenotypes other than telomere length for genome-wide significance, except the 3 SNPs with the strongest associations with leukocyte telomere length (rs10936600[P = 6.42E-51],rs77005526 [P = 4.82E-45], and rs2853677 [P=3.12E-31]), we excluded the other 13 SNPs that had relatively weak associations with telomere length and performed a sensitivity analysis.

CKD(Chronic kidney disease) Gen outcome data for CKD(Chronic kidney disease)

CKD(Chronic kidney disease) Gen Consortium is the largest GWAS meta-analysis for CKD outcome to date, and it mainly includes individuals of European ancestry.5 There was partial overlap of the population in CKDGen Consortium and that in the GWAS for genetic instrument development (e.g., the TWINGENE or EGCUT studies).

Summary statistics for CKD(Chronic kidney disease) （eGFR<60 ml/min per 1.73 m²）risk are available in the public domain. Data for individuals of European ancestry (N =480,698 with 41,395 CKD cases)were downloaded as outcome data for CKD in the summary-level MR.

Summary-level MR analysis

The multiplicative random-effect inverse variance–weighted method is the main summary-level MR method.18 In addition, a series of robust MR sensitivity analyses were performed to yield valid causal estimates even under a condition of unbalanced pleiotropy.19 We implemented weighted median,20 simple medians,20 MR-Egger,21 MR–robust adjusted profile scores,22 contamination mixture,23 and MR–Pleiotropy RESidual Sum and Outlier (PRESSO)24 analyses, and the details are described in the Supplementary Methods. Single-SNP and leave-one-out analyses were also performed to investigate the presence of a disproportionate effect from one SNP. The summarylevel MR analysis was performed by the TwoSampleMR package, and a 2-sided P < 0.05 was considered signifificant.25

CISTANCHE BENEFIT: IMPROVE ADRENAL FUNCTION

Individual-level MR for CKD(Chronic kidney disease) with UK Biobank data

We additionally assessed causal estimates in the UK Biobank data, which are independent of those of CKDGen, as a replicative analysis. UK Biobank comprises a prospective cohort of individuals aged 50 to 65 years throughout the United Kingdom (N > 500,000).16,26 The data have strength because the study participants did not overlap with the population included in the GWAS for genetic instrument development, enabling a 2-sample MR without any overlapping samples and the availability of individual-level covariate information.

BENEFIT OF CISTANCHE

In this study, we included 337,138 unrelated White British individuals who passed the sample quality control filter.11 Among them, 321,024 individuals had available serum cystatin C and creatinine levels to determine kidney function outcomes. The PGS was calculated by multiplying the effect size b values with the gene dosage matrix by PLINK2.0.27 Then, the PGS was regressed to CKD outcomes with adjustment for age, sex, and the 10 genetic principal components. For sensitivity analysis, a multivariable logistic regression model was constructed, including the covariates assessed by the above-mentioned GWAS.28

We first evaluated CKD at stage ≥3 as the outcome, including eGFR < 60 ml/min per 1.73 m2 calculated from the cystatin C/ creatinine-based Chronic Kidney Disease–Epidemiology Collaboration equation.29 We also examined CKD outcomes based on creatinine-based eGFR values.

Reverse-direction MR to assess causal estimates from kidney function on telomere attrition

We further explored reverse-direction causal effects, including genetic instruments for kidney function, as developed by the above-mentioned CKDGen GWAS meta-analysis,1and investigated outcome data for telomere length (log-transformed and standardized to a Z score) developed from UK Biobank data (available in 326,075 White British participants). The CKDGen GWAS meta-analysis revealed 256 index SNPs with genome-wide significance within a 1-Mb segment for log-transformed eGFR values in 567,460 individuals of European ancestry. As in our previous bidirectional MR analyses,2.10 we excluded one nonoverlapping SNP and 115 SNPs that did not reach the Bonferroni-adjusted significance level(P<0.5/256)with eGFR values based on cystatin C level or with different directions of βvalues, as assessed by GWAS with UK Biobank data (Supplementary Table S2). The remaining 140 SNPs consisted of the genetic instrument for kidney function. To conservatively reach the independence assumption, we, in sensitivity analysis, assessed associations for potential confounders as above and additionally excluded 42 SNPs that showed a strong association (P<1×10-5) with any of the confounders. The above GWAS and summary-level MR analysis followed the above-mentioned methods. The directions of the summary statistics were aligned toward eGFR"decrease" and telomere"shortening" As there is a lack of other large-scale public summary data for telomere length, additional replication analysis was not performed for reverse-direction MR analysis.

RESULTS

From telomere attrition to CKD: summary-level MR results from MR by the inverse variance-weighted method revealed significant causal estimates for telomere attrition with a higher risk of CKD (odds ratio,1.20; 95% confidence interval,1.08-1.33; P<0.001)(Figure2 and Table 1). Moreover, the causal estimates remained significant with all implemented MR sensitivity analyses, and the MR-Egger intercept P-value supported an absence of significant directional pleiotropy (P ¼ 0.31). The leave-one-out analysis or single-SNP analysis also suggested an absence of a notable disproportionate effect from a given SNP in the causal estimates (Supplementary Figure S1). The results were similar when we excluded SNPs with a relatively weak association with telomere length but a strong association with other phenotypes, and only MR-Egger regression analysis provided nonsignificant causal estimates.

From telomere attrition to CKD: individual-level MR results

In PGS analysis with UK Biobank data, a higher PGS for telomere attrition was significantly associated with a higher risk of CKD, with both determined using creatinine-based eGFR or cystatin C/creatinine-based eGFR < 60 ml/min per 1.73 m2 (Figure 2 and Table 2). Furthermore, the causal estimates remained significant when we adjusted for multiple clinical covariates

From kidney function impairment to telomere attrition:summary-level MR results

In reverse-direction MR, we identified that genetic predisposition toward reduced kidney function was significantly associated with a higher degree of telomere attrition (Table 3 and Figure 2). Single-SNP and leave-one-out analyses showed the absence of a disproportionate effect from an individual SNP(Supplementary Figure S2). These results were supported by most pleiotropy-robust MR sensitivity analyses, except that MR-Egger regression did not provide significant causal estimates along with the MR-Egger intercept, indicating the possibility of directional pleiotropy(pleiotropy P=0.04).In sensitivity analysis conservatively excluding potentially confounder-associated SNPs, the causal estimates obtained by inverse variance-weighted,median-based, MR-robust adjusted profile score, contamination mixture methods and MR-PRESSO analysis were significant. Although MR-Egger regression again provided nonsignificant causal estimates, the MR-Egger intercept P-value(pleiotropy P=0.62)indicating directional pleiotropy was nonsignificant.

DISCUSSION

In this study using population-scale genetic data sets, we demonstrate that telomere attrition may have a causal impact on the risk of CKD. Furthermore, reverse-direction investigation supported that decreased kidney function impairment may be causative for accelerated telomere attrition. With ef-forts to reach MR assumptions, this study supports the causal linkage between kidney function impairment and telomere attrition.

Previous observational studies have suggested an association between telomere length and CKD. One study reported that coronary heart disease patients with CKD have shorter leukocyte telomere lengths, although the association was dependent on age," and leukocyte-derived relative telomere length was associated with the progression of CKD in 2 prospective cohorts."In addition, short telomere length correlates significantly with the progression of diabetic nephropathy in type 1 diabetic patients. However, as observational findings are inevitably affected by reverse causation or confounders, particularly in the elderly population with multiple comorbidities, a causal interpretation was limited in previous reports. Therefore, MR investigation is warranted to investigate CKD risk regarding telomere attrition. Although a previous MR analysis, which provided genetic instruments for the current analysis, "did assess CKD risk, the study reported null results for CKD outcome. However, in the study, only an unrealistically low number of CKD events from self-reports or a limited number of Inter-national Classification of Diseases, Tenth Revision(ICD-10), diagnostic codes were induced, which might result in false-negative findings. As CKD is best determined by laboratory evidence, as stated in the current guidelines and other MR studies in the nephrology field, we reinvestigated the causal effects of telomere attrition on the risk of CKD stage ≥3. In the current study, we identified significant causal estimate trends toward higher CKD risk due to telomere attrition in 2 large-scale genetic data sets assessing kidney function traits. We also demonstrated that decreased kidney function may have a causal effect on a higher degree of telomere attrition. Therefore, this study supports that telo-mere attrition and kidney function impairment have causal effects on each other.

This study highlights the importance of cellular senescence in CKD pathophysiology and provides a scientific background for the emergence of CKD in elderly individuals. It has been debated whether the current definition of CKD should be age-adjusted because there may be"normal" kidney function decline as individuals age. This notion is supported by the fact that the eGFR equation itself includes chronological age and the high prevalence of CKD diagnosis among the elderly population may be, in part, because of the limitation of the current methods to calculate eGFR. In addition, some researchers have considered that there is no pathological effect of aging on the kidneys and that the high CKD prevalence in the elderly population is due to the high prevalence of comorbidities, such as hypertension or diabetes,2 which would be supported if a direct effect of biological aging on the risk of CKD is absent. However, in the current study, telomere attrition was a significant causal factor for CKD, independent of chronological age or major comorbidities. Thus, the high CKD prevalence in the elderly population may reflect this causal impact of cellular senescence on kidney function. Furthermore, the results suggest that if an aging-related process, such as telomere attrition, can be modified, which may be achieved by healthy lifestyles or future interventions,33 age-related kidney function decline may not be inevitable or may even be preventable.3 Therefore, the current study advocates the maintenance of the current CKD definition for the elderly population, as this may capture the differences in biological age among individuals. This idea is supported by the fact that telomere attrition has causal impacts on important health outcomes, including cardiovascular diseases or neurodegenerative disorders6,7,14; thus, the elderly population with such accelerated biological aging should be diagnosed for the presence of CKD.

In addition, the current findings suggest that telomere length may be a targetable factor in the development of novel preventive measures to reduce the risk of CKD, Although a practical intervention directly modifying telomere length is not yet available, there have been efforts, and CKD can be considered a target disease for such a future strategy. On the other hand, the potential benefits of telomere lengthening should be pursued with caution, as longer telomere length has been suggested to be causative for various cancers, including kidney malignancies.

Reverse-direction MR supports that kidney function impairment may lead to a higher degree of telomere attrition. Thus, kidney function impairment may increase the risk of degenerative diseases by accelerating cellular aging, similar to other previously suggested factors, including smoking, which again obesity, or lack of physical activity, underscores the importance of efforts to prevent the progression of kidney function impairment in clinical practice.

Analytically, MR-Egger regression provided some nonsignificant causal estimates. However, weak statistical power is the limitation of this method.37 In addition, potential pleiotropic effects, as reflected by the significant MR-Egger intercept P-value, in reverse-direction MR were reduced in sensitivity analysis, as were the consistently significant results by other MR methods. Thus, we believe some null findings by MR-Egger regression do not disregard the overall results, particularly when considering the possibility of false-negative bias in 2-sample MR, although caution with regard to potential bias by pleiotropic effects in the results is necessary.

Some limitations in this study should be noted. First, there were some overlapping samples in the CKDGen data and GWAS meta-analysis that provided the genetic instruments, which may cause a bias toward observational associations.38 However, as most of the samples did not overlap and the UK Biobank outcome data were independent, such potential bias is unlikely to alter the main findings of this study. Second, we used leukocyte telomere length, which is a proxy for telomere length in kidney-indwelling cells. Although the measurement of telomere length in leukocytes is the most common approach used, further studies should investigate the direct effects of kidney cell telomere length on kidney function. Third, the direct biological mechanism mediating the causal link could not be investigated in this MR study. Hence, future studies are warranted to investigate whether the link between telomere attrition and CKD can be modified, not only by directly targeting telomere length but also by targeting related mechanistic pathways (e.g., inflammation). Last, the generalizability of the results should be considered, as the study data included individuals of European ancestry because of the availability of these data.6 In addition, healthy volunteer bias is present in the UK Biobank data,26 as reflected in the low prevalence of CKD, although this is not likely to alter the main result of this study, as the findings were similar to the CKDGen data.

In conclusion, telomere attrition is a causal factor for CKD, and a decrease in kidney function may accelerate telomere attrition. Clinicians should keep in mind the causal link between telomere length, representing the biological aging process, and kidney function impairment.

CISTANCHE BENEFIT: TREAT CHRONIC KIDNEY DISEASE

REFERENCES

1 Global, regional, and national burden of chronic kidney disease, 1990- 2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2020;395:709–733.

2. Glassock RJ, Delaney P, Rule AD. Should the definition of CKD be changed to include age-adapted GFR criteria? YES. Kidney Int. 2020;97: 34–37.

3. Levey AS, Inker LA, Coresh J. “Should the definition of CKD be changed to include age-adapted GFR criteria?”: con: the evaluation and management of CKD, not the definition, should be age-adapted. Kidney Int. 2020;97:37–40.

4. Benetos A, Okuda K, Lajemi M, et al. Telomere length as an indicator of biological aging: the gender effect and relation with pulse pressure and pulse wave velocity. Hypertension. 2001;37:381–385.

5. Haycock PC, Burgess S, Nunu A, et al. Association between telomere length and risk of cancer and non-neoplastic diseases: a Mendelian randomization study. JAMA Oncol. 2017;3:636–651.

6. Li C, Stoma S, Lotta LA, et al. Genome-wide association analysis in humans links nucleotide metabolism to leukocyte telomere length. Am J Hum Genet. 2020;106:389–404.

7. Zhan Y, Song C, Karlsson R, et al. Telomere length shortening and Alzheimer disease–a Mendelian randomization study. JAMA Neurol. 2015;72:1202–1203.

8. Fyhrquist F, Tiitu A, Saijonmaa O, et al. Telomere length and progression of diabetic nephropathy in patients with type 1 diabetes. J Intern Med. 2010;267:278–286.

9. Rosenberger J, Kollerits B, Ritchie J, et al. Association of relative telomere length with the progression of chronic kidney disease in two cohorts: effect modification by smoking and diabetes. Sci Rep. 2015;5:11887.