Telomere Length in Celiac Disease in a Population-Based Cohort: A HUNT Study

Abstract

INTRODUCTION:

Celiac disease (CeD) occurs in genetically predisposed individuals who are exposed to dietary gluten proteins. Genome-wide association studies have found CeD to be associated with a genetic predisposition to shorter telomere length. It was therefore of interest to measure leukocyte telomere length in a large cohort of patients with CeD.

METHODS:

This study was population-based and included a defined cohort of patients with CeD in the Trøndelag Health Study. Leukocyte telomere length was measured by quantitative polymerase chain reaction (PCR) in patients and controls 1:2 matched for age, sex, alcohol, and tobacco use.

RESULTS:

The study population consisted of 58.7% women, with mean age 43 (SD 14.5) years, where half were never smokers, and the majority consumed little or no alcohol. Telomer length in 1,077 patients with CeD did not differ significantly from 2,147 matched controls, 10.39 kb (SD 2.21) vs 10.52 kb (SD 2.25), P = 0.088. In a mixed linear model, we found higher age and alcohol consumption, but not tobacco use or sex, to be associated with shorter telomere length. The telomere length did not differ between patients with newly diagnosed and known CeD.

DISCUSSION:

This large population-based study provides evidence that CeD is not associated with clinically meaningful telomere shortening.

INTRODUCTION

Celiac disease (CeD) is an immune-mediated inflammatory disorder of the small intestinal mucosa that occurs in genetically predisposed individuals who are exposed to dietary gluten proteins of wheat (gliadin and glutenin), barley (hordein), or rye (secalin) origin (1,2). Various host factors may influence the risk of CeD, including genetic variants predisposing to shorter telomere length (3). Telomeres are the natural ends of eukaryotic chromosomes and consist of short tandem DNA repeats of the 6-nucleotide unit sequence TTAGGG, which protects the chromosomes against damage. Cells with critically short telomeres become senescent and, ultimately, lose their ability to divide. The accumulation of cells with short telomeres drives tissue ageing that with time is believed to contribute to numerous degenerative or age-dependent diseases (4,5). Telomere length is maintained by an enzyme complex called telomerase, and mutations which cause reduced function increase the risk of several diseases. Whereas telomeres undergo shortening by 60 bp/year in healthy individuals, the shortening may be 120 bp/year in patients with telomere syndromes. Single-gene inactivating mutations are known in 11 human genes to date, and these genes encode either a telomerase component (TERC, TERT, DKC1, NOP10, NHP2, WRAP53) or a telomere-binding protein found to have an essential protective function (TINF2, RTEL1, POT1, CTC1, TPP1) (4,6). Mutations in these genes are estimated to explain 50–70% of the short telomere Mendelian phenotype.

Although HLA-DQ2/8 carrier status is necessary for CeD development, non-HLA and unexplained genetic variation accounts for most of the genetic risk (7). Non-HLA genetic variation could include genes involved in cytokine signalling and B-cell and T-cell function (8–10). Various immune and bone marrow manifestations are also common in patients with telomere syndromes (4,5), and genetic variation in genes encoding telomerase components could possibly affect immune response and affect the risk of CeD.

Longer genetically predicted telomere length was found to be associated with lower risk of CeD in 2 large genome-wide association studies using Mendelian randomization (MR) (3,11). Studies where leukocyte telomere length has been measured in patients with CeD have shown conflicting results (11–13). Since MR estimates reflect lifetime genetic predisposition to telomere length, it was also of interest to measure actual telomere length in patients with newly diagnosed or known CeD.

In this study, we aimed (i) to determine telomere length in patients with CeD compared with controls and (ii) to examine the association between telomere length and CeD phenotype, i.e., if patients have previously known or new diagnosis of CeD.

METHODS

Design

The study was population-based and included a defined cohort of patients with CeD in the Trøndelag Health Study (HUNT) (14–16). It is a substudy of the Coeliac disease in HUNT, which has been approved by the Regional Committee for Medical and Health Research Ethics (ID # 943).

Subjects

In HUNT, residents in Nord-Trøndelag County, Norway, from 20 years of age were invited to participate in repeated health surveys. The second survey in 1995–1997 (HUNT2), the third survey in 2006–2008 (HUNT3), and the fourth survey in 2018–2019 (HUNT4) included questionnaires on health-related factors, standardized clinical examinations, and blood samples on 65,000, 51,000, and 56,000 individuals (response rate 70%, 54%, and 54%), respectively (17). Participants in HUNT are considered responders if they have completed at least the main questionnaire or taken part in the clinical examination. Most responders have completed all questionnaires, attended the clinical examination, and provided blood samples stored in the HUNT biobank. The biobank's organization, procedures for sample collection, its contents, and storage conditions have been described previously (18). Using the unique 11-digit person identification number assigned to every Norwegian resident, HUNT data can be linked to patient records at any Norwegian hospital and health registry. General characteristics of the HUNT participants have been compared with the Norwegian population (19), and only minor differences were found. In HUNT3, nonparticipants (n = 6,922) were found to have slightly lower socioeconomic status, higher mortality, and higher prevalence of some diseases such as diabetes mellitus, cardiovascular disease, and psychiatric disorders (20). The frequencies of common diseases found in patient registries at a national and regional level in Nord-Trøndelag have been compared with HUNT4 participants overall (19), but only small differences in the overall prevalence estimates of common diseases were found when also including data from nonparticipants. Specific data about participants and nonparticipants with CeD are not available in these studies. In conclusion, the HUNT participants provide a good representation of the general population at a regional and national level.

A cohort of patients with CeD has been identified by serological analyses (1), and all participants with positive serology were invited to clinical assessment and upper endoscopy with duodenal biopsies (2 from the duodenal bulb and 4 from the descending part), as previously described (15,16). Patients with CeD were diagnosed by demonstration of intraepithelial lymphocytosis (>25/100 enterocytes) and villous blunting, corresponding to Marsh grade 3 following current guidelines (21,22). In addition, participants with CeD diagnosed before HUNT4 participation were also identified through health registries, with a subsequent review of medical records including serology and histology, and these were added to the group of patients with “known CeD,” while participants diagnosed as part of HUNT4 participation were labelled “new CeD” (Figure 1, simplified from Lukina et al [15]). “Potential CeD” was defined as seropositive participants with histological Marsh grade <3 (15). The use of nonsteroidal anti-inflammatory drugs and Helicobacter pylori (H pylori) infection were assessed to rule out false positives, and participants with H pylori infection and recent nonsteroidal anti-inflammatory drugs use with Marsh grade 3 were in this study categorized as “potential CeD.” Seropositive HUNT4 participants who did not undergo upper endoscopy with biopsies were excluded from further analyses. Participants with “known CeD” were on a gluten-free diet, whereas exact compliance could not be evaluated at an individual level in this population-based study. Participants with new CeD consumed a regular, gluten-containing diet at the time of blood-sampling and biopsy collection.

Figure 1.

Flow diagram of participants in HUNT and formation of the study cohort of 1,077 patients with CeD who underwent analysis of leukocyte telomere length. CeD, celiac disease; HUNT, Trøndelag Health Study.

Previous publications have suggested that in addition to age, also tobacco smoking (23,24), and alcohol consumption (25) may shorten telomere length, and we aimed to reduce the influence of these potential confounding factors by study design. Patients with CeD and controls (1:2) were therefore matched for age, sex, smoking status (never, former, current, or missing information), and alcohol consumption (0, 1–3, 4–7, 8–14, 15+ units/week, or missing information) reported in the HUNT-study when DNA was collected for biobanking.

Analyses

Telomere length was analyzed in leukocyte DNA stored in the HUNT biobank from the cohort of participants with CeD (Figure 1) and matched controls (1:2). DNA was available from the first HUNT participation for each participant with CeD, and DNA was selected from the same HUNT study (2, 3, or 4) for matched controls to reduce theoretical effects of DNA extraction and storage. DNA was available from HUNT2 in 1,052 participants, from HUNT3 in 1,050 participants, and from HUNT4 in 1,122 participants. Extraction of DNA in HUNT2 was performed using Gentra Puregene Blood Kit (Qiagen GmbH, Hilden, Germany), whereas in HUNT3 and HUNT4, DNA extraction was performed using the Buffy Coat DNA Purification Protocol (Epicentre, Madison, WI) by the MasterPure DNA purification Kit for blood. Both extraction methods have yielded high DNA quality (18) appropriate for telomere length determination. In cases of low DNA concentrations in samples from controls, additional matched controls were analyzed in a second round to achieve the intended 1:2 matching ratio. As a result, all but 7 patients with CeD had 2 matched controls with available telomere length measurements.

Telomere length analysis by quantitative polymerase chain reaction

DNA isolation, quantification, and overall quality assessments were performed as described by Næss et al (18). Then, the DNA samples were diluted to 25 ng/μl and stored at −80°C till the quantitative polymerase chain reaction (qPCR) for telomere length was run. Genomic DNA (gDNA) isolated from HeLa cells, using the DNeasy Blood and Tissue kit (Qiagen), was used for generating a standard curve, run in technical quadruplicate, for both the telomere primer set and the reference gene 36B4. The standard curve ranged from 30 kb to 0.5 kb, and it was included in every qPCR run. The PCR efficiencies for both the primer sets were within 95%–105% in all the experiments. The gDNA from HeLa, used as standard and reference DNA throughout the runs, was aliquoted immediately after the isolation and stored at −80°C. Each aliquot was single used only to preserve the integrity of the gDNA.

Average telomere length relative to HeLa was measured from total gDNA samples by using the qPCR method previously described by Cawthon et al (26) and O'Callaghan and Fenech (27) with minor modifications. Briefly, each reaction included 5 μL 2× Power SYBR Green PCR Master mix (Applied Biosystems, Foster City, CA), 0.5 μL forward and 0.5 μL reverse primers (10 μM), 2 μL water, and 2 μL gDNA (5 ng/μL) to yield a 10 μL reaction. The single-copy gene 36B4 was used as reference. A QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems) was used with reaction conditions of 95°C for 15 minutes, 2 cycles at 94°C for 15 seconds, 58°C anneal for 15 seconds, and 74°C extend for 15 seconds and followed by 35 cycles of data collection at 94°C for 15 seconds, 62°C anneal for 10 seconds, and 74°C extend for 15 seconds. A standard curve was made using HeLa gDNA, and their known telomere length of 4.6 kb was used to calculate average telomere length in the samples. The following primer sequences were used: hTelo forward (5'-CGGTTTGTTTGGGTTTGGGTTTGGGTTTGGGTTTGGGTT-3'); hTelo reverse (5'-GGCTTGCCTTACCCTTACCCTTACCCTTACCCTTACCCT-3'); h36B4 forward (5'-CAGCAAGTGGGAAGGTGTAATCC-3'), and h36B4 reverse (5'-CCCATTCTATCATCAACGGGTACAA-3'). Triplets (1 case and 2 matched controls) were positioned on the same plate to reduce effects of interassay variation.

Statistical methods

Telomere length was analyzed with a linear mixed model with a triplet-specific random intercept, CeD status (patients versus controls) as fixed effect and adjusted for sex, age, smoking status, and alcohol consumption. The effect of disease phenotype was assessed within the group of patients with CeD using a linear model adjusted for the same covariates. Analyses were performed using mySQL version 8.0.42 (Oracle, Austin, TX) and SPSS version 30.0 (Armonk, NY).

RESULTS

Characteristics of the study population are summarized in Table 1. From the original cohort of HUNT4 participants with suspected CeD based on serology, individuals were excluded because endoscopy (n = 325) or biopsy (n = 8) for histological evaluation had not been performed, resulting in 1,077 patients (Figure 1) and 2,147 matched controls.

Table 1.

Characteristics of the study population (n = 3,224), consisting of patients with celiac disease (n = 1,077) and matched controls (n = 2,147)

	Celiac disease n = 1,077	Controls N = 2,147
Men, n (%)	445 (41.3)	890 (41.5)
Women, n (%)	632 (58.7)	1,257 (58.5)
Agea in years, mean ± SD	42.8 ± 14.1	43.0 ± 13.7
Smoking status, n (%)
Never	556 (51.6)	1,111 (51.7)
Former	341 (31.7)	680 (31.7)
Current	164 (15.2)	327 (15.2)
Missing	16 (1.5)	29 (1.4)
Alcohol consumption (units/wk), n (%)
0	233 (21.6)	462 (21.5)
1–3	575 (53.4)	1,150 (53.5)
4–7	165 (15.3)	328 (15.3)
8–14	36 (3.3)	72 (3.4)
15+	3 (0.3)	6 (0.3)
Missing	65 (6.0)	129 (6.0)

^aAge at DNA sample collection used for leukocyte telomere length measurement.

The final cohort of patients with CeD who had leukocyte telomere length measured comprised 58.7% women, with mean age of 42.8 (SD 14.1) years, where half (51.6%) were never smokers, and the majority consumed little or no alcohol (75.0% of the patients with CeD consumed 0 or 1–3 units per week).

The telomere length in patients with CeD did not differ significantly from matched controls 10.39 (SD 2.21) kb vs 10.52 (SD 2.25) kb, P = 0.088 (Figure 2 and Table 2).

Figure 2.

Leukocyte telomere length in 1,077 patients with celiac disease and 2,147 controls.

Table 2.

Leukocyte telomere length in patients with celiac disease and controls assessed by mixed linear model and estimates of fixed effect size

Parameter	Estimate	95% CI	P-value
Celiac disease	0.112	−0.017 to 0.241	0.088
Sex (male)	−0.036	−0.251 to 0.178	0.178
Age (yr)	−0.027	−0.035 to −0.020	<0.001
Tobacco smoking			0.799
Alcohol (units/wk)			0.025
0	Reference
1–3	−0.424	−0.684 to −0.164	0.001
4–7	−0.323	−0.670 to 0.024	0.068
8–14	−0.737	−1.34 to −0.135	0.016
15+	−0.920	−2.35 to 0.510	0.207
Missing	−0.438	−0.904 to 0.027	0.065

Among the 1,077 patients with CeD, 380 (35.3%) had known CeD, 465 (43.2%) had new CeD, and 232 (21.5%) had potential CeD. Telomere length in patients with known CeD was 10.48 (SD 2.26) kb, in patients with new CeD 10.44 (SD 2.14) kb and potential CeD 10.21 (SD 2.23) kb. In an analysis including age, sex, tobacco, and alcohol consumption, disease phenotype did not seem to affect telomere length (P = 0.77), while older age was still significantly associated with shorter telomere length (P < 0.001).

DISCUSSION

In this study of a population-based cohort of patients with CeD, no difference in telomere length was found compared with controls matched for factors previously reported to influence telomere length.

This result is contrary to several previous reports. Two previously published large studies aiming to assess the causal direction and magnitude of association between telomere length and CeD used data from genome-wide association studies and MR (3,11). These MR studies found an inverse association between genetically increased telomere length and CeD (odds ratio 0.42 [0.28–0.61] (3) and 0.71 [0.57–0.89] (11) per SD change), suggesting that CeD could be one of the most common diseases associated with shorter telomere length. The data from the 2 studies have also been used in a meta-analysis of 7,762 cases where a odds ratio of 0.62 (0.51–0.76 per SD change) (28). In the U.S. National Health and Nutrition Examination Survey population, it was found that the 25 patients with CeD identified by serology had shorter leukocyte telomere length than 3,914 controls (13), also when adjusted for age, sex, race/ethnicity, family income, serum ferritin, and folate. However, in another study of 20 patients with CeD, it was found that leukocyte telomere length did not differ from controls (12). Codd et al (11) also reported that shorter measured leukocyte telomere length was associated with CeD (n = 1,181) with hazard ratio 0.84 (0.77–0.92) per SD in an analyses adjusted for age and sex. Findings from MR studies and leukocyte telomere length analyses may differ because MR estimates reflect the effect of lifelong, genetically determined differences in telomere length, whereas telomere length measurements are typically assessed at a single time point during an individual's life. Furthermore, the effect size in MR studies may be too small to be detected in a phenotype study of 1,077 patients with CeD.

In this study, patients with CeD and controls were matched for age and sex, but also for tobacco smoking (29) and alcohol consumption (25,30), factors which previously have been reported to be associated with shorter telomere length. In our population, we found effects of alcohol consumption in a linear mixed model where increasing alcohol consumption was associated with shorter telomeres. This is in concordance with findings from an MR study (25), whereas other studies have found telomere shortening mainly in patients with alcohol abuse disorders (31). However, we did not find significant effects of tobacco smoking on telomere length in a linear mixed model analysis. Although many studies have found an association between tobacco smoking and shorter telomeres (23,24), studies with a longitudinal design suggest that tobacco consumption does not seem to accelerate telomeres shortening (32). Finally, we did find that increasing age was associated with shorter telomere length, which was expected from numerous other reports.

A considerable proportion of patients with CeD in our cohort were newly diagnosed after a positive serological screening and confirmed by biopsy (15), and we also performed a separate analysis of telomere length according to whether patients had previously known or newly diagnosed CeD because this would give information about the effect of being on a gluten-free diet. We did not find that patients with known CeD had telomere length differing from patients with newly diagnosed disease, suggesting that having had untreated CeD did not cause shorter leukocyte telomere length.

This study had several strengths including a population-based design, and the cohort with CeD had a verified diagnosis with high data completeness. We measured leukocyte telomere length and also matched controls for variables believed to influence telomere length. The study was limited by less systematic documentation of the CeD diagnosis in participants with a known diagnosis before HUNT4, and this included incomplete information regarding e.g., number of duodenal biopsies and the specific serological assays used. We did not measure telomere length for individual leukocyte subpopulations, and potential batch effects were not assessed during the pPCR analyses. However, triplets were positioned on the same plate to reduce effects of interassay variation. Finally, selection bias of subjects was possible because HUNT4 participants who accepted the invitation to clinical assessment may e.g., have had more symptoms than those who did not chose to participate.

In conclusion, this large population-based study provides evidence that CeD is not associated with a clinically meaningful telomere shortening.

CONFLICTS OF INTEREST

Guarantor of the article: Reidar Fossmark.

Specific author contributions: Conceptualization: R.F., H.H.H., E.N.J. Methodology: E.N.J., H.L.N., L.L. Software: Ø.S., T.L., M.V., L.L. Formal analysis: L.L., T.L., M.V., Ø.S., H.L.N., Y.L. Investigation: E.N.J., R.F., H.L.N. Data curation: L.L., T.L., M.V., R.F. Writing—original draft: R.F. Writing—review and editing: all. Project administration: R.F.

Financial support: Felles forskningsutvalg, St. Olav's Hospital/Norwegian University of Science and Technology (NTNU).

Potential competing interests: None to report.

Data availability statement: Data are available upon reasonable request. Researchers with a PhD associated with a Norwegian research institute can apply for the use of HUNT data and samples, given approval by a Regional Committee for Medical and Health Research Ethics. Researchers from other countries are welcome to apply in cooperation with a Norwegian principal investigator. Access to the requested HUNT material is given after the application is approved by HUNTs Data Access Committee (DAC) and an agreement is signed. The agreement gives the researcher(s) the right to study a specific topic for a limited time period and to publish a decided upon number of articles. For more information contact: HUNT Research Centre, Forskningsveien 2, 600 Levanger, Norway. Phone +47 74 07 51 80 | + 47 74 01 92 40 | + 47 407 90 010. E-mail: kontakt@hunt.ntnu.no.

Ethics statement: At participation, the HUNT participants gave a broad written consent to future research, including registry and hospital record linkage. The CeD in HUNT project, including genetic analyses, has been approved by the Reginal Committee for Medical and Health Research Ethics in Central Norway (reference number 7943).

Study Highlights.

WHAT IS KNOWN

• ✓ Genome-wide association studies have found celiac disease to be associated with shorter telomere length.

WHAT IS NEW HERE

• ✓ In a large population of patients with celiac disease and matched controls, leucocyte telomere length measured by qPCR did not differ significantly.

• ✓ In a mixed linear model, we found higher age and alcohol consumption, but not tobacco use or sex, to be associated with shorter telomere length.

ABBREVIATIONS:

CeD

Celiac disease

DNA

deoxyribonucleic acid

gDNA

genomic deoxyribonucleic acid

H pylori

Helicobacter pylori

HLA

human leucocyte antigen

HUNT

Trøndelag Health Study

MR

Mendelian randomization

PCR

polymerase chain reaction

qPCR

quantitative polymerase chain reaction

SD

standard deviation