Ulcerative colitis is a disease of accelerated colon aging: evidence from telomere attrition and DNA damage

INTRODUCTION

Telomeres protect the end of the chromosomes from end-to-end fusions, degradation, and recombination. Telomeres shorten approximately 100 base pairs in each cell division due to their incomplete replication ¹, and also as a consequence of oxidative damage ². Telomere attrition plays two opposing roles in cancer: as a tumor-suppressor mechanism when DNA damage checkpoints are intact ³, and as a mechanism to generate chromosomal instability when DNA damage checkpoints are dysfunctional ⁴. In addition, cross-sectional studies suggest that telomere shortening also plays an important role in aging: not only does telomere length decrease with age in most human tissues ⁵, but shorter leukocyte telomeres are associated with mortality, diseases of aging, and factors that predispose to aging ^{6, 7}. Although the risk of most cancers increases with age ⁸, it is still not known to what extent telomere shortening is responsible for this association. Support for the role of telomeres as a key factor linking cancer and aging comes from some premature aging syndromes, such as dyskeratosis congenita and aplastic anemia, characterized by accelerated rate of telomere attrition with age, chromosomal instability and predisposition to cancers (reviewed in ⁹).

Critically short, uncapped telomeres lead to replicative senescence by activation of a DNA damage response that involves phosphorylation of the histone H2AX (γH2AX) ¹⁰. This response is normally triggered by DNA double-strand breaks to arrest cell cycle progression until the DNA damage has been repaired, and it can be visualized by fluorescence microscopy as nuclear accumulation or foci of γH2AX and other DNA repair factors ¹¹. In the presence of uncapped telomeres, growth arrest is permanent and γH2AX foci are persistent and a marker of cellular senescence ¹². Although the DNA damage response that links short telomeres, H2AX phosphorylation, and senescence is increasingly well characterized in vitro ¹³, less is known of how telomere shortening translates into aging and disease in humans.

Ulcerative colitis (UC) is a chronic inflammatory disease of the colon that predisposes to colorectal cancer ¹⁴. The risk of colorectal cancer increases with duration, extent and early onset of disease ¹⁵, and with the severity of inflammation ¹⁶. We and others have previously demonstrated telomere shortening in colonocytes of UC patients, especially in those progressing to dysplasia and cancer ^{17, 18}. In addition, we have shown that telomere shortening is associated with chromosomal instability and anaphase bridges (a result of end-to-end chromosomal fusion), providing a mechanistic connection between telomere shortening and chromosomal damage leading to cancer ¹⁸. The mechanism that triggers UC is, however, unknown, but inflammation and oxidative damage are thought to play a crucial role ¹⁹. Because telomeres shorten with cell replication ¹ as well as oxidative damage ², we hypothesized that, in UC, telomeres might shorten faster than in normal colon. We have previously demonstrated that in normal colon telomere length decreases with age until approximately age 60. Beyond that age, the average telomere length increases, presumably reflecting the selective survival of patients with longer colonocyte telomeres ²⁰. Therefore, by comparing colonocyte telomere length and γH2AX damage-related foci in UC and normal individuals across a wide range of ages, we aimed to study the role of telomere attrition and DNA damage in this disease and as a potential link between cancer and aging. In addition, we analyzed telomere length in stroma adjacent to colonocytes and in leukocytes of UC patients to determine whether telomere attrition in UC patients is systemic or cell and tissue specific.

MATERIAL AND METHODS

RESULTS

DISCUSSION

We have demonstrated an accelerated shortening of colonocyte telomere length with age in UC patients compared to normal controls. On average, by age 33 years, UC patients have acquired very short colonocyte telomeres, comparable to that reached in normal individuals by age 60 years. Moreover, this rapid attrition occurs within the 8 first years of the disease, although patients with late-onset disease have short colonocytes telomeres irrespective of disease duration. These findings have a parallel in clinical practice: colon cancer risk becomes elevated in UC only after 8 years disease duration ²⁶, which marks the beginning of colorectal cancer screening and, interestingly, the acquisition of short colonocytes telomeres. Similarly, in the general population, colorectal cancer screening starts at age 50 years, which is concordant with the age at which short colonocyte telomere length is reached in normal individuals. Thus, rapid telomere shortening in UC colonocytes might reflect accelerated colon aging of these patients, predisposing them to colorectal cancer at an earlier age. Indeed, the average age of colorectal cancer incidence in UC patients (age 47) is more than two decades earlier than in the general population (age 70) ²⁷, and the lifetime risk is 19 times higher ²⁶.

It is well known that telomere length decreases with age in most human tissues, including colon ²⁰, and it has been hypothesized that short telomeres might partially explain the connection between cancer and aging ^{8, 9}. When telomeres become critically short, and in the absence of proficient DNA damage checkpoints, anaphase bridges are formed in mitosis as a consequence of end-to-end chromosome fusions, which initiates cycles of breakage-fusion-bridge. These cycles facilitate the accumulation of genetic changes and chromosomal instability. To emerge from this crisis cells need to acquire mechanisms of telomere maintenance, which usually consist on reactivation of telomerase. Telomerase prevents further accumulation of chromosomal instability and confers unlimited replicative potential, setting the stage for cancer progression ⁴. We previously demonstrated that, in UC, colorectal cancer progression is associated to shorter colonocyte telomeres, chromosomal instability and anaphase bridges ¹⁸. In this expanded cross-sectional study, we demonstrate that, in UC, age-related telomere shortening is accelerated. Our results also suggest that there is a minimal viable length of colonocyte telomeres, consistent with data from both human cell lines ²⁸ and telomerase deficient mice ²⁹. When this critical length is reached, UC colonocytes defective for DNA damage checkpoints can continue to proliferate, but with increased chromosomal instability. Since p53 and p16 are frequently altered in UC non-neoplastic epithelium ^{30, 31}, this could be a common pathway of tumor progression in this disease. Further research in needed, however, to characterize the mechanisms of activation and abrogation of senescence in UC, and to determine the timing of these events.

An important question that arises in the light of our results is whether accelerated telomere shortening in UC is due to an inherent telomere maintenance defect or is a consequence of the chronic inflammation characteristic of this disease. If the former was true, then we would expect shorter telomeres not only in colonic epithelium, but also in other tissues. Our data shows that shorter telomeres are observed only in the epithelium, but not in the adjacent stroma of UC colon biopsies. On the other hand, leukocyte telomere length is on average 6.6% shorter in UC than in normal controls. While this difference is significant, it is considerably smaller than the telomere reduction observed in syndromes with an inherent telomere maintenance defect, such as dyskeratosis congenita ³². The small shortening observed in leukocytes of UC patients is more similar to that previously reported in cardiovascular diseases, where affected patients have average telomere length reductions of 3–12% compared to normal controls ^33–35. These small differences require of a large number of cases to be detected, especially given the large variability of telomere length between individuals, which might explain why shortened leukocyte telomere length was not reported in a previous study that included 19 UC patients and 18 controls ³⁶. However, lymphocytes from UC patients were found to have higher frequency of telomeric associations and chromosome aberrations than normal controls ³⁷, which may be a consequence of the telomere attrition observed here. Multiple associations between short leukocyte telomere and diseases of aging have been reported ^{6, 7} and, although this could reflect a predisposition of individuals with constitutionally short telomeres to age-related diseases, recent results suggest that these associations are more likely due to the ability of leukocyte telomere lengths to reflect the history of inflammation and oxidative stress of the individual ^{23, 38}. We thus hypothesize that inflammation and proliferation are responsible for telomere shortening in UC: the largest shortening is observed in colonocytes, which are highly proliferative cells located in the areas of inflammation; second to this, leukocytes, which are also proliferative and frequently travel through the inflamed colon, experience moderate shortening; lastly, stroma cells, which have very low proliferative rates, seem to be protected from shortening.

Another way to study the role of telomere shortening in the colon of UC patients is to compare right and left biopsies. In UC, inflammation can be confined to the rectum (proctitis) or extend throughout the entire colon (pancolitis). We found comparable telomere shortening in both right and left colon epithelium, probably due to the fact that most of the patients analyzed had pancolitis. Interestingly, the only patient that had exclusively left disease showed longer telomeres in the right than in the left colon. Kinouchi et al. analyzed telomere length in rectal and cecal mucosa of 17 UC patients, 7 of which had left disease, and found significantly shorter telomeres in the rectal mucosa, especially in those patients with dysplasia ¹⁷. Regarding telomerase activity, Usselmann et al. reported a decrease along the colon of UC patients and a negative correlation with inflammation ³⁹. However, two other studies reported no such association ⁴⁰, but rather an increase in activity corresponding to closer proximity to dysplasia ⁴¹. In spite of the conflicting results, these studies suggest a possible link between the extent and severity of inflammation, the presence of dysplasia, and colonocyte telomere dynamics, that is worth investigating.

To further explore the effects of telomere shortening in UC, we also measured the intensity of γH2AX, which is an indication of the degree of phosphorylation of the histone H2AX, an event that occurs as a response to DNA double-strand breaks and critically short, uncapped telomeres. Phosphorylation of H2AX is known to be one of the triggers in the cascade of events that leads to senescence ¹⁰. We showed that the levels of γH2AX were higher in the colons of UC patients than in normal individuals, across all decades of age. This confirms previous reports of higher levels of DNA damage in UC colonocytes compared to normal controls, especially in UC cases with dysplasia and cancer ^{42, 43}. It is important to note that the normal individuals in this study include some diagnoses that involve colonic inflammation (e.g., diverticular disease or infection) and possible increased oxidative damage, so the γH2AX signal in truly healthy colon might be even lower. Surprisingly, in UC there was neither an age-related pattern, nor a correlation between γH2AX foci and telomere length, contrary to what was observed in normal colon. Although short telomeres in UC colon are likely to contribute to high levels of γH2AX signal, our data indicate that there must be other large sources of DNA damage. Double strand breaks caused by reactive oxygen and nitrogen species secondary to chronic inflammation are a very likely cause ¹⁹. While activation of DNA damage pathways that lead to senescence acts as a tumor suppressor mechanism, it is also a hallmark of premalignant tissues and a sign of potential cancer progression, as tumors can eventually bypass senescence and acquire proliferative ability in the presence of high levels of DNA damage ⁴⁴.

Secondary to our main goal, but highly relevant to the field of aging, we found that normal colonocytes not only shorten telomeres with age, but they also accumulate γH2AX foci, up to approximately age 67 years, after which the trend reverses. This two-segment model for age-dependent increase in DNA damage foci in normal colon is the inverse of the age-dependent changes in telomere length previously reported ²⁰, and there is a significant correlation between these variables: the shorter the telomeres, the higher the mean intensity of γH2AX foci. Older individuals (above 60 years old) tend to have longer telomeres and less γH2AX signal, which, as previously hypothesized, probably reflects the selective survival of individuals with constitutionally longer telomeres, better antioxidant defense, or less cumulative exposure to inflammation and oxidative stress in their lifetimes.

In summary, we have shown accelerated telomere shortening and accumulation of DNA damage in the colon of UC patients, which might be secondary to inflammation, increased cell turnover, and oxidative stress. In light of these results we postulate that UC could be considered a disease of premature aging of the colon, which might explain the high and early incidence of colorectal cancer in these patients. Further studies are needed to clarify the role of senescence in UC and to determine the specific contribution of colonocyte turnover and oxidative damage in the shortening of telomeres.