Journal of the National Cancer Institute, June 6, 2012

     Age is the single greatest risk factor for cancer.  About 60% of cancer diagnoses  occur in the 13%  of the population that is 65 years or older.  Why does aging lead to an increasing incidence of cancer?

     One view holds that it’s not aging but time that produces cancer.  Only with the passage of time, typically decades, can cells accumulate the DNA mutations and other alterations that cause the common epithelial cancers of the lung, breast, prostate and colon.  From this perspective, cancer is coincident with aging, but not caused by it.  But over the last fifteen years our concept of aging has evolved, from an accumulation of damage that “just happens,” to a consensus that aging is a regulated process that’s at least partially genetically determined.   And cancer is now commonly viewed as mechanistically linked to this aging process.  

     But there is no universally accepted view of how aging causes cancer.   One theory holds that a progressive loss of telomeres, the caps of noncoding DNA that protect the ends of human chromosomes, helps propel the aging process, and that genomic instability caused by such telomere dysfunction is a major driver of malignancy.  Another theory holds that an age-dependent deterioration of mitochondria, organelles that produce energy for the cell, causes both aging and cancer by generating reactive oxygen species (ROS) that damage DNA and proteins.  Recent mouse studies from the laboratory of Ron DePinho, M.D., until recently at the Dana-Farber Cancer Institute in Boston and now president of the M.D. Anderson Cancer Center in Houston, are raising the possibility that the two processes are linked in aging, and perhaps also in cancer.  “We provided a ‘unified field theory’ for aging,” said DePinho, “by providing a direct molecular link.”   Broad acceptance of DePinho’s theory hinges on definitive validation in humans, including the success of interventions designed to forestall aging and to prevent cancer.

Linking Short Telomeres to Cancer

     There is room for a new theory, because there is no consensus on what causes aging and what links it to cancer.  And much about telomeres remains mysterious.   Telomeres have been observed since the 1930s, but only in recent decades has their biology been worked out in any detail.  In 1978 Elizabeth Blackburn, Ph.D., then at Yale University in New Haven, Connecticut, studying the protozoan Tetrayhymena thermophilia, found that telomeres consisted of hundreds of base pairs of repeat sequences rich in guanine bases.  In 1985 Carol Greider, Ph.D., a graduate student in Blackburn’s lab at the University of California, Berkeley, discovered telomerase, the enzyme that synthesizes these telomeric repeats and thus maintains telomeres.  (Blackburn and Greider, together with Jack Szostak, Ph.D., of Harvard, received the 2009 Nobel Prize in physiology or medicine.)   Biochemist Cal Harley, Ph.D., then at McMaster University in Hamilton, Ontario, reported in 1990 that telomeres shorten as human fibroblasts divide in culture, explaining the “Hayflick limit,” the observation that cells stop growing after a fixed number of divisions—typically 60-70 doublings.    

     Telomeres and telomerase burst on the public imagination in 1998, when Science published a paper showing that simply adding the reverse transcriptase component of telomerase to normal cells could immortalize them.  The Science paper caused a furor; much of the lay press and public embraced telomerase as the secret of eternal life.   “They thought we were going to cure aging,” recalled Jerry Shay, Ph.D., of the University of Texas Southwestern in Dallas, one of the Science paper authors.  In fact, few biogenerontologists take telomere loss seriously as the main cause of aging at the organism level.  To begin with, many small mammals, including mice, have very long telomeres and plenty of telomerase, and only live a few years, whereas humans have short telomeres and scant telomerase and can live a century or more.  More importantly, most human cells never approach the Hayflick limit.  “Normally humans don’t undergo enough cell divisions to get towards the point at which our telomeres are getting either critically short or anywhere near short,” said Peter Hornsby, Ph.D., an aging researcher at the University of Texas Health Science Center in San Antonio.   And cells that do divide, Hornsby said, typically have sufficient telomerase.  On the other hand, human population studies have shown an association between short telomere length in blood cells and many diseases of aging.   “Telomere biology is going to explain maybe 10 or 15 percent of aging,” said Shay.  “It’s important, [but] there are many ways to develop an aging phenotype.”

     As in aging, the role of telomeres and telomerase in cancer is complex and still being worked out.   Telomerase can be detected in about 90% of human tumors, so until the year 2000 the enzyme was thought to be necessary for tumor development.  That year DePinho’s group reported that telomerase- and p53-deficient transgenic mice, instead of abolishing cancer (mice normally develop lymphomas and carcinomas) surprisingly developed human-like epithelial cancers.  These tumors displayed so-called “breakage-fusion-bridge” (BFB) cycles, in which the shortened telomeres of chromosomes fuse, are then pulled apart during cell division when chromosomes move to opposite poles of the cell, with the fused chromosomes forming bridges, until finally breaking.  Then the chromosomes rearrange to create the translocations that characterize many human tumors.  “That creates amplifications of oncogenes and deletions of tumor suppressor genes,” said DePinho. 

     In other words, by engineering mice without telomerase, DePinho created a model of what may happen in premalignant human cells that lose telomere function and then go through “telomere crisis.”  Unable to replicate, most senesce or die, but  a tiny few find a way to avoid keep dividing, and emerge as malignant tumor cells.  “It took Ron’s lab to figure out how to make a mouse model reflect what’s really going on in human cancer,” said Shay.

Linking Telomeres to Mitochondria

     The recent telomere-mitochondria connection emerged from work by DePinho’s group on aging.  Mice engineered to lack telomerase surprisingly have sick  hearts and livers, quiescent organs not likely to wear down their telomeres.   To understand how loss of telomerase damaged these organs, DePinho examined patterns of gene expression.  These pointed strongly to the repression of pathways related to mitochondria, including PGC-1α and PGC-1β, master regulators of mitochondrial biogenesis and function.  DePinho’s group worked out the following sequence of events: Telomere dysfunction activated p53, which—in addition to a major role in DNA repair—binds to the PGCs, repressing their activity and impairing mitochondria.  Impaired mitochondria pour out ROS, which can damage DNA, especially the G-rich sequences in telomeres, in a self-reinforcing cycle.  “You create a lot of ROS, which will create a feed-forward loop,” explained DePinho.  “That increases damage, increases p53, further represses PGC, further messes up mitochondria and oxidative defense.”   This model of aging for the first time linked telomeres to mitochondria—DePinho’s “unified field theory.”

            But why were telomeres failing in these nondividing cells?    “Telomere dysfunction is not simply the shortening of telomeres,” explained DePinho.  Mutations in telomeres that disrupt their function, or defects in the “shelterin” complex of proteins that binds them, could also be responsible, he speculated.

            Researchers studying the biology of aging have cautiously welcomed the paper, published last year in Nature.  By connecting telomeres to mitochondria and oxidative stress, DePinho is “linking two things that are very… central to the aging community,” said Toren Finkel, M.D., Ph.D., who studies aging at the National Heart, Lung and Blood Institute in Bethesda, Maryland.  Such unifying hypotheses are badly needed in the fractured aging field, Finkel added, but proof that telomere dysfunction leads to a cascade of mitochondrial pro-aging effects in humans is still lacking.   Finkel points out that mitochondrial deterioration over time can be explained without telomeres.  “To what degree in the [human] heart, the brain and the liver this is caused by telomeres eroding or just the sort of ravages of time I think is still an open question,” he said.

     The aging work set the stage for even newer research linking telomeres and mitochondria in cancer.   In research published in February in the journal Cell, DePinho’s group took a mouse model of lymphoma and engineered these mice so that telomerase could be turned off and on at will.  Consistent with the “telomerase off, tumors on” model he established in 2000, DePinho’s mice develop tumors, but feeble ones, since they lack the necessary telomerase for unchecked proliferation.  Turning telomerase back on led to highly aggressive tumors.  A second Cell paper from this group showed a similar tumor explosion, this one leading to bone metastases, in a mouse model of prostate cancer.  These aggressive tumors, in DePinho’s view, are  “compelling evidence” that telomerase does more than just permit tumors to grow–it actually drives that process.  Thanks to telomerase, such tumors “can acquire new events that can make them even more malignant,” he said.

     But the most novel finding came when DePinho’s researchers turned telomerase off again: Tumor growth slowed, then renewed, as tumors learned to lengthen their telomeres without telomerase, via a process known as “alternative lengthening of telomeres” (ALT).   In these tumors, PGC-1β was up, along with mitochondrial numbers.  “The cell… tries to make more of those mitochondria,” said DePinho.  “And therefore they upregulate PGC.  And you see massive upregulation of many, many antioxidant defense genes as well.”  In effect, these tumor cells are reversing the mitochondrial hallmarks of aging seen in telomerase-knockout mice.  They are becoming young again, even as they set out to kill their host.  DePinho’s lab is now investigating if mitochondrial loss and recovery takes place not just in tumor cells with ALT but in the more typical tumors that reactivate telomerase in order to grow and spread. 

Validating the Telomere Hypothesis

            Work is already underway to verify that telomere-related mitochondrial changes occur not just in mice, but in humans.  For example, Rosa Ana Risques, Ph.D., in the pathology department at the University of Washington medical school in Seattle, has been studying bowel tissue from patients with ulcerative colitis (UC).  About ten percent of UC cases eventually progress to colorectal cancer.   UC is a good model for understanding human colorectal cancer progression because most patients undergo annual colonoscopies, allowing researchers to track changes as cells progress through various stages of dysplasia to cancer.  In a poster presented at the annual meeting of the American Association for Cancer Research in April, Risques’s group reported that a marker of mitochondrial function went down more often in UC patients who later progressed to cancer, compared to those who did not progress, and that the mitochondria recovered function once cancer appeared.   PGC-1α also fell and then rose.  “Our data aligns with DePinho’s data,” said Risques.  “It matches, in theory… with these ‘turn off, turn on’ telomeres, and mitochondria at the same time.”  Risques is now checking to see if these patients’ telomeres follow the predicted pattern.

     Such validation is crucial.  “Mouse studies are fantastic for us to learn cause and effect,” Risques said.  “But then I think it’s necessary to move to human biopsies to see what is happening in the real tissue.”  Jerry Shay agreed.  “You’ve got to separate mouse models where you knock out a gene from what really goes on in human beings,” he said.  Researchers, including Risques and her Washington colleague Peter Rabinovitch. M.D., Ph.D., and especially Alan Meeker, Ph.D., at Johns Hopkins University in Baltimore, have over the course of the past decade studied telomeres in human tissue for evidence backing the telomere hypothesis of cancer.  That is, loss of telomeres leading to chromosome fusion and breakage, translocations and genomic instability, escape from telomere crisis, and finally telomerase reactivation and full-blown tumors.  Almost uniformly, these studies have shown that changes in human tissues examined at different time points are consistent with this sequence of events.

     “There must be a dozen papers that provide absolutely direct evidence that there is a fusion-bridge-breakage process at the time that telomeres are eroded in human cancers,” DePinho said.  Loss of telomere function may not cause cancer cells to initially transform, DePinho said, but it’s crucial for full malignancy.

     But the case for telomeres driving human cancer is not closed.   “The evidence is largely circumstantial, as it has been impossible to know what happens during actual tumor progression,” said Robert Weinberg, Ph.D., a cancer researcher at the Whitehead Institute in Cambridge, Massachusetts, by email.  A demonstration in real time, he said, would help:  “Extract cells from a human premalignant growth and demonstrate that such cells, upon culturing ex vivo, would go through the [breakage-fusion-bridge] cycles, crisis, and telomerase activation.”   Weinberg acknowledged that such an experiment would be difficult.  (Putting tumor cells in culture, Hornsby noted, leads to selection for cell subpopulations best adapted to culture conditions, and loss of others, distorting results.)

     DePinho argued that there is already sufficient evidence to validate the telomere hypothesis of cancer. “The genetic evidence, as well as the genomic evidence, showing that it is causal to the development of cancer is without question at this point.”

            DePinho stresses that telomeres are not the whole story in cancer and aging.  “Telomere dynamics are incredibly important,” he said.  “However, they’re not the only reason [for cancer].  The accumulation of alterations, changes in the epigenome, a variety of things are going to undoubtedly conspire.”  Complete acceptance of the telomere hypotheses of cancer and aging will probably not come until telomere-based interventions succeed.  These range from telomerase inhibitors now in clinical development for treating cancer (See JNCI 2010; 102, 520-521), to telomere-measuring diagnostic tests, to strategies for boosting telomerase to prevent cancer or even slow aging.  In the meantime, DePinho’s work plausibly explains how aging leads to cancer: telomere dysfunction is the common element.  “It may not be a driver of cancer neoplastic initiation,” he said.  “[But] I would say it’s the main driver of the benign to malignant transition in cancer.”

Dr. DePinho is a consultant to Metamark Genetics in Cambridge, Massachusetts.  Dr. Shay is a consultant to Life Link in Madrid, Spain.