I still recall my disappointment one day in 2015. I received the tumor DNA sequencing results for one of my patients, whose cancer had spread from her lungs to her brain. I saw not a single genetic alteration that could point us to a targeted treatment approach. Yet something else caught my eye in the data: there were numerous changes in the structures and numbers of nearly every chromosome.
Instead of having two copies of each chromosome, as is typical for normal cells, the cancer cells had anywhere from one to as many as five or six copies; and sometimes there were parts of chromosomes on their own, in one or multiple copies. The phenomenon that produces such genomic chaos is known as chromosomal instability (CIN). Cancer researchers and clinicians have long appreciated CIN as a hallmark of advanced and metastatic cancers; and yet I did not expect to see...
Instead of having two copies of each chromosome, as is typical for normal cells, the cancer cells had anywhere from one to as many as five or six copies; and sometimes there were parts of chromosomes on their own, in one or multiple copies. The phenomenon that produces such genomic chaos is known as chromosomal instability (CIN). Cancer researchers and clinicians have long appreciated CIN as a hallmark of advanced and metastatic cancers; and yet I did not expect to see it present to this extent in my 59-year-old patient, given her recent cancer diagnosis and the fact that CIN is known to arise progressively over time.
To oncologists’ frustration, there are no good therapies that target cancers with CIN. Patients may simply receive multiple rounds of systemic chemotherapies and radiation. My patient was no different; the only option available to treat her metastases was radiation to the brain, which runs the risk of neurological toxicity, sometimes resulting in memory loss or other cognitive deficits, among other side effects.
This dilemma resonated with me given my background as a cell biologist. A decade earlier, my doctoral work had centered on trying to understand how chromosomes in normal cells are evenly distributed between the two daughter cells during mitosis. The process is so intricate and so frequent—in many tissues, cell division happens faithfully and on a daily basis—that evolution has devised multiple backup mechanisms to ensure it takes place without errors. On the rare occasion that errors do occur, cells with abnormal chromosome numbers are rapidly cleared. But cancer is different. Tumor cells are remarkably tolerant of chromosomal abnormalities, and such large-scale genomic changes often go hand in hand with disease progression. Yet, whether CIN plays an active role in promoting cancer progression and metastasis has remained unclear, as have the mechanisms by which such gross chromosomal abnormalities might exert these effects.
A few years ago, I set out to determine if CIN drives cancer progression or if it is simply a separate phenomenon that goes along for the ride. To do this, I teamed up with Lewis Cantley, who at the time was leading the Meyer Cancer Center at Weill Cornell Medicine and is now at the Dana-Farber Cancer Institute. We genetically manipulated various metastatic, chromosomally unstable cancer cells, reducing their level of CIN without affecting the other genetic abnormalities they carried. Remarkably, cancer cells that lost CIN also lost their ability to spread. And to our surprise, we found that CIN promotes cancer spread by generating persistent and smoldering inflammation. It is the body’s own immune response that ultimately enables these cells to break free from the site of the primary tumor and colonize distant organs.
Unlike cancer cells, normal cells do not tolerate errors in chromosome segregation.
This finding, which we published in Nature in 2018, suggested that the act of being unstable is itself critically important to cancer’s ability to evolve. It was also an untapped opportunity: could we target the processes that lead to cells with abnormal chromosomes in order to treat CIN-ful cancers? Could we halt metastasis by somehow stabilizing the genome, or by alleviating the chronic inflammation CIN causes? Could we reeducate the immune system to clear cells with abnormal chromosome numbers?
To address these critical questions, my laboratory at Memorial Sloan Kettering (MSK) Cancer Center has employed an interdisciplinary approach rooted in cell biology while combining single-cell genomics, mathematical modeling, and clinical sampling. We believe that through such an integrated approach we will understand how CIN alters cancer cell behavior and promotes adaptability to sustain cancer progression. Furthermore, we aim to uncover the cellular pathways that enable cancer cells to tolerate CIN and to target those pathways for therapeutic benefit.
In 2018, to complement our academic efforts to understand CIN’s role in cancer, Cantley and I, along with another colleague, Olivier Elemento, cofounded Volastra Therapeutics, where researchers are now working to develop CIN-targeting therapies for a range of cancer types. With these broad and collaborative efforts, we hope to expedite the development of new treatments for patients suffering from chromosomally unstable cancers.
An overlooked cancer hallmark
Ever since researchers sequenced the first cancer genomes in 2006, our understanding of the genetic alterations that promote cancer formation has steadily evolved. Along the way, scientists have developed therapies that target individual genetic drivers of tumor progression, with the underlying assumption that if these drivers are inhibited, the tumor can be halted in its tracks. This was the reason we sent the tumor DNA of my patient for genomic sequencing, and doing so has become a standard approach for many oncologists treating cancer patients at MSK. But when sequencing fails to turn up a genomic alteration that would direct us to a targeted treatment, the limitations of personalized oncology become clear: while successful for a few, it remains limited in its success for most patients with advanced malignancies.
Even if a targeted therapy is applicable, it may only be effective at first. That’s because cancer evolves, and often finds ways to evade the drugs we throw at it. One of the most powerful weapons that cancer cells have at their disposal is CIN, which involves randomly shuffling their chromosomes each time they divide. (See illustration on opposite page.) Errors in chromosome segregation propagate and multiply. This results in a population of cancer cells with a vast amount of heterogeneity in their chromosomal composition and copy number, a phenomenon known as aneuploidy. Indeed, high levels of CIN and aneuploidy are features of advanced tumors that have relapsed after multiple rounds of therapy and are not responding to a drug that inhibits a single mutated gene, even if the patient’s cancer had at one point been beaten back by the therapy.
While sequencing DNA purified from bulk tumor samples enables a much more detailed look at the genetic code, this tactic fails to provide a large-scale map of the overall structure of the genome.
Researchers have known for decades that aneuploidy is a characteristic feature of human cancer, but it wasn’t until 1997 that Christoph Lengauer and Bert Vogelstein of Johns Hopkins University School of Medicine first demonstrated the role of CIN in promoting cancer cell heterogeneity. Their work led to the immediate appreciation that CIN has the potential to actively fuel cancer evolution and progression by tuning the number of copies of each chromosome, and as a result, the copies of the genes encoded on these chromosomes. More-recent work done by Stephen Elledge at Harvard Medical School and colleagues revealed that human cancers indeed increase their fitness by maximizing the number of chromosomes that harbor oncogenes and minimizing the ones bearing tumor suppressor genes.
Despite its importance to human cancer, CIN has taken a backseat to genetic mutations in the laboratory, thanks to the methodological revolution brought about by the widespread adoption of next-gen sequencing. This development focused the attention of the cancer research community squarely on the contribution of individual genes in cancer. This led to important discoveries that expanded our knowledge of the role that many genes play during tumor initiation. But the approach overlooked the impact of large-scale chromosomal abnormalities and how they might affect gene function and cancer cell behavior. While sequencing DNA purified from bulk tumor samples enables a much more detailed look at the exact genetic code of the chromosomes, this tactic fails to provide a map of how alterations in DNA sequences fit in the overall structure of the genome, and it obfuscates cell-to-cell heterogeneity in chromosomal copy number.
Over the past decade, researchers have begun to assign more weight to large-scale chromosomal changes. A key observation came in 2010 from Robert Benezra and colleagues at MSK. They revealed that cancers that have acquired CIN no longer depend on the oncogene that gave rise to the cancer in the first place. Indeed, when the investigators promoted the formation of lung cancers that were driven by the oncogene KRAS in mice, they observed tumor regression when oncogene was withdrawn using genetic manipulations. However, this regression effect was abrogated when the tumors were further engineered to be chromosomally unstable. This work has important implications for our understanding of how cancers can acquire resistance to targeted therapies, which by definition aim to inhibit the oncogenic drivers, such as KRAS.
More recently, Charles Swanton of University College London and the Crick Institute and his group were able to robustly establish the importance of CIN in human cancer. In a 2017 study that followed lung cancer patients, the team demonstrated that CIN, rather than the number of individual mutations that a tumor harbors, was associated with reduced overall survival. The researchers went on to show that CIN likely plays critical roles in almost every facet of tumor biology, from metastasis to the ability of tumor cells to evade immune surveillance. This work has revealed that stepwise changes in chromosome copies each time cancer cells divide provide human tumors with the ability to evolve under various selective pressures.
Thanks to these and other studies illustrating the role of CIN in cancer, the divide separating the cancer genomics and cell biology fields has progressively eroded. While chromosomes carry the genetic code that can be deciphered using technologically sophisticated genomic approaches, their life cycle and segregation during
cell division is fundamentally a physical process that can be followed with relatively high resolution under the light microscope. For instance, chromosomes that undergo segregation errors during mitosis end up in small DNA-containing structures called micro-nuclei, separate from the primary nucleus. Micronuclei have long been appreciated as a unique feature that distinguishes cancer cells from their normal surrounding tissues. Multiple groups have shown that envelopes surrounding micronuclei often rupture, spilling chromosomes into the cytoplasm, where they are exposed to enzymes that can break down DNA as well as other proteins. This in turn leads to the shattering of chromosomes.
Following these widespread chromosomal breaks, some pieces are lost, while others are randomly patched together out of order or in the wrong orientation, leading to the birth of new, highly abnormal chromosomes. This process is known as chromothripsis, and researchers have recently identified it as a critical mechanism fueling cancer progression. In addition to stepwise changes in chromosome numbers, chromothripsis can lead to a jackpot for the cancer through wholesale rearrangements of entire chromosomes at once. In this way, chromothripsis can rapidly amplify oncogenes and dispose of tumor suppressor genes. We now also know that this process can position oncogenes next to highly active regions of the chromosome and promote the formation of extrachromosomal circular DNA, both of which have been found to promote rapid resistance to targeted therapies.
Despite a long-standing recognition of chromosomal chaos in cancer, our understanding of how chromothripsis takes place did not materialize until the field combined state-of-the-art genomics and cutting-edge microscopy techniques. In 2015, David Pellman and colleagues used a microscopy-based technique by which they captured individual cancer cells that showed evidence of chromosome segregation errors for subsequent genomic analysis. Using this approach, which they termed Look-seq, the researchers demonstrated that the complex rearrangement patterns often seen in human cancer genomes can arise within a single cell cycle. This was probably the case with my patient whose cancer contained large-scale chromosomal abnormalities despite the recency of her diagnosis. In these ways and more, it seems, CIN can promote progressive as well as rapid and punctuated evolution of the cancer genome.
Mechanisms of Micronuclei Formation
Numerous errors in chromosome segregation during cell division can lead to the formation of micronuclei, even if there isn’t actual mis-segregation of the chromosomes. These events are not mutually exclusive, nor are they independent, as each of these only serves to fuel chromosomal chaos.
© andrew swift, iso-form
When microtubules from each pole of a dividing cell attach to a single centromere, that chromosome lags behind the others and is often encapsulated in a micronucleus, even if it ends up in the intended cell.
© andrew swift, iso-form
If mis-segregation does occur, whether due to an errant microtubule attachment or another reason, the mis-segregated chromosome can similarly get encapsulated. If it doesn’t, the resulting aneuploid cell is at an increased risk of a lagging chromosome and micronucleus formation.
© andrew swift, iso-form
Shortened or broken telomeres can leave chromosomes vulnerable to fusion events that can lead to chromosomes with two centromeres, called dicentric chromosomes. When the cell divides, microtubules attach to both centromeres, often fracturing and separating the dicentric chromosomes into the daughter cells. These broken chromosomes can get sequestered in micronuclei immediately or after a subsequent cell division, due to impaired replication.
Instability and inflammation
Focusing on the role of CIN in cancer metastasis, we were surprised to learn how CIN drives cancer’s spread. In particular, we made the surprising observation that cancer cells with CIN displayed activation of pathways related to inflammation, causing these cells to produce and secrete many inflammatory molecules known to be involved in cancer metastasis. This was puzzling at first, given that these cancer cells were being cultured in the lab and had not yet been introduced into animals—and thus had not encountered any immune cells. So what was the source of this inflammation?
After spending many hours looking through the microscope, we observed not only that cells with CIN had a preponderance of micronuclei, but that those containing ruptured micro-nuclei harbored an immune-related enzyme called cGAS. Discovered by James Chen at University of Texas Southwestern in 2013, cGAS is a sensor of double-stranded DNA in the cytoplasm. We thus wondered if the rupture of micro-nuclei and the subsequent exposure of chromosomes to the cytoplasm might be interpreted by the cancer cells as a danger signal, in much the same way that cells might react to the DNA of an invading pathogen. Sure enough, we found that ruptured micronuclei were potent activators of cGAS and its partner protein STING, leading to innate immune activation. But unlike acute viral infection, which only lasts for a few days before it’s cleared, the cancer cell cytoplasm is continuously exposed to bursting micronuclei, leading to persistent pathway activation and chronic inflammation.
It has therefore become apparent to Cantley and myself, among others, that cancer cells must have coopted a protective immune pathway to their own advantage. While activation of innate immune signaling might play a protective role during early tumor development by preventing many cancers from arising in the first place, at some point tumor cells override these safeguards, develop tolerance to CIN-driven inflammation, and chronically leverage these pathways to drive tumor growth. The ability of cancer cells to sustain ongoing levels of inflammation is critical to their spread from one organ to another. Immune cells are some of the most mobile cell types in the body; within hours of sensing an infection or a wound, they can travel through the vasculature and migrate against elevated hydrostatic pressures present in inflamed tissues to reach the site of injury. This process, which is vital for organismal survival, is mimicked by cancer cells during metastatic progression and is enabled by ongoing CIN and the genomic abnormalities it produces.
The link between chronic inflammation and cancer is well established. In fact, all the cardinal signs of inflammation first described by the Roman encyclopedist Aulus Cornelius Celsus—blush, heat, pain, and swelling—apply to cancer, and clinicians through the ages have often referred to tumors as non-healing wounds due to their persistent and unremitting inflammation. What role inflammatory signaling plays in cancer progression is yet to be fully elucidated, but by linking intrinsic genomic abnormalities such as CIN with ongoing inflammation in cancer, we have shown that CIN not only drives genetic heterogeneity but also fuels cancer spread through mechanisms other than genetics inheritance.
How Bursting Micronuclei Promote Cancer
Micronuclei have fragile nuclear envelopes that often rupture, causing chromosomes to spill out into the cytoplasm. There, they encounter nucleases that pulverize the DNA into small fragments that can be lost, randomly linked, or looped into circles known as circular extrachromosomal DNA. This process, known as chromo-thripsis, produces complex rearrangements that can drive cancer.
At the same time, the presence of DNA in the cytoplasm triggers the cGAS-STING inflammatory pathway thought to have evolved as a form of immune defense against viral infection. The enzyme cGAS binds DNA from the ruptured micro-nucleus, catalyzing the formation of 2’3’-cyclic GMP-AMP (cGAMP), which subsequently activates STING and downstream inflammation. When chronically activated due to abundant micronuclei in cancer, this inflammation can drive tumor growth and metastasis.
© andrew swift, iso-form
Targeting chromosomal instability
Unlike cancer cells, normal cells do not tolerate errors in chromosome segregation. Work led by the late Angelika Amon at MIT has revealed that aneuploidy is associated with multiple cellular defects including metabolic and mitochondrial dysfunction, as well as cellular stress induced by protein misfolding. In fact, humans have evolved various mechanisms that ultimately lead to the clearance of aneuploid cells. Work done by Duane Compton and colleagues at the Geisel School of Medicine at Dartmouth revealed that normal cells rapidly activate p53, a master tumor suppressor, in response to chromosome segregation errors, thus halting future cell division and the propagation of aneuploid cells. These important safeguards are in place to preserve the genomic integrity of the organism, and they usually work. In cancer, however, they are often breached. Understanding how cancer cells cope with deleterious consequences of CIN could result in an important therapeutic opportunity.
Indeed, there has been a burgeoning interest in understanding and undermining cancer cells’ ability to tolerate CIN. Multiple academic groups have recently tackled this question using genetic screens, and some have identified genes and cellular processes whose loss is selectively lethal to cancer cells with high levels of CIN. One such target is a kinesin protein called Kif18a, which plays a role in chromosome movement during mitosis and was independently found by three groups to be required for cell division in cancer cells with CIN but not those without. Interestingly, mice lacking functional Kif18a are viable with minimal defects, suggesting that targeting this kinesin might indeed provide a valuable and safe therapeutic. A Phase 1 clinical trial sponsored by Amgen is now testing Kif18a inhibition in patients with advanced cancers.
Another therapeutic strategy that has been explored by a number of academic groups, including our own, is the inhibition of targets that enable cancer cells to cope with chronic inflammation. ENPP1, for example, was initially identified by Harvard Medical School’s Timothy Mitchison and Stanford University’s Lingyin Li (then at Harvard), and found by my group to be selectively upregulated in chromosomally unstable cancer cells. ENPP1 is an enzyme located at the outer surface of cancer cells; it degrades an immune stimulatory metabolite called cGAMP that is produced when cGAS encounters DNA in the cytoplasm of tumor cells. If left intact, cGAMP can spread between cells and initiate a robust antitumor immune response. By degrading this molecule in the extracellular space, cancer cells prevent immune cells from sensing the inflammation that arises from CIN, thus providing them with much-needed cover. Not only does this shield tumor cells from attack by the immune system, but the breakdown of cGAMP ultimately generates another metabolite, called adenosine, that promotes immune cell dysfunction and enhances the ability of cancer cells to migrate. This was an eye-opening example of how cancer cells can turn a foe (cGAMP) into a friend (adenosine), thus bending inflammation to their own benefit.
To augment these and other approaches, researchers at Volastra, where I continue to serve as a scientific advisor, are pursuing a deep and fundamental understanding of the biology of CIN. Out of this research, which involves both computational and genetic screens that can inactivate individual genes one at a time, a number of therapeutic strategies have emerged. The company’s lead drug candidate targets microtubule attachments to chromosomes and is selectively lethal to cancer cells with CIN while sparing normal cells; Volastra plans to advance this one to clinical trials in 2023. Additional treatment strategies we’re exploring include modulating mitotic spindle formation, altering chromosome organization during cell division, and harnessing CIN-driven inflammation.
Identifying targetable pathways linked to chromosomal instability is exciting, as it creates an opportunity to therapeutically intervene to tackle a feature of cancer that was hitherto considered undruggable. CIN is a particularly attractive drug target because it is only present in cancer cells, meaning CIN-targeting therapies should selectively kill cancer cells while sparing normal cells—the holy grail of cancer therapy. The progressive rapprochement among the fields of cell biology, genomics, and cancer biology over the past decade will fuel novel discoveries that can only be made through a multidisciplinary approach as well as robust collaboration between academia and industry. The ultimate goal is to serve patients, such as mine who developed brain metastasis, for whom there are currently very limited therapeutic options.