When Cancer is Just the Beginning

rasping her hands together to stop the tremors, Mary Slattery crinkles her brow. Her neck muscles ripple. Her lips and tongue labor to propel her words across her dining room table. "Basically, I'm empty," she says.

Despite the effort, her words are slightly slurred. But her meaning is clear: Cancer surgery in 2005, at the age of 43, removed all of Slattery's reproductive organs. But that didn't slow her for long, and the New Jersey mother soon returned to shuttling her daughters to sports and Girl Scouts, playing tennis, skiing, gardening, bike riding, and volunteer work. That all abruptly ended 15 months later.

While vacationing in Hawaii, Slattery suddenly developed vertigo. Her vision went blurry. At times, she felt like she was sinking in quicksand. ("That is an interesting sensation," she says wryly.) Her oncologist's telephone guess was an inner ear infection from snorkeling. The symptoms quickly worsened. Her eyeballs began oscillating. She had trouble walking and reached out to steady herself, but a chair or wall was always closer than she thought. Her cognition was unaffected, but she struggled to speak. Her intonation flattened into a monotone. Eventually, her head began bobbing front to back. Her arms and hands trembled.

Diagnostic tests at Memorial Sloan-Kettering Cancer Center in New York were inconclusive, but she clearly had something worse than an ear infection. Someone called Robert Darnell, the head of the molecular neuro-oncology lab at Rockefeller University, just down the street. He did a lumbar puncture, finding Slattery's spinal fluid awash in an antibody to cdr2, a protein expressed by ovarian cancer cells.

Up to 60% of ovarian cancer patients have tumors that produce the cdr2 antigen, Darnell says.¹ But the antibody in Slattery's spinal fluid signified something more sinister: An immune response to her cancer, theoretically arrested more than a year earlier, was so ferocious that it had chewed through the blood-brain barrier and into her cerebellum. Specifically, it had gnawed on Purkinje neurons that, like ovarian tumor cells, also express cdr2. Slattery had paraneoplastic cerebellar degeneration.

PCD is one of about a dozen paraneoplastic neurologic disorders (PNDs).² Some, including PCD, are defined by an immune response to a malignancy that remains at a distance from the brain. PCD, associated with ovarian, breast, and small-cell lung cancer, is rare; accepted estimates put it at 1 in 10,000 cancer patients.

Significant improvement in its neurological condition is also rare. More often than not, the level of disability hits a plateau within just a few weeks. The menu of symptoms is long: vertigo; severe bilateral loss of coordination in the arms, legs, neck and trunk; a jerky gait; oscillating eyeballs; speech abnormalities from swiftly worsening weakness in facial and mouth muscles; sensory disturbances. Slattery's deterioration was typical. Within three months, she devolved from using a cane to a walker to a wheelchair. "I have to ask for everything," she says. Her trembling left hand reaches out but misses her blue sippy cup. "It's a degrading disease." She means PCD, not cancer. On the second attempt, she lifts the wobbling straw to her lips and swallows purposefully.

PCD patients generally do better against their cancers than non-PCD patients, but many still see their cancers return. They're devastated by brain damage, and all die of one or the other disease.

"They're desperately ill, and I have to tell them the truth," says Darnell, a Howard Hughes Medical Institute investigator and a chaired professor and senior physician at Rockefeller. "But I also tell them they're special because their bodies are doing something critically important." Specifically, doctors suspect PCD's defining immune response can keep a malignancy at bay for years, or in few cases kill it off altogether. Slattery may have been one of those few cases; a month before the onset of PCD symptoms, her doctors had confirmed that a previously detected, suspicious nodule in her lung had vanished. It's an idea that casts a silver lining around an otherwise dark situation. Slattery says she wants to believe her immune system was doing something right all along, that maybe it bought her some blissfully cancer-free years before everything went haywire.

A graduate student in Darnell's lab demonstrated last year that a PCD patient's immune system may indeed be fighting off the cancer. She aimed T cells with an engineered mouse receptor at cdr2-expressing tumor cells; the T cells killed the cancer. ³

The fascinating paradox of PCD, says Darnell, is that despite its rarity, figuring out how and why it works could open a back door for a broad range of fields such as tumor immunology, immune diseases in the brain, normal brain function, cancer biology and therapies, and provide answers, albeit indirectly, to some of those fields' big questions. "It's not always true that the most obvious path to a solution of the big problem is to directly head-on attack that problem." Research in PCD may even one day lead to one of biology's most sought-after targets: A new therapeutic cancer vaccine. If PND patients have an immune response to their cancers, could we harness it to treat the same cancer in other patients?

ernard Brouwer, a professor of neurology in the Netherlands, first identified PCD in 1919. He became one of a succession of physician-researchers who translated clinical experience into incremental lab discoveries, handing off tiny revelations about the nature of PNDs to others who then developed diagnostic tools and treatments.

In 1957, immunologists Lewis Thomas, a future Memorial Sloan-Kettering Institute president, and Frank Macfarlane Burnet, a future Noble laureate, separately conjectured tumor cells might behave like viruses. "It is by no means inconceivable that small accumulations of tumour cells may develop and because of their possession of new antigenic potentialities provoke an effective immunological reaction, with regression of the tumour and no clinical hint of its existence," Burnet wrote.⁴ But how do you test for an undetectable phenomenon?

Confirmation that neurons and tumor cells share an antigen came in 1987 from the lab of Jerome Posner, the first chairman of Memorial Sloan-Kettering's neurology department. He led the team that discovered at least half a dozen antibodies and their associated onconeuronal antigens, and cloned and sequenced the genes that code for some of those antigens. Asked how that all began, he shrugs. "You get involved because you see patients," he says.

Posner's group isolated and cloned cdr2, which reacted strongly against a PCD patient's immunoglobulin G. The sequencing revealed a repeating pattern of nucleotides that, as they wrote in the Proceedings of the National Academy of Science, was "unlike any previously described eukaryotic gene."⁵ Additional experiments demonstrated that cdr2 expression was restricted to normal brain tissue and to certain cancer cells. And that individual anti-Purkinje antibodies signaled the presence of specific cancers.

For patients, those discoveries meant faster PND diagnoses and tumor identification and treatment. For Posner, though, it wasn't enough. "I had always hoped we'd utilize these not only for diagnostic purposes but also for treating the paraneoplastic disorder," he says. "We didn't get that far, but Bob Darnell is working on it."

n a table in Darnell's office rests his black leather physician's bag, a worn reminder of his mission to carry what he learns from his patients into the lab, and vice versa. Darnell never wanted to be "an ivory tower scientist." That was how, as a young man, he viewed his father, James E. Darnell Jr., who worked on RNA biosynthesis and regulation, winning a National Medal of Science in 2002. The son wanted "to do something related to the world, to people and their diseases." In medical school, he looked for a translational specialty where cancer-causing genes meet brain disease. He found Posner in 1987.

While a neurology resident, Darnell saw a patient with cerebellar degeneration and ovarian cancer but none of the Yo antibodies that help define PCD. He characterized a new antibody⁶ and named it anti-Nb for his own index patient, a woman with 12 adopted children and a husband in prison. Her Western blot showed a clearly demarcated black band at 150 kiloDaltons; the Yo antibody sits at 52 kD. Anti-Nb was unique, and indicated the patient might have a different type of PND. Darnell saw a signpost for his specialty at the intersection of cancer and the brain. "All the pieces of the career puzzle," he says. "She got me hooked."

Darnell became the first among PND physician-researchers to start from a patient and clone a gene that coded for a protein under attack in that patient. Now about a dozen PNDs have been characterized and cloned, leading to reagents - the cloned antigens for cdr2 and Hu are among the best known - and assays that rule a disorder in or out.

Before this, tumor immunity was just an interesting idea. "But having the gene there in your hands, being able to make the protein or peptides that predicted proteins, allows you to do specific immunologic assays," Darnell says. "To ask, 'what is the nature of this immune response?' It's very 'black box.' No one has ever studied this in a human before."

In the mid-1990s, Darnell set up a lab at Rockefeller near his father's, and his team began synthesizing cdr2 peptides. A student, Matthew Albert, put those peptides on the cell surface to investigate whether a female patient with PCD had receptor-equipped T cells that could recognize the peptides. She did, and the result, Darnell recalls, was "an acute killing."

Recalling the Thomas-Burnet "immune surveillance" hypothesis, Darnell notes that, similar to a virus, tumor cells display protein fragments as a peptide, and immune cells "find" them by carrying peptide-specific receptors. Neurons are believed to do the same as tumor cells. But long-lasting viral immunity is regulated by dendritic cells. So how do the dendritic cells acquire the tumor antigen?

Albert pushed on, investigating what turns a T cell into a cancer killer. The dilemma: memory of viral immunity is handed off in the lymph nodes, but PCD antigens are found in the brain and tumor. Nonetheless, the T cells were somehow activated to the cancer, and Darnell says his team was finding memory T cells, depending on how long and how severely a patient had been ill.

In 1996, Darnell predicted apoptotic death in tumor cells - triggered by any one of a host of possible actions, one example being the p53 cancer-suppressor gene - would lead immature dendritic cells through phagocytosis to "eat" those dead tumor cells. The dendritic cells degrade the tumor antigens and incorporate the resulting peptides into the groove of MHC class II molecules. While in the lymph node, those peptides are presented to T cells as foreign invaders, activating memory T cells. So Albert and others on Darnell's team tried it in vitro using a PCD patient's immature dendritic cells. The end result was a massive activation of T cells.⁷

That was the full-throated answer to Thomas and Burnet's previously untestable hypothesis: Tumor cells in PCD patients do, indeed, behave like viruses - they spark an immune response that achieves lasting, measurable effect. Those discoveries led Darnell's team to embark on three small-scale trials for vaccines to treat prostate cancer (two are now completed) and one for brain cancer. It took years to develop the vaccines, and the results of the first prostate cancer trial were still being analyzed this summer. So far, Darnell says, the patients' prostate-specific antigen (PSA) curves were showing "a favorable effect."

hree years ago, Bianca Santomasso had a major revelation while tending bar at a conference on new tumor immunotherapies in Keystone, Colorado. An MD/PhD candidate in a program that links Cornell University, Rockefeller University, and Sloan-Kettering, Santomasso was recruited to pour drinks during a happy hour. She handed a glass to Mark Dudley, the head of the National Institutes of Health Center for Cancer Research's cell production facility, and began telling him about her work.

Santomasso had spent years as a biomedical fellow in Darnell's lab working up a PCD mouse model. She immunized and gave cdr2-expressing tumors to transgenic mice and identified the fragment of cdr2 that could be seen by the mouse's immune system. Her intention was to find a way to help PCD patients, she says. Although she briefly considered that a mouse model would have applications for immunotherapy in general, she worried about whether injecting a non-PND volunteer with a PND-derived vaccine could transmit the disorder, should the work someday extend to neurologically normal human subjects.

Santomasso went to Keystone looking for ways to supercharge her T cells and push them into mice brains, hoping an animal model of PCD-induced brain damage would answer some crucial questions about how PCD works in humans. While at the conference, she heard about isolating mouse T-cell receptors and transferring genetic material into human T cells to turn them into melanoma killers. During the cocktail hour, she mentioned to Dudley and his colleagues the challenge of getting mouse T cells to cross the blood-brain barrier. They assumed she would, of course, isolate the receptor.

"Why?" she asked.

The response: "You could cure breast and ovarian cancer." She described the potential risk of transmitting PCD, but it hadn't happened to the mice. Look for a way to transfer those T cells to humans, they advised.

She went back to the lab refocused on the idea that a PCD immune response, the cerebellar damage notwithstanding, might be doing something right - battling the cancer. Understanding this process could illuminate cancer immunotherapy in general. Santomasso cloned the A and B receptor genes from a cdr2-specific T cell on the first shot, guided by an NIH-developed protocol. Santomasso says she stepped back in amazement when, after inserting the genetic material from the mouse receptor into human T cells, she saw they recognized her cdr2 peptide. "I said something boring like 'Wow' and then ran around the lab telling people," she says.

About a week later, Santomasso turned the engineered T cells on cdr2-expressing tumor cells. She found herself staring at an assay image of two things: the release of interferon-gamma, an indication of how activated the T cells were, and dead tumor cells at the bottom of the dish. Her T cells had slaughtered the cancer. She jumped up and down in excitement. "That was a great moment," she recalls. "It told me this reagent could actually do something important." Darnell didn't hear the news until nearly a week later, at a lab meeting. He smiled, and then peppered her with questions.

After writing up her experiments for PNAS, Santomasso finished medical school, then picked up in the lab where she left off. Having shown that T-cell receptors could kill ovarian cell lines, Santomasso now wants to tackle cancer from a patient. And she wants to clone receptors from other epitopes she's already identified to produce reagents for immunotherapy. "The one [reagent] in the PNAS paper is a great start, but you may need more than one," she says.

Darnell's lab is now investigating how an immune response to a tumor, virus, or multiple sclerosis, crosses into the brain. Furthermore, they're exploring why an immune system that viewed a brain protein as part of the self perceives that same protein as a foreign invader when it's expressed by a tumor. And his team is studying the potential risk of transmitting PCD to neurologically normal vaccine volunteers. "All we want to do is to mimic what the bodies of PCD patients are already doing right: fighting the cancer," Darnell says. "One can imagine ways to build in safety catches."

But T cells aren't adept at crossing the blood-brain barrier. "It's not generally understood how they get in there," Darnell says. Although his lab has tried to achieve this in animal models, by July they'd succeeded in just one animal.

ary Slattery began an experimental treatment of two immune suppressor drugs within a month of the onset of PCD symptoms. But after three rounds, she showed no long-term response. She later tried a more standard high-dose intravenous immunoglobulin therapy. No response. In November 2007, she called it quits on any further PCD treatment. But she's still battling the cancer. In early July, she was on a fifth round of chemotherapy; the fourth had done little to shrink a tumor in her stomach. The 46-year-old is now one organ away from stage IV cancer.

"My perspective has changed," she says, slowly and deliberately. "Now I just want my daughters to know courage."

Yes I was going to ask that. It must be compounded by the plane taken in the scan. There are some puzzling differences that hint at far more than just cerebrellum damage. Indeed that is intact on both images but there is no spinal cord on the second image.

There is much more going on. I find it very hard to believe as suggested that it is only in cancer sufferers that the immune system recognises and attacks the virus - this normally happens. I am also surprised that it was in any way unexpected that DC's would be the key to recognising the cancer or that anti-viral mechanisms would be the key ones involved. This is how normal inflammation, infection and immune response is involved. If the cancerous tissues 'fly under the radar' so to speak DC's would be attracted to sites of damage by tiisue inflammation and then identify anything that appears to be wrong, provided that a number of factors are present. In most cancer the problem seems to be not lack of immune aggresion, but the fact that cancer cells dont go around giving tags of their abnormal status. Under normal tissue damage, immune cells will add to the inflammatory impetus themselves, so need to be switched off. In this womans unfortunate condition, the reason the cancer was probably put into remission seems to me more the case that the immune system does not have this throttle-off control, rather than the presence of an anti-cancer ability, but that the immune system, not necesarilly not normally involving this part of the immune system, identifies abnormal cells very early on and deals with them normally. In all those cases that we never hear of it is because the abnormal tissue was cleared at the earlier stages of metastatsis, abnormal tissue and lumps must often and routinely be cleared. DC would bring in other parts of the immune response only when tissue damage from the cancer attracted them in sufficient numbers and for sufficient time to bring about a more generalised attack and result in antibodies detectible in plasma or CSF. This requires certain thresholds to be breached. After this is succesful the immune system has some sort of refractory period so arrests the attack before it can get a chance to invade tissues with similar markers. Automimmune control is the order of the day. In those few who develop this particular disease, the reason it is so rare is, I would conclude from the data, either - that not because these few have such automimmunity, but because these few have a lack of immune regulation, and thus when an opportunity is presented to cross the blood brain barrier, it results in omgoing problems, or that perhaps a co-factor is present in these cases that increases ongoing immune aggresion, such as a viral infection, which may also cause invasion into the neuro-epithelium. Either way, it is more down to bad luck in breaching the BBB than by design that such attacks on the CNS occur.

Therefore I conclude that such anti-tumour activity is normal, but that it appears rare in cancer simply because in cancer it is that relatively uncommon disease that in the first place results from poor immune patrolling and auo-immune activity. The auto-immune system is always at the core of removing abnormal cells and virally infected cells, so it is not surprising that this incorrectly describef 'anti-viral' wing of the immune system would be used to remove any abnormally growing tissue that was damaged or damaging at the margins. Abnormal growth and form naturally results from viral infections, and DC's are needed to investigate directly what is different. They do not just use markers they recognise, they require other harmful changes to be present in the environment. They will go after, and not be down-regulated aftwards, those markers that are associated with specialist tissues as these are the most distinct, different, and in the brains case, we see why there is a barrier there in the many different cell types, and the need for a seperate immune system regulated directly by neurons, but again like DCs, whose capacity to identify targets depends on other damaging circumstances to the tissue and the immune cell. The damage attracts the immune-active cell, then the cell too is damaged, and a pro-inflammatory response is initiated and can sustain. The question to me is what normally downregulates this to prevent the widsespread and perpetuting recognition by accident of the bodies own markers, like myelin proteins. DC are little known to be involved iin switching off response as well as initiating targets to attack. Ironically, and potentialy here too, neurons also play major roles in this too. We need to know other things, like the ratio of TRegs as well. We also need to understand why the lower nervous system seems the principal target of attack in those photos. Another possible linkage is that glia on those nerves along with secreted compounds must normally help regulate immune responses by the wider system, as neurons will attract immunocompetant cells (unusual surface markers, meuro-endocrine secretion and accesibility by blood cells, and electrical fields) quite possibly by acting on local and central APC. An infection for example may act on the peripheral nervous system first, gaining access to the purjinke neurons in the cerebellum via this route, as it is only a short hop. It could travel directly up the nerves or in fluids. An infectious agent (purely by way of example) could be present to explain the pattern and sustained attack in the patient shown, and that something broader happening, by whatever stimulus, would seem to be the most consistent explanation with the changes actually shown in the scan. It appears to describe a much broader tissue change and possible immune activity. The apparently general heightened aggresion may be a viscious cycle of already inflammaing one central target, but I think that the core problem is lack of regulation in the immune system, and that can also come about from a consistent source of stimulation to it.

Just to confirm, we double-checked with Dr. Posner, and both brain images we posted are indeed of Mary Slattery, even though they look quite different.

Thanks,

Alison McCook
Deputy Editor

Perhaps the researchers should subject their mice to several rounds of radiation/chemotherapy prior to attempting to find out how immune cells cross the BBB.

The brain scans are at different levels in the skull making direct comparison impossible.

New immunotherapeutic anticancer strategies are being developed with immunomodulatory antibodies directed against T-cell costimulators and inhibitors such as CD137, OX40, CTLA-4, and PD-1.

This is a very interesting and potentially promising area of active preclinical and now clinical research--check it out.

I enjoyed reading this article today about Mary Slattery and her journey. The last thing the article stated was Mary's current status as far as chemo and other medical interventions.

I would like to offer a possibility to Mary which may help her to assist her body to heal. My website is www.amyfreundbodytalk.com and would like Mary to know that there are LOTS of people in her area who are offering BodyTalk sessions. There is a big BodyTalk Center in New York City, called, The BodyTalk Center! The website is www.thebodytalkcenter.com

I would be happy to speak with Mary, if she is interested.

hugs,
amy

I agree with the previous post.

These scans are not from the same person.

The two brains scans illustrating this story seem to me very different. I cannot believe these are from the same person. Not so much the difference in the cerebellum, but the bony structures of skull and nose do not resemble at all.