May 2012
Feature Story

William Check, PhD

“The fact that a small cancer hidden in the chest, abdomen, or pelvis could destroy or damage portions of the nervous system, such as cerebellar Purkinje cells or cholinergic synapses, has intrigued neurologists since paraneoplastic syndromes were first described.” So wrote two neuro-oncologists in 1999 (Dalmau JO, Posner JB. Arch Neurol. 1999;56:405–408), and so it is today. Despite almost a half century of clinical and research investigations since the first description of what we now call paraneoplastic neuronal autoantibodies (Wilkinson PC, Zeromski J. Brain. 1965;88:529–583), major questions remain about this curious disease entity. At the same time, important new discoveries are being made that have expanded its scope and raised awareness among clinicians and laboratorians.

A sign of continuing interest in paraneoplastic syndromes was the invited lecture at the 2011 meeting of the Association of Medical Laboratory Immunologists, presented by John E. Greenlee, MD, professor and vice chair of the Department of Neurology at the University of Utah Health Science Center and the Salt Lake City VA Medical Center. Even though it is now known that the combination of certain neurological syndromes plus neuronal autoantibodies strongly suggests the presence of a tumor, Dr. Greenlee said in his lecture, it is often difficult or impossible to find the neoplasm. “Some patients [with paraneoplastic syndromes] can have high titers [of neuronal autoantibodies] and the tumor will be impossible to find,” he told CAP TODAY. When a tumor is diagnosed, it can be up to five years after the paraneoplastic syndrome initially presented (Mathew RM, et al. J Neurol Sci. 2006;250:153–155).

Among the prevailing questions are what determines production of neuronal autoantibodies by a tumor, and how can autoantibody production be stopped. The latter question is particularly important when the tumor cannot be found and removed. Also debated is the etiologic role of antibodies, as opposed to cytotoxic T cells.

In his lecture Dr. Greenlee also addressed the role of antibody testing in the diagnosis of paraneoplastic syndromes, saying, “The role of the clinical laboratory is absolutely critical.” Laboratory testing can delineate neuronal autoantibodies and help monitor response to therapy by following quantitative serum or CSF antibody titers. Because of the large number of neuronal autoantibodies—“We are now up to maybe 40 and still counting,” Dr. Greenlee said—the laboratory must offer a screening panel.

One service that has been in the forefront of both clinical screening and discovery of new neuronal autoantibodies is the Neuroimmunology Laboratory at the Mayo Clinic, directed by Vanda A. Lennon, MD, PhD, professor of laboratory medicine and pathology, who trained in neurobiology and immunology. “We test 500 patients per day for autoantibodies to neuronal antigens,” Dr. Lennon said in an interview. Consenting patients positive for neuronal autoantibodies are followed long term, and serum is stored in a serum bank at Mayo Clinic. Clinical-serological correlative information enhances interpretive support to clinicians. On a research basis, Dr. Lennon says, “With patients’ consent, we can follow them for many years, even sometimes to autopsy. We have gleaned correlative data between what we find in the antibody profile and what occurs in patients.”

Such data are helpful in managing patients, says Sean J. Pittock, MD, professor of neurology at Mayo and co-director of the Neuroimmunology Laboratory. “The profile of neuronal autoantibodies can guide clinicians in how aggressive they want to be in looking for cancer.” For example, with antibody to an antigen called PCA-1 (Purkinje cell cytoplasmic antigen 1, also called anti-Yo), “there is greater than a 90 percent probability that a patient has a malignancy of ovary or breast,” Dr. Pittock says.

Having a practicing neurologist like Dr. Pittock integral to the Neuroimmunology Laboratory contributes to new findings, Dr. Lennon notes, pointing to recent reports of autoimmune epilepsy and dementia from Dr. Pittock and their colleagues. “What’s being recognized,” Dr. Pittock says, “is that among patients with progressive neurological disorders thought to be untreatable, a subset have reversible, sometimes paraneoplastic autoimmune conditions. Markers we see in the lab can help advise clinicians looking for a specific cancer as well as provide an indication that a patient may have a disease that is treatable and reversible.” He calls this finding “very exciting,” and adds, “We are just at the tip of the iceberg. There is rapid growth in the number of neuronal autoantibodies being found.”

Parallel with growth in the number of neuronal autoantibodies has come heightened awareness of paraneoplastic syndromes. “The whole of the United States and North America has become much more aware of paraneoplastic syndromes in the last three years,” Dr. Lennon says. Adds Dr. Pittock, “The concept of autoimmune neurology becoming a standalone discipline is gaining momentum.” For the second year Mayo Clinic faculty will give a course on autoimmune neurology at the meeting of the American Academy of Neurology. And for the first time, the American Neurological Association will have a national special interest group session in autoimmune neurology.

Dr. Greenlee’s interest in paraneoplastic syndromes was sparked in the early 1970s, during his last year of residency, by a patient with ovarian cancer and classic cerebellar degeneration—a severe cerebellar deficit characterized by rapid or subacute development of extreme ataxia, dysarthria, and nystagmus. A few years later antibody to cerebellar Purkinje cells was found in a woman with cerebellar degeneration and Hodgkin lymphoma, in whom radiation therapy for the lymphoma halted progression of neurological symptoms (Trotter JL, et al. Arch Neurol. 1976;33:660–661). Subsequently, Dr. Greenlee found antibody to Purkinje cells in his initial patient, and he had a second ovarian cancer patient who also had cerebellar degeneration (Greenlee JE, Brashear HR. Ann Neurol. 1983;14:609–613). A few years later he published the finding of anticerebellar antibodies in a patient with lung cancer and paraneoplastic cerebellar degeneration, or PCD (Greenlee JE, Lipton HL. Ann Neurol. 1986;19:82–85). In this case, the authors reported, “[I]ntrathecal antibody synthesis was suggested by serum CSF antibody ratios, CSF IgG index, and CSF IgG synthesis rate.”

We now know that the antibodies in patients with PCD and breast or ovarian cancers are directed against intracellular cytoplasmic components of Purkinje cells (anti-Yo), while those in patients with another form of paraneoplastic syndrome, encephalomyelitis associated with lung cancer, are directed against nuclear and cytoplasmic antigens of all neurons (anti-Hu). Many other antineuronal immunoglobulins associated with paraneoplastic syndromes, such as anti-Ri and anti-Ma, are also known.

Antibodies against neuronal cell surface receptors are only sometimes paraneoplastic. In contrast, Dr. Greenlee says, “Antineuronal antibodies to intracellular proteins are not commonly found in the absence of neoplasia.” He reported on a patient who presented with PCD and anti-Yo antibodies but no evidence of tumor. Two years after PCD onset an infiltrating mammary carcinoma was diagnosed (Greenlee JE, et al. West J Med. 1992;156:199–202). Dr. Greenlee and his co-authors concluded that this case showed “the need for vigorous, prolonged follow-up in antibody-positive patients in whom the search for an occult neoplasm is initially unsuccessful.”

A greater difference between auto-antibodies against cell surface proteins and those versus intracellular antigens is that the etiologic role of the former is accepted. “We don’t know whether antibodies against intracellular proteins cause disease,” Dr. Greenlee says. “It’s widely held that antibodies can’t get into cells so they don’t cause disease; they are just markers.” He did an in vitro study with rat brain tissue slices attempting to give credence to the notion that antibodies against intracellular proteins can cause cerebellar degeneration. In this experiment, living Purkinje cells took up anti-Yo antibodies, which killed the cells (Greenlee JE, et al. J Neuropathol Exp Neurol. 2010;69:997–1007). Dr. Greenlee believes that the much lower ratio of antibody in serum to that in CSF in patients with cancer (10 to 20) relative to the serum/CSF ratio in healthy people (200) suggests synthesis within the brain. Still, he acknowledges that his hypothesis is controversial.

When it comes to treatment of paraneoplastic syndromes associated with antibodies to intracellular proteins, Dr. Greenlee says, “There is no proven effective therapy.” (Paraneoplastic syndromes caused by antibodies to cell surface proteins are often improved by immunosuppressive therapy.) Two years ago he wrote a review article on this topic (Greenlee JE. Curr Treat Options Neurol. 2010;12:212–230). “The therapy section was very hard to write,” he says. “There are a few moderate-sized studies—around 20 patients—then a whole bunch of case reports.” One trial that he calls “probably the best of the lot” treated 20 patients with early disease due to anti-Hu or anti-Yo with plasma exchange plus either antitumor therapy or cyclophosphamide. After six months half of the patients improved (Vernino S, et al. Neuro Oncol. 2004;6:55–62). “Aggressive immunosuppression early in the clinical course should be considered in patients who have paraneoplastic neurological disorders, even when there is no evidence of active malignancy,” the authors concluded.

Rituximab showed modest benefit in a small trial of patients with anti-Hu or anti-Yo antibodies (Shams’ili S, et al. J Neurol. 2006;253:16–20).

Because paraneoplastic syndromes are rare, a convincing trial of therapy will require a consortium, Dr. Greenlee says. “To date, though, no group of laboratories has been able to work together to carry out such a study,” he adds.

Rarity of paraneoplastic syndromes also dictates that only reference laboratories can offer testing for neuronal autoantibodies, says Thomas R. Haven, PhD, research and development scientist II at the ARUP Institute for Clinical and Experimental Pathology, who works with Dr. Greenlee in developing and validating tests for neuronal autoantibodies. Even a large hospital would see only a small number of these patients. “The population you are testing has a low positive yield,” Dr. Haven says. “For instance, we have had maybe a 10 percent yield for antibody to NMDA [N-methyl-D-aspartate] since we came on line with that assay.”

Setting up assays for onconeural antibodies can be a challenge. For some there is no commercial vendor, Dr. Haven says. Others are so rare it is difficult to get enough positive samples to validate the assay. For those assays that ARUP does offer, clinicians can order standalone or panel testing. Shorter turnaround time is an advantage of the panel. “If you order them all together, you get much faster results than if you order each sequentially,” Dr. Haven says. “You have to balance cost-efficiency with rapidity of diagnosis or prognosis.”

Mayo’s Neuroimmunology Laboratory offers a comprehensive screening evaluation for autoantibodies to both surface and intracellular targets (www.mayomedicallaboratories.com/interpretive-guide/index.html?alpha=P&unit_code=83380). Immunofluorescence assays are done on a composite of sections of normal tissue—brain, gut, and kidney—to distinguish neuron-restricted or glial-restricted antibodies from nonspecific antibodies. Western blot confirmation is done for intracellular antibodies. “Highly sensitive criteria allow technologists to distinguish one antibody from another,” Dr. Lennon says. “Two technologists read every slide, and a trained neurologist confirms their interpretation.”

What makes their practice unique, she says, is that they don’t allow clinicians to order a single autoantibody: “Cancer-derived autoimmune processes lead to many antibodies, and you can’t guess which one you are looking for.”

In addition to providing clinical support, the laboratory has identified several novel neuronal autoantibodies and autoimmune neurologic disorders, including some associated with neoplasia. For example, patients with neuromyelitis optica (NMO), an inflammatory demyelinating disease sometimes misdiagnosed as multiple sclerosis, have been found to produce a serum antibody to aquaporin-4, an astrocytic water channel protein (Lennon VA, et al. Lancet. 2004;17:2106–2112). Aquaporin-4 is sometimes found in a paraneoplastic context. Patients with NMO do much worse if given an immunomodulatory treatment commonly used for multiple sclerosis.

Dr. Pittock has led investigations of autoimmune subtypes of dementia (Flanagan EP, et al. Mayo Clin Proc. 2010;85:881–897) and epilepsy (Quek AM, et al. Arch Neurol. 2012;March 26 [Epub ahead of print]). If epilepsy is controlled on standard medication(s), it doesn’t need further workup, Dr. Pittock says. “However,” he adds, “in a patient with intractable epilepsy not controlled on two anti-epileptic drugs, and with no obvious cause, such as a tumor, it is worth investigating an autoimmune etiology. If an autoimmune etiology is found, the patient has a high likelihood of having a beneficial response to immunotherapy.”

For paraneoplastic syndromes associated with antibodies to intracellular targets, Dr. Lennon comes down strongly on the side of cytotoxic T cell-mediated immunity as the cause. “Antibodies to intracellular antigens are markers of neurologic disorders, not effectors of neurologic syndromes,” she says. “Antibodies directed to intracellular nuclear or cytoplasmic antigens are helpful in establishing a diagnosis of an autoimmune neurologic disorder, but they are markers for T cells being generated against peptides derived from the same proteins.”

Dr. Lennon cites several pieces of evidence to support her assertion. For instance, in a 2005 article on paraneoplastic accompaniments of autoimmunity to amphiphysin, one patient who could not walk had paraneoplastic spinal cord myelopathy at autopsy (Pittock SJ, et al. Ann Neurol. 2005;58:96–107). In this patient, she says, “Our pathologist found CD8+ T cells invading anterior horn cells. They were found in the act of causing myelopathy.”

In a report on patients with paraneoplastic autoimmune optic neuritis, one patient was thought from imaging studies to have multiple sclerosis. On the basis of a positive result for CRMP-5 autoantibody, she says, “we predicted cancer. At autopsy the pathologist found a tiny malignant lymph node in the mediastinum.” The pathologist harvested the optic nerve and spinal cord. “You can see cytotoxic T cells in sections of the optic nerve,” she says (Cross SA, et al. Ann Neurol. 2003;54:38–50).

Immunopathology of patients with paraneoplastic syndromes is complex, Dr. Pittock says. They commonly have antibodies to both surface antigens and intracellular targets. “Probably there are both T cell-mediated and antibody-mediated processes,” he says. “Generally, patients who have T cell-mediated paraneoplastic syndromes are not very responsive to immunotherapy, while those with antibody targeted to cell membrane antigens potentially have reversible neurologic disease.” In terms of the immunopathology of these disorders, he says, there is probably a lot of overlap.

To complement the Neuroimmunology Laboratory’s clinical and investigative work on paraneoplastic syndromes, Mayo Clinic has established an autoimmune neurology fellowship program, for which Drs. Lennon and Pittock are director and co-director, respectively. When this program was set up in 2006, it was the first of its kind in the United States, Dr. Pittock says. “In 10 years departments of neurology throughout the U.S. will have autoimmune neurology clinics staffed by neurologists with specialized training,” he predicts. Andrew McKeon, MB, BCh, who did a fellowship in the program, is now an autoimmune neurologist at Mayo Clinic with a specialization in movement disorders.

Dr. Pittock emphasizes the value of having a neurologist working in the Neuroimmunology Laboratory. “Because of the complexity of these profiles it is important to have advice from a clinician, somebody familiar with neurologic diseases,” he says. Every specimen that tests positive is interpreted by a neurologist, and a report is tailored for that patient. “We provide practicing neurologists with an estimate of the likelihood of malignancy, plus what type of malignancy and what type of neurologic disorders one might see with that profile. That’s where the field is moving—to disease-specific profiles.” At this time the laboratory produces two disease-specific reports: an autoimmune dysautonomia evaluation and an autoimmune gastrointestinal evaluation. In the next six months it will introduce three additional evaluations, Dr. Pittock says—for autoimmune epilepsy, dementia, and encephalopathy.

Dr. Greenlee, too, foresees rapid expansion for the field of paraneoplastic syndromes. “Additional antineuronal antibodies will unquestionably be discovered,” he says. He sees also a continued elucidation of autoimmune subsets of standard conditions. He points, as an example, to the California Encephalitis Project: “They just reported that a surprising number of patients with encephalitis had no infection but did have autoimmune disease,” especially antibodies to NMDA (Gable MS, et al. Clin Infect Dis. 2012;54:899–904).

Dr. Greenlee calls all of this “exciting to watch.” His own early work and that of other pioneers in this field “opened a door into an area that people hadn’t looked at,” he says, adding, “It’s going to grow. The immune system probably has a lot more to do with neurological disease than we thought.”