Subscribe to eTOC

A Novel Sequencing Technology Helps Identify an Autoimmune Paraneoplastic Disorder Associated with Testicular Cancer

Article In Brief

With the help of a novel sequencing technique, researchers have identified an autoimmune paraneoplastic neurological disease associated with testicular cancer in men. The technique enables neurologists to diagnose the disease sooner, giving their patients a better prognosis.

Scientists at University of California, San Francisco (UCSF) and Mayo Clinic have used cutting edge technology to identify and describe a novel autoimmune paraneoplastic neurological disease associated with testicular cancer in men.

The disease is characterized by a subacute onset and progressive cerebellar ataxia, double vision, slurred speech, vertigo, and balance problems.

The findings from the study, published July 4 in The New England Journal of Medicine, could enable neurologists who see men with unexplained neurological symptoms to diagnose this new form of paraneoplastic encephalitis associated with testicular cancer.

The sooner men are properly diagnosed and treated with immune suppressing therapies, the better their prognosis, said study co-author Michael R. Wilson, MD, FAAN, the Debbie and Andy Rachleff distinguished professor of neurology and associate professor of neurology in the division of neuroimmunology and glial biology at UCSF.

It seems as if the immune system's reaction to the testicular cancer sets off a T-cell mediated attack on kelch-like protein 11 (KLHL11) that is found in the testes and in parts of the brain, particularly the cerebellum and brainstem.

This discovery was possible because of a single patient at UCSF whose unusual case landed in Dr. Wilson's office. The neurologist has been doing sequencing studies looking for infections that cause meningitis and encephalitis. What he and his colleagues have learned is that a percentage of people whose cerebrospinal fluid (CSF) samples are sent for sequencing do not have infections but rather have an autoimmune related issue.

That was the case with Glenn Sauber, now 40. At the time, in 2015, Dr. Wilson suspected he had an autoimmune disease but he had no clue about the possible trigger. There were high levels of immunoglobulin G (IgG) antibody in his CSF. Dr. Wilson prescribed an anti-inflammatory therapy for three weeks and some of the symptoms improved. Then, Sauber's health insurance changed, and he had to seek help outside of the UCSF system. But he agreed to enroll in research.

Dr. Wilson, UCSF biochemist Joseph DeRisi, PhD, and their colleagues set out to hunt for antibodies in the patient's CSF. They had been enhancing a technique called programmable phage display originally developed in the Elledge lab at Harvard (see “The Science Explained”), and a few months later they were able to identify an antibody that was made against the KLHL11 protein. It still was a puzzle since no one really knows what this protein does in the brain. But there was a major clue: the protein appears in the testes as well as in the brainstem and cerebellum.


Panel A shows immunoprecipitation from a human phage display (1010 plaque-forming units [PFU] per milliliter) in CSF or serum samples (asterisks) obtained from the patients and healthy controls. Panel B shows detection of immunoprecipitated KLHL11 by Western blotting with an anti-FLAG antibody. Panel C shows colocalization of the immunofluorescence signal from the IgG of Patient 11 (left) and anti-FLAG antibody (middle) in a cell-based assay with KLHL11–MYC–FLAG overexpression. The scale bar corresponds to 10 μm. Panels D and E show colocalization of the immunofluorescence signal from the KLHL1 IgG of Patient 11 (Panel D) and Patient 3 (Panel E) with a commercial antibody to KLHL11 on mouse brain tissue. The insets in Panels D and E show the degree of red and green channel overlap, or “yellow” signal at a higher magnification. (Areas represented in the insets are shown with dashed lines.) The scale bars correspond to 50 μm.

Sauber had a history of testicular cancer—and his neurological symptoms seemed to be emanating from the cerebellum and brainstem, which control movement, balance, and eye movements.

“We were looking at an immune response to the tumor that was having off-target effects in the brain,” said Dr. Wilson.

The neurologist called Sauber to tell him they found an antibody that led them to this novel protein target. But Sauber had deteriorated. Subsequent visits to more than half a dozen doctors led to the same diagnoses; the last doctor said Sauber had a brain tumor and wanted to irradiate his brain and start chemotherapy. Dr. Wilson received permission from the man's insurance company to treat him again. He found no brain tumor but did observe chronic damage from the inflammatory process. Dr. Wilson resumed the steroid treatment and added intravenous IgG and later rituximab. Again, the man improved.

The Hunt to Identify Other Patients

The UCSF group had just started collaborating with Sean J. Pittock, MD, the Marilyn A. Park and Moon S. Park, MD, director of the Center for Multiple Sclerosis and Autoimmune Neurology and director of Mayo's neuroimmunology laboratory.

Dr. Pittock and his colleagues were entrenched in the new field of autoimmune neurology and his group was receiving serum and CSF samples for neural autoantibody testing from all over the world. They were hunting for autoantibodies by applying human serum and CSF samples onto mouse brain tissue and looking for human antibodies that bind to brain antigens; these antibodies can potentially explain mysterious neurological symptoms. By 2018, they were logging in more than 150,000 samples a year. The Mayo team currently identifies two to three new neural autoantibodies annually.

About 18 years ago, the Mayo Clinic group identified a novel antibody that bound to mouse brain tissue in a pattern, sparsely visible, which reminded them of stars in the night sky. Over subsequent years they identified additional patients with the same pattern of staining, which they termed “sparkles.”

They noticed the pattern was only seen in men. When physicians were contacted for clinical information, the patients all had progressive cerebellar ataxia and double-vision, and many had a known history of testicular cancer. The specificity of the “sparkles pattern” was so strong that the description of the phenotype and the association with testicular cancer was shared with clinicians. Dr. Pittock and his colleagues recommended that any suspected patients should be evaluated for the presence of testicular cancer.

At the same time, Dr. Pittock was collaborating with Dr. Wilson and Dr. DeRisi on potential applications of programmable phage display for epitope mapping of neural antibodies and detection of neural antibodies targeting both intracellular and cell surface antigens.

Dr. Wilson and Dr. DeRisi shared the story of their patient with a similar phenotype to that of the Mayo Clinic cohort of men with testicular cancer and ataxia. At Dr. Wilson's suggestion, the UCSF patient's sample was applied to mouse tissue and the sparkles pattern was identified. “Immediately, we all recognized the potential of this finding,” said Dr. Pittock.

The Mayo Clinic team sent serum and CSF samples from 11 Mayo Clinic male patients to UCSF for analysis with the programmable phage display technique. All of the tests came back positive for KLHL11 antibody. They ran samples from 300 people with other types of cancers and all those tests were negative. Over the ensuing months, both teams used multiple antibody detection platforms (western blot, immunoprecipitation, cell-based assays, and monoclonal antibody colocalization experiments) to prove that the “sparkles” antibody targeted KLHL11. They have now identified 37 cases.

“This is a nice example of how two teams at different institutions can come together collaboratively, bringing their respective strengths to discover and define a novel syndrome,” said Dr. Pittock. “Untreated, this disease can have devastating consequences.”

The Mayo scientists have developed a high throughput assay which is currently being validated before it can be offered as a diagnostic test. This test will likely be added to Mayo Clinic Laboratory's “autoimmune encephalopathy” and “autoimmune movement disorders” evaluations that assist clinicians in the diagnostic work-up of patients with suspected paraneoplastic encephalitis. According to the NEJM paper, an epidemiologic study of this autoimmune neurologic syndrome in Olmsted County revealed that this newly described disorder may be a relatively common cause of paraneoplastic encephalitis in men: 2.8 in every 100,000.

Dr. DeRisi, Dr. Pittock, and Dr. Wilson have a patent pending for KLHL11 autoantibodies as a biomarker of paraneoplastic encephalitis associated with testicular cancer.

These types of cases have led to the growth of a new sub-specialty in autoimmune neurology. It turns out that different antibodies have predilections to certain tumors. Dr. DeRisi, who is also president of the Chan Zuckerberg Biohub, explained that the immune system does its job of trying to kill the tumor and sometimes the immune system acquires inappropriate targets, such as the brain in these testicular cancer patients. “It's like friendly fire,” he explained. “There are several examples in cancer. The tumors elicit a T-cell response and the T cells target the wrong cells (in the brainstem and cerebellum in this case) and kill them.”


Panel A shows a comparison of filamentous phage and T7 phage virions. Filamentous phages are single-stranded circular DNA viruses and have been engineered to display foreign amino acid sequences on coat protein III and coat protein VIII. T7 phages are distinguished by their icosahedral head (capsid) and linear double-stranded DNA genome. They have been engineered to display polypeptides on coat protein gp10B. Panel B shows phage-immunoprecipitation sequencing (PhIP-Seq),2 the method used in the study by Mandel-Brehm et al. 3 A T7 phage-display library was used to identify the cause of a paraneoplastic encephalitis, a condition that is presumed to be autoimmune. They created this library from DNA encoding the human proteome. The peptide-coding sequences were synthesized as pools of 49–amino acid oligonucleotides on DNA arrays (step 1), and the released oligonucleotide pools were amplified by polymerase chain reaction (PCR) (step 2). The DNA pools were cloned into DNA encoding coat protein gp10B of T7 phage to create the library (step 3), made up of phages, each expressing a single peptide. The library was selected by screening against samples from patients (serum, as well as cerebrospinal fluid that presumably contained IgG autoantibodies) (step 4). Phage-bound antibodies were captured or immunoprecipitated on agarose beads coated with a protein that binds IgG (step 5). The beads were washed, ridding the surfaces of the beads of unbound or loosely bound phages, and the bound phages were then eluted (step 6). Eluted phages were amplified in host Escherichia coli (step 7) and purified (step 8). The amplified phages were used in another round of selection (step 9). The DNA sequence of the output phage was then obtained, which allowed for the identification of the antigen (step 10).

Expert Commentary

“This is a classic story in the field of paraneoplastic neurologic diseases,” said Josep Dalmau, MD, PhD, research professor at the University of Barcelona and adjunct professor of neurology at the University of Pennsylvania. The KLHL11 antibody-associated syndrome belongs to the category of disorders associated with antibodies against intracellular proteins, which is different (almost the opposite in many ways) to the category of disorders mediated by antibodies against cell surface proteins (NMDAR and other diseases).

“This exciting technology reminded me of a ‘back to the future approach. The authors used a programmable phage display technique, instead of cDNA library screening, to identify the KLHL11 antigen. This new technique has many advantages but shares with cDNA library screening the limitation of not being optimal to identify cell surface proteins that need the native conformation of the antigen.

“The syndromes associated with KLHL11 antibodies are extremely rare (less than one patient per year identified at the neuroimmunology laboratory at the Mayo Clinic), yet the epidemiological study in Olmsted County shows a prevalence of 1.4 per 100,000 persons. This is more than twice the prevalence of anti-NMDAR encephalitis reported by the same investigators, which is puzzling and will need validation studies.”

“These types of diseases are becoming more recognized,” added Steven A. Vernino, MD, PhD, FAAN, the Dr. Bob and Jean Smith Foundation distinguished chair in neuromuscular disease research in the department of neurology and neurotherapeutics at UT Southwestern Medical Center in Dallas. Dr. Vernino has been involved in the field of autoimmune neurology and antibody discovery research for the past 20 years.

“By itself, the identification of a new autoantibody does not have huge clinical impact but the techniques used to find it using an antigen display is novel. The fact that this antibody was correlated with particular neurological symptoms and testicular cancer is very interesting.

“If I had a patient with this neurological presentation, I would look carefully for testicular cancer. The technology used in this way is novel and will probably be utilized in other cases of suspected autoimmune diseases. It may allow us in the future to test for antibodies against a large number of neurological antigens quickly and cost-effectively.”


Dr. Pittock disclosed these fees, paid directly to the Mayo Clinic: research grants, consulting. and travel expenses from Alexion Pharmaceuticals and research grants from the Autoimmune Encephalitis Alliance; consulting fees from Euroimmun; grants from Grifois USA LLC; the Guthy-Jackson Charitable Foundation; research grants; consulting fees, and travel expenses from Medimmune; and NIH grants. The Mayo Clinic holds patents for intellectual property— some pending—for various autoantibodies for autoimmune conditions. Dr. Wilson receives grants from F. Hoffmann-La Roche and also has a patent pending, through UCSF, for the KLH11 as neurological autoimmunity and paraneoplastic marker.

Link Up for More Information

• Mandel-Brehm C, Dubey D, Kryzer TJ, et al. Kelch-like protein 11 antibodies in seminoma-associated paraneoplastic encephalitis N Engl J Med 2019; 381(1):47–54.