By Martin Ingelsson, co-organizer of the meeting and one of the speakers.
This is Part 1 of a two-part series. See Part 2.
16 July 2009. “Neurodegenerative Disorders—Immunotherapy and Biomarkers,” an international conference organized by Lars Lannfelt, Martin Ingelsson, and Lars Nilsson from Uppsala University, Sweden, was held 28-29 May 2009. During the two-day symposium, scientists came together to share new data and discuss important developments in this emerging research field. In addition to immunotherapy, the meeting focused on recent progress in the development of novel disease biomarkers, an important aspect of successful implementation of immunotherapeutic strategies.
Peter Seubert from Elan Pharmaceuticals, San Francisco, California, was the logical first speaker. Elan pioneered immunotherapy against Alzheimer disease in 2000 with a clinical trial of AN-1792, a synthetic Aβ antigen preparation. Currently, Elan has, together with Wyeth, entered a large Phase 2I study on Bapineuzumab, an Aβ antibody which has recently shown promising effects in a Phase 2B study (see ARF related news story). Moreover, Elan is pursuing other active immunization protocols, one of which is now in Phase 2 (see ARF related news story). Seubert did not report any new data from the ongoing trials, but instead presented interesting preclinical observations. For example, PDAPP mice treated with 3D6, an antibody against the N-terminal portion of Aβ, demonstrated markedly reduced pathology. Treated 12-18-month-old mice, i.e., well beyond the onset of plaque build-up, displayed up to 90 percent plaque clearance, reduction of neuritic dystrophy, and an increase in synaptic integrity. Moreover, behavioral treatment effects could be observed as treated mice outperformed non-treated mice in a contextual fear conditioning (CFC) setup. These results, Seubert speculated, demonstrate the efficacy of antibodies against the N-terminus of the Aβ molecule. Interestingly, he also pointed out that the antibodies raised among responders in the AN-1792 study were found to predominantly react against this part of Aβ.
Unfortunately, the immunization protocols are not entirely problem-free. The AN-1792 vaccination study had to be halted due to development of T cell-mediated meningoencephalitis in a minority of treated patients (see ARF related news story). Another concern with both the active and passive protocols is the risk for microhemorrhages and vasogenic edema. Seubert hypothesized that such unwanted treatment effects can be explained by amyloid removal from the brain vasculature, causing transient weakening of the vessel walls. Interestingly, the smooth muscle layer of the vessels where vascular amyloid was cleared appears to recover over time and adopt a normal phenotype.
Overall, Seubert concluded that both preclinical and clinical data now strongly indicate that immunotherapy can efficiently ameliorate plaque pathology both in mice and possibly patients. With further research and development, it can hopefully be learned how to avoid disturbing side effects.
Next, Bernd Bohrmann from F.Hoffmann-La Roche, Basel, Switzerland, presented efforts to generate Aβ plaque-specific antibodies. By phage display, Bohrmann and colleagues selected different Fab fragments, which they then optimized and finally converted into human IgGs with sub-nanomolar binding affinities for Aβ40/42. The researchers characterized the antibodies thoroughly, both in vitro and in vivo, to ensure that they had optimal binding properties and targeted physiologically relevant Aβ species. For example, they recently developed an Aβ antibody recognizing a conformational dual Aβ epitope that was shown to attenuate Aβ oligomer-induced toxicity and trigger phagocytosis of Aβ by macrophages in cell culture. Moreover, this antibody was efficient in reducing plaques from the brains of PS2APP mice. Notably, the most efficient removal was seen for the small plaques.
Bohrmann also described the development of a novel mouse model, which is a cross between the PS2APP mouse and a mouse overexpressing both the heavy and the light chains of a human anti-Aβ antibody under the peripherally active MHC-1 promoter. These double-transgenic mice displayed substantially reduced plaque pathology as compared to mice with the PS2APP transgene alone, demonstrating that endogenously expressed Aβ antibodies from peripheral tissues can be effective against amyloid deposition in brain.
N-terminal Aβ antibodies, such as those developed by Seubert and colleagues, seem to be a good choice for the removal of plaque pathology. But will plaque removal also lead to a concomitant improvement in disease symptoms, or do we instead need to target smaller Aβ species before they deposit as plaques? Lars Nilsson from Uppsala University, Sweden, has taken the latter approach. By trying to mimic the aggregation process in a test tube, Nilsson and colleagues identified a prefibrillar oligomeric Aβ species of >100 kDa, which they defined as an Aβ protofibril. Nilsson, Anna Lord, and other colleagues from Uppsala University then raised an antibody, mAb158, against this Aβ (Englund et al., 2007), and recently completed their first animal study with it.
Nilsson showed that treatment with mAb158 effectively prevented Aβ accumulation in younger, and selectively lowered levels of soluble protofibrils in older, amyloid-depositing mice. The mouse model used has been developed by Nilsson and expresses a combination of the Arctic and Swedish APP mutations under the control of the Thy1 promoter (see Lord et al., 2006; Philipson et al., 2008). These mice display high levels of Arctic Aβ, which have a propensity to form Aβ protofibrils. A similar type of antibody, with a conformational epitope, is now being developed for clinical use by Bioarctic Neuroscience AB, Stockholm, Sweden, in alliance with the pharmaceutical company Eisai Co Ltd., Tokyo, Japan.
The next speaker, Bengt Winblad from the Karolinska Institute, Stockholm, Sweden, reported from his group’s ongoing clinical immunotherapy trials. Winblad and his team have, together with different pharmaceutical companies, entered into as many as seven trials. With Novartis, Basel, Switzerland, they are investigating an active immunization protocol, utilizing Aβ1-6. This approach successfully cleared plaques in APP23 transgenic mice and did not cause microhemorrhages. In a safety study, two different doses, 50 μg and 150 μg, were compared to placebo on 58 patients, and 30-82 percent of the participants in the different cohorts developed Aβ-specific antibodies in serum. No severe side effects were seen, and Novartis is now gearing up to perform a larger study on the efficacy of this vaccine.
Winblad finished by emphasizing several practical aspects of the clinical trials that the companies need to consider. Overall, he believed that each patient visit (up to six hours) is too long and demanding for the often already fragile dementia patients. In particular, the numerous cognitive tests (which are typically computerized nowadays) and the MRI/PET scans are very stressful for the patients. Furthermore, Winblad believed that the exclusion/inclusion criteria commonly used are not properly defined and make the recruitment unnecessarily complicated. For example, for add-on studies he suggested that the requirement for at least six months’ treatment with acetylcholine esterase inhibitors should be changed to three months and that the use of any acetylcholine esterase inhibitor—not just donepezil—should be allowed. Finally, he criticized the informed consent documents in use for often being extremely long and detailed, and emphasized that the consent primarily should serve the patient’s purpose and not be used as a legal disclaimer for the companies to avoid lawsuits.
Switching to Parkinson disease and other α-synucleinopathies, Eliezer Masliah from the University of California, San Diego, discussed his findings published in PLoS ONE last year (see ARF related news story on Tsigelny et al., 2008), suggesting that Aβ can interact with α-synuclein and thereby induce the formation of Lewy bodies. Masliah and his team were able to demonstrate that human α-synuclein expression via a lentiviral vector promotes α-synuclein aggregation and degeneration in APP transgenic (tg) mice. Moreover, he speculated that the interactive effects may take place in lysosomes, where Aβ and α-synuclein have been shown to be colocalized in cell-based experiments. These and other observations raise the intriguing possibility that α-synucleinopathies could be targeted by immunotherapy against Aβ. Preliminary findings suggest that this indeed seems to work. By immunizing hAPP x hα-synuclein double transgenic mice with Aβ, Masliah and colleagues showed that not only Aβ, but also α-synuclein pathology, was reduced.
However, Masliah’s team also devotes efforts to a more direct approach. By intracranial administration of different α-synuclein antibodies, they found that C-terminal antibodies with high affinity against membrane-bound α-synuclein were more efficient in clearing existing pathology in α-synuclein tg mice. Based on these findings, they set out to treat mice with intraperitoneal injections of a C-terminal α-synuclein antibody and found α-synuclein brain pathology was reduced. Moreover, by fluorescently labelling them, they demonstrated that the antibodies indeed pass the blood-brain barrier and are taken up by lysosomes. The molecular mechanisms involved in lysosomal uptake and degradation are now being intensely studied by Masliah, with the aim of targeting them therapeutically. One such candidate has already been found, beclin-1, as viral transduction of a beclin-1-expressing vector was shown to initiate the autophagy pathway and thereby reduce α-synuclein accumulation.
Next, Martin Ingelsson from Uppsala University presented ongoing work on α-synuclein oligomerization. Together with Joakim Bergström and other colleagues at Uppsala University, they generated various α-synuclein oligomers from recombinant protein in vitro and generated pure oligomeric sample preparations that they use as antigens to raise conformation-specific monoclonal antibodies. Once generated, such antibodies will be tested for their efficacy on transgenic α-synuclein mice and also adopted for ELISA measurements on patient blood and cerebrospinal fluid to investigate whether such soluble intermediate pre-aggregates are increased early in the disease process of Parkinson disease and dementia with Lewy bodies.
As for antigen characterization, different stabilization protocols have been developed to generate highly stable oligomeric samples with distinct morphological and functional properties. One such species of ring-formed oligomers, approximately 30-120 nm in diameter and 4-8 nm in height, correspond to between 50 and 75 α-synuclein monomers in size. These protofibrils are also highly toxic in an MTT assay and are now being used as antigens to raise antibodies.
Although multiple sclerosis (MS) has not traditionally been regarded as a neurodegenerative disease, Jan Hillert from the Karolinska Institute, primarily known for his vast research on underlying genetic disease mechanisms, presented an overview on the current status of immunotherapy against this disease. Diffuse neuronal loss and cortical degeneration are features commonly seen along with the demyelination in MS.
Immunomodulatory treatment for MS has been available since the mid-1990s, first with β-interferon and somewhat later with glatiramer acetate. This first generation of MS treatments had a partial effect, reducing the number of relapses by about 30 percent and preventing new inflammatory lesions seen on MRI by 40-80 percent.
The first drug of a new generation of highly efficacious MS treatment, Natalizumab from Elan Pharmaceuticals, was approved by the FDA for treatment of multiple sclerosis in 2004. In patients who also received β-interferon, an increased incidence of progressive multifocal leukoencephalopathy, a rare viral condition, led to a temporary withdrawal, but the drug is available again since 2006, and to date approximately 50,000 patients have been treated. Natalizumab in monotherapy, or together with β-interferon, decreases the number of relapses 60-70 percent, whereas the number of plaque lesions decline over 90 percent. These effects are significantly better than with β-interferon monotherapy. Although Natalizumab, which binds to extracellular matrix receptors α4β1-integrin on mononuclear leukocytes, often causes milder side effects, it is a valuable contribution to the drug options for multiple sclerosis.
Hillert continued to describe other antibodies now being tested against multiple sclerosis. Rituximab, originally approved for treatment of non-Hodgkin lymphoma, is an anti-CD20 antibody directed against surface antigen present on B cells. Yet another antibody, Alemtuzumab, traditionally used as a second-line therapy for chronic lymphatic leukemia, is now being evaluated for multiple sclerosis treatment. Like Rituximab, Alemtuzumab targets a lymphocyte cell surface antigen, CD-52.
All of the antibodies currently in use or in evaluation against multiple sclerosis seem to affect only the remitting-relapsing type of disease. Thus, an immunotherapy effective against the progressive form of multiple sclerosis still needs to be developed.
Amyotrophic lateral sclerosis (ALS) treatment options were discussed by Neil Cashman from the University of British Columbia. Cashman is the chief scientific officer and co-founder of Amorfix Life Sciences, a Canadian company focused on developing novel therapeutics against neurodegenerative disorders. In their program against (ALS), they have identified several so-called disease-specific epitopes (DSEs) exposed by misfolding of superoxide dismutase 1 (SOD1), the protein aggregating in the spinal cord and brain in familial cases caused by SOD1 mutations. Two of these epitopes, DSE1 and DSE2, were found to be immunogenic and antibodies raised against them specifically recognized misfolded SOD1, both in familial and sporadic cases (although for the latter, SOD1 inclusions are less prominent). Thus, these epitopes could constitute a relevant immunotherapeutic target both in familial and sporadic disease forms.
Cashman’s research team has successfully targeted two of these SOD1 epitopes with immunotherapy. By actively immunizing SOD1 G93A tg mice with either DSE1 or DSE2, they found that although treatment did not delay onset of symptoms, immunized mice had up to a 50 percent increase in survival compared to sham-treated mice. Passive immunotherapy with a DSE2 monoclonal antibody also demonstrated increased survival in G37R mice. Neuropathologically, no consistent pattern correlated with increased survival of the DSE immunized mice.
Finally, Cashman presented evidence that SOD1 may propagate pathology via prion-like mechanisms. By developing antibodies against wild-type (wt), mutant, and misfolded SOD1, they studied the conversion of SOD1 in cell cultures. Intriguingly, they found that cells expressing mutant human G127X SOD1 also harbored misfolded wt SOD1, whereas cells expressing only wt SOD1 did not contain misfolded protein. In additional experiments, the researchers demonstrated that extracellular immunoneutralization could block disease-propagating effects, whereby misfolded protein seemed to spread from cell to cell.
Anne Messer from the Wadsworth Center, New York State Department of Health, Albany, New York, has a wide experience in disease mechanisms for Huntington disease. Messer has developed the concept of treatment with specifically engineered single-chain Fv (sFv) antibodies against carefully selected epitopes of proteins that misfold intracellularly, especially huntingtin and α-synuclein. When such antibodies are expressed as intrabodies in cultured cells overexpressing mutant human huntingtin exon 1, the intranuclear aggregates are greatly reduced. Moreover, dual transgenic Drosophila flies expressing both intrabody and mutant huntingtin have a significantly increased life span as compared to transgenic flies overexpressing only mutant huntingtin.
Intrabodies against the disease-causing pathological protein seem to be a feasible strategy also for Huntington disease. But how should these antibodies be administered in the clinical situation? Messer presented intracranial antibody delivery via AAV vectors as one possibility, since her group found transgenic mice with widespread intrabody expression to have many fewer huntingtin aggregates overall. Notably, in the transduced neurons (approximately 20 percent of all neurons) they observed a significant reduction in aggregates. As an alternative approach, combinations of intrabody gene therapy with either small molecular therapies or chaperone genes are being tested in flies and in mice. The first published fly studies show additive effects of intrabody plus Hsp70.
Finally, Messer cautioned that therapies aimed at increasing huntingtin aggregation (which have been suggested to trap and thereby defuse harmful intermediate moieties) could be detrimental for polyglutamine disorders. In a new paper published in PLoS ONE, Messer and colleagues, using a construct isolated by Michael Sierks at Arizona State University, Phoenix, have found that a single-chain antibody against fibrillar α-synuclein targets a pathological huntingtin conformation and increases its aggregation. Cells treated with this intrabody displayed increased oxidative stress along with decreased survival (see Kvam et al., 2009).
In the last presentation of the day, Adriano Aguzzi from the University Hospital in Zürich, Switzerland, described some intriguing findings on mechanisms of peripheral prion replication and neuroinvasion. By studying sheep in Sardinia, it was found that prions could transfer via macrophages in the milk duct, transmitting prions to suckling lambs. Similar mechanisms are at play also in mice in which prion proteins are presented by B cells to macrophages or follicular dendritic cells (FDC). Lymphotoxin signaling plays an important role in maintenance of FDC and, if lacking, prion replication is impaired.
Aguzzi and his team have further explored the molecular mechanisms involved in the interplay between B cells and macrophages and found that Mfge8 (phosphatidylserine-binding protein milk fat globule epidermal growth factor 8) can block prion uptake. His team has shown that Mfge8 is normally produced by follicular dendritic cells, rather than by macrophages—as was previously thought.
Aguzzi finished by describing work on organotypic brain slices, illustrating the influence of other factors in prion conversion and disease propagation. For instance, if microglia are depleted, prions replicate much faster. This indicates that microglia are beneficial rather than detrimental in prion diseases. It will be exciting to investigate whether this positive function of microglia is true also for other neurodegenerative diseases.
This is Part 1 of a two-part series. See Part 2.