2 April 2010. At the 11th International Geneva/Springfield Symposium on Advances in Alzheimer Therapy, this latter goal—advances in therapy—saw little sprigs of news pop up in talks that otherwise took stock of the current status of immunotherapy as one of the major new approaches in that field. Several speakers started by showing tables listing ongoing immunotherapies that have sprung up since their first demonstration in mice (Schenk et al., 1999). “There is a huge amount of activity on immunotherapy going on in industry,” said Dale Schenk of Elan and Janssen Alzheimer Immunotherapy. Roger Nitsch of the University of Zurich noted that at present, some 10,000 patients are enrolled in more than 40 ongoing trials. The Phase 3 trials—Bapineuzumab from Elan et al., Lilly’s Solanezumab, or Baxter’s Gammagard—are not expected to report results this year. Smaller earlier-stage trials may do so, however. These include one by Genentech, one by GlaxoSmithKline, as well as the ACC-001 active vaccine that represents a second generation to Elan/Wyeth’s AN1792 and is now being developed by Janssen following an agreement between these two companies in late 2009. Several of the current immunotherapy trials are “loaded” with biomarker measurements, said Ronald Black of Pfizer. Solanezumab even advanced to Phase 3 solely on the strength of its safety and biomarker data, without clinical efficacy signals, said Ronald DeMattos of Eli Lilly and company. In fact, so many immunotherapy trials are now being done that companies might be well advised to share experiences on what works and what needs tweaking, said Bengt Winblad of the Karolinska Institute in Huddinge, Sweden.
Generally speaking, immunotherapy trials have begun adopting more innovative designs. They measure CSF Aβ levels and brain amyloid load as well as more distal disease markers such as CSF tau, brain volume by MRI, or, in some cases, brain function by means of FDG PET, though these measures are not usually primary endpoints. The safety record of immunotherapies is a topic of much discussion. For his part, Nitsch said that so far this novel type of therapy appears to be acceptably safe despite some challenges. For active vaccines, pro-inflammatory T cell responses and adjuvant reactions must be avoided. For antibody therapies, vasogenic edemas and microbleeds can occur that appear to be related to dose and to ApoE genotype, and require careful monitoring by MRI. On efficacy, preliminary data suggest removal of brain amyloid and hints of improvement in some responders (Hock et al., 2003; Vellas et al., 2009; Salloway et al., 2009). This latter point is controversial (Holmes et al., 2008) and will remain unsettled until more definitive data come in. Some important differences between active immunization and passive (i.e., antibody) therapies are that in the former, only a minority of patients respond and it cannot be stopped easily, whereas the latter requires more injections and would likely be more expensive, Winblad said.
While new trials are ongoing, scientists are mining the older one for further insights on biomarker data, and that is where some news tidbits were to be had in Geneva. For example, Nitsch previewed a study in press in Brain, in which Alberto Serrano-Pozo and colleagues in Brad Hyman’s laboratory at Massachusetts General Hospital in Charlestown analyzed brain tissue of five more people from the AN1792 trial who had since passed away. Like previous AN1792 autopsy reports, this study documented extensive removal of amyloid from the hippocampus after immunization. One of the five patients had developed meningoencephalitis; this person had some amyloid left in the brain regions Serrano-Pozo examined, and the others had hardly any at all, Nitsch told the audience. Areas of total plaque and of dense core plaques were markedly down. None of these five patients had increased cerebral amyloid angiopathy (CAA) in their brains compared to autopsies of untreated patients, Nitsch added (see also Schenk talk below).
Moreover, Serrano-Pozo et al. noticed that immunotherapy appears to have treated the neuritic dystrophy known in AD. By quantifying how curved (i.e., abnormal) the neurites were in treated compared with untreated patients’ brains, he found that the treated neurites straightened out both in plaque-free areas and near the remaining compact plaques. This would suggest a degree of morphologic recovery of hippocampal neurons, Nitsch said. In addition, the scientists found that immunotherapy appears to have reduced PHF1 immunostaining in the hippocampus of treated patients, particularly those who had not suffered the inflammatory side effect. PHF1 staining reflects hyperphosphorylated tau. In contrast, the scientists saw no change in staining with Alz50, which is thought to reflect a misfolded form of pathologic tau.
Nitsch also spoke about work being done at Neurimmune Therapeutics AG, a start-up biotech company located at the University of Zurich. Scientists there try to develop future therapeutics from naturally occurring antibodies that protect people against AD. Using a trademarked technology called reverse translational medicine (RTM), the Swiss scientists isolate such antibodies from normally aging humans and clone them in vitro. In preclinical experiments, some of these antibodies have lowered the level of pathology in mouse models of neurodegenerative disease, including models of α-synuclein pathology, and delayed disease onset in the SOD1-transgenic model of amyotrophic lateral sclerosis, Nitsch said.
In his talk, Elan’s Dale Schenk noted that because each antibody is slightly different, the 40+ ongoing clinical trials are likely to generate diverging results, as well. Elan is trying to understand its clinical antibodies with crystallography studies (see Basi et al., 2010; ARF related Eibsee story).
Schenk emphasized that as analysis of the first clinical immunotherapy studies continues, its results are beginning to suggest to his mind that mice and humans do show some similarities in their response to immunotherapy, after all. “That's my theme here today,” he said. Previously, mice had been viewed at predicting human outcomes poorly. Specifically, that’s because mice gave no advance warning of the meningoencephalitis that derailed the AN1792 vaccine, and more generally, it’s because clinical failure following murine success stories has become a disheartening theme across neurodegeneration research in recent years (see Alzforum Live Discussion). By now, however, it looks to Schenk as if a growing number of immunotherapy’s established effects in mice—reductions of plaques, of vascular amyloid, of soluble Aβ, of tau pathology, straightening out neurites, behavioral improvement—do seem to occur in humans as well.
Vascular amyloid has become a topic of intense interest, Schenk said. It comes in different forms. Studying amyloid pathology on capillaries of immunized mice, Elan scientist Dora Games and others found that capillary Aβ initially forms during the clearance of plaques and then decreases again over the course of treatment. “This is a temporary phenomenon. We do not know what it means clinically, but we need to factor it in,” Schenk said. Pre-existing vascular amyloid on vessels larger than capillaries corresponds to what in humans is called CAA. Schenk said the 3D6 antibody that formed the basis for Bapineuzumab temporarily increases, but eventually clears, this pathology, noting, “It all depends on when you look.” Microhemorrhages colocalize with vascular Aβ clearance in a subset of blood vessels. In mice, their appearance as measured by hemosiderin staining also depends on dose, and it wanes over time. Finally, Schenk claimed that cerebrovascular amyloid in mice leaves a morphological fingerprint in that the vessel wall is abnormally thicker in some places than in others and that this, too, resolves over time with immunotherapy.
In this context, it is worth noting that Elan/Wyeth in April 2009 dropped the highest dose of Bapineuzumab in its Phase 3 trial of ApoE non-carriers. (ApoE4 carriers, who are assumed to have a higher amyloid burden, only received a single low dose to begin with. See ARF related news story.) Commentators said that Elan took this measure to reduce the risk of MRI flare signals called vasogenic edema; these are thought to be dose-dependent vascular side effects possibly having to do with clearance of Aβ or with inflammation. AD clinicians told this reporter that some of the edemas generate no symptoms, whereas others caused transient dizziness or gait disturbance.
Regarding tau pathology, clinical hints that immunotherapy changes have come from several studies, Schenk said. That AN1792 reduced CSF tau and phospho-tau staining, but not tangles, in autopsy tissue of responders, just like previously in mice, has been reported in published papers (e.g., Gilman et al., 2005), and again by Nitsch in Geneva. Bapineuzumab appears to be doing a similar thing. In its Phase 2 study, CSF levels of phospho-tau decreased as well, though that finding stayed just shy of reaching statistical significance, said Pfizer’s Ronald Black, who presented further clinical data on biomarkers. The hope is that if anti-Aβ immunotherapy indeed reduces tau pathology in CSF and brain, as well as other AD markers, then that would be regarded as evidence for the claim that immunotherapy modifies the underlying disease.
On Aβ reduction itself, autopsy research recently got company when an amyloid imaging biomarker study reported that PiP-PET imaging was able to measure a decrease in living people who had received Bapineuzumab (Rinne et al., 2010; ARF related news story). In Geneva, this finding came up frequently in presentations and hallway discussions. It is considered pharmacodynamic proof that the antibody is bioactive in the brain, though it does not by itself mean the patient will get better.
MRI measurement in immunotherapy trials has been more complicated, Black said. Prior research has established that patients lose volume in the hippocampus and whole brain as their disease progresses (e.g., Barnes et al., 2009); however, in the AN1792 trial, responders unexpectedly lost more brain volume, not less, than non-responders six months after treatment (see ARF related news story). Since then, researchers have tried to understand this finding better, Black said (e.g., Fox et al., 2005). They found, for example, that the areas where the responders’ brains shrank the most were also the areas where there was the most amyloid. “This fits with autopsy data,” Black said. This volume loss stopped one year after treatment. Overall, the researchers suspect that brain areas might temporarily contract because the amyloid is taking water with it as it is clearing out. But the last word is far from spoken, because in the Bapineuzumab Phase 2 trial, MRI data came out as expected. That is, the strongest responders showed less volume loss than did people on placebo. “This was different than in AN1792,” Black said, qualifying that this is preliminary because it came up in post hoc analysis, which can bias the sample.
As for plasma Aβ measurements, this marker went up one day after patients received the Bapineuzumab injection, Black said. Through this kind of research, scientists are learning to distinguish between rapid markers of target engagement, such as a 24-hour plasma signal, and slower, more distal markers hinting at disease modification. In his talk, Ronald DeMattos of Eli Lilly and company also focused on biomarker studies meant to establish separate points: that the brain gets exposed to therapeutic, that the therapeutic engages its target, and that the therapeutic might change disease. Lilly’s research on its Phase 3 antibody Solanezumab has developed different biomarkers for each of these steps, DeMattos said (see prior ARF conference story).
All of these signals are preliminary hints awaiting more definitive clinical data. In the meantime, big unsolved questions remain, Schenk said. Above all, no one yet knows whether reducing Aβ oligomers, plaques, vascular amyloid, and phospho-tau is sufficient to help people clinically with their illness. If so, does it help at the mild to moderate stage, at the early stage, or for prevention? There is a movement afoot to get preclinical studies off the ground. Its supporters, however, have not settled on a specific therapy; they are considering all options including immunotherapy (see ARF API story; see ARF DIAN clinical trials story).
Bengt Winblad of the Karolinska Institute in Huddinge, Sweden, offered the experience of an academic clinician-researcher whose center has already conducted seven different immunotherapy studies for various companies. His site has seen that many participants are highly motivated to contribute to research and help future generations of patients. Their idealism is tested by several practical shortcomings, however. For example, neuropsychiatric testing puts them through a tiring six-hour wringer; shorter visits might improve their compliance. The informed consent form can cause irritation, Winblad noted. Frequently 18 pages long, it serves as much as a legal indemnification package for the sponsor as to inform the patient. Six pages for this form should be enough, Winblad said. Furthermore, many trials require that patients be stable on acetyl cholinesterase inhibitor drugs for six months prior to enrollment. In practice, this may miss a window of opportunity because it means that patients who might be motivated at diagnosis to join a trial lose that interest by the time these drugs have stabilized their symptoms. Some protocols are irksome in requiring an add-on design specifically to donepezil. “We use all the acetyl cholinesterase inhibitor drugs, not just this one,” Winblad said. Participating sites find recruitment difficult when the list of exclusion criteria becomes long and overly restrictive. For example, some trials exclude patients whose blood pressure exceeds 140/80, Winblad said.
Echoing the biomarker theme that pervaded all talks, Winblad recommended that trial designers include baseline labs, CSF and imaging data for each person as a reference for comparison with later findings. He emphasized the importance of having a contact person available around the clock for patients as well as for sites to contact the safety monitoring board. Finally, Winblad urged sponsors to compare notes, especially on practical issues that are independent of their proprietary drug. “As one center running several ongoing clinical trials, it can be dismaying to see that different sponsors make similar planning errors.”—Gabrielle Strobel.