Live Discussion: Alzheimer Immunotherapy Trial Grounded: Time to
Reassess Safety and Vaccine Design
Updated 17 October 2002
Live Discussion held Tuesday, March 5, 2002, 12 pm EST.
Participants: Claude Nicolau, Ruth Itzhaki, Einar Sigurdsson, Blas Frangione, David Morgan, Bekka Solomon, Ken Ugen, Craig Atwood, and David Weiner
See Transcript of Live Discussion - Posted 3 May 2002
11 February 2002. Last month, the frontrunner in the effort to develop a therapeutic vaccine for Alzheimer's disease suffered a widely noted setback when Elan Pharmaceuticals and American Home Products, the companies sponsoring the first human trials, announced their decision to temporarily halt dispensing any more trial drug until they had found out why four patients in France had come down with cerebral inflammation (See ARF story.)
The effort to develop a vaccine is being closely observed, partly because the concept of treating a neurodegenerative disease with a vaccine is radically new and controversial, and in part because the approach has so rapidly moved into the clinic since the first paper describing it appeared three years ago (Schenk et al. 1999). Hopes and stakes are high, and by now, many academic labs and pharmaceutical companies are working on modifications of Elan's original approach.
While the research community is awaiting results of Elan's investigation into the causes of these patients' inflammatory response, the Alzheimer Research Forum is hosting a live discussion into the underlying issues and future directions this incident points to. -Gabrielle Strobel.
Question and Answer session with Blas Frangione, New York University.
Posted 11 February 2002.
Q: What was your initial reaction when you read the press release announcing that Elan Pharmaceuticals was temporarily suspending dosing in their trial of the Ab vaccine?
A: This was not unexpected. When you immunize humans with something they have, in this case Ab1-42, you can produce autoimmune disease, that was known. Of course this did not happen in the transgenic animals. So one question was: will it happen in humans?
It might be an antibody against your own protein that induces inflammation. An antibody against Ab1-42 could conceivably also recognize AbPP.
Alternatively, it could be that the protein that was given to the patients intramuscularly went into the brain. We and other groups have shown that the peptide goes into and out of the brain very well. Ab1-42 could have precipitated in the brain, caused inflammation, and made plaque deposition worse. Which of these two possibilities occurred in these four patients, if any, I do not know. But these certainly can happen, so I was not too surprised.
Q: Could the inflammation be a beneficial part of the reaction to the vaccine? Dave Morgan saw temporary CNS inflammation in vaccinated mice.
A: Could be, but you cannot test that in humans. It creates this ethical dilemma that when someone has a cerebral inflammation you cannot go on vaccinating for months to find out if it is a beneficial or a detrimental reaction to the trial vaccine. Remember that transgenic mice are not humans. These sorts of data in animals do not reflect very well what will happen in humans. This is a general phenomenon, and a big problem. Elan was perfectly correct in stopping dosing for now.
Usually when you vaccinate, especially in children, you do get a local inflammation at the injection site. It lasts a few days and goes away. In this case it is probably not that simple. The press release did not explain for how long the inflammation went on. That is an important question. Has this gone on for months? We do not know. Also, the inflammation here was probably not local. It was in the brain, which means that it went through the blood-brain-barrier.
Q: What's most important now?
A: The important thing is to find out why this happened. For example, if it is an autoimmune response, these patients will have a high titer of antibody against the injected antigen. You can find that out with a spinal tap and blood sample, which you react with the injected antigen.
If, however, this is an aggravation of plaques, they can find out with MRI or other methods. It is possible to find out in the next few months. It is very important to study these two mechanisms, and the third possibility that desired improvements would happen via a temporary inflammation.
Q: It is difficult for a drug company to be first to develop a new approach, and vaccination is a radically new therapeutic concept in Alzheimer's disease. Competitors who are a year or more behind can learn from the stumbles of the leader and refine their method. Do you think these other approaches waiting in the wings stand a better chance?
A: Yes I do. Of course I am partial to an approach we published earlier this year (see ARF news story). Personally, I worried about Elan's approach from the beginning. When they published the initial paper on injecting Ab1-42 more than two years ago (Schenk et al., 1999), I was very surprised they did it in that way. At least in vitro, 1-42 aggregates in 20 minutes and also is very toxic, killing cells within a few days.
That is why we decided not to use the carboxy terminal of Ab, which aggregates and produces the in-vitro toxicity. We only use the antigenic determinant, which is known to be in the amino terminus of the peptide between position 1 and 28. That way we hoped to reduce toxicity.
Elderly patients generally have a poor immune response, so we added polylysine to increase the antigenicity. Polylysine has two properties: it diminishes the probability of Ab being converted into a b-pleated form, which is the toxic, fibrillogenic material, and secondly it increases antigenicity, which has been shown in guinea pigs and mice.
At this point there is really no good scientific argument to use Ab1-42, because in our paper we have the same antibody titer as the Elan group.
Q: Some advocate passive immunization, that is, injection with anti-Ab antibodies, as a better approach?
A: I disagree. Passive immunization works well for hepatitis B and other acute diseases, including even prion disease. Because of the short half-life of immunoglobulin G, you have to inject antibody every three weeks. But if you do that for months, you can induce serum sickness or an anaphylactic reaction. AD is a chronic process that takes10 or 20 years, and passive immunization for such long time is difficult.
An exception would be if you used an approach called one-chain antibody, which means you make a small, roughly 100-residue piece of antibody that contains the antigen-recognition site. Those you can inject for long periods of time. Or you could make a humanized antibody, there are many options to pursue.
Q: Can CNS inflammation appear suddenly?
A: Making the diagnosis of cerebral inflammation in a patient with AD is not simple. Patients cannot express themselves well, they do not understand what you are asking them and don't necessarily describe their symptoms clearly. Only the caregiver sees that something is wrong. Then you need to do a spinal tap, or MRI to see edema in brain. It does not become apparent in one week.
Commentary to Background Reading by Pasinetti et al., 2002 - Posted 26 February 2002
Inflammatory consequences: benevolent, or virulent?
Pasinetti et al.'s main point is that there is an intriguing paradox: prevention of inflammation appears to be beneficial in respect to development of Alzheimer disease (AD), but the inflammation caused by vaccination with Ab is beneficial too, at least in mice. The protection anti-inflammatory agents afford presumably is due to the prevention of the induction of pro-inflammatory cascades-microglial activation and generation of free radicals-which are thought to cause neuronal injury. Passive or active immunization by intraperitoneal injection of Ab antibodies or of Ab peptides appears to remove Ab plaques by activating pro-inflammatory microglial cells via immunoglobulin receptor signalling.
Whether vaccination is beneficial also for humans is yet to be determined, and in fact there has recently been an interesting development during the progress of a clinical trial: four patients treated with the AN-1792 vaccine, a synthetic version of Ab, have suffered from inflammation of the brain. The company press release notes that viral infection can cause inflammation, and furthermore that in some of these patients (presumably two or three of the four examined), a virus - herpes simplex virus type 1 (HSV1) (D.Schenk, personal communication) - was found in cerebrospinal fluid (CSF).
HSV1 has long been known to reside in a latent state in the PNS in most humans, and in our lab we showed, using polymerase chain reaction, that it is present also in the brain in a high proportion of the elderly, including AD patients [5,6], although not in the brain of most younger people (Wozniak, Lin, Cairns, Mann, Itzhaki, in preparation). In the virus' dormant state, only its genome is present, there is only one set of transcripts-those termed latency-associated transcripts-and no viral proteins have yet been detected. The latent virus can reactivate in the peripheral nervous system (PNS) as a result of inflammation and other forms of stress, and this results in an acute infection, i.e. synthesis of whole virus, which causes cold sores in some people.
We have proposed that a similar process can occur in the central nervous system (CNS) and that the consequence is a localized encephalitis. In fact, we now have direct evidence that reactivation does occur in the CNS and we infer that the whole viral genome must therefore be present. We have detected antibodies to HSV1 in CSF that are not due to leakage across the blood-CSF barrier (Wozniak, Klapper, Combrinck, Esiri, Wilcock, Itzhaki, in preparation). For an in-depth discussion of our hypothesis, visit our entry in the Alzheimer Hypotheses/Online Forums section.
Presumably, in the clinical trial, the vaccine caused inflammation that reactivated HSV1 residing in the brain. HSV may, in turn, have caused damage in the way that viruses normally do-through the inflammatory response. Following our finding that HSV1 in brain, together with carriage of the type 4 allele of the apolipoprotein E gene (apoE-e4), is a strong risk factor for AD (and that apoE-e4 is a risk for cold sores), we suggested that the extent of damage caused by HSV1 is greater-or the extent of subsequent repair is less-in the presence of the apoE4 isoform [2,4,7]. We attributed the protective action of anti-inflammatory agents to their preventing reactivation episodes caused by inflammation.
Another relevant finding is that certain vaccines appear to protect against AD development. A study of 3,682 cognitively normal people and 183 people newly diagnosed with AD found that after adjustment for age, sex, and education, past exposure to vaccines against diphtheria or tetanus, poliomyelitis, and influenza was associated with lower risk of AD [11]. This raises the possibility that viruses other than HSV1 might be involved in AD. Alternatively, it could be that the vaccines reduce the frequency of inflammatory events as a consequence of preventing infection with the target virus. They also support the possibility that vaccination against HSV1 might be feasible in future, as do our findings that a vaccine of mixed viral glycoproteins protects HSV1-infected mice from establishment of latency by the virus in brain [8].
Pasinetti et al. point out the need for considering possible effects of Ab immunization therapy on neurofibrillary tangles (NFT) and neuronal loss, especially as these features do not occur in animal models of AD. Also, the authors discovered that a raised Il-6 level in severe AD correlates with level of NFT pathology [9]. Thus, immunization might increase the numbers of NFT. Incidentally, it may be highly relevant that the main cytokine induced by HSV1 infection is IL-6 [1,10], and that NFT have been described in two viral disorders [3]: subacute sclerosing panencephalitis, caused by measles virus, and herpes simplex encephalitis. A further major difference between animal models and humans is that, unlike humans, mice and other animals are not natural hosts for HSV1. This might explain the absence of NFT and neuronal loss in Ab-immunized mice.
Pasinetti et al. suggest that selective anti-inflammatory agents that interfere only with harmful pro-inflammatory responses might be developed. Alternatively, if the putative effects of the vaccine on HSV1 reactivation are confirmed as the cause of the inflammation in the AN-1792 trial, usage of antiviral agents in combination with the vaccine should perhaps be considered.
Clearly, further information about disease states, the role of specific cytokines at each stage, and the mechanisms involved in plaque clearance, as well as modification of the antigen, might enable conditions to be chosen which resolve the dilemma. -Ruth Itzhaki, Matthew Wozniak, Curtis Dobson, University of Manchester, UK.
References
[1] Baker M, Noisakran S. The relationship between interleukin-6 and herpes simplex virus type 1: implications for behavior and immunopathology. Brain Behav Immun 1999; 13: 201-211.
Abstract
[2] Dobson CB, Itzhaki RF. Herpes Simplex virus type 1 and Alzheimer's disease. Neurobiol Aging 1999;457-465.
Abstract
[3] Esiri MM. Typical and atypical viruses in the aetiology of senile dementia of the Alzheimer type. Interdiscipl Topics in Gerontol 1988;25:113-139.
Abstract
[4] Itzhaki RF, Lin W-R, Shang D, Wilcock GK, Faragher B, Jamieson GA. Herpes simplex virus type 1 in brain and risk of Alzheimer's disease. Lancet 1997;349:241-244.
Abstract
[5] Jamieson GA, Maitland NJ, Wilcock GK, Itzhaki RF. Latent herpes simplex virus type 1 in normal and Alzheimer's disease brain. J Med Virol 1991;33:224-227.
Abstract
[6] Jamieson GA, Maitland NJ, Wilcock GK, Yates CM, Itzhaki RF. Herpes simplex virus type 1 DNA is present in specific regions of brain from aged people with and without senile dementia of the Alzheimer type. J Pathol 1992;167:365-368.
Abstract
[7] Lin W-R, Graham J, MacGowan SM, Wilcock GK, Itzhaki RF. Alzheimer's disease, herpes virus in brain, apolipoprotein E4 and herpes labialis. Alzheimer's Repts. 1998;1:173-178.
[8] Lin WR, Jennings R, Smith TL, Wozniak MA, Itzhaki RF, Vaccination prevents latent HSV1 infection of mouse brain. Neurobiol.Aging 2001;22:699-703.
Abstract
[9] Luterman JD, Haroutunian V, Yemul S, Ho L, Purohit D, Aisen PS. Cytokine gene expression as a function of the clinical progression of Alzheimer's disease. Nature 2000;57:1153-60.
Luterman JD, Haroutunian V, Yemul S, Ho L, Purohit D, Aisen PS
[10] Oshima M, Azuma H, Suzutani T, Ikeda H, Okuno A. Direct and mononuclear cell mediated effects on interleukin 6 production by glioma cells in infection with herpes simplex virus type 1. J. Med. Virol 2001;63:252-258.
Abstract
[11] Verreault R, Laurin D, Lindsay J, De Serres G, Past exposure to vaccines and subsequent risk of Alzheimer's disease. Canad Med Assoc J 2001;165:1495-8.
Abstract
Comments
11 February 2002. The release from Elan certainly raises more questions for me than it answers. For example, what is the "marker of CNS inflammation" it is referring to?
Our first experiments in mice anticipated exactly the outcome that the Aß vaccine would provoke brain inflammation and lead to premature memory deficits (Morgan et al., 2000). We did not find the memory deficits.
In more recent work, we measured brain inflammation by microglia activation in response to the vaccine (Wilcock et al, 2001). There, we find a slow development of an inflammatory reaction, perhaps related to the slow rise in antibody titers, that was maximal in mice given five inoculations, roughly one per month. Interestingly, in mice given nine inoculations, the microglial activation had abated (it was 50% lower, but this was not statistically significant).
We can only hope that the same will prove true in the human trials. It may be that, initially, the vaccine causes an inflammatory reaction that normalizes after the microglia succeed in phagocytosing the amyloid deposits.
A vaccine trial continues whether you wish it to or not. You cannot reverse the vaccination (which is why we have suggested passive immunization for clinical trials). Still, this will likely permit us to determine if the same course of reaction occurs in humans.
Of course, it is entirely possible the inflammatory reaction in these four people is unrelated to the vaccine. Without more details, I cannot evaluate the likelihood of that outcome.
We would all like to see the vaccine work. Still, I am confident that if this one does not succeed, one of the half dozen other approaches currently underway will eventually prove beneficial." -David Morgan, University of South Florida, Tampa.
Comment: Vaccine Disrupting "Scab"? Ab as a Vascular Sealant that Protects Against Hemorrhage
10 February 2002. "The question whether removing amyloid-ß from the brain is therapeutic has been raised again by the recent interruption of the Elan/AHP trial due to clinical signs of inflammation. Numerous papers suggest that Aß has neuroprotective properties (reviewed in Atwood et al., 2001), but this literature is being overlooked in the great rush to find better treatments. Ab's protective functions include metal chelation, antioxidant activity and, perhaps most significantly in the present context, sealant properties that we think help to maintain the integrity of the blood-brain barrier and parenchymal structures (Atwood et al., 1998).
Consistent with Aß's proposed role as an antioxidant and its role in maintaining structural integrity under stress conditions are data showing that Aß binds copper under acidotic conditions, and that it possesses hydrophobic and hydrophilic regions, which both span the plasma membrane and bind to extracellular matrix molecules (Atwood et al., 1998; Chan et al., 1999; Kontush et al., 2001; Narindrasorasak et al., 1991). Together with Aß's ability to aggregate under inflammatory conditions, these properties make it an excellent candidate for a molecule that could form an intracranial 'scab'. The aggregated Cu,Zn-Aß would serve as a superoxide-scavenging, solid-phase matrix, which disassembles when Zn and Cu levels lower as tissue damage resolves (Pluta et al., 1999). This may explain the acute-phase generation and rapid cortical deposition of Aß in stroke and following head trauma (Roberts, et al., 1994), an important physiological response to limit loss of terminally differentiated neurons (Smith et al., 2000; Perry et al., 2000).
If indeed Aß acts as a seal to maintain the blood-brain barrier, then we would expect its removal to cause leakage of serum components into the brain, resulting in an immune or autoimmune response characterized by inflammation. A severe consequence would be the occurrence of mini-strokes. Notably, this breakdown of the blood-brain barrier is consistent with the presence of a virus within the cerebrospinal fluid, as was reported in some of the four patients under investigation. The adverse effects of Aß vaccination has halted the current clinical trial, but a more balanced assessment of the current literature would have cautioned against such trials in the first place." -Craig S Atwood, Glenda M. Bishop, George Perry, and Mark A Smith, Case Western Reserve University, Cleveland, OH.
References:
Atwood, C.S., Huang, X., Moir, R.D., Smith, M.A., Tanzi, R.E. Roher, A.E., Bush, A.I. and Perry, G. (2001). Neuroinflammatory Responses in the Alzheimer's Disease Brain Promote the Oxidative Post-translational Modification of Amyloid Deposits. In: Alzheimer's Disease: Advances in Etiology, Pathogenesis and Therapeutics pp. 341-361, Eds. K. Iqbal, S.S. Sisodia and Bengt Winblad, John Wiley & Sons, Ltd., UK.
Chan, C.-W., Dharmarajan, A., Atwood, C.S., Huang, X., Tanzi, R.E., Bush, A.I., and Martins, R.N. (1999) Anti-apoptotic action of Alzheimer Aß. Alzheimer's Reports 2, 1-6.
Perry, G., Nunomura, A., Raina, A.K., Smith, M.A. (2000). Amyloid-beta junkies. Lancet. 355, 757.
Comment:
5 March 2002. "It is difficult to discuss the reasons of the "inflammation"
observed in some of the patients having received Elan vaccine as long as detailed
data are not available. Is it the length of the peptide used as antigen, which
may reduce the specificity of the antibodies elicited, is it the activation
of complement, is it the location of antibody production, is it the immune reaction
itself and its mechanism suggested by Bard
et al, 2000 that are responsible, or a combination thereof? Schenk
et al., 1999 had shown that the mouse antisera from animals immunized with
the human synthetic Ab1-42 cross-reacted with mouse
amyloid-b peptide. The studies performed on mice
did not report any inflammation and neither did the Phase I clinical studies.
I would like to mention that in our study, using tetra -palmitoylated Ab1-16
reconstituted in liposomes-lipid A, no such cross-reactions with mouse amyloid
could be detected (see ARF
news story.) The mechanism of action of the anti-amyloid antibodies probably
plays a significant role in the development of inflammation. If the mechanism
proposed by Bard et al really occurs in vivo, inflammation appears as a distinct
possibility. On the other hand, mechanisms of antibody action such as those
suggested by DeMattos et al. (see
ARF news story), who raised antibodies directed against the central domain
of Ab, and by ourselves (loc.cit) with our antibodies
directed against liposomal tetra-palmitoylated Ab1-16,
may not necessarily lead to inflammation. Currently, we are conducting studies
aiming to better understand the mechanism of action of the anti-amyloid antibodies."
Claude Nicolau, Beth Israel Deaconess Medical Center, Boston.
Comment: Alzheimer vaccine: lest we forget.
5 March 2002. "At latest count (Saturday, March 2, 2002) 15 people in
the phase IIA trials of Elan-Wyeth's Alzheimer vaccine had developed symptoms
of encephalitis and/or meningitis, and the trials of this particular drug (AN-1792)
have been permanently discontinued. Elan's spokesperson, Ivan Lieberberg, has
defended the outcome in the press by claiming that these serious side effects
were totally unexpected.
Not all of us share this view. The AN-1792 vaccine was specifically designed
to generate an auto-immune response to amyloid-b peptide,
which is not only the principal constituent of senile plaques, it is also abundant
throughout the extracellular fluid of the brain. Indeed, it is present throughout
life at levels that are as high, or higher, than found in the brains of Alzheimer's
patients (for discussion see Robinson and Bishop, 2002). The ubiquity of this
peptide in normal individuals indicates that either amyloid-b
is not neurotoxic in its soluble form, or that efficient mechanisms exist for
the detoxification of amyloid-b. Proponents of amyloid-b
neurotoxicity have never adequately explained why the brain produces such a
toxin. Amyloid-b is metabolically expensive to produce
(42 amino acids), and requires a sustained rate of production to replace amyloid-
that is being constantly flushed from the cerebrospinal fluid. The investment
of so much energy in the production of a peptide that is toxic to the brain
would appear to defy the laws of natural selection, yet there are other examples
of neurotoxic substances that are released extracellularly in the brain. A distinguishing
feature of such neurotoxic compounds is that they are also essential for brain
function.
One example is glutamate, the principal excitatory neurotransmitter in the
nervous system. If glutamate neurotransmission is blocked, entire pathways in
the brain are inactivated, leading to sensory loss, amnesia and death; yet excessive
extracellular accumulation of glutamate can cause a state of excitotoxicity
that overactivates neurons and damages them. Excitotoxicity rarely occurs because
neuroprotective mechanisms, particularly in astrocytes, help to minimise any
neuronal damage that might result (Hertz et al., 1999). Another example is found
in ammonia, which is a metabolic byproduct that is highly neurotoxic. However,
ammonia is rapidly combined with glutamate to form the non-toxic amino acid,
glutamine. Thus, the brain neutralises two neurotoxins by combining them. Glutamine
serves as the main precursor for glutamate biosynthesis, and it can also be
harmlessly secreted across the blood-brain-barrier to decrease the brain's ammonia
load (for review see Robinson, 2001).
If vaccines were developed to eliminate glutamate or ammonia from the brain,
few researchers would be surprised if those vaccines caused brain dysfunction
or brain damage. However, our collective obsession with the supposed toxicity
of amyloid-b has blinded us to the obvious: that
this peptide would not be produced if its toxicity outweighed its contribution
to brain function. Indeed, the notion that soluble amyloid- might be important
for normal brain function (Robinson and Bishop, 2002) does not appear to have
been considered by Elan-Wyeth or the regulators of their vaccine trials. The
only published tests of the amyloid-b vaccine approach
have been in transgenic mice that overexpress human APP in addition to their
normal amyloid-b. When injected with human amyloid-b
the plaques in their brains are cleared, and the mice reportedly show no ill-effects.
However, such mice retain normal levels of their endogenous amyloid-b.
The appropriate control experiments, which involve injecting normal mice (or
primates) with their endogenous amyloid, have not been published (Check, 2002).
Due to this oversight, the first experiments to investigate the importance of
endogenous amyloid-b for brain function may have
been unwittingly conducted in humans in the AN-1792 trials.
Many contributors to today's online forum will argue that the vaccine approach
is still sound, and that we simply have to clear up a few glitches concerning
the method of delivery and the permeability of the blood-brain-barrier. They
may say that we need to fine-tune the strength of the inflammatory response
and increase the specificity of the antibodies. These proponents of amyloid-
toxicity will down-play the fact that any amyloid-b
vaccine is also likely to eliminate soluble amyloid-b,
and they will dismiss the fact that we have no idea what normal functions are
served by this peptide.
The AN-1792 trials have already shown that a little knowledge can be dangerous.
Scientific objectivity, caution, and patient health must surely be our paramount
concerns in the race to find a cure for Alzheimer's disease. The research community
has a responsibility to ensure that these concerns are addressed. The most effective
way of ensuring this outcome would be to place a moratorium on human trials
of ALL anti-amyloid vaccines and drugs until:
1) the reasons for, and nature of, the side-effects of AN-1792 are properly
understood and subjected to scrutiny in peer reviewed journals;
2) extensive investigations have been conducted on the effects of anti-amyloid
drugs in species other than mice, and in old primates.
I do not expect this proposal to receive strong support from the research community.
Hype about a new miracle vaccine will soon cloud our collective judgement, and
the lessons that we should have learned will soon be forgotten." - Stephen
R. Robinson, Monash University, Australia
References:
Check, E. (2002) Nerve inflammation halts trial for Alzheimer's drug. Nature,
415, 462. Pubmed
record
Hertz, L., Dringen, R., Schousboe, A., and Robinson, S.R. (1999) Astrocytes:
Glutamate producers for neurons. J. Neurosci. Res. 57, 417-428. Abstract
Robinson, S.R. (2001) Changes in the cellular distribution of glutamine synthetase
in Alzheimer's disease. J. Neurosci. Res. 66, 972-980. Abstract
Robinson, S.R. and Bishop, G.M. (2002) Amyloid beta as a bioflocculant: Implications
for the amyloid hypothesis of Alzheimer's disease. Neurobiol Aging 23, (In press).
|
Home
The Alzheimer's Vaccination Story, Continued
