ducanumab, Biogen's candidate for Alzheimer immunotherapy, is back.
The pharma giant says they've re-analyzed their data and discovered that their futility analysis was wrong, so they're planning to submit an NDA. Here's what that means to Alzheimer researchers and to patients.
Their cognitive testing results, available in their Third Quarter 2019 presentation, claim a 23% reduction in their primary end point (CDR-SB) at week 78. This, they say, is statistically significant with N=547 treated subjects and 548 placebos, p=0.010.
| Cognitive test | Low dose | p-value | High dose | p-value | |
|---|---|---|---|---|---|
| CDR-SB | −14% | 0.117 | −23% | 0.010 | |
| MMSE | + 3% | 0.690 | −15% | 0.062 | |
| ADAS-Cog13 | −14% | 0.167 | −27% | 0.010 | |
| ADCS-ADL-MCI | −16% | 0.156 | −40% | 0.001 |
CDR-SB means Clinical Dementia Rating Sum of Boxes. This is a scale from 0, which means normal, to 18, which means severe dementia. An Alzheimer patient typically progresses at around 1.4 points per year. The progression is nonlinear and follows a sigmoidal shape.
They also looked at three other cognitive tests. MMSE had no effect, while ADAS-Cog and ADCS-ADL had a statistically significant effect. The numbers differ because the tests are not very precise.
These numbers mean that the high dose produces between 15 and 40 percent slower decrease in cognitive function over 1.5 years, or a 10–26% slower rate of decline per year. Naturally, we'll need the actual numbers before we can make any sense of this. If true, aducanumab would be the first drug ever to affect the course of the disease. There is little expectation that it would ever reverse the dementia, but if the FDA approves it, it will become the principal treatment for AD.
What does this mean for patients
Patients will have to be genotyped for apolipoprotein E (ApoE), because a small number of patients with the ε4 allele experience a side effect called ARIA-E, which is a buildup of fluid in the brain caused by weakening of the blood-brain barrier. Patients with ApoE4 will be put on a lower dose.
Neurologists will probably continue to prescribe the four existing treatments for AD (donepezil, galantamine, rivastigmine, and memantine) in conjunction with aducanumab. These treatments are purely symptomatic and do not affect disease progression.
The FDA will be under enormous pressure to approve it, but they will also be aware that doing so will dampen any corporate research into alternative models of the disease. They will also be concerned that higher doses might induce ARIA-E. Thus, there is a good chance that the FDA might require yet another Phase III trial to verify that the higher dose is safe and the numbers are real.
Repeating a trial is common when the effect only shows up in a post-hoc analysis, which is what happened here. Post-hoc analyses need extra scrutiny because once the results have been collected it is usually possible to re-partition the subjects into new groups to make the results statistically significant.
The benefit to patients, if the drug really works, will be to increase the survival time by about 20 percent. The patient will still die from Alzheimer's disease. Like most antibody drugs, this one is likely to be quite expensive.
Other therapies: hormone replacement therapy (HRT) and NSAIDs
There are twice as many women as men with AD, and women have a higher incidence of AD in old age. Estrogen replacement therapy is said to decrease the risk (in women) by 6% per year, if it is started within five years of menopause. If the patient waits more than five years, HRT actually increases the risk.
Men also have estradiol in their brains. It is made there from testosterone and both molecules have a function in the brain unrelated to sex. Even so, AD correlates with lower brain estradiol only in women, while AD correlates with lower brain testosterone only in men. Why this is so remains a mystery.
HRT may work by raising brain levels of BDNF, a protein that induces the growth of new synapses. There is evidence that HRT is of little benefit in ApoE4 carriers. Thus, your doctor will want to have you genotyped if you plan on getting HRT.
NSAIDs have a small effect in delaying the onset of AD, but only in patients with ApoE4. Whether this effect is real is not clear.
There are many other theories. One is that depression, which is seen in AD patients, might be a risk factor and not an effect of the disease as previously thought. This theory has yet to be proven.
And it may be small consolation, but if you get AD, you're 32% less likely to be diagnosed with cancer.
What does this mean for researchers
The beta-amyloid (Aβ) theory is strongly debated. Although there are strong reasons to think Aβ is involved in AD, there is overwhelming evidence that Aβ is not the sole cause of AD, for the following reasons:
- The time course of Aβ in various brain regions doesn't match the time course of dementia. AD starts in the entorhinal cortex and spreads to the hippocampus and then to the cerebral cortex, while Aβ starts in the cerebral cortex and spreads to the hippocampus and then to the striatum and brainstem. AD pathology correlates reasonably well with tau inclusions, but it is not clear whether tau is a cause of AD or an effect of axonal degeneration.
- Other treatments, including BACE1 inhibitors (such as verubecestat) and gamma-secretase inhibitors (such as semagacestat), effectively block production of Aβ in vivo, but have no effect on AD. Many of them make the disease worse.
- None of the 7 other antibody-based passive immunotherapies has worked despite more than 12 phase 3 trials involving (last time I added them up) 18,563 subjects.
What might be different about aducanumab
There are three types of Aβ: monomers, which are harmless; plaques, which are mostly harmless, or at least don't correlate with memory loss; and oligomers, which are small clumps of Aβ that are still soluble. Oligomers are neurotoxic above 200 nM. This is a fairly high concentration for a protein. Unlike most other anti-Aβ antibodies, Aducanumab binds preferentially to oligomers.
The requirement for high doses is typically a red flag that a drug is acting non-specifically. Aducanumab binds to residues 3–7 at the N-terminal end of Aβ when Aβ is in an oligomer and in an extended conformation. The amino acid sequence of this region is EFRHD, which is a unique sequence in humans, but lots of bacteria proteins (in streptomyces, corynebacterium, and pseudomonas species of soil bacteria) also contain this sequence.
The need for high doses could mean that there is a new form of Aβ that is stuck somewhere that the antibody can't reach it effectively. Or it could mean that Aducanumab is actually doing something else in addition to binding to Aβ.
Thus, even if Biogen's data hold up, there is still a pressing need to understand what is going on. Aducanumab doesn't cure AD, and keeping a patient in a state of severe dementia alive indefinitely is not a desirable outcome.
There's a growing body of data that suggests that generalized activation of the innate immune system in the brain, possibly triggered by a bacterium, virus, toxin, or Aβ itself, is the culprit. The two strongest risk factors for AD (besides age) are TREM2, which is a pattern recognition receptor on the microglia, and apolipoprotein E, which is a cholesterol carrying protein. These two proteins bind to each other in a complicated biochemical pathway that produces neuroinflammation, which we know can destroy synapses and kill neurons.
Biogen's data, if they're correct, means that there is still a role for Aβ somewhere in this pathway. But it is clear that Aβ cannot be acting alone. Other drugs that can prevent the disease from starting are still needed.
Let's hope Biogen is right. If they're not, the Aβ “bad protein” theory of AD could go on forever.
oct 30 2019, 4:29 am. last edited nov 11 2019, 9:35 am
