The $10 billion unicorn in INmune Bio: Leqembi and Kisunla’s MoA is microglia-related, and amyloid beta is a cytokine

Introduction

Amyloid beta has been the primary hallmark of Alzheimer’s disease for years. On the one hand, numerous trials focusing on amyloid beta have failed, but on the other hand, two anti-amyloid antibodies have had some modest success on cognitive decline recently.

Inflammation and glial function are more recent targets for the treatment of Alzheimer’s disease, and other neurodegenerative diseases.

The question then arises whether targeting inflammation and glial function is useful, given the afore-mentioned successes.

After first setting out what exactly amyloid beta in Alzheimer’s is and how it may be toxic in the framework of Alzheimer’s disease, the present blog explores the relationship between amyloid beta, inflammation and glial function. The reader will note that glial function plays a primary role in the mechanism of action of anti-amyloid antibodies, their side effect ARIA, and that amyloid beta triggers microglia to move to a pro-inflammatory state.

Part A: what is amyloid beta or Aβ?

There are several diseases with problematic amyloid

Alzheimer’s disease is probably the best known disease of which ‘amyloid’ is the main characteristic, but it is not the only disease in which this protein plays a role. There are numerous other diseases in which forms of amyloid appear to be toxic, such as cerebral amyloid angiopathy and different types of amyloidosis.

And in fact, not amyloid is considered to be the original problem, but amyloid beta. So what is amyloid beta?

Structure of beta amyloid

Amyloid beta or Aβ is a product of the cellular metabolism. It is derived from the amyloid precursor protein or APP, and synthesized in the endoplasmic reticulum, transported to the Golgi complex, and finally transported to the plasma membrane. At the plasma membrane, it is cleaved by β-secretase and γ-secretase to generate Aβ. Aβ is either released to the extracellular space or remains associated with the plasma membrane.

Amyloid fibrils are self-assembled, fibrillar structures made from proteins that fold into an alternative, β-rich form. For many proteins, the most stable conformation in physiological conditions is the native state, allowing the protein to refold into its native state should it be unfolded. But proteins can take on different structural conformations, some of which allow for assembly into fibrillar structures known as fibrils. This is the case for amyloid. It is believed that the ability to form amyloid fibrils is a generic property of proteins because many unrelated proteins can also make amyloid fibrils.

Function

The normal function of Aβ is not fully understood to date. Aβ is known to have roles in kinase activation, protection from metal oxidation damage, cholesterol transport and ion channels. APP is a highly conserved protein that is expressed in almost all mammalian tissues, where the highest levels appear in the brain and kidney.

Among these and other possible biological roles for Aβ, one compelling (and potentially unifying) idea is that Aβ may be an immune modulator. Indeed, evidence suggests that Aβ may activate pro-inflammatory signaling pathways in multiple CNS cell types.

Amyloid plaques

Aβ forms amyloid fibrils by folding from the native random-coil rich state to a α-helical rich intermediate, and finally to a β-sheet rich amyloid monomer that self-assembles into the fibrils.

Amyloid fibrils can accumulate to form deposits or plaques, and can do so in various organs in the body. When they are most seen in the brain and coincide with cognitive decline, Alzheimer’s disease is often the diagnosis. The mechanism of action of amyloid-related diseases in the periphery and the central nervous system is not identical, however.

Amyloid plaques in Alzheimer’s disease first appear in parts of the isocortex, and spread throughout the brain in later stages.

This is a timeline of the +100-year history of Aβ in Alzheimer’s disease.

The picture below shows where in the cell amyloid originates, how it aggregates into monomers, oligomers, protofibrils, fibrils and then plaques, and how fibrils and plaques are thought to interact with the (inflammatory) immune response in the CNS, and with other cell functions.

The most cytotoxic species from amyloid fibril formation pathway is generally believed to be the soluble oligomeric state, as the oligomers have been shown to be more cytotoxic than the mature fibrils for many proteins. These soluble species can permeate cell membranes and cause an influx of calcium ions into the cytosol from the surroundings and the endoplasmic reticulum of the cell, this eventually results in cell death. This effect is believed to be a generic, nonspecific mechanism for the cytotoxicity for all oligomeric amyloids.

Alternative theory: is (some) beta-amyloid protective?

One camp of scientists, led by Alberto Espay, is positing a theory following which (some species of) beta-amyloid protects the brain. The multitude of failures and the fact that g-secretase and b-secretase inhibitors, which prevent beta-amyloid from being produced, actually appeared to worsen cognitive performance, even in cognitively unimpaired participants at risk of developing Alzheimer’s disease, are in support of that theory.

Perhaps the idea needs to be that treatments for Alzheimer’s disease may need to increase the amount of this “good” beta-amyloid in the brain.

From a timeline perspective, the relationship between amyloid accumulation in the brain and dementia risk is also not well established. Whereas amyloid accumulation goes up significantly from the age of 60, the incidence of dementia only increases considerably from the age of 85.

Amyloid has already reached very high levels far before the onset of symptoms of dementia. Some Alzheimer's disease biomarker studies found amyloid changes 20 years or more in advance of expected symptoms, while cognitive changes lagged for more than a decade. Others established that it took 6.4 years to transition from amyloid negative to positive, and another 13.9 years to the onset of MCI. Even more, Aβ deposition rates began to slow only 3.8 years after reaching the positivity threshold, meaning about 10 years before the onset of MCI.

Leqembi (lecanemab) and Kisunla (donanemab)

With Leqembi and Kisunla now approved, at least in the United States with more hesitation abroad given the safety profile of these treatments, the clearance of Aβ deposits has finally proven to have some efficacy, though not noticeable by patients and caregivers.

Leqembi and Kisunla are antibodies that remove amyloid beta from the brain.

Leqembi targets small chains of beta-amyloid proteins called protofibrils. It targets amyloid plaques indirectly by preventing these protofibrils from merging.

Kisunla only targets the beta-amyloid plaques. 

As antibodies, Leqembi and Kisunla count on the body’s immune system to do the rest of the work. Concretely, the anti-amyloid antibodies are supposed to stick to beta-amyloid, and flag to the brain’s immune cells, the microglia, to eat and destroy beta-amyloid. 

Anti-amyloid antibodies are thought to remove amyloid beta by two mechanism of action: microglia-mediated phagocytosis and transport of amyloid out of the brain. Microglia-mediated phagocytosis is a process by which the brain’s immune system, particularly microglia, bind to amyloid to engulf and degraded it. Amyloid clearance from the brain is further facilitated by the glymphatic system along perivascular pathways, and by transporting amyloid across the blood-brain barrier through transporters like LRP1.

ARIA is caused by glial cells

A well-known side effect of anti-amyloid antibodies, ARIA or amyloid-related imaging abnormalities: brain swelling or brain hemorrhage), affects 41% of treated patients, and is life-threatening at times. The incidence and seriousness of ARIA as seen in patients treated with Leqembi and Kisunla have led to criticism and refusal of approval of Leqembi and Kisunla in territories outside of the US.

ARIA is considered to be caused by glial cells, as they are being activated and move to a higher pro-inflammatory state caused by the anti-amyloid antibodies.

Other unsuccessful approaches to target amyloid beta

Other than targeting amyloid plaque formation itself, other possible approaches to combating amyloid diseases that have been unsuccessful are modifying the production of the APP protein, amyloid vaccines, inhibition of β-secretase or γ-secretase, or increase the activity of α-secretase to reduce the amount of Aβ.

The probably (very) relative relevance of the Aβ42/Aβ40 ratio

Amyloid-beta is not a single molecule and does not come in one conformation. It is a collection of peptide fragments of different lengths, all derived from the amyloid precursor protein (APP). These fragments vary in size, typically between 37 and 49 amino acids. Aβ42 and Aβ40 are the most studied.

In several studies, the Aβ42/Aβ40 ratio is reported as an outcome. This is based on the following observation.

Aβ42 is more prone to aggregation than other amyloid beta peptides. That makes it more likely to form oligomers and fibrils. Insofar as these oligomers and fibrils are considered toxic, then Aβ42 is considered to contribute to such neurotoxicity. 

Because Aβ40 is considered to be less aggregation-prone and more abundant, compared to Aβ42 which is considered to be more prone to aggregation and therefore more neurotoxic, the Aβ42/Aβ40 ratio is often used as a diagnostic marker as well as a market to detect any treatment effect. A reduced ratio in CSF would then suggest Alzheimer's pathology.

However, amyloid peptides like Aβ37, Aβ38, Aβ43, and longer forms up to Aβ49 also exist. Some of these have been suggested to either promote or inhibit aggregation, influencing disease progression. 

The Aβ42/Aβ40 ratio is therefore, in my opinion, a human invention that is very much embedded in the entire historical idea that amyloid is the primary culprit of Alzheimer’s disease. It may be very relative because other conformations of amyloid may also be neurotoxic and more or less abundant, in any given individual, without those being measured.

We know from earlier blog posts that the amyloid hypothesis of Alzheimer’s disease has had an unseen amount of failures. It also has several flaws, the most prominent of which being that a high percentage of non-dementing people have high levels of amyloid (for further info, see here or here). Furthermore, removing amyloid betayields an all-in-all small effect, even if one does it only in people with dementia who have amyloid PET positivity (i.e. the recipe for trial success of anti-amyloid trials – see this blog post).

In view of all of the above, if even useful at all, I think looking at the Aβ42/Aβ40 ratio may be of very relative value.

Part B. Neuroinflammation and glial cells

On cytokines

Cytokines are messenger proteins. They can be anti-inflammatory and pro-inflammatory. In the framework of the central nervous system, they are considered of vital importance for the initiation, maintenance, and regulation of the immune response, and for cell-to-cell communication. They enable long-range intercellular communication, allowing for the global regulation of inflammation rather than just localized effects.

Pro-inflammatory cytokines Pro-inflammatory cytokines are a group of signaling molecules primarily produced by immune cells that promote inflammation and contribute to the immune response. The well-known proinflammatory cytokines are IL-1β, IL-6, IL-12, IL-23 and IL-17.

They also all interact with each other. IL-1β, for example, exacerbates neurodegeneration by recruiting and activating other immune cells. The upregulation of the proinflammatory cytokine IL-1β has been reported as an early indicator of various neuropathogenic diseases, including Alzheimer’s disease.

Anti-Inflammatory cytokines Another group of cytokines possesses anti-inflammatory properties. The well-known ones are IL-2, IL-3, IL-33, and IL-35. IL-3, for example, prevents cell death. IL-3 modulates microglia activity by catalyzing the microglial clearance mechanism. IL-35 regulates the production of anti-inflammatory regulatory T-cells and regulatory B-cells. IL-2 promotes Treg generation, survival, and activity, but also plays a role in promoting the generation and proliferation of their counterpart, pro-inflammatory effector T cells.

During dysfunction, IL-2 recruits astrocytes around Aβ plaques and activates them via the JAK/STAT3 pathway. The activated astrocytes protect the neurons by forming a physical barrier around plaques, reducing Aβ deposition by degrading and internalizing amyloid.

Chronic neuroinflammation

Chronic neuroinflammation is a chronic state in which a higher than normal amount of pro-inflammatory cytokines is found in the central nervous system. Neuroinflammation plays a vital role in the neuropathological changes in Alzheimer’s disease. It is mainly attributed to activated microglia and astrocytes, which produce numerous pro-inflammatory cytokines.

Astrocytes are the most abundant, make up 40-50% of all glial cells, and about 30-40% of all the brain cells.

Microglia are the resident macrophages within the CNS and comprise around 10–15% of all the cells in the brain. 

As the most abundant glial cells in the CNS, astrocytes play an essential role in the communication with neurons and regulation of synapse formation and function. Under pathological conditions, astrocytes become reactive, which are characterized by cell hypertrophy as well as the release of cytotoxins.

In a healthy adult brain, microglia are in a resting state and highly ramified morphology with small somas, maintaining the development and homeostasis of the central nervous system. However, when they recognize the insults of the central nervous system, they respond to the injury or invasion by a morphological change, resulting in cell enlargement and migration. 

TNF, and traditional TNF inhibitors

I discuss TNF separately as it is considered to be the master regulator of inflammation. TNF has been implicated in the progression of various diseases, including rheumatoid arthritis, Crohn’s disease, inflammatory bowel disease, psoriatic arthritis, ulcerative colitis, ankylosing spondylitis, type I diabetes, multiple sclerosis, uveitis, and atherosclerosis. Inhibiting TNF has been considered the most effective way to treat a number of inflammatory diseases.

Until 2024, AbbVie’s TNF inhibitor Humira was the top selling drug for six years straight for many of these indications. The global TNF inhibitor market was forecasted to reach $43 billion in 2025. In comparison, the Alzheimer’s market currently had a market size of $ 4.82 billion in 2023, and is expected to increase to $ 8.18 billion by 2032.

Traditional TNF inhibitors lead to up to 72% lower risk of developing Alzheimer’s risk

Traditional TNF inhibitors are too big to pass the blood-brain-barrier, and come with several side effects. However, they somehow do show some efficacy in preventing Alzheimer’s, as has been confirmed repeatedly in several meta analyses.

To quote some of these studies:

- A study of 8.5 million insured adults in the United States (US) reported increased risk for Alzheimer’s disease in patients with rheumatoid arthritis. It also reported that etanercept, a TNF inhibitor, was associated with reduction of risk for AD.

- A  large, retrospective case-control study from 56 million unique adult patients showed again that patients with rheumatoid arthritis, psoriasis, ankylosing spondylitis, inflammatory bowel disease, ulcerative colitis and Crohn’s disease had an increased risk for Alzheimer’s disease. The risk for Alzheimer’s disease in patients with rheumatoid arthritis and psoriasis was lower among patients treated with different TNF inhibitors.

-  A study of 2510 patients with rheumatoid arthritis reported an association of TNF inhibitor use and reduced dementia risk, consistent as the study period increased from 5 to 20 years after diagnosis. TNF inhibitor use showed a long-term effect in reducing the risk of Alzheimer’s disease during the 20 years of follow-up.

- Here, here and here are some further meta analyses.

And we’re not talking about any risk reduction here. One study reported that patients with rheumatoid arthritis have treated with etanercept, adalimumab, or infliximab had a 66%, 72%, and 48% lower risk of developing Alzheimer's disease, respectively, compared to those not treated with these TNF inhibitors. Another study mentioned a 30% risk reduction.

Even if one cannot compare apples to oranges, that seems much higher than any efficacy the anti-amyloid antibodies have ever shown. And this comes from patients with inflammation, probably the right target, knowing patients with inflammation generally also progress faster.

Various TNF-blocking agents including etanercept, infliximab and adalimumab have furthermore demonstrated the ability to reduce microgliosis, neuronal loss, tau tangles, and Aβ accumulation. Both pre-clinical and clinical studies have indicated their potential in improving cognitive function.

The evidence that TNF-inhibition, even non-selectively, impacts risk and therefore probably also progression of Alzheimer’s, is therefore too big to ignore. It is, in my opinion, so much clearer than the all-in-all muddy waters in which the amyloid hypothesis has been developed.

And again, this efficacy is seen in non-selective TNF inhibitors. They are too large to get into the brain. Any efficacy probably comes from immune crosstalk from the periphery to the brain, and may be negligeable compared to a drug that could directly enter the central nervous system and exert its function there. Even non-selectively, I believe a TNF inhibitor that could pass the blood-brain-barrier may have a significantly higher impact than the existing TNF inhibitors. XPro passes the blood-brain-barrier and is selective for soluble TNF.

That said, I expect the largest and most immediate efficacy to be tied to the fact that XPro allows remyelination, which is related to its selectivity for soluble TNF (see blog post here).

Transmembrane and soluble TNF in Alzheimer’s disease

TNF comes in two forms, a transmembrane form which sticks to the cell membrane and a soluble form cleaved from the cell membrane.

The soluble form promotes inflammation, signals mostly via the TNF type 1 receptor (TNFR1), and less via the TNF type 2 receptor (TNFR2). The transmembrane form is anti-inflammatory and signals mostly via TNFR2.

In Alzheimer’s disease, the binding of soluble TNF to TNFR1 triggers the NF-κB/MAPK signaling pathway, inducing inflammation, tissue degeneration, and cell death, which greatly contributes to the development of AD. TNF modulates the functionality of microglia and other glial cells by altering the way they function and express themselves, i.e. by changing their phenotype.

In the brain, the above-mentioned cytokines are secreted mostly by different immune cells, particularly by microglia.

Part C. Amyloid is a cytokine

Inflammation related to Leqembi and Kisunla may reduce efficacy

So far, the story of amyloid in Alzheimer’s disease could be read mostly as a story independent of that of inflammation in Alzheimer’s disease.

The only relationship there is so far relates to the anti-amyloid antibodies; they need microglia to engulf amyloid. In doing so, they move microglia to a pro-inflammatory state, even causing ARIA and death.

From that angle, anti-amyloid antibodies promote a pro-inflammatory state, which may by itself promote cognitive decline…which would then need to be countered by the efficacy related to the removal of amyloid.

Also, if one were to enable amyloid clearance without causing ARIA in patients whose dementia is amyloid-related, higher efficacy from anti-amyloid treatments could be seen. It is also possible that a combination therapy of a (selective) TNF inhibitor and amyloid antibodies may yield higher efficacy.

Amyloid acts as a cytokine

And this is where it gets really interesting. In fact, the theories collide to a certain extent.

The neurotoxic cytokine milieu that develops with aging occurs as a result of allostatic inflammatory load, i.e. the accumulation of inflammatory triggers over a lifetime. These may also include pathogens. That milieu causes chronic neuroinflammation and is caused by glial cells.

Some amyloid aggregates acts as pro-inflammatory triggers, and may in so doing exacerbate neurodegeneration. Scientists have proven this throughout several studies:

- As Alzheimer’s disease progresses, microglia continue to cluster around Aβ plaques, but their efforts to clear them become less effective, driving in more microglia.

- When exposed to Aβ, astrocytes activate certain genes that produce inflammatory molecules. This response increases levels of toxic Aβ in the brain, creating a cycle of inflammation. To compensate for the poor clearance of Aβ, immune cells from outside the brain, called macrophages, are recruited. However, their presence can further intensify inflammation, worsening Alzheimer’s disease.

- Microglia release inflammatory molecules like IL-1α, TNF, and C1q, which push astrocytes into a harmful, neurotoxic state. In turn, these reactive astrocytes further promote inflammation and nerve cell damage. Aβ itself also stimulates the release of inflammatory molecules from microglia, mimicking the behavior of a cytokine (a signaling protein involved in immune responses). Even tiny amounts of Aβ can work alongside existing inflammatory molecules to trigger a strong immune response in astrocytes.

- Microglia recognize Aβ in the brain and respond by increasing receptors for IL-3, a molecule produced by astrocytes. This IL-3 response enhances microglia’s ability to move, cluster around Aβ deposits, and clear them—although this process is not always sufficient to stop disease progression.

- Aβ can substitute for a key immune system protein (C1q) that normally activates astrocytes, driving them toward a harmful, inflammatory state.

- When glial cells (microglia and astrocytes) are exposed to the most toxic form of Aβ (Aβ1−42), they undergo changes that indicate immune activation. Furthermore, certain forms of Aβ interact with microglia through a receptor called TLR4, leading to the production of inflammatory molecules that contribute to nerve cell damage in Alzheimer’s disease.

- A mix of Aβ aggregates triggers a strong immune response with the release of large amounts of TNF. However, when Aβ monomers or fibrils do not provoke the same level of inflammation.

However, there are also many persons with high amyloid burden that do not have cognitive decline, so thinking that amyloid alone would cause cognitive decline is erroneous. It acts as a pro-inflammatory cytokine, but in those with cognitive decline, it is probably just one of several other pro-inflammatory triggers.

The above again proves that removing the inflammation, and having the glial cells return to their nurturing task and homeostasis, should be the primary goal of any treatment against inflammatory neurodegeneration.

Conclusion

Treating Alzheimer’s disease from the angle of removing amyloid beta remains a story with many flaws and nuances in my view. First of all, amyloid may not always be the trigger of the disease, and may certainly not be the only one. Amyloid beta refers to a number of amyloid aggregates, which do not seem to have a direct cellular function, and some of which may be protective, in any case not neurotoxic.

The Aβ42/Aβ40 ratio is a measure based on that understanding which is supposed to give an indication of disease progression, but as a human creation that neglects so many other conformations and potential triggers of neurodegeneration, it appears very fallible and related to the amyloid hypothesis to me.

Anti-amyloid antibodies rely in part on microglia to clear amyloid from the brain. In doing so, they promote inflammation, ARIA, and potentially ensuing neurodegeneration.

However, amyloid may not be the cause of dementia, as there are cognitively unimpaired individuals with high levels of amyloid in the brain. Also, the timeline of accumulation of amyloid in the brain does not coincide at all with the onset of cognitive impairment.

Microglia and astrocytes have difficulties clearing amyloid beta from the brain. Over time, they cluster around Aβ plaques and end up in a pro-inflammatory feedback loop. In that sense, amyloid beta acts as a cytokine, namely as a pro-inflammatory messenger protein clump.

The cognitive efficacy of Leqembi and Kisunla appears to be all-in-all limited, which makes sense as they only remove one inflammatory trigger, and promote inflammation at the same time. Higher efficacy may be obtained if one were to remove several inflammatory triggers at once; unfortunately it is unclear which inflammatory triggers each person has.

Lowering inflammation as a whole allows glial cells to return to their nurturing function and homeostasis. In fact, this has already been proven; several meta analyses in patients with inflammatory diseases have shown a reduced risk of developing dementia. That reduction is probably  around 50% on average, which seems to outperform efficacy of anti-amyloid antibodies. And these studies only relate to existing TNF inhibitors, which cannot pass the blood-brain-barrier, are non-selective, and do not allow remyelination.

XPro should outperform those by far.

If one understands all of that, INmune’s goal of stabilizing cognition may not be overstating XPro’s potential. If so, a dormant $5 billion market may be for the taking.

---

Link to INMB investors discord server: https://discord.gg/JEA8r7wCGY