Retention of normal glia function by an isoform-selective protein kinase inhibitor drug candidate that modulates cytokine production and cognitive outcomes

Alzheimer’s disease (AD) is the most common form of dementia, affecting >5.3 million people in the USA alone [1]. Current treatments stabilize symptoms temporarily without slowing progression of the disease [2]. Therefore, there is a pressing need for neuroprotective agents for AD and related dementias.

An array of complementary clinical and preclinical reports implicate discrete aspects of early stage neuroinflammation in disease progression. For example, genome-wide and genetic investigations reveal associations between inflammation-related genes and AD risk (for recent reviews, see [3‐5]). Complementary to such associations are experimental outcomes from clinical pathology and animal model investigations that together provide a pathophysiology link between early stage neuroinflammation and AD risk (for reviews, see [6‐8]). A specific form of neuroinflammation and dysregulated proinflammatory cytokine production, was shown to be a viable drug discovery pathway if intervention was in early stages when basal cytokine levels were progressively increasing prior to or coincident with synaptic dysfunction [9].

The p38 mitogen-activated protein kinase (MAPK) family of serine-threonine protein kinases, especially p38αMAPK (MAPK14), are key regulators of proinflammatory cytokine production in the brain [10]. Activation of microglia p38αMAPK and the downstream overproduction of proinflammatory cytokines such as TNFα leads to synaptic protein loss, neurite degeneration, and neuronal death in microglia-neuron co-cultures [11]. Activation of p38αMAPK in neurons occurs in response to a variety of CNS disease-relevant stressors, and inhibition of neuronal p38αMAPK can be neuroprotective [12, 13]. Taken in its entirety, the increasing body of evidence is consistent with activation of p38αMAPK activity in both neurons and glia in response to disease-relevant stressors, raising the possibility that selective dosing with a brain-penetrant, isoform-selective p38αMAPK could be a viable approach to disease modification or attenuation of disease progression susceptibility.

The recent description [14] and initial validation of a highly selective p38αMAPK inhibitor, MW150, that is efficacious in two different AD models provided both a mechanistic precedent and a novel candidate for development. Previous efforts at validating p38αMAPK as a viable CNS target used mixed-kinase inhibitors whose results were difficult to interpret or lacked blood brain barrier penetrance and bioavailability to allow exposure to the brain-activated p38αMAPK [15]. The development of MW150 solved those problems. Briefly, MW150 is a unique small molecule drug candidate that is the most highly specific p38αMAPK inhibitor reported to date. MW150 is orally bioavailable, CNS-penetrant, and efficacious in rescuing hippocampal-dependent associative and spatial memory deficits in mouse models of AD-related pathology [14]. MW150 outcomes in a series of pharmacological screens forecast lower risk for development and documented the absence of safety problems that plagued previous CNS p38MAPK inhibitor campaigns. The battery of critical pharmacological and efficacy results qualified this unique, isoform-selective, p38αMAPK inhibitor as a candidate for investigational new drug (IND) development required for first-in-human clinical studies. There is no comparable drug or drug candidate with the combination of documented specificity and safety, in vitro and in vivo pharmacological features, and target recognition [14, 15]. Therefore, it is essential that the glial effects of this exceptional inhibitor with in vivo efficacy be addressed in order to anticipate future use in clinical investigations.

The goal of the current study was to determine in more depth the effects of MW150 on glial biology and proinflammatory cytokine dysregulation in the AD-relevant APP/PS1 knock-in (KI) mouse model previously used to show MW150 efficacy in suppression of hippocampal-dependent associative and spatial memory deficits [14]. To further probe the selective action of MW150, we complemented the in vivo studies by live cell imaging analysis of the murine microglia BV2 cell line to monitor proliferation, migration, and phagocytosis activities. Our results demonstrate that MW150 has a selective role in modulation of neuroinflammatory responses without pan-suppression of normal physiological functions of microglia.

There are two major findings in this study. First, treatment with an established efficacy dose of MW150 attenuated age-related increases in brain IL-1β and TNFα without affecting the overall amount of microglia or Aβ levels. A surprising finding was that MW150 treatment increased the microglia closely associated with amyloid plaques. Second, MW150 treatment did not suppress protective microglia cell responses such as migration, proliferation, or phagocytosis over concentrations that attenuate proinflammatory cytokine production. These key findings document the selective effect of efficacious doses of MW150 on disease-linked glia processes with retention of responses considered protective.

The results reported here extend our previous validation [14] of MW150 as a candidate for development by showing that MW150 in the efficacious dose range suppresses IL-1β and TNFα overproduction in vivo in the APP/PS1 KI mice. We did not detect a change in IL-6 levels, consistent with the prevailing view that IL-6 cytokine induction generally occurs via signaling pathways independent of p38αMAPK [26].

We previously reported [14] that MW150 showed no detectable effects on amyloid plaque burden in either the APP/PS1 KI mice or the APP/PS1 transgenic mice assessed by standard immunohistochemistry with 6E10 anti-Aβ antibody. We further extended the analyses here by using quantitative Aβ ELISA and confocal microscopy. Aβ ELISA demonstrated that MW150 had no detectable effect on the levels of PBS-soluble or formic acid-soluble Aβ40 and Aβ42 in the APP/PS1 KI mouse cortex, consistent with our previous study that reported no change in plaque load by immunohistochemistry [14]. Confocal microscopy analysis showed that MW150 treatment of APP/PS1 KI mice had no significant effect on the total Aβ plaque volume in the cortex and on the amount of internalized Aβ inside each microglia. However, even though the results did not reach significance, there was a trend toward decreased Aβ plaque volume and increased Aβ inside microglia in the APP/PS1 KI mice treated with MW150. These results are also consistent with a small increase in bioparticle phagocytosis activity seen in BV2 cells after MW150 treatment. Overall, our findings raise the intriguing possibility that MW150 stimulates microglia phagocytosis. However, these data must be interpreted with caution because the in vivo results with Aβ plaque volume did not reach statistical significance, and in vitro results in the BV2 cell line with bioparticle phagocytosis cannot be directly equated to Aβ phagocytosis in vivo. It will be important in future studies to test potential effects of MW150 on microglia phagocytosis in more detail. Nevertheless, it is clear that MW150 did not impair the ability of the plaque-associated microglia to phagocytose Aβ and did not lead to a significant change in Aβ40 or Aβ42 levels.

An unexpected finding in the results reported here was that MW150 treatment significantly increased the microglia in close proximity to amyloid plaques. Whether this reflects an effect of MW150 on microglia migration is unclear. There was no statistically significant effect of MW150 on migration of BV2 cells in response to a scratch wound, but similar studies with migration to amyloid deposits were not done because of limited responsiveness of BV2 cells to synthetic Aβ. The functional significance of the increase in plaque-associated microglia is also not known, but a logical hypothesis to test in future studies is whether there is a relationship between modulation of p38αMAPK signaling, and the ability of microglia processes to surround amyloid plaques and promote their compaction. A demonstrated linkage would provide mechanisms whereby microglia could decrease compaction (p38αMAPK mediated) or increase compaction (e.g., TREM2 mediated; [27]). The unexpected finding, therefore, raises the potential of subtle effects of p38αMAPK inhibition on microglia—plaque interactions that will require further exploration.

There is some general confusion in the literature about the pharmacodynamic effects of p38αMAPK inhibitor drug candidates. Specific to this report, there is an unresolved issue in terms of the relationship between p38αMAPK and amyloid pathology in preclinical mouse models of AD. Our results here and elsewhere [14] using multiple experimental approaches clearly show that the isoform-specific p38αMAPK inhibitor MW150 does not significantly affect overall Aβ levels or plaque burden. However, previous studies with some p38αMAPK inhibitors that hit multiple kinases, such as VX-745 and CNI-1493, report inhibition of Aβ production or amyloid plaque deposition [28‐30]. Whether the effects on these inhibitors on amyloid reflect engagement of targets other than p38αMAPK [31, 32], such as Abl [33], is not known. Further, a full evaluation of the differences is limited by the type of animal model used in efficacy and pharmacodynamic endpoint analyses. For example, APP/PS1 transgenic mouse models that overexpress APP exhibit high levels of amyloid, whereas the APP/PS1 KI mouse uses endogenous promoters and demonstrates progressive AD pathology without the forced APP overproduction. Regardless, future clinical studies are required to resolve any therapeutic significance of these preclinical pharmacodynamic differences.

In summary, our results demonstrate the selective effects on glial inflammatory responses to treatment with MW150, a unique, isoform-selective, p38αMAPK inhibitor drug candidate that attenuates cognitive impairment in AD-relevant mouse models. The findings show that MW150 has a selective role in modulation of neuroinflammatory responses without pan-suppression of normal physiological functions of microglia. The refined insight into glial biology presented here, combined with the safety and efficacy pharmacological profile of MW150 [14], reinforce the need to move this novel therapeutic candidate into clinical development for AD and related disorders.