We have another entry in the “Is Alzheimer’s caused by infectious disease?” drawing, and it’s a good one. A large multicenter team reports that Porphyromonas gingivalis, which is the key pathogen in gingivitis (gum disease) may be the actual causative agent in Alzheimer’s, which is a bold claim indeed. But they have several lines of evidence to back it up, and it’s a hypothesis that has to be taken seriously.
This particular idea has been developing out there already. Here’s a 2016 study looking at Alzheimer’s patients with peridontitis and finding that infected patients had more rapid cognitive decline. And this 2015 paper reported that gum disease was associated with higher amyloid levels in the normal elderly as well. In 2017, it was shown that deliberate infection gum-disease bacteria caused amyloid pathology and cognitive decline in a transgenic mouse model. This new paper draws quite a few more connecting lines – here we go:
The bacteria secrete some virulence factors called gingipains, which are cysteine proteases. This paper’s immunoassays for these proteins show much greater reactivity in Alzheimer’s brain tissue as compared to controls (up to 96% of the AD samples, 39% of the controls). There is a strong correlation between gingipain levels and tau, and also between gingipain and ubiquitin, and there’s a physical localization of the gingipain antibody staining and tau tangles as well. In fact, the samples show a spectrum of all these proteins, with the Alzheimer’s samples invariably on the high side and the controls invariably on the low. The authors suggest that they are in fact seeing preclinical Alzheimer’s showing up in the controls, and that this could be a diagnostic marker all by itself. There were no such correlations with samples from patients with Huntington’s, Parkinson’s, or ALS. They also demonstrate the presence of P. gingivalis DNA in post-mortem Alzheimer’s brain samples (both hippocampus and cerebral cortex) and in the CSF of living patients who are believed to have the disease as well. These PCR readouts matched the immunoreactivity results.
Infecting tau-expressing cell cultures with P. gingivalis led to breakdown of the tau protein, apparently through the actions of those gingipain proteases. Incubation of the pure protein with the proteases confirmed this, with numerous cleavage sites, and some of the fragments generated have previously been detected in Alzheimer’s CSF and proposed as possible disease markers. Meanwhile, exposing SH-SY5Y neuronal cell cultures to gingipains led to cell aggregation and other morphological changes, but if the proteases were covalently inactivated with iodoacetamide beforehand, no changes were seen.
There’s more. The team members from Cortexyme, a biotech in the Bay Area, developed more gingipain inhibitors through a med-chem effort. Looking at some of their patents, these appear to be tetrafluorophenyl esters as covalent inhibitors – two compounds in particular (COR271 and COR286) are the subject of this paper. These compounds block the cell death of the SH-Sy5Y cells, whereas broad-spectrum antibiotics were ineffective. Lilly’s gamma-secretase inhibitor semagacestat also did nothing in this assay (which to be sure is pretty much how it performed in the clinic). Injecting mice with gingipains into the hippocampus led to neuronal degeneration, but this too could be abrogated by pretreatment with the small-molecule inhibitors.
Update: I’ve heard from a longtime colleague in the industry who believes that I’m being unfair to Lilly’s efforts on semagacestat. I take his point – it was an excellent compound, and that in an area where it’s very hard to get excellent compounds. Lilly’s chemists and preclinical scientists did great work in getting it to the clinic, but (unfortunately) it did nothing for Alzheimer’s patients themselves. . .a result that anyone in this AD field is all too familiar with.
The they tried infecting aged mice with oral exposure to the P. gingivalis bacteria. Six weeks of exposure led to brain infection in all the mice so treated, along with elevated amyloid levels. Infection with bacteria that had had their gingipains genetically deleted showed no such effects. The amyloid response, as they note, is consistent with the idea that it’s an antimicrobial peptide that is induced by infection, and indeed the bacteria themselves were shown to be sensitive to exposure to beta-amyloid protein in cell culture. But oral dosing of their inhibitor compounds significantly protected the mice from infection in that model. The amyloid-as-antibiotic hypothesis is quite appealing intellectually, as it ties the well-documented amyloid/Alzheimer’s connection to alternate explanations of the disease. It also brings in the immune system and inflammation components of the disease, as many have noted. An improved compound (COR388) is currently in human clinical trials. It has already made it through Phase I, and enrollment in Phase II is expected this year.
OK, this is a pretty solid paper. You can find some gaps in it – for example, there is some switching between tau and amyloid markers along the way (the authors do propose that the documented spread of tau pathology is consistent with the idea of neuron-to-neuron bacterial infection). But overall, this is impressive work, and it is the most thorough case I’ve yet seen for an infectious-disease component to Alzheimer’s and the identification of an actual Koch-postulate causative agent. I am very happy to hear that this hypothesis is being put to the test in the clinic, and I am similarly very eager to see what happens. This is the best shot at something new in Alzheimer’s in a long time.