This post is an update for an earlier report on a self-experiment with flagellin immunization, tested as an approach to adjust the balance of microbial populations in the aging gut microbiome in a favorable, more youthful direction. Commentary and data from the earlier report are repeated, with the addition of a new assessment of the gut microbiome taken two years after the end of the experiment. In summary, changes from this short and simple intervention were largely favorable, and largely sustained over this period of time.
Flagellin is the protein that makes up bacterial flagellae, and it is hypothesized that there is a sizable overlap between populations of gut microbes that possess flagellae and populations of gut microbes that are harmful rather than helpful. The harmful microbes are largely a problem because they contribute to chronic inflammation, while helpful microbes are largely beneficial due to the metabolites that they produce. The gut microbiome changes with age, shifting towards more harmful and fewer helpful microbes.
If the immune system can be roused to do a better job of eliminating the problem microbes, then perhaps this could lead to improved health. Flagellin immunization has been trialed in humans as a vaccine adjuvant, and shown to be safe in the small studies conducted to date. In recent years, researchers tested its ability to adjust the gut microbiome in mice, with favorable results. In 2020, I posted a potential study outline for a self-experiment in flagellin immunization as a prompt for discussion, and in 2021 I published a report from one adventurous self-experimenter who gave it a try.
Setting Expectations
The motivation for this self-experiment was curiosity: would human data be similar to the mouse data? After a couple of years, the results continue to be, on balance, positive. The mouse data doesn't cover this sort of time span, but it is worth noting that in killifish adjustments to the gut microbiome made by fecal microbiota transplant are lasting, at least on the relatively short scale of a killifish life span. The question of whether results from an intervention to change the gut microbiome will last is of course quite an important one! A useful, lasting intervention is a great deal more valuable than one that does not last. This is a self-experiment in which there is an unusually clear readout for the outcome of interest, in the form of the Viome gut microbiome assay. This is nonetheless a study population of one. The results should be taken as interesting, but not supportive of any particular conclusion beyond the desire to run a larger and more formal study.
Schedule for the Self-Experiment
The self-experiment ran for ten weeks. Weekly intramuscular injections of 10 μg flagellin in 0.5ml phosphate-buffered saline were used, with Viome gut microbiome assays performed (a) beforehand, (b) at 10 weeks, (c) at 8 months, (d) and finally at 28 months.
- Week 1: Viome gut microbiome assessment.
- Week 1: Intramuscular injection of 10 μg of flagellin.
- Week 2: Intramuscular injection of 10 μg of flagellin.
- Week 3: Intramuscular injection of 10 μg of flagellin.
- Week 4: Intramuscular injection of 10 μg of flagellin.
- Week 5: Intramuscular injection of 10 μg of flagellin.
- Week 6: Intramuscular injection of 10 μg of flagellin.
- Week 7: Intramuscular injection of 10 μg of flagellin.
- Week 8: Intramuscular injection of 10 μg of flagellin.
- Week 9: Intramuscular injection of 10 μg of flagellin.
- Week 10: Intramuscular injection of 10 μg of flagellin.
- Week 10: Viome gut microbiome assessment.
- Week 34: Viome gut microbiome assessment.
- Week 122: Viome gut microbiome assessment.
Subject Details
The subject for the self-experiment was in the 45-55 age range, healthy and without chronic conditions, with a BMI of ~22 throughout the duration of the experiment. Diet and exercise were described as "relatively consistent" across this time, including the six month and two year follow up assessments. I feel that one should always be relatively skeptical of that sort of claim, however, no matter how formal or informal the study.
Summary of Results
Viome does not provide raw data on species and prevalence of gut microbes and their biochemistry, but rather a set of scores derived from that raw data. The algorithm used isn't public, meaning that one can't really dispute any of their choices or the studies used to support those choices, unfortunately. The algorithm is, nonetheless, consistent between assays at different times, and so can be used as a point of comparison for the purposes of a self-experiment, at least.
Over the course of the self-experiment, Viome summary scores improved for Inflammatory Activity, Digestive Efficiency, Gut Lining Health, Protein Fermentation, and Gas Production. The summary scores declined for Metabolic Fitness and Active Microbial Diversity. The gains (largely bad scores transforming into good scores) were larger than the declines (bad scores becoming worse scores). Some of these areas of function are likely more important than others to health, though expect arguments over which and why. My bias would be to prioritize Inflammatory Activity and Gut Lining Health when it comes to interactions between the gut microbiome and the processes of aging, but this is certainly a viewpoint that could be challenged.
Viome Data - Overall Score
Gut Microbiome Health:
Before: 27
After: 43
Week 34: 49
Week 122: 42
Viome Data - Summary Scores
Inflammatory Activity (lower is better):
Before: 50
After: 45
Week 34: 28
Week 122: 31
Metabolic Fitness (higher is better):
Before: 25
After: 29
Week 34: 21
Week 122: 25
Digestive Efficency (higher is better):
Before: 0
After: 57
Week 34: 68
Week 122: 52
Gut Lining Health (higher is better):
Before: 12
After: 64
Week 34: 69
Week 122: 75
Protein Fermentation (lower is better):
Before: 87
After: 49
Week 34: 33
Week 122: 54
Gas Production (lower is better):
Before: 83
After: 48
Week 34: 35
Week 122: 35
Active Microbial Diversity (higher is better):
Before: 34
After: 15
Week 34: 15
Week 122: 5
Viome Data - Other Ratings
Ammonia Production Pathways
Before: Not Optimal
After: Average
Week 34: Good
Week 122: Not Optimal
Bile Acid Metabolism Pathways
Before: Average
After: Good
Week 34: Good
Week 122: Average
Biofilm, Chemotaxis, and Virulence Pathways
Before: Not Optimal
After: Not Optimal
Week 34: Good
Week 122: Not Optimal
Butyrate Production Pathways
Before: Average
After: Average
Week 34: Not Optimal
Week 122: Average
Flagellar Assembly Pathways
Before: Not Optimal
After: Not Optimal
Week 34: Average
Week 122: Average
LPS Biosynthesis Pathways
Before: Average
After: Average
Week 34: Average
Week 122: Average
Methane Gas Production Pathways
Before: Good
After: Not Optimal
Week 34: Good
Week 122: Average
Oxylate Metabolism Pathways
Before: Average
After: Not Optimal
Week 34: Not Optimal
Week 122: Not Optimal
Putrescine Production Pathways
Before: Not Optimal
After: Not Optimal
Week 34: Average
Week 122: Average
Salt Stress Pathways
Before: Average
After: Average
Week 34: Average
Week 122: Average
Sulfide Gas Production Pathways
Before: Not Optimal
After: Average
Week 34: Average
Week 122: Good
TMA Production Pathways
Before: Good
After: Good
Week 34: Good
Week 122: Average
Uric Acid Production Pathways
Before: Not Optimal
After: Not Optimal
Week 34: Not Optimal
Week 122: Good
Anecdotal Notes
The first few injections of flagellin produced a minor injection site reaction that lasted a few days: red and tender. That was reduced with each injection, and later injections produced no reaction. Beyond that, no perceptible change in health or digestion, positive or negative, was observed as a result of the self-experiment.
Conclusion
Coupled with the animal data, and the existing human trial data for safety, the results here suggests that someone should run a formal, controlled trial of flagellin immunization in older people, 65 and over. The goal would be to see whether (a) this sort of outcome holds up in a larger group of people, and (b) there is a meaningful impact on chronic inflammation and other parameters of health that are known to be affected by the aging of the gut microbiome.
The most interesting part of the data is perhaps the decline in microbial diversity, when considered against the gains elsewhere. Microbial diversity correlates with better health in epidemiological studies, but there isn't a good mechanistic understanding as to why this is the case, or what factors provoke diversity versus a lack of diversity.