This post describes the results of an attempt to run a self-experiment using senolytic drug candidates derived from chemotherapeutics. Based on animal study data, these pharmaceuticals are potential means to selectively destroy some fraction of senescent cells in some tissues. This is of interest because the accumulation of senescent cells is one of the root causes of aging. The pharmaceuticals have been tested for the ability to kill senescent cells in mice, and tested in humans for other uses, but as of yet are only just now entering initial human trials to test their ability to replicate the destruction of senescent cells observed in mice. This particular effort at self-experimentation involved a combination of dasatinib, quercetin, and venetoclax, with an outline somewhat close to the suggested approach to self-experimentation that I published not so very long ago.
This is a description of a learning experience, rather than of any compelling or interesting data on the efficacy of presently available senolytics in humans. The points to take away are near entirely related to the utility (or lack thereof) of various approaches to gathering data, and our expectations regarding the size of effects in different age groups.
The only changes observed in the recorded data were temporary and related to the side-effects of the senolytics. The only relevant outcome of the test that might be related to removal of senescent cells should be treated as an anecdote, as it was neither anticipated nor recorded in any way. Read through the details, learn from this, and do better if considering an attempt.
In hindsight, the chief challenges were that (a) the subject wasn't old enough to reasonably expect sizable effects to emerge in cardiovascular measures, (b) the commercially available methods of obtaining cardiovascular data, popularized by the quantified self movement, are too unsteady to detect anything other than sizable effects, and (c) user error sabotaged the most potentially interesting of the tests. Even simple things are hard, it seems.
Test Schedule and Dosing
The test schedule and oral dosing of senolytics is described below. The amounts below are scaled to a 60kg human for convenience. You might look over the past outline on self-experimentation for a lengthy dose scaling discussion. Note the use of an initial low test dose - the idea being that if this produces a reaction, the self-experiment is aborted. Caution is the watchword.
- Day 1: Bloodwork and an epigenetic clock test.
- Day 1-14: Take baseline cardiovascular metrics, daily.
- Day 15: Test dose of 80mg venetoclax, 20mg dasatinib, 250mg quercetin.
- Day 16-22: 400mg venetoclax.
- Day 16: 100mg dastanib, 1200mg quercetin.
- Day 20: 100mg dastanib, 1200mg quercetin.
- Day 23: Bloodwork.
- Day 23-36: Repeat cardiovascular metrics, daily.
- Day 36: Bloodwork and an epigenetic clock test.
- Day 36-49: Repeat cardiovascular metrics, daily.
The subject for the self-experiment is in the 45-50 age range, BMI 21.4 at the outset, and BMI of 21.0 following the experiment. See the anecdotal notes for an explanation of that change.
- Blood pressure and heart rate using an Omron 10 armband device.
- Heart rate variability using a Polar H10 chest band device.
- Pulse wave velocity using the iHeart fingertip device.
Measurements were taken at a consistent time of day, using left and right arm for the Omron 10 and iHeart, and twice in succession for the Polar H10.
These tests were picked on the basis of (a) being easy to run, and (b) matching up with physiological changes in aging that have been tied to the growing presence of senescent cells, such as stiffening of blood vessels and consequent effects on the cardiovascular system. There are other possible tests that could be attempted, but most are expensive, require a physician, or both. See the past outline on self-experimentation or an earlier discussion of testing for more details on this front.
Heart Rate Variability
The Polar H10 system is well spoken of, but here the output was garbage: non-physical values, varying constantly. Thus heart rate variability was abandoned as an option fairly quickly.
Pulse Wave Velocity
Pulse wave velocity from a fingertip device is roughly consistent between two measures on left and right fingers taken at the same time, but jumps around over a large range from day to day. Reading the research literature on this class of metric and the devices used, it is expected that peripheral measurements will be less reliable and subject to more influences than core measurements, taken closer to the heart. Unfortunately, without access to specialized medical equipment and a second pair of hands, peripheral devices are all that is available.
For a 45-50 year old, the reference range for pulse wave velocity - measured centrally rather than peripherally - is 6.0-8.5 m/s or so. In that context, the following is an example set of data for left fingertip / right fingertip taken over consecutive days prior to the test, while attempting to keep lifestyle, position, and other factors consistent. As one can see, it is quite varied from day to day:
- Day 1: 8.08 / 8.39
- Day 2: 8.01 / 7.52
- Day 3: 7.95 / 7.7
- Day 4: 6.76 / 6.89
- Day 5: 8.29 / 8.42
- Day 6: 7.17 / 7.34
- Day 7: 8.15 / 7.99
- Day 8: 7.13 / 6.75
- Day 9: 8.47 / 8.14
- Day 10: 7.7 / 7.24
- Day 11: 7.35 / 6.84
- Day 12: 7.27 / 7.07
- Day 13: 7.19 / 7.87
- Day 14: 7.98 / 7.85
The data following the test is more of the same, with little difference. It would require a fairly sizable change, say much larger than the standard deviation of ~0.5 m/s, to be visible, and more than two weeks of measurements to have confidence in that result. No change of that size, or any size that could be more than just chance, was apparent.
Heart Rate and Blood Pressure
The Omron 10 is a solid device. Heart rate and blood pressure came out as follows, showing no meaningful change.
- Baseline: systolic BP 113 ± 6.0, diastolic BP 69 ± 4, heart rate 58 ± 3
- First repeat: systolic BP 109 ± 4.0, diastolic BP 66 ± 3, heart rate 60 ± 4
- Second repeat: systolic BP 111 ± 5.0, diastolic BP 67 ± 4, heart rate 60 ± 3
The interesting changes in bloodwork are noted below, which all take the form of an alteration following the dosage that is mostly recovered a month later. Other values remained more or less consistent - any change of less than 15% was ignored for the purposes of this list.
- Baseline: Apo B 66 mg/dL, Eosinophil Count (absolute) 0.072 x 10^3/μL, LDL cholesterol 83 mg/dL, Lp(a) 14 nmol/L, Lymphocyte Count (absolute) 1.576 x 10^3/μL, TSH 1.67 mIU/L, Total Cholesterol 159 mg/dL, White Blood Cell Count 4 x 10^3/μL
- First repeat: Apo B 53 mg/dL, Eosinophil Count (absolute) 0.03 x 10^3/μL, LDL cholesterol 56 mg/dL, Lp(a) 10 nmol/L, Lymphocyte Count (absolute) 1.086 x 10^3/μL, TSH 1.24 mIU/L, Total Cholesterol 136 mg/dL, White Blood Cell Count 3.3 x 10^3/μL
- Second repeat: Apo B 57 mg/dL, Eosinophil Count (absolute) 0.049 x 10^3/μL, LDL cholesterol 70 mg/dL, Lp(a) 14 nmol/L, Lymphocyte Count (absolute) 1.391 x 10^3/μL, TSH 1.45 mIU/L, Total Cholesterol 144 mg/dL, White Blood Cell Count 3.8 x 10^3/μL
The temporary cull of immune cells is entirely expected from the action of these chemotherapeutics. The drop in LDL cholesterol is not expected. The mechanisms by which LDL cholesterol is maintained at a given level and how that changes in response to circumstances and aging are in fact not well understood in detail. There is an interesting open access paper that covers the present state of knowledge. A possible hypothesis is that the change is the result of dramatic temporary alteration in gut microbiota due to the impact of dastinib and loss of immune cells. A more honest hypothesis for this particular case is a shrug; it is simply impossible to know. Nonetheless, it is interesting, albeit not relevant to the point of the exercise, to see a temporary effect on the same scale as that produced by some lesser statins.
The epigenetic clock test failed. A few weeks after sending off the blood sample, the company responded to say there wasn't enough blood in the sample to run the tests. They were kind enough to supply a new kit with updated instructions - which are the same as the old instructions, except that the part discussing blood now heavily emphasizes that more is better than less. Unfortunately, by that point it was too late. This is very aggravating, as with the benefit of hindsight, the epigenetic clock was probably the most useful item in the lineup of tests.
Anecdotal Experiences with the Chemotherapeutics
Quercetin is innocuous. The low test doses of venetoclax and dasatinib produced no evident effects at all. At the higher dose, venetoclax results in a subtle running down of energy and wellbeing; at the end of a week, one feels worn. At the higher dose, dasatinib has the sort of immediate outcomes one might expect from an assault on the microbial population of the gut, such as due to stomach flu or a heavy dose of indiscriminate antibiotics. That lasts half a day to a day. Anti-nausea medication is recommended. The effects of dastinib mean that weight will likely be lost between the start and end of the dosage schedule.
One of the items that has been reported, anecdotally, by a couple of the self-experimenters using dasatinib is the improvement of skin lesions, troublesome scars, and the like. This happened here, for an patch of skin that failed to heal correctly after an injury years ago, and that as a result has been persistently annoying and painful since then. It improved notably within a week following the end of the dosage; not completely, but more than enough to feel that something useful was achieved at the end of the day. It has remained thus improved in the time since.
While we can speculate as to mechanisms involving senescent cells in troublesome lesions, one should of course assign this report exactly the same weight as the others, which is to say zero. It wasn't the aim of the experiment to look at results there, nothing was measured or recorded, and you have only the assertion in this post to go by. Nonetheless, I think that someone should consider running a formal study on the varieties of non-healing wounds in older individuals with the presently available senolytics; even absent anecdotal reports, there is a decent theoretical foundation for thinking that it might be beneficial.
The primary conclusion is that the measurement devices available to consumers for cardiovascular metrics are not as useful as hoped. Greater and longer experimentation with measurement techniques prior to running a self-experiment would have been most helpful in this case. Realizations on the usefulness of various measurement techniques could have been reached beforehand, and the strategy adapted in response. My current thinking on cardiovascular measures is that they might be useful for older individuals, where we could expect to observe larger changes, but they are not useful in the 45-50 age range, and the commercial tools are largely not up to the task at this time.
Secondly, the effects of current senolytic pharmaceuticals for people in the younger age range, just starting to show visible signs of aging, at doses equivalent to the mouse studies in which 25-50% of senescent cells were removed in multiple tissues, are modest in size. They are not large enough to be detected using standard approaches to measurement of cardiovascular and blood chemistry. If using better and more reliable medical devices, the story might be different, but even there I think, based on the results here, that the odds of observing changes would only be significant in people considerably older than 50, and much more impacted by cellular senescence. We shall see how the current trial of dasatinib at Betterhumans pans out, where the participants are in the 65+ age range.
If it were possible to start over, the approach would be adjusted to focus primarily on epigenetic age measures, or other protein expression tests such as that offered by AgeCurve, or some form of assay that actually measures counts of senescent cells. (Though the latter would be hard to obtain at this point, given that there is only the one useful test so far as I know, and it isn't yet commercially available to consumers). There are considerable advantages to tests in which you take a sample and send it off - assuming that one is actually competent to accomplish that simple task, which seems in doubt at this end of the world. All of the complexity and sophistication of the test is baked in, and requires little to no effort on the part of the self-experimenter.