Homeostasis and “Root Causes” in Aging

Let’s start with a stylized fact: al­most ev­ery cell type in the hu­man body is re­moved and re­placed on a reg­u­lar ba­sis. The fre­quency of this turnover ranges from a few days (for many im­mune cells and cells in the gas­troin­testi­nal lin­ing) to ten years (for fat, heart, and skele­ton cells). Only a hand­ful of tis­sues are be­lieved to be non-re­new­ing in hu­mans—e.g. eggs, neu­rons, and the lens of the eye (and even out of those, neu­rons are de­bat­able).

This means that the num­ber of cells of any given type is de­ter­mined by “home­o­static equil­ibrium”—the bal­ance of cell re­moval and re­place­ment. If an ul­cer de­stroys a bunch of cells in your stom­ach lin­ing, they’ll be re­placed over a few days, and the num­ber of stom­ach cells will re­turn to roughly the same equil­ibrium level as be­fore. If a healthy per­son re­ceives a bunch of ex­tra red blood cells in a trans­fu­sion, they’ll be bro­ken down over a few months, and the num­ber of blood cells will re­turn to roughly the same equil­ibrium level as be­fore.

As or­ganisms age, we see a change in the home­o­static equil­ibrium level of many differ­ent cell types (and other pa­ram­e­ters, like hor­mone and cy­tokine lev­els). In par­tic­u­lar, a wide va­ri­ety of symp­toms of ag­ing in­volve “de­ple­tion” (i.e. lower ob­served counts) of var­i­ous cell types.

How­ever, hu­man ag­ing hap­pens on a very slow timescale, i.e. decades. Most cell counts equil­ibrate much faster—for in­stance, im­mune cell counts equil­ibrate on a scale of days to weeks. So, sup­pose we see a de­crease in the count of cer­tain im­mune cells with age—e.g. naive T cells. Could it be that naive T cells just wear out and die off with age? No—T cells are re­placed ev­ery few weeks, so a change on a timescale of decades can­not be due to the cells them­selves dy­ing off. If the count of naive T cells falls on a timescale of decades, then ei­ther (a) the rate of new cell cre­ation has de­creased, or (b) the rate of old cell re­moval has in­creased (or both). Either of those would re­quire some “up­stream” change to cause the rate change.

More gen­er­ally: in or­der for cell counts, or chem­i­cal con­cen­tra­tions, or any other phys­iolog­i­cal pa­ram­e­ter to de­crease/​in­crease with age, at least one of the fol­low­ing must be true:

  • the timescale of turnover is on the or­der of decades (or longer)

  • rate of re­moval in­creases/​decreases

  • rate of cre­ation de­creases/​increases

If none of these is true, then any change is tem­po­rary—the cell count/​con­cen­tra­tion/​what­ever will re­turn to the same level as be­fore, de­ter­mined by the re­moval and cre­ation rates.

Of those three pos­si­bil­ities, no­tice that the sec­ond two—in­crease/​de­crease in pro­duc­tion/​re­moval rate—both im­ply some other up­stream cause. Some­thing else must have caused the rate change. Sooner or later, that chain of cause-and-effect needs to bot­tom out, and it can only bot­tom out in some­thing which equil­ibrates on a timescale of decades or longer. (Feed­back loops are pos­si­ble, but if all the com­po­nents equil­ibrate on a fast timescale then so will the loop.) Some­thing some­where in the sys­tem is out-of-equil­ibrium on a timescale of decades. We’ll call that thing (or things) a “root cause” of ag­ing. It’s some­thing which is not re­placed on a timescale faster than decades, and it ei­ther ac­cu­mu­lates or de­cu­mu­lates with age.

Now, the main crite­ria: a root cause of ag­ing can­not be a higher or lower value of any pa­ram­e­ter sub­ject to home­osta­sis on a faster timescale than ag­ing it­self. Ex­am­ples:

  • Most cell types turn over on timescales of days to months. “De­ple­tion” of any of these cell types can­not be a root cause of ag­ing; ei­ther their pro­duc­tion rate has de­creased or their re­moval rate has in­creased.

  • DNA dam­age (as op­posed to mu­ta­tion) is nor­mally re­paired on a timescale of hours—some­times much faster, de­pend­ing on type. “Ac­cu­mu­la­tion” of DNA dam­age can­not be a root cause of ag­ing; ei­ther the rate of new dam­age has in­creased or the re­pair rate has de­creased.

  • DNA mu­ta­tions can­not be re­paired; from a cell’s per­spec­tive, the origi­nal in­for­ma­tion is lost. So mu­ta­tions can ac­cu­mu­late in a non-equil­ibrium fash­ion, and are a plau­si­ble root cause un­der the home­osta­sis ar­gu­ment.

Note that the home­osta­sis ar­gu­ment does not mean the fac­tors ruled out above are not links in the causal chain. For in­stance, there’s quite a bit of ev­i­dence that DNA dam­age does in­crease with age, and that this has im­por­tant phys­iolog­i­cal effects. How­ever, there must be changes fur­ther up the causal chain—some other long-term change in the or­ganism’s state leads to faster pro­duc­tion or slower re­pair of DNA dam­age. Con­versely, the home­osta­sis ar­gu­ment does not im­ply that “plau­si­ble root causes” are the true root causes—for in­stance, al­though DNA mu­ta­tions could ac­cu­mu­late in prin­ci­ple, cells with cer­tain prob­le­matic mu­ta­tions are be­lieved to be cleared out by the im­mune sys­tem—so the num­ber of cells with these mu­ta­tions is in equil­ibrium on a fast timescale, and can­not be a root cause of ag­ing.

For any par­tic­u­lar fac­tor which changes with age, key ques­tions are:

  1. Is it sub­ject to home­osta­sis?

  2. If so, on what timescale does it turn over?

  3. If it is sub­ject to home­osta­sis on a timescale faster than ag­ing, then what are the pro­duc­tion and re­moval mechanisms, and what changes the pro­duc­tion and re­moval rates with age?

Th­ese de­ter­mine the ap­pli­ca­bil­ity of the home­osta­sis ar­gu­ment. Typ­i­cally, any­thing which can nor­mally be fixed/​re­placed/​un­done by the body will be ruled out as a root cause of ag­ing—the timescale of ag­ing is very long com­pared to prac­ti­cally all other phys­iolog­i­cal pro­cesses. We then fol­low the causal chain up­stream, in search of plau­si­ble root cause.