PIEZO1 and the mechanism of the long circulatory longevity of human red blood cells

Human red blood cells (RBCs) have a circulatory lifespan of about four months. Under constant oxidative and mechanical stress, but devoid of organelles and deprived of biosynthetic capacity for protein renewal, RBCs undergo substantial homeostatic changes, progressive densification followed by late density reversal among others, changes assumed to have been harnessed by evolution to sustain the rheological competence of the RBCs for as long as possible. The unknown mechanisms by which this is achieved are the subject of this investigation. Each RBC traverses capillaries between 1000 and 2000 times per day, roughly one transit per minute. A dedicated Lifespan model of RBC homeostasis was developed as an extension of the RCM introduced in the previous paper to explore the cumulative patterns predicted for repetitive capillary transits over a standardized lifespan period of 120 days, using experimental data to constrain the range of acceptable model outcomes. Capillary transits were simulated by periods of elevated cell/medium volume ratios and by transient deformation-induced permeability changes attributed to PIEZO1 channel mediation as outlined in the previous paper. The first unexpected finding was that quantal density changes generated during single capillary transits cease accumulating after a few days and cannot account for the observed progressive densification of RBCs on their own, thus ruling out the quantal hypothesis. The second unexpected finding was that the documented patterns of RBC densification and late reversal could only be emulated by the implementation of a strict time-course of decay in the activities of the calcium and Na/K pumps, suggestive of a selective mechanism enabling the extended longevity of RBCs. The densification pattern over most of the circulatory lifespan was determined by calcium pump decay whereas late density reversal was shaped by the pattern of Na/K pump decay. A third finding was that both quantal changes and pump-decay regimes were necessary to account for the documented lifespan pattern, neither sufficient on their own. A fourth new finding revealed that RBCs exposed to levels of PIEZO1-medited calcium permeation above certain thresholds in the circulation could develop a pattern of early or late hyperdense collapse followed by delayed density reversal. When tested over much reduced lifespan periods the results reproduced the known circulatory fate of irreversible sickle cells, the cell subpopulation responsible for vaso-occlusion and for most of the clinical manifestations of sickle cell disease. Analysis of the results provided an insightful new understanding of the mechanisms driving the changes in RBC homeostasis during circulatory aging in health and disease.


Reviewer #1.
The manuscript proposed a very interesting question about how red cell volume changed in response to Piezo1 activation during capillary transit. In this case, computational approach was used to simulate red cell volume upon piezo1 activation and demonstrated that cell volume increased following by shrinkage during capillary transit and the magnitude of such volume change was small.

Reply:
We appreciate the deft account of the question and the answer in the first paper! Comments:

1) The kinetic of such process is not clear. For example, start from time 0 where red cells enter capillary, how long it takes to activate piezo1 channel and what is the timescale in other related signaling pathways that eventually lead to cell volume change (figure 4)? This kinetic analysis is critical to understand the question proposed by the authors.
Reply: As explained in the paragraphs headed "The reference protocol", "Ingress of a RBC in a capillary is simulated with an instant CVF transition from 0.00001 to 0.9 together with PIEZO1 activation.". The time scale of the pathways leading to volume changes is illustrated in Figs 3 and 6. These show almost instant onset to peak, peak durations of between a few ms to about ½ s, and durations of recovery phase varying greatly between a few ms to ~ 10s, all variations depending on the parameter values attributed to PIEZO1-mediated permeabilities, to PMCA Vmax, to JS-mediated cotransport rate, to Gardos channel PKmax or to combinations thereof. The mechanisms involved are discussed and analysed in detail in the text, and illustrated in the diagram of

Second paper, Lifespan
In terms of the second manuscript entitled "PIEZO1 and the mechanism of the long circulatory longevity of human red blood cells", the description was comprehensive but it would be better if it could be more focused. For example, it would be interesting to focus on quantitatively how decay in different channel activities that regulated cell volume contribute to red cell longevity and what were the relative roles of each channel in this process.

Reply:
We assume the Reviewer meant "decay in the different pump activities". The quantitative effects of the parameters controlling the longevity pattern of RBCs are shown in Fig 5. Curves 1, 2 and 5 show the effects of varying the decay rate of the PMCA, and curves 1, 3 and 4 illustrate the effects of variations in the onset time of Na/K pump decay and on its rate of decay. These results show how relatively minor parameter differences have powerful quantitative effects. It is hard to decide between "comprehensive" and "focus" bypassing details. Still, we tried by editing the text in the Abstract to focus on the specific effects of pump decay, as follows: "The first unexpected finding was that quantal density changes generated during single capillary transits cease accumulating after a few days and cannot account for the observed progressive densification of RBCs on their own, thus ruling out the quantal hypothesis. The second unexpected finding was that the documented patterns of RBC densification and late reversal could only be emulated by the implementation of a strict time-course of decay in the activities of the calcium and Na/K pumps, suggestive of a selective mechanism enabling the extended longevity of RBCs. The densification pattern over most of the circulatory lifespan was determined by calcium pump decay whereas late density reversal was shaped by the pattern of Na/K pump decay. A third finding was that both quantal changes and pump-decay regimes were necessary to account for the documented lifespan pattern, neither sufficient on their own. A fourth new finding…"

Reviewer 2. Dr Anna Bogdanova
Dr Bogdanova's comments address issues in the Guide and in both manuscripts.

1.
I would like to thank the authors for taking time to develop a tool that is available for testing as an open-source software. I consider myself as a potential user of the software to try to design experiments of my own and try to predict their possible outcome. It is an interesting instrument that has to be tested. Experimental biophysics is in need of such modelling approaches if they prove valuable (it needs substantially more time to tell).

Reply:
We appreciate Dr Bogdanova's interest in testing the model and in recognizing the need of such tools in experimental biophysics. We also expect the new model applications to prove as valuable as the earlier ones did. The current applications were designed to explore the changes in RBC homeostasis during single capillary transits and throughout the circulatory lifespan of RBCs. The results essentially define our current knowledge of the circulatory behaviour of RBCs based on the body of early and recent information encoded in the model. Further refinements may be expected as additional information becomes available in the future.

2.
However, some statements, such as a definition of initial condition as "a collection of RBC clones in plasma-like medium" make researchers involved in actual experiments worry already now. What we see suggests that the cells are no "clones" at all, and are not equipped with the same number of channels and pumps. As a result, we are dealing with a collection of sub-populations of cells with different properties, that include, but are not reduced to cell age.

These sub-populations vary in abundance and maximal level of activity PIEZO1 channels, responsiveness to mechanical and chemical stressors and with different activity/abundance of PMCAs. We regularly observe "non-responders to stretch, making 10-40% of all cells. PMCA decay is not a continuous process as the enzyme, when undergoing cleavage by calpain, changes its activity in a "quantum" way, going from uncontrolled up-regulation in Vmax to its complete inactivation or loss, or incorporation into the inside-out vesicles (in cells of patients with sickle cell disease this process is very pronounces, as the author of this paper has successfully demonstrated a while ago). How does the model address this process, would it change a lot in the outcome?
Reply: Thank you for pointing out the mention of "clones" in the Guide, a misleading description. We thought we had removed it from all texts, but obviously the "clones" in the Guide escaped our attention. We have now removed all mention of "clones" from the Guide. (Blood, 1995, 86, 334-341). These CVs combine differences from endowments "at birth" and from cell aging. In the model, on the other hand, experiments are simulated on a RBC defined by the constitutive properties specified when entering the initial conditions. When using the mean values of a normally distributed cell population for defining the constitutive properties of the modelled cell, predictions and experimental results were repeatedly shown to agree to within a 5 to 10% margin in all measured variables, as expected from the observed CVs, extensively referenced in the Guide.

The point Dr Bogdanova makes about RBC differences and about how our models deal with distributions and heterogeneities is important and applies to both papers. Let's consider first the case of a fresh RBC sample from a healthy adult. At the bench, experiments are performed on a cell population in suspension, with coefficients of variation (CVs) of between ~7% and ~13% in the relevant homeostatic variables
To study variability in more heterogeneous populations it is necessary to run multiple identical simulations on RBCs defined with the hypothesized deviations within the sets of parameters and variables considered in the model.  Fig 4D). In Blood, 2007; 110:1334-1342, it was additionally shown that the huge Vmax variation was due to PMCA decline with cell age. For the results in the current modelling papers, this is the only information that counts for the kinetic representation of the PMCA.

Whether the decline is due to glycation, cleavage by calpain or both is simply not addressed by the model. Discussing molecular mechanisms beyond model competence may convey the misleading impression that model results contribute information of value in that area.
The effects of PMCA and PIEZO1 variations on model outcomes was studied in detail as summed up below.

PMCA variations
Capillary transits (first paper). The effects of PMCA variations within the observed Vmax range are shown in Fig 6. Their interpretation and complex mechanism is explained in the text and illustrated in Fig 5. PMCA Vmax variations within the measured range were shown to have powerful effects on the volume recovery phase of the biphasic response following PIEZO1 activation, and minor effects on initial peak height.

Lifespan model (second paper). As explained in the fourth paragraph under Results
and Analysis, only an exponential decay function of the PMCA could deliver a right skew distribution, as measured (figure above). We have added a heading to the paragraph where this is reported in the paper to avoid confusion with the unrelated previous heading (Testing the quantal hypothesis). The new heading reads: "Analysis of pump-decay patterns." Unexpectedly, compliance with observed densification patterns could be implemented only with a very restricted set of pumpdecay parameter values (Fig 4, curves 1, 7 and 8; Fig 5, curves 1, 5 and 6).