Very interesting and, again, paints a dismal picture. Yates et al. mentions that immune activation is significantly reduced during antiretroviral therapy, at which point homeostatic compensation is presumed to take over the responsibility of CD4+ cell turnover--assuming that this turnover allows for a faster accumulation of viral drug resistance through sustained viral generation, I am wondering if suppressing homeostatic compensation during this period would extend the period of ART effectiveness, possibly preventing a viral rebound. I also had a question regarding the assumptions of the model that CD4+ cells replicate consistently--though Yates et al. mentions that these cells can decline in proliferation, I am wondering if the actual variance in replications per cell would significantly reduce the strength of the model's predictions.
So had some difficulty following along with this paper because of the vocabulary but here are a few things: Early on in the paper (page 949) they tell us how they estimate 'a,' the rate at which resting cells are signaled to begin dividing and say that 2 types of cells compete for these signals, the effector CD4+ T cells and central memory cells- and according to my understanding they just average published 'a' values for each type of cell. This is a critical number in the model- it tells us how often immune cells divide. Am I wrong, or is this a really critical step where you wouldn't want to lump your data?
And, second, I am confused by Figure 6 and need some help understanding why high immune activation would cause such a huge drop in the healthy memory pool. It seems counterintuitive to me
1. Ben mentioned that immunologists tend to be splitters, after reading this paper I'm starting to see what he means. It's difficult to keep track of all the cell types, proliferation states, infectivity levels, etc. I'm not complaining at all, I just find it fascinating. The more I read about this subject the more skeptical I am that an accurate model will be developed to explain the slow decline of CD4 T-cells. It seems as though the splitters generate more and more information for the modelers to work through, which leads to ever more complex models attempting to account for all the potentially relevant variables. As model complexity increases I would imagine assumptions build up and reduce the accuracy of the model. What a tricky problem!
2. I find the hypothesis that senescence plays a role in declining CD4 T-cells interesting (see pg. 0953 par. 4 and supplemental). It seems logical to me that cells in infected patients will fatigue over time as they are taxed to maintain homeostasis. HIV must expedite the aging process of CD4 T-cells by forcing them to divide much faster than normal. It would be interesting to culture these cells and see how many divisions they're capable of in health. Maybe HIV is pushing the CD4 T-cell line to the edge of it's life span?
I found figure 1 in Grossman et al. 2002 (http://medicina.unmsm.edu.pe/farmacologia/archivos/ARTÍCULOS/CD4%20T-CELL%20DEPLETION%20IN%20HIV%20INFECTION.pdf) to be helpful in visualizing T-cell population dynamics. Also, on page 322 of the same paper the authors note that "if virus or CTLs [cytotoxic T-cells] eliminate a substantial proportion of activated CD4+ cells, too few activated CD4+ cells may be converted into resting memory cells to replace those lost by activation. In this case, the pool of CD4+ memory T cells might be gradually depleted."
In the Yates paper....
I'm curious about how the authors arrived at an infectivity parameter estimate of p=200 for Figure 7. Perhaps decreasing the estimate of p might decrease the loss of dividing cells to infection, thus extending the time course to CD4+ cell depletion. (Also why in Figure 3 is the threshold value of p set at 100?)
On page 951, the authors state "...our results are insensitive to the choice of gamma1 and gamma 2." Do they mean the model results are insensitive to the values of gamma1 and gamma2? What if gamma2 is very close to a*? This might also be an adjustment that could be made to extend the time course to CD4+ cell depletion.
Also might be worth taking a look at the corrections. http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.0050011
Christ! What the hell is ART?!? I can't find what it stands for any where. Oh I see, antiretroviral therapy. Excuse me.
Why would we expect the 'runaway hypothesis'? It appears that the virus has mechanisms in place to slow the deterioration of the host. Isn't this what we would expect? A virus can't just kill its host, it needs to be transmitted. So yeah, the results suggest a slow process (i.e. viral adaptation) must be active in the HIV-induced depletion of CD4+ cells, as opposed to a rapid one.
Obviously, the more important questions is 'how' and these authors provide a bunch of hypotheses. Interestingly, the supporting info S1 shows a model that more so reflects the observed phenomenon. While I like the publication of a non-significant result, it seems that S1 has some important info.
Okay, got to drive to school. See you all in class.
I found this paper very interesting and specially the last comment about it (the reader response: "The Hypothesis that Wasn't")!!! When I read the paper, I found that it was lacking of how could the model run when the body is infected by a different pathogen, other than HIV (as a control)!!! So, that is my question!!! How could this model run with other viruses, for example H1N1?!
Could be possible that T cells behavior against weird particles in the body needs a negative (-) feedback? Could the HIV turn the negative feedback to positive? could this new positive feedback collapse the immune system?
However, I think that these authors published something reasonable, system modeling is not an easy topic. Specially when the system is a complex one.
Hi all, Excuse me for posting my comment so late but I found this paper a difficult one to read...
I just wonder how much their model could be generalized to other virus infections that also affect CD4+ T cells (all?, is HIV the only one?) or other type of cell, and why should it be considered specific to HIV. By the little amount of knowledge that I have gathered on this topic just by our last 2 readings (not much though...), I guess that what makes HIV very special is its very high rate of reproduction and mutation, so I think that a HIV-specific model should also consider the dynamics of the virus besides the dynamic of the affected cells and the interaction between these two processes. That's it...
Lets give this a shot Tim. I checked my graphs and I think I had it right: x on the x and y on the y. I actually flipped them around and it looks like a teepee.
Very interesting and, again, paints a dismal picture. Yates et al. mentions that immune activation is significantly reduced during antiretroviral therapy, at which point homeostatic compensation is presumed to take over the responsibility of CD4+ cell turnover--assuming that this turnover allows for a faster accumulation of viral drug resistance through sustained viral generation, I am wondering if suppressing homeostatic compensation during this period would extend the period of ART effectiveness, possibly preventing a viral rebound. I also had a question regarding the assumptions of the model that CD4+ cells replicate consistently--though Yates et al. mentions that these cells can decline in proliferation, I am wondering if the actual variance in replications per cell would significantly reduce the strength of the model's predictions.
ReplyDelete-Mike
So had some difficulty following along with this paper because of the vocabulary but here are a few things:
ReplyDeleteEarly on in the paper (page 949) they tell us how they estimate 'a,' the rate at which resting cells are signaled to begin dividing and say that 2 types of cells compete for these signals, the effector CD4+ T cells and central memory cells- and according to my understanding they just average published 'a' values for each type of cell. This is a critical number in the model- it tells us how often immune cells divide. Am I wrong, or is this a really critical step where you wouldn't want to lump your data?
And, second, I am confused by Figure 6 and need some help understanding why high immune activation would cause such a huge drop in the healthy memory pool. It seems counterintuitive to me
1.
ReplyDeleteBen mentioned that immunologists tend to be splitters, after reading this paper I'm starting to see what he means. It's difficult to keep track of all the cell types, proliferation states, infectivity levels, etc. I'm not complaining at all, I just find it fascinating. The more I read about this subject the more skeptical I am that an accurate model will be developed to explain the slow decline of CD4 T-cells. It seems as though the splitters generate more and more information for the modelers to work through, which leads to ever more complex models attempting to account for all the potentially relevant variables. As model complexity increases I would imagine assumptions build up and reduce the accuracy of the model. What a tricky problem!
2.
I find the hypothesis that senescence plays a role in declining CD4 T-cells interesting (see pg. 0953 par. 4 and supplemental). It seems logical to me that cells in infected patients will fatigue over time as they are taxed to maintain homeostasis. HIV must expedite the aging process of CD4 T-cells by forcing them to divide much faster than normal. It would be interesting to culture these cells and see how many divisions they're capable of in health. Maybe HIV is pushing the CD4 T-cell line to the edge of it's life span?
Hey Christine,
ReplyDeleteI found figure 1 in Grossman et al. 2002 (http://medicina.unmsm.edu.pe/farmacologia/archivos/ARTÍCULOS/CD4%20T-CELL%20DEPLETION%20IN%20HIV%20INFECTION.pdf) to be helpful in visualizing T-cell population dynamics. Also, on page 322 of the same paper the authors note that "if virus or CTLs [cytotoxic T-cells] eliminate a substantial proportion of activated CD4+ cells, too few activated CD4+ cells may be converted into resting memory cells to replace those lost by activation. In this case, the pool of CD4+ memory T cells might be gradually depleted."
In the Yates paper....
I'm curious about how the authors arrived at an infectivity parameter estimate of p=200 for Figure 7. Perhaps decreasing the estimate of p might decrease the loss of dividing cells to infection, thus extending the time course to CD4+ cell depletion. (Also why in Figure 3 is the threshold value of p set at 100?)
On page 951, the authors state "...our results are insensitive to the choice of gamma1 and gamma 2." Do they mean the model results are insensitive to the values of gamma1 and gamma2? What if gamma2 is very close to a*? This might also be an adjustment that could be made to extend the time course to CD4+ cell depletion.
Also might be worth taking a look at the corrections.
http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.0050011
Christ! What the hell is ART?!? I can't find what it stands for any where. Oh I see, antiretroviral therapy. Excuse me.
ReplyDeleteWhy would we expect the 'runaway hypothesis'? It appears that the virus has mechanisms in place to slow the deterioration of the host. Isn't this what we would expect? A virus can't just kill its host, it needs to be transmitted. So yeah, the results suggest a slow process (i.e. viral adaptation) must be active in the HIV-induced depletion of CD4+ cells, as opposed to a rapid one.
Obviously, the more important questions is 'how' and these authors provide a bunch of hypotheses. Interestingly, the supporting info S1 shows a model that more so reflects the observed phenomenon. While I like the publication of a non-significant result, it seems that S1 has some important info.
Okay, got to drive to school. See you all in class.
I found this paper very interesting and specially the last comment about it (the reader response: "The Hypothesis that Wasn't")!!! When I read the paper, I found that it was lacking of how could the model run when the body is infected by a different pathogen, other than HIV (as a control)!!! So, that is my question!!! How could this model run with other viruses, for example H1N1?!
ReplyDeleteCould be possible that T cells behavior against weird particles in the body needs a negative (-) feedback? Could the HIV turn the negative feedback to positive? could this new positive feedback collapse the immune system?
However, I think that these authors published something reasonable, system modeling is not an easy topic. Specially when the system is a complex one.
other last thing:
ReplyDeleteThe immune system has two responses, primary and secondary.... It seems that these authors assume them as only one!
----
Finally, this paper inspire me to learn more about modeling! This is a fascinating topic that could be a very useful tool for my interests in biology!
Hi all,
ReplyDeleteExcuse me for posting my comment so late but I found this paper a difficult one to read...
I just wonder how much their model could be generalized to other virus infections that also affect CD4+ T cells (all?, is HIV the only one?) or other type of cell, and why should it be considered specific to HIV. By the little amount of knowledge that I have gathered on this topic just by our last 2 readings (not much though...), I guess that what makes HIV very special is its very high rate of reproduction and mutation, so I think that a HIV-specific model should also consider the dynamics of the virus besides the dynamic of the affected cells and the interaction between these two processes. That's it...
Lets give this a shot Tim.
ReplyDeleteI checked my graphs and I think I had it right: x on the x and y on the y.
I actually flipped them around and it looks like a teepee.