This course will introduce the student to contemporary Systems Biology focused on mammalian cells, their constituents and their functions. Biology is moving from molecular to modular. As our knowledge of our genome and gene expression deepens and we develop lists of molecules (proteins, lipids, ions) involved in cellular processes, we need to understand how these molecules interact with each other to form modules that act as discrete functional systems. These systems underlie core subcellular processes such as signal transduction, transcription, motility and electrical excitability. In turn these processes come together to exhibit cellular behaviors such as secretion, proliferation and action potentials. What are the properties of such subcellular and cellular systems? What are the mechanisms by which emergent behaviors of systems arise? What types of experiments inform systems-level thinking? Why do we need computation and simulations to understand these systems? The course will develop multiple lines of reasoning to answer the questions listed above. Two major reasoning threads are: the design, execution and interpretation of multivariable experiments that produce large data sets; quantitative reasoning, models and simulations. Examples will be discussed to demonstrate “how” cell- level functions arise and “why” mechanistic knowledge allows us to predict cellular behaviors leading to disease states and drug responses.
Case Studies II
Loading...
From the course by Icahn School of Medicine at Mount Sinai
Introduction to Systems Biology
279 ratings
Explore Related Videos
At Coursera, you will find the best lectures in the world. Here are some of our personalized recommendations for you
Icahn School of Medicine at Mount Sinai
279 ratings
Course 1 of 6 in the Specialization Systems Biology and Biotechnology
From the lesson
Emergent Properties: Ultrasensitivity and Robustness | Case Studies
Module description goes here.
Meet the Instructors
Ravi Iyengar, PhDDorothy H. and Lewis Rosenstiel Professor
Department of Pharmacology and Systems Therapeutics, Systems Biology Center New York (SBCNY)
[MUSIC]
Okay.
Continuing, the study and, of, signaling micro domains within cells.
We tried to use the partial differential equation based
models to understand the origins of the micro domains.
What we found was that both cellular geometry and enzymatic activity
of negative regulators are needed for the formation of these micro domains.
You might remember in one of the last slides of the previous lecture.
I showed you that ten dendrites, de, dendrites with the small
diameter favored the formation of sharp gradients of cyclic AMP.
As compared to thick dendrites.
So we explored this property further with thin and thick dendrites
but now adding another variable which was, reading concentrations.
Or reading levels of the negative regulator phosphodiesterase [LAUGH].
Also called PDE or PDE4 in this case, varying levels of the phophodiesterase.
So when there is no phosphodiesterase or negative regulator there's a lot
of cyclic AMP can be formed and then loss signal flows through.
So there are no gradients really observable.
At moderate levels of phosphodiesterase, one starts to see nice sharp gradients.
And at high levels of phosphodiesterase all the cyclic AMPs
dissipate, and all the gradients get to be dissipated, as well.
Now, you can see that the, the key in, this happens,in the case
of the thin dendrites, but not so well in the case of thick dendrites.
Although at one particular concentration of phosphodiesterase
there is some gradient that's seen here.
This effect of the phosphodiesterase which degrades cyclic AMP.
on, on which is out here, on MAP kinase, which is down here
it's sort of distill prediction of how this network is configured.
And so, we try to use this prediction to test the whether one could see
this in real life experimentally and for this we used brain slices.
Stimulated the slices with a bit [INAUDIBLE] receptor ligand.
And ask the question, does, is their MAP kinase
activity in the dendrites, or, or the cell body.
And as you can see that when the phosphodiesterase
is inhibited, with the chemical inhibitor, there are no gradients.
Similar to the condition here and there is for a
MAP kinase both in the cell body and the dendrites.
Otherwise, the MAP kinase is largely localized to the dendrites
and there is very little in the cell body of each.
Here are the corresponding simulations for this.
Okay.
So if these micro domains sort of require both enzymatic
activity of the negative regulators and the shape or size of the cell.
One can sort of ask the question.
Is this, are these conditions enough to transmit the
spatial information from upstream components to downstream components?
So in, in this particular set of simulations we
follow the flow of spatial information within the dendrite.
And it's this one white dendrite in this particular neuron.
And starting with cyclic AMP the signal goes to protein kinase A.
And from there it goes down to MAP kinase and to to a couple of pathways.
So you can see that the PKA gradient is similar to the MAP kinase gradient.
However, when the signal that goes to the Mac.
Neither RAF nor MEK show good gradients.
In contrast, the protein kinase A gradient,
is recapitulated by this second negative regulator, PTP.
Which is which is phosphotyrosine phosphatase, which is the
regulator of the tyrosine phosphorylation side of MAP kinase.
So you can see that the gradient, this
looks similar to this this looks similar to this.
Indicating that the flow of gradient was likely to be to this PTP pathway.
Again since although there was a cyclic
PKA gradient because there was no MEK gradient.
We tested whether this was observable in experimentally.
And we saw indeed that MEK is activated both
in the dendrites and the cell body in these experiments.
So, from these experiments what one can sort of conclude
is that the flow of information regarding the activity state.
Which means protein kinase A activates RAF.
B-RAF activates MEK.
MEK activates cyclic AMP forced through this RAF MAPK kinase pathway.
However, the flow of special information.
That is how much MAP kinase is activated at a certain location.
In response to the similar spatially
restricted activation of PKA flows through PTP.
You can again clearly see this is the same simulation as before.
The PKA gradient matches the PTP gradient that matches the MAP kinase gradient.
So what does simulation which was indeed a very surprising finding
was that the flow of spatial information is through a different arm.
As compared to the flow of information with regard to the activity.
Of course, the model of the computation predicts, the
per that PTP is going to be very important.
And we tested, therefore, the importance of PTP for loo, by looking at
what happens when you ablate the PTP by using antisense oligos.
And what we found is that when you
knock out the PTP, the MAP kanise gradient disappears.
And you get MAP kinase all over the cell body,
which is what is shown here, as well as the dendrites.
So, the study was very useful in several ways.
Because it told us that, in as a true systems level situation.
Multiple factors including sub-cellular geometry and
enzymatic activity, they're responsible for microdomains.
You can see here that some of the physical determinants.
So that lead to surface to volume ratio control.
Microdomains upstream.
And this in turn is controlled by negative feedback regulation.
And further down there are feed forward motives that control it.
So a variety of factors control the flow
of information to various taps in the signaling pathways.
The other, I think, most surprising finding from these
modeling studies was that spatial information is a distinct entity.
As compared to information regarding activity states.
And this can be separately transmitted
to different proteins through selected pathways.
And of course, the most general of
these conclusions is that these partial differential models.
Actually do provide deep insight into the dynamic of
these subcellular processes of how microdomains are formed and dissipated.
Okay, let's move on to our third case
study which is [LAUGH] a more complicated system.
And in this study, the researchers, mostly in Julie Theriot and Alex Mogilners
lab were interested in understanding a pretty high level question.
Which is how do moving cells determine their shape.
As cells move along a gradient or even swim randomly, they,
they are capable of preserving their shape even though they are moving.
And the question is how do they do this?
For this the researchers used fish skin epithelial cells.
These are called keratocytes.
People use these cells to study these kinds of
stuff because they are fast moving cells and cultures.
And when they move, even during the
movement, they maintain both their shape and speed.
So one can study these cells in a way that is mm, we
can look at the shape of the cell and also look how they're moving.
And ask the question what are the processes that are
involved in sort of preserving this shape during the dynamics of movement.
So there are four characteristics that provide a nearly
complete description of the shape of these ce, cells.
And the four characteristics are shown here.
There are cells that are D shaped.
There are cells that are elongated and canoe shaped.
There are sev, the third mode tells you something
about where the position of this cell body is.
Which is down here, with respect to the leading edge
and the final mode tells you something about right left symmetry.
So by looking at all the characteristics that go with cell shape.
One can identify the four top, if you
want to call that characterize a population of these keratocytes.
So what these authors also found that for sharp, for [LAUGH] fast moving cells.
The speed of the cell could be correlated with the shape of the cells.
So here is a correlation between area and time.
Here is a correlation between aspect ratio.
Which is the long, which is the short access on time.
Here is correlation between the front of the cell and with respect to time.
And here is a correlation of the speed of
the cell with which it moves with respect to type.
Now, the different correlations are plotted here in a summary fashion.
And I just want you to pay attention just to the aspect ratio.
One which gives you the cell, and you can see that the
aspect ratio one correlates with speed and also correlate with front, roughness.
And of course, it correlates with actin ratio.
So these kinds of correlation allow the cells to be
basically put into two categories, which they call coherent and incoherent.
And all of these experimental data allow them to now develop a mathematical model.
That will be useful for studying, this complex behavior.
So what does the mathematical model do?
The mathematical model provides a direct link between the
distribution of actin filaments and the overall cell morphology.
In this figure shown here in A, one, one
can look at cell shape and actin filament density.
By, measured, by binding of the fluorescent probe kabiramide.
Which stains the actin binding or the actin
filamentary the region where geosolve binds to actin filaments.
A, B, and C, sort of show the
quantification of shape and actin filament density.
And indeed the shape is maintained by the actin filament density.
So what they did was for they develop some equations to describe this
relationship between the filaments and the shape and so on.
And then, they plotted the comparison between the model
which is in the red line and the experiments.
The simulations which is in the red [LAUGH] line a, a, a,
and the experiments the which are the smooth mean and the blue line.
And you can see that the match between the
calculated values and the experimentally observed values is pretty good.
This is done for a very large number
of cells and these individual dots represent individual cells.
So this mathematical formulation allows them to correlate the shape of
the leading edge, along with the speed of the, with which the cell moves.
So even a model such as this has a number of assumptions.
And these authors have published a very nice
Table of Model Assumptions, the rationale for their assumptions.
The level of confidence they have in assuming this, and
how critical their assumption is, both in the model and experiment.
I'm just going to draw your attention to one of them which warrants us speculation.
And here we can see that this is speculation that the growing
filaments are stored or indeed buckled at the sides of the cell.
They clearly state that this is a speculative assumption.
And since the model calculations agree with experimentally observed
relationships it is reasonable to sort of assume that this speculative.
Assumption is likely to be correct.
However I should caution you there is no direct experimental proof just
visualization of stored or buckled filaments at the, the sides of the cells.
So this is something we need to wait further
to understand how whether it really occurs within these cells.
At the next level, what they did was to now that they were able to
calculate out the sort of protruding shape.
They started to look at whether they could use for
each cells both are measured, shape and a calculated shape.
And what the relationship between the two were for a good,
695 cells and they found that this relationship is very good.
And then, they were able to take this shape calculate out the aspect
ratio then look at the relationship between the shape.
As denoted by the aspect ratio and the speed at which they move.
And they have a predicted behavior and they find that experimentally.
Observed system behaves very closely, enter for this very same behavior.
There's been compelling experiments and, model.
They found the match was actually very good.
And so, one could say that they had sort of like a predictive understanding
of how cell shape is related to the, the speed of cells in these moving cells.
So what are the conclusions from this cell spreading model?
The conclusions are, that relatively simple mathematical
models can accurately describe complex cellular behaviors.
In this case, they did not actually
build a differential equation based dynamical model.
But rather they, they, they found values and use relatively
simple relationships, simple models to relate one parameter to the others.
All of these models however simple or
complex they are, they all have underlying assumptions.
And like, like how these researchers had done in their, in their paper.
It is good to clearly state
the assumptions, underlying assumptions for the model.
And let the reader judge for herself or himself, whether
this is appropriate and how the model is to be interpreted.
The relationship between the speed of movement of
whole cells and the shape can be accurately predicted.
By using the dynamics of actin cytoskeleton
and its interaction with the plasma membrane.
What this tells us is that is that to, even when
you want to build models of complex cells, per cellular processes.
You don't need to put in everything that's known about the cell.
You just need to put in the relevant details and this
will allow you to build sort of good quantitative models of cellular behavior.
So, sort of to get a handle on the feasibility of building complex model.
It always helps to know what detail that nee, need to be included for the model.
Okay, so this concludes this part.
Thank you.
[SOUND]
[BLANK_AUDIO]
Explore our Catalog
Join for free and get personalized recommendations, updates and offers.
Coursera provides universal access to the world’s best education,
partnering with top universities and organizations to offer courses online.
© 2019 Coursera Inc. All rights reserved.
