Which is.

So.

All.

All right thank you very much.

But it's a pleasure to be back.

A Georgia Tech the last time I was here
are you guys when the old building.

Tells you how long ago.

It's been since since I was here and
it's good to see old friends or

new friends in the in the audience and

I really am grateful to the organizers for
asking me to come and

give this presentation before I
get too far into the presentation.

I know that it typically happens
that we get so far into the work and

we forget to give credit to the people
who actually helped me get the work done.

So I'm going to do that right away so

that I don't forget my students chant is
the one that did most of this work and

he's actually as we speak right now
trying to finish up is this is.

Unfit miles is the Ph D. student in
the biology course sciences as you will

see this is a joint collaboration between
people in chemical engineering and

people in biological sciences Cooper
psychs are still in the University

of Delaware the biological sciences
department and the Center for

translational research in cancer
signifier cross in unfortunately has

left Delaware and she's gone to rice but
these are the people that did the work.

With me and
I'd like to give them a lot of credit.

Before I get into the whole.

Presentation I think it might be a good
idea to tell you the moral of the story

that I'm going to tell before we don't
get into all the details the moral of

the story is the following if there
is a control system involved in

any system be careful how
you analyze the data you

collect because if you don't know that
there's a control system involved.

You are more likely to
misinterpret your data.

Let me give you a good example.

If you have a thermostat in your
house that regulates the temperature.

And you have somebody
constantly opening the door.

Especially during the winter and closing
the door and open it and closing the door.

You will find that if your
thermostat works very well

the temperature inside of
your house will not change.

Whereas the temperature outside
is very cold and it's changing

the temperature inside of the House
doesn't change if you then collect data.

And you try to correlate
outside temperature

with inside temperature you will
come to the wrong conclusion

that outside temperature has no
effect on the inside temperature.

The reason it appears that way is
because you have a control system

inside of the house that is working
very well that is the central

story behind what I'm going to
tell you now if a control system

is involved in something with very
careful how you interpret the data.

I'm going to start then
by showing you a cartoon

of a cell because this is where all
the action is going to be taking place and

what you will find is that
inside of this membrane.

Any collection of imagining
collection of proteins

that do all kinds of different things so
that if a signal

comes from outside and in particular
going to focus on that aspect of it.

If a signal comes from outside the amount

of these are like and changes.

It would bind to the receptor and
then set into motion

a whole sequence of events
that are at the end you

would change Jenna expression and then it
will cause a few other things to happen.

OK So and there are normal
circumstances especially when it

comes to growth there is
a signal that says it's time for

a cell to grow and this will start
to grow and divide and then another

signal will come and say stop growing and
then the cell will stop growing.

Problems come when you tell it tell the
cells to stop and it doesn't stop growing.

That's how cancer develops but
even then along the way there are all

kinds of checks and balances to make
sure that even if one fails there's

something else that would that would take
over and at least try to bring things

under control what we're going
to be looking at is how T.G.F.

beta placed a certain role in cancer and
that really has been misunderstood and

how by using mathematical modeling and
including process control.

We have been able to resolve this paradox.

This is a diagram from her hand on
Weinberg in the year two thousand and

recently as a few months ago they
came up with a newer version and

the reason I put in this up is that one
hundred and one Burg scientists their

cancer biology is there but they think
like engineers one of the things

that they have now done is to put this
in terms of subsystems and as you

can see I'm not going to spend a lot of
time on any of this but this is the subs.

System that we're going
to be looking at for

this presentation and
so what is T.G.F. beta.

Let's start from there it is well
actually there's no such thing as

a single to appear this a family of pages
per for the sake of this discussion.

We're going to say it is a cytokine
it's a lie again that is involved in

many processes proliferation of
Potosi differentiation maturity

depending on the conditions
depending on context they can

be involved in all of this
under normal circumstances.

However it turns out that when

excel becomes cancerous the role that E.G.

a better place is not very well
understood what we have or what.

What is available is a lot of data and
people have looked at the data and

what I'm going to discuss with you
are some of the things that people have

thought are happening with T.G.F.
beta and here is one of them.

This is real data and I'm not going to
bore you with a lot of the details that

under normal circumstances T.G.F.
beta is a very potent to most suppress or

in other words if it appears as if
cells are going to become cancerous.

There is a release of T.G.F.
beta and within a short period of time.

It brings the cell population
back under control.

Unfortunately us.

If you look at what happens
with med a static cancer and

specially when you look at the worst
cases the cases with the was

prognosis you find that
the level the amount of T.G.F.

beta is the highest in cases
with the worst provosts is now.

What kind of sense does
that make if T.G.F.

beta is supposed to be a tumor suppressor
how come that this level is higher.

In cases where.

It's the worst prognosis.

So the idea then is that maybe T.G.F.

beta starts out as a tumor suppressor and
then somewhere down the lie.

It changes his mind it becomes
a tumor promoter now for

an engineer that does make sense.

So let me give you a quick.

Analogy.

We all know that if the temperature
inside of your reactor is going up.

If you increase the cooling
water flow rate.

What happens to the temperature
brings the temperature down.

So everybody understands that and
that doesn't cause any problems.

So if somebody then says well if
you increase cooling water flow

rate your temperature is going to go up
and genius will rather will say something

is wrong with your reasoning because
this does not make any sense.

However imagine a situation in
which you have sea water supply and

you have calcium deposits
around your your reactor.

So it changes the heat
transfer characteristics.

Not only that it takes away the volume
from your cooling capacity.

So the temperature starts going up
inside of your ex atomic reactor.

OK.

You increase the cooling water flow rate
it tries to bring the temperature down but

it's not able to bring it down completely.

Temperature is hotter reaction goes
faster because there's a thermic and

they're used to add more cooling water and
luckily water but

you are unable to keep up your huffing and
puffing to keep up.

What you care at temperature
keeps going up and

then the reactor melts
down after a meltdown.

You then go back and you do a postmortem
you take the stripped out recorder for

those of you who were
stripped recorders or.

You take the structure recorder and
you look at the temperature going up and

you look at they could have
got a flow that you see Paso.

And this is the discourse that you
know because you understand better.

We thought that that was
what was going on but

we needed to be able to convince people.

So this is really the.

Basic concept behind what we've got going
to talk about so I feel it's important for

me to let that out of the bag and

then we can look at the details
a little bit later on.

So here are the questions we're
trying to ask how can a single sided

changes might switch from being a tumor
suppressor to be to a tumor promoter

that doesn't seem to make sense but let's
see if it's possible for that to happen.

Why is the amount of teacher
pay down usually high in those

cases where you get the worst prognosis.

When it is supposed to be a tumor
suppressor and growth inhibitor.

All right to help us with the approach
that we're going to take it is

important to realize two things
one cancer is fundamentally

a cellular disease in the sense that
it is a single cell go in crazy and

then the next one goes crazy and
the next one goes crazy that causes cancer

although if the treatment is at
the physiological level so far.

If you do.

Radiation therapy or
if you do chemotherapy.

It's still at the.

Physiological level what that means is if
you want to understand what's going on

with it to study the single cells because
that's really where the problem mechanism

is but to be able to understand clinical
observation we need to we need to be able

to look at the macroscopic description so
what you're going to see is I'm going to

do both and
this is the reason that I do both.

So here's my outline the first thing
I'm going to talk about is single cell

modeling the good news about this
is that the paper is already out so

I can not talk about all the details and
if anybody is interested in the details

you can you can look at the paper on
Fortunately the macroscopic model.

The paper is not yet out and

sound is working on it as we are as we
are finishing but the stuff you've taken

control from a whole host of people
I'm looking at in this audience.

You'll be able to understand the control
system block diagram and the modern and

the simulation that we're going to do and
then I'll summarize and conclude it was

OK So let's talk about the single cell
modeling and again I apologize for

throwing something like this up for those
of you who may not be familiar with this

kinds of cartoons of what
happens with signal transduction

here is essentially what goes
on with a signal the lie

guaranteed you have paid or
when it comes close to its receptor

the type two receptor it binds to it and
you get this kind of.

Thing how it's a cartoon.

But it helps understand
what's what's happening.

It gets first correlated and

their recruits the type one receptor
they both now form in receptor complex.

Now once you form the receptor complex
that's what's going to go on the go into

psychosis goes into the side apply and
then

Smart two will become forceful related and
a whole bunch of things happen.

OK let's without going into a lot of
details you get this mad too smart for

complex which is then able to go in
trance look at into the nucleus and

then kick off.

Transcription changes engine expression.

So what we're interested in finding out
is what happens when I make a change

in my arm out of T.G.F.
beta and what is the response.

In the amount of smart too smart for
complex that I get in the nucleus.

So the if I could make this change
what happens here because this

is what's going to help us understand
the signaling propensity of T.G.F.

beta and what we want to do is
use this mathematical model to

understand what happens
under regular conditions and

what happens on the cancerous conditions
to be able to see how this hour.

Signaling system response.

OK And if we do that and by the way this
is the paper it appeared in by physical

Journal in two thousand and
nine and the mathematical

model is a system of seventeen would
ease with thirty seven parameters.

And the first question that I think
everybody should ask me is where do you

get these parameters from OK with
thirty seven parameters and seventeen.

Odis this is this is a lot and
there's not much data.

OK so I'm going to take you through not
in great detail but with enough to give

you an understanding of what we're
doing so here's what we do we do.

What's an initial rough
estimation what does that mean

we look in the literature to
find values that we can use for

some of these kinetic parameters and
then we do a very important thing

it parametric sensitivity analysis
what does that mean it means we asked

the question if I change this
parameter by is certain amount.

How does my mathematical model respond so

it's a it's a partial
derivative of D Y D P.

And you generate this matrix you
do this sensitivity analysis.

Unfortunately it's local
It has to be local.

And then we what this allows
us to do is to find those

parameters which you can change the value
by so much and it doesn't change much.

So it's not very sensitive so don't waste
your time trying to pin that down after we

find out the parameters that are really
the most sensitive then we do list Quest

feed into data and after that we
do parameter identify ability.

Now you might be asking why are we doing
this backwards why don't we do the primate

identify ability first which by the way
means even if I had perfect data.

Can i don't defy this parameter can
I get a value for this parameter.

Why do we have to do this backwards.

Well unfortunately you need some in.

Parameter estimates to be able to do this.

That's why we do it
this form in any event.

They're the identify
parameters we refine them and

a lot of this information is in the paper.

So I'm not going to drag you through this.

Let me just show you some results and

before I tell you about
the results this is.

I don't want to call it a dirty little
secret full of people who do signal

transduction modeling but
maybe it is you would notice that all of

the day that I have here
their maximum values is one.

OK that is because any time you
collect data and you do Western blot.

The the amount of protein that you have
your protein abundance is relative.

Nobody knows what
the absolute amount is so

unfortunately the only way we can
use data from different labs to.

Estimate parameters is to normalize.

And for those of you or control people
right away you know that this means that

we don't know what the gain is for

the system you will never
be able to know the game.

Well that's a small price to pay for being
able to get this kind of information.

OK So let me just put that out there so

that you all know that this is what we
have to do because of the nature of

the data until we improve
the measurement capability.

This is all that we have so

anyway having said all of that
here is data from different labs.

And this is the relative concentration
of phosphor alluded to in the nucleus.

This is the concentration of force for
illiterates part two in the side applies

and and cytoplasm and smiled for in
the nucleus and you can see that the model

fits well now don't forget this is how we
estimated the parameter simultaneously.

So we have this set of data and

we estimated the parameters
by fit in the model to this.

OK So that is modified in the next
thing we did was the model validation.

We took at.

Completely different set of data from
a completely different set of labs and

we used to validate the dataset.

So essentially what we
did was we had data and

we decided we're going to use this
to fit the model where they're going

to use this to validate the model and
at this point there was no.

We don't change any parameters.

So deep the model fits really well and so

now we want to use the model to
do some analysis and you can

see some of this in the paper but the some
of the things that we did was to look at

which parameters
are the most sensitive to or

it's the other way around which
parameters affect the response.

The most and
here's a list of of this parameters.

The interesting thing about what I've
listed here is that if you look in

the literature and you look at
some of the mechanisms of the bill

the signaling of T.G.F. beta nuclear
cytoplasmic shuttling of smart for

was not meant to be
something that important.

It wasn't highlighted as
something that's important.

What we've identified that is very
important and I would talk about

some of the other things in the paper
what those implications are for

this presentation I'm going to skip.

The rest of it.

What I'm going to do now is have
an validated the model that he works very

well single cell model we now want to ask
some questions we want to investigate

the behavior of this single cell
under conditions that are normal and

other conditions that are cancerous to be
able to see if we can understand what T.J.

paid a signal in can tell us
about the observed behavior.

So here are some of the things
information that's known

the everything that is in green
is known to be mutated or

down regulated in cancer
cells everything that is.

Yellow is a modified and you don't see
anything here anything that's red is over

EXPRESS So for example T.G.F.
beta is over expressed in cancer cells and

I started by telling you that that you see
too much to your theta and the question is

Why do you see tumor to appear when it's
supposed to be suppressing the tumors

the things that we also
see is that this receptor

are down regulated that gives us
a hint it's a is a lot of cancer cells

have lost a significant amount of
functional T.G.F. beta receptor.

So that's the first thing that
we want to investigate and

keep out of the back of your mind.

We want to investigate what happens
when you have a ten fold decrease

in the initial amount off.

Your teacher better receptor.

OK And the amount you put is so
in other words I have a cell and

I reduce the amount of functional to
better receptors by a factor of ten and

I ask what do I see very
interesting things.

This is the response of of
a system to a pause of T.G.F.

beta what you get is the your receptor
your first formulated receptor complex.

This is under normal conditions
under cancerous conditions.

It is attenuated you don't see
a lot of it when you now go into

the smart too smart for
first formulated in the nucleus.

When you look at it under
normal conditions you've got

what's called a sustained.

Response.

But when you under cancerous conditions.

It is transit in the sense that it
goes up and it comes back down.

Now if you one of the postulates that
we then gave was that T.G.F. beta for

it to do what it needs to do may
have to get to a certain amount pick

value to trigger the trance transcription
and translation of all the things.

That would help do the anti cancer.

Responses.

So what if you don't get to that threshold
and it doesn't do what it needs to do

because this is saying that the threshold
that you get to the maximum value is

is small so we do what's called a two
year period a dose response and

the idea is if you got a certain
response from your cell.

How much T.G.F. beta do you need
to be able to get that response.

OK so here is a plot and I apologize for
this that it doesn't seem to

it doesn't seem to show very well let
me take an example if I want a certain

peak value of the piece much too smart for
in the nucleus the blue line

is what I get from normal cells the red
line is what I get from my cancer cell.

So if I want a certain value
you will see that I get

ten times I need ten times as much.

T.G.F. beta asked to get
the Senate response and

it ten times as much response from.

Ten times as much did you have
paid off from cancerous cells

than from normal cells or
if you flip it the other way around.

Yeah that's exactly what it is to
get to get this you need no T.G.F.

beta to elicit the same response
from a cancerous self and

now the control people can start
to see where this might be going.

This is if I want to
get the same behavior.

If I want to get the same response.

I need to use more T.G.F. beta.

It's like I need more cooling
water to get the same response

that I got before I got fouling
inside of my reactor and

I was unable to get the same
temperature control and so

on the basis of that on the basis of
the fact that you need more T.G.F. beta

to get normal response from a cancerous
cell we postulate at this hypothesis that.

There's a control system.

There is a cellular
control system that uses

to achieve its objective of
regulating cellular growth

this control system functions
effectively in normal cells.

But with cancerous cells
this this control system

is still intact there's
nothing wrong with it.

The only problem is that in now
needs to secrete more T.G.F.

beta because of because T.G.F.

beta because the curse Roselli is no
longer as responsive to it anymore.

And therefore the increased
level of T.G.F.

beta is a consequence of this
acquired a beta resistance.

Not because of it.

Now you can postulate that that's from
looking at the single cell model but

we're not dealing with just a single
cell what happens when you put itself

population together.

So now we want to shift gears and

say Let us now take this and
look at what the repercussions of this

hypothesis would be if we then put
this microscopic model together

of a control system and
we postulate what it is and.

Look at a simulation of its
behavior can we reproduce

the observed clinical responses.

If we can then we would
move to the next step and

try to identify this clinically All right
so here's what we're going to do we want

to provide a quantitative explanation
of this contradictory role and

the way I was going to do this is
to use mathematical modeling and

use Process control because when
we postulate a control system.

So here's what we're going to do.

We're going to start by identifying
different components of the process

develop models for them and
then put them together and

do a simulation so
what process I was going to use

where specifically look at the prostate
gland why because the people we were.

With.

Prostate cancer people and

the center that where we are doing this
work really focuses on prostate cancer.

So we're going to be looking
specifically at the prostate gland

the epithelial as well as
the stronger compartment.

So that's our process and
as you probably know that when young boys.

Before they reach puberty
things are still growing.

So there's development and T.G.F. beta is
part of the regulation of what goes on

during that period and
what happens is stressed around

is continually used to stimulate
proliferation of prostate cells.

But by the same time.

You.
You have to it's like up a gas pedal and

a brake pedal.

You press the gas if you want to move
forward you press the brakes when you want

to stop so but by combining both you
can get development in which you

you generate more cells and when you
you've had enough you can tell it to stop.

So this energy half
are the growth factors that help.

The growth and then you stop with T.G.F.
beta in not ourselves

induces differentiation inhibits so
proliferation in normal cells.

When cells begin to
undergo only usual growth

has the sequence of events
the break The Best Man membrane and

counter Stroma and there are people here
who actually deal with in information.

So you get information and
you get a response right away.

Part of the response is T.G.F. beta
producing cells would produce T.G.F. beta.

However this is like one of those times
when you want to delete something on your

computer and you get the signal says
Do you really want to delete this.

OK When you produce T.G.F.

beta because it's very potent
it is produced in form of.

Large little complexes.

It is not value of deliberate
away then you have

to go through a whole sequence of
activation to be able to release them so

they become bioactive In other words you
produce them and then you ask you Do you

really want this and only when you really
want it will it make it available to you.

So buy a villa Bertie itself
is regulated by a sequence of

it's a very complicated steps.

It's very complicated
sequence of events and

in the early stages this is
enough to kill off the cancer.

All right so this is what the process and
this script a few things but

what I've done is now put together
a control system block diagram.

So here's your process is the process
of self proliferation and death

specifically in the prostate gland your
growth factors act as the disturbances.

So for those of you who've done process
control before the things that you see in

parenthesis are the ones that
you should be able to relate to.

So your growth factors
are like the disturbances so

you put in a growth factor and
the cells start to grow.

You're controlled output is
the total cell population.

We're not sure what the sensors
are we're not entirely sure

what the controllers are but
they are sort of together and

what happens is that they just like
I said you break the membrane.

I think the sense that there is
there is a matter of pressure now

on the on the basement membrane and

therefore they start to activate
the production of T.G.F. beta.

But you have is produced in
an inductive form just like I said and

then you have to activate them by previous
Is this are these are things that cleaves

the cliff that the proteins
to release T.G.F.

beta and then integrates also play a role
and so then you get your activation.

So think about this like your
actuator you're control valve and

then you release your T.G.F.
beta than then.

On The System.

Now that we have something like this
we develop a mathematical model for

each one on the basis of what we know
what's available in the literature and

then to show you a few things about
how that those models are developed.

The differential equation for
growth is actually relatively simple

it is the rate at which it's been
proliferated and the rate of death and

the difference between them multiply
by the population where the rate

of proliferation depends on your growth
factor and T.G.F. beta according to these

hills function type model
like I said we're not

entirely sure how the sensor in and
the control is been done.

The only thing we know is that this
biological control systems have

is sigmoid a response function.

What that means is that as your
stimulus grows and grows to infinity.

You don't release something in your
control action is not infinite.

For those of us who deal
with linear control systems

we typically deal in this linear range
where your proportional controller

says that if you give
me this error signal.

I'll give you this response but

in this political systems they
tend to have this signal model.

C. model function and then
the activation system unfortunately we

don't have a lot of time to talk
about all the details of how this.

This takes place and like I like
I explained earlier what you have

a pretty house is the cliff and
then you have

the the influence of into grains and
without going into a lot of the details.

We what we do is represent
by a a differential equation

of this form that takes the last little
complex is with into grains and.

You do you add to version and you obtain
your T.G.F. beta by breaking the.

Can it down.

Let me say something real quickly
about how we obtain the parameters for

this model.

OK So this this the kind of model
that we've put together now

is a physiological model and
we have parameters

that have something to do with the rate
of proliferation the rate of death.

So let me show you an example of how we
are temporary measures for proliferation.

This is.

This is from an experiment performed
where they took some cells and

the put in in-vitro and
they spiked them with a group factor.

OK And so these are the concentrations
of the growth factors and

this is the cell count.

So for every for over a period of or
No hundred minutes or.

Two hours you you look at your
cell number as a function

of time and I see a function of
the concentration so this is what happens.

I look Onsen tradition.

This is what happens at
high concentration and

I want you to pay attention to the fact
that you actually get a little cane.

This is what happens when you don't put
any growth factor into your cells they

start to die and when you put the growth
factor in it and in reverse is that in

some cases you can you can suppress the
death in some cases you actually turn them

around and they start to grow but with
this you can you can obtain the rate of

proliferation in each case as a function
of growth part of concentration and

I would take all this information and
we put it into an expression for

how the proliferation rate as
a function of growth factor.

Is obtained from from this.

So we did this for sequence of of this.

Parameters.

Here is a is summary of the of

the specific parameters that we used and
I would put this together.

And I'm going to skip
the sensitivity analysis.

Because that's that's essentially
just tells us which parameters

affect our response is the most
here's what I'm going to do.

We're going to do a simulation
under normal conditions

when you spike your
system with good factors.

How does this simulated system
respond to a change is step

input of one hundred nine meters in growth
factor when you have a nominal controller.

And here is what we see.

So this is this is color coded
this is the growth factor

you get it a step in your growth factor
yourself population starts out this is

and it comes back to where it was before
you would expect that if your control.

Is working fine.

How does the controller do this.

This is what your ballot active T.G.F.

beta responses like and this is what
happens with your latent T.G.F. beta.

So you make a change in
the a mother growth factor and

it's able to bring it back to where it
was before by increasing the amount of of

little bioactive T.G.F. beta that
you that you put in if I change my

controller characteristics if I if I make
the controller either stronger or weaker.

What kind of response do I get all right.

If if under.

Not conditions is the blue line that I
showed you earlier nominal conditions.

If I have a controller that is less

active if it's if I have a controller
that's less active myself

population doesn't quite get back to where
it was before but it doesn't run away.

You get you get some kind of are not set.

And this is the behavior of your little
better if I increase my control activity.

You can actually get yourself
population to be lower than

what it was before because
you are now being.

In a little over active period.

Now.

We're going to look at cases where
your system is now cancerous.

And we're going to look at this.

We're going to simulate these like I
said initially by reducing the number of

of receptor is that a functional and
we're going to reduce

the number of functional receptor
us in steps to see what happens.

As you reduce the number of receptors and

we do that by changing these
parameters P two and D two and

we're going to change them one point five
fold two fold and three fold respectively.

And what we do here is the responses
that you need to see I had to

go to in love blogs scale here because
now we're dealing with with things on

on a very large scale the blue line
is when you have one hundred percent.

Response.

OK this is what you've
seen before it goes up and

then it comes back down
yourself population.

You're probably is your body of T.G.F.
beta when you lose about

a third of the functional receptor
you get the green line with

the cell population goes up but
it's not anything

to worry about this actually turns out to
be cases the clinical cases where people

have enlarged prostates it's still benign
but is and large and this is now you're

starting to inch in that direction by
the time you reduce it by fifty percent.

It gets a little larger but
you still stable you get to thirty three

percent in other words you'd lost two
thirds of the functionality and you get

things that goes completely unstable and
you are no longer able to stabilize it.

And so
now you see some population going up.

You see your bio active T.G.F. beta.

Going up and now if you didn't know that
there was a control system involved.

It is very easy to say at the end where
this is going up and this is going up.

So this was because in that happened and

this is the point I were trying to get
across to people that it is very likely

that it is because of
the feedback control and

because of the stability characteristics
the inability to stabilize the system that

is causing all of this elevated T.G.F.
beta in late cancers right now.

So this is what I've just said under
normal circumstances the controller

regulates growth and regulate it.

Well under cancerous conditions
the return of the job.

It hasn't changed the problem is that
it is no longer able to keep up with

the fast growth because the cells
are no longer as responsive to T.G.F.

beta and in fact I was talking to
somebody today and I forget who says

it's almost essential in the same thing
as what happens with type two diabetes.

When you become insulin resistant and
that's that's the sort of thing that.

That we want people to start to think
about I'm not going to drag you

through this to go to analysis but
let me just tell you.

Roughly what we're trying to look at here.

The question the theoretical question
we're now asking is the following and

what condition.

Do you cross over from being
able to stabilize your system

to not been able to stabilize the system
if you think about this roughly if

I balance proliferation and
death myself population

remains constant if proliferation
is lower than death.

I kill off all the cells.

If death is less than
proliferation the ground.

Definitely I mean this is this is
essentially what it is so essentially you

have a what the control people here will
understand as an open upon stable system.

OK it's open upon stable and you keep it
stable through active control the question

is At what point do you cross that line
and so we have a mathematical model.

And we can go through
all of this analysis and

the good news about stability analysis is
that it doesn't matter how Nalini it is.

You actually take Jacoby in and
the first every day and

you say the call to zero and
under those conditions.

You can actually find out
a lot about the stability so

even though it looks really
complicated not of that kind of stuff.

The bottom line is when you look at the
Jacoby and we can actually map this out.

We can map out the behavior of why

it is that when we got to thirty three
point three percent the system was already

unstable I mean the boundary was somewhere
around here somewhere about thirty five or

thirty six percent is is when when you
can no longer stabilize the system.

So this was two to show mathematically
when when those conditions happen.

So what are the implications.

Don't forget we're working
with people in center for

Translational Research cancer research.

What are the implications.

Well the question now really is and
forgive me for

going back to control again when
a system that involves a control system.

Stops working.

Is it the process.

That's the problem is it the sensor.

That's the problem is it the controller
that's the problem or is it the actuator.

We need to ask that question
because fix only what is wrong and

see this is part of the problem that when
you see if you see a logical problem

if it involves a control system.

Most people don't even know
to ask those questions.

It turns out in this particular
case the problem is not the sensor.

The problem is not the controller.

The problem is not the actuator
The problem is the process itself

the process itself is simply on able
to respond to the corrective action.

So the question I'm asking my my my third.

Biology is is it possible to
resell it cancerous cells.

I don't know the answer to that but
if it's possible.

Because that's really the problem.

Think about it.

Process with a process going.

That just went down and it's no longer
able to to respond can we change the gain

of the process and that's the question
that we're still asking them right now.

So you've seen this analogy and
I'm just going to move on so

that we can have some time for questions.

So to summarize and I'll tell you
what we're going to be doing next.

But let me summarize the single
cell model that we developed and

that the model that we've
published told us one thing that

if you have if and
we don't know that this is the case but

there is evidence to say that if the first
four are not as smart too smart for

complex in the nucleus is
the true business end or

did you have better signaling
when if you have a cancer cell

to to get the same smart
mediated activity.

You need more T.G.F. beta.

OK so that's what that told us then
we moved to the next stage and

we said OK part

of the reason that we think
we're seeing too much.

T.G. have paid for cancer cells where
things have gone completely crazy

is because there's a control system
because we understand control of stable

systems we postulated that there's
a control system and under that postulate

we're able to show that if you have
a control system that is using T.G.F.

beta and if your cell truly has
lost its function or receptor

is there what you're seeing clinically
is what you should expect to see which

is consistent with the fact that if you
have better hasn't changed its role.

So then blunted you have better.

The problem is not if you have
better if there's a control system.

So the microscopic marrow
that says that that.

It now makes more sense it is
counter-intuitive to say that E.G.

a bit or changes his mind but if there is
a control system then it makes sense that

E.G. a bit keeps doing what he's doing but
are you a control system is working.

You're so you're so is the one
that is not working very well so

implications clinical observations are not

paradoxical what you see is what you
are supposed to see it is all consistent.

OK wish you the level of
teacher better should increase

if your if your cancerous cell
is no longer responding to

what is supposed to be responded
to what are the consequences.

I don't want to mention the name of
the pharmaceutical company because this

is been recorded with their
pharmaceutical companies that have started

to look at anti T.G.F. beta.

Treatment for cancer.

OK And that is not.

According to what we're seeing here.

That's not the way to go.

The problem is now with Egypt better.

OK So current approaches of targeting T.G.

a bit alike and
therapeutically may have to be abandoned.

And we need to start
looking at something else.

Now what is that something else.

I don't know I don't know if cells can
be recessing times but if they can be if

there's a will to reintroduce
functional T.G.F. beta receptor.

And we can figure out a way to to
transfer them again so that they can have

these they can express it again then
maybe they can start responding to T.G.F.

beta and that might work.

So what do we need to do.

Well one of the things that
we need to do is to sort of.

First of all in vitro demonstrate what
will happen if I take a bunch of cells.

Cell lines that these
are normal epithelial cells and

then this our cells that are on their
way towards becoming cancerous.

And this ourselves that
I will over the line.

OK And then we treat each one of
them with growth factor titrate that

would growth factor and
watch them grow and then use.

The training with T.G.F.

beta The hypothesis is that the ones that
are no no you can control the growth.

The ones that are halfway there.

You control the growth but they would if
they were not come back to normal and

the ones that are over the line.

You can't control them anymore.

And then finally don't forget now this is
all in vitro involute we need to be able

to find out if this is true or not.

And this is where things get really really
complicated and this is where I'm glad

that I'm working with people who
have access to those kinds of things

that I don't have to be they want to
worry about doing those experiments.

So here's what I was telling you
about the experiments that song is.

Actually finishing up right now you
take the B.P.H. one cell and the L. and

cap this is totally malignant and then you
put growth factor in each one of them and

then you start titrating with did you have
paid and this is what we expect to see.

OK And if we do then we'll be able
to wrap that up and publish that and

then have to wait for the in vivo.

In Vivo work.

So again let me acknowledge.

Our cousin and Sykes and funding
from all kinds of various places and

I'll be happy to answer any questions
that you may have like very much.

Very

good question.

And yes.

The forward for the benefit of those
who may not have heard the question.

Essentially is.

You see all this complicated network of
things and we just kind of traced a path

down one of them do other things interfere
with this and the answer is yes.

Natalie and it's called CROSSTALK.

And there is a significant crosstalk with
E.G.F. with with the with the E.G.F.

pathway and
we're currently actually at the last

meeting we presented to preliminary
results of the influence of.

The E.G.F. response on it and
it does modify things a little

bit especially when it comes to the total.

Response.

And so there's certain issues about time
in and if you if you give G.D. are paid.

At this time and can you rescue
they say and things of that nature.

It does affect it and but
the reason we were not.

It's almost like if you have a.

Flow sheet with several processes.

You can do the full blown thing in or

you can look at the reactor and
look at how the reactor behaves.

You're not ignoring the fact
that you have a separator.

That's going to recycle something back
later on but you want to do this and

understand and then when you get
your recycle that comes back.

You already have this or

you can now investigate the influence
of having this with it.

So that's that's what we're doing
right now and the the the paper on

the influence of the E.G.F.
because of the growth where we're putting

that together to see how it moderates the
signals Well that's an excellent question.

You.

That's.

That's that's a that's a very good
question and the reason is we

didn't postulate this we know
that cancer cells have actually

lost functional T.G.F.
beta receptors in other words

if you take if you do an assay and
you figure out how many.

Functionality you have better receptors
you have in your normal cell and

in a cancer cell.

You can see the difference and
and and as you note is

the be the translation of
the signal from T.G.F.

beta into the nucleus depends on your.

And I just said to you a better receptor
is you have type two and type one and

combination of or both of them.

So you don't have enough of
those it affects your signal.

So it's a physiologically established
fact that that they've lost.

And it's a question of how much they've
lost and can they regenerate it.

That's that what we don't know.

Well it it should be able to help.

The question that we have to ask is so

with him with a mathematical
model I can do anything I want.

So I can I can I can
propose all kinds of things

which one of them is
physiologically possible.

So for example I can say if I
were to increase the number of.

Functional to a better receptor
as by a certain amount.

This is the behavior I would get and
this will pull me out of that instability

regime the questions that I'm looking to
my people my collaborators in the biology

department a medical school is OK so
how do you do that physiologically and

so you have the model can help us
in terms of if that's not possible.

What's another alternative that we can use
to amplify the signal and that actually in

the question you ask and maybe we can come
through some of the other things that

interact with it because all we really
need is to be able to amplify the signal.

If if if it is not going to respond to
the original signal to a certain point.

Can we fight.

So some other way and we can
investigate the mathematical model and

that would be very helpful for us but
we are constantly being reminded by our

friends in the translational area and
I don't get carried away with mathematics.

What is this going to mean
physiologically so yes.

That is a fantastic idea.

For for
people who may not get what's going on and

when you manufacture
something within the cell.

Typically you have to go through
transcription and translation and then

if it's a protein you have to follow that
you have to do all kinds of things and

maybe even sometimes add some sugar story
to make it do what it needs to do and

all of those things take time.

So the dynamics of how that is produced
might actually affect this is very

thank you very much.

That's a.

That was free of charge right.

Like that you've got the dynamics
of did you have better production

of might in fact be something to look into

in our model we assume that you
can get it instantaneously.

And then to buy activity.

Can it become by active is
what we spend some time on.

But the production itself might be
something interesting to look at.

Thank you.

Cross validation in the sense of.

Yes no kidding.

Yeah.

That's what we did.

Maybe I didn't.

Maybe didn't come out that way
we we used the first set that

I showed you the one with the red dots.

We used those to fit and then the ones
that I showed you in yellow dots was

the one that we used to do the crossword
edition because yes you're right there's

a huge error and in fact we did
the data is not abundant enough.

And so we had to be careful to
select what we what we used for

fit in and what we use for validation.

But if you look into the paper if you're
interested in that the paper that

we actually go into
details about how we did.

The model fitting and
the model validation itself.

Now.

Thank you.