David Baltimore (Cal Tech) Part 3: The Grand Challenge: Engineering Immunity

Uploaded by ibioseminars on 01.04.2010

Hello, I'm David Baltimore, a Professor of Biology at California Institute of Technology.
This is the third in this series of talks about HIV.
Are there ways that we can ultimately conquer it?
Today's talk will be about something that we call the grand challenge,
an attempt to engineer the immune system to do things that it ordinarily doesn't do,
and in particular, to see if we can adapt this form of thinking to HIV.
So, what do we mean by reprogramming the immune system?
What we need to do is to make vaccines or therapies that can bring into play
antibodies or other kinds of antiviral proteins...
for that matter, T cells, that ordinarily are not there.
We need to genetically alter cells of the immune system
by bringing in new genes to those cells,
such that they will make proteins that can allow the immune system
to attack things it ordinarily refuses to deal with
or ordinarily is not able to deal with.
An approach like that liberates us to use modern protein design methods
to create antibody-like proteins that then we can program into
the immune system and get the immune system to make for us.
The vectors that we'll use should be able to get into the cells
that make antibodies, which are called B cells,
but the route into it is probably through the hematopoetic stem cell,
because that gives rise to B cells continually.
So, our overall goal for this project, and it actually is funded
by the Gates Foundation Global Health Program
and is considered by them a grand challenge.
The overall goal is to direct blood stem cells
to develop into B cells which are capable of making therapeutic molecules
against HIV, and if we can do that, against other pathogens.
Our immune systems are remarkable.
They can make reactions against molecules that have never been seen before.
So, if Union Carbide or Dupont goes out and makes a new kind of molecular entity
never before seen on the planet,
you put that into a rabbit, the rabbit will make antibodies against that molecular determinant.
So, it's got to be able to learn about molecules it's never seen before...
the species has never seen before.
And it works. It works terrifically well.
The repertoire of antibodies that we'll make is large, but it's not infinite.
There are holes in it.
Particularly, if the antigen doesn't present itself in a way that an antibody can see it
because the antigen is hidden away inside a molecule,
then antibodies can't get in there, because they're just big things.
And even if the immune repertoire includes the ability to make a particular kind of antibody,
we may not be able to elicit that antibody with any available antigen.
So, there are lots of situations you can imagine in which you could do better
than this remarkable system does already.
So, we have as a goal in my own laboratory
to engineer immune cells to do new, valuable things
like making antibodies or T cells or using intracellular immunogens, like
RNAi, which I talked about in the previous talk.
So, this is what we call our engineering immunity project.
And we've got 3 pieces of it, in fact.
One piece is to use intracellular immunization with RNAi.
I talked about that.
Mainly, that will be directed against HIV.
And, the second is that we want to engineer T cells
so they can react to tumor antigens
because T cells have the ability to kill tumor cells.
But, in fact, all tumors that grow up in humans have eluded immune attack
in one way or another, and we believe that there are ways of programming
the immune system so that it can do things it doesn't do ordinarily.
And we are very close to having a Phase I trial of this idea.
And finally (and it turns out to be much more difficult)
we want to engineer B cells to make pre-specified either monoclonal antibodies
or antibody-like proteins that will target HIV.
So, all of these have a common methodology.
We're making retrovirus vectors... generating lentivirus vectors,
of the kind I described previously,
to program particular abilities into hematopoetic stem cells
and then as those stem cells give rise to B cells or T cells
or, actually, other kinds of cells.
They can then carry out the specific functions that we're interested in.
So, it's a gene therapy approach to stem cell modification
with the idea of generating immunotherapeutic molecules.
It's this kind of triumvirate, as shown here:
Gene therapy, leading to stem cell therapy, leading to immunotherapy.
And, we can call it instructive immunotherapy.
It's extremely difficult. It's extremely difficult because
nobody really has gene therapy working in a reproducible manner yet,
we don't have stem cell therapies working in a reproducible manner,
and we don't have immunotherapy working.
Alright. What are the details of this?
Well, the details of it are... let's take the case of attacking cancer
through T cells.
We've got to make a vector which has none of the viral genes in it,
but has the gene of interest in it, and in this case, the gene of interest
is a gene that will direct the specificity of T cells, way down here,
to make a particular response to a tumor.
We've got to put that vector into a virus.
We've got to infect that virus into a cell.
And there it has to make proteins that will have this therapeutic activity.
What are the cells?
Well, we want hematopoetic stem cells.
Experimentally, we find those in the bone marrow of mice.
So, we take mouse bone marrow, we expose it to viruses like this,
and some fraction of the cells become infected -- the red ones,
shown through here.
We then transfer this mixture of cells into a mouse.
It's actually a mouse that's been irradiated, so it's own blood-forming system is gone.
The new cells now replace the old cells in the bones of these animals,
and the animal now makes all the different blood elements --
red blood cells, white blood cells, platelets.
But now, it makes them from this population of partially genetically modified cells.
And some fraction of those cells should be lymphocytes.
And those lymphocytes will be able to target themselves to tumor cells.
So, I show you this, not because it has to do with HIV,
but because this we have working in an experimental system in mice.
And we have been able to protect mice against really quite lethal tumors
using this form of methodology.
So, it's not totally theoretical. It has actually worked in at least one experimental setting.
So, for HIV, it's basically the same thing,
but now we want to put in genes that will be expressed in B lymphocytes
to make antibodies.
So, we put those now into a virus, the virus goes into the cells.
We now imagine this in humans.
We take out bone marrow cells.
We expose them to this virus. A fraction get infected.
We put them back into the person.
The person's immune system now derives from these modified cells.
And we get lymphocytes... B lymphocytes in this case,
which we hope will target HIV.
What we would really like to do...
and this is now carrying it to another realm of technology...
is to get rid of that whole intermediate stage where we take out the cells,
put them in cell culture, infect them with a virus, and then put them back.
What we'd rather do is put a virus into a person... a viral vector
into a person. So, same thing... we'll make our vector,
we'll put it into a virus, but now we want to put that virus
directly into a human being or into an animal,
and get it to infect the appropriate cells in the bone marrow.
Those cells then will be modified, and exactly the same consequence will occur
at the other end. That's very difficult.
It's difficult to get a virus that will find its way into the bone.
But, we're working on that, and there are ways of actually getting
these cells out of the bone into the circulation,
so we might not have to get the virus into the bone.
And there are actually ways of getting the virus in the bone
that have worked for us.
We have done this in animals, and let me show you what this took.
So, here is our virus.
And viruses have, as I've described to you in an earlier talk,
two abilities -- an ability to bind to the specific of the surface of the cell
and an ability to penetrate that cell.
But there's a whole class of viruses that do that penetration in a two-step mode.
So, they bind here.
Here it is bound.
And then the cause the cell to encircle the bound virus
in what's called an endosome
to carry that inside the cell so that it makes a complete membrane around it.
The cell then acidifies that compartment,
because that's what cells do with endosomes -- they make them acid.
The virus sees an acid environment,
and that triggers the fusion of the virus with the endosome membrane,
releasing the core of the virus into the cell.
So, what we've done is to say, let's separate that into two parts.
One part is the binding to the cell.
The second part is the fusion.
So, viruses do this with one molecule. What we did is to do it with two molecules.
And we can then make the binding molecule a very specific molecule...
an antibody, in fact.
And an antibody to this surface receptor.
That enables us to target the virus to specific cells.
So, for instance, in an experimental system, we targeted it to be lymphocytes
by making an antibody that was specific for a protein
only found on B lymphocytes, called CD20.
So, now this interaction is very specific with only one kind of cell,
and we showed that if you inject the virus into a mouse,
and the mouse carries human lymphocytes,
only the B lymphocytes from humans will be infected.
Mouse cells are not infected,
and non-B cells are not infected.
So, you can actually do this in an animal. Can you do it in a person?
Will it work with bone marrow stem cells?
Those are the problems we're investigating now.
How do you make such a virus that can bind that way?
Well, what you have to do is to get a virus packaging cell --
so, this is a cell that makes virus --
and modify it so that it has antibody on its surface,
so it has an acid fusion molecule on its surface.
And we do that by making a lot of plasmids that can be put into the cell
that encode all the necessary proteins.
So, a huge sort of engineering process, but it works... it works at least in one kind of cell line.
And we have made lots of these kinds of viruses that have specific targeting ability.
And, the kind of data that you get out is here.
This is just looking at whether the virus comes out with all the relevant components in it.
And you can see that 97% of the virus or 99.7% of the virus comes out
with the relevant components in it.
And, I won't go into the details of all of these plots.
This is the experiment, actually, that shows targeting of primary human B cells in vivo.
And, what you can see down here is that the human B cells get infected.
That's because this curve moves over.
Human T cells do not, and all the other cells -- mostly mouse cells -- do not.
So, this is a project in the lab.
And, it has, in fact, 5 components to it.
First of all, the hardest component is
that we have to engineer antibody-like proteins
that will work to inactivate HIV.
And actually, a colleague of mine at Caltech, Pamela Bjorkman
and her laboratory, is working on this issue.
And what you see here is a picture of a 4-headed antibody.
We're thinking about all sorts of things -- things with multiple heads,
things with various sorts of modifications in them.
And then we have to get them expressed.
So, we have to perfect the lentiviral expression system
so that it will make antibodies and antibody-like proteins
... will encode them ...
And that is a challenge in itself.
We have to learn how to program hematopoetic stem cells
so they develop into B cells that make the right antibodies.
And we need experimental systems in which we can do that.
The best system would be to put human cells -- human stem cells --
either in a mouse or in a culture system
and to develop all the way to secreting antibody.
And we're working on that problem.
We also need to work on this targeting methodology
so that we can get targeting to specific kinds of cells,
which will increase safety and will allow perhaps use directly in patients.
And finally, we're trying to develop a mouse..
actually adapt a development that other people have made
of a mouse that will hold the human immune system
and allow us to manipulate it in an experimental fashion.
What'll it take for this actually to go into humans?
A lot. First of all, we have to make this kind of transplant system
routine, safe, and reasonably priced.
And reasonably priced is very important, because we've promised
the Gates Foundation that we will try to make this kind of therapy available
at a reasonable price
to people who couldn't afford very high-tech medicine ordinarily.
And then, we have to be able to show that this is safe,
because other people doing these kinds of gene therapy experiments
have accidentally caused cancer, and we have to be careful that we don't do that.
We think we know how to avoid that problem,
but we're not sure.
The logistic work to get antibody secretion is ongoing,
but is daunting. All of it is daunting, but the potential benefits are enormous
because we can see ways now of attacking cancer,
HIV, and other infectious diseases that might not have been available
if this technology were not developed.
All of this actually is about therapy.
To make it into a vaccine is conceivable, but then we have to be extremely safe,
and we certainly can't take out blood stem cells into culture.
We have to be able to target directly in a person.
And so, if I was to encapsulate everything I've said to you here,
it is that we're taking big risks in this work,
and I must say a lot of junior people working in my laboratory are taking those risks,
but I don't think there's any harm in thinking big when thinking small
has led us to so little advance in these very important areas of targeting the immune system
to HIV and to cancer.
Thank you very much.
Let me just acknowledge, finally, the people who have been involved in this work:
Pamela Bjorkman's laboratory, as I said, in the protein work,
Lili Yang as the major contributor in my laboratory
who, working very closely, particularly on the targeting experiments,
with Pin Wang, who is at the University of Southern California.
We got initial funding for this whole thing from the Skirball Foundation
when it was really just a vague dream,
and it's still mostly a dream.
And the Gates Foundation has seen fit to give us a $14 million grant
over 5 years to realize this dream.
Thank you very much.