T Cell Receptors For Potential Cancer Immunotherapy Applications: Single Chain T Cell Receptors Circumvent Mispairing with Endogenous TCR and Mediate T Cell Activity
Authors: David H. Aggen*, Adam S. Chervin, Carolina M. Soto, Thomas M. Schmitt, Boris Engels, Jennifer D. Stone, Sarah A. Richman, Kurt H. Piepenbrink, Brian M. Baker, Philip D. Greenberg, Hans Schreiber, and David M. Kranz
To improve the immune response to cancer, gene therapy with tumor-specific T cell-receptors (TCRs) provides an attractive approach to effectively arm patients T cells for tumor cell destruction. A high-affinity TCR specific for an antigenic peptide-major histocompatibility complex (pepMHC) can be introduced into patients T cells in vitro, and transfer of transduced T cells could kill cancer cells. The promise of this approach is potentially hindered by mispairing of exogenous TCR chains with endogenous TCR chains that could generate mixed TCR-heterodimers with graft-vs-host specificities. To address the issue of 'mixed' TCR-heterodimers, we investigated the use of single-chain-TCRs (scTv) that lack constant-domains to avoid mispairing and compared activity of T cells expressing scTv relative to full-length TCR in vitro and in a murine tumor model. We demonstrated that the murine scTv m33, specific for a model tumor antigen pepMHC, mediated the activity of both CD4+ and CD8+ T cells with similar sensitivity to full-length m33 TCR. The scTv avoided mispairing with endogenous alpha-beta full-length TCRs and allowed for endogenous TCR surface expression at high levels. In a murine tumor model, CD8+ m33 scTv T cells persisted, whereas CD8+ m33 full-length TCR T cells were deleted, presumably by peripheral tolerance mechanisms.
And thank you to the committee for organizing this session.
I think it's an excellent way to showcase the things Illinois
has to offer to the cancer community.
So today, I'd like to give you a little overview
of the things we've done towards engineering
single-chain T-cell receptors for cancer immunotherapy,
and specifically for adoptive therapy applications.
I'm going to start by giving you a brief introduction
to T-cell antigen recognition and adoptive T-cell therapy,
and then tell you about how we've
developed these high-affinity single-chain T-cell receptor
fusions.
And finally, I'll show you some data
on the activity of these constructs
in adoptive therapy.
So first, a brief overview of T-cell antigen recognition.
T-cells are responsible for surveying
the body for virally infected cells
and determining whether or not cells are self or non-self.
And they mediate this recognition
through an alpha-beta T-cell receptor, which
recognizes antigen in the form of peptide major histocompatibility
complexes.
And if this recognition is of sufficient affinity,
the T-cell will be activated, it will proliferate,
and it will ultimately release perforins and granzymes,
which will kill the target cell.
So a little bit more about antigen processing.
We'll take the example of a virally infected cell.
Peptides from the virus are continually
turned over by the proteasome, and they're
presented on this protein called the major histocompatibility
complex, or MHC protein.
That peptide is then presented by the virally infected cell
to the T-cell, and recognition is
mediated through this alpha-beta T-cell receptor.
And as I said, if recognition is of sufficient affinity,
ultimately this leads to death of the target cell.
So typically, the interaction between T-cell receptor
and peptide MHC has an affinity on the range of 10
to 200 micromolar, which is quite low affinity compared
to an antibody, which would be, say, a nanomolar range
of affinities.
I just want to point out a few features of the T-cell
receptor.
It consists of an alpha chain and a beta chain,
and recognition is mediated by the variable domains.
There are two membrane proximal constant domains,
but all that's required for recognition
is this variable alpha and variable beta domain.
In the context of cancer, however,
cancer antigens present a particularly difficult problem
to the immune system, in that during development,
T-cells mature in the thymus, and they're
exposed to all the cell peptide MHCs in the body.
So T-cells against many self-antigens
are deleted in this process called negative selection.
And for cancer, this is a problem
because many cancer antigens that are on the cell surface
are derived from overexpressed self-proteins.
So the T-cells that escape negative selection
have very low affinity for cancer antigens.
And our solution to that approach, this problem,
is to generate higher affinity T-cell receptors
against cancer antigens, and then
reintroduce those high affinity receptors back
into the T-cells, and let the T-cells target the tumor.
So I want to introduce you to this idea of adoptive T-cell
therapy.
So as I said, this problem of the low affinity TCR,
we've overcome by using yeast display
to engineer T-cell receptors for improved affinity.
So today, I'm going to talk about the 2C T-cell receptor,
which has low affinity for a model tumor antigen.
And Phil Holler, a former member of our lab,
was able to use yeast display and fluorescence-activated cell
sorting to engineer a version of the 2C T-cell receptor called
M33, which has 1,000-fold higher affinity for a model tumor
antigen.
In the setting of adoptive therapy,
the idea is that you'd be able to isolate the patient's T-cells
and then, using a retroviral vector,
deliver this high affinity T-cell receptor into those T-cells.
You can then expand them ex vivo and reintroduce the T-cells
back into the patient.
So this provides the potential for an individualized cancer
therapy.
Ideally, for every cancer antigen,
you'd have a specific receptor, and you'd
be able to type the tumor and use that specific receptor
on each patient.
And they've done some clinical trials at the NIH
using a T-cell receptor specific for the Mk1 melanoma antigen.
These are melanoma metastases to the brain.
And you can see, before infusion of the T-cells,
they have these metastases.
And one month later, after T-cell infusion,
the metastases have gone away.
So they've shown proof of concept
that you can use a T-cell receptor in this adoptive
therapy setting to eradicate tumor metastases.
However, the efficacy of these studies is limited.
They only saw an objective response with the Mk1 TCR
in 30% of patients.
In addition, there's autoimmune side effects
of using these specific receptors,
because they're choosing a self-antigen.
So they saw problems with toxicity
in skin, in the eye, and in the auditory system.
So we think that if you could optimize the TCR,
you could improve clinical outcomes
by improving the affinity or by selecting an optimal target
antigen.
So now that I've introduced you to adoptive T-cell therapy,
I want to describe our approach of using
high-affinity, single-chain T-cell receptor fusions.
So as I said, the TCR is an alpha-beta heterodimer.
And all that's required is this V-alpha and V-beta
for recognition of this peptide MHC.
So we developed an approach where
we made a receptor consisting of only the variable domains
of the T-cell receptor fused to intracellular signaling
domains.
And we termed these single-chain TCR variable fragments SCTV,
by analogy to the antibody single-chain FV composed
of variable heavy and variable light chain.
So these single-chain T-cell receptors
provide some advantages.
One, they can be expressed at high levels on the T-cell
surface relative to the full-length T-cell receptor.
Secondly, they avoid mispairing with the endogenous TCR.
And there's some evidence that the mispairing
between the introduced TCR and the endogenous TCR
can create receptors of unknown specificity
that would cause other side effects.
Finally, these single-chain TCRs can
mediate signaling independent of the endogenous signaling
molecule.
So in one protein, you have all that's
required to drive T-cell activity.
So using the 2C SCTV, which is low affinity,
has an affinity of 30,000 nanomolar.
And the M33 single-chain TCR, with an affinity
of 30 nanomolar that's high affinity,
we created these single-chain TCR fusions
that consist of the single-chain receptor fused
to a hinge and a number of intracellular signaling
molecules, including the CD3 zeta subunit.
We then compared these to full-length TCR constructs
that V alpha, V beta heterodimer I showed you earlier
in the crystal structure.
And as I said, the advantage of the single chain
is that the CD3 subunits are built
into the single-chain receptor.
Whereas in the full-length TCR, the CD3 subunits
that are endogenous are required to bring
that receptor to the surface.
So the surface levels of the full-length TCR
are limited by the number of CD3 subunits
that exist in the cell in this case.
But in our case, it's totally dependent
on the introduced protein.
So we wanted to see if these constructs
did afford advantages in surface expression,
and if they could mediate T-cell activity.
So we introduced these cells
into the 58 minus minus T-cell hydrodome,
and looked at surface levels of murine V beta eight,
the V beta used by the 2C and M33 receptor.
And the full-length receptors are shown here
in red and green, as in comparison
to the single-chain receptors,
which were on the surface
at five to 10-fold higher surface levels.
So we think that this could improve
the potency of the approach.
Subsequently, we did an in vitro IL-2 release assay,
where we exposed cells transduced
with single-chain TV or full-length TCR
to antigen-presenting cells.
And as I said, M33 is the high-infinity version.
When you look at the single-chain and full-length,
they released IL-2 at similar concentrations of antigen.
SiY is the model tumor antigen that 2C and M33 recognize.
In comparison, the low-affinity 2C receptors
were not activated and did not release IL-2.
So what this says is that our high-affinity single-chain
and full-length TCRs have similar sensitivity.
So now I'd like to show you
the use of these single-chain receptors
in an adoptive therapy setting.
So the next two experiments I'm gonna show
were performed by Adam Shervin in the lab.
He has a poster later this afternoon.
I'd encourage you to go look at it.
He took primary murine T cells
and transduced them with either single-chain
or full-length TCR.
He then incubated those cells
with peptide-loaded target cells
that had SiY peptide on the surface,
and then measured target cell lysis.
So in cytotoxic CD8 T cells,
the cells that are typically thought to lyse tumor cells,
the full-length and single-chain TV receptors
were able to specifically lyse cells expressing SiY,
whereas with the null peptide OVA or no peptide,
there was no killing.
So this says that the M33 single-chain
is able to redirect T cell-mediated killing.
Subsequently, Adam did the same experiment
in CD4-positive T cells,
which are usually thought to be the helper class of T cells
that release cytokines.
And surprisingly, the CD4-positive T cells
were able to kill SiY targets
with both the full-length and single-chain TV.
So we were able to redirect another class of T cells
to a model tumor antigen.
So then we wanted to test out these receptors in vivo,
and Carolina Soto, a graduate student at Roy's lab,
took a rag mouse and injected it
with B16 surgical-expressing melanoma cells.
B16's a very aggressive murine melanoma cell line.
And on tumor day five, after the tumor had been established,
she injected C57 black six T cells
that had been transduced
with either single-chain T cell receptor,
with single-chain T cell receptor
into CD4 or CD8 T cell populations
and measured tumor volume at various days.
And so you can see that the M33 single-chain CD8
and CD4-positive T cells
were able to control tumor growth
better than unmodified mock-transduced T cells.
We were also surprised to see
that the CD4-positive T cells
were able to control the tumor
better than the CD8-positive T cells.
And we think that there may be
some peripheral tolerance mechanisms that are occurring
that allow the CD8-positive high-affinity T cells
to lose potency over time.
And so the advantage of the CAR format
or the single-chain T cell receptor format
is that you can modify those intracellular signaling domains
I glossed over earlier.
So you can add other signaling components
such as 401BB is becoming more popular
in chimeric antigen receptors to try and improve potency.
But this says that our single-chain approach
would work with other tumor antigen.
So in conclusion, single-chain TB
can be expressed at high surface levels on T cells
relative to full-length TCR.
They avoid mispairing with endogenous alpha and beta chains.
I didn't have time to show that data,
but we have various lines of evidence
using close cytometry that this is the case.
We also were able to show that the single-chain TB
can mediate signaling through a high-affinity TCR
and mediate killing.
And finally, that the CD4-positive T cells
control tumor growth and persist.
And we think that using the single-chain approach
may allow us to escape these peripheral tolerance mechanisms
that, in some cases,
cause the T cells to become aneurgic over time.
And with that, I would like to thank members
of the Krantz and Roy Lab, past and present,
specifically Adam Shervin and Carolina Soto,
whose data I showed today.
They also have posters at the poster session
later this afternoon.
And Jennifer Stone, a postdoc in our lab,
she's giving a talk later this afternoon
on the chimeric antigen receptor approach,
where instead of using a single-chain TB,
you'd use a single-chain FV derived from an antibody
to target cancer antigens.
I'd also like to thank our collaborators,
Hans Schreiber at the University of Chicago,
Phil Greenberg at the University of Washington,
and Brian Baker at the University of Notre Dame,
and the Medical Scholars Program for their support.
And with that, I'll be happy to take any questions
you might have.
Thank you.
Any questions from people in the room?
Yes.
Have you engineered any T cells yet for ultimate?
So we have a number of TCRs in the lab
directed against human antigens.
And one of those is 868,
which is not specific for a cancer antigen,
but this is a human high affinity T cell receptor
that's specific for the SL9-GAG peptide.
And so we introduced this 868 single-chain TCR
into humans, into murine T cells,
but it's a human receptor.
And we've shown that it's specific for the GAG peptide
and the HLA-A2-MHC.
So we've extended it to HIV.
In the lab, we also have receptors directed against MART1.
We're working on receptors against WT1
and Survivin as other targets.
Certainly.
And the other thing I didn't mention
is that with the full-length TCR,
there is the advantage that you have CD8 helping
in that interaction.
So it's possible that you don't need
extremely high affinities with full-length TCRs.
With single-chain TCRs, it's more important.
But we think that eventually,
once we've elucidated all the targets,
we'll have receptors against all these antigens.
So it's a matter of what the best approach is
to use those antigens to attack the tumor.
All right, there's no further questions.
Thank you.