T cell therapies

Question 1

Animal evidence for immune system role in cancer

Answer 1

1. Mice with impaired innate or adaptive systems have increased spontaneous carcinogen-induced cancers
2. In wt mice, cancer cells with immunogenic mutations are rapidly eliminated
3. Cancers that develop in immunodeficient mice retain immunogenic mutations and are rapidly rejected by normal mice

Question 2

Human evidence for immune system role in cancer

Answer 2

1. Patients with AIDs or on immunosuppressants have more cancers (including relapse of long-dormant cancers)
2. Donor-origin tumours develop in immunosuppressed recipients of transplanted organs from asymptomatic donors

Question 3

Evidence for equilibrium between cancer and immunity

Answer 3

Injected MCA carcinogen into mice -> some develop cancer.

Some develop small tumours that do not grow progressively but if immunosuppressive, they grow.

Question 4

CTLA4 inhibitors

Answer 4

Ipilimumab in clinical trials caused cancers to get bigger, so trials stopped, however doctors realised treated patients did better. Growth due to initial inflammation, then regress.

Question 5

PD1/PDL1 inhibitors

Answer 5

[Hodi et al 2010] - when work, checkpoint inhibitors induce stable remission

Question 6

Factors contributing to efficacy of checkpoint inhibitors

Answer 6

Is PDL1 present in the tumour? Mutational load, T cell tumour infiltrate

Question 7

Increase efficacy of checkpoint inhibitors by pre-treating cancer

Answer 7

Experiment: Oncolytic virus mutated so does not infect normal cells (due to tumour suppressor genes). Inject into tumour -> infects -> inflammation and cell death -> increased antigen presentation of tumour cell peptides

Question 8

Cancer vaccines

Answer 8

1. Vaccinate against classical tumour antigens

2. Neoantigens: sequence exome/ RNA of cancer and compare vs normal cells. Then identify those highly expressed on MHC.

Question 9

Evidence for neoantigen vaccines

Answer 9

In mouse mutagen model, neoantigen-expressing tumour cells provoke T cell response and are eliminated.

In human tumours, there is a correlation between mutational load and T cell infiltration/ killing

Checkpoint inhibitor efficacy correlated with mutational load

Question 10

Sahin et al 2017

Answer 10

Clinical trial of RNA mutanome neoantigen vaccine
Identify mutations by sequencing, predicting MHC presentation.
Chose 8 mutations then made RNA vaccine specific to patients -> inject -> compare frequency of metastatic events before and after (since no other control).
Now in phase 2 clinical trials.

Question 11

Redirected T cell therapy strategies

Answer 11

1. Bispecific T cell engagers (BiTE)
2. Immune mobilising T cells against cancer (ImmTAC)
3. TCR engineered T cells
4. Other engineered T cell strategies

Question 12

BiTE

Answer 12

Soluble protein which connects T cell and cancer cell

Question 13

Blinatumomab

Answer 13

First and only approved BiTE. Is an anti-CD19 and anti-CD3 single chain antibody, use to treat B cell malignancies.
Engages CD19 in low nM range (requires 1000s molecules to target robustly). Binds CD3epsilon in low nM range.

Recruits CD4 and CD8 T cells independently of TCR. The response is not induced unless both ends of blinatumomab are engaged.

Leads to T cell activation, proliferation and cytokine production, with CD19 target cell killing. T cells serially kill.

Question 14

Dreier et al 2003

Answer 14

Preclinical data for blinatumomab.

Co-cultured blinatumomab with T cells and CD19+ cells => potent killing.

• 2:1 effector:target killing
• 1000s of CD19 molecules on target cells required for killing

Prevented tumour formation in vivo

• cohorts of immunodeficient mice inoculated with NALM6 (tumour) cells
• human PBMCs from healthy donors given (vs control)
• different doses of blinatumomab or PBS vehicle control given d0, 1, 2, 3, 4 following tumour cell/ PBMC injection

Question 15

ImmTAC

Answer 15

Soluble TCR-based targeting

Question 16

Challenges of ImmTAC

Answer 16

1. TCRs are membrane bound => engineer to be soluble
2. TCRs have natural low affinity to targets, esp cancer targets (due to -ve thymic selection)
   HLA expression down regulated in cancers
   => increase affinity of TCR for its pMHC targets.
3. Soluble TCRs lack avidity provided by coreceptors.
   A high affinity stable TCR on its own could target a cancer call but not kill it.
   => fuse TCR to an effector molecule (which binds CD3)

Question 17

How to solubilise TCRs for ImmTAC

Answer 17

Truncate TM domains and relocate stabilising inter-chain disulphide bond.
Produce in bacteria so can increase quantity.
- the two mTCR chains, alpha and beta, are made in E coli as inclusion bodies, then refolded in vitro, purified
- mTCRs are stable at room temperature for months

Question 18

How to enhance TCR affinity for ImmTAC

Answer 18

TCR contacts pMHC via 6 CDRs- CDR3s mainly contact the peptide and CDR2s the HLA.

[Yi et al 2005] - Phage display to increase affinity. Random mutation of different CDR loops, then select CDR combinations.

Affinity enhanced TCRs show increased antigen residence times.
- Wt half-life 6s vs 8hrs for 3 mutant CDRs combined. Wt binds in micro molar range vs pico molar range for engineered TCR.

[Crean et al 2020] - Affinity-enhanced TCRs have increased buried hydrophobic content at the binding interface (likely entropically favourable due to expulsion of H2O molecules that typically surround hydrophobic groups). So brings existing TCR-pMHC contacts closer rather than making more contacts.

Question 19

How to enhance TCR affinity for ImmTAC

Answer 19

TCR contacts pMHC via 6 CDRs- CDR3s mainly contact the peptide and CDR2s the HLA.

[Yi et al 2005] - Phage display to increase affinity. Random mutation of different CDR loops, then select CDR combinations.

Affinity enhanced TCRs show increased antigen residence times.
- Wt half-life 6s vs 8hrs for 3 mutant CDRs combined. Wt binds in micro molar range vs pico molar range for engineered TCR. Does this actually increase sensitivity? -> modified version of KPR (modified receptors less sensitive than wt -> TCR internalisation?)

[Crean et al 2020] - Affinity-enhanced TCRs have increased buried hydrophobic content at the binding interface (likely entropically favourable due to expulsion of H2O molecules that typically surround hydrophobic groups). So brings existing TCR-pMHC contacts closer rather than making more contacts.

[Holland et al 2020] - TCR have different pMHC binding mode to Ab TCR mimetic scaffolds.
- Contact points much more dispersed with TCR vs Ab mimetic -> specificity advantage (e.g. over BiTE)

Question 20

How fuse TCR to effector molecule in ImmTAC

Answer 20

TCR fused to anti-CD3epsilon scFV (binds CD3 with nM affinity).
Non cancer-specific T cell then kills the cancer cell in an antigen-specific manner.

Question 21

Liddy et al 2012

Answer 21

ImmTAC potency is driven by affinity -> increases sensitivity => can target cancers expressing only a few target molecules.

Looked at dose-response (target peptide conc vs IFNy release from CD8 T cells).

Question 22

Damato et al 2019

Answer 22

ImmTAC Tebentafusp redirects CD8 T cells to kill melanoma cells, sparing antigen-negative bystander cells.
Expected to be approved this year for uveal melanoma.

Question 23

What next for soluble T cell redirection agents (BiTE and ImmTAC)?

Answer 23

1. Solid tumours- targets different e.g. EpCAM, PSMA, CEA for BiTEs. Or cancer testis antigens (NY-ESO1, MAGE, PRAME) for soluble TCRs.
2. Combination treatments to enhance effect
   - enhance T cell access to tumours
   - relieve inhibitory tumour microenvironment

Question 24

TCR engineered cells

Answer 24

CAR-T cells

Adoptive TCR T cell therapy

Question 25

Adoptive TCR T cell therapy

Answer 25

Extract patient immune cells, isolate T cells, edit to express chosen TCRs (e.g. via viral vector) -> expand -> deliver to patient.

4-6 weeks

TCR editing by phage display to increase affinity. Must test specific potency to avoid off-target effects.

Optimal affinity where enhance TCR recognition of cancer targets without compromising specificity

Question 26

Linette GP et al 2015

Answer 26

NY-ESO-1 specific TCR transducer T cells cross-reactive to Titin protein in HEP2 heart cells.
Measured T cell activation by cytokine production e.g. IL2, IFNy

Question 27

Afami-cel

Answer 27

Expected to be first TCR-engineered T cell therapy approved this year for synovial carcinoma.

Target Ag is MAGE-A4.

Overall 34% response rate (16/47), including 36% (14/39) synovial carcinoma and 25% for MRCLS (2/8)

Question 28

Challenges of adoptive TCR T cell therapy

Answer 28

1. Efficacy in solid tumours
=> most TCR-T cell therapies target HLA class I, so can engineer CD4 T cells to also express CD8

2. Off-target toxicity
   => comprehensive preclinical safety analysis (empirical testing and in silico analysis to identify molecular mimicry peptides)
   * Mispairing? Possible that transgenic TCR chains may mispair with endogenous TCR chains -> off-target toxicity. Observed in mice, resulted in severe autoimmunity [Bendle et al 2010]
3. Want off-the-shelf treatments
   => universal/ allogeneic T cells more scalable, cheaper, immediate access vs autologous T cells. But challenge of this is GVHD

Question 29

Other T cell redirection strategies

Answer 29

1. HiT T cell

2. TRuC T cell

Question 30

HiT T cell

Answer 30

Natural TCRs that bind cell surface targets (not pMHC) independently of TCR expression.
Use phage display to optimise target affinity.

Preclinical data in mouse models demonstrate tumour response with mesothelia-targeting HiT T cells

Question 31

TRuC T cell

Answer 31

Fuse Ab binding domain to CD3epsilon.

Hypothesis: engagement of entire TCR complex may facilitate a broad and controlled response, allowing more potent tumour cell killing, faster migration to tumour and longer persistence.

A mesothelia TRuC is in clinical development with responses observed in ovarian cancer patients [Patrick A et al 2019]

Question 32

Discovery early 1990s

Answer 32

Zeta chain on TCR discovered. Has 3 ITAMs + short EC domain so difficult to study function by giving antibody.

Instead, fused zeta chain to other proteins for which they had Abs. Then activate T cell this way without engaging the TCR.

Question 33

Eshhar et al 1993

Answer 33

Made chimeric genes = single chain Fv domain of Ab linked to gamma or zeta chains.
Expressed chimeric genes as functional surface receptors in a cytosolic T cell hybridoma.
Triggered IL2 secretion upon entering Ag and mediated non-MHC restricted hasten-specific target-cell lysis.
Used FACS, immunoblotting/ gel electrophoresis.

Question 34

CAR T cell therapy success in CD19+ B cell leukaemia

Answer 34

Cancer caused by genetic modifications e.g. generation of fusion genes creating active kinases -> ^B cells.
Chemo to kill malignant BM cells + imatinib (kinase inhibitor).
Median 6 month survival in adults.

Question 35

Zhao et al 2015

Answer 35

Efficacy of CAR T cells with different signalling domains.

• EC domain targeting CD19 fused to zeta chain
• Murine model shows increased survival using CAR-T cells with co-stimulation
• CD28 co-stimulation allows faster cancer killing by CAR-T cells but have shorter persistence
• Authors used flow cytometry to show CAR and LNGFR expression. Measured cytotoxic activity using a (51)Cr release assay and bioluminescence assay (NALM6 as target cells)

Question 36

Brentjens et al 2013

Answer 36

Flow cytometry showed that after 11 days of treatment with CAR T cells for acute lymphoblastic leukaemia, B cell markers not detected.
This observation persists for 50 days.

Question 37

FDA approved 2 CAR T cell therapies for B cell leukaemia in 2017

Answer 37

Kymriah for ‘ALL’

Yescarta for B cell lymphoma

Question 38

CAR T cell therapy is the culmination of advances in several fields

Answer 38

• Genetic engineering
• Target selection (e.g. CD19 paradigm)
• Tumour immunology
• Synthetic receptor design
• Cell manufacturing

Question 39

Park et al 2018

Answer 39

Combined analysis of all CAR-T cell patients in a particular hospital (n = 53).
14/53 -> severe cytokine release syndrome
44/53 -> complete remission
Overall median survival = 12.9 months (i.e. patients relapse)
Patients with a low disease burden (<5% of B< are active B cells) have median survival of 20.1 months

Question 40

Roybal et al 2016

Answer 40

Want to increase specificity of CAR T cells to selectively kill cancer cells.

(in vivo murine model)
Authors took Ag fragment, fused to IC domain of notch receptor (TF that is cleaved when recognises ligand) rather than zeta chain. Transduced a CAR in this cell.

CD19 -> up regulation of CAR. Mesothelia -> activate custom response.

No IL2 produced if just target one of the two antigens. High IL2 production when target both. Showed T cell activation through CAR-GFP expression and CD69.

Problems:

1. If T cell in tissue expression both Ag, even if by different cells
2. Whenever T cell expresses CAR, is liable to response to any cell expressing mesothelin. So authors measured CAR surface expression half life (with CAR-GFP) -> 8hrs.

Question 41

Wu et al 2015

Answer 41

CAR T cell therapy can cause life threatening cytokine storms so can increase safety though remote-controlled CAR T cell activation. For example, CAR with no ITAMs but has heterodimerisation domain so can only activate cell in presence of heterodimeriser small molecule ‘Rapalog’.

Authors took mice, injected with CD19+ and tumour cells (both with fluorescent protein), then give CAR T cells, with aforementioned architecture, that target CD19. Reported robust IL2 release only when had CD19 and dimeriser.

Question 42

Eyquem et al 2017

Answer 42

Want to increase potency, since many patients relapse. T cells thought to become exhausted by chronic stimulation, and normally TCR down regulation is a mechanism of reducing stimulation. Expression of TCRs (+CARs) partly regulated by production and Ag-induced internalisation/degradation. Viral transduction of CAR into T cells -> random integration (decoupled from normal TCR expression).

Authors used CRISPR in mouse model of B cell leukaemia to KO TCR and introduce CAR to TRAC locus so have TCR-CAR+ T cells.
Improved survival in mouse model of B-ALL. Normal CAR T cells -> mice die after ~20days; mice survive > 60 days when under control of TCR promotor. These CARs maintain a lower level of expression -> may prevent exhaustion.

Question 43

Extend Ag repertoire by targeting intracellular Ag

Answer 43

1. Generate Ab specific for pMHC
   [Maus et al 2016] - made Ab that binds HLA-A2 expressing NY-ESO-1 peptide
2. Generate a TCR-based CAR.
   [Stone et al 2014] - Studied in vitro and in vivo effectiveness of single chain TCR => growth of metastatic tumours significantly inhibited. Tumour cells which escaped were Ag-loss variants.

Question 44

Boice et al 2016

Answer 44

Want to target solid tumours. CAR T cells can be engineered to manipulate the tumour microenvironment, acting as ‘micro pharmacies’.

Used B cell lymphoma system. HVEM (TNF receptor) expression is dysregulated. HVEM binds the inhibitor receptor BTLA on B cells.
Knock-down of HVEM exacerbates the cancer.
Soluble HVEM treatment reduces lymphoma burden.
CAR-T cells engineered to constitutively secrete HVEM +> reduced tumour volume between than standard CAR T cells.

Question 45

Decrease costs of CAR T cell/ adoptive TCR T cell therapies

Answer 45

Autologous T cells expensive and not always available (e.g. if T cells depleted by radio/chemotherapy).
Allogeneic T cells -> rejection or GVHD?

[Qasim et al 2017]

1. Deplete patients own cells with alemtuzumab which targets CD52
2. Prepare T cells with CD52-KO, transduced with CAR and ‘suicide gene’ so could deplete specifically if adverse reaction.

Want target Ag to be expressed on cancer but not healthy cells. And ideally not be completely unique to one patient.

• High specificity e.g. neoantigens, viral antigens, gremlin antigens e.g. NY-ESO-1 testis antigen (developmental Ag expressed in some cancers)
• Low specificity e.g. tumour differentiation antigens (CD19, CEA), over expressed antigens (ERBB2 over expressed in epithelial tumours)