CAR T cells Flashcards
CAR T cells
• Patient’s own T cells are modified to express a chimeric antigen receptor (CAR) that targets a particular cancer antigen
• CAR components
1. Antigen recognition domain = single chain variable fragment (scFv) composed of heavy and light chains of variable region (Fab) of monoclonal antibody, coupled by a linker
2. Hinge region = provides flexibility
3. Transmembrane domain
4. Intracellular T cell signalling domain = usually composed of CD3-zeta (contains ITAMs), combined with signalling domain from co-stimulatory molecule (e.g. CD28 or 4-1BB/CD137)
• Binding of scFv to tumour-associated antigen activates signalling pathways normally activated by TCR-pHLA binding facilitates T cell activation and expansion
production of CAR T cells
• Production of CAR T cells
- T cells isolated from patient by leukapheresis
- Treated with antibody-coated beads (e.g. anti-CD3, anti-CD28) and IL-2 promotes activation and expansion of T cells
- T cells genetically modified through addition of lentiviral vector containing gene encoding desired CAR
- Patient subjected to chemotherapy to deplete own T cells = generates space for CAR T cells, reduces competition for cytokines such as IL-7 and IL-15
- Modified T cells then infused back into patient
early evidence that CAR T cells could kill cancer
Eshhar et al (1993) = early evidence that CAR T cells could mediate selective killing in vitro
Transduced a CTL hybridoma with a gene encoding CAR, made up of scFv of a TNP antibody linked to FcR-gamma chain or CD3-zeta chain
Expression of CAR confirmed through immunofluorescence staining and immunoblotting of cell lysates using anti-TNP mAbs
Modified CTL hybridomas interacted specifically with TNP-modified target cells activated (detected by increase in IL-2 secretion) subsequently killed target cells
anti CD19 CARs
• 2 CAR T cell therapies now approved for treatment of B cell leukaemia = both target CD19, but contain different co-stimulatory domains
- Kymriah = scFV against CD19, signalling domains from CD3-zeta + 4-1BB
- Yescarta = scFv against CD19, signalling domains from CD3-zeta + CD28
- CD19 is expressed during B cell development, from the pro B cell stage to the mature B cell stage = not expressed on plasma cells
- CD19 also expressed on cancer cells derived from B cells at these developmental stages (e.g. pre-B acute lymphoblastic leukaemia, mature B cell leukaemia and lymphomas)
- Not expressed on HSCs or other essential cell types
- Depletion of B cells managed by monthly administration of IV Igs
trial of anti-CD19 CAR T cells
• Maude et al (2014) = anti-CD19 CAR-T cells displayed considerable efficacy in 30 patients with relapsed acute lymphoblastic leukaemia = desperate clinical need
- complete remission in 90% of patients 1 month after infusion
- event-free survival rate of 67% at 6 months
- however, all patients in the trial developed cytokine release syndrome, consisting of a systemic inflammatory response driven by increased T cell activation, with 30% of patients requiring management intensive care
- no control arm, small patient sample
why not CARs in solid tumours?
Translation into treatment for solid tumours proven challenging
- CAR-T cells unable to function effectively in immunosuppressive microenvironment
- Difficult to select suitable antigen high chance that solid tumour antigens are expressed in healthy tissues so greater ‘on-target, off-tumour’ toxicity
2nd gen CARs
Second generation CARs- combining the CAR with costimulatory domains to increase efficacy of the treatment.
Tammana et al 2010- combined the CAR with two costimulatory molecules to make a second-generation CAR. Cd28 and 4-1BB. Superior function and anti-cancer properties than the first-generation CARs.
improving specificity
CD19 CAR-T cell therapy eliminates B cell cancers expressing CD19 but also destroys all normal B cells, which also express CD19 B cell aplasia
Need to develop more effective ways for T cell therapeutics to distinguish tumours from normal tissue
Using multiple receptors to detect multiple antigens in combination
Roybal et al (2016) = developed two-antigen AND-gate circuits
Model
•Genetically engineered T cells to constitutively express synthetic Notch receptor that recognises antigen A = consisted of anti-A scFv coupled to core and transcription domains of Notch receptor
Also inserted gene encoding CAR for antigen B under control of promoter activated by Notch CAR only expressed in cells in which synNotch receptor has been activated T cells only activated in presence of antigen A AND B
Efficacy in vitro
SynNotch receptors drive CAR expression localised to tumours
•Injected CD19+ Daudi B cell lymphoblastic tumours into immunocompromised NOD SCID IL-2Ry-/- mice left flank injected with WT Daudi cells, right flank with Daudi cells also expressing surface GFP
•Infused mice with primary CD4+ and CD8+ T cells expressing anti-GFP synNotch, anti-CD19 CAR and IRES-enhanced firefly luciferase reporter (both under control of Notch-inducible promoter), which gave measure whether T cells had begun expressing CAR
luciferase expression, and hence CAR expression and T cell expansion, selectively detected in GFP+ Daudi tumour and not control GFP- tumour
Modified T cells drove antigen-specific tumour clearance in mice
implanted mice with K562 tumour expressing just CD19, or expressing CD19 and GFP, and treated with CAR AND-gate T cells or untransduced control T cells
mice with dual antigen tumours treated with CAR AND-gate T cells all survived, tumours were all cleared by day 25
mice with single antigen tumours treated with CAR AND-gate T cells showed no difference in prognosis to those treated with untransduced T cells
evidence that synNotch-gated CAR expression confines activity of T cells only to in response to multiple antigens
POSSIBLE CONCERN with AND-gate therapy
• synNotch receptor of AND-gate T cells could be activated by antigen A in one tissue, facilitating expression of CAR T cells could then migrate to another tissue expressing only antigen B and be activated to kill single-antigen bystander tissue
Study indicated this is unlikely to be a concern
• injected mice with 2 tumours = 1 expressing only CD19, 1 expressing only GFP = followed by CAR AND-gate T cells or untransduced control T cells
• growth of CD19+ tumour was identical in 2 groups = no evidence that CAR T cells can be primed by GFP+ cells then migrate to kill bystander CD19+ cells elsewhere
• also found that half-life of CAR expression after removal of synNotch stimulus was only 8 hours = unlikely to be enough time for migration and CAR-mediated T cell activation
improving safety of CART
- Safety
• CD19 CAR-T cell therapy associated with cytokine release syndrome = high fevers, hypertension, tachycardia, end organ failure
• Neurological toxicity also occurs in ~1/4 patients
• Associated with massive expansion of activated T cells release pro-inflammatory mediators (e.g. TNF, IFNy) that activate macrophages release cytokines (e.g. IL-1, IL-6) CRS and neurotoxicity
• Treatment = tocilizumab (anti-IL-6) for CRS, high-dose steroids for neurotoxicity
Wu et al (2015) = engineered ON/OFF CAR T cells, dependent on both antigen and an exogenous small molecule drug for activation
Designed CAR construct consisting of 2 components (1 = extracellular anti-CD19 antigen-binding domain, 1 with intracellular CD3ζ signalling domains), which only assemble in presence of 500nM rapalog
In vitro
• Transduced CAR construct into primary human CD4+ T cells = cells only activated to proliferate and produce IL-2 and IFNy in the presence of both CD19 cognate antigen and rapalog, either alone induced only minimal cytokine production
• Transfected CAR construct into primary human CD8+ T cells = both CD19 and rapalog required for cell-mediated cytotoxicity
In vivo
• Implanted CD19+ and CD19- tumour cells into peritoneal cavity of mice
• Selective killing of CD19+ population in mice treated with conventional anti-CD19 CAR T cells, and mice treated with ON-switch CAR T cells and rapalog
• No selective killing seen in control mice or mice treated with ON-switch CAR T cells without rapalog
ON-switch CAR T cells could improve safety of CAR therapy = allow infused T cells to be selectively regulated in temporal and titratable manner to reduce toxicity
Prevent ‘first pass’ toxicity seen with anti-HER2 CAR T cells
• first concentrate in liver, lungs and heart where they cause toxicity
• could delay activation of T cells until distributed throughout body = reduce off-target effects
Problems with ON-switch technology
• Efficacy not yet tested in humans = much more extensive investigation required
• Cannot control distribution of small molecule drug technology does not permit spatial control of CAR T cells
- Facilitated by Syn-Notch CARs or NOT Gates (CAR T cells only activated if tumour cell lacks expression of healthy antigen)
Potency of CAR T cells
- Potency
• Subset of patients on CD19 CAR-T therapy relapse
• Often due to phenomenon called antigen escape cancer cells downregulate target antigen to evade CAR T cells and escape killing
• Could generate T cells expressing CAR that recognises 2 antigens CAR T cells still effective against tumour cells that mutate and no longer express 1 of the markers
Zanetti et al (2022) = generated tandem CAR that recognises both CD19 and CD22 (pan-B marker maintained in CD19+ and CD19- relapses)
• Tested ability of CD19 CARs and CD19-CD22 tandem CARs to eliminate WT and CD19/CD22 KO acute lymphoblastic leukaemia cells in vitro
- Both eliminated WT and CD22 KO cells, and failed to eliminate double Kos
- Only tandem CARs also eliminated CD19 KO
• Tandem CARs were as efficient at eradicating leukaemia in mice as CD19 CARs targeting CD19 and CD22 simultaneously did not compromise anti-cancer efficacy
• Tandem CARs facilitated better long-term control of leukaemia in mice than CD19 CARs
- Intravenously infused CD19+ CD22+ leukemic cells into mice and treated with mock CAR, CD19 CAR or tandem CAR T cells
- All mice treated with mock CARs = quickly succumbed to disease
- All mice treated with CD19 or tandem CARs controlled leukaemia within 4 weeks, but mice treated with CD19 CARs were more prone to relapse = disease-free survival of tandem CAR-treated was double that of CD19-treated at week 13 (86% vs 43%)
Need to investigate efficacy in humans
Is targeting 2 antigens sufficient? decrease risk of antigen escape but unlikely to prevent relapse in highly heterogenous tumours
expanding repertoire
- Current CAR T cells are dependent on variable domains from antibodies so limited to extracellular protein antigens could expand range of therapy if we could target intracellular antigens/pHLA complexes
- CARs containing variable domains of antibody specific for pMHC
- TCR-based CARs
