Patent Application
20190002521
The present invention relates to nucleic acids encoding novel chimeric antigen receptors (CARs), as well as to the CARs themselves, cells incorporating the nucleic acids and their use in therapy, in particular to methods in which they are used to facilitate a T-cell response to a selected target.
Chimeric antigen receptors (CARs), which may also be referred to as artificial T cell receptors, chimeric T cell receptors (TCR) or chimeric immunoreceptors are engineered receptors, are well known in the art. They are used primarily to transform immune effector cells, in particular T-cells, so as to provide those cells with a particular specificity. They are particularly under investigation in the field of cancer immunotherapy where they may be used in techniques such as adoptive cell transfer. In these therapies, T-cells are removed from a patient and modified so that they express receptors specific to the antigens found in a particular form of cancer. The T cells, which can then recognize and kill the cancer cells, are reintroduced into the patient.
First generation CARs provide a TCR-like signal, most commonly using CD3 zeta (z) and thereby elicit tumouricidal functions. However, the engagement of CD3z-chain fusion receptors may not suffice to elicit substantial IL-2 secretion and/or proliferation in the absence of a concomitant co-stimulatory signal. In physiological T-cell responses, optimal lymphocyte activation requires the engagement of one or more co-stimulatory receptors (signal 2) such as CD28 or 4-1BB. Consequently, T cells have also been engineered so that they receive a co-stimulatory signal in a tumour antigen-dependent manner.
An important development in this regard has been the successful design of ‘second generation CARs’ that transduce a functional antigen-dependent co-stimulatory signal in human primary T cells, permitting T-cell proliferation in addition to tumouricidal activity. Second generation CARs most commonly provide co-stimulation using modules derived from CD28 or 4-1BB. The combined delivery of co-stimulation plus a CD3 zeta signal renders second generation CARs clearly superior in terms of function, when compared to their first generation counterparts (CD3z signal alone). An example of a second generation CAR is found in U.S. Pat. No. 7,446,190.
More recently, so-called ‘third generation CARs’ have been prepared. These combine multiple signalling domains, such as CD28+4-1BB+CD3z or CD28+OX40+CD3z, to further augment potency. In the 3rd generation CARs, the signalling domains are aligned in series in the CAR endodomain and placed upstream of CD3z.
In general however, the results achieved with these third generation CARs have disappointingly represented only a marginal improvement over 2nd generation configurations.
The use of cells transformed with multiple constructs has also been suggested. For example, Kloss et al. Nature Biotechnology 2012, doi:10.1038/nbt.2459 describes the transduction of T-cells with a CAR comprising a signal activation region (CD3 zeta chain) that targets a first antigen and a chimeric co-stimulatory receptor (CCR) comprising both CD28 and 4-1BB costimulatory regions which targets a second antigen. The two constructs bind to their respective antigens with different binding affinities and this leads to a ‘tumour sensing’ effect that may enhance the specificity of the therapy with a view to reducing side effects.
It is desirable to develop systems whereby T-cells can be maintained in a state that they can grow, produce cytokines and deliver a kill signal through several repeated rounds of stimulation by antigen-expressing tumour target cells. Provision of sub-optimal co-stimulation causes T-cells to lose these effector functions rapidly upon re-stimulation, entering a state known as “anergy”. When CAR T-cells are sequentially re-stimulated in vitro, they progressively lose effector properties (eg IL-2 production, ability to proliferate) and differentiate to become more effector-like—in other words, less likely to manifest the effects of co-stimulation. This is undesirable for a cancer immunotherapy since more differentiated cells tend to have less longevity and reduced ability to undergo further growth/activation when they are stimulated repeatedly in the tumour microenvironment.
The applicants have found that effective T-cell responses may be generated using a combination of constructs in which multiple co-stimulatory regions are arranged in distinct constructs.
According to a first aspect of the present invention, there is provided an immuno-responsive cell expressing
The applicants have found that the efficacy of this system is good and in particular may be better than that achieved using conventional third generation CARs having similar elements. Constructs of the type of the invention may be called ‘parallel chimeric activating receptors’ or ‘pCAR’.
In addition, the proliferation of the cells, their ability to maintain their cytotoxic potency and to release IL-2 is maintained over many repeated rounds of stimulation with antigen-expressing tumour cells.
Without being bound by theory, the arrangement of the elements in the pCARs may be facilitating activity. For example, by definition, one of the co-stimulatory modules in a 3rd generation CAR must be placed away from its natural location close to the inner leaflet of the plasma membrane. This may cause it not to signal normally owing to impaired access to obligate membrane-associated partner molecules. Alternatively, close proximity of 2 co-stimulatory signalling modules in a 3rd generation CAR might lead to steric issues, preventing full engagement of one or more downstream signalling pathways. Both of these issues are avoided in the arrangement of the invention. Both the signalling moieties (b) and (e) may be fused directly to a transmembrane domain, ensuring that they are both adjacent to the plasma membrane within the cell. Furthermore, they may be spaced at distinct sites within the cell so that will not interact sterically with each other.
Suitable immuno-responsive cells for use in the first aspect of the invention include T-cells such as cytotoxic T-cells, helper T-cells or regulatory T-cells and Natural Killer (NK) cells. In particular, the immuno-responsive cell is a T-cell.
Suitable elements (a) above may include any suitable signalling region, including any region comprising an Immune-receptor-Tyrosine-based-Activation-Motif (ITAM), as reviewed for example by Love et al. Cold Spring Harbor Perspect. Biol 2010 2(6)1 a002485. In a particular embodiment, the signalling region comprises the intracellular domain of human CD3 [zeta] chain as described for example in U.S. Pat. No. 7,446,190, or a variant thereof.
In particular, this comprises the domain, which spans amino acid residues 52-163 of the full-length human CD3 zeta chain. It has a number of polymorphic forms (e.g. Sequence ID: gb|AAF34793.1 and gb|AAA60394.1), which are shown respectively as SEQ ID NO 1 and 2:
As used herein, the term ‘variant’ refers to a polypeptide sequence which is a naturally occurring polymorphic form of the basic sequence as well as synthetic variants, in which one or more amino acids within the chain are inserted, removed or replaced. However, the variant produces a biological effect which is similar to that of the basic sequence. For example, the variant mentioned above will act in a manner similar to that of the intracellular domain of human CD3 [zeta] chain. Amino acid substitutions may be regarded as “conservative” where an amino acid is replaced with a different amino acid in the same class with broadly similar properties. Non-conservative substitutions are where amino acids are replaced with amino acids of a different type or class.
Amino acid classes are defined as follows:
As is well known to those skilled in the art, altering the primary structure of a peptide by a conservative substitution may not significantly alter the activity of that peptide because the side-chain of the amino acid which is inserted into the sequence may be able to form similar bonds and contacts as the side chain of the amino acid which has been substituted out. This is so even when the substitution is in a region which is critical in determining the peptide's conformation.
Non-conservative substitutions may also be possible provided that these do not interrupt the function of the polypeptide as described above. Broadly speaking, fewer non-conservative substitutions will be possible without altering the biological activity of the polypeptides.
In general, variants will have amino acid sequences that will be at least 70%, for instance at least 71%, 75%, 79%, 81%, 84%, 87%, 90%, 93%, 95%, 96% or 98% identical to the basic sequence, for example SEQ ID NO 1 or SEQ ID NO 2. Identity in this context may be determined using the BLASTP computer program with SEQ ID NO 2 or a fragment, in particular a fragment as described below, as the base sequence. The BLAST software is publicly available.
The co-stimulatory signal sequence (b) is suitably located between the transmembrane domain (c) and the signalling region (a) and remote from the binding element (d). Similarly the co-stimulatory signal sequence (e) is suitably located adjacent the transmembrane domain (f) and remote from the binding element (g).
Suitable co-stimulatory signalling regions for use as elements (b) and (e) above are also well known in the art, and include members of the B7/CD28 family such as B7-1, B7-2, E7-H1, B7-H2, E7-H3, B7-H4, B7-H6, B7-H7, BTLA, CD28, CTLA-4, Gi24, ICOS, PD-1, PD-L2 or PDCD6; or ILT/CD85 family proteins such as LILRA3, LILRA4, LILRB1, LILRB2, LILRB3 or LILRB4; or tumour necrosis factor (TNF) superfamily members such as 4-1BB, BAFF, BAFF R, CD27, CD30, CD40, DR3, GITR, HVEM, LIGHT, Lymphotoxin-alpha, OX40, RELT, TACI, TL1A, TNF-alpha or TNF RII; or members of the SLAM family such as 2B4, BLAME, CD2, CD2F-10, CD48, CD8, CD84, CD229, CRACC, NTB-A or SLAM; or members of the TIM family such as TIM-1, TIM-3 or TIM-4; or other co-stimulatory molecules such as CD7, CD96, CD160, CD200, CD300a, CRTAM, DAP12, Dectin-1, DPPIV, EphB6, Integrin alpha 4 beta 1, Integrin alpha 4 beta 7/LPAM-1, LAG-3 or TSLP R.
The selection of the co-stimulatory signalling regions may be selected depending upon the particular use intended for the transformed cells. In particular, the co-stimulatory signalling regions selected for (b) and (e) above are those which may work co-operatively or synergistically together. For example, the co-stimulatory signalling regions for (b) and (e) may be selected from CD28, CD27, ICOS, 4-1BB, OX40, CD30, GITR, HVEM, DR3 or CD40.
In a particular embodiment, one of (b) or (e) is CD28 and the other is 4-1BB or OX40.
In a particular embodiment, (b) is CD28.
In another particular embodiment (e) is 4-1BB, or OX40 and in particular, is 4-1BB. In another embodiment, is CD27.
The transmembrane domains of (c) and (f) above may be the same or different but in particular are different to ensure separation of the constructs on the surface of the cell. Selection of different transmembrane domains may also enhance stability of the vector since inclusion of a direct repeat nucleic acid sequence in the viral vector renders it prone to rearrangement, with deletion of sequences between the direct repeats. Where the transmembrane domains of (c) and (f) are the same however, this risk can be reduced by modifying or “wobbling” the codons selected to encode the same protein sequence.
Suitable transmembrane domains are known in the art and include for example, CD8α, CD28, CD4 or CD3z transmembrane domains.
Where the co-stimulatory signalling region comprises CD28 as described above, the CD28 transmembrane domain represents a suitable option. The full length CD28 protein is a 220 amino acid protein of SEQ ID NO 3
where the transmembrane domain is shown in bold type.
In particular, one of the co-stimulatory signalling regions is based upon the hinge region and suitably also the transmembrane domain and endodomain of CD28. In particular, which comprises amino acids 114-220 of SEQ ID NO 3, shown below as SEQ ID NO 4.
In a particular embodiment, one of the co-stimulatory signalling regions (b) or (e) above is a modified form of SEQ ID NO 4 which includes a c-myc tag of SEQ ID No 5.
The c-myc tag is well known and is of SEQ ID NO 5
The c-myc tag may be added to the co-stimulatory signalling region (b) or (e) by insertion into the ectodomain or by replacement of a region in the ectodomain, which is therefore within the region of amino acids 1-152 of SEQ ID NO 3.
In a particularly preferred embodiment, the c-myc tag replaces MYPPPY motif in the CD28 sequence. This motif represents a potentially hazardous sequence. It is responsible for interactions between CD28 and its natural ligands, CD80 and CD86, so that it provides potential for off-target toxicity when CAR T-cells encounter a target cell that expresses either of these ligands. By replacement of this motif with a tag sequence as described above, the potential for unwanted side-effects is reduced.
Thus in a particular embodiment, the co-stimulatory signalling region (b) of the construct is of SEQ ID NO 6
Furthermore, the inclusion of a c-myc epitope means that detection of the CAR T-cells using a monoclonal antibody is facilitated. This is very useful since flow cytometric detection had proven unreliable when using some available antibodies.
In addition, the provision of a c-myc epitope tag could facilitate the antigen independent expansion of targeted CAR T-cells, for example by cross-linking of the CAR using the appropriate monoclonal antibody, either in solution or immobilised onto a solid phase (e.g. a bag).
Moreover, expression of the epitope for the anti-human c-myc antibody, 9e10, within the variable region of a TCR has previously been shown to be sufficient to enable antibody-mediated and complement mediated cytotoxicity both in vitro and in vivo (Kieback et al. (2008) Proc. Natl. Acad. Sci. USA, 105(2) 623-8). Thus, the provision of such epitope tags could also be used as a “suicide system”, whereby an antibody could be used to deplete CAR T-cells in vivo, in the event of toxicity.
The binding elements (d) and (g) will be different and will bind the same, overlapping or different epitopes. In a particular embodiment the first and second epitopes are associated with the same receptor or antigen. Thus the first and second epitopes as described above may, in some cases, be the same, or overlapping so that the binding elements (d) and (g) will compete in their binding. Alternatively, the first and second epitopes may be different and associated with the same or different antigens depending upon the particular therapy being envisaged. In one embodiment, the antigens are different but may be associated with the same disease such as the same specific cancer.
As used herein, the term ‘antigen’ refers to any member of a specific binding pair that will bind to the binding elements. Thus the term includes receptors on target cells.
Thus suitable binding elements (d) and (g) may be any element which provides the pCAR with the ability to recognize a target of interest. The target to which the pCARs of the invention are directed can be any target of clinical interest to which it would be desirable to induce a T cell response. This would include markers associated with cancers of various types, including for example, one or more ErbB receptors or the αvβ6 integrin, markers associated with prostate cancer (for example using a binding element that binds to prostate-specific membrane antigen (PSMA)), breast cancer (for example using a binding element that targets Her-2 (also known as ErbB2)) and neuroblastomas (for example using a binding element that targets GD2), melanomas, small cell or non-small cell lung carcinoma, sarcomas and brain tumours. In a particular embodiment the target is one or more ErbB dimers as described above or the receptor for colony stimulating factor-1 (CSF-1R) or the αvβ6 integrin, all of which have been implicated in several solid tumours.
The binding elements used in the pCARs of the invention may comprise antibodies that recognize a selected target. For convenience, the antibody used as the binding element is preferably a single chain antibody (scFv) or single domain antibody, from a camelid, human or other species. Single chain antibodies may be cloned from the V region genes of a hybridoma specific for a desired target. The production of such hybridomas has become routine, and the procedure will not be repeated here. A technique which can be used for cloning the variable region heavy chain (VH) and variable region light chain (VL) has been described in Orlandi et al., Proc. Natl Acad. Sci. (USA) 86: 3833-3837 (1989). Briefly, mRNA is isolated from the hybridoma cell line, and reverse transcribed into complementary DNA (cDNA), for example using a reverse transcriptase polymerase chain reaction (RT-PCR) kit. Sequence-specific primers corresponding to the sequence of the VH and VL genes are used. Sequence analysis of the cloned products and comparison to the known sequence for the VH and VL genes can be used to show that the cloned VH gene matched expectations. The VH and VL genes are then attached together, for example using an oligonucleotide encoding a (gly4-ser)3 linker.
Alternatively, a binding element of a pCAR may comprise ligands such as the T1E peptide (binds ErbB homo- and heterodimers), colony-stimulating factor-1 (CSF-1) or IL-34 (both bind to the CSF-1 receptor). The T1E peptide is a chimeric fusion protein composed of the entire mature human EGF protein, excluding the five most N-terminal amino acids (amino acids 971-975 of pro-epidermal growth factor precursor (NP_001954.2)), which have been replaced by the seven most N-terminal amino acids of the mature human TGF-α protein (amino acids 40-46 of pro-transforming growth factor alpha isoform 1 (NP_003227.1)).
In another embodiment, a binding element of a pCAR comprises an αvβ6 integrin-specific binding agent. The integrin αvρ6 is now regarded as a target in cancer as it has been found to be strongly upregulated in many types of cancer. αvβ6 has been identified as a receptor for foot-and-mouth disease virus (FMDV) in vitro by binding through an RGD motif in the viral capsid protein, VP1. As a result, as described for example in U.S. Pat. No. 8,383,593, a range of peptides derived from FMDV and in particular, peptides originating from the VP1 protein of FMDV and comprising an RGD motif showed increased binding potency and binding specificity. In particular, these peptides comprise the sequence motif
wherein LX5X6L or LX5X6I is contained within an alpha helical structure, wherein X5 and X6 are helix promoting residues, which have a conformational preference greater than 1.0 for being found in the middle of an [alpha]-helix (from Creighton, 1993 and Pace C. N. and Scholtz J. M. (1998), Biophysical Journal, Vol. 75, pages 422-427). In particular such residues are independently selected from the group consisting of Glu, Ala, Leu, Met, Gln, Lys, Arg, Val, Ile, Trp, Phe and Asp.
Specific examples of such sequences include SEQ ID Nos 9-11 or variants thereof:
These peptides may form a particular group of binding elements for the CARs of the present application.
For selected malignancies such as Hodgkin's lymphoma and some breast cancers, two natural ligands are CSF-1 and IL-34 and these form particularly suitable binding elements for (d) and (g). They do however bind with different affinities. The affinity of binding can impact on the activity observed. It may be beneficial in this case to ensure that the binding element with the lower binding affinity is used as binding element (d) and that with the higher binding affinity is used as binding element (g). In particular, in an embodiment, the relative affinity of the second generation CAR (i) for its cognate target is lower than that of the partnering TNFR-based chimeric co-stimulatory receptor (ii). This does not preclude the use of high or low affinity targeting moieties in each position provided that this relative affinity relationship is maintained. Thus in the case of the present invention, in a particular embodiment, binding element (d) is CSF-1 which has a relatively low binding affinity, whilst binding element (g) comprises IL-34 which has a higher binding affinity.
Suitably the binding element is associated with a leader sequence which facilitates expression on the cell surface. Many leader sequences are known in the art, and these include the macrophage colony stimulating factor receptor (FMS) leader sequence or CD124 leader sequence.
In a further embodiment, the cells expressing the pCAR are engineered to co-express a chimeric cytokine receptor, in particular the 4αβ chimeric cytokine receptor. In 4αβ, the ectodomain of the IL-4 receptor-α chain is joined to the transmembrane and endodomains of IL-2/15 receptor-β. This allows the selective expansion and enrichment of the genetically engineered T-cells ex vivo by the culture of these cells in a suitable support medium, which, in the case of 4αβ, would comprise IL-4 as the sole cytokine support. Similarly, the system can be used with a chimeric cytokine receptor in which the ectodomain of the IL-4 receptor-α chain is joined to the transmembrane and endodomains of another receptor that is naturally bound by a cytokine that also binds to the common γ chain.
As discussed, these cells are useful in therapy to stimulate a T-cell mediated immune response to a target cell population. Thus a second aspect of the invention provides a method for stimulating a T cell mediated immune response to a target cell population in a patient in need thereof, said method comprising administering to the patient a population of immuno-responsive cells as described above, wherein the binding elements (d) and (g) are specific for the target cell.
In a third aspect of the invention, there is provided a method for preparing an immuno-responsive cell according to any one of the preceding claims, said method comprising transducing a cell with a first nucleic acid encoding a CAR of structure (i) as defined above; and also a second nucleic acid encoding a CAR of structure (ii) as defined above.
In particular, in this method, lymphocytes from a patient are transduced with the nucleic acids encoding the CARs of (i) and (ii). In particular, T-cells are subjected to genetic modification, for example by retroviral mediated transduction, to introduce CAR coding nucleic acids into the host T-cell genome, thereby permitting stable CAR expression. They may then be reintroduced into the patient, optionally after expansion, to provide a beneficial therapeutic effect. Where the cells such as the T-cells are engineered to co-express a chimeric cytokine receptor such as 4αβ, the expansion step may include an ex vivo culture step in a medium which comprises the cytokine, such as a medium comprising IL-4 as the sole cytokine support in the case of 4αβ. Alternatively, the chimeric cytokine receptor may comprise the ectodomain of the IL-4 receptor-α chain joined to the endodomain used by a common γ cytokine with distinct properties, such as IL-7. In this setting, expansion of the cells in IL-4 may result in less cell differentiation, capitalizing on the natural ability of IL-7 to achieve this effect. In this way, selective expansion and enrichment of genetically engineered T-cells with the desired state of differentiation can be ensured.
In a fourth aspect of the invention, there is provided a combination of a first nucleic acid encoding a CAR of (i) above and a second nucleic acid encoding a CCR of (ii) above. As indicated previously, this combination is referred to as a pCAR. Suitable sequences for the nucleic acids will be apparent to a skilled person. The sequences may be optimized for use in the required immuno-responsive cell. However, in some cases, as discussed above, codons may be varied from the optimum or ‘wobbled’ in order to avoid repeat sequences. Particular examples of such nucleic acids will encode the preferred embodiments described above.
In order to achieve transduction, the nucleic acids of the fourth aspect of the invention are suitably introduced into a vector, such as a plasmid or a retroviral vector. Such vectors including plasmid vectors, or cell lines containing them form a further aspect of the invention.
The first and second nucleic acids or vectors containing them may be combined in a kit, which is supplied with a view to generating immuno-responsive cells of the first aspect of the invention in situ.
Parallel chimeric activating receptors (pCAR) encoded by the nucleic acids described above form a further aspect of the invention.
The invention will now be particularly described by way of example and with reference to the following Figures in which:
A panel of CARs targeted against the CSF-1 receptor (encoded by c-FMS), which is over-expressed in Hodgkin's lymphoma, anaplastic large cell lymphoma and some solid tumours such as triple negative breast cancer were prepared and are illustrated schematically in
The constructs SFG C28ζ and SFG CTr were cloned in the SFG retroviral vector as NcoI/XhoI fragments, ensuring that their start codons are at the site of the naturally occurring NcoI site, previously occupied by the deleted env gene. Gene expression is achieved from the Moloney murine leukaemia virus (MoMLV) long terminal repeat (LTR), which has promoter activity and virus packaging of the RNA is ensured by the MoMLV ψ packaging signal, which is flanked by splice donor and acceptor sites.
All other constructs were designed and cloned using the Polymerase Incomplete Primer Extension (PIPE) cloning method. PIPE cloning method is a PCR-based alternative to conventional restriction enzyme- and ligation-dependent cloning methods. It eliminates the need to incorporate restriction sites, which could encode additional unwanted residues into expressed proteins. The PIPE method relies on the inefficiency of the amplification process in the final cycles of a PCR reaction, possibly due to the decreasing availability of dNTPs, which results in the generation of partially single-stranded (PIPE) PCR products with overhanging 5′ends. A set of vector-specific primers was used for PCR vector linearization and another set of primers with 5′-vector-end overlapping sequences then used for insert amplification, generating incomplete extension products by PIPE. In a following step, the PIPE products were mixed and the single-stranded overlapping sequences annealed and assembled as a complete SFG CAR construct. Successful cloning was confirmed by diagnostic restriction digestion. DNA sequencing was performed on all constructs to confirm that the predicted coding sequence was present, without any PCR-induced mutations (Source Bioscience, UK).
The panel included two “dual targeted” Chimeric Activating Receptors (pCARS) in which CSF-1 or IL-34 are coupled to 28z and 4-1BB, or vice versa. The dual targeted pCAR combinations were then stoichiometrically co-expressed in the same T-cell population using a Thosea Asigna (T)2A-containing retroviral vector. One of these CARs was designated ‘C34B’ (CSF1-28z plus IL34-41BB) and the other was named ‘34CB’ (IL34-28z plus CSF1-41BB).
In these dual targeted CAR T-cells, both co-stimulatory motifs (CD28/4-1BB) are placed in their natural location, close to the membrane, physically separated from each other and co-expressed in the same T-cell.
All CARs were co-expressed with an IL-4 responsive 4αβ receptor using an additional T2A element in the vector. This enables enrichment/expansion of T-cells using IL-4, making it easier to compare the function of these diverse cell populations after selection.
The main focus of the experiments was to test the behaviour of the T-cells on repeated re-stimulation with tumour target cells that either express or lack the FMS/CSF-1 receptor target. In each cycle, 1 million of the indicated IL-4 expanded CAR T-cells were suspended in RPMI+human AB serum and cultured with a confluent monolayer (24 well dish) of the antigen-expressing target (T47D FMS) or antigen null target (T47D).
Thereafter, if the CAR T-cells had persisted and destroyed the monolayer, 1 million T-cells were removed and re-stimulated in an identical manner each week. Total cell number was extrapolated at each time-point depending on the expansion of T-cells that occurred in each weekly cycle.
Throughout all of these experiments, T-cells were cultured in the absence of any exogenous cytokine such as IL-2 or IL-4—so they had to make their own cytokines in order to persist and expand. Cytokine (IFN-γ and IL-2) production was measured by ELISA in supernatants harvested from T-cell/tumour cell co-cultures, providing a second marker of effective co-stimulation.
It was found (
A representative re-stimulation experiment is shown in
In the later cycles however, the C34B dual pCAR combination (a CSF-1 targeted 28z second generation CAR co-expressed with an IL-34 targeted 4-1BB co-stimulatory motif) consistently emerged as clearly superior.
In the experiment shown in
Number of viable T-cells post monolayer destruction on consecutive cycles of Ag-stimulation was also monitored and the results are shown in
Also, so-called “exhaustion markers” on these T-cells (PD1, TIM3, 2B4 and LAG3) were also measured by flow cytometry. The results are shown in
In summary, the pCAR approach of the invention seems to maintain the cells in a state whereby they retain responsiveness to antigen through more cycles of re-stimulation. There are indications that it may retard differentiation beyond controlled memory state and it appears to delay the onset of anergy while retaining the ability of the cells to make IL-2 upon activation.
A panel of CARs used in Example 1 above were tested for anti-tumour activity using a highly aggressive in vivo xenograft model in which the CSF-1 receptor target is expressed at low levels and in which disease is disseminated throughout lymph nodes (
SCID/Beige mice were randomised into 6 groups (9 animals per group combined over two independent experiments) and were inoculated intravenously (IV) with 2×106 K299 tumour cells, re-suspended in 200 μL PBS. On day 5, the groups were treated with one of the therapeutic regimens indicated below:
Tumour growth was monitored using bioluminescence imaging (BLI) at appropriate time-points for the duration of the study.
The results are shown in
Animals were weighed throughout the experiment and no significant toxicity was noted (
Selection of Targeting Moieties to Engineer pCARs that Elicit T-Cell Activation in an αvβ6-Dependent Manner.
A panel of CARs that target αvβ6 integrin alone or together with the extended ErbB family were prepared and are shown schematically in
To create the pCAR of the invention (named TIE-41BB/A20-28z), A20-28z was co-expressed with a chimeric co-stimulatory receptor comprising a pan-ErbB targeted peptide (T1E) fused to a CD8a transmembrane and a 41BB endodomain.
Where indicated, CARs were co-expressed with the 4αβ chimeric cytokine receptor to allow for IL-4-mediated enrichment in vitro. Equimolar co-expression of the IL-4-responsive 4αβ chimeric cytokine receptor, in which the IL-4 receptor α ectodomain is fused to the transmembrane and endodomain of the shared IL-2/15 receptor β, was achieved using a Thosea Asigna (T)2A ribosomal skip peptide. These chimeric molecules were expressed in human T-cells by retroviral gene transfer.
The integrin expression pattern of cancer cell lines A375 was assessed using flow cytometry (
These data show that A20-28z CAR T-cells kill all target cells that express αvβ6 integrin (Bxpc3 and A375 β6 puro), but spare targets that lack this integrin (Panc1 and A375 puro). Secondly, the control CARs C20-28z and A20-Tr are inactive in these assays. Thirdly, T-cells that express the T1E-41BB/A20-28z pCAR cause efficient killing of target cells that express αvβ6 integrin (Bxpc3 and A375 β6 puro). All of these results are as expected. Notably however, T-cells that express the T1E-41BB/A20-28z pCAR also cause the killing of target cells that lack αvβ6 (Panc1 and A375 puro). This indicates that, within a pCAR configuration, the ability of the A20 peptide to bind non-αvβ6 integrins with low affinity is sufficient to trigger the activation of these engineered T-cells.
Production of IFN-γ by the pCAR and control engineered T-cells was then assessed. Tumour cells that lacked αvβ6 (Panc1 and A375 puro) or expressed αvβ6 (Bxpc3 and A375 puro β6) were co-cultured with genetically engineered T-cells at an effector:target ratio of 1:1 and supernatant was collected after 24, 48 or 72 hours. Levels of IFN-γ were quantified by ELISA (eBioscience). The results are shown in
Next, the CAR T-cell populations were re-stimulated bi-weekly in the absence of IL-2 support on Panc1 (αvβ6 negative) or Bxpc3 tumour cells (αvβ6 positive). Tumour cells were co-cultured with CAR T-cells derived from a patient with pancreatic ductal adenocarcinoma (PDAC) at an effector:target ratio of 1:1 (
Results are shown in
Overall, the results clearly showed that the pCAR comprising A20-28z/T1E-41BB exhibits enhanced in vitro functionality compared to a 2nd generation CAR targeted against αvβ6. Furthermore, the A20-28z/T1E-41BB+ T-cells also undergo activation by Panc1 or A375 puro cells, which express minimal to undetectable levels of this integrin. Taken with the findings obtained using the C34B pCAR (examples 1 and 2), this indicates that the pCAR configuration allows T-cell activation to occur upon serial re-stimulation when a high affinity binding interaction occurs with the 41BB CCR while a lower affinity interaction occurs with the 28z 2nd generation CAR.
Use of an Alternative TNF Receptor Family Member, CD27 to Engineer a Functional pCAR.
Using the A20-28z/T1E-41BB pCAR as starting material, additional pCARs were engineered in which the 41BB module was replaced by alternative members of the TNF receptor family, namely CD27 or
CD40. Control pCARs were engineered in which endodomains were truncated (tr). Target cells that express (Bxpc3) or lack (Panc1) αvβ6 were plated at a density of 5×104 cells per well of a 24 well plate. After 24 hours, 5×104 pCAR T-cells were added to target cells or empty wells (“unstimulated”), without exogenous cytokine support. After a further 72 hours, T-cells were harvested from the wells and were counted (
These data once again confirm the superior functionality of the A20-28z/T1E-41BB pCAR when T-cells are stimulated on Panc1 target cells, indicated by sustained T-cell proliferation and tumour cell killing. This provides further confirmation that low affinity binding of the A20 peptide to non-αvβ6 integrins is sufficient to trigger the activation of these engineered T-cells. Notably however, the A20-28z/T1E-CD27 pCAR achieved the greatest level of proliferation (
| Number | Date | Country | Kind |
|---|---|---|---|
| 1513540.3 | Jul 2015 | GB | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/GB2016/052324 | 7/28/2016 | WO | 00 |
