HYPOXIA-RESPONSIVE CHIMERIC ANTIGEN RECEPTORS

20220195009 · 2022-06-23

    Inventors

    Cpc classification

    International classification

    Abstract

    The present invention relates to therapeutic agents, particularly to therapeutic polypeptides and nucleic acids having the capacity for selective expression under conditions of hypoxia, cells incorporating the nucleic acids and their use in therapy, in particular in methods requiring selective expression under conditions of hypoxia, such as typically found in solid cancers. The nucleic acids encode novel hypoxia-responsive chimeric antigen receptors (CARs). The invention also relates to hypoxia-responsive regulatory nucleic acids.

    Claims

    1. A nucleic acid molecule comprising: a. a polynucleotide encoding a Chimeric Antigen Receptor (CAR), wherein the CAR comprises: (i) one or more Oxygen-Dependent Degradation Domains (ODD); and (ii) at least one polypeptide with anti-tumour properties; and b. a hypoxia-responsive regulatory nucleic acid, wherein said CAR-encoding polynucleotide is operably linked to said hypoxia-responsive regulatory nucleic acid.

    2. The nucleic acid molecule of claim 1, wherein said hypoxia-responsive regulatory nucleic acid comprises a plurality of hypoxia-responsive elements (HREs), wherein each individual HRE of said plurality of HREs independently comprises (i) an HIF binding site (HBS): 5′-(A/G)CGT(G/C)-3′ (SEQ ID NO: 1); and optionally (ii) an HIF ancillary site (HAS): 5′-CA(C/G)(G/A)(T/C/G)-3′ (SEQ ID NO: 2); or (iii) an HNF-4 site: 5′-TGACCT-3′ (SEQ ID NO: 3).

    3. The nucleic acid molecule of claim 2, wherein said HBS and HAS if present are separated by a linker, optionally wherein said linker is at least 6 nucleotides in length.

    4. The nucleic acid molecule of claim 2, wherein said plurality of HREs comprises at least one or a plurality of sequences selected from SEQ ID NOs 5-17 or sequences having at least 70%, 75%, 80%, 85%, 90%, 95% or more sequence identity to any of SEQ ID NOs 5-17.

    5. The nucleic acid molecule of claim 2, wherein said plurality is at least two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty or more individual HREs, which may be sequentially positioned or which may be spatially separate.

    6. (canceled)

    7. The nucleic acid molecule of claim 2, wherein said hypoxia-responsive regulatory nucleic acid comprises a sequence of SEQ ID NOs 18, 19, 20, 21, 22, 23, 24, 25, 26 or 27 or functional fragment thereof or homologues thereof.

    8. The nucleic acid molecule of claim 2, wherein said hypoxia-responsive regulatory nucleic acid comprises a sequence of SEQ ID NO 19 or 26 or a homologue thereof having at least 70%, 75%, 80%, 85%, 90%, 95% or more sequence identity thereto.

    9. The nucleic acid molecule of claim 2, wherein the hypoxia responsive regulatory nucleic acid is comprised in a retroviral or lentiviral vector, optionally an SFG retroviral vector.

    10. The nucleic acid molecule of claim 9, wherein the retroviral or lentiviral vector comprises an enhancer region, wherein the enhancer region comprises a plurality of HREs, optionally wherein the plurality is nine HREs which may be sequentially positioned or which may be spatially separate.

    11. (canceled)

    12. The nucleic acid molecule of claim 2, wherein said HREs are derived from any one or more of the following oxygen-responsive genes or from orthologues or paralogues thereof: erythropoietin (EPO), vascular endothelial growth factor (VEGF), phosphoglycerate kinase (PGK), glucose transporters (e.g. Glut-1), lactate dehydrogenase (LDH), aldolase (ALD), Enolase (e.g. ENO3), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), nitric oxide synthetase (NOS), Heme oxygenase, muscle glycolytic enzyme pyruvate kinase (PKM), endothelin-1 (ET 1).

    13. The nucleic acid molecule of claim 1, wherein said ODD has the sequence of SEQ ID NO: 28: X.sup.1X.sup.2LEMLAPYIXMDDDX.sup.3X.sup.4X.sup.5, where “X.sup.1-5” can be any amino acid residue, optionally wherein X.sup.1 is “L” or any conservative substitution; X.sup.2 is “D” or any conservative substitution, X.sup.3 is “F” or any conservative substitution, X.sup.4 is “Q” or any conservative substitution, X.sup.5 is “L” or any conservative substitution.

    14. The nucleic acid molecule of claim 1, wherein said ODD has the sequence of SEQ ID NO: 29, 30 or 31 or homologue thereof having at least 70%, 75%, 80%, 85%, 90%, 95% or more sequence identity thereto and comprising SEQ ID NO: 28 or the sequence of SEQ ID NO: 5 or variant thereof having at least 70%, 75%, 80%, 85%, 90%, 95% or more sequence identity to SEQ ID NO: 5, wherein said variant comprises SEQ ID NO: 4.

    15. (canceled)

    16. The nucleic acid molecule of claim 1, wherein said polypeptide with an anti-tumour property comprises: a. an extracellular antigen-specific targeting region, or b. a protein for delivery to a tumour, selected from immune stimulating antibodies; surface or intracellular receptors that confer cell activation and tumour-killing capability; T-cell Receptor (TCR); immunomodulatory cytokines (for example, IL-12, IL-15), decoy antibodies (for example, PD axis-interacting antibodies), and a protein that alters host cell function (for example, Lck, TCR zeta chain, ZAP70).

    17. (canceled)

    18. (canceled)

    19. The nucleic acid molecule of claim 1, wherein hypoxia is a condition with O.sub.2 concentration below 5%, preferably below 3%, or reduced O.sub.2 availability relative to O.sub.2 availability or partial pressure of the corresponding non-cancerous organ, tissue or cells.

    20. (canceled)

    21. The nucleic acid molecule of claim 1, wherein said CAR is selected from a first, second, third, fourth generation CAR, a split CAR design, and armoured CAR.

    22. The nucleic acid molecule of claim 1, wherein said CAR has specificity towards the ErbB family of receptors.

    23. An immunoresponsive cell comprising said nucleic acid molecule of claim 1.

    24. (canceled)

    25. (canceled)

    26. A method for the preparation of a modified immunoresponsive cell, comprising: a isolating lymphoid or myeloid-derived cells from a subject; b. modifying said cells to introduce the nucleic acid molecule of claim 1; c. expanding said modified cells ex-vivo; and d. obtaining expanded cells capable of expressing said nucleic acid molecule under conditions of hypoxia.

    27. (canceled)

    28. (canceled)

    29. A method for treatment of haematological or solid cancer, comprising administering the immunoresponsive cell of claim 23 to a patient in need thereof.

    30. (canceled)

    31. (canceled)

    32. (canceled)

    33. A pharmaceutical composition comprising the immunoresponsive cell of claim 23.

    34. (canceled)

    35. (canceled)

    36. (canceled)

    37. (canceled)

    38. (canceled)

    39. (canceled)

    40. (canceled)

    41. The method of treatment of claim 29, further comprising a preceding step of: a monitoring co-expression of at least two, three, four or all five the following genes: PGK1, SLC2A1, CA9, ALDOA and VEGFA, wherein co-expression of said genes in said subject is indicative of the subject's suitability for treatment, b. immunohistochemical staining of a tumour biopsy from the subject and assessing HIF stabilisation in the tumour or stoma, or c. monitoring T-cell infiltration (and/or of other immunoresponsive cells) to HIF stabilised regions of the tumour, wherein infiltration of the immunoresponsive cells to HIF stabilised regions of the tumour is indicative of a subject's suitability for treatment.

    Description

    BRIEF DESCRIPTION OF THE DRAWINGS

    [0247] One or more embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

    [0248] FIG. 1 shows a schematic representation of CAR T-cell immunotherapy. In the practice of CAR T-cell immunotherapy, T-cells are isolated from the cancer patient and genetically modified ex-vivo, for example using retro- or lentiviral particles or RNA electroporation. By this means, the T-cells are engineered to express a chimeric receptor (CAR) with specific binding affinity to a tumour antigen of interest. Following this genetic modification, the resultant CAR-expressing T-cells are expanded using appropriate cytokines and the expanded population is re-infused back into the patient leading to T-cell-mediated targeting of the cancer.

    [0249] FIG. 2 shows a schematic representation of oxygen sensing in the mammalian cell. Under conditions of normoxia (left), HIF1α is hydroxylated by PHD enzymes in a process that requires oxygen. Hydroxylated HIF1α is then able to bind to pVHL ubiquitin ligases, which add ubiquitin on the HIF1α molecule causing its proteasomal degradation. Under conditions of hypoxia (right), due to the lack of oxygen, HIF1α hydroxylation and degradation is blocked leading to the stabilisation of the HIF1α. Stabilised HIF1α then translocates to the nucleus, where it forms a complex with HIF1β and other molecules (such as P300 and CBP). This complex is then able to bind to HIF-binding sites (HREs) that are present upstream of hypoxia-inducible genes and activates their transcription.

    [0250] FIG. 3 shows a schematic representation of the system of the present invention in which a cytotoxic T-lymphocyte (CTL), which when in the circulation or in tissue under normal oxygen tension, will not express on its surface any artificial receptor. However, when it is located in a hypoxic region, the CTL will express a cell surface CAR that will have specific binding affinity for a cancer antigen of interest. Therefore, CTL-mediated killing will happen only when both hypoxia and the antigen of interest are present, owing to the presence of the hypoxia-responsive regulatory nucleic acid.

    [0251] FIG. 4 shows the frequency logos of nucleotides in HIF-binding or ancillary sites: A. Frequency of HIF-binding nucleotides in human hypoxia-inducible genes B. Frequency of HIF-binding nucleotides in mouse hypoxia-inducible genes C. Frequency of HIF-ancillary nucleotides in hypoxia-inducible genes. The height of each letter is representative of the frequency of occurrence of the corresponding nucleotide in each position.

    [0252] FIG. 5 shows an example of a 3 tandem HRE design. The human erythropoietin (hEPO) HRE includes 3 HREs in tandem, wherein each single HRE includes HIF-binding-linker-HIF-ancillary sequences derived from the human EPO gene. The human vascular endothelial growth factor A (hVEGFA) HRE includes 3 HREs in tandem, wherein each single HRE includes HIF-binding-linker-HIF-ancillary sequences derived from the human VEGFA gene. The human glucose transporter 3(hGLUT3) HRE includes 3 HREs in tandem, wherein each single HRE includes HIF-binding-linker-HIF-ancillary sequences derived from the human GLUT3 gene.

    [0253] FIG. 6 shows a linear map representation of constructs used to optimise the technology: A. The long terminal repeat (LTR) unmodified SFG reporter retroviral construct containing click beetle luciferase (cbluc) and enhanced green fluorescent protein (eGFP) cDNAs (reporter SFG), B. A modified reporter SFG vector in which the hEPO HRE has been inserted within the 3′ LTR, C. A modified reporter SFG vector in which the hVEGF HRE has been inserted within the 3′ LTR, D. A modified reporter SFG vector in which the hGLUT3 HRE has been inserted within the 3′ LTR.

    [0254] FIG. 7 shows the HIF1α amino acid sequence (UniProt database).

    [0255] FIG. 8 shows a linear map representation of further constructs used to optimise the technology: A. Reporter SFG vector containing cbluc luciferase-ODD fusion, B. Reporter SFG vector containing cbluc luciferase-ODD fusion and hEPO HRE LTR modification, C. Reporter SFG vector containing cbluc luciferase-ODD fusion and hVEGFA HRE LTR modification, D. Reporter SFG vector containing cbluc luciferase-ODD fusion and hGLUT3 HRE LTR modification.

    [0256] FIG. 9 shows Western blot results. These results represent HIF1α protein levels (and β-Actin reference) detected in cell lines (293T, HT1080, T47D and Jurkat) following their incubation in 0.1% oxygen and 20% oxygen. Bar chart depicts the intensity of HIF1α bands. This was calculated by plotting the bands and calculating the area under the curve (AUC) using ImageJ.

    [0257] FIG. 10 shows the gating strategy and determination of transduction efficiency (of the unmodified SFG reporter construct) by measuring eGFP fluorescence signal of transduced cells (FIGS. 6 and 8). 7-AAD negative cells (viable cells) are gated and used in evaluating eGFP fluorescence in the histogram.

    [0258] FIG. 11 shows qPCR assay validation. The graph shows the linear (y=x) relationship between thermal cycle number and the DNA amount (Log scale) in nanograms for detecting genomic TATA-box binding protein gene (TBP) and luciferase (luc; encoded by the constructs) in the transduced cells.

    [0259] FIG. 12 shows relative light unit (RLU) data obtained from 293T cells following 18 hours of 5% (A), 1% (B) and 0.1% (C) oxygen (right bars) incubation compared to their respective normoxic condition (left bars) for the indicated constructs.

    [0260] FIG. 13 shows relative light unit (RLU) data obtained following culture of 293T cells for 18 hours in 100 or 0 μM cobalt chloride for the indicated constructs.

    [0261] FIG. 14 shows mRNA expression of ErbB receptor (egfr and erbb2-4) and integrin β6 (intgb6) genes in healthy mouse tissue. In total, 13 tissues were analysed in this experiment. Tissues are ranked according to their expression level of each mRNA relative to the house keeping gene, Tbp.

    [0262] FIG. 15 shows the effect of 3 and 9 HRE copies versus the control (constitutive) in the expression of luciferase under conditions of normoxia. The inclusion of the HREs significantly silenced the expression of the downstream reporter transgene (luciferase). NT: non-transduced; Constitutive: wild-type non-HRE modified LTR; 3HRE: LTR modified to contain 3 tandem HRE elements; 9HRE: LTR modified to contain 9 tandem HRE elements. The HRE elements were derived from human EPO gene promoter. By modifying the LTRs (retroviral promoter) to contain multiple HREs, the expression of luciferase was significantly reduced under conditions of normoxia.

    [0263] FIG. 16 shows the fold induction of luciferase expression under conditions of hypoxia (calculated by dividing gene expression under conditions of hypoxia with that observed under conditions of normoxia). Constitutive: wild-type non-HRE modified LTR; 3HRE: LTR modified to contain 3 tandem HRE elements; 9HRE: LTR modified to contain 9 tandem HRE elements. Under hypoxic conditions (0.1% O.sub.2), the expression of luciferase correlates with the number HREs included in the promoter.

    [0264] FIG. 17 shows the effect of fusing different lengths of the human HIF1α ODD (amino acid numbers are indicated) onto the C-terminus of click beetle luciferase in SFG vectors containing an unmodified LTR. Gene expression was assessed in normoxic conditions. Constitutive: no ODD addition vs fusion of different indicated lengths of ODD to luciferase.

    [0265] 17A: Constructs containing variable ODDs fused on the C-terminus of Click Beetle luciferase.

    [0266] 17B: T47D cells transduced with constructs shown in A, non-transduced (NT) or constitutive transduced (wild type non ODD modified Click Beetle luciferase) were exposed in hypoxia (0.1% oxygen) for 18 h. Fold induction is the luciferase expression induction seen in hypoxia in relative to the normoxic expression in each construct. N=3 Line=mean and error bars SEM.

    [0267] FIG. 18 shows the combination of the 9 HRE promoter architecture with the human HIF1α ODD (amino acids 401-603) fused onto the C-terminus of luciferase. This dual oxygen sensing system showed no detectable expression of luciferase under conditions of normoxia, but was switched on in hypoxic conditions (0.1% oxygen).

    [0268] FIG. 19 shows that T4-CAR T-cells reside in the liver and lung acutely after i.v. infusion. T4-CAR T-cells co-expressing a luciferase reporter were injected i.v. into NSG immunocompromised mice bearing an established subcutaneous SKOV3 tumour. CAR T-cells were tracked using an IVIS bioluminescence imager.

    [0269] (a) Shows the detected light (shown in blue/green on the picture) from the luciferase that is expressed within the T4-CAR T-cells in three mice bearing established SKOV3 human ovarian tumours implanted subcutaneously (left) and the dissected organs/tumour from a representative mouse (right), 4 days post infusion.

    [0270] (b) Quantitation of the luciferase signal in each indicated organ (n=6 individual mice). As can be seen at the 4 day timepoint post infusion, these cells preferentially reside in the lung and liver rather than the tumour.

    [0271] (c) T4-CAR T-cells have specificity for 8 homo- and heterodimers formed by the Erbb receptor family, which are expressed by most, if not all, epithelial cells. Analysis of the vital organs for mRNA expression of the Erbb family (presented relative to the housekeeping gene Tbp), demonstrated that both the lung and liver, where T4-CAR T-cells initially accumulate, are both rich sources of the CAR ligands. N=6 (biological replicates combined).

    [0272] FIG. 20 shows the median fluorescence intensity (MFI) of CAR expression on the gated detectable CAR T-cells in the indicated groups at 20 h post exposure to 0.1% oxygen (e.g. hypoxic conditions; n=3 individual CAR preparations). ‘T4’ is expressed using the standard SFG vector (LTR-based retroviral promoter) and ‘HRE-CAR’ is expressed using a modified SFG vector (9×HRE elements inserted into the LTR of the SFG vector). The encoded HRE CAR does not contain an additional ODD. Unexpectedly, the median fluorescence intensity (MFI) of CAR expression was greater in the HRE-CAR group.

    [0273] FIG. 21 shows that HypoxiCAR T-cell effector function is stringently restricted to hypoxic conditions: (a) Schematic diagram depicting the CAR constructs (with 9×HREs in tandem, not shown), and their modular arrangements (when integrated in the genome) that were transduced into human T-cells; LTR-Long terminal repeat. (b) Representative flow cytometry dot plots evaluating surface CAR and CD8α (to identify CD8.sup.+ T-cells). Data demonstrates CAR expression by live (7AAD.sup.−) CD3.sup.+ T4-CAR, HypoxiCAR or non-transduced T-cells that had been maintained in normoxic or hypoxic (0.1% O.sub.2) conditions for 18 h prior to staining and flow cytometry analysis. (c-h) Healthy donor CD3.sup.+ T-cells (n=6) were transduced to generate T4-CAR or HypoxiCAR T-cells and (c) placed into 0.1% O.sub.2 hypoxic conditions for up to 18 h prior to being transferred back to normoxic conditions, where CAR expression was evaluated at the indicated times using flow cytometry analysis. (d) The median fluorescence intensity (MFI) of CAR expression on T4-CAR and HypoxiCAR T-cells at 18 h of exposure to 0.1% O.sub.2 hypoxia from panel (c). (e) Detectable surface CAR expression on HypoxiCAR T-cells after 18 h exposure to decreasing concentrations of O.sub.2; statistical significance was evaluated in comparison to expression under normoxic conditions. (f) In vitro SKOV3 tumour cell killing by T4 CAR, HypoxiCAR, or CD3-truncated HypoxiCAR (CD3□ endodomain removed to prevent intracellular signalling) T-cells at the indicated times in normoxic and 0.1% O.sub.2 hypoxic conditions. Quantification of IL-2 (g) and IFN-γ (h) released from the T-cells used in (f). ELISA analysis was performed on media collected 72 h post exposure to SKOV3 cells, making comparison with co-cultures performed using untransduced T-cells (“T-cells”). All statistical comparisons that were conducted are shown. Bar on the bar charts shows the group mean and each dot represents an individual healthy donor in the group. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

    [0274] FIG. 22 shows in panel A) a schematic of the HypoxiCAR retroviral construct when integrated into the genome of the T-cells. HypoxiCAR T-cells were injected either i.v. or i.t. into HN3 tumour bearing NSG mice. B) 24 hours after infusion, tumours were excised, enzyme-digested and stained for markers of interest prior to flow cytometry analyses. Shown are gated HypoxiCAR T-cells (CD3.sup.+ and CD45.sup.+) residing in the indicated tissues, which were assessed for cell surface CAR expression (x axis of histogram). CAR expression was only detected in T-cells residing in the tumour. C) Quantification of surface CAR expression (as seen in B) where each dot represents an individual mouse for each respective tissue.

    [0275] FIG. 23 shows that HypoxiCAR provides tumour-selective CAR expression in SKOV3 and LL2 tumours: (a) Growth curve of SKOV3 tumours grown in NSG mice (n=6 mice). (b) Representative stacked histograms showing detectable cell surface HypoxiCAR expression in enzyme-dispersed tissues and blood of a SKOV3 tumour bearing mouse that had been injected i.v. and i.t. with HypoxiCAR T-cells 24 h prior to sacrifice. Histograms show gated live (7AAD.sup.−) Ter119.sup.− CD45.sup.+ CD3.sup.+ T-cells alongside a CAR isotype stained tumour (grey histogram) (left) and full cohort quantification of percent T-cells with detectable CAR in the respective tissues (across n=6 individual mice). (c) Equivalent experiment to that described in b, but with LL2 tumour bearing Rag2.sup.−/− mice, showing representative cell surface HypoxiCAR expression by T-cells within the respective tissues (left) and quantification of the percent CAR expressing HypoxiCAR T-cells (right) in the respective tissues (across n=3 individual mice). Bar on the bar charts shows the group mean and each dot represents an individual healthy mouse in the group. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

    [0276] FIG. 24: T4-CAR T-cells cause inflammation in healthy organs. (A) Diagram depicting T4-CAR. (B) Representative histogram showing cell surface CAR expression on live (7AAD.sup.−) CD3.sup.+ T4-CAR or non-transduced human T-cells, assessed using flow cytometry. (C-E) Day 13 post subcutaneous HN3 tumour cell inoculation, mice were infused i.v. with vehicle or 10×10.sup.6 non-transduced or T4-CAR T-cells (n=5). (C) Schematic diagram depicting the experiment. (D) Weight change of the mice. Arrow denotes T-cell infusion; cross indicates an animal that was culled because a humane endpoint had been exceeded. (E) Serum cytokines 24 h post-infusion. (F) Low-dose human ErbB-CAR/Luc T-cells (4.5×10.sup.6) were infused i.v. into SKOV3 tumour bearing NSG mice and 4 days later, bioluminescence imaging was performed on the whole body and dissected organs. (G) Quantification of the photons/s/unit area as percent of all organs (n=6), LN-inguinal lymph node, SI-small intestine. (H,I) H&E stained sections (left) and quantitation of myeloid infiltration (right) in the lung (H) and liver (I) 5 days post infusion i.v. of low-dose (4.5×10.sup.6 cells) T4-CAR or untransduced T-cells or vehicle. Arrows indicated myeloid infiltrates. (J,K) Immunohistochemistry (IHC) staining of tissue sections for reductively-activated pimonidazole in tumour bearing NSG mice (J) and quantitation of the staining (K). All experiments are representative of a biological repeat. Line charts, the dots mark mean and error bars represent s.e.m. Bar charts show mean and dots individual mice. * P<0.05, ** P<0.01.

    [0277] FIG. 25: HypoxiCAR T-cell effector function is stringently restricted to hypoxia. (A) Diagram depicting HypoxiCAR under conditions of normoxia and hypoxia. (B) Representative histograms to show cell surface CAR expression on live (7AAD.sup.−) CD3.sup.+ T4-CAR, HypoxiCAR and non-transduced human T-cells in normoxic or 18 h hypoxic (0.1% O.sub.2) conditions, assessed using flow cytometry. (C) Surface CAR expression on HypoxiCAR T-cells at the indicated times under conditions of hypoxia (0.1% O.sub.2) or normoxia assessed using flow cytometry analysis. Values were normalized to those seen at 18 h hypoxia (n=6). (D) Surface CAR expression on HypoxiCAR T-cells after 18 h exposure to 0.1, 1, 5%, 20% O.sub.2 (n=6). Values were normalized to those seen in 0.1% O.sub.2. (E-G) In vitro SKOV3 tumour cell killing by T4-CAR, HypoxiCAR, CD3□-truncated HypoxiCAR (CD3.sup.−; to prevent intracellular signalling) and non-transduced T-cells (CAR.sup.+ effector to target tumour cell ratio 1:1) in normoxic and 0.1% O.sub.2 hypoxic conditions. (F) Quantification of IL-2 and (G) IFNγ released into the media from the respective T-cells after 24 h and 48 h exposure to SKOV3 cells respectively, under normoxic and 0.1% O.sub.2 hypoxic conditions. Bar on charts shows mean and dots represent each individual healthy donor. Datapoints were collected in parallel and are representative of a biological repeat. In line charts, the dots mark mean and error bars represent s.e.m. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

    [0278] FIG. 26: HypoxiCAR T-cells provide anti-tumour efficacy without systemic toxicity. (A-C) Subcutaneous HN3 tumour-bearing NSG mice were injected both i.v. and i.t. with human HypoxiCAR T-cells (2.5×10.sup.5 cells i.t. and 7.5×10.sup.5 cells i.v.) 72 h prior to sacrifice. (A) Schematic diagram depicting the experiment. (B) Representative histograms showing surface CAR expression on live nucleated cells (7AAD.sup.−, Ter119.sup.−), CD45.sup.+ CD3.sup.+ HypoxiCAR T-cells in the indicated enzyme-dispersed tissues and blood and (C) quantification in the respective tissues across n=9 individual mice. (D-F) Sixteen days post subcutaneous HN3 tumour cell inoculation, mice were infused i.v. with either vehicle or 10×10.sup.6 T4-CAR, HypoxiCAR or non-transduced human T-cells (control) (n=4 mice). (D) Schematic diagram depicting the experiment. (E) Weight change of the mice. (F) Serum cytokines 24 h post-infusion. (G,H) low dose (4.5×10.sup.6) T4-CAR or HypoxiCAR T-cells were infused i.v. into NSG mice. Five days later the indicated tissues were excised, and myeloid infiltration was scored in the lung (G) and liver (H). (I) HN3 tumour growth curves from (D-F), arrow marking the point of CAR T-cell infusion. All experiments are representative of biological repeat. Bar charts shows the mean and each dot an individual mouse. In line charts, the dots marks the mean and error bars represent s.e.m. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

    [0279] FIG. 27: T-cells are not excluded from HIF1α stabilized regions of hypoxic squamous cell carcinomas of head and neck (SCCHN)s. (A-C) An HRE-regulated gene signature was constructed from known HRE-regulated genes in SCCHN tumours (n=528). (A) Heatmap displaying the Pearson correlation coefficient for the individual genes. (B) Signature expression based on tumour (T) stage (T1 n=48, T2 n=136, T3 n=99, T4 n=174). (C) Survival curve for patients with Stage 3 and 4 SCCHN for high and low expression of the HRE-regulated gene signature (n=87 respectively). (D) Representative IHC stained SCCHN section for HIF1α (red) and CD3 (brown) (n=60). (E-F) Abundance of intra-epithelial T-cells (IETs) in SCCHN tumours was grouped as low/absent (n=40) and high (n=55). An example of an IET is marked by a black arrow in (D). IET number was assessed against the HIF1α stabilization (H)-score of the tumour (E). For tumours in which high numbers of IETs were present, tumour infiltrating lymphocytes directly infiltrating HIF-1α stabilized regions of the tumour (H-TILs) were grouped as absent (n=6 of 55 tumours) or present (n=46 of 55 tumours). Examples of H-TILs are marked by white arrows in (D). H-IET number was assessed against the H-score of the tumour (F). (G) Confocal images of an oral tongue carcinoma stained with DAPI (nuclei; blue) and antibodies against CD3 (green) and HIF1α (red); white denotes CD3 and HIF1α co-localization. Box plots show median and upper/lower quartiles, whiskers show highest and lowest value. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

    [0280] FIG. 28: HypoxiCAR T-cells provide anti-tumor efficacy against established SKOV3 tumours. Human HypoxiCAR T-cells (10×10.sup.6 i.v.) or non transduced control T-cells were injected i.v. into NSG mice bearing established subcutaneous SKOV3 tumours. Chart shows the growth curves of the respective cohorts of mice. The arrow marking the point of CAR T-cell infusion. The dots mark the mean and error bars s.e.m.

    [0281] FIG. 29: T4 (constitutive non-HRE-modified) or HRE-modified (HRE alone, lacking ODD) CAR T-cells were cultured for 24 h with SCOV3 target cell lines at the indicated CAR+ effector to target ratios in normoxic (20% oxygen) or hypoxic conditions (0.1% oxygen). A. Shows the % of viable targets in the co-cultures following the 24 h co-culture and B. Shows the IL-2 released in the co-cultures following antigen-specific stimulation of T-cells by the targets. Data shown are means from n=4 independent experiments using T-cells from 4 independent donors for panel A, and means from n=3 independent experiments using T-cells from 3 independent donors for panel B. Error bars show SEM.

    EXAMPLES

    [0282] The invention will now be described with reference to the following examples.

    [0283] Materials and Methods

    [0284] Constructs

    [0285] Three HRE sequences, each containing three in tandem HBS from human EPO, VEGFA and GLUT3, were synthesized by GeneArt (ThermoFisher Scientific) and flanked by a NheI and an XbaI restriction sites. These sequences were sub-cloned and replaced the natural NheI/XhoI sequence within the 3′ LTR of the SFG Moloney murine leukemia virus plasmid. Specific modification of the 3′ LTR was achieved by the synthesis of a XhoI/EcoRI-flanked intermediate fragment, which contained the HREs, achieved using primers that contained the restriction enzyme sites and complementary sequences to the respective HRE cassettes. Overlapping PCR and sub-cloning of the fragment achieved insertion into the SFG vector. Next, a protein-coding sequence coding for green-emitting variant of click beetle luciferase and green fluorescent protein separated by a P2A was cloned into NcoI/XhoI site of the SFG. Restriction digestions were performed at 37° C. using enzymes and buffers purchased from New England Biolab. DNA was detected in ethidium bromide stained 1.2% agarose gels and bands of appropriate sizes as assessed according to the DNA ladder were excised and extracted from gels using QIAquick Gel Extraction Kit (Qiagen). Sticky end ligations were catalysed by T4 DNA ligase (ThermoFisher Scientific) at 16° C. for 1 hour.

    [0286] CAR/Reporter Construct Cloning

    [0287] Human T1E CAR containing SFG retroviral vector was modified to generate the constructs utilized in this study. The full-length ODD cDNA encoding amino acids 401-603 (SEQ ID NO: 29) from human HIF1α was synthesis as a gBlock® (Integrated DNA Technologies) and was appended onto the C-terminus of the CD3ζ within the T1E CAR through overlap PCR using Platinum Pfx DNA polymerase (Thermo Fisher Scientific) according to the manufacturer's instructions with the primers; 5′-TCCAGCGGCTGGGGCGCGAGGGGGCAGGGCC-3′ (SEQ ID NO: 38) and 5′-GGCCCTGCCCCCTCGCGCCCCAGCCGCTGGA-3′ (SEQ ID NO: 39). PCR products were run on 1.2% Agarose (Sigma-Aldrich) gels and product size was estimated against a 1 kb Plus DNA ladder (Thermo Fisher Scientific). Fragments of the expected size were excised and purified using the QIAquick® Gel Extraction kit. T1E CAR-ODD was cloned into the SFG vector using AgeI and XhoI restriction endonucleases (New England Biolabs) to cleave AgeI and XhoI restriction enzyme sites in the SFG plasmid and those which had been built into the T1E CAR-ODD cDNA. Vector and constructs that had been restriction endonuclease digested were purified using QIAquick PCR purification kit (QIGEN) and ligated using T4 ligase (Thermo Fisher scientific) prior to transformation into One Shot Stb13™ chemically competent E. coli (Thermo Fisher Scientific).

    [0288] Transformed E. coli were selected using ampicillin (Santa Cruz Biotechnology) containing Luria Bertani (LB) Agar (Sigma-Aldrich) plates. Transformed colonies were there grown up in LB broth (Sigma-Aldrich) with 100 μg/ml ampicillin and then purified using either QIAGEN Plasmid Midi or Maxi kits. Final constructs were sequence verified (Source BioScience). Using a similar approach, the following additional modifications were made: The constitutive reporter construct was generated using a Click Beetle Luciferase (Luc) and eGFP, separated by a viral P2A sequence, reporter construct previously generated in the lab. This was achieved by PCR amplification using Platinum Pfx DNA polymerase (Thermo Fisher Scientific) according to the manufacturer's protocol with the forward primer 5′-CCATGGTGAAGCGTGAGAAAAATG-3′ (SEQ ID NO: 40) and the reverse primer 5′-CTCGAGTTACTTGTACAGCTCGTCCATGC-3′ (SEQ ID NO: 41). The amplified product was digested with NcoI and XhoI (New England Biolabs) and cloned into the SFG vector using the NcoI and XhoI and T4 DNA ligase (Thermo Fisher Scientific). Full length ODD (as described above) was also appended onto the C-terminus of Luc from the reporter construct by overlap PCR using the primers: forward 5′-GAGAAGGCCGGCGGTGCCCCAGCCGCTGGA-3′ (SEQ ID NO: 42) and reverse 5′-CCTCAAAGCACAGTTACAGTATTCCAGGGAAGCGGAGCTACTAACTTCAG-3′ (SEQ ID NO: 43) to amplify the ODD flanked with complimentary overhangs. Subsequently, overlapping fusion PCR using primers: forward 5′-CCATGGTGAAGCGTGAGAAAAATG-3′ (SEQ ID NO: 44) and reverse 5′-CTCGAGTTACTTGTACAGCTCGTCCATGC-3′ (SEQ ID NO: 45) was performed to generate a fragment encoding Luciferase-ODD-P2A-eGFP flanked by NcoI and XhoI restriction sites, which were used to insert Luciferase-ODD-P2A-eGFP into the SFG vector. The HRE modification was targeted in the 3′ LTR of the SFG retroviral vector, as the 3′ LTR region gets copied to the 5′ LTR upon integration. DNA containing 9 tandem 5′-GGCCCTACGTGCTGTCTCACACAGCCTGTCTGAC-3′ (SEQ ID NO: 27) HRE motifs containing both HIF-binding and ancillary site was synthesized as a gBlock® (Integrated DNA Technologies) and sub-cloned into the 3′ LTR of the SFG vector between the NheI and XbaI restriction endonuclease sites using the NheI and Xba1 restriction endonucleases (New England Biolabs). The T1E CAR CD3.sup.− truncated control construct was synthesized as a gBlock® (Integrated DNA Technologies) with flanking SbfI and XhoI restriction sites and sub-cloned into the HRE-modified SFG vector using SbfI and XhoI restriction endonucleases (New England Biolabs). To generate the bicistronic Luciferase-T2A-CAR construct, a gBlock® (Integrated DNA Technologies), which was designed to include Luciferase-T2A-T1E peptide binder flanked with AgeI and NotI restriction sites, was inserted into the T1E CAR construct.

    [0289] Bacterial Transformation

    [0290] One Shot Stb13 Chemically Competent E. coli (ThermoFisher Scientific) were used for transformations. 5 μl of the ligation mixture was added into a vial of One Shot Stb13 cells that were thawed on ice. Cells were subsequently incubated on ice for 30 minutes. Next, the cells were heat-shocked (45 seconds, 42° C.), placed on ice for 2 minutes then 250 μl of S.O.C. Media was added and the vial incubated in a 37° C. bacterial shaker. The cells were spread on ampicillin (100 μg/ml) agar plates and incubated overnight at 37° C. in a humidified bacterial incubator. Colonies were picked and grown in 3 ml LB broth containing 100 μg/ml ampicillin. DNA was extracted from bacteria using QIAprep Miniprep Kit (Qiagen) according to the manufacturers protocol. DNA was quantified by nanodrop spectrophotometer at 280 nm and sequenced by Source BioScience. SnapGene software was used for sequencing alignments and verification.

    [0291] Cell Lines

    [0292] All cell lines were grown at 37° C. and 5% CO.sub.2 in a humidified incubator. Human embryonic kidney (HEK) 293, Phoenix-ECO (gift from Sandra Diebold), human fibrosarcoma cell line HT1080, BW5147.G.1.4 (purchased from ATCC), Jurkat (Clone E6-1) (ATCC) were maintained in RPMI 1640 medium (Gibco) supplemented with 10% foetal calf serum (FCS; Thermo Fisher Scientific). T47D cells were maintained in RPMI 1640 medium (Gibco) supplemented with 10% FCS and insulin (0.2 U/ml).

    [0293] SKOV3 human ovarian adenocarcinoma cells were originally purchased from ATCC and were re-authenticated for this study by ATCC. HN3 human head and neck adenocarcinoma were acquired from Ludwig Institute for Cancer Research, London and grown in D10 medium, Dulbecco's modified Eagle's medium (DMEM; Gibco) supplemented with 10% FCS and GlutaMAX (Thermo Fisher Scientific). Murine Lewis Lung carcinoma (LL2) cells were purchased from ATCC and were cultured in RPMI 1640 supplemented with 10% FCS. Cell lines were confirmed to be free of Mycoplasma for this study using the MycoAlert® Mycoplasma Detection Kit (Lonza).

    [0294] Mice

    [0295] NSG (NOD-scid IL2Rgamma.sup.null) mice were purchased from Charles River and bred internally. Balb/c Rag2.sup.−/− mice were a gift from Professor Adrian Hayday (KCL). Male mice were used for studies involving HN3 and female mice were used for studies involving SKOV3 and LL2 studies. All mice used for ectopic tumor studies were 6-8 weeks old and approximately 22 g in weight.

    [0296] Generation of Retrovirus

    [0297] To produce retrovirus with tropism for human cells, RD114 pseudotyped transient retroviral particles were generated by triple transfection using (per well of a six well plate) 1.5 μg of Peq-Pam plasmid (Moloney GagPol), 1 μg RDF plasmid (RD114 envelope) and 1.5 μg of the SFG plasmids using FuGENE HD transfection reagent into 50%-60% confluent HEK 293T cells (Promega, US). Peq-Pam, RDF and SFG plasmids were incubated in plain RPMI 1640 media (Gibco) for 15 minutes at room temperature (RT) and then added drop-wise onto the 293T cells. Retrovirus-containing supernatant was harvested after 48 hours and used to transduce human cell lines.

    [0298] Hypoxic Conditions

    [0299] A hypoxia chamber was purchased from STEMCELL Technologies (Canada) and purged with certified gas supplied by BOC containing 0.1%, 1% or 5% O.sub.2, with constant 5% CO.sub.2 and using N2 as a balance. The chamber was re-purged 1 hour after the first purge according to the manufacturer's protocol. Equal numbers of cells plated on two parallel plates where one was exposed to hypoxic conditions and the other maintained at normoxia for 18 hours. Luciferase activity was then measured using a luciferase assay (Promega, US) according to the manufacturer's protocol on a Perkin Elmer Fusion α-FP plate reader (Life Sciences). Incubation time for assessing hypoxia responsive gene expression was based on known studies. Hypoxic conditions were also mimicked using cobalt (II) chloride (Sigma-Aldrich, US) (PHD inhibitor) at a final concentration of 100 μM.

    [0300] Western Blot Analysis

    [0301] Cells were lysed in Western lysis buffer (2.5 ml 1M Tris pH 6.8, 1 g SDS, 5 ml glycerol, 17.5 ml water) containing a 1× concentration of a protease inhibitor cocktail (Thermo Scientific). Total protein in cell lysate was quantified using Pierce BCA Protein Assay Kit (ThermoFisher Scientific). 10 ug of protein from each lysate alongside with SeeBlue pre-stained protein ladder (ThermoFisher Scientific) were separated using 12% sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS PAGE) at 150V and transferred onto an activated PVDF nitrocellulose membrane (Thermo Scientific, Pierce) at 30V for 2 hours. The membrane was blocked with 1% milk in PBS 0.1% Tween-20 for 1 h at RT and then incubated with rabbit anti-HIF1α antibody (Novus Biologicals, Littleton, Colo.) in 1% milk (1:2000) overnight at 4° C. or polyclonal anti-β-Actin (1:5000; Abcam). After washing, the membrane was incubated with a secondary anti-rabbit horseradish peroxidase (HRP) goat anti-rabbit IgG antibody in 1% milk (1:5000; Invitrogen). Next, the HRP substrate 3,3′,5,5′ tetramethylbenzidine (TMB) was added to the PVDF membrane and the signal was read using a CL-XPosure Film (Thermo Scientific) and Western blot X-ray analyser.

    [0302] Quantitative PCR

    [0303] Genomic DNA was extracted from cell lines using a DNeasy Blood & Tissue Kit (QIAGEN, Germany) according to manufacturer's protocol and measured with nanodrop spectrophotometer at 280 nm absorbance. qPCR was performed using KiCqStart SYBR Green qPCR ReadyMix with ROX (purchased from Sigma-Aldrich, US) according to the manufacturer's protocol using custom designed primers to generate amplicons from Tbp, Luc or T2A sequences in the genome. The primers used were: murine Tbp 5′-TGTCTGTCGCAGTAAGAATGGA-3′ (SEQ ID NO: 46) and 5′-AAAATCCCAGACACGGTGGG-3′ (SEQ ID NO: 47), human Tbp 5′-TTTGGTGTTTGCTTCAGTCAG-3′ (SEQ ID NO: 48) and 5′-ATACCTAGAAAACAGGAGTTGCTCA-3′ (SEQ ID NO: 49), Luc 5′-ATTTGACTGCCGGCGAAATG-3′ (SEQ ID NO: 50) and 5′-AAGATTCATCGCCGACCACAT-3′ (SEQ ID NO: 51), T2A 5′-CGGAGAAAGCGCAGC-3′ (SEQ ID NO: 52) and 5′-GGGTCCGGGGTTCTCTT-3′ (SEQ ID NO: 53). Amplifications of the genes of interest were detected on an ABI 7900HT Fast Real Time PCR instrument (ThermoFisher Scientific).

    [0304] Quantitative Reverse Transcription PCR

    [0305] Healthy female C57BL/6 mice were sacrificed and the following organs were extracted: mammary gland, fat, liver, kidneys, colon, small intestine, stomach, skeletal muscle, lung, heart, brain, olfactory bulb and eyes (n=13). Organs were submerged in RNAlater (Sigma-Aldrich, US) reagent to stabilise and protect cellular RNA and kept overnight at 4° C. RNA was isolated from the tissues using PrepEase RNA Spin Kit (Affymetrix, US) according to the manufacturer's protocol and quantified using NanoDrop spectrophotometer at 280 nm. Erbb1-4 and Integrin β-6 mRNA expression was analyzed in purified mRNA by quantitative reverse transcriptase PCR using the EXPRESS One-step Superscript qRT-PCR kit (ThermoFisher Scientific), alongside assays on demand for the genes of interest which included: Egfr Mm01187858_m1, Erbb2 Mm00658541_m1 Erbb3 Mm01159999_m1, Erbb4 Mm01256793_m1, Itgb6 Mm01269869_m1, Tbp Mm01277042_m1. qRT PCR was performed using an ABI 7900HT Fast Real Time PCR instrument (ThermoFisher Scientific) and data analysis was done in Excel. RNA was stored at −80° C. Expression of all genes is represented relative to the house-keeping gene Tata-binding protein (Tbp).

    [0306] List of Primers Used:

    TABLE-US-00017 Primer name Sequence Fwd EPO HRE 5′-CCA CCT GTA GGT TTG GCA AGC TAG CGT CCG GGA AAC-3′ (SEQ ID NO: 54) Fwd GLUT3 HRE 5′-CCA CCT GTA GGT TTG GCA AGC TAG CCA CGC CTG TAA TC-3′ (SEQ ID NO: 55) fwd VEGFA HRE 5′-CCA CCT GTA GGT TTG GCA AGC TAG CCC CCC TTT GGG-3′ (SEQ ID NO: 56) Fwd frag 3 5′-GAA CCA TCA GAT downstream GTT TCC AGG-3′ Xba HRE (SEQ ID NO: 57) Fwd frag A 5′-ATC CGC CAC AAC binds in eGFP ATC GAG-3’ (SEQ ID NO: 58) Rev EPO HRE 5′-CCT GGA AAC ATC TGA TGG TTC TCT AGA CCT CAG GCC CGG-3′ (SEQ ID NO: 59) Rev frag 3 5′-GCG GGC CTC TTC downstream GCT ATT A-3′ EcoRI (SEQ ID NO: 60) Rev frag A 5′-TTG CCA AAC CTA upstream Nhe CAG GTG G-3′ HRE (SEQ ID NO: 61) fwd HRE from 5’-GGT GGT ACC GGT p3 p4 p5 CTG TAG GTT TGG CAA GCT AGC-3′ (SEQ ID NO: 62) fwd primer 5′-GAA AGA CCC CAC seq genome to CTG TAG GTT T-3′ verify (SEQ ID NO: 63) orientation of HRE fwd puro plus 5′-GCC ACG ACC GGT AgeI plus GCC GCC ACC ATC CCC buffering TGA CCC ACG CC-3′ (SEQ ID NO: 64) fwd tataa 5′-GGG TAT ATA ATG linker gilbert GAA GCT CGA ATT CTA overlap GCG-3′ (SEQ ID NO: 65) fwr HRE overlap 5′-CGA AAG GAG CGC and skip ACG ACC AAT TCA ATT Nco GGC CCT ACG TG-3′ (SEQ ID NO: 66) gagSFG seq primer 5′-CGG ATG GCC GCG AGA-3′ (SEQ ID NO: 67) qPCRfwd Luc 5′-ATT TGA CTG CCG GCG AAA TG-3′ (SEQ ID NO: 68) qPCRfwdrefmouseTBP 5’-TGT CTG TCG CAG TAA GAA TGG A-3′ (SEQ ID NO: 46) qPCRreffwdhumanTBP 5′-TTT GGT GTT TGC TTC AGT CAG-3′ (SEQ ID NO: 48) qPCRrefrevhumanTBP 5′-ATA CCT AGA AAA CAG GAG TTG CTC A-3′ (SEQ ID NO: 49) qPCRrefrevmouseTBP 5′-AAA ATC CCA GAC ACG GTG GG-3’ (SEQ ID NO: 47) qPCRrev Luc 5′-AAG ATT CAT CGC CGA CCA CAT-3′ (SEQ ID NO: 69) rev GLUT3 HRE 5′-CCT GGA AAC ATC TGA TGG TTC TCT AGA TTT GGC CAT GTT GAC TAG-3′ (SEQ ID NO: 70) rev VEGFA HRE 5′-CCT GGA AAC ATC TGA TGG TTC TCT AGA GTT CCG GGG TTA GTC AGT-3′ (SEQ ID NO: 71) rev primer seq 5′-CAC CAA AGA GTC orientation CTA AAC GAT C-3′ HRE (SEQ ID NO: 72) rev puro skip 5′-CAC GTA GGG CCA Nco site ATT GAA TTG GTC GTG CGC TCC TTT CG-3′ (SEQ ID NO: 73)

    [0307] Cell Viability

    [0308] Cells were washed twice with cold Dulbecco's Phosphate Buffered Saline (DPBS) (Gibco) and resuspended in 1× Binding Buffer supplied in the PE Annexin V Apoptosis Detection Kit (BD Biosciences). Cells were then stained with PE Annexin V and 7-Amino-Actinomycin (7-AAD) according to PE Annexin V Apoptosis Detection Kit protocol (BD Biosciences) for 15 minutes at RT in the dark, washed and resuspended in 1× Binding Buffer and analysed by flow cytometry (FACSCanto II Flow cytometer, BD Biosciences). Flow data were analysed using FlowJo software. PE Annexin V and 7-AAD negative cells are considered viable, PE Annexin V positive and 7-AAD negative cells are in early apoptosis and PE Annexin V and 7-AAD positive cells are in late apoptosis or dead.

    [0309] T-Cell Isolation

    [0310] For isolating human T-cells; blood was obtained from healthy volunteers under approval of the Guy's and St Thomas' Research Ethics Committee (REC reference 09/H0804/92). Blood was collected into Falcon tubes containing anti-coagulant (10% Citrate), mixed at 1:1 with RPMI 1640 and layered over Ficoll-Paque Plus (GE Healthcare). Samples were centrifuged at 750 g for 30 mins at 20° C. to separate the peripheral blood mononuclear (PBMC) cell fraction. The interface between the plasma and the Ficoll layer, which contained the PBMCs, was harvested using a sterile Pasteur pipette and washed in RPMI 1640. T-cells were purified from the PBMC fraction using human Pan T-cell isolation kit (Miltenyi Biotec) and isolated using a MidiMACs™ separator and LS columns (Miltenyi Biotec) according to the manufacturer's protocol. Purified human T-cells were activated using CD3/CD28 Human T-Activator Dynabeads (Gibco) at a 1:1 cell to bead ratio and seeded in tissue culture plates at 3×10.sup.6 in RPMI 1640 supplemented with 5% human serum (Sigma-Aldrich) and 1× penicillin/streptomycin. The following day, 100 IU/ml recombinant human IL-2 (PROLEUKIN) was added to the cultures.

    [0311] T-Cell and Cell Line Transduction

    [0312] To produce retrovirus with tropism for human cells, RD114 pseudotyped retroviral particles were generated by triple transfection, using Peq-Pam plasmid (Moloney GagPol), RDF plasmid (RD114 envelope) and the SFG plasmid of interest, using FuGENE HD transfection reagent (Promega), of HEK 293T cells as previously described. To produce retrovirus with murine cell tropism, Phoenix-ECO retrovirus producer cells were transfected using FuGENE HD (Promega) with the relevant plasmid. Supernatant containing viral particles were harvested and incubated with the cells of interest for at least 48 h to allow their transduction. T-cells were transduced in non-tissue culture treated plates that were pre-coated with 4 μg/cm.sup.2 RetroNectin (Takara Bio) overnight at 4° C. Prior to the retroviral transduction of human T-cells, CD3/CD28 Human T-Activator Dynabeads (Gibco) were removed and fresh IL-2 was added as stated in the T-cell isolation section. In the case of T-cell transduction with the bicistronic 4αβ-T2A-CAR construct, following T-cell transduction, human IL-4 (Peprotech) at 30 ng/ml final concentration was added to the culture to enrich the transduced T-cell population. Adherent cell lines, including SKOV-3 and HN3, were transduced with retrovirus, produced as indicated before, in media solution containing Polybrene (Santa Cruz Biotechnology Inc) at 4 μg/ml final concentration to increase infection efficiency. Cells modified to express Luc/eGFP were purified by cell sorting using BD FACSAria III (BD Biosciences) based on their eGFP fluorescence.

    [0313] In Vitro Studies

    [0314] In vitro hypoxia was achieved using a hypoxia incubator chamber (Stemcell Technologies) purged at 25 L/min for 4 mins with gas containing either; 0.1, 1, 5% O.sub.2, 5% CO.sub.2 and nitrogen as a balance (BOC), after which the chamber was sealed. This process was repeated again after 1 h. Hypoxia-mediated HIF1α stabilization was, in some cases, mimicked by using the chemical CoCl.sub.2 (Sigma-Aldrich), which inhibits HIF1α hydroxylation, at 100 μM final concentration, unless otherwise stated. In vitro cytotoxicity assays 1×10.sup.4 Luc/eGFP-expressing SKOV3 cells were seeded in 96-well tissue culture plates and transduced or non-transduced T-cells were added in the well at the indicated effector to target ratios. Co-cultures were incubated for 24, 48 and 72 h time points and target cell viability was determined by luciferase quantification (in normoxic conditions, following the addition of 1 μl of 15 mg/ml XenoLight D-luciferin (PerkinElmer) in PBS per 100 μl of media. Luminescence was quantified using a FLUOstar Omega plate reader (BMG Labtech). At the 24 and 48 h co-culture time points a sample of media was taken from the co-culture and subsequently used for IL-2 and IFNγ quantification, respectively. IL-2 was quantified using Human IL-2 ELISA Ready-SET-Go! Kit, 2nd Generation (eBioscience) as per manufacturer's protocol. IFNγ was quantified using Human IFN-gamma DuoSet ELISA kit (Bio-Techne) as per manufacturer's protocol. In both ELISAs cytokine concentration was determined by absorbance measurements at 450 nm on a Fusion alpha-FP spectrophotometer (Perkin-Elmer).

    [0315] In Vivo Studies

    [0316] Tumour cell lines (2.5×10.sup.5 cells in PBS) were inoculated by subcutaneous (s.c.) injection into female (for SKOV3 and LL2) and male (for HN3) mice that were six to eight weeks of age. Once tumours were palpable, digital caliper measurements of the long (L) and short (S) dimensions of the tumour were performed every 2 or 3 days. Tumour volume was established using the following equation: Volume=(S.sup.2×L)/2. Blood samples were taken from mice in EDTA-coated Microvette™ tubes (Sarstedt) and plasma was extracted by centrifugation of these samples at 2,000 g for 5 mins. The indicated doses of CAR T-cells were injected in 200 μl PBS through the tail vein using a 30 G needle. Tumour tissue, and other organs, for flow cytometry analyses were enzyme-digested to release single cells as previously described. In brief, tissues were minced using scalpels, and then single cells were liberated by incubation for 60 mins at 37° C. with 1 mg/ml Collagenase 1, from Clostridium Histolyticum (Sigma-Aldrich) and 0.1 mg/ml Deoxyribonuclese I (AppliChem) in RPMI (Gibco). Released cells were then passed through a 70 μm cell strainer prior to staining for flow cytometry analyses. Viable cells were numerated using a haemocytometer with trypan blue (Sigma-Aldrich) exclusion.

    [0317] Bioluminescence Imaging

    [0318] To assess luciferase bio-distribution in vivo, mice were injected intraperitoneally (i.p.) with 200 μl (15 mg/ml) XenoLight D-luciferin (PerkinElmer) in sterile PBS 10 mins prior to imaging. Animals were anesthetized for imaging and emitted light was detected using the In vivo Imaging System (IVIS®) Lumina Series III (PerkinElmer) and data analysed using the Living Image software (Perkin Elmer). Light was quantified in photons/second/unit area.

    [0319] Flow Cytometry

    [0320] Flow cytometry was performed as previously described. The following antibodies were purchased from eBioscience and were used at 1 μg/ml unless stated otherwise: anti-human CD3ε Brilliant Violet 421™ (SK7; Biolegend®), anti-human CD8α Alexa Fluor 488 (RPA-T8), anti-human CD4 PE (RPA-T4), anti-human CD45 Brilliant Violet 510™ (H130 Biolegend®), anti-mouse CD4 FITC (Clone: RM4-5), anti-mouse CD8α eFluor®450 (Clone: 53-6.7), anti-mouse CD3ε PE (Clone: 145-2C11), neutralizing anti-mouse CD16/CD32 (Clone: 2.4G2). Background staining was established using fluorescence minus one stained samples. T1E CAR was stained with a biotinylated anti-human EGF antibody (Bio-Techne: BAF236) and detected using Streptavidin APC. eGFP was detected by its native fluorescence. Dead cells and red blood cells were excluded using 1 μg/ml 7-amino actinomycin D (Cayman Chemical Company) alongside anti-Ter-119 PerCP-Cy5.5 (Ter-119; eBioscience). Data were collected on a BD FACS Canto II (BD Biosciences). Data was analyzed using FlowJo software (Freestar Inc.).

    [0321] Statistics

    [0322] Normality and homogeneity of variance were determined using a Shapiro-Wilk normality test and an F-test respectively. Statistical significance was then determined using a two-sided unpaired Students t test for parametric or Mann-Whitney test for nonparametric data using GraphPad Prism 6 software. When comparing paired data, a paired ratio Students t test was performed. A Welch's correction was applied when comparing groups with unequal variances. Statistical analysis of tumour growth curves was performed using the “CompareGrowthCurves” function of the statmod software package. No outliers were excluded from any data presented.

    [0323] Results

    [0324] HRE Design

    [0325] Based on analysis of genomic data obtained from the Ensembl database, putative HIF1-binding site (HBS), which is conserved between species and between hypoxia-induced genes, were identified. We compared the putative 6 nucleotide (nt)-long HBS from different oxygen-sensitive genes in human, mouse and rat based on the frequency of each nucleotide in each position in the 6-nt sequence, which binds HIF, and a sequence logo was constructed for human and mouse HBS (FIGS. 4A and 4B). Outside of the HBS element there is also a sequence 8 nts downstream of the genomic HBS sequence, which is associated with oxygen-controlled transcription. This site is known as HIF ancillary site (HAS) (FIG. 4C).

    [0326] The HRE design included an HBS and a HAS site separate by a 8 nt linker region taken from the genomic sequence. In the first instance, 3 sequential HBS-HAS sequences were used. Also, to see whether different HBS sequences have different sensitivities to HIF, three constructs were initially designed, each containing 3 sequential HBS-HAS (HRE for simplicity) sequences. The difference between these constructs was that the HBS in each construct was derived from different genes (FIG. 5). These genes were human Epo, human VEGFA and human GLUT-3.

    [0327] HREs in the LTR

    [0328] To stably integrate the construct into the host cell's genome we used the SFG retroviral vector with modified LTRs as previously described. The SFG vector is derived from the Moloney murine leukaemia virus (MMLV). We attempted to modify the retroviral enhancer region within the LTRs without affecting the integration of the transgene into the host cell genome. This has previously been achieved by cloning HREs in to the NheI/XbaI site of the LTR, which is upstream the viral promoter. In order to avoid inactivating the vector or its ability to integrate into the host genome, we replaced the NheI/XbaI region with a fragment of similar length.

    [0329] DNA sequences containing our HREs sequences that include 5′ NheI and 3′ XbaI restriction sites were synthesized by GeneArt. These sequences were sub-cloned in the NheI/XbaI site in the 3′ LTR of the SFG MMLV vector. We modified the 3′ LTR but not the 5′ LTR as, when reverse transcription occurs, the modified 3′ LTR U3 region is copied to the 5′ LTR. Due to the fact that NheI/XbaI were not unique restriction sites in the SFG, we synthesised a fragment in several steps using sequential overlapping PCR, which contained unique restriction sites (XhoI/EcoRI) in order to achieve specific modification of the NheI/XbaI site in the 3′ LTR. To make an oxygen-sensing reporter construct, green-emitting variant of click beetle luciferase and green fluorescent protein separated by a P2A peptide (self-cleaving peptide) were cloned into NcoI/XhoI site of the SFG vector. The resulting constructs are shown in FIG. 7.

    [0330] ODD Addition

    [0331] We simultaneously cloned an additional set of vectors that had an ODD domain attached to the luciferase reporter to facilitate protein degradation under conditions of normoxia. HIF1α stability is controlled by oxygen-dependent hydroxylation of prolines (p402 and p564) in the ODD. This sequence was fused with a protein of interest to make the degradation of the protein oxygen-dependent. Based on the UniProt database, the ODD domain (highlighted in FIG. 7) of human HIF1α is 203 amino acids long while the mouse orthologue consists of 213 amino acids. Using overlapping PCR, we fused the amino acid sequence 557-574 from HIF1α (in bold in FIG. 7) to the C-terminus of luciferase. The exact amino acid sequence 557-574 (LDLEMLAPYIPMDDDFQL (SEQ ID NO: 74)) is conserved in human and mouse. The resulting fragment (luciferase-ODD fusion) was inserted into the LTR-modified and LTR-unmodified SFG reporter constructs, as depicted in FIG. 8.

    [0332] In subsequent experiments we fused SEQ ID Nos 29, 30, 31. All three SEQ ID Nos conferred oxygen sensitivity to the fusion partner, with optimal results being obtained with SEQ ID NO: 29, i.e. whole ODD (401-603) (FIG. 17).

    [0333] HIF1α Stability Under Normoxia or Hypoxia in Different Cell Lines

    [0334] Cell lines were cultured for 18 hours in normoxic or hypoxic conditions, 20% or 0.1% O.sub.2, respectively. The following human cell lines were screened under these conditions: HEK293 T, HT1080, T47D and Jurkat (Clone E6-1). Immediately after the 18-hour exposure, cells were lysed and a Western blot was performed to quantify HIF1α as described in the methods. In all cell lines tested, HIF1α was found to be stabilised under hypoxic conditions (0.1% O.sub.2), when compared to normoxia (20% O.sub.2) (FIG. 10). Protein was quantified using densitometry using ImageJ Software. 293T cells and HT1080 cells had the highest amount of HIF1α under hypoxic conditions, however in these cell lines there was also some HIF1α detected in normoxic conditions. T47D and Jurkat cells both had detectable HIF1α protein under hypoxic conditions but no detectable HIF1α band was seen for T47D and Jurkat cells under normoxic conditions.

    [0335] Cell Choice

    [0336] We chose to use 293T cells in initial experiments for three reasons. First, HIF1α Western blot analysis showed that 293T cells had strong expression of HIF1α protein under hypoxic conditions, at levels 5-fold higher than found in normoxia. Second, we observed that 293T are fast-growing cells when compared to T47D, allowing multiple experiments to be performed in a short time period. Third, 293T cells are the packaging cell lines that we use to produce the retrovirus. Therefore, transfection of 293T cells to produce retrovirus results in an auto-transduction of the 293T cells themselves.

    [0337] Transduction Efficiency Based on Flow Cytometry

    [0338] Since the expression of transgene in our constructs is oxygen-sensitive, we cannot rely on flow cytometry to determine accurate transduction efficiency. Flow cytometry analysis of 293T cells, which had been transduced with the constitutive luciferase-P2A-GFP construct (SFG Reporter construct), revealed a transduction efficiency in the live cell population (7-AAD negative) of 83% (FIG. 10). These results indicated that retroviral transduction method we used worked efficiently.

    [0339] Sequencing to Verify Post-Integration HRE Orientation within the LTR

    [0340] To confirm that the modifications in the 3′ LTR had been duplicated to the 5′ LTR and were correctly orientated in the integrated provirus we sequenced the 5′ LTR region after transduction. Genomic DNA was isolated from transduced 293T cells and the 5′ LTR region was amplified via PCR and run on a 1.2% agarose gel. The band of the correct length was excised, gel purified and then sequenced. Sequence analysis revealed that the HRE modifications to the 3′ LTR were correctly copied and had the correct orientation in the 5′ LTR.

    [0341] Establishment of Copy Number Assay/qPCR (Copy Number) Assay Validation

    [0342] For our assay in which we would quantitate luciferase expression under hypoxic conditions, we need to normalise our data, as not every cell would be transduced and some cells may have contained multiple copies of the reporter construct. To permit this we utilised quantitative PCR (qPCR) using the amplification of a reference gene (TBP), which is present as 2 copies in every cell (native genomic DNA), as well as that of the transgene (luciferase) to allow us to calculate the number of integrated transgenes. To design the qPCR primers, we screened multiple possible primer sequences in silico using the Ensembl database to ensure high specificity of binding. We chose primers that bind to unique sites in the genes of interest so that the amplicons produced by PCR would be indicative of reference and transgene gene amount. We designed a primer set that binds to click beetle luciferase and human and mouse TBP (since we are using both human and mouse cell lines). Using this approach, the following three sets of primers were designed: forward mouse TBP (5′-TGT CTG TCG CAG TAA GAA TGG A-3′ (SEQ ID NO: 46)) and reverse mouse TBP (5′-AAA ATC CCA GAC ACG GTG GG-3′ (SEQ ID NO: 47)) that amplify a 94 nt fragment specifically from the mouse TBP gene, forward human TBP (5′-TTT GGT GTT TGC TTC AGT CAG-3′ (SEQ ID NO: 48)) and reverse human TBP (5′-ATA CCT AGA AAA CAG GAG TTG CTC A-3′ (SEQ ID NO: 49)) that amplify a 103 nt fragment specifically from the human TBP, and forward luciferase (5′-ATT TGA CTG CCG GCG AAA TG-3′ (SEQ ID NO: 68)) and reverse luciferase (5′-AAG ATT CAT CGC CGA CCA CAT-3′ (SEQ ID NO: 69)), which amplify specifically a 90 nt fragment from luciferase transgene.

    [0343] To determine primer binding specificity (a single amplified product), we performed qPCR on genomic DNA extracted from cells and run the PCR product on an agarose gel. All PCR products gave a single band of appropriate length demonstrating that the primers were specific.

    [0344] To validate the copy number assay, genomic DNA was extracted from non-transduced cells and from cells transduced with the construct containing the click beetle luciferase. 200 ng of DNA was serially diluted (1:2) and qPCR was performed using the designed primers. Each reaction was performed in triplicate. As expected, no luciferase amplicon was detected in the DNA extracted from non-transduced cells. qPCR data generated using DNA extracted from the transduced cells demonstrated that there was a linear relationship between the qPCR signal from both luciferase and TBP primer sets and the cycle number of the reaction, validating the assay. 18-hour incubation of 293T cells in 20%, 5%, 1% and 0.1% oxygen 293T cells were transduced with retrovirus and transduction efficiency was determined by qPCR. Non-transduced 293T cells and 293T cells transduced with luciferase constructs 1-8 (A, B, C and D from FIGS. 6 and 8) were seeded and cultured in 5%, 1% and 0.1% oxygen and normoxia (20% oxygen). Following an 18-hour incubation under these conditions, luciferase expression, and cell viability were determined. Raw relative light unit (RLU) data obtained following 18 h incubation of 293T cells in 5% oxygen and normoxia indicate that an oxygen-controlled luciferase expression system had been generated (FIG. 14A). All HRE and/or ODD modified constructs gave a modest increase in RLU in hypoxia (5%) compared to normoxia, however this was not seen at lower oxygen concentrations. In general, LTR HRE modified constructs gave lower RLU compared to their LTR wild type counterparts when cells were maintained at 0.1% oxygen. Based on previous publications, more severe hypoxia tends to increase the fold induction in protein expression under the hypoxic vs the normoxic condition. However, we did not see this trend in our data (FIG. 12C).

    [0345] The effect of adding the ODD domain within the construct is best assessed by comparing the constitutively expressing unmodified LTR construct +/−ODD. See FIGS. 17 and 18. The addition of the ODD, across the experiments only modestly decreased the detection of luciferase in the conditions. It remained a possibility that the absence of a significant induction of hypoxia might have been a result of the apparatus or experimental procedure, so to exclude this, we stimulated the transduced 293T cells with 100 μM Cobalt chloride for 18 hours which mimics hypoxic conditions (by cobalt-mediated inhibition of HIF1α degradation). However, we did not observe luciferase induction in the presence of 100 μM Cobalt chloride compared to the absence of Cobalt chloride (FIG. 13).

    [0346] FIG. 29 demonstrates the superiority of the HRE promoter versus the wild type. We observed that HRE modification leads to a superior promoter, which in a hypoxic, e.g. tumour environment, drives better expression of the downstream gene in comparison with its non-modified wild type counterpart in the same conditions. FIGS. 29 A and B demonstrate that HRE-modification alone leads to superior target killing and activation capacity in T-cells in a hypoxic (solid tumour) environment at all effector:target ratios (even at low E:T such as 1:2). This is extremely important as usually the effector to target ratio in an established solid tumour in patients is low, thus the ability of HRE-CAR to be efficient at low E:T ratios is crucial and may determine CAR T-cell immunotherapy outcome. In addition, this enhanced CAR expression will only happen within the solid tumour because of its hypoxic status and therefore as the enhanced expression will be tumour-specific it would not pose any risk of off tumour toxicities higher than the risk from the WT CAR.

    [0347] Hypoxia Inducibility in the Presence of Increasing Numbers of HRE Elements in the Promoter

    [0348] As shown in FIGS. 15 and 16, hypoxia-inducibility increases with increasing numbers of HRE elements in the promoter. By modifying the LTRs (retroviral promoter) to contain multiple HREs, expression of luciferase in conditions of normoxia was effectively silenced.

    [0349] Luciferase Stability in Normoxia (+/−ODD)

    [0350] A variety of ODD segments were fused to the C-terminus of luciferase and the results are shown in FIGS. 20 and 21. SEQ ID NO: 29: ODD segment 401-603, SEQ ID NO: 30: ODD segment 530-603 and SEQ ID NO: 31: ODD segment 530-653 were tested. Addition of each of the three ODD segments resulted in reduced expression in normoxic conditions, with the combination of the 9 HRE promoter architecture with SEQ ID NO: 29 (the 401-603 ODD) showing no expression of luciferase in normoxia, but which was switched on in hypoxia (FIG. 18).

    [0351] In Vitro and In Vivo T4-CAR Results

    [0352] We utilised a pan-ErbB CAR T1E28z which has specificity towards 8/10 of the possible ErbB homo- and hetero-dimers in both mice and humans. We modified the CAR construct to concurrently co-express a reporter Click Beetle luciferase (Luc) to permit in vivo tracking once transduced into T-cells. ErbB-CAR/Luc T-cells were i.v. infused into immunocompromised NSG mice bearing subcutaneous SKOV3 ovarian cancer xenografts. The bio-distribution of the CAR T-cells was analysed 4 days post infusion. At this early time point, the majority of cells were seen to reside in the lungs and liver, while there was minimal uptake in the tumour (FIG. 19b). Profiling of organs for ErbB1-4 mRNA expression confirmed that all receptors from the family were expressed across all vital organs, including the lungs and liver where CAR T-cells were observed to accumulate post infusion.

    [0353] As hypoxia differentiates the tumour microenvironment from healthy tissues, we sought to exploit this to create a hypoxia-sensing T4-CAR. T4 is a next generation anti-ErbB CAR co-expressed with a chimeric IL-4 receptor delivering an intracellular IL-2/IL-15 signal upon binding of IL-4 to the extracellular domain, thereby providing a means to selectively enrich CAR T-cells during ex vivo expansion without affecting the CAR-dependent killing capacity of the T-cells. We engineered the anti-ErbB CAR to contain a C-terminal 203 amino acid ODD and modified the CAR promoter in the long terminal repeat to contain a series of 9 HREs, rendering the CAR selectively responsive to hypoxia when transduced into T-cells (Schematic FIG. 21). In vitro, this CAR, named ‘HypoxiCAR’, demonstrated stringent hypoxia-specific surface CAR expression in both CD4.sup.+ and CD8.sup.+ T-cell populations (FIG. 21b).

    [0354] CAR expression was highly dynamic and represented a switch that could be turned ‘on’ and ‘off’ in an O.sub.2-dependent manner (FIG. 21c). The HRE proved to be a robust promoter as, in hypoxic conditions, only slightly less total cell surface CAR expression was observed compared to the parental T4-CAR, despite equivalent transduction efficiency and equal CD4/CD8.sup.+ T-cell ratio (FIG. 21d). HypoxiCAR demonstrated a favourable sensitivity of response to environmental O.sub.2, where CAR expression was absent at O.sub.2 concentrations found in healthy organs (≥5%) but detectable at O.sub.2 levels seen in the tumour microenvironment (≤1%). Moreover, CAR expression positively correlated with the severity of hypoxia (FIG. 21e).

    [0355] Having validated HypoxiCAR's ability to sense hypoxia, we sought to investigate its ability to elicit hypoxia-dependent killing of target cells. For this, SKOV3 ovarian cancer cells were used which express ErbB1-4. Cells were seeded onto culture plates and co-incubated with T4-CAR or HypoxiCAR under normoxic (20% O.sub.2) and hypoxic (0.1% O.sub.2) conditions. Despite equivalent transduction efficiencies, HypoxiCAR displayed efficient hypoxia-dependent killing of the SKOV3 cells with no significant killing under normoxic conditions. Target-cell killing was CAR-mediated as when HypoxiCAR's intracellular tail was truncated to prevent signalling (CD3.sup.−), killing was abrogated (FIG. 210. We also assessed the secretion of both IL-2 (FIG. 21g) and IFNγ (FIG. 21g-h) in these co-cultures, two cytokines which play important role in the T-cell response. Cytokine production by hypoxiCAR T-cells was also stringently regulated such that detectable levels were only found under hypoxic conditions.

    [0356] To translate these observations in vivo, we evaluated whether HypoxiCAR could circumvent off-tumour toxicity of ErbB-CAR T-cells. This is a major hurdle that precludes their systemic administration in the clinic. To evaluate this technology in the tumour setting, HypoxiCAR T-cells were injected concurrently i.v. and i.t. in HN3 tumour-bearing NSG mice. By this means, we achieved a rapid accumulation of these cells in tumour and vital organs for ex vivo investigation (FIG. 22A). Four days after HypoxiCAR infusion, tissues were harvested, enzyme-digested and T-cells were assessed for CAR expression using flow cytometry. HypoxiCAR achieved tumour-selectivity of expression and only presented surface CAR molecules within the hypoxic tumour microenvironment, with an absence of CAR expression when T-cells were located in the blood, lungs, and liver (FIG. 22B-C). This observation was not model specific as it was also observed in NSG mice bearing SKOV3 tumours and Rag2.sup.−/− mice bearing murine Lewis lung carcinoma (LL2) tumours.

    [0357] The results show a stringent hypoxia-sensing CAR T-cell approach which achieves selective expression of a panErbB-targeted CAR within a solid tumour, a microenvironment characterized by an inadequate oxygen supply. Despite widespread expression of ErbB receptors in healthy organs, the approach provides anti-tumour efficacy without off-tumour toxicity in murine xenograft models. This dynamic oxygen-sensing safety switch potentially facilitates unlimited expansion of the CAR T-cell target repertoire for treating solid malignancies.

    [0358] Identifying approaches to circumvent off-tumour toxicity has the potential to unlock an entirely new repertoire of CAR antigen targets for carcinomas, which are currently limited.

    [0359] To investigate this issue, we utilized a 2.sup.nd generation pan-anti-ErbB CAR T1E28z which has specificity towards 8/10 of the possible ErbB receptor homo- and hetero-dimers and crosses the species barrier binding both mice and human receptors equivalently. This CAR is currently undergoing Phase I evaluation by intra-tumoural (i.t.) delivery in patients with SCCHN. The CAR is co-expressed with a chimeric cytokine receptor (4αβ) which delivers an intracellular IL-2/IL-15 signal upon binding of IL-4 to the extracellular domain (FIGS. 24 A and B), providing a means to selectively enrich CAR T-cells during ex vivo expansion, but however does not affect the CAR-dependent killing capacity of the T-cells. This combination is referred to as T4-immunotherapy. Although i.t. delivery of T4-CAR T-cells has proven safe in man, i.v. infusion is desirable as this permits these cells to home to both the primary tumour and metastasis. I.v. infusion of human T4-CAR T-cells into immunocompromised NSG mice bearing HN3 tumours (FIG. 24C) which express ErbB1-4 resulted in lethal toxicity, evident by a rapid loss of weight in these animals (FIG. 24D). As observed clinically, analysis of the blood of these mice revealed evidence of an increase in pro-inflammatory cytokines (FIG. 24E). In an attempt to resolve the biodistribution of CAR T-cells, we modified the CAR construct to concurrently express a luciferase (Luc) reporter to permit in vivo tracking of transduced T-cells (FIG. 24F). Imaging analysis four days post i.v. infusion of a sub-lethal dose of reporter human CAR T-cells revealed that the majority had accumulated in the lungs and liver, while only a minority were present in the tumour despite the expression of ErbB1-4 (CAR targets) on these cells (FIG. 24G). The accumulation in the liver and lung was not an artefact of the xenograft system as, when murine T-cells were transduced to express the same reporter CAR and infused i.v. into Rag2.sup.−/− mice (FIG. 24C), they accumulated in the same tissues and in the spleen (Fig. S3). Notably, murine T-cell accumulation in the liver, but not the lung, was CAR-dependent as T-cells expressing the Luc reporter alone were significantly less prevalent at this location. The CAR-independent T-cell accumulation in the lung was likely due to an integrin-dependent interaction. Profiling of ErbB1-4 mRNA expression confirmed that all four receptors were expressed in all vital organs, including the lungs and liver. To investigate for direct evidence of T4-CAR T-cell-mediated tissue damage, a sub-lethal dose of human T4-CAR T-cells was infused i.v. into NSG mice and a pathohistological examination using haematoxylin and eosin (H&E) stained tissue sections of the liver and lung was conducted after 5 days. This analysis revealed the presence of myeloid cell infiltrates in the lungs and liver (FIGS. 24H and I), representing a surrogate marker of CAR-mediated inflammation. The infiltrate was observed both in a perivascular distribution and scattered throughout the parenchyma, consisting of both neutrophils (polymorphonuclear cells) and large mononuclear cells with abundant cytoplasm, likely to be macrophages. Hepatocyte necrosis/apoptosis was also seen in some animals. T4-CAR T-cells accumulated in the kidney at a lower level (FIG. 24G) with no significant evidence of inflammation in this tissue. These data indicate that the liver and lung represent the two key organs for off-tumour CAR T-cell activation.

    [0360] Hypoxia is a characteristic of most solid tumours. The proliferative and high metabolic demands of the tumour cells, alongside inefficient tumour vasculature, result in a state of inadequate oxygen supply (<2% O.sub.2) compared to that of healthy organs/tissues (5-10% O.sub.2) (FIGS. 24 J and K). As hypoxia differentiates the tumour microenvironment from that of healthy, normoxic tissue, it represents a desirable marker for the induction of CAR T-cell expression (FIGS. 24 J and K). To create a stringent hypoxia-regulated CAR expression system, we developed a dual-oxygen sensing approach for the T4-CAR (FIG. 25A). This was achieved by appending a C-terminal 203 amino acid ODD onto the anti-ErbB CAR while concurrently modifying the CAR promoter in the long terminal repeat (LTR) enhancer region to contain a series of 9 consecutive HREs, rendering CAR expression selectively responsive to hypoxia. In vitro, this CAR, named ‘HypoxiCAR’, demonstrated stringent hypoxia-specific presentation of the CAR molecules on the cell surface of human T-cells (FIG. 25B). We demonstrated that the dual-oxygen sensing system proved superior to variants in which either the 9 HRE cassette or ODD were used alone. In both cases, these alternative approaches displayed leakiness of CAR expression under conditions of normoxia, permitting tumour cell killing under normoxic conditions. HypoxiCAR's expression of the CAR was also highly dynamic and represented a switch that could be turned both ‘on’ and ‘off’ in an O.sub.2-dependent manner (FIG. 25C). In further in vitro characterization, exquisite O.sub.2 sensitivity of HypoxiCAR was confirmed as CAR expression was absent under O.sub.2 concentrations consistent with healthy organs (5%) but became detectable on the cell surface at O.sub.2 concentrations equivalent to those found in the tumour microenvironment (1%) (FIG. 25D). Tumour-infiltrated T-cells have been demonstrated to egress from the tumour microenvironment, highlighting a potential safety concern if hypoxia-experienced HypoxiCAR T-cells expressing CAR were to re-enter healthy normoxic tissue. However, as cytolytic T-cell mediated killing of a target cell may take up to 6 hours, within which time in normoxia it might be expected that approximately 62±8% of HypoxiCAR's surface CAR may have already degraded (FIG. 2C), any off-tumour killing by egressed HypoxiCAR T-cells would be expected to be limited. Moreover, once HypoxiCAR has expressed sufficient CAR to kill a target, cell egress would be limited as has been demonstrated that CD8.sup.+ T-cell migration ceases in regions where it encounters a tumour cell expressing its cognate antigen.

    [0361] Having validated HypoxiCAR's ability to sense hypoxia, we sought to investigate its ability to elicit hypoxia-dependent killing of tumour target cells. SKOV3 ovarian cancer cells were seeded onto culture plates and co-incubated with T4-CAR or HypoxiCAR under normoxic and hypoxic (0.1% O.sub.2) conditions. Despite equivalent transduction efficiencies and CD4.sup.+:CD8.sup.+ T-cells ratios, HypoxiCAR T-cells displayed efficient hypoxia-dependent killing of the SKOV3 cells, almost equivalent to T4-CAR T-cells, with no significant killing observed under normoxic conditions (FIG. 25E). Target-cell destruction was strictly CAR-dependent as when the intracellular tail of HypoxiCAR was truncated to prevent CD3□ signalling, killing was abrogated (FIG. 25E). In addition, HypoxiCAR-provided stringent hypoxia-restricted T-cell secretion of both IL-2 (FIG. 25F) and IFNγ (FIG. 25G), two cytokines which play an important role in the T-cell response.

    [0362] To evaluate whether HypoxiCAR could provide tumour-restricted CAR expression in vivo, human HypoxiCAR T-cells were injected concurrently i.v. and i.t. in NSG mice bearing HN3 tumours. These tumours had an approximate volume of 500 mm.sup.3 (FIG. 26A), in which the presence of hypoxia was confirmed (FIG. 24J,K). Four days after HypoxiCAR T-cell infusion, tissues were harvested, enzyme-digested and T-cells were assessed for CAR expression using flow cytometry. As predicted by the in vitro analyses (FIG. 25), HypoxiCAR T-cells did not express detectable cell surface CAR molecules when recovered from the blood, lungs, or liver of the mice post infusion, but they did express CAR molecules on the cell surface within the hypoxic tumour microenvironment (FIG. 26B,C). This finding was not model specific as similar observations were made in both NSG mice bearing SKOV3 tumours and in Rag2.sup.−/− mice bearing murine Lewis lung carcinoma (LL2) tumours. To establish if the ‘Hypoxi’ construct elements would be active across different stages of tumour growth, a Hypoxi-luciferase reporter was developed in which the HRE promoter was used to drive expression of a luciferase-ODD. This reporter was stably transduced directly into the SKOV3 and HN3 cell lines. Luciferase-ODD, despite not being detectable in tumour cells under normoxic conditions, was detected in vivo at all stages of tumour growth, even prior to the tumour becoming palpable, in both SKOV3 and HN3 tumours. This suggested that HypoxiCAR T-cells might be active even against early stage tumours. To test this, HypoxiCAR T-cells were infused into mice at day 16 post injection of HN3 tumour cells, just prior to the tumours becoming palpable. In keeping with the absence of CAR expression on the T-cells in normoxic tissues, HypoxiCAR also circumvented the treatment-limiting toxicity seen using following i.v. infusion of high-dose T4-CAR T-cells. Indeed, mice infused i.v. with human HypoxiCAR T-cells displayed no acute drop in weight post infusion (FIG. 26D,E), no evidence of pro-inflammatory cytokines in the systemic circulation (FIG. 26F), nor were there any signs of tissue damage in the lung, liver or kidney (FIG. 26G,H). Importantly, while mice infused i.v. with human T4-CAR T-cells all reached their humane endpoints at 28 h (FIG. 26E), the HypoxiCAR T-cell infused mice showed no signs of off-tumour toxicity and prevented tumour growth (FIG. 26I). As such, HypoxiCAR overcomes a major hurdle that currently precludes the systemic administration of CAR T-cells targeting antigens that are expressed in normal tissues throughout the body.

    [0363] Hypoxia has been extensively studied in SCCHN. To assess which patients might be most appropriate for HypoxiCAR T-cell immunotherapy, we firstly generated an HRE-regulated gene signature using patient tumour transcriptomic data. Known HRE-regulated genes were analyzed for co-expression, and a refined signature utilizing the genes PGK1, SLC2A1, CA9, ALDOA and VEGFA was chosen as we observed a significant positive correlation between these genes (FIG. 27A). There was no difference in expression of this signature across the different SCCHN subtypes (hypopharynx, larynx, oral cavity, and oropharynx). However, expression of this 5-gene signature, significantly increased with tumour size (T-score; FIG. 27B) and was also prognostic of poorer survival in stage 3 and 4 HNSCC patients (FIG. 27C). Utilizing an HRE-regulated gene signatures to predict hypoxia from biopsy material could provide a simple means to assess those patients which might respond best to HypoxiCAR therapy.

    [0364] Immunohistochemistry staining of SCCHN tumour sections for stabilized HIF1α, the master transcription factor for HypoxiCAR's CAR expression, revealed large regions of the tumours where HIF1α had become stabilized (FIG. 27D). Although several factors can stabilize HIF1α, hypoxia represents the most probable explanation for this observation. Heterogeneity in both HIF1α stabilization and intra-tumoural T-cell infiltration was seen between patients. Encouragingly however, those tumours with the highest prevalence and/or intensity of HIF1α stabilization did not exclude T-cells from entering the intra-epithelial space nor from entering HIF1α stabilized regions of the tumour (FIG. 27E,F). Using immunofluorescence, we also demonstrated that CD3.sup.+ T-cells infiltrating HIF1α stabilized tumour regions also stabilized HIF1α themselves, suggesting that in these environments HypoxiCAR T-cells would become activated (FIG. 27G). These observations suggest that HypoxiCAR could find clinical application in hypoxic tumour types such as SCCHN, where gene expression (FIG. 27A-C), staining of biopsy samples for HIF1α/CD3 (FIG. 27D-G) and imaging techniques such as PET/CT using a hypoxia-radiotracer such as .sup.64Cu-ATSM might provide biomarkers to confirm the presence of a hypoxic tumour microenvironment and guide patient selection.

    [0365] Approaches to improve tumour-specificity of CAR T-cells have been developed, such as T-cell receptor-mimetic CARs with specificity for HLA-presented antigens, combined targeting of tumour antigens, or tuning of CAR affinity to preferentially target high density antigens. This study demonstrates an alternative approach to achieve cancer-selective immunotherapy, exploiting one of the most innate characteristics of the tumour microenvironment. The ‘dual hypoxia-sensing’ system described here achieves compelling anti-tumour efficacy while abrogating off-tumour toxicity of a CAR that recognizes multiple targets in normal tissues. The hypoxia-sensing HRE module and the ODD appended onto the CAR act synergistically to provide stringent hypoxia-specific target killing (FIG. 25E). This approach restricts both transcription (HRE) and stability (ODD) of the CAR under conditions of normoxia and, when these two systems are utilized concurrently, they overcame the leakiness observed when either system was used alone.

    [0366] The hypoxic tumour microenvironment is not conducive to efficient immune reactions. Hypoxia can activate immune-suppressive programmes in stromal cells such as macrophages, regulate the expression of immune checkpoint molecules and promote a more aggressive tumour cell phenotype. However, encouragingly we found that hypoxia did not negatively affect T-cell effector function directly in vitro (FIG. 25E-G), which is in agreement with that observed by others. HypoxiCAR T-cells also were able to prevent the growth of hypoxic tumours (FIG. 26I) suggesting that, in the in vivo models tested, the tumour microenvironment was not a complete barrier to HypoxiCAR's ability to deliver in vivo anti-tumour therapeutic efficacy. There is also the potential in the future to combine HypoxiCAR T-cell therapy with microenvironment modifying agents, such as immune checkpoint inhibitors, which may further improve the ability of these cells to target the tumour. Furthermore, as T-cells are not excluded from HIF1α stabilized regions of human tumours (FIG. 27D-F) it is likely that HypoxiCAR T-cells should be able to access the appropriate microenvironments to activate CAR expression. Although we did not observe evidence of treatment-limiting toxicity in mice infused with high dose HypoxiCAR T-cells (FIGS. 26E and I), there are microenvironments in healthy tissues such as the intestinal mucosa where ‘physiologic hypoxia’ has been observed. Such tissues might represent sites where off-tumour activation of HypoxiCAR

    [0367] T-cells could take place. As such, a suicide switch could be incorporated into HypoxiCAR to provide an additional level of safety for the most pervasive CARs. Although the ‘HypoxiCAR’ dual oxygen sensing system was exemplified using a pan-ErbB-targeted CAR, the broadly applicable strategy may be used to overcome the paucity of safe targets available for the treatment of solid malignancies.