FIELD OF THE INVENTION

The invention relates to the use of genetically modified CD 14-expressing blood cells (CDl4 ⁺ ), such as monocytes, as agents for engraftment into tumours stimulated by tumour-associated signals to produce and secrete anticancer therapeutics.

BACKGROUND TO THE INVENTION

In spite of recent medical advances, cancer continues to represent a significant health problem worldwide. Cancer immunotherapy has demonstrated exciting clinical results in the setting of numerous solid tumours and hematologic malignancies. The endogenous immune system is typically non-reactive to malignant cells, or can be actively immunosuppressed with respect to the body's reaction to the presence of malignant cells. One way to enhance treatment of tumours is to force tumour recognition by the immune system through genetic engineering of immune cells.

Macrophages and other cells of the myeloid lineage frequently constitute a major component of solid tumours, sometimes representing as much 80% of the total tumour mass. It is therefore unsurprising that a great deal of research has focused on uncovering the role of such tumour-associated myeloid-derived cells (TAMDCs) and whether they support or hinder tumour growth. As a pleiotropic cell type, TAMDCs include a variety of cells such as infiltrating monocytes (including Tie-2-expressing monocytes), tumour- associated macrophages, myeloid-derived suppressor cells (MDSC) and polymorphonuclear leukocytes (neutrophils, eosinophils and basophils). The variety and function of tumour-associated myeloid cells is widely characterised and reviewed in the literature, for example by Elliot et al, Frontiers in Immunology, volume 8, article 86 (2017). TAMDC display diverse features and properties and play important roles in regulating tumour growth. For example pro inflammatory macrophages (sometimes referred to as‘Ml’ macrophages) and dendritic cells may play a role in stimulating immune recognition of tumours, while anti-inflammatory macrophages (sometimes referred to as‘M2’ macrophages) and MDSC can play roles in suppressing immune recognition and supporting malignancy. Indeed, many of the“hallmarks of cancer” have been shown to be supported by TAMDCs. Most notably, macrophages are central to the normal wound-healing process and they may play important roles in supporting tumours, described as‘wounds that never heal’ (Dvorak, New England J Medicine, 315(26): 1650- 9, 1986). Clinical data has shown that high macrophage infiltration and density within solid tumours frequently correlates with a poor prognosis. The combination of increased macrophage infiltration and poor prognosis is suggestive that in many tumour types (such as breast, bladder and ovarian carcinoma) macrophages can be actively recruited by tumours to facilitate their growth and indeed, this has been shown to be the case. Similarly MDSC often play important immunosuppressive roles in tumour growth and development. They can be roughly divided into monocytic (CDl4 ⁺ ) MDSC and polymorphonuclear (CDl5 ⁺ ) MDSC, and both types are capable of suppressing T cell function within tumours.

Once established within tumours, TAMDCs actively subvert and suppress potential local anti-tumour immune responses through a range of secreted cytokines and chemokines. Secretion of broad acting immunosuppressive cytokines such IL-10 and TGF-b allows TAMDCs to dampen inflammatory immune responses and encourage the tolerisation of other immune cells. TAMDCs secrete a range of chemokines that aid recruitment of many suppressive cell types, most notably regulatory T cells (Tregs) for example through CCL22 production. Through physical interaction with T cells, TAMDCs are also able to inhibit local T cell proliferation and activation through the up -regulation of a number of inhibitory receptors such as B7-H4.

TAMDCs are thought to originate partly from bloodbome CDl4 ⁺ monocytes and partly from tissue resident myeloid cells derived from yolk sac-derived‘early’ and Tate’ erythro-myeloid progenitors, as reviewed by Franklin and Li, Trends Cancer 2(1): 20-34, 2016. The yolk sac route is considered particularly important for the production of microglia and many myeloid cells within glioma tumours, while the monocyte route reflects more of the wound healing functionalities found in many carcinoma tumours. We define tumour-associated cells derived from bloodborne monocytes as myeloid-derived tumour infiltrating cells (MDTIC) and they are thought to give rise to many types of TAMDC, including wound healing immunosuppressive‘M2’ macrophages, more pro- inflammatory ‘Ml’ macrophages, dendritic cells, and monocytic myeloid-derived suppressor cells. A need exists in the art for more effective compositions and methods that treat cancers by improving delivery of the therapeutic agents into tumour sites and also by increasing their anticancer selectivity to improve the therapeutic index (defined as the anticancer activity divided by the unwanted toxicity against other tissues). The present invention relating to the use of modified monocytes as agents for active engraftment into tumours where they will be stimulated by tumour-associated signals to produce and secrete anticancer protein therapeutics which fulfils this need. A range of mechanistic approaches is described that all achieve this effect, ultimately leading to tumour-induced production of anticancer agents deep within the tumour itself.

SUMMARY

The present invention relates to treatments for cancer using CDl4 ⁺ blood cells, engineered with a nucleic acid encoding one or more therapeutic agent(s), the expression of which is activated (or if already expressed increased) in response to tumour-associated micro-environmental signals when the CDl4 ⁺ cells infiltrate into tumours and become MDTICs. The expression of the therapeutic agent(s) is preferably controlled, directly or indirectly, by a promoter. The promoter must be differentially active within MDTICs compared to bloodbome CDl4 ⁺ cells or cells derived from them within normal tissues such as normal breast tissue.

The present inventors have surprisingly shown that it is possible to express therapeutic agents in a tumour-selective manner based on differentiation and/or polarisation of a CD 14+ cell in response to tumour associated-microenvironmental signals, and have identified a variety of promoters and microRNA binding sites that may be used alone or in combination to sensitively control therapeutic agent expression in a tumour, in tumour-associated/infiltrating CD14+ cells. The inventors’ therapeutic approach is based on administration of modified CD 14+ cells and advantageously uses tumour-associated differentiation and polarisation within this class of cells, avoiding undesired activation of expression of a therapeutic agent in circulating cells outside the tumour context. The inventors approach is also broadly applicable to all CD 14+ cells infiltrating/associated with tumours, maximising the available cell populations for use in therapy.

Accordingly, the invention provides a modified CD 14+ cell comprising a nucleic acid encoding at least one therapeutic agent whose expression is activated within a tumour in response to tumour-associated micro-environmental signals. The invention also provides:

- a population of modified CD 14+ cells of the invention, typically a cell population comprising a plurality of modified CD 14+ cells of the invention, wherein said population comprises at least two different subclasses of modified CD14+ cells;

- a method for making a modified CD 14+ cell comprising a nucleic acid encoding therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated within a tumour in response to tumour-associated micro-environmental signals which consists of: (a) isolating bloodbome cells from an individual, (b) culturing the isolated bloodborne cells ex vivo, and (c) transfecting or transducing cultured CD 14+ cells from the bloodbome cells to introduce a vector encoding a therapeutic agent;

- a method of treating a disease in a human by administering the modified CD 14+ cell of the invention or the cell population of the invention;

- the modified CD 14+ cell or cell population of the invention for use in a method of treating a disease in a human, the method comprising administering to the human the modified CD 14+ cell;

- a method for treating cancer, wherein the method produces a modified CD 14+ cell comprising a nucleic acid encoding anticancer therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated in MDTICs or TAMDC and consists of: (a) isolating peripheral blood mononuclear cells (PBMC) from a patient, or a suitable donor, using leukapheresis, (b) culturing the isolated blood mononuclear cells ex vivo, (c) transducing cultured CD 14+ cells from the PBMCs to introduce a vector encoding a bispecific T-cell activating antibody, optionally binding EpCAM, and further encoding one or more chemokines selected from CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20 , CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12 or CXCL16, (d) introducing the modified CD14+ cells into a patient, and (e) allowing the engraftment or localisation of the CD 14+ cells within a tumour(s) optionally by the application of radiotherapy or chemotherapy; and

- a vector for use in a method for treating cancer, wherein the vector encodes a bispecific T-cell activating antibody, optionally binding EpCAM and further encodes one or more chemokines selected from CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20 , CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12 or CXCL16, and wherein the method produces a modified CD14+ cell comprising a nucleic acid encoding an anticancer therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated in MDTICs or TAMDC and consists of: (a) isolating peripheral blood mononuclear cells (PBMC) from a patient, or a suitable donor, using leukapheresis, (b) culturing the isolated blood mononuclear cells ex vivo, (c) transducing cultured CD 14+ cells from the PBMCsto introduce the vector, (d) introducing the modified CD 14+ cells into a patient, and (e) allowing the engraftment or localisation of the CD 14+ cells within a tumour(s) optionally by the application of radiotherapy or chemotherapy.

DESCRIPTION OF THE FIGURES

Figures 1-5 illustrate some of the molecular strategies that maybe used to activate expression of the nucleic acid encoding the therapeutic agent(s) following differentiation of CDl4 ⁺ cells in the tumour microenvironment. In some instances more than one therapeutic agent can be encoded within the same mRNA using a 2A site to allow translation of two or more separate proteins. Enzyme cleavage sites (such as furin cleavage sites) may be included to allow spontaneous removal of unwanted amino acid stubs.

Figure 1 shows simple direct regulation by an‘Upregulated Target Promoter’ driving expression of the therapeutic construct only within CDl4 ⁺ cells exposed to the tumour microenvironment (A) and not in CDl4 ⁺ cells within normal tissues (B).

Figure 2 shows an indirect system, where an Upregulated Target Promoter within CDl4 ⁺ cells, in response to tumour microenvironment (A) but not within CDl4 ⁺ cells in non-cancer tissues (B), causes transcription of the VPl6-TetRepressor transactivator, which then binds TetO sites upstream of a minimum CMV promoter and activates it to give high level expression of the nucleic acid encoding the therapeutic agent(s). This system can be used to amplify the signal generated by the Upregulated Target Promoter without compromising selectivity. In addition, this system provides a safety switch for use in the event of gene product-related toxicity (C) because addition of doxycycline will bind the VPl6-TetRepressor and prevent transactivation of the Tet-Operator sequences, inhibiting expression of the nucleic acid encoding the therapeutic agent(s).

Figure 3 shows a down-regulated Target Promoter being used to drive expression of a short hairpin RNA (shRNA) capable of binding and degrading the nucleic acid encoding the therapeutic agent(s). In the presence of the tumour microenvironment (A) the Target Promoter will be inhibited and no shRNA produced, leading to the expression of the nucleic acid encoding the therapeutic agent(s), whereas in non-cancer tissues (B) the Target Promoter will be active, shRNA will be expressed and the nucleic encoding the therapeutic agent(s) will not be translated.

Figure 4 shows simple exploitation of an endogenous microRNA that is present in CD 14+ cells within non-cancer settings (B) but lost following exposure of the CDl4 ⁺ cells to the tumour microenvironment(A). By including cognate microRNA target sites into the 3’UTR of the nucleic acid encoding the therapeutic agent(s), their mRNA can be selectively destabilised within non-cancer settings but stabilised (and translation activated) following exposure to the tumour microenvironment.

Figure 5 shows an example of how these technologies can be combined. In this example an Upregulated Target Promoter (which will be active within CDl4 ⁺ cells within the tumour microenvironment) drives transcription of the VPl6/TetR transactivator leading to enhanced transcription of therapeutic agents (as described for Figure 2). By introducing microRNA sites into the 3’UTR of the nucleic acid encoding the therapeutic agent(s) and placing the corresponding microRNA under control of a Downregulated Target Promoter, the nucleic acid encoding the therapeutic agent (s) can be selectively stabilised within tumour cells and translation will proceed (as in Figure 3). This combination of two Target Promoters (one upregulated and one downregulated) can combine to give potent and highly selective transgene expression only within CD 14 ⁺ cells exposed to the tumour microenvironment.

Figure 6 A - Illustrates the primary monocyte sub-populations within human blood, as defined by CD14 and CD16 expression. Briefly, peripheral blood monocytes were stained for CD14, CD16, and TIE-2, expression was determined by flow cytometry. Figure 6B - Illustrates the expression of TIE -2 by each primary monocyte sub-population. Figure 6C - Illustrates the relative levels of lentiviral infection, as defined by GFP expression, in the CD14+CD16- sub-population compared to the total primary monocyte population (CD14+CD16+). Peripheral CD14+CD16- monocytes were isolated from peripheral blood mononuclear cells (PBMC) first by CD 16- depletion, followed by CD 14+ positive selection. The total monocyte population was isolated by CD 14+ positive selection only. Both populations were rested overnight and infected with lentiviruses by spinoculation. GFP expression was driven by either the MCR1 promoter region, or spleen focus forming virus (SFFV) promoter. Both the percentage of expression and the geometric mean fluorescent intensity (Geo.MFI) of the GFP+ cells are displayed.

Figure 7 shows the variety of vectors that can be used to transfect and transduce primary human CD 14 cells, and the large number of cells that can be genetically modified in this way, hot restricted to certain subsets or lineages. Figure 7A - Illustrates the efficiency of transduction methods in monocytes or THP-l cells. le5 or 2e5 cells were electroporated with 0.4, 0.6 or 0.8 pg of a GFP plasmid. 24 hours later the level of GFP expression was read by flow cytometry (first graph) and the number of viable cells were counted (second graph). Figure 7B - Illustrates that primary monocyte can be efficiently infected and transduced by HIV 1 -based lentiviral vectors, but that VPX protein is required. Primary monocytes were isolated and infected with a GFP lentivirus vector at the given multiplicity of infection (MOI), ± VPX accessory protein-containing viral-like particle. After 96 hours GFP expression was measured by flow cytometry. Figure 7C - illustrates that primary monocytes can be efficiently infected by adenoviral vectors. 2e5 isolated CD 14+ cells were infected with GFP adenoviral vector at 461 MOI. GFP expression was measured by flow cytometry over a time course of 72 hours.

Figure 8 A - Illustrates monocytes can be differentiated and polarized in vitro into M0, Ml and M2 macrophages. Isolated CD 14+ primary monocytes were incubated in media containing 1% human serum. On day 4, the media were exchanged for polarizing media containing: M0: media+l% Human serum, Ml : IFNy (25ng/ml) + LPS

(lOng/ml), M2: IL4 (50ng/ml) + IL10 (50ng/ml). Archetypal markers of these populations were then measured by flow cytometry at 24, 72 hours and at day 7. Figure 8B - Illustrates that malignant ascites fluid and cancer cell line conditioned media (CCM, produced by A549 human lung carcinoma, MDA468 human breast carcinoma and SKBR3 human breast carcinoma)) can polarize primary monocytes towards a TAM/M2 phenotype and not Ml or M0. Primary CD 14+ monocytes were isolated from PBMC and differentiated to M0, prior to 3 days polarization with CCM or ascites. The phenotype was determined by measuring expression of archetypal markers, described above, by flow cytometry.

Figure 9A - Schematic illustrating the protocol for measuring candidate promoter activity in M0, Ml and M2 macrophages. Figure 9B - Heatmap illustrating the selective activity of candidate promoters in M2 macrophages compared to M0 and Ml macrophages. Isolated primary CD 14+ monocytes were infected with lentiviral vectors, containing molecular constructs were GFP was driven by candidate promoters, then differentiated and polarized into M0, Ml or M2 macrophages. Promoter activity was then determined by measuring GFP expression in MO, Ml or M2 macrophages by flow cytometry. Data shown is the fold change in iMFI (Geo.MFI multiplied by the percentage of GFP+ cells). Figure 9Ci - Schematic illustrating the protocol for measuring candidate promoter activity in TAM/M2 macrophages generated by culturing in ascites or CCM. Figure 9Cii - Heatmap illustrating the selective activity of candidate promoters in TAM/M2 macrophages, generated by culturing cells in ascites at 80% (A80) or 50% (A50) or with different cancer cell line conditioned media (CCM), compared to Ml macrophages, generated by polarizing with IFNy(25ng/ml)+ LPS(l0ng/ml). Isolated primary CD 14+ monocytes were infected with lentiviral vectors containing GFP driven by candidate promoters, then differentiated and polarized into TAM/M2 or Ml macrophages. Promoter activity was then determined by measuring GFP expression in the different macrophage populations generated. Data shown is the fold change in iMFI (Geo.MFI multiplied by the percentage of GFP+ cells). Figure 9Di - illustrates candidate promoter activity in macrophages generated by polarizing cells in different breast cancer cell line conditioned media. Isolated primary CD 14+ monocytes were infected with GFP lentiviral vectors, where GFP was driven by candidate promoters, then differentiated and polarized into TAM/M2 by culturing in CCM, or M2 macrophages by stimulation with IL4 (50ng/ml)+ IL10 (50ng/ml). Promoter activity was then determined by measuring GFP expression in the different macrophage populations generated. Figure 9Dii shows a similar analysis performed in media conditioned by human cells of different, non breast cancer, types. Figure 9E - illustrates the durability of promoter activity. Isolated primary CD 14+ monocytes were infected with lentiviral vectors containing GFP driven by candidate promoters, then differentiated and polarized into M0, Ml or M2 macrophages. Promoter activity was then determined by measuring GFP expression in M0, Ml or M2 macrophages by flow cytometry after 10 days or 17 days of culture (the left and right bar of each pair, respectively). Figure 9F - Illustrates the ability of ex vivo tumour sections to induce promoter activity and drive GFP expression. Isolated primary CD 14+ monocytes were infected with lentiviral vectors containing GFP driven by the candidate promoters, then differentiated and polarized into TAM/M2 macrophages by co-culture with ex vivo tumour sections. After 3 days of co-culturing, promoter activity was determined by measuring GFP expression by flow cytometry. They are compared with the same batch of monocytes that underwent standard M0 or M2 polarising treatments for comparison. Figure 10 - Volcano plot showing microRNA molecules that are differentially expressed between primary human macrophages (derived from CD 14 monocytes) that are polarised either with Ml -inducing cytokines or supernatants from human breast cancer cells, determined by microRNAseq. These data show both the fold change (along the x axis) and the significance of the difference (y value). The microRNAs that are downregulated in cells treated with CCM compared to Ml (i.e. those of value in the practice of this invention) are towards the left.

Figure 11A - Schematic illustrating the protocol for measuring the activity of candidate miR target sites in M0, Ml and M2 macrophages. Here infection occurs post- differentiation but prior to polarization into M0, Ml and M2 macrophages. Figure 11B - Heatmap illustrating the selective activity of candidate miR target sites in M2 macrophages compared to M0 and Ml macrophages. Isolated primary CD 14+ monocytes were differentiated, infected with GFP lentiviruses, where the constructs contain individual candidate miR target sites, then polarized into M0, Ml or M2 macrophages. Loss of specific miRs was determined by the gain of GFP expression in M0, Ml or M2 macrophages, as measured by flow cytometry. Data shown is the fold change in iMFI (Geo.MFI multiplied by the percentage of GFP+ cells). Figure 11 C - Schematic illustrating the protocol for measuring the activity of candidate miR target sites in TAM/M2 macrophages. Here infection occurs post-differentiation but prior to polarization into TAM/M2 macrophages. Figure 11D - Illustrates the activity of miR targets in TAM/M2 macrophages generated in breast cancer cell line conditioned media (CCM). Levels of GFP expression are shown next to levels seen in M2 macrophages (red line). Isolated primary CD 14+ monocytes were differentiated, infected with GFP lentiviral vectors containing candidate miR target sites, then polarized into TAM/M2 or M2 macrophages. Loss of specific miRs was determined by the gain of GFP expression, as measured by flow cytometry. Figure 11E shows the same analysis for macrophages treated with CCM conditioned by non-breast human cancer cell lines, showing similar effects.

Figure 12A - Illustrates that adenovirus-infected primary CD 14+ monocytes actively migrate towards cancer cell line conditioned media (CCM). Freshly isolated primary CD 14+ monocytes were infected with an adenovirus-GFP vector and cultured overnight. The following day 2e5 adenovirus infected monocytes were placed in the top well of a transwell, and 20m1 of CCM was added to 180m1 of media in the lower well. After 90 minutes, the number of monocytes that had migrated into the lower well were counted. Figure 12B - Infiltration of ex vivo tumour sections by GFP lentivirus infected monocytes. Isolated primary CD 14+ monocytes were infected with GFP lentiviral vectors and then co-cultured for 9 days in the presence of a human tumour section ex vivo. Infiltration of GFP+ cells into the section was determined by microscopy, and the fluorescence image (lower graph) shows multiple bright green punctate signals, indicating that transduced monocytes have actively engrafted into the tumour section. Figure 12C shows two SCID mice with subcutaneous human MDA468 breast cancer xenografts, inoculated intravenously with primary human monocytes that had been infected with an adenovirus expressing luciferase under control of the CMV immediate early promoter. After 48h the mice were administered luciferin intraperitoneally and light emission was measured using an Ivis camera. Apart from the background luminescence in the snout and paws (thought to reflect the luminescence of the animal feed) and the injection site in the tail, the main signal is coming from the region of the tumour suggesting appreciable accumulation of monocytes at the tumour site. Figure 12D shows a bright field image of five MDA-468 xenograft tumour-bearing mice (tumours are encircled). Mice had received intravenous treatment with human primary CD14 cells infected with lentiviruses expression luciferase under the regulatory control of various promoters as shown. Figure 12E shows the luciferase expression from the mice shown in Figure 12D. Apart from luminescence associated with the paws, snout and injection site (thought to reflect the luminescence of the animal feed and injection technicalities respectively) the main signals are coming from the tumour site. The image on the left include a circle shown around the tumour, and the histogram illustrates the quantified light signal coming form the region of interest (the tumour region) compared to an equivalent region on the other flank.

Figure 13A - Illustrates the ability of resting and activated T cells to migrate towards chemokines. Primary CD3+ T cells were isolated fresh blood, 2e5 cells were added to the top transwell, and chemokines were added to the lower well at a final concentration of 50ng/ml. Cells were then incubated for 90 minutes and the number of cells migrating to the lower well were counted by flow cytometry. The T cells were either unstimulated (Figure 13AΪ) or pre-stimulated with antiCD3/antiCD28 beads (Figure l3Aii). Figure 13Bi - Schematic of a plasmid construct containing an EpCAM BiTE, CXCL10 and CCL3L1 chemokines, all with signal peptides, under regulatory control of a minimal CMV promoter with upstream TET-operator sites, together with CD19. Figure l3Bii - 293T cells were transfected with the plasmid containing the CXCL10-CCL3L1 molecular construct (shown in Figure 13BΪ). Supernatants were collected at 48 and 72 hours. 2e5 Peripheral blood mononuclear cells were placed in the top well of a transwell, and 20m1 of supernatants (S/N) from the transfected 293T cells were added to 180m1 of media in the lower well. Supernatants from non-infected 293T cells was used as a negative control, and 50ng/ml of CXCL10+CCL3L1 acted as a positive control. The number of cells migrating into the lower well was determined by FACS. Figure 13Biii - Illustrates the ability of 293T and primary CD 14+ monocytes to express chemokines derived from a lentiviral vector containing CXCL10 and CCL3L1 under a SFFV promoter. Concentration of CXCL10 and CCL3L1 in supernatants of 293T cells or primary CD14+ monocytes where determined by ELISA. Figure 13C - Illustrates the ability of primary CD 14+ monocytes to express large quantities of IFNy from a lentiviral construct under a SFFV promoter. Isolated primary CD 14+ monocytes were infected with lentiviral vectors containing the SFFV-IFNy construct, then differentiated and polarized into TAM/M2 by CCM or ascites (As), or M0, Ml, or M2 macrophages - as previously described. IFNy concentration was determined after 3 days of polarization by ELISA. Primary CD 14 cells from three individuals are shown as representative examples. Figure 13D - Illustrates the ability of primary CD 14+ monocytes to express IFNy from a lentiviral construct under regulatory control of candidate promoters. Isolated primary CD 14+ monocytes were infected with lentiviral vectors containing the promoter-IFNy constructs, then differentiated and polarized into TAM/M2 by CCM or ascites (As) stimulation or M0, or M2 macrophages - as previously described. IFNy concentration was determined after 3 days of polarization by ELISA. Figure 13E - Illustrates that genetically-expressed EpCAM bispecific T cell activator can activate T cells in the presence of EpCAM positive cancer cell lines. 0.5e6 Isolated primary CD3+ T cells were cultured in the presence of the A549, MDA468 and SKBR3 cancer cell lines (all known to be EpCAM -positive), at an E:T of 5: 1, with and without the EpCAM bispecific T cell activator. After 16 hours co-culture, T cell activation was determined by flow cytometry analysis of T cells expressing the markers CD25 and CD69. Analysis explored activation on the different T cell subsets. Figure 13F - Illustrates EpCAM bispecific T cell activator induces T cells degranulation and activation in the presence of EpCAM positive cancer cell lines. Isolated primary CD3+ T cells were cultured in the presence of the SKBR3 cancer cell lines (either 0.5e5 cells or le5 cells) ± the EpCAM bispecific T cell activator at two concentrations of the 293A supernatant (1/1000 and 1/500). After 16 hours co culture, T cell degranulation and activation was determined by flow cytometry analysis of T cells expressing the markers CD 107a and CD69. Analysis explored activation in the different T cell subsets, CD4+ and CD8+. Figure l3Gi - Illustrates EpCAM bispecific T cell activator expression, under SFFV and candidate promoter control and produced by 293 cells, can induce T cell activation and degranulation when co-cultured with the non breast cancer cell line SKOV3. le5 T cells were co-cultured with SKOV3 cells for 6 hours, at an E:T 5: 1. T cell expression of CD69 and CDl07a was determined by flow cytometry. Figure l3Gii - Illustrates EpCAM bispecific T cell activator expression, under SFFV and candidate promoter control and produced by M2 macrophages, can induce T cell activation and cytotoxicity when co-cultured with the breast cancer cell line MCF7. 5e4 T cells were co-cultured with MCF7 cells at an E:T of 5: 1 for 16 hours. T cell expression of CD69 was determined by flow cytometry. T cell cytotoxicity (target cell death) was determined by CytoTo-Glo™ (Promega Corporation) analysis of cell culture supernatants. Figure 13H shows combination of target promoter and polarisation-induced microRNA sites to give selective expression of gamma interferon in M2 -polarised macrophages. Expression of gamma interferon was determined in macrophages polarised to M0, Ml and M2 states. The cells were pre-infected with lentiviruses expressing gamma interferon driven by the target promoter CD200R and with three copies of binding sites for the polarisation-induced microRNA 155 (5P) contained in its 5’ UTR. As can be seen cells with interferon production regulated by the CD200R promoter alone showed greater expression in M2 cells than Ml or M0, as expected, but this differential became much greater when expression was additionally regulated by the presence of the l55p microRNA sites. Figure 131 shows combining target promoter and polarisation- dependent microRNA sites to regulate expression of a bispecific T cell activator that targets T cell cytotoxicity to EpC AM-positive cancer cells. Cytotoxicity was determined against EpCAM-positive human cancer cells, mediated by primary T cells that are activated by supernatants taken from macrophages infected with lentiviruses expressing bispecific T cell activators and polarised to Ml or M2. Expression of the bispecific T cell activators is regulated by the MRC1 promoter and three binding sites for microRNA 155 are included within the 3’ UTR of its mRNA, to promote degradation of the mRNA within Ml -polarised macrophages. T cell cytotoxicity is far greater when the macrophages are polarised to M2, indicating that substantially more bispecific T cell activator is produced from M2 polarised than Ml -polarised macrophages, as expected.

Figure 14 shows use of a TET repressor switch system to amplify transgene expression using weak promoters, and also to allow silencing of expression following administration of doxycycline. Figure 14A1 shows a schematic outline of the concept, where expression of a TETR-VP16 fusion protein from a relatively weak (tumour- responsive) promoter can bind TETO sites upstream of a minimal CMV promoter and allow the VP 16 component to activate strong transgene expression. Addition of doxycycline inactivates the fusion protein and prevent binding to TETO sites. Figure l4Aii shows that weak promoters (in this case IL1B and VEGF promoters in 293 cells) can be amplified using this switch system, while strong promoters are not enhanced. Figure 14BΪ shows a schematic structure of an adenovirus genome including the components of a TETR-VP16 transactivator switch and three therapeutic genes as payload (EpCAM bispecific T cell activator, CCL3L1 and CXCL10). Figure l4Bii shows that the virus produces significant quantities of both the EpCAM bispecific T cell activator and CXCL10, measured by western blotting and ELISA, respectively, and expression of both can be silenced by the addition of doxycycline as expected.

DETAILED DESCRIPTION

Modified CD14+ cells

The present invention provides a modified CD 14+ cell comprising a nucleic acid encoding at least one therapeutic agent whose expression is activated within a tumour in response to tumour-associated micro-environmental signals. In other words, the CD 14+ cell of the invention comprises a nucleic acid that encodes at least one therapeutic agent whose expression is selectively activated when the CD 14+ cell locates within a tumour. Location of the CD 14+ cell within the tumour exposes the CD 14+ cell to tumour- associated microenvironment signals. The tumour-associated microenvironment signals activate expression of the therapeutic agent.

Preferably, expression of the therapeutic agent in the CD 14+ cell is essentially or entirely inactive before the cell locates within a tumour and is exposed to tumour-related microenvironment signals. That is, the CD 14+ cell may not express the therapeutic agent, or may express a biologically insignificant amount of the therapeutic agent, before the cell locates within a tumour and is exposed to tumour-related microenvironment signals. Expression of the therapeutic agent may therefore be undetectable or minimally detectable in the CD 14+ cell prior to its location in the tumour. Methods for determining expression of a therapeutic agent by a cell are well-known in the art. CD 14+ cells

The modified CD 14+ cell may be any CD 14+ cell present in peripheral blood, preferably a monocyte. Further surface markers for identifying monocytes are known in the art. The monocyte may, for example, express CD 16. The monocyte may have high expression of CD 14 and low expression of CD 16 (CDl4 ^hl CDl6 ^l0 ). The monocyte may have high expression of CD14 and mid expression of CD16 (i.e. is a CDl4 ^hl CDl6 ^imd ). The monocyte may have low expression of CD 14 and high expression of CD 16 (CDl4 ^lo CDl6 ^hl cell). The monocyte may express detectable levels of Tie2. The monocyte may not express detectable levels of Tie2. The monocyte may be a classical, intermediate or non-classical monocyte. The CD 14+ cell may be a monocyte-derived suppressor cell. The CD 14+ cell may be a circulating CD 14+ monocyte. Typically the CD 14+ cell will be present in the context of a population of a plurality of different subclasses of CD 14+ cells, as discussed further below.

The CDl4 ⁺ cells may be isolated by purifying peripheral blood mononuclear cells (PBMC) from whole blood by leukapheresis followed by positive-selection for CD 14 expression using affinity beads or column or negative selection to remove other cell types and leave purified CD 14 cells behind

The CDl4 ⁺ cells of the invention may be derived from a patient or from a donor - the invention therefore envisages both autologous and allogeneic treatments.

Tumour-associated micro-environmental signals

Tumour-associated micro-environmental signals that may activate expression of the therapeutic agent may, for example, include cytokines. For instance, the tumour- associated micro-environmental signals may comprise IL-l, IL-2, IL-3, IL-4, IL-6, IL-7, IL-8, IL-10, IL-12, IL-13, IL-15, IL-17, IL-18, IL-21, IL-23, VEGF, FGF2, TGFbeta, CSF, MCSF, GCSF, GMCSF, TNFalpha, lymphotoxin, MIF, Fas, Fas ligand, type 1 interferons (particularly alpha-interferon and beta-interferon), type 2 interferons (gamma- interferon), type 3 interferons (gamma 1 -interferon, gamma2-interferon, gamma3- interferon), TRAIL, FLT3 ligand, Lymphotactin and chemokines, alone or in any combination.

Promoter

Expression of the at least one therapeutic agent may be controlled by a promoter. Expression may be directly controlled by the promoter. In this case, the promoter may be operably linked to the therapeutic agent. Expression may be indirectly controlled by the promoter. Direct and indirect control are discussed in more detail below. The promoter may represent the full-length wild-type promoter for a given gene or any functional portion or mutant thereof able to control expression of the at least one therapeutic agent.

When expression of the therapeutic agent is directly or indirectly controlled by a promoter, exposure of the CD 14+ cell to the tumour-associated microenvironment signals may increase the activity (e.g. the transcriptional activity) of the promoter. In other words, the promoter may be differentially active when the CD 14+ cell is within a tumour compared to when the CD 14+ cell is not within the tumour. In particular, the promoter may be more active when the CD 14+ cell is within a tumour compared to when the CD 14+ cell is not in the tumour. In other words, promoter activity may be induced by location of the CD 14+ cell in the tumour and, consequently, exposure of the CD 14+ cell to the tumour microenvironment.

Increased promoter activity may, for example, be mediated by an increased presence of one or more stimulatory transcription factors following exposure to tumour- associated micro-environmental signals. Increased promoter activity may, for example, be mediated by a reduced presence of one or more inhibitory transcription factors following exposure to tumour-associated micro-environmental signals.

The promoter typically is not natively associated with the therapeutic agent, where said agent is a naturally occurring gene product. The promoter may be a promoter which, natively functions to control expression of a secreted gene product. For example the promoter may be a 1110, 116, CCL3, CCL8, CCL22, FN1, SPP1 or VEGF promoter. The promoter may be a promoter which natively functions to control expression of a cytokine. For example, the promoter may be an ill 0 promoter, an U6 promoter or a VEGF promoter. The promoter may be a promoter which natively functions to control expression of a component of the extracellular matrix. For example the promoter may be a FN1 or SPP1 promoter. The promoter may be a promoter which natively functions to control expression of a cytokine receptor. For example, the promoter may be an 117 RB promoter. The promoter may be a promoter which natively functions to control expression of a chemokine. For example, the promoter may be a CCL3, CCL8, or CCL22 promoter. The promoter may be a promoter which natively functions to control expression of a chemokine receptor. For example, the promoter may be a CXC3R1 promoter. The promoter may be a promoter which natively functions to control expression of a cargo receptor. For example, the promoter may be a MRC1, ITGAM, MARCO or CD 163 promoter. The promoter may be a promoter which natively functions to control expression of a homodimeric protein. For example, the promoter may be a VEGF, CTSC, CCL3 or FN1 promoter. The promoter may be a promoter which natively functions to control expression of a folic acid receptor. For example, the promoter may be a Folr2 promoter. The promoter may be a promoter which, natively functions to control expression of a sialic acid-binding immunoglobulin-type lectins (Siglec). For example, the promoter may be a Siglec 1 promoter.

The promoter may be a CCL2, SR-A, SR-B, CXCR7, CX3CR1, CD32, CD23 (FCER2), CD200R1, PD-L2 (PDCD1LG2), PD-L1 (CD274), CSF1R, I11RA (IL1RN), I11R2, IL4R, CCL4, CCL13, CCL20, CCL17, CCL18, CCL22, CCL24, LYVE1, VEGFA, VEGFB, VEGFC, VEGFD, EGF, Cathepsin A (CTSA), CTSB, CSTC , CTSD, TGFB1, TGFB2, TGFB3, MMP14, MMP19, MMP9, CLEC7A, WNT7b, FASL, TNFSF12, TNFSF8, CD276 (B7-H3), VTCN1 (BH-H4), MSR1 (CD204), FN1, IRF4, IL10RB, BEST1, FOSL2, MTRNR2L12, CYBA, HSPA1B, HSPA1A, COL1A1, IGF1, EGR1, SLC40A1, USP53, SRGAP2, DUSP6, SESN1, AXL, GPR34, SLC15A3, SLC02B1, FAM53B, DAB2, EPB41L2, TMEM176B, CHKA, MS4A7, FOLR2, SEPP1, EIF3M, TRIM22, MAF, PLD3, C1QA, WASF2, NR3C1, MBNL1, TXNIP, OGFRL1, TNFRSF1A, FGL2, C1QC, DUSP1, PLXDC2, CIRBP, CSF1R, FOS, DPYSL2, MAFB, C1QB, HLA.E, SEC 11 A, PABPC4, SRSF5, KLF6, CCNI, HERPUD1, RNF130, MARCKS, NCOA4, SLA, RPS20, FCGRT, ITM2B, RPL22, CXCL16, MS4A6A, JUNB, CEBPD, TPT1, CD68, MT.ND2, RPL5, CST3, CD74, RPS23, HLA.DPB1, RPL37A, RPS11, MALAT1, RPS4X, FTL ,MT.NDl, RPL19, RPL37, RPS14, MT.C03, RPL27A, MT.ND2, MT.ATP6, RPL3, UBC, RPS7, GNAS, MT.ND3, RPS10, TMBIM6, MT.ND5, HLA.DRB1, HSP90AA1, TXNIP, ZFP36L1, CALR, UBB, MT.ND4, MCL1, CTSB, STK17B, PNRC1, SRSF5, HNRNPA3, BTG1, HSP90B1, SRSF2, BRD2, LCP2, MT.C03, MZT2B, ITM2B, PREX1, NFKBIA, PTMA, SPINT2, HSPA5, ALOX5AP, CITED2, ATP1A1, RBM3, CSF1R, ARL4C, RGS10, HLA.DRB5, SRSF7, CD63, TUBA1B, PPP1R15A, ADAP2, MS4A7, HLA.DQA1, CEBPD, CANX, KLF2, PDIA3, COX6C, HLA.DQB1, TUBB, PEBP1, MT2A, PIM1, CTSC, CD74, JUND, GLUL, ZFAND5, BTG2, KLRD1, EVL, AP1B1, IL32, CD163, NUCKS1, DNAJB1, DUSP1, HLA.DRA, KLF4, ARL6IP1, FOS, CD81, MAFB, XBP1, TRAC, SAMSN1, ETS2, ID2, CIRBP, ZFP36L2, HLA.DPA1, IFNGR1, ETS1, SLA, MARCKS, IER3, KLF6, GADD45B, HLA.DPB1, ZFP36, PLIN2, CXCL16, FOSB, CCL5, HSPB1, FKBP5, CPM, HLA.DOA, IGKC, HERPUD1, CSF1, NR4A2, IL7R, DUSP2, TSC22D3, CD83, IL1B, TNFAIP3, GSN, DDIT4, HLA.DQA2, PER1, RHOB, TSC22D1, CXCR4, GPR183, VWA5A, NR4A1, C3AR1, CCL3L3, PDK4, SGK1, JUN, CCL3, MIR24.2, MRC1, CCL4, CD69, EGR1, PLTP, HBB, CLU, SLC18A2, SPARC, LMNA, IGFBP7, LGMN, MAF, AREG, DAB2, RNASE1, C3, CCL4L2, FN1, ATF3, USP53, C1QA, VSIG4, FCGBP, RGS1, MS4A2, LTF, SCGB2A2, KIT, GATA2, LYVE1, CTGF, COL6A3, ADIRF, AQP1, STC2, COL3A1, COL6A1, CXCL12, ACKR1, COL1A2, CP A3, TIMP3, IGFBP4, FABP4, APOE, SLC02B1, C1QC, A2M, SEPP1, DCN, MGP, C1QB, S100A12, S100A8, S100A9, SELL, LILRA2, CSTA, S1PR4, FCN1, VC AN, CFP, VNN2, TKT, MY01G, CRISPLD2, AGTRAP, LYZ, PLBD1, IRF1, IL17RA, TALDOl, CLEC12A, ANKRD13D, FGR, ICAM3, ANPEP, TNFAIP8L2, CASP1, NCF2, G6PD, SLC7A7, PYCARD, STXBP2, MEGF9, CYP1B1, APOBEC3A, CTSS, PRAM1, CORO 1 A, MNDA, PGD, LTA4H, CCDC69, TMEM176A, POU2F2, CSF3R, SC02, MYOlF, SERPINA1, CD300E, MPEG1, WARS, TSPO, PILRA, LSP1, LILRB3, LIMD2, NFAM1, GLRX, CDC42EP3, CEBPA, LST1, LILRB2, LRRC25, CPPED1, TYMP, CD52, PTPN6, PRKCD, HCK, METTL9, TNFSF10, GMFG, CD48, STAT6, RXRA, SHKBP1, ARAP1, TNFSF13B, SERPINB1, S 100A6, RAB27A, TNFAIP2, ZNF385A, RHOG, CD37, APOBR, USP15, GIMAP1, IT GAL, TNFRSF1B, RTN3, FMNL1, ZYX, PRKCB, ARHGAP26, FGL2, AP1S2, TNIP1, SASH3, FAM65B, SPI1, IL6R, CYTH4, S100A4, ATG16L2, TMEM176B, SCPEP1, STAT1, BIN2, SH3BGRL3, NINJ1, ITGB2, GIMAP4, RAB32, GAPDH, VASP, GPSM3, ARPC1B, HRH2, CNPY3, PLP2, TYROBP, EVI2B, MX2, EHD1, RAC2, PLEKHOl, PLEC, DOK2, ARHGAP27, JAML, FLNA, PKM, ANP32A, HK1, PLXNB2, PTAFR, FKBP1A, PTK2B, EMP3, EHBP1L1, C140RF2, LGALS1, PVL, BRI3, PRELID 1, CHMP2A, PRKCSH, STK10, GRN, LGALS9, MYD88, CYBB, LCP1, SKAP2, S100A11, TAPBP, ZNF106, FCER1G, GLIPR2, FPR1, TRIM38, CCND3, NHSL2, IT GAM, S100A10, CD36, TSC22D4, APLP2, OAZ2, LYN, TCEB2, EVI2A, PCSK7, CD44, RPS13, ALDH2, RNF166, GSTP1, EFHD2, DOCK2, CYBA, AIF1, RNF213, SLC43A2, SAMHD1, COTL1, ARHGDIB, CECR1, GRK2, ARRB2, ZEB2, THEMIS2, CHMP4B, PLEK, ACTB, CNN2, HCST, PSAP, CAPG, SERF2, OAZ1, ARHGAP30, PTP4A2, CAPNS1, ZNF207, RPL28, ATP5E, EIF3G, ATP5B, MSN, ARPC3, PSME1, TPP1, LRRFIP1, CD68, ASAH1, ENOl, LAMTOR1, COX7C, EIF3F, UQCRH, WAS, RSRP1, PECAM1, Sep-09, PPP1R18, EIF3E, PFDN5, SH3BGRL, GRB2, ZBTB7A, UQCRB, SLC25A6, NEAT1, EIF3K, CFLAR, C10RF162, RPS11, HLA.C, GNAI2, MYH9, ANXA2, LAPTM5, RPL6, PCBP1, ACTR2, C10ORF54, FXYD5, CAP1, GABARAP, UBA52, RPL23A, FTL, RPL29, CLIC1, RPS14, CCNI, RPL38, PFN1, RPL13A, MT.ND1, ATP5G2, RPL37A, HLA.B, RPS23, MT.COl, RPL23, FTH1, PABPC1, RPS15, RPL10, RPS16, RPS24, RPLP2, HLA.A, TMSB10, RPL32, RPL7A, RPL18, CD14, RPL11, RPL19, RPL13. IL8, PDGFRB, IL23A, TEK, IL17C, PDGFRL, PDGFA, CCR6, IL13RA2, HSPA1A, CCR2, HIF1A, IL13RA1, MIF, IL10RA, FGF2, IL13, IL4, PITPNM3, CCL14, or CLEC10A promoter.

Preferably, the promoter is a promoter listed in Table 5 below. Other preferred promoters may be selected from a IL10, IL6, IL1B, FOLR2, MRC1 (CD206), CCL3, TGFB1, VEGFA, MARCO or CD 163 promoter. Particularly preferably, the promoter may be a CCL3, CCL8, FN1, SPP1 or SIGLEC1 promoter. Particularly preferably, the promoter is a CCL3 promoter. Particularly preferably, the promoter is a CCL8 promoter. Particularly preferably, the promoter is a FN1 promoter. Particularly preferably, the promoter is a SPP1 promoter. Particularly preferably, the promoter is a SIGLEC1 promoter.

The promoter may be a fully synthetic promoter designed to respond to transcription factors and/or repressors, expression of which is induced or suppressed within the CDl4 ⁺ cell in response to tumour-associated microenvironmental signals. Transcription factors with differential activity in CD14+ cells located within a tumour and/or exposed to tumour-associated micro-environmental signals compared to their bloodborne precursor cells may include (but are not limited to) IRF2, IRF7, IRF9 and STAT2.

When expression of the therapeutic agent is indirectly controlled by a promoter, exposure of the CD 14+ cell to the tumour-associated microenvironment signals may decrease the activity (e.g. the transcriptional activity) of the promoter. In other words, the promoter may be differentially active when the CD 14+ cell is within a tumour compared to when the CD 14+ cell is not within the tumour. In particular, the promoter maybe less active when the CD 14+ cell is within a tumour compared to when the CD 14+ cell is not in the tumour. In other words, promoter activity may be reduced by location of the CD 14+ cell in the tumour and, consequently, exposure of the CD 14+ cell to the tumour microenvironment.

In some embodiments, the promoter is not arginase-l . In some embodiments, the promoter is not Tie2. microRNA

Expression of the therapeutic agent may be controlled by one or more microRNA binding sites comprised in the nucleic acid. At least two, at least three, at least four microRNA sites may be present (including up to twelve or more sites), to allowing for binding of multiple microRNAs. The nucleic acid may include multiple copies of identical sites (e.g. two, three, four or more identical sites) for binding of the same microRNA. The nucleic acid may include multiple binding sites for binding of different microRNAs (e.g. two, three or four or more different microRNAs), each of which may be present in one or more copies (e.g up to four different copies). The nucleic acid may thus comprise two, three, four or more binding sites for two, three, four or more different microRNAs. The microRNA binding site(s) may be located in the 3’UTR of the mRNA encoding the therapeutic agent. The microRNA binding sites may be provided in either orientation. When the nucleic acid comprises one or more microRNA binding sites to control expression of the therapeutic agent, exposure to the tumour-associated micro- environmental signals may reduce the amount of microRNA molecules in the CD 14+ cell that are capable of binding to the microRNA binding site. In this way, microRNA- mediated inhibition of expression of the therapeutic agent may be reduced. In more detail, the mRNA encoding the therapeutic agent may be selectively stabilised when the CD 14+ cell is located within the tumour and/or exposed to tumour-associated micro environmental signals, due to a reduction in the amount of microRNA molecules capable of degrading the mRNA. To achieve this effect, the nucleic acid may be designed such that the mRNA encoding the therapeutic agent contains a recognition site for an endogenous microRNA molecule that is naturally decreased in CD 14+ cells located within a tumour or exposed to tumour-associated micro-environmental signals compared to CD 14+ cells in the blood or normal (non-tumour) tissue. The microRNA may decrease on differentiation and/or polarisation of a CD14+ cell in response to tumour-associated micro-environmental signals. The microRNA is not a microRNA that decreases on differentiation of a HSC or HPC, such as mir-l30a or mir-l26.

Expression of the at least one therapeutic agent may be controlled by the one or more microRNAs or both by the microRNA binding site(s) comprised in the nucleic acid and a promoter. For example, the nucleic acid may be designed such that the mRNA encoding the therapeutic agent contains a recognition site for a short hairpin RNA (shRNA) typically also encoded in the nucleic acid whose expression is controlled by a promoter that becomes less transcriptionally active when the CDl4 ⁺ cell locates within a tumour and/or is exposed to tumour-associated micro-environmental signals.

In other preferred aspects, expression of the at least one therapeutic agent is controlled both by increase of the activity of the promoter for the therapeutic agent and by reduction of the amount of microRNA molecules capable of binding to the one or more microRNA binding sites, on exposure to the tumour-associated micro-environmental signals.

MicroRNAs for which binding sites may be provided include any of the of those listed in Table 4 below. The microRNAs are preferably selected from one or more of miR-l55-5p, miR-l55-3p, miR-l47b-3p, miR-328-3p, miR-3 l8l-5p, miR-4773-3p, miR-7702-5p, miR-505-3p, miR-449b-5p, miR-l9a-5p, miR-l93b-3p, miR-532-5p and miR-50l-3p. A particularly preferred microRNA is miR-l55-5p. One or more binding sites for any of the above one or more microRNAs may be employed with any promoter of the invention, including any promoter from Table 5 and in particular a CCL3, CCL8, FN1, SPP1 or SIGLEC1 promoter.

Differentiation and polarisation

Expression of the at least one therapeutic agent may be activated by differentiation and/or polarisation of the CD14+ cell in response to the tumour-associated micro environmental signals. The tumour-associated micro-environmental signals may, for example, induce differentiation and/or polarisation of the CD 14+ cell such that activity of a promoter controlling expression of the at least one therapeutic agent is increased. The tumour-associated micro-environmental signals may, for example, induce differentiation and/or polarisation of the CD14+ cell such that microRNA-mediated inhibition of expression of the at least one therapeutic agent is reduced. The tumour associated micro-environmental signals may induce differentiation and/or polarisation of the CD 14+ cell such that activity of a promoter controlling expression of the at least one therapeutic agent is increased, and microRNA-mediated inhibition of expression of the at least one therapeutic agent is reduced (typically by a reduction of the amount of microRNA molecules capable of binding to the one or more microRNA binding sites). Preferably, the CD 14+ cell is a monocyte and expression of the therapeutic agent is activated by differentiation or polarisation of the monocyte in response to the tumour- associated micro-environmental signals, typically into a tumour-associated myeloid- derived cell (TAMDC) or myeloid-derived tumour infiltrating cell (MDTIC). The tumour-associated micro-environmental signals may induce differentiation of the monocyte into a macrophage. The tumour-associated micro-environmental signals may induce polarisation of the monocyte into a M2 macrophage. The tumour-associated micro-environmental signals may induce differentiation and polarisation of the monocyte into tumour-associated macrophage (TAM). The tumour-associated micro environmental signals may induce differentiation and/or polarisation of the CD 14+ cell into a myeloid derived suppressor cell.

In one aspect, the present invention provides a modified CD 14-expressing (CDl4 ⁺ ) cell comprising a nucleic acid encoding at least one therapeutic agent, expression of which is/are increased within the CDl4 ⁺ cell in response to tumour-associated micro environmental signals as the cell engrafts into tumour tissue and becomes an MDTIC. This increased expression can result from activation or increased transcription due to direct or indirect transcriptional control by a promoter in particular an Upregulated Target Promoter that is activated within the modified CDl4 ⁺ cells as they become MDTICs by transcription factors that are induced (or repressors that are suppressed) by tumour- associated micro-environmental signals.

Without limiting the invention it is anticipated that reinfused modified CDl4 ⁺ cells (predominantly monocytes) will engraft into tumours as MDTIC and may become TAMDCs that assist with immunosuppression and/or wound healing within tumours. Factors secreted by tumours that promote engraftment of CD 14+ cells include, but are not limited to ccl2/mcp-l and csfl/mcsf.

Tumour-associated micro-environmental signals that may cause the modified MDTICs or modified TAMDCs to activate expression of the therapeutic agent(s) include but are not limited to cytokines such as IL-l, IL-2, IL-3, IL-4, IL-6, IL-7, IL-8, IL-10, IL- 12, IL-13, IL-15, IL-17, IL-18, IL-21, IL-23, VEGF, FGF2, TGFbeta, CSF, MCSF, GCSF, GMCSF, TNFalpha, lymphotoxin, MIF, Fas, Fas ligand, type 1 interferons (particularly alpha-interferon and beta-interferon), type 2 interferons (gamma-interferon), type 3 interferons (gamma 1 -interferon, gamma2 -interferon, gamma3 -interferon), TRAIL, FLT3 ligand, Lymphotactin and chemokines.

Therapeutic agent

The nucleic acid comprised in the modified CD 14+ cell encodes at least one therapeutic agent. For example, the nucleic acid may encode at least two, at least three, at least four, at least five, at least six, at least seven, at least eight, at least nine or at least ten therapeutic agents.

The therapeutic agent is preferably an agent having therapeutic benefit in an individual having a tumour. For example, the therapeutic agent may be an anti-cancer agent.

The therapeutic agent(s) may be selected from, but are not limited to, antibodies (including IgGs, IgAs, IgD, IgEs and IgMs, fragments of antibodies, single chain antibodies such as camelid or shark antibodies, single domain antibodies including V _H H and VNAR fragments, bispecific and trispecific antibodies made from formats such as single chain Fv (scFv) regions, nanobodies or V _H domains coupled together, darpins, cytokines, chemokines, enzymes, toxins (including fragments of toxins) or composite molecules containing more than one functionality (for example antibody-toxin fusion proteins).

The therapeutic agent may comprise a bispecific antibody. The therapeutic agent may comprise a chemokine. In one embodiment, the therapeutic agent is a bispecific antibody or a chemokine. In a further embodiment, at least two therapeutic agents are expressed. The therapeutic agents may both be anticancer agents.

Bispecific antibodies that are useful in the invention include but are not limited to bispecific T cell activating antibodies (which may bind any tumour surface antigen and any T cell activating surface antigen), such as bispecific T cell engagers (BiTEs), one end of which recognises an antigen on the surface of an immune cell, such as CD3 on the T cell surface, and the other end recognises an antigen on the surface of a target cell. Although simple bispecific BiTEs cannot cause clustering of the CD3 on their own, when they bind the CD3 to an antigen on another cell, multiple BiTEs can cause formation of an immunological pseudo-synapse, which does cluster CD3 and leads to signalling and T cell activation. In this way T cells are activated by crosslinking of their T cell receptors (TCR) through the associated CD3s to a cell surface target on another cell, independent of the intrinsic specificity of the TCR. Presentation of peptides on HLA molecules, and binding to cognate TCRs, is not involved in this process. This means that instead of receiving some 50-100 activatory signals in the physiological setting (reflecting the number of identical peptides thought to be presented on HLA by each cell), BiTEs can deliver activatory signals to every CD3 on the T cell surface, estimated at many thousands. In this way T cells activated by BiTEs may be regarded as‘superactivated’ and may be able to overcome signals that would otherwise induce anergy. More particularly a bispecific T-cell activating antibody is capable of crosslinking CD3s on the surface of a T cell to epithelial cell adhesion molecule (EpCAM) on the surface of an epithelial or carcinoma cell (such as an EpCAM BiTE). The antiCD3s component of the bispecific T-cell activating antibody may be a single chain variable fragment (scFv) known as OKT3 although many other suitable scFv and nanobodies that can bind CD3s exist. EpCAM BiTEs show very powerful EpCAM-specifie cytotoxicity in vitro and are therefore potential anticancer agents. Versions of EpCAM BiTE include MT110 (Solitomab) which has been evaluated in a multicentre dose escalating clinical trial given intravenously to patients with advanced and refractory EpCAM-positive solid tumours that were not suitable for standard therapies. Sixty-five patients were treated with solitomab at doses up to 96 pg/day for >28 days. Solitomab showed a circulatory half life of 4.5 hours. Fifteen patients had dose-limiting toxicities (DLTs) and the maximum tolerated dose was 24 pg/day. The trial concluded that use of solitomab in this way was associated with DLTs, including severe diarrhoea and increased liver enzymes, which precluded dose escalation to potentially therapeutic levels. EpCAM BiTEs therefore offer significant anticancer potential but are limited by systemic toxicities when administered intravenously as free proteins; they therefore comprise promising agents for targeted expression within tumours as described herein, as this approach maximises their anticancer activity and minimises the toxicities resulting from systemic exposure.

Other bispecific T-cell activating antibodies (such as BiTEs) that may be useful in accordance with the invention include but are not limited to those linking T cells to NKG2D ligands (for example nanobody recognising CD3 epsilon fused to NKG2D) or CD47 (eg. nanobody recognising CD3 epsilon fused to nanobody recognising CD47), or to other cancer targets (PDL1, PDL2, CEA, Her2, EGFR, folate receptor (FOLR1), prostate-specific membrane antigen, PSMA) or to antigens on stromal cells within the tumour microenvironment (eg. folate receptor beta (FOLR2), fibroblast-activation protein, CD206).

Chemokines that may be useful in the invention include but are not limited to chemokines known to attract T cells of various types, including CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20, CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12, CXCL16.

Nucleic acid The form of the nucleic acid encoding the therapeutic agent(s) may be selected from, but not limited to, an adenovirus, an adeno-associated virus, a simian lentivirus (as described by Muhlebach et al., Molecular Therapy 12(6) 1206-16, 2005), a circular plasmid (including CpG-deleted plasmid), a linear plasmid, a circular DNA minicircle, a linear DNA minicircle or rnRNA. Preferably the form is a plasmid. Most preferably, the form is an adenovirus or lentivirus, i.e an adenoviral or lentiviral vector is used. An HIV1- based lentiviral vector may be used, preferably in combination with delivering VPX protein either before or simultaneously with the HIV 1 lentiviral vector (to address effects of SAMHD1).

Expression

The present invention provides a modified CD 14+ cell comprising a nucleic acid encoding at least one therapeutic agent whose expression is activated within a tumour in response to tumour-associated micro-environmental signals. As set out above, expression of the therapeutic agent in the CD 14+ cell is preferably inactive before the cell locates within a tumour and is exposed to tumour-related microenvironment signals. That is, the CD 14+ cell may not express the therapeutic agent, or may express a biologically insignificant amount of the therapeutic agent, before the cell locates within a tumour and is exposed to tumour-related microenvironment signals. Expression of the therapeutic agent may therefore be undetectable or minimally detectable in the CD 14+ cell prior to its location in the tumour. Methods for determining expression of a therapeutic agent by a cell are well-known in the art.

Expression of the therapeutic agent may be activated in a non-lineage specific manner. That is, expression of the therapeutic agent may be activated in a manner that is not specific to a particular lineage of CD14+ cell. For example, when the CD14+ cell locates within the tumour and/or is exposed to tumour-associated micro-environmental signals, expression of the therapeutic agent may be activated irrespective of its subtype (e.g. Tie2+ monocyte, Tie2- monocyte, CDl4 ^hl CDl6 ^l0 monocyte, CDl4 ^hl CDl6 ^imd monocyte, or CDl4 ^lo CDl6 ^hl monocyte) or the subtype of any cell into which it differentiates or is polarised (e.g. Tie2+ macrophage, or Tie2- macrophage) advantageously allowing for use of CD 14+ cells of any subtype or subclass. For instance, expression of the therapeutic agent may be activated irrespective of whether or not the CD 14+ cell expresses Tie2 before or after its location within a tumour and/or exposure to tumour-associated micro-environmental signals. When a plurality of modified CD 14+ cells locate within the tumour and/or are exposed to tumour-associated micro-environmental signals, expression of the therapeutic agent may be activated in all (or substantially all) of the modified CD 14+ cells, rather than only a subset thereof (e.g. Tie2+ cells, or Tie2- cells). For example, expression of the therapeutic agent may be activated in 95% or more (such as 96% or more, 97% or more, 97.5% or more, 98% or more, 98.5% or more, 99% or more, or 99.5% or more) of the modified CD14+ cells located within the tumour and/or exposed to tumour-associated micro-environmental signals.

Activation of expression in a non-lineage specific matter is advantageous because it allows the therapeutic agent to be expressed broadly within the modified CD 14+ cell population. Expression is not limited to a subset of CD 14+ cells. This allows more cells to express the therapeutic agent, improving the amount of therapeutic agent delivered to the tumour and, therefore, the efficacy of treatment.

Expression of the therapeutic agent may be activated in a non-tissue specific manner. That is, expression of the therapeutic agent may be activated in a manner that is not specific to the tissue from which the tumour is derived. The therapeutic agent is capable of expression in response to micro-environmental signals from more than one tumour type. For example, the therapeutic agent may be capable of expression when the CD 14+ cell is located within and/or exposed to micro-environmental signals from a first tumour type (e.g. breast cancer), and also when the CD14+ cell is located within and/or exposed to micro-environmental signals from a second tumour type (e.g. non small cell lung cancer).

Activation of expression in a non-tissue specific matter is advantageous because it allows the design of a single modified CD 14+ cell to treat more than one tumour type. Put another way, a given modified CD 14+ cell has broad anti-cancer efficacy, irrespective of the tissue of origin of the tumour.

The expression of the therapeutic agent(s) is preferably controlled, directly or indirectly, by a promoter. To be useful in the invention, promoters must be differentially active within MDTICs compared to bloodborne CD 14+ cells or cells derived from them within normal tissues such as normal breast tissue. These promoters are “Target Promoters”, and the expression of the therapeutic agent(s) can result from increased transcription due to direct or indirect transcriptional control by a Target Promoter that is activated within the MDTICs by the actions of transcription factors and repressor proteins that are regulated by tumour-associated micro-environmental signals (‘Upregulated Target Promoter’) and/or increased translation reflecting selective mRNA stabilisation due to decreased levels of microRNA molecules capable of its degradation, in response to tumour-associated micro-environmental signals. This latter strategy may result from designing the therapeutic agent encoded mRNA to contain recognition sites for endogenous microRNA molecules that are naturally decreased in MDTICs compared to their bloodborne precursors (or their equivalent in normal tissues) or may result from designing the therapeutic agent encoded mRNA to contain recognition sites for short hairpin RNAs (shRNA) encoded within the mRNA encoding the therapeutic agent under the control of a Target Promoter that become less transcriptionally active as the CDl4 ⁺ precursor cells become MDTICs, in response to tumour-associated micro-environmental signals (‘Downregulated Target Promoter’).

The differential Target Promoter activity within MDTIC compared to their counterparts in normal tissue and blood is predominantly in response to tumour microenvironmental signals. Target Promoters are predominantly a subset of the group of promoters that are differentially expressed within TAMDC compared to myeloid cells within healthy tissue. This larger group of differentially-expressed TAMDC promoters are “Candidate Promoters”, from which Target Promoters are identified for use in accordance with the invention.

Candidate Promoters that may be deemed useful in the invention include but are not limited to IL10, IL6, IL1B, FOLR2, MRC1 (CD206), CCL3, TGFB1, VEGFA, MARCO and CD 163. Other candidate promoters include CCL2, SR-A, SR-B, CXCR7, CX3CR1, CD32, CD23 (FCER2), CD200R1, PD-L2 (PDCD1LG2), PD-L1 (CD274), CSF1R, I11RA (IL1RN), I11R2, IL4R, CCL4, CCL13, CCL20, CCL17, CCL18, CCL22, CCL24, LYVE1, VEGFA, VEGFB, VEGFC, VEGFD, EGF, Cathepsin A (CTSA), CTSB, CSTC , CTSD, TGFB1, TGFB2, TGFB3, MMP14, MMP19, MMP9, CLEC7A, WNT7b, FASL, TNFSF12, TNFSF8, CD276 (B7-H3), VTCN1 (BH-H4), MSR1 (CD204), FN1, IRF4, IL10RB, BEST1, FOSL2, MTRNR2L12, CYBA, HSPA1B, HSPA1A, COL1A1, IGF1, EGR1, SLC40A1, USP53, SRGAP2, DUSP6, SESN1, AXL, GPR34, SLC15A3, SLC02B1, FAM53B, DAB2, EPB41L2, TMEM176B, CHKA, MS4A7, FOLR2, SEPP1, EIF3M, TRIM22, MAF, PLD3, C1QA, WASF2, NR3C1, MBNL1, TXNIP, OGFRL1, TNFRSF1A, FGL2, C1QC, DUSP1, PLXDC2, CIRBP, CSF1R, FOS, DPYSL2, MAFB, C1QB, HLA.E, SEC 11 A, PABPC4, SRSF5, KLF6, CCNI, HERPUD1, RNF130, MARCKS, NCOA4, SLA, RPS20, FCGRT, ITM2B, RPL22, CXCL16, MS4A6A, JUNB, CEBPD, TPT1, CD68, MT.ND2, RPL5, CST3, CD74, RPS23, HLA.DPB1, RPL37A, RPS11, MALAT1, RPS4X, FTL ,MT.NDl, RPL19, RPL37, RPS14, MT.C03, RPL27A, MT.ND2, MT.ATP6, RPL3, UBC, RPS7, GNAS, MT.ND3, RPS10, TMBIM6, MT.ND5, HLA.DRB1, HSP90AA1, TXNIP, ZFP36L1, CALR, UBB, MT.ND4, MCL1, CTSB, STK17B, PNRC1, SRSF5, HNRNPA3, BTG1, HSP90B1, SRSF2, BRD2, LCP2, MT.C03, MZT2B, ITM2B, PREX1, NFKBIA, PTMA, SPINT2, HSPA5, ALOX5AP, CITED2, ATP1A1, RBM3, CSF1R, ARL4C, RGS10, HLA.DRB5, SRSF7, CD63, TUBA1B, PPP1R15A, ADAP2, MS4A7, HLA.DQA1, CEBPD, CANX, KLF2, PDIA3, COX6C, HLA.DQB1, TUBB, PEBP1, MT2A, PIM1, CTSC, CD74, JUND, GLUL, ZFAND5, BTG2, KLRD1, EVL, AP1B1, IL32, CD 163, NUCKS1, DNAJB 1 , DUSP1, HLA.DRA, KLF4, ARL6IP1, FOS, CD81, MAFB, XBP1, TRAC, SAMSN1, ETS2, ID2, CIRBP, ZFP36L2, HLA.DPA1, IFNGR1, ETS1, SLA, MARCKS, IER3, KLF6, GADD45B, HLA.DPB1, ZFP36, PLIN2, CXCL16, FOSB, CCL5, HSPB1, FKBP5, CPM, HLA.DOA, IGKC, HERPUD1, CSF1, NR4A2, IL7R, DUSP2, TSC22D3, CD83, IL1B, TNFAIP3, GSN, DDIT4, HLA.DQA2, PER1, RHOB, TSC22D1, CXCR4, GPR183, VWA5A, NR4A1, C3AR1, CCL3L3, PDK4, SGK1, JUN, CCL3, MIR24.2, MRC1, CCL4, CD69, EGR1, PLTP, HBB, CLU, SLC18A2, SPARC, LMNA, IGFBP7, LGMN, MAF, AREG, DAB2, RNASE1, C3, CCL4L2, FN1, ATF3, USP53, C1QA, VSIG4, FCGBP, RGS1, MS4A2, LTF, SCGB2A2, KIT, GATA2, LYVE1, CTGF, COL6A3, ADIRF, AQP1, STC2, COL3A1, COL6A1, CXCL12, ACKR1, COL1A2, CP A3, TIMP3, IGFBP4, FABP4, APOE, SLC02B1, C1QC, A2M, SEPP1, DCN, MGP, C1QB, S100A12, S100A8, S100A9, SELL, LILRA2, CSTA, S1PR4, FCN1, VC AN, CFP, VNN2, TKT, MYOlG, CRISPLD2, AGTRAP, LYZ, PLBD1, IRF1, IL17RA, TALDOl, CLEC12A, ANKRD13D, FGR, ICAM3, ANPEP, TNFAIP8L2, CASP1, NCF2, G6PD, SLC7A7, PYCARD, STXBP2, MEGF9, CYP1B1, APOBEC3A, CTSS, PRAM1, COROlA, MNDA, PGD, LTA4H, CCDC69, TMEM176A, POU2F2, CSF3R, SC02, MYOlF, SERPINA1, CD300E, MPEG1, WARS, TSPO, PILRA, LSP1, LILRB3, LIMD2, NFAM1, GLRX, CDC42EP3, CEBPA, LST1, LILRB2, LRRC25, CPPED1, TYMP, CD52, PTPN6, PRKCD, HCK, METTL9, TNFSF10, GMFG, CD48, STAT6, RXRA, SHKBP1, ARAP1, TNFSF13B, SERPINB1, S100A6, RAB27A, TNFAIP2, ZNF385A, RHOG, CD37, APOBR, USP15, GIMAP1, ITGAL, TNFRSF1B, RTN3, FMNL1, ZYX, PRKCB, ARHGAP26, FGL2, AP1S2, TNIP1, SASH3, FAM65B, SPI1, IL6R, CYTH4, S100A4, ATG16L2, TMEM176B, SCPEP1, STAT1, BIN2, SH3BGRL3, NINJ1, ITGB2, GIMAP4, RAB32, GAPDH, VASP, GPSM3, ARPC1B, HRH2, CNPY3, PLP2, TYROBP, EVI2B, MX2, EHD1, RAC2, PLEKHOl, PLEC, DOK2, ARHGAP27, JAML, FLNA, PKM, ANP32A, HK1, PLXNB2, PTAFR, FKBP1A, PTK2B, EMP3, EHBP1L1, C140RF2, LGALS1, PVL, BRI3, PRELID 1, CHMP2A, PRKCSH, STK10, GRN, LGALS9, MYD88, CYBB, LCP1, SKAP2, S100A11, TAPBP, ZNF106, FCER1G, GLIPR2, FPR1, TRIM38, CCND3, NHSL2, IT GAM, S100A10, CD36, TSC22D4, APLP2, OAZ2, LYN, TCEB2, EVI2A, PCSK7, CD44, RPS13, ALDH2, RNF166, GSTP1, EFHD2, DOCK2, CYBA, AIF1, RNF213, SLC43A2, SAMHD1, COTL1, ARHGDIB, CECR1, GRK2, ARRB2, ZEB2, THEMIS2, CHMP4B, PLEK, ACTB, CNN2, HCST, PSAP, CAPG, SERF2, OAZ1, ARHGAP30, PTP4A2, CAPNS1, ZNF207, RPL28, ATP5E, EIF3G, ATP5B, MSN, ARPC3, PSME1, TPP1, LRRFIP1, CD68, ASAH1, ENOl, LAMTOR1, COX7C, EIF3F, UQCRH, WAS, RSRP1, PECAM1, Sep-09, PPP1R18, EIF3E, PFDN5, SH3BGRL, GRB2, ZBTB7A, UQCRB, SLC25A6, NEAT1, EIF3K, CFLAR, C10RF162, RPS11, HLA.C, GNAI2, MYH9, ANXA2, LAPTM5, RPL6, PCBP1, ACTR2, C10ORF54, FXYD5, CAP1, GABARAP, UBA52, RPL23A, FTL, RPL29, CLIC1, RPS14, CCNI, RPL38, PFN1, RPL13A, MT.ND1, ATP5G2, RPL37A, HLA.B, RPS23, MT.COl, RPL23, FTH1, PABPC1, RPS15, RPL10, RPS16, RPS24, RPLP2, HLA.A, TMSB10, RPL32, RPL7A, RPL18, CD14, RPL11, RPL19, RPL13. IL8, PDGFRB, IL23A, TEK, IL17C, PDGFRL, PDGFA, CCR6, IL13RA2, HSPA1A, CCR2, HIF1A, IL13RA1, MIF, IL10RA, FGF2, IL13, IL4, PITPNM3, CCL14, CLEC10A

Alternatively the promoter may be a fully synthetic promoter designed to respond to transcription factors and/or repressors, expression of which is induced or suppressed within CDl4 ⁺ cell-derived MDTICs by signals from the tumour microenvironment. Transcription factors with differential activity in MDTICs compared to their bloodborne precursor cells may include (but are not limited to) IRF2, IRF7, IRF9 and STAT2.

Ideally the expression of said anticancer therapeutic agent by the Target Promoter is activated only when the modified CDl4 ⁺ cell enters the tumour environment and differentiates towards a wound-healing phenotype such as a TAMDC.

In addition or alternatively to the promoter driven approach the increased expression of therapeutic agent(s) can result from increased translation reflecting selective mRNA stabilisation due to decreased levels of microRNA molecules capable of its inactivation or degradation, again in response to tumour-associated micro- environmental signals. To enable this strategy, the nucleic acid is mRNA designed to contain recognition sites for endogenous microRNA molecules that are naturally decreased as the CDl4 ⁺ cells become MDTICs (in response to tumour-associated microenvironmental signals) leading to decreased inactivation or degradation of the mRNA in MDTICs and greater translation of therapeutic proteins.

The expression of the anticancer therapeutic can be regulated by the incorporation of binding sites for endogenous microRNA molecules into the 3’ untranslated region of the mRNA(s) encoding the anticancer therapeutic(s). Sites can be included that are recognised by microRNA molecules that are expressed within undifferentiated CDl4 ⁺ monocytes or CDl4 ⁺ monocytes associated with non-cancerous tissue, but are lost as the CDl4 ⁺ monocytes infiltrate into tumours and differentiate into MDTICs. In this way the mRNA encoding the therapeutic agent(s) can be degraded or neutralised within undifferentiated monocytes but can become more stable when the monocytes differentiate as they engraft into the tumour. Candidate microRNA include but are not limited to miR- 223, miR-454, miR-93, miR-l50, miR-27a, miR-l8lc, miR-532, miR-l30bp, miR-4772, miR-3 l77, miR-l25ap, miR-67l .

An alternative embodiment of this strategy would be to encode short hairpin RNA(s) (shRNA) capable of binding selectively to the mRNA encoding the therapeutic agent(s) and preventing its translation, encoded within the mRNA encoding the therapeutic agent(s) under control of a Downregulated Target Promoter’ that becomes less transcriptionally active as the CDl4 ⁺ precursor cells become MDTICs in response to tumour-associated micro-environmental signals.

In addition to direct regulation of therapeutic gene expression, using target promoters and/or microRNA sites, a further embodiment of the invention uses indirect regulation where the tumour microenvironment-associated signals regulate expression of an effector protein that then regulates expression of the therapeutic construct. Such protein effectors could include Tet-on and tet-off systems (including repressors such as TETO fused to transactivator domains such as VP 16), cumate-on and cumate-off systems etc. Several of these possible constructions, including both direct and indirect mechanisms of regulating expression of the nucleic acid encoding the therapeutic agent(s) are shown in the Figures.

Population

The invention provides a cell population comprising a plurality of modified CD 14+ cells of the invention wherein said population comprises at least two different subclasses of modified CD 14+ cells. At least three, at least four, at least five, or at least six or more different subclasses of modified CD 14+ cells may be present The subclasses may comprise monocyte subclasses, monocyte-derived suppressor cells and any other subclasses of CD 14+ cells present in peripheral blood. The population may comprise all subclasses of CD 14+ cells present in peripheral blood. The population may represent at least 25%, at least 30%, at least 35%, at least 40%, at least 45% or at least 50% of the CD 14+ cells present in peripheral blood derived from a patient. The population may comprise classical, classical and intermediate, classical and non-classical, or classical, intermediate and non-classical modified monocytes. The population may comprise modified CD14+ Tie- cells.

The population represents a starting population of modified CD 14+ cells which are administered to a human and then undergo differentiation and/or polarisation to other cell types which express the therapeutic agent. Preferably, the population comprises two or more different subclasses of modified CD 14+ monocytes of the invention; wherein the tumour-associated micro-environmental signals induce differentiation of each subclass of CD 14+ cell into a macrophage and/or polarisation of each subclass of CD 14+ cell into a M2 macrophage or tumour-associated macrophage (TAM); and wherein differentiation or polarisation in each subclass of CD 14+ cell each activate expression of the therapeutic agent.

The starting monocytes may comprise Tie2- cells. The induced macrophages may comprise Tie2- cells. Both the starting monocytes and the induced macrophages may comprise Tie2- cells.

Expression of the therapeutic agent may be activated within all, or substantially all, of the induced macrophages derived from the modified CD 14+ cells. In other words, activation of expression may not be limited to only a subset of the induced macrophages derived from the modified CD 14+ cells (such as Tie2+ macrophages, or Tie2- macrophages). Rather, expression of the therapeutic agent may be activated broadly within the population of macrophages induced from the monocyte population. For example, expression of the therapeutic agent may be activated in 95% or more (such as 96% or more, 97% or more, 97.5% or more, 98% or more, 98.5% or more, 99% or more, or 99.5% or more) of the induced macrophages. Substantially all modified CD14+ cells in the starting population (such as 70%, 75%, 80%, 85%, 90%, 95% or all) may be capable of differentiating and/or polarising to activate expression of the therapeutic agent in response to the tumour-associated micro-environmental signals. Activation of expression within all, or substantially all, of the induced macrophages derived from the modified CD 14+ cells is advantageous because it allows the therapeutic agent to be expressed broadly within cells derived from the population of modified CD 14+ cells. This allows more cells to express the therapeutic agent, improving the amount of therapeutic agent delivered to the tumour and, therefore, the efficacy of treatment.

Medicaments, methods and therapeutic use

The invention provides a method of treating a disease in a human by administering a modified CD 14+ cell of the invention, or population of CD 14+ cells of the invention. The invention also provides a modified CD 14+ cell of the invention, or population of CD 14+ cells of the invention, for use in a method of treating a disease in a human. Preferably, the disease is cancer. Preferably, the cancer is a solid tumour, such as a carcinoma. The cancer may, for example, be cancer of the breast, lung, ovary, bladder, pancreas, prostate, or gastrointestinal tract including the colon, rectum, stomach and oesophagus.

In a further embodiment, the modified CDl4 ⁺ cells are used to treat cancer in humans, livestock or companion animals. Cancers which are anticipated to benefit from the treatments according to the invention are those which contain high levels of infiltrating monocytes and TAMDCs. These include many types of solid tumour, including but not limited to carcinomas of the breast, lung, ovary, bladder, pancreas, prostate, gastrointestinal tract including the colon, rectum, stomach and oesophagus. The invention provides for the treatment of cancer by administering modified CDl4 ⁺ cells by: a) infusing said modified CDl4 ⁺ cells into the patient

b) allowing the transfected or transduced modified CDl4 ⁺ cells to engraft within a target site(s) naturally or encouraging them to do so by the application of radiotherapy or chemotherapy.

Radiotherapy and chemotherapy have the effect of encouraging engraftment of bloodborne monocytes and other myeloid cells into tumours by causing damage to the tumour and encouraging it to activate wound healing processes. While the phenomenon has been long documented in murine models, clinical data are now emerging that show similar effects in humans (Miller et al. Science Translational Medicine, 2017, Vol. 9, Issue 392). These myeloid cells, predominantly TH2 polarised, play an important role in assisting tumour recovery and growth (Shiao SL et al. Cancer Immunol Res. 3(5):518- 525) and play important paracrine roles, such as the stimulation of angiogenesis (Ahn G- O et al. Proceedings of the National Academy of Sciences May 2010, 107 (18) 8363- 8368).As described above, the nucleic acid may be designed such that expression of the therapeutic agent is activated when the CD 14+ cell is exposed to tumour-associate microenvironment signals. For example, expression of the therapeutic agent may be activated only when the modified CD 14+ cell enters the tumour and matures into a MDTIC or TAMDC.

The invention also provides a method for treating cancer, wherein the method produces a modified CD 14+ cell comprising a nucleic acid encoding anticancer therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated in MDTICs or TAMDC and consists of: (a) isolating peripheral blood mononuclear cells (PBMC) from a patient, or a suitable donor, using leukapheresis, (b) culturing the isolated blood mononuclear cells ex vivo and optionally purifying a subsection of CD 14+ cells, (c) transfecting or transducing cultured CD 14+ cells from the PMCs (optionally a purified subsection) (optionally using electroporation or viral infection) to introduce a vector encoding a bispecific T-cell activating antibody (optionally binding EpCAM, such as an EpCAMBiTE) and a chemokine selected from CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20 , CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12 or CXCL16, (d) introducing the modified CD14+ cells into a patient, and (e) allowing the engraftment or localisation of the CD 14+ cells within a tumour(s) naturally or encouraging them to do so by the application of radiotherapy or chemotherapy.

The invention further provides a vector for use in a method for treating cancer, wherein the vector encodes a bispecific T-cell activating antibody, optionally binding EpCAM (optionally an EpCAMBiTE) and a chemokine selected from CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20 , CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12 or CXCL16, and wherein the method produces a modified CD 14+ cell comprising a nucleic acid encoding an anticancer therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated in MDTICs or TAMDC and consists of: (a) isolating peripheral blood mononuclear cells (PBMC) from a patient, or a suitable donor, using leukapheresis, (b) culturing the isolated blood mononuclear cells ex vivo and purifying a subsection of CD 14+ cells, (c) transfecting the purified subsection of CD 14+ cells using electroporation to introduce the vector, (d) introducing the modified CD 14+ cells into a patient, and (e) allowing the engraftment or localisation of the CD 14+ cells within a tumour(s) naturally or encouraging them to do so by the application of radiotherapy or chemotherapy.

The modified CD 14+ cell or cell population may be provided as a pharmaceutical composition. The pharmaceutical composition preferably comprises a pharmaceutically acceptable carrier or diluent. The pharmaceutical composition may be formulated using any suitable method. Formulation of cells with standard pharmaceutically acceptable carriers and/or excipients may be carried out using routine methods in the pharmaceutical art. The exact nature of a formulation will depend upon several factors including the cells to be administered and the desired route of administration. Suitable types of formulation are fully described in Remington's Pharmaceutical Sciences, 19th Edition, Mack Publishing Company, Eastern Pennsylvania, USA.

The modified CD 14+ cell, cell population or pharmaceutical composition may be administered by any route. Suitable routes include, but are not limited to, the intravenous, intramuscular, intraperitoneal, subcutaneous, intradermal, transdermal and direct intratumoural routes.

Compositions may be prepared together with a physiologically acceptable carrier or diluent. Typically, such compositions are prepared as liquid suspensions of modified CD14+ cells. The modified CD14+ cells may be mixed with an excipient which is pharmaceutically acceptable and compatible with the active ingredient. Suitable excipients are, for example, water, saline, dextrose, glycerol, of the like and combinations thereof.

In addition, if desired, the pharmaceutical compositions may contain minor amounts of auxiliary substances such as wetting or emulsifying agents, and/or pH buffering agents.

The modified CD 14+ cells are administered in a manner compatible with the dosage formulation and in such amount will be therapeutically effective. The quantity to be administered depends on the subject to be treated, the disease to be treated, and the capacity of the subject’s immune system. Precise amounts of modified CD14+ cells required to be administered may depend on the judgement of the practitioner and may be peculiar to each subject.

Any suitable number of modified CD 14+ cells may be administered to a subject. For example, at least, or about, 0.2 x 10 ⁶ , 0.25 x 10 ⁶ , 0.5 x 10 ⁶ , 1.5 x 10 ⁶ , 4.0 x 10 ⁶ or 5.0 x 10 ⁶ modified CD 14+ cells per kg of patient may administered. For example, at least, or about, 10 ⁵ , 10 ⁶ , 10 ⁷ , 10 ⁸ , 10 ⁹ modified CD14+ cells may be administered. As a guide, the number of modified CD 14+ cells to be administered may be from 10 ⁵ to 10 ⁹ , preferably from 10 ⁶ to 10 ⁸ .

Methods of making a CD14+ cell

A further aspect of the invention is a method for making a modified CDl4 ⁺ cell comprising a nucleic acid encoding the therapeutic agent or agents being selectively activated within a tumour in response to tumour-associated micro-environmental signals which consists of:

(a) isolating bloodborne cells from an individual.

(b) culturing the isolated bloodborne cells ex vivo and purifying said cells to produce a CDl4 ⁺ enriched cell population

(c) transfecting or transducing the CDl4 ⁺ enriched cell population subsection of cells to introduce a vector encoding a therapeutic agent or agents.

Isolation of said bloodborne cells may for example be by apheresis or leukapheresis. The bloodborne cells may be isolated from a patient to which it is intended to administer the modified CD 14+ cell produced by the method. Alternatively, the bloodborne cells may be isolated from a donor other than a patient to which it is intended to administer the modified CD 14+ cell produced by the method.

Transfection or transduction may be used to introduce the nucleic acid encoding the therapeutic agent using electroporation or more preferably using a viral vector such as adenovirus, adeno-associated virus or lentivirus (including a simian lentivirus capable infecting cells in GO phase of the cell cycle). The process may be undertaken using industrial-based cGMP-compatible systems such as the Miltenyi Prodigy system (Fraser et al, Cytotherapy. 2017 Sep; 19(9): 1113-1124).

In one method for making modified CD 14+ cells of the invention the following steps may be undertaken:

(a) isolating peripheral blood mononuclear cells (PBMC) from a patient, or an acceptable donor, using leukapheresis

(b) culturing the isolated cells ex vivo and optionally purifying a subsection of CDl4 ⁺ cells

(c) transfecting or transducing cultured CDl4 ⁺ cells from the bloodborne cells (optionally a purified subsection of CDl4 ⁺ cells), optionally

using electroporation or viral infection, to introduce a plasmid or vector encoding a bispecific T-cell activating antibody, optionally binding EpCAM (such as an EpCAMBiTE) and a chemokine selected from this list: CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20 , CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12, CXCL16. Said bispecific T-cell activating antibody and chemokine are under the control of a promoter and/or one or more micro RNA binding sites as described herein.

FURTHER ASPECTS OF THE INVENTION:

1. A modified CDl4 ⁺ cell comprising a nucleic acid encoding at least one

therapeutic agent the expression of which is activated within the CD 14+ cell in response to tumour-associated micro-environmental signals.

2. A modified CDl4 ⁺ cell according to item 1 wherein expression of said

therapeutic agent is directly or indirectly controlled by a promoter.

3. A modified CDl4 ⁺ cell according to item 2 wherein the promoter is a Target Promoter.

4. A modified CDl4 ⁺ cell according to item 1, wherein expression of said

therapeutic agent is, controlled by the incorporation of binding sites for microRNA molecules into the nucleic acid which encodes for said therapeutic agent.

5. A modified CDl4 ⁺ cell according to any of item 1-4, wherein at least one

therapeutic agent is an anti-cancer agent.

6. A modified CD l4 ⁺ cell according to item 5, wherein expression of said

therapeutic agent is activated when the modified CDl4 ⁺ cell enters a tumour and undergoes differentiation.

7. A modified CDl4 ⁺ cell according to any of the preceding items , wherein a therapeutic agent is a bispecific antibody or a chemokine, or any combination thereof.

8. A modified CDl4 ⁺ cell according to item 7, wherein a bispecific antibody is a T cell-activating bispecific antibody such as a BiTE.

9. A modified CDl4 ⁺ cell according to item 8, wherein a T cell-activating

bispecific antibody is an EpCAM BiTE.

10. A modified CDl4 ⁺ cell according to item 7, wherein a chemokine is CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20 , CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12 or CXCL16. 11. A modified CDl4 ⁺ cell according to any of the preceding items, wherein at least two therapeutic agents are expressed.

12. A method for making a modified CDl4 ⁺ cell comprising a nucleic acid

encoding therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated within a tumour in response to tumour-associated micro-environmental signals which consists of:

(a) isolating bloodborne cells from an individual

(b) culturing the isolated bloodborne cells ex vivo and purifying said

bloodborne cells to produce a CDl4 ⁺ enriched cell population

(c) transfecting said CDl4 ⁺ enriched cell population of cells to introduce a vector encoding a therapeutic agent.

13. A method according to item 12 wherein the vector is a viral vector.

14. A method of treating a human by administering the modified CDl4 ⁺ cell

according to any of items 1 to 13.

15. A method of treating cancer by administering a modified CDl4 ⁺ cell according to any of items 1-11 which involves:

a) reinfusing the modified CDl4 ⁺ cells into a cancer patient

b) allowing the modified CDl4 ⁺ cell to engraft within a target site(s)

naturally or encouraging them to do so.

16. A method of treating cancers according to item 15, wherein expression of said anticancer therapeutic agent is activated only when the modified CDl4 ⁺ cell enters the tumour and matures into a MDTIC or TAMDC.

17. A method for treating cancer involving making a modified CDl4 ⁺ cell

comprising a nucleic acid encoding anticancer therapeutic agent or agents wherein the expression of the therapeutic agent(s) is directly or indirectly under transcriptional regulation of a promoter that is selectively activated in MDTICs or TAMDC which consists of:

(a) isolating peripheral blood mononuclear cells (PBMC) from a patient, or a suitable donor, using leukapheresis

(b) culturing the isolated blood mononuclear cells ex vivo and purifying a

subsection of CDl4 ⁺ cells

(c) transfecting the purified subsection of CDl4 ⁺ cells using electroporation to introduce a vector encoding an EpCAMBiTE and a chemokine selected from CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20, CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12 or CXCL16.

(d) introducing the modified CDl4 ⁺ cells into a patient

(e) allowing the engraftment or localisation of the CDl4 ⁺ cells within a tumour(s) naturally or encouraging them to do so by the application of radiotherapy or chemotherapy

Examples

Example 1- cGMP-compliant preparation of CD14 ⁺ cells from patients or volunteers

Leukapheresis of peripheral blood for mononuclear cells (MNCs) was carried out using an Optia apheresis system by sterile collection. A standard collection program for MNC was used, processing 2.5 blood volumes.

Isolation of CDl4 ⁺ cells was carried out using a GMP-compliant functionally closed system (CliniMACS Prodigy system, Miltenyi Biotec). Briefly, the leukapheresis product was sampled for cell count and an aliquot taken for pre-separation flow cytometry. The percentage of monocytes (CDl4 ⁺ ) and absolute cell number were determined, and, if required, the volume was adjusted to meet the required criteria for selection (<20 x 10 ⁹ total white blood cells; <400 ^c 10 ⁶ white blood cells/mL; <3.5 ^c 10 ⁹ CD 14 cells, volume 50-300 mL). CDl4 ⁺ cell isolation and separation was carried out using the CliniMACS Prodigy with CliniMACS CD14 microbeads (medical device class III), TS510 tubing set and LP-14 program. At the end of the process, the selected CDl4 ⁺ positive monocytes were washed in PBS/EDTA buffer (CliniMACS buffer, Miltenyi) containing pharmaceutical grade 0.5% human albumin (Alburex), then re-suspended in TexMACS (or comparator) medium for culture.

The cells are then ready for transfection to introduce the nucleic acid, preferably DNA encoding the therapeutic agent(s). Transfection is performed using electroporation to introduce plasmid encoding the therapeutic agents under appropriate regulatory control, all performed using the CliniMACS-compatible Miltenyi electroporation system according to the manufacturer’s specification.

Modified cells are then maintained ex vivo in the Miltenyi Prodigy system for up to 96 hours before reinfusion to patients following demonstration of release criteria.

Example 2: Isolation of CD14 ⁺ cells from blood cones and transfection with DNA for research purposes

Separately, peripheral blood mononuclear cells (PBMC) are isolated from huffy coats using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation. Primary human CDl4 ⁺ monocytes are then isolated from PBMC using CD 14 beads, typically using 20m1 MACS CD 14 MicroBeads per 107 cells. CDl4 ⁺ cells are then purified using a positive selection MS+/RS+ Column (Miltenyi Biotec) as per manufacturer’s instructions. The plasmids are then introduced into the primary human CD14 cells, by electroporation (using the Miltenyi electroporator or the Amaxa Nucleofector, according to the respective manufacturer’s instructions), and incubated in non-differentiating culture conditions (typically Iscove’s modified Dulbecco’s medium (Lonza, Basel, Switzerland) supplemented with 10% (vol/vol) heat- inactivated human pooled serum, penicillin (100 U/ml; Invitrogen, Carlsbad, CA), streptomycin (100 pg/ml, Invitrogen) and ciproxin (5 pg/ml; Bayer, Leverkusen, Germany) at 37°C in a humidified atmosphere with 5% C02. After 12-120 hour incubation, cells are studied by flow cytometry (Attune NxT, ThermoFisher) and expression of EGFP (488 nm; 530/30) and EBFP (405 nm; 440/50) measured by median fluorescence intensity (MFI).

Example 3. Isolation of CD14 ⁺ cells from tumour biopsies to determine their patterns of gene expression in order to define Candidate Promoters

Fresh primary human breast cancer biopsies (invasive ductal carcinoma, not selected by stratification) are obtained either from local surgical resections or from commercial suppliers. Tissues are maintained in Dulbecco’s Modified Eagles Medium (DMEM) with FBS (10 % v/v) and penicillin/streptomycin (1%) at 4°C and visually checked for viability. Medium is removed, the tissue is weighed and lml of serum-free PBS added. Manipulations are done above a bed of ice to keep the temperature low. The tissue is then diced into tiny pieces using a razor blade, as small as possible to maximise subsequent enzymatic digestion, and the total volume is then made up to 3ml with serum-free PBS. Then is added 30 mΐ of Liberase DL stock solution 28 U/ml (Roche), 120 mΐ of Liberase TL stock solution 14 U/ml (Roche), and 30 mΐ of DNase I (15 mg/ml stock solution Sigma l00x), mixed by vortexing and incubated at 37°C for 45min with continuous rotation. The digestion is checked visually and the sample should appear smooth (if lumpy, digestion time is extended for an additional 15 min but not longer). Digestion is then halted by addition of 10 ml of PBS containing bovine serum albumin (BSA, 2% w/v), and the sample is then filtered through a 100 pm cell strainer and the total volume adjusted to 40 ml with PBS (including 2% w.v BSA) before centrifugation at 500 ref (5 min, 4°C) and the supernatant discarded. The pellet is then resuspended in 3ml of red blood cell lysis buffer (eBioscience), mixed well and incubated on ice for 10 min before addition of 30ml PBS (including 2% w/v/ BSA). The sample is then centrifuged at 500 ref (5 min, 4°C) and the supernatant discarded and pellet resuspended in PBS (including 2% w/v BSA, 900 mΐ). The suspended cells are then subject to purification of CDl4+ cells using CD14 beads (MiltenyiBiotec) to isolate several types of myeloid cells. For this clumps are removed using density gradient centrifugation using Ficoll-Paque and the cells are suspended at a concentration of 107 cells/80pl of buffer and then mixed with 20m1 MACS CD 14 MicroBeads per 107 cells. CDl4+ cells are then purified using a positive selection MS+/RS+ Column (Miltenyi Biotec) as per manufacturer’s instructions.

For comparison, CDl4+ cells are also prepared from normal breast tissue. Mammoplasty reductions are cut into chunks and washed extensively with PBS to remove blood. Samples are then cut into tiny pieces using a razor blade, resuspended in PBS and centrifuged at 500 ref for 5 min at 4°C. The pellet and supernatant fat are then resuspended in serum-free DMEM. To each ml of solution is added 10 mΐ of Liberase DL stock solution 28 U/ml, 20 mΐ of Liberase TL stock solution 14 U/ml, and 10 mΐ of DNase I (15 mg/ml stock solution 100c). Samples are capped with parafilm, vortex mixed and incubated for 10-18h at 37°C with continuous rotation until appearing as a greasy brown broth. The sample is then filtered through a 1 OOmih cell strainer and centrifuged at 500 ref (4°C, 7 min) and the pellet (without the fat) is isolated. The pellet is then resuspended in 3ml of red blood cell lysis buffer (eBioscience), mixed well and incubated on ice for 10 min before addition of 30ml PBS (including 2% w/v/ BSA). The sample is then centrifuged at 500 ref (5 min, 4°C) and the supernatant discarded and pellet resuspended in PBS (including 2% w/v BSA, 900 mΐ). The suspended cells are then subject to purification of CDl4+ cells using CD 14 beads (MiltenyiBiotec) to isolate several types of myeloid cells. For this clumps are removed using density gradient centrifugation using Ficoll-Paque and the cells are suspended at a concentration of 107 cells/80l of buffer and then mixed with 20m1 MACS CD14 MicroBeads per 107 cells. CDl4+ cells are then purified using a positive selection MS+/RS+ Column (Miltenyi Biotec) as per manufacturer’s instructions.

The enriched CDl4+ cell pools are then subject to analysis of expression by RNAseq or Nanostring using established protocols. RNAseq approaches include strategies such as that described in Schuster, S. C. Next-generation sequencing transforms today’s biology. Nature Methods, 5(1), 16-18 (2008). Nanostring analysis is performed using the nCounter® PanCancer Immune Profiling Panel (NanoString Technologies, USA) system in accordance with Nanostring® guidelines. Background thresholding is performed, then data are normalised using the mean of the internal NanoString positive controls, and differential expression determined with reference to uninfected cells.

The resulting data are then analysed, and results from multiple different patients compared with each other and also against the expression profiles of primary circulating human monocytes, in order to identify which genes are differentially expressed within TAMDCs. The promoters regulating transcription of these genes are possible‘Target Promoters’, subsequently tested as described below, to define Candidate Promoters for use in accordance with the invention.

Example 4. Identification of TAMDC promoters by single cell sequencing: An alternative approach to defining Candidate Promoters is to use single cell sequencing (SCS) to identify genes whose expression is increased or decreased within TAMDC, ideally with TAMDC that have derived recently from bloodbome CDl4 ⁺ monocytes. This is achieved by using single cell sequencing (SCS) to study CDl4 ⁺ TAMDC. Accordingly freshly resected primary breast tissues, kept in RNALater, are disaggregated and enzymatically digested, as described above, but instead of CDl4 ⁺ cells being purified, a mixed cell population is used for SCS in order to obtain linked data on expression profiles from many different cells including cells from a variety of lineages. These data are then subject to dimensionality reduction by principle component or t-SNE analysis, and can then be further stratified to identify other expression features and profiles. In this way it is possible to associate the expression of specific genes together within subsets of broader categories of cells, and thereby to infer Candidate Promoters for evaluation as described below. One example of SCS for the identification of expression profiles in immune cells associated with breast cancer is shown by Azizi et ah, Cell, 174, 1293-1308 (August 2018).

Example 5. Identification of Candidate Promoters by single cell sequencing: Data from Azizi et al. is publicly available at

These data were further analysed to allow identification of genes that are differentially expressed within CDl4 ⁺ cells within tumours compared to CDl4 ⁺ cells within normal breast tissue or blood. The Azizi et al. study profiled 45K immune cells from 8 breast carcinomas with single-cell sequencing. Some patients had matched normal tissue, blood or lymph nodes profiled. The publication released three types of data on the GEO database:

• raw expression count data before cell filtering

• raw expression count data after cell filtering (used in this analysis)

• corrected expression count data based on clustering normalisation

The raw expression count data after cell filtering was used for this analysis because of the following two reasons:

1. This set of cells is the result of a filtering process where bad quality cells are removed and this set also represent the presented cell atlas in the paper. The cell filtering process consists of:

a. Molecule size selection, which corrects the surplus of barcoding beads loaded into inDrop per cell

b. Coverage selection, which removes high error barcoded molecules c. Filtering of dead or dying cells based on mitochondrial RNA content d. Low complexity cell infiltration where cells are excluded that detect less genes than expected based on the number of measured molecules

2. The normalised data was not used for this analysis, because that normalisation is optimised to best separate cell clusters based on gene expression, which is the paper’s primary focus. Normalisation happens simultaneously with clustering using the Biscuit algorithm. Since clustering is not the focus, data is normalised from scratch.

The data that was used in this analysis is the GSE1 l4725_rna_raw.csv.gz file from the GEO database at The

underlying comma separated values (csv) file consists of an unnormalized count matrix with 47017 cells as rows and 14846 genes. Additional meta-data for each cell consists of patient ID, type of tissue, replicate number, cluster ID (as presented in Figure 3 of the paper) and cell ID.

Analysis: The analysis scripts were written in R (version 3.5.1). The following pipeline was implemented to derive the differentially expressed genes: (i) The data was loaded into R as one matrix. New row names were assigned to the cells based on patient, tissue, replicate and cell id, to make every cell ID unique and traceable. Meta-data was split from the gene expression data.

(ii) Only cells were selected that originate from samples that have the relevant tissue types for our comparison - For tumour vs normal tissue: BC1, BC2, BC3, and for tumour vs blood: BC1, BC4

Within these samples, cells were selected that showed expression of CD14. Because these data are based on single cell sequencing using unique molecular identifiers (UMI) that allows detection of unique mRNA transcripts at low frequencies, a high sensitivity is expected for this method. Therefore, we select all cells that have any expression of CD 14 as being CDl4 ⁺ cells. These are not many cells because the count data is very sparse. This is due to the inefficiency of mRNA capture (DropSeq is predicted to capture about 10% at best of each cell’s mRNA), which results in the expression matrix being mostly filled with zeros because of these dropout events.

To reduce the likelihood of false positives we first filter on the genes:

a) Genes are removed that do not show any expression counts across the selected cells.

b) Keep only genes that have at least one count in at least 5 cells

c) Only carry on with the top most variable genes. Low variability genes are not expected to be differentially expressed. For tumour vs normal tissue we take the genes with standard deviation of counts above the median of this standard deviation across all remaining genes. Because there seems to be a lot of signal in the tumour vs blood comparison, we take a more conservative gene list here, i.e. the top 1000 most variable genes based on standard deviation of expression.

In this analysis, a common method to do bulk RNA differential expression analysis, called DESeq2, is used with adapted parameters in order to be compatible with single-cell sequencing data This method estimates

gene dispersion (the concept that highly expressed genes, or longer genes, are more likely to be differentially expressed) and applies binomial generalized linear models. The DESeq2 model internally corrects for library size, which means that unnormalized counts can be used as input for the DESeq2 algorithm.

The following parameters were used in the DESeq algorithm, • test="LRT" (the test used for significance testing. Default it is the Wald-test, but in multiple single-cell benchmarks it was observed that likelihood ratio test, LRT, performs better on single cell data).

• sfType- 'poscounts" (base the estimation of size factors only on the non-zero counts)

• parallel = T (the computation was done using multiple cores simultaneously to speed up the calculations)

• useT=TRUE, minReplicatesForR minmu=1 e-6 (as recommended by

the DESeq2 developers).

It was tried to directly model the zero inflation of the counts, and take account of these in the DESeq2 model, using the zinbwave package. However, the limited number of CDl4 ⁺ cells did not allow for the most optimized model training. Besides that, if zero inflation was taken into account, the resulting weights used in DESeq2 did not alter the resulting gene hits significantly. Therefore it was omitted in the final analysis.

DESeq2 has two more features that will affect the final output of the differential expressed genes:

• Fold change shrinking: this process looks at the largest fold changes that are not due to low counts and uses these to inform a prior distribution. So the large fold changes from genes with lots of statistical information are not shrunk, while the imprecise fold changes are shrunk. This allows you to compare all estimated logged FCs across experiments, for example, which is not really feasible without the use of a prior.

• Checking for expression outliers within cells: The DESeq function calculates, for every gene and for every sample, a diagnostic test for outliers called Cook’s distance. Cook’s distance is a measure of how much a single sample is influencing the fitted coefficients for a gene, and a large value of Cook’s distance is intended to indicate an outlier count. The results function automatically flags genes which contain a Cook’s distance above a cut-off for samples which have 3 or more replicates. The p-values and adjusted p-values for these genes are set to NA. The number of cells and genes in each step of the algorithm for both the tumour vs normal and tumour vs blood comparisons are shown in Table 1.

Table 1: Cell and gene numbers for the total analysis

The following measures are reported in the Excel files that list the differentially expressed genes: • log2FC (original): the calculated fold change of gene expression by DESeq

• log2FC (shrink): the shrunk fold change as described on page 4.

• Pvalue: the significance of the LTR test for differential expression

• Padjusted: The P-value corrected for multiple testing using the Benjamin

Hochberg approach

• MeanExpr: the mean raw count of the data for each of the involved tissue types.

• RankExpr is a percentage (0-100%) reporting how the particular gene ranks in mean expression in the set of genes that were considered for the DE analysis (that means after gene filtering, considering only genes that have sufficient abundance and that have sufficient variance)

This analysis yielded these Candidate Promoters, which are upregulated in tumour CDl4 ⁺ cells compared CDl4 ⁺ cells in normal breast tissue: IL10RB, BEST1, FOSL2, MTRNR2L12, CYBA, HSPA1B, HSPA1A. Similarly the following genes are less expressed in tumour-associated CDl4 ⁺ cells compared to those in normal breast tissue, identifying them also as Candidate Promoters: COL1A1, IGF1, EGR1, SLC40A1, USP53, SRGAP2, DUSP6, SESN1, AXL, GPR34, SLC15A3, SLC02B 1, FAM53B, DAB2, EPB41L2, TMEM176B, CHKA, MS4A7, FOLR2, SEPP1, EIF3M, TRIM22, MAF, PLD3, C1QA, WASF2, NR3C1, MBNL1, TXNIP, OGFRL1, TNFRSF1A, FGL2, C1QC, DUSP1, PLXDC2, CIRBP, CSF1R, FOS, DPYSL2, MAFB, C1QB, HLA.E, SEC11A, PABPC4, SRSF5, KLF6, CCNI, HERPUD1, RNF130, MARCKS, NCOA4, SLA, RPS20, FCGRT, ITM2B, RPL22, CXCL16, MS4A6A, JUNB, CEBPD, TPT1, CD68, MT.ND2, RPL5, CST3, CD74, RPS23, HLA.DPB1, RPL37A, RPS11, MALAT1, RPS4X, FTL ,MT.NDl, RPL19, RPL37, RPS14, MT.C03, RPL27A.

This analysis also yielded these possible Candidate Promoters, which are upregulated in tumour CDl4 ⁺ cells compared CDl4 ⁺ cells in blood: MT.ND2, MT.ATP6, RPL3, UBC, RPS7, GNAS, MT.ND3, RPS10, TMBIM6, MT.ND5, HLA.DRB1, HSP90AA1, TXNIP, ZFP36L1, CALR, UBB, MT.ND4, MCL1, CTSB, STK17B, PNRC1 , SRSF5, HNRNPA3, BTG1, HSP90B1, SRSF2, BRD2, LCP2, MT.C03, MZT2B, ITM2B, PREX1, NFKBIA, PTMA, SPINT2, HSPA5, ALOX5AP, CITED2, ATP1A1, RBM3, CSF1R, ARL4C, RGS10, HLA.DRB5, SRSF7, CD63, TUBA1B, PPP1R15A, ADAP2, MS4A7, HLA.DQA1, CEBPD, CANX, KLF2, PDIA3, COX6C, HLA.DQB1, TUBB, PEBP1, MT2A, PIM1, CTSC, CD74, JUND, GLUL, ZFAND5, BTG2, KLRD1, EVL, AP1B1, IL32, CD 163, NUCKS1, DNAJB1, DUSP1, HLA.DRA, KLF4, ARL6IP1, FOS, CD81, MAFB, XBP1, TRAC, SAMSN1, ETS2, ID2, CIRBP, ZFP36L2, HLA.DPA1, IFNGR1, ETS1, SLA, MARCKS, IER3, KLF6, GADD45B, HLA.DPB1, ZFP36, PLIN2, CXCL16, FOSB, CCL5, HSPB1, FKBP5, CPM, HLA.DOA, IGKC, HERPUD1, CSF1, NR4A2, IL7R, DUSP2, TSC22D3, CD83, IL1B, TNFAIP3, GSN, DDIT4, HLA.DQA2, PER1, RHOB, TSC22D1, CXCR4, GPR183, VWA5A, NR4A1, C3AR1, CCL3L3, PDK4, SGK1, JUN, CCL3, MIR24.2, MRC1, CCL4, CD69, EGR1, PLTP, HBB, CLU, SLC18A2, SPARC, LMNA, IGFBP7, LGMN, MAF, AREG, DAB2, RNASE1, C3, CCL4L2, FN1, ATF3, USP53, C1QA, VSIG4, FCGBP, RGS1, MS4A2, LTF, SCGB2A2, KIT, GATA2, LYVE1, CTGF, COL6A3, ADIRF, AQP1, STC2, COL3A1, COL6A1, CXCL12, ACKR1, COL1A2, CP A3, TIMP3, IGFBP4, FABP4, APOE, SLC02B1, C1QC, A2M, SEPP1, DCN, MGP, ClQB,COLlAl and these possible Candidate Promoters that are downregulated in tumour-associated CDl4 ⁺ cells compared to CD14 ⁺ cells in blood: S100A12, S100A8, S100A9, SELL, LILRA2, CSTA, S1PR4, FCN1, VC AN, CFP, VNN2, TKT, MYOlG, CRISPLD2, AGTRAP, LYZ, PLBD1, IRF1, IL17RA, TALDOl, CLEC12A, ANKRD13D, FGR, ICAM3, ANPEP, TNFAIP8L2, CASP1, NCF2, G6PD, SLC7A7, PYCARD, STXBP2, MEGF9, CYP1B1, APOBEC3A, CTSS, PRAM1, CORO 1 A, MNDA, PGD, LTA4H, CCDC69, TMEM176A, POU2F2, CSF3R, SC02, MYOlF, SERPINA1, CD300E, MPEG1, WARS, TSPO, PILRA, LSP1, LILRB3, LIMD2, NFAM1, GLRX, CDC42EP3, CEBPA, LST1, LILRB2, LRRC25, CPPED1, TYMP, CD52, PTPN6, PRKCD, HCK, METTL9, TNFSF10, GMFG, CD48, STAT6, RXRA, SHKBP1, ARAP1, TNFSF13B, SERPINB 1 , S100A6, RAB27A, TNFAIP2, ZNF385A, RHOG, CD37, APOBR, USP15, GIMAP1, IT GAL, TNFRSF1B, RTN3, FMNL1, ZYX, PRKCB, ARHGAP26, FGL2, AP1S2, TNIP1, SASH3, FAM65B, SPI1, IL6R, CYTH4, S100A4, ATG16L2, TMEM176B, SCPEP1, STAT1, BIN2, SH3BGRL3, NINJ1, ITGB2, GIMAP4, RAB32, GAPDH, VASP, GPSM3, ARPC1B, HRH2, CNPY3, PLP2, TYROBP, EVI2B, MX2, EHD1, RAC2, PLEKHOl, PLEC, DOK2, ARHGAP27, JAML, FLNA, PKM, ANP32A, HK1, PLXNB2, PTAFR, FKBP1A, PTK2B, EMP3, EHBP1L1, C140RF2, LGALS1, PVL, BRI3, PRELID 1 , CHMP2A, PRKCSH, STK10, GRN, LGALS9, MYD88, CYBB, LCP1, SKAP2, S100A11, TAPBP, ZNF106, FCER1G, GLIPR2, FPR1, TRIM38, CCND3, NHSL2, IT GAM, S100A10, CD36, TSC22D4, APLP2, OAZ2, LYN, TCEB2, EVI2A, PCSK7, CD44, RPS13, ALDH2, RNF166, GSTP1, EFHD2, DOCK2, CYBA, AIF1, RNF213, SLC43A2, SAMHD1, COTL1, ARHGDIB, CECR1, GRK2, ARRB2, ZEB2, THEMIS2, CHMP4B, PLEK, ACTB, CNN2, HCST, PSAP, CAPG, SERF2, OAZ1, ARHGAP30, PTP4A2, CAPNS1, ZNF207, RPL28, ATP5E, EIF3G, ATP5B, MSN, ARPC3, PSME1, TPP1, LRRFIP1, CD68, ASAH1, ENOl, LAMTOR1, COX7C, EIF3F, UQCRH, WAS, RSRP1, PECAM1, Sep-09, PPP1R18, EIF3E, PFDN5, SH3BGRL, GRB2, ZBTB7A, UQCRB, SLC25A6, NEAT1, EIF3K, CFLAR, C10RF162, RPS11, HLA.C, GNAI2, MYH9, ANXA2, LAPTM5, RPL6, PCBP1, ACTR2, C10ORF54, FXYD5, CAP1, GABARAP, UBA52, RPL23A, FTL, RPL29, CLIC1, RPS14, CCNI, RPL38, PFN1, RPL13A, MT.ND1, ATP5G2, RPL37A, HLA.B, RPS23, MT.COl, RPL23, FTH1, PABPC1, RPS15, RPL10, RPS16, RPS24, RPLP2, HLA.A, TMSB10, RPL32, RPL7A, RPL18, CD14, RPL11, RPL19, RPL13.

Example 6. Histological assessment to confirm Candidate Promoters behave reproducibly between different patients.

Candidate Promoters deduced as described above may not be induced or suppressed reproducibly between different patients and different tumours, and the reproducibility of their activation or inactivity is therefore assessed using tissue microarrays. Breast cancer tissue microarrays are obtained either from local tissue banks or from commercial suppliers. Tissues are then stained histologically to identify tumour-associated myeloid- derived cells (TAMDC) using antibodies recognising CD 163 or CD68, CD206, or any M2-associated marker. Consecutive sections are stained with antibodies recognising the gene products corresponding to the promoters identified above as possible Target Promoters, to determine their expression in TAMDCs. When possible normal breast tissues are also included in the study and their contained myeloid cells are stained similarly to identify whether activation of the Candidate Promoters is cancer-selective. Where suitable antibodies are not available to the gene product, sections are prepared and analysed by RNA scope or FISH (fluorescence in situ hybridisation).

Example 7. Driving differentiation of transfected CD14 ⁺ to assess whether Candidate Promoters are Target Promoters

To determine whether Candidate Promoters are also Target Promoters, Candidate Promoters are cloned into reporter constructs and their possible activation by tumour-like conditions is assessed following transfecting them into CDl4 ⁺ cells isolated from human blood as described above. The sequences of the Candidate Promoters is found by using Biomart to extrapolate 1500 base pairs upstream of the gene transcribed. Candidate Promoters are synthesised (Genewiz) and placed into a custom cassette, pMXl, which is made by Oxford Genetics. pMXl contains EGFP, driven by the candidate promoter and, in the opposite orientation, EBFP driven by a reference promoter normally spleen focus forming virus promoter (SFFV).

Breast cancer and normal tissue cores are embedded in UltraPure low melting-point agarose (4% w/v, Thermo Fisher, UK), and 300 pm tissue slices are prepared using a vibratome (Leica VT 1200S, Leica Microsystems, Germany). Each ex vivo tissue slice is transferred to a 0.6 cm ² PTFE insert (Millipore, UK) in 24-well plates containing 1 mL of cultivation media. After overnight culture, the media are replaced, and tissue slices are used as described below. In addition breast cancer spheroids are formed as described in example 11.

To identify possible Target Promoters, promoter induction or suppression following exposure to tumour-like microenvironmental conditions is measured. Results are analysed by normalising the MFI of EGFP to EBFP for each experiment, to control for transfection efficiency in order to allow comparison of the activity of different promoters in different cell preparations. A basal EGFP/EBFP ratio is defined in CD 14+ cells in non- differentiating medium and the change of EGFP/EBFP ratio between different conditions is then calculated. These conditions include exposure to M2 -polarising media (DMEM containing 10% FBS, 37°C, 5% C0 ₂ , containing IL-4 at 50 ng/ml (Peprotech) and/or IL- 10 at 50 ng/ml (Peprotech), or breast cancer cell-conditioned media, primary malignant peritoneal ascites (isolated directly from the clinic), allowing them to engraft into breast cancer spheroids or allowing them to engraft within primary breast cancer biopsies maintained alive ex vivo. After 1-7 days the transfected CDl4 ⁺ cells are analysed by flow cytometry and their MFI measured. Although CDl4 ⁺ cells are easily recovered from most experimental systems (particularly if low-adherence plates, bags or shaking cultures are used), they are recovered from spheroids and tissues by disaggregation coupled with cell sorting. This identifies the promoter(s) that give the greatest induction or greatest suppression of EGFP expression when monocytes are subject to tumour microenvironmental conditions (by measuring the MFI of transduced cells and controlling for the level of EBFP). The promoters with the greatest level of change, measured as an absolute increase or decrease in the level of EGFP/EBFP ratio in test conditions compared to basal conditions, as well as those with the greatest selectivity (measured as the ratio of EGFP/EBFP in test conditions compared to basal conditions) are then identified as target promoters for use in this invention. Those causing increased expression are‘Upregulated Target Promoters’ and those showing downregulation are ‘Downregulated Target Promoters’

Example 8. Design of synthetic Target Promoters using a library approach.

RNA sequence data from TAMDC (including tumour-associated macrophages), blood monocytes and myeloid cells in normal tissues is analysed to identify genes that are selectively transcribed or silenced within TAMDC. In general this selective expression is due to either chromatin configuration changes or to differential transcriptional activators/repressors. Focusing on the latter, and aiming to identify the relevant transcription factors and repressors, the sequences of the promoters driving the differential expression are analysed for binding sites for known transcription factors and repressors. When promoters are not fully defined, we analyse the 1500 bases of DNA immediately upstream of the transcription start site for the relevant gene. In this way we identify a series of candidate transcription factors and repressors, binding sites for which are then be introduced into a synthetic plasmid expression library. Candidate transcription factors include IRF2, IRF7, IRF9 and STAT2. In this library, the EGFP reporter gene is positioned downstream of minimal CMV promoter itself downstream of a combinatorial library of transcription factor binding sites. The plasmid library is then transfected by electroporation into primary (non-differentiated) human CDl4 ⁺ monocytes purified from PBMC as described above. Cells that immediately express EGP (without differentiation) are isolated for use in defining promoters that are suppressed during differentiation, whereas cells not immediately expressing GFP are isolated for use in defining promoters that are activated during differentiation. The cells are exposed to tumour-like conditions (including exposure to M2-polarising media (DMEM containing 10% FBS, 37°C, 5% C02, containing IL-4 at 50 ng/ml (Peprotech) and/or IL-10 at 50 ng/ml (Peprotech)), or breast cancer cell-conditioned media, primary malignant peritoneal ascites (isolated directly from the clinic), allowing them to engraft into breast cancer spheroids or allowing them to engraft within primary breast cancer biopsies maintained alive ex vivo.) and the cells showing the strongest induction or suppression of EGFP expression (after 1-7 days) are recovered and the promoters sequenced/ recovered by PCR for generation of two refined libraries - one for activation and one for repression. Sequential application of this approach allows reiterative selection of promoter sequences for testing as described below. Example 9. Defining microRNA molecules that show increased expression during differentiation of monocytes towards MDTICs.

Isolation of monocytes and cell culture. Monocytes are obtained from huffy coats from healthy blood donors, all with written informed consent. To do this PBMC are isolated from huffy coats using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation and monocytes isolated either using CD 14 beads (as described above) or by adherence to plastic and cultured in Iscove’s modified Dulbecco’s medium (Lonza) containing 10% [vol/vol] heat-inactivated human serum, penicillin (100 U/ml; Invitrogen, Carlsbad, CA), streptomycin (100 _g/ml, Invitrogen) and ciproxin (5 pg/ml; Bayer) for 5 days in the presence of different cytokines: IL-4 at 50 ng/ml (Peprotech), IL-10 at 50 ng/ml (Peprotech), or medium alone at 37°C in a humidified atmosphere supplemented with 5% C02.

Isolation of TAMDCs: Fresh breast cancer tissues are checked for viability and then disaggregated to allow isolation of TAMDCs by cell sorting using CD 14 beads (MiltenyiBiotec), as described above. cDNA Library preparation and sequencing. A range of methodologies can be used to identify expression of microRNA within monocytes, and how that changes with differentiation. See for example Cobos-Jimenez et al. (Physiol Genomics 46: 91-103, 2014) who produced the data published at EMBL-EBI ArrayExpress public database (http://www.ebi.ac.uk/arrayexpress/) under the accession number EMTAB-1969. In a typical scenario, total RNA is isolated from TAMDCs, PBMC-derived monocytes and 5- day cultured macrophages (unstimulated or cytokine -polarized), using TriPure Isolation Reagent (Roche) according to the manufacturer’s instructions. RNA quality is checked using NanoDrop Spectrophotometer (ThermoFisher). All samples used for sequencing have an A260/280 value >1.8 and good RNA integrity. The small (15-40 nt) RNA fraction is enriched using 15% TBE-urea acrylamide gels, hybridised and ligated for cDNA synthesis. cDNA libraries are purified and after a size-selection step on 10% TBE-urea gels, cDNA samples are amplified using 17 cycles. cDNA libraries are purified (for example using the PureLink PCR Micro kit (Invitrogen)) according to the manufacturer’s instructions. cDNA libraries are pooled together at equimolar ratios and used for the emulsion PCR reaction. The libraries are sequenced to 35 bp read length using Multiplex Fragment Sequencing reagents and the SOLiD 4 Analyzer. Data are then analysed by standard bioinformatics techniques to identify candidate microRNA that show decreased expression as human monocytes differentiate towards immune-suppressive TAMDCs.

Testing of candidate microRNA sites: When a shortlist of candidate microRNA are identified (for example miR-223, miR-454, miR-93, miR-l50, miR-27a, miR-l8lc, miR- 532, miR-l30bp, miR-4772, miR-3l77, miR-l25ap, miR-67l), showing decreased expression as CDl4 ⁺ monocytes differentiate towards TAMDCs, their suitability is assessed by introducing their cognate target sites into non translated regions of reporter EGFP constructs, cloned into expression plasmids. Typically four copies of a microRNA binding site are introduced into the 3’UTR (untranslated region) of the EGFP mRNA. Transcription of the EGFP is typically regulated by the CMV immediate early promoter. Plasmids are grown up in bacteria, purified and then electroporated into PBMC-derived CDl4 ⁺ cells. EGFP expression is then monitored over a period of 1-7 days by flow cytometry, measured as the median fluorescence index (MFI) of transduced cells, and cells are maintained in either medium alone (DMEM containing 10% FBS, 37°C, 5% C0 ₂ ), or medium containing cytokines chosen to promote polarisation towards an immunosuppressive phenotype (IL-4 at 50 ng/ml (Peprotech) and/or IL-10 at 50 ng/ml (Peprotech), or breast cancer cell-conditioned media, primary malignant peritoneal ascites (isolated directly from the clinic), allowing them to engraft into breast cancer spheroids or allowing them to engraft within primary breast cancer biopsies maintained alive ex vivo. MicroRNA sites that show greater EGFP expression when monocytes have been differentiated towards a TAMDCs phenotype are selected for further development. Promising microRNA sites are then introduced into new plasmid constructs in combination with each other, and tested again using the same system, in order to define the most potent combinations.

Example 10: Define molecular choice of EpCAM BiTE.

Five alternative BiTEs are designed, encoding either the OKT3 antiCD3 single chain Fv (scFv) or a nanobody recognising CD3 with different antiEpCAM scFv or nanobody coding regions. The Bites are all encoded to contain a C terminal hexahistidine or FLAG domain. These BiTEs are all encoded into an expression plasmid (OG4728) under transcriptional control of a CMV immediate early promoter. These plasmids are then grown up in E coli and purified using standard protocols. They are then transfected individually into primary monocytes (isolated by CDl4-based purification from PBMC as described above), THP human monocyte cell line and also into 293 cells, using Lipofectamine 2000 and also by electroporation. The amount of BiTE produced by each system is assessed and compared by dot blotting serial dilutions of supernatant (isolated after 24h) and visualising with anti-hexahistidine antibody. The BiTEs are then all applied, at serial dilutions of standardised equal concentrations (measured by dot blotting) to EpCAM -positive human colorectal DLD tumour cells and CHO-EpCAM cells (Shines Hamster Ovary cells engineered to express EpCAM) in the presence of primary T cells isolated from human PBMC. EpCAM-negative cells (parental CHO cells) are used as a control. T-cell activation is measured by staining for surface expression of activation markers (CD69, CD25) and analysed by flow cytometry. To study T-cell proliferation, T-cells are labelled with 5 mM CFSE dye (Thermo Fisher, UK) prior to culturing with target cells. After five days, T-cells are harvested and analysed by flow cytometry. To measure T-cell degranulation, the extemalisation of CD 107a is assessed by adding a CD 107a antibody directly to the well at the start of the experiment. After one -hour incubation, GolgiStop (6 pg/mL, BD Biosciences, USA) is added, followed by flow cytometry analysis after an additional five hours. IL-2 and IFNy quantities are measured using the Human IL-2 ELISA MAX kit (Biolegend, UK) or Human IFNy ELISA MAX kit (Biolegend, UK). A flow cytometric multiplex bead immunoassay is performed using LEGENDplex Th Kit (Biolegend, UK).

To assess cytotoxicity of the BiTEs towards EpCAM -positive target cells, release of LDH into the supernatant (CytoTox 96 Non-Radioactive Cytotoxicity Assay, Promega, USA) or MTS viability assay (CellTiter 96 Cell Proliferation Assay, Promega, USA) are used. To determine viability of specific cell types, total cells are harvested by cell-dissociation buffer, and residual number of viable target cells measured by flow cytometry using an amine-reactive fluorescence live-dead stain and selective antibody staining. For observation of cell viability in real-time, xCELLigence technology (Acea Biosciences, USA) is used. TGF[ and VEGF quantities are measured using TGF beta-l Human/Mouse ELISA Kit (Thermo Fisher, UK) and LEGENDplex Growth Factor Kit (Biolegend, UK), respectively. In this way the ability of each BiTE to mediate T cell cytotoxicity towards EpCAM- positive cells is assessed. By studying activity against EpCAM-negative cells it is also possible to assess their antigen-selectivity. In this way the candidate EpCAM BiTEs are ranked and the most powerful and selective taken forward for further development.

Example 11: Define preferred chemokine capable of attracting optimal T cell subsets

(a) Using a Simple Migration Assay

Chemokines needed to induce T cell chemotaxis are defined through the use of Transwell/semipermeable membrane technologies. 600 mΐ of media containing chemokines (CCL3, CCL3L1, CCL4, CCL4L1, CCL13, CCL14, CCL17, CCL19, CCL20, CCL21, CCL22, CCL28, CXCL4, CXCL9, CXCL10, CXCL11, CXCL12, CXCL16) at concentrations of 300 ng/ml, 100 ng/ml, 30 ng/ml, and 0 ng/ml (media alone is used to determine the spontaneous migration across the membrane) are added to wells in a 24 well receiving plate. PBMC are isolated from whole blood using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation. PMBC are resuspended at 2xl0 ⁶ cell/ml in a chemotaxis buffer of HBSS containing 0.1% BSA , 1 mM CaCl2, and 0.5 mM MgCl2 (all from Sigma-Aldrich). 10 ⁵ PBMCs in 100 mΐ are transferred to a Boyden Chamber (transwell insert, 3 mih pore size , 6.5 mm Transwell® (#3422), Coming®, NY, USA), placed in the wells containing the chemokines or buffer only. The plates are incubated at 37°C for 4 hours to allow the cells to migrate from the insert into the lower well before the chemokine equilibrates. The Boyden Chamber is careful removed and the migrated cells are collected from the wells of the receiving plate, carefully washing the well with buffer to ensure all cells are collected. Supernatants and washes are pooled and 10 mΐ AccuCheck Counting Beads (Thermo Fischer) are added. Cell and beads are spun down and resuspended in 100 mΐ and stained for CD3, CD4, CD8, CD56, CD19, CD14, CD45RA, CCR7. Cells are counted and characterised by flow cytometry. A migration index is calculated by (percentage of chemokine-induced migration)/(percentage of spontaneous migration).

(b) Using an Invasion Assay

Using Cultrex® Cell Invasion Assays (RnD Systems) the ability for chemotaxis through extracellular matrix in response to the chemokines identified in the migration assay is assessed. The defined chemokines are added to the receiving well of a 24 well plate in a concentration gradient from 300 ng/ml, 100 ng/ml, 30 ng/ml, and Ong/ml (media alone is used to determine the spontaneous migration across the membrane). PBMC are isolated from whole blood using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation. PMBC are resuspended at 2xl0 ⁶ cell/ml in a chemotaxis buffer and 10 ⁵ PBMCs, in IOOmI, are transferred to the inserts pre-coated with basement membrane extract (BME), which are placed in the wells containing the chemokines or buffer only. The plates are incubated at 37°C for up to 24 hours to allow the cell to migrate through the BME into the lower well. The insert is careful removed and the migrated cells are collected from the underside of the insert and the wells of the receiving plate, carefully washing the well with buffer to ensure all cells are collected. Supernatants and washes are pooled and 10 mΐ AccuCheck Counting Beads (Thermo Fischer) are added. Cell and beads are spun down and resuspended in 100 mΐ and stained for CD3, CD4, CD8, CD56, CD19, CD14, CD45RA, CCR7. Cell are counted and characterised by flow cytometry. A migration index is calculated by (percentage of chemokine-induced migration)/(percentage of spontaneous migration).

(c) Using a Spheroid Invasion and Cytotoxicity Assay

Spheroid generation: Formation of tumour spheroids is achieved by creating a single cell suspension from adherent tumour lines grown in T25 flasks (DLD human colorectal cancer cells and SKBr3 human breast cancer lines). Prior to dissociation, ensure that the cultures have reached 90% confluence. Wash the monolayers two times with PBS, then add 2 mls of 0.05% trypsin-l mM EDTA and incubate at 37°C until cells detach. The trypsinization process is stop by the addition of 2 mls of complete medium. The cell suspension is mixed gently by pipetting until there are no large clumps visible. Smaller clumps can be dissociated by adding 40 mΐ of a 10 mg/ml DNAse stock and incubated at 37°C for 5 minutes. The cell suspension can then be gently vortexed, and cells pelleted by centrifugation at 200 xG for 5 minutes. The cell pellet is resuspended by gentle pipetting and washing with fresh complete medium, followed by centrifugation at 200 xG for 5 minutes. This is repeated two times. The tumour cells are then counted and adjusted to a concentration of 2 x 10 ⁴ cells/ml. Transfer 4000 cells, 200 mΐ of the cell suspension, to a 96-well round bottom ultra-low attachment (MLA) plates (Corning® Costar®). Incubate the cells for 4 days, or until spheroids of the desired size have formed (usually 300 - 500 pm in diameter). CD14+ cell invasion assay: Label CDl4 ⁺ monocytes with 5 mM CFSE (Thermo Fisher), washing them 5 times using complete medium containing 10% FCS, and using a fresh tube for each wash round to limit excess CSFE transfer into the medium. Resuspend the labelled cells at 10 ⁵ cells/ml.

Very gently remove 100 mΐ/wcll of medium from each well containing spheroids by placing the pipette tip angled against the side of the well, and not at the bottom where the spheroids are located. Gently, pipette 100 mΐ of the labelled CDl4 ⁺ cell suspension into each well containing the spheroids, again by pipetting slowly against the side wall of each well. Incubate the multicell spheroid culture for a further 24 hours. After the 24-hour incubation, as previously described, very carefully aspirate 150 mΐ of the culture medium and replace with PBS, repeat 10 times to ensure the culture medium is fully replaced by PBS, and free or loosely associated labelled CDl4 ⁺ cells are removed too. Remove 100 mΐ of PBS and carefully replace with 8% (w/v) paraformaldehyde in PBS - thus, giving a final concentration of 4% (w/v) paraformaldehyde. Incubate at room temperature for a further 30 minutes. CDl4 ⁺ cell invasion of the spheroids can be assessed by either dissociating the spheroids and FACS analysis, or by freezing in OTC and sectioning the spheroids and analysing by fluorescent microscopy.

T cell Invasion assay: As described above, spheroids are formed and labelled CDl4 ⁺ cells, transfected with plasmids containing the defined chemokines under either a general CMV promoter (to act as a positive control) or using a promoter defined previously, are allowed to invade the spheroids for 24 hours. After which, the spheroids are washed 10 times with complete media to remove free or loosely associated CDl4 ⁺ cells. Then, 100 mΐ of either isolated T cells or PBMC, both at a concentration of 10 ⁵ cells/ml, are added to the spheroids and incubated for a further 24 hours. T cell invasion is then assessed by either dissociating the spheroids and staining the cell suspension for T cell markers (e.g. CD3, CD4, CD8, CD45RA, CCR7, Granzyme B, perforin) and analysed by flow cytometry, or by freezing and sectioning the spheroids, staining for CD3 and analysing by fluorescent microscopy. For a greater, high-dimensional characterisation of the T cells, use a tissue-cytometry approach, such as Chip-Cytometry (ZellKraftWerk), staining T cells for markers such as CD3, CD4, CD8, CD45RA, CCR7, Granzyme B, perforin, CD69, HLA-DR, and CD71. Spheroid T Cell Cytotoxicity Assay: Repeat the T cell invasion assay but use CDl4 ⁺ cells expressing both the defined chemokines and BiTEs (as a negative control use cells not expressing BiTEs). Cytotoxicity is assessed after the T cells have been co-cultured with the spheroids for 24 hours, by measuring the dissociation of the spheroids by microscopy.

(d) Assessment of the Suitability of Attracted T cells for Activation by bispecific T cell activators

T-cell activation is measured by staining for surface expression of activation markers (CD69, CD25, and HLA-DR) and analysed by flow cytometry. To study T-cell proliferation, T-cells are labelled with 5 mM CFSE dye (Thermo Fisher, UK) prior to culturing with target cells. After five days, T-cells are harvested and analysed by flow cytometry. As a surrogate for proliferation in mixed cell populations, total T-cell number per well is determined using precision counting beads (Biolegend, UK). To study T-cell proliferation at early time points, the intracellular presence of Ki67 and phospho-Histone- H3, as markers of active proliferation, are determined by flow cytometry. To measure T- cell degranulation, the extemalisation of CD 107a is assessed by adding a CD 107a antibody directly to the well at the start of the experiment. After one -hour incubation, GolgiStop (6 pg/mL, BD Biosciences, USA) is added, followed by flow cytometry analysis after an additional five hours. Excess CD 107a is removed by washing, and the intracellular IL-2, IFNy and TNF-alpha quantities are determined by formaldehyde fixation and cell membrane permeabilization using Fixation/P ermeabilization solution (BD Biosciences) and intracellular staining of IL-2, IFNy and TNF-alpha. flow cytometer, where population responses are defined by percentage positive for the given markers, and the level of expression of each marker is determined by the geometric mean fluorescence intensity. Precise measurement of cytokine concentrations are measured in the cellular supernatants, where GolgiStop has not been added, using the Human IL-2 ELISA MAX kit (Biolegend, UK) or Human IFNy ELISA MAX kit (Biolegend, UK). A flow cytometric multiplex bead immunoassay is performed using LEGENDplex Th Kit (Biolegend, UK).

(e) Assessing Cytotoxic Activity of the Attracted T Cells Following Activation by bispecific T cell activators To assess target cell cytotoxicity by bispecific T cell activator-activated T cells, target cell lines (DLD human colorectal cancer cells and SKBr3 human breast cancer lines) are prepared with cell-dissociation buffer to preserve cell surface antigens (notably EpCAM). If adherent tumour cells are also required, cell-dissociation buffer is used to detach them from plate surface. Target cells are counted using precision counting beads (Biolegend, UK) and then divided into two populations of 2xl0 ⁶ cell/ml. The first population, true targets, is cultured with the bispecific T cell activators (30-300 ng/mL) for 20 minutes at 37°C and then labelled with 5 mM CFSE (Thermo Fisher). The true target cells are washed 5 times using complete medium containing 10% FCS, using a fresh tube for each wash round to limit excess CSFE transfer to into the medium. The second, control targets, is labelled with 5 pM CTV (Thermo Fisher). Again, these cells are washed five times with complete medium containing 10% FCS, using a fresh tube for each wash round to limit excess CTV transfer to into the medium. These populations are mix together at a 1 : 1 ratio and plated out into a 96 well plate at 2xl0 ⁴ cells/well. The effector T cells are added so effector to true target cell ratios (E:T) of 0: 1 (as a control), 1 : 1, 5: 1, 10: 1, 20: 1, and 40: 1 are achieved. Each set of ratios are repeated in triplicate. The cells are cultured for 4 hours at 37°C. The level of cytotoxicity is determined by dividing the number of remaining true targets cells by the number of control target cells. Additionally, the effector T cell population responsible for the cytotoxicity can be determined and characterised by measuring the levels of target cell membrane transfer acquired during target cell to T cell synapse formation. Target cell positive T cells are characterised by staining for CD3, CD4, CD8, CD45RA, CCR7, CD69, CD25, CD71. Alternatively, release of LDH into the supernatant (CytoTox 96 Non-Radioactive Cytotoxicity Assay, Promega, USA) or MTS viability assay (CellTiter 96 Cell Proliferation Assay, Promega, USA) is used. For observation of cell viability in real-time, xCELLigence technology (Acea Biosciences, USA) is used. Where appropriate, CD3/CD28 Dynabeads (Thermo Fisher, UK) are included as positive controls for T-cell activation. T-cells are subsequently recovered (after 12-120 h) by pooling the culture media with a PBS wash and stained for surface and intracellular markers as described below.

Using T cells collected in the migration assay, their ability to be activated by BiTEs and kill EpCAM positive target cells is assessed, thereby defining the chemokine/s able to induce the migration of T cells with the greatest BiTE -mediated cytotoxicity. (f) Assessing Subtypes of the Attracted T Cells by Flow Cytometry

To classify different cellular populations, antibodies specific for CDl lb (ICRF44), EpCAM (9C4), CD3 (HIT3a), CD4 (OKT4), CD8a (HIT8) are used. To analyse T-cell populations, the following antigens are used: CD69 (FN50), CD25 (BC96), IFNy (4S.B3), IL-2(MQl-l7Hl2), TNFa(MAbl), CD 107a (H4A3), PD1 (H4A3), CD45RA(HIl00), CD45RO(UCHLl), CCR7(G043H7), Granzyme B (GB11), Perforin(B-D48), FASL(NOK-l), TRAIL(RIK-2), Ki67(l 1F6), Histone-pH3(l 1D8), HLA-DR(L243). The appropriate isotype control antibody is used in each case combined with fluorescence minus one staining. All antibodies are acquired from Biolegend (UK) unless stated otherwise. Surface markers are first stained for prior to cytoplasmic and nuclear markers. For intracellular staining (ICS), cells are first fixed and permeabilized using Fixation/P ermeabilization solution (BD Biosciences). Analysis is performed on a Attune flow cytometer (ThermoFisher, Waltham USA) and data processed with FlowJo vl0.0.7r2 software (TreeStar Inc., USA). In this way the types of T cells attracted by the different chemokines can be defined.

Example 12. Define optimal radiation protocol to enhance infiltration of therapeutic monocytes and other immune cells into tumours

Nod SCID Gamma (nsg) mice are mutant mice strain that do not express Prkdc nor the X linked Il2rg genes. These mice strain permit the engraftment of primary human immune cells. They are obtained from Jackson Labs via Charles River.. The study was approved by the Animal Welfare and Ethical Review Body and undertaken in accordance with the UK Home Office guidelines. Mice are bred in house in a temperature-controlled facility, and supplied with food and access to water. They are fed a standard diet until 8 weeks. Following this, Nod SCID Gamma (NSG) mice are implanted subcutaneously with 2xl0 ⁶ SKBr3 human EpCAM+ breast cancer cells (mixed with Matrigel), and tumours allowed to develop to a size of 5-8 mm diameter. This normally takes about three weeks. The developing tumours are then exposed to external beam radiation (at single doses ranging from 5 Gy to 21 Gy) using a Caesium source with lead shielding or a SAARP small animal radiotherapy device, and animals are left to recover. CFSE-labelled human CDl4 ⁺ monocytes (purified from PBMC as described above) are administered intravenously or intraperitoneally between four and seven days after the radiation. To minimise clogging of the pulmonary beds, following intravenous delivery, cells are resuspended thoroughly to prevent formation of clumps. Animals are sacrificed and tumours recovered five-ten days following radiation exposure, and immediately cut into portions and snap frozen in preparation for fluorescence microscopy to assess the frequency of CFSE-labelled cells that have engrafted into tumour tissue. In addition some tumour samples are stored in RNAlater to allow quantification of the infiltrating monocytes by Q-RT-PCR against monocyte genes. When the human monocytes are male, infiltration is also studied by QPCR measuring DNA sequences such as the sex-determining region, SRY, contained only in the Y chromosome using these primers: forward primer: 5'- AGTTTCGC ATT CT GGGA TTCTCT-3', reverse primer: 5 '-GCGACCCATGAACGC A TT-3'.

Example 13. Demonstrate therapeutic mechanism and activity

Nod SCID Gamma (NSG) mice are implanted subcutaneously with 2xl0 ⁶ SKBr3 human EpCAM+ breast cancer cells (mixed with Matrigel), and tumours allowed to develop to a size of 5-8 mm diameter. This normally takes about three weeks. Tumours are then exposed to radiotherapy, typically 5-21 Gy in a single administration, as optimised above.

Separately, PBMCs are isolated by density gradient centrifugation (Boyum, 1968) from whole blood leucocyte cones obtained from the NHS Blood and Transplant UK (Oxford, UK). Blood is diluted 1 :2 with PBS and layered onto Ficoll (1,079 g/ml, Ficoll-Paque Plus, GE Healthcare) before centrifugation at 400 g for 30 min at 22°C with low deceleration. After centrifugation, PBMCs are collected and washed twice with PBS (300 g for 10 min at room temperature) and resuspended in RPMI-1640 medium supplemented with 10% FBS. CD3 -positive T cells are then purified from PBMCs by depleting non-CD3 cells using Pan T Cell Isolation Kit (Miltenyi Biotec, #130-096-535), according to the manufacturer's protocol. Either whole PBMC (2xl0 ⁷ cells/mouse) or purified T cells (lxlO ⁷ cells/mouse) are then administered i.v. to tumour-bearing NSG mice, immediately following radiotherapy and 5-7 days before administration of genetically modified monocytes.

Human peripheral blood mononuclear cells (PBMC) are isolated from huffy coats using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation. Primary human CDl4 ⁺ monocytes are then isolated from PBMC using CD 14 beads, typically using 20m1 MACS CD14 MicroBeads per 10 ⁷ cells. CDl4 ⁺ cells are then purified using a positive selection MS+/RS+ Column (Miltenyi Biotec) as per manufacturer’s instructions. Plasmids encoding our therapeutic agents (EpCAM BiTE plus T cell chemokine, both with signal peptides to ensure secretion) are then introduced into the primary human CD 14 cells, by electroporation (using the Miltenyi electroporator or the Amaxa Nucleofector, according to the respective manufacturer’s instructions), and incubated in non-differentiating culture conditions (typically Iscove’s modified Dulbecco’s medium (Lonza, Basel, Switzerland) supplemented with 10% (vol/vol) heat- inactivated human pooled serum, penicillin (100 U/ml; Invitrogen, Carlsbad, CA), streptomycin (100 mg/ml, Invitrogen) at 37°C in a humidified atmosphere with 5% C0 ₂ . After 3-6 hours, the transduced CDl4 ⁺ cells are administered i.v. to mice at doses ranging from 10 ⁵ to 10 ⁷ cells/mouse. Genetically modified monocytes are normally administered to mice approximately one week following radiotherapy and delivery of PBMC/T cells.

Activation of genetically modified T cells within the tumour is measured by expression of therapeutic mRNA (determined by QRT-PCR or RNAscope or Nanostring) and production of therapeutic proteins (EpCAM BiTE and chemokine), leading to T cell activation (measured by expression of CD25/CD69/CDl07a) and production of gamma interferon, leading in turn to tumour cell killing and therapeutic outcome demonstrated by shrinkage of the SKBr3 tumours.

Example 14 Isolation of CD14+ cells from blood cones and the transduction process of these cells to contain therapeutic DNA

CD 14 cells were isolated from whole blood collected from healthy volunteers or blood cones provided by NHS blood and transport. In short, PBMC were isolated from blood through centrifugation using Ficoll gradients and subsequently CD 14+ monocytes were extracted using positive selection by magnetic beads (CD 14+ positive selection using the Miltenyi Biotec Bead system) (Figure 6). Monocyte isolation by such processes captures the full range of blood monocyte subsets (classical, intermediate and non-classical) albeit in ranging quantities as would be expected; i.e. that the blood monocyte pool largely consists of classical monocytes (CDl4 ^h V CD16 ¹⁰ ), whilst rarer subsets such as Tie2+ positive monocyte numbers are quite negligible. Accordingly, a broad swathe of blood monocytes and not only specific subsets are able to be transduced. We identified a number of viral and non-viral methodologies by which DNA could be introduced into monocytes for therapeutic purposes (Figure 7). We demonstrated that electroporation of blood monocytes either unstimulated (“M0”), or stimulated with pro- inflammatory (“Ml”) or anti-inflammatory (“M2”) cytokines yielded significant numbers of viable, GFP positive cells, although this approach was most successful for relatively small plasmids. Alternatively, transduction with adenovirus and lentivirus yielded larger numbers of GFP positive, viable monocytes.

Example 15 Identifying target promoters

Several candidate promoters can be identified from the published literature, for example Cassetta et al., and Table 2 shows a reanalysis of published literature to identify some of the genes showing strongest and most selective expression within breast tumour macrophages compared to circulating monocytes from cancer patients. These genes are driven by‘candidate promoters’. The column headed‘Induction’ shows the selectivity for expression within TAMs compared to monocytes, and the column headed‘Level’ shows the relative mRNA abundance by RNAseq, both calculated from the data of Cassetta et al. We have shown that some of these candidate promoters, but not all, can be induced to mediate powerful increased expression of their cognate genes within primary CD 14 cells following exposure to tumour-like microenvironments (using either tumour cell-conditioned media or ascites samples) and therefore constitute ‘target promoters’ that could be suitable for use in performance of the invention. Other candidate promoters are not activated within primary CD 14 cells under these conditions, and are not suitable for use as a target promoter in accordance with the invention. In Table 2 the column headed‘Mono’ indicates the relative abundance of mRNA found by RNAseq in undifferentiated primary CD14 monocytes and the columns headed‘CCM’ and‘ascites’ show the relative abundance of the cognate mRNA in primary CD 14 cells that had been treated with CCM or malignant ascites, respectively, and allowed to differentiate and polarise. Among target promoters, those that are preferred show high selectivity for expression within TAMs compared to naive monocytes combined with good potency. In the analysis below we have introduced an expression threshold of 1000 (calculated as the average of expression within CD14 cells exposed to CCM and ascites). Those promoters that emerge as‘target promoters’ are indicated in the final column. It is notable that several of the target promoters are natively associated with secreted proteins, suggesting that such promoters may often combine selectivity with potency.

Preferred target promoters are those that can be captured effectively within a DNA sequence suitable for incorporation into a therapeutic vector, typically 2kb or less in length.

Table 2. Identification of target promoters from candidate promoters that are expressed in tumour macrophages but not in monocytes from people with cancer

t

Alternatively target promoters may be identified from a list of candidate promoters that are known to be expressed in tumour-associated macrophages but not within resident macrophages in normal tissue. A list of such candidate promoters, identified from data published by Cassetta et al. (Cancer Cell, 35, 588-602. elO (2019)), is shown in Table 3. This table uses the same headings as Table 2 and the data are obtained and analysed in the same way. Once again it is possible to use the levels and selectivity of expression reported in the literature combined with RNAseq analysis of gene activity within primary CD 14 cells exposed to CCM and ascites to determine which of these candidate promoters may be useful as target promoters, being those with substantial and selective levels of expression within primary DC 14 monocytes exposed to tumour micro-environmental conditions. Preferred target promoters based on these data include CCL3, FN1, SPP1, CCL8, and SIGLECl.

Preferred target promoters are those that can be captured effectively within a DNA sequence suitable for incorporation into a therapeutic vector, typically 2kb or less in length.

Example 16 Driving differentiation of transduced CD14 ⁺ to identify Target

Promoters Our concept aims to generate therapeutic monocytes, which upon reaching the tumour and undergoing the processes of differentiation and polarisation into tumour associated macrophages, will produce therapeutic agents locally in solid tumours.

To be able to screen candidate promoters that would meet the criteria required for our concept, we needed to ensure that our methodology was capable of inducing both differentiation of monocytes into macrophages, and their polarisation into a TAM phenotype. Aligned with the intended application use of our technology, we aimed to infect our monocytes quickly following isolation from blood and ensure sufficient time and conditions for differentiation and polarisation into tumour associated macrophages, mimicking the clinical scenario where monocytes would infiltrate a tumour and begin to differentiate and polarise towards tumour associated macrophages

Culturing monocytes in the presence of human serum drives the monocyte-to- macrophage differentiation process across a 3-6 day period, differentiating monocytes into macrophages (Figure 8 A and 9A). Further stimulation with defined cytokines/ culture medias during this period can further modify these monocytes to produce macrophages of defined polarisation states. Most defined are the archetypal pro- and anti inflammatory macrophage phenotypes, denoted Ml and M2 respectively, which we are able to generate across the culture period. As general anti-inflammatory cells, M2 macrophages are considered to share a broadly similar phenotype to that of Tumour- associated macrophages. However, the generation of M2 macrophages is based on the addition of one or two“M2” cytokines into culture media (normally IL-4 + IL-10) and cannot therefore truly encapsulate the full range of factors present in the tumour milieu that drive tumour associated macrophage (TAM) generation. To engineer more relevant culturing model systems for TAMs, we chose to stimulate our macrophages with supernatants from cancer cell lines (denoted“CCM”) or with peritoneal ascitic fluid from patients with advanced cancer (Denoted“Asc”). As demonstrated (Figure 8B), such culturing conditions yield macrophages with polarisation states similar to M2 as evidenced by prototypic M2 marker (CD 163 and CD206) expression.

When performing candidate promoter screening we chose to infect monocytes within a day of isolation to ensure that infection was performed at the monocyte state (Figure 9 A), aligning with our proposed therapeutic strategy of using monocytes, not macrophage, as the therapeutic product. In providing polarisation conditions at day 6 and assessing GFP expression at days 9/10 we are identifying promoters that are activated in the differentiated macrophage form, which have been stimulated by TAM- or TAM-like polarisation conditions.

To confirm the validity of our screening method, we first performed a simplistic screen in the archetypal Ml and M2 macrophage polarisation states using a number of promoters (Figure 9B). Here we showed high expression of TNF promoter driven GFP in Ml conditions and high expression of Mannose receptor driven GFP in M2 conditions confirmed that our promoter strategy could enable expression of GFP in selective polarisation states, and that our method of infecting monocytes with viruses containing said promoters validated that we have a process by which we can capture promoters that are microenvironmentally induced following differentiation into macrophages and polarisation into defined phenotypes. We could further show that promoter-driven expression by such methods is durable (Figure 9E)

Whilst our screen showed a number of promoters having high activity in M2 polarised cells, M2 macrophages do not fully recapitulate the tumour-associated macrophage phenotype. As described, we sought to better generate tumour-associated macrophage profiles by culturing in the presence of conditioned media from cancer cell lines (CCM). We performed screening of a number of promoters in CCM and compared their selection induction and intensity of expression against M2 polarisation (Figure 9Ciii + 9D). For a number of promoters, high activity in M2 conditions does not correlate with increased expression in CCM such as is the case of CD200R. Conversely, a number of promoters appear relatively weak in M2 but show significantly higher activation when in CCMs such as FN1, CCL3 and CCL8. Such cases could not have been identified by M2 based screening alone and are only discoverable through further screening in more tumour relevant conditions as described. Further, by screening in a number of breast cancer CCMs (Figure 9Ciii) and CCMs from cancers of a range of tissue types (Figure 9D), we can gain insights into tissue-independent, tumour-specific promoters.

To further examine whether our CCM screening constitutes a represent model system in which to test potential promoters, we are able to test candidate promoters in sections of human tumour tissue obtained from patients with cancer through ex vivo culture. In one such representative example, we demonstrated that patterns observed in CCM screening held true ex vivo; CD200R activity in both tumour biopsies and M0 conditions was lower than M2, while FN1 was higher than M0 in both biopsies and M2-polarising conditions (Figure 9F).

Example 17 Defining microRNA molecules that show increased expression during differentiation and polarisation of monocytes into MDTICs

MicroRNA sites that are useful in this invention include sites that are recognised by microRNA molecules that are differentially regulated between tissue macrophages and tumour-associated macrophages, most notably those which are expressed in MO or Ml macrophages but lost in M2. It is recognised that this terminology is insufficient to describe the complex biology of macrophage polarisation, but the terms are used herein to indicate extreme polarisation states, with Ml macrophages showing strong proinflammatory properties and M2 macrophages showing more immunosuppressive and wound healing-associated properties typical of tumour-associated macrophages. Hence microRNA sites that are regulated during the process of macrophage polarisation, a feature of their physiological plasticity, may be useful in the practice of this invention. Figure 10 shows microRNA that are differentially regulated between Ml polarised macrophages and macrophages allowed to polarise following incubation in cancer cell- conditioned supernatant (CCM), using MicroRNAseq analysis.

Using microRNAseq to compare microRNA molecules that are expressed in macrophages polarised to Ml or not polarised (M0) compared to those polarised in CCM or ascites allows several microRNAs to be identified that may be exploited in performing the inventive concept, by incorporating their cognate microRNA binding sites into the transgene (or mediator) nucleic acid. In this way microRNA binding sites for the following microRNAs (Table 4) are particularly relevant in the practice of the invention by incorporating them into the mRNA of genes expression of which is desirable within tumour-associated macrophages, typically as one or more microRNA sites incorporated into the 3’ UTR.

We have taken several of these microRNA sites and incorporated them into the 3’ UTR of various transgenes including GFP reporter gene and therapeutic proteins. The data in Figure 10 shows the effects of microRNA sites incorporated into the 3’UTR of a GFP gene under transcriptional control of the SFFV promoter, following transduction of CD 14 cells and their exposure to differentiation- and polarisation-inducing cytokines to drive them towards Ml or M2 phenotypes.

It can be seen that many of the microRNA sites are capable of selectively destabilising reporter gene expression in M0 or Ml polarised monocyte/macrophages, verifying their utility in the practice of the invention - see Figure 11B.

CD 14 cells infected with lentiviruses encoding GFP reporter under transcriptional control of the SFFV promoter and with microRNA sites within their 3’UTR were exposed to cancer cell conditioned media prepared using a range of human breast cancer cell lines. Most cancer cell-conditioned media were capable of inducing high levels of transgene expression, usually higher than that seen even in M2 -polarising conditions (which is known to be usually considerably higher than the level in M0 or Ml conditions). Hence it follows that when exposed to these cancer cell conditioned media the macrophages are downregulating expression of the relevant microRNAs and thereby stabilising expression of the transgene product - see Figure 11D.

Similarly CD 14 cells infected with lentiviruses encoding GFP reporter under transcriptional control of the SFFV promoter and with microRNA sites within their 3’UTR were exposed to cancer cell conditioned media prepared using a range of human cancer cell lines that are not breast cancer. Again it can be seen that virtually all of the conditioned media induced high levels of GFP expression, indicating selective loss of the cognate microRNA leading to stabilisation of the transgene product. See Figure 11E.

Genes whose expression may be usefully regulated in this way may encode therapeutic proteins, for secretion into the tumour microenvironment, or may encode molecules (such as proteins or microRNAs) that can positively regulate the expression (directly or indirectly) of other genes whose expression is desirable. Such therapeutic proteins include antibodies, notably bispecific antibodies based on pairs of scFv or nanobodies, or regulatory proteins such as VP 16-TETR that can bind to TETO sites within the regulatory region of other desired genes and positively influence their expression. In a preferred embodiment, sites for microRNAs suitable for use in accordance with the invention are present within the mRNA encoding a regulatory molecule (such as VP 16-TETR) under promoter control of one of the promoters useful in the practice of the invention and also within the mRNA encoding a therapeutic gene under control of a promoter construct that contains TETO (tet-operator) binding sites. In this way microRNA can be used to degrade mRNA within unpolarised macrophages and thereby to give selective stabilisation following cell polarisation of both the mRNA encoding VP 16-TETR and also the mRNA encoding the therapeutic gene that is induced following binding of the VP 16-TETR protein to TETO sites.

Example 18 Therapeutic application of genetically-modified monocytes for cancer therapy

For studies using lentivirus, 4e6 CD 14 monocytes isolated from a single healthy human donor were treated with VPX virus-like particles and infected in a 6 well plate with lentiviral vectors containing a luciferase or GFP transgene driven by various promoters. Cells were then spinoculated for 2 hours at 33°C l500rpm. Following inoculation cells were incubated at 37°C in a C02 -controlled incubator overmight. The following day cells were isolated by washing with PBS 5mM EDTA, to detach them and then collected by gentle centrifugation and counted. For studies using adenovirus, CD 14 monocytes isolated from a single healthy donor were infected with a type 5 adenovirus vector expressing luciferase under control of a CMV promoter, and cells were incubated overnight prior to use.

To assess migratory capacity, the ability of lentivirus -transduced monocytes to migrate towards various tumour/ cancer models was assessed to mimic our proposed mode of treatment. Using monocytes transduced with a GFP expressing virus (with GFP driven by the viral SFFV promoter), a transwell migration assay was performed to confirm that transduced monocytes retain cancer-homing properties. Using culture media from a range of different cancer cell lines, transduced monocytes demonstrated increased migration towards cancer cell line media as compared to media baseline (Figure 12A).

To better recapitulate the tumour, we extended this model to include migration toward biopsies of human breast cancer samples. Lentivirus transduced monocytes with SFFV promoter-driven GFP were able to migrate towards the tumour biopsy, demonstrating that tumour-homing functionality of these monocytes is retained post-transduction (Figure 12B).

To mimic intravenous injection of our therapeutic cells, in vivo studies were performed using a breast cancer xenograft model where human MDA-468 breast cancer cells were injected subcutaneously in NSG mice. Tumour cells were inoculated subcutaneously on the flank and tumours were allowed to reach a size of l00mm3 prior to commencement of the experiment. Adenovirus-infected CD 14 monocytes were then administered intravenously and after 48 hours, mice were imaged by IVIS imaging which demonstrated substantial accumulation of therapeutic monocytes at the tumour site (Figure 12C).

To examine the in vivo biodistribution and transgene expression properties of monocytes containing luciferase regulated by candidate promoters, tumour-bearing mice were injected intravenously with lentivirus-transduced monocytes containing candidate promoters. On days 2, 5, 6, 7 and 12 post injection mice were injected with luciferin and imaged for light emission on an IVIS camera (Figure 12D and E). No Luciferase was detected at any point prior to day 6, reflecting the time taken for lentivirus to infect primary monocytes. ROI was determined by drawing a circle around the area of the tumour and backgrounds were determined by copying the ROI circle to another part of the mouse on the same flank. The luminescence associated with the snout and feet is seen routinely, including in the control animal which received no treatment, is thought to reflect background light emission of components of the animal feed.

It is notable that only SFFV, IL6 and CD200R showed light emission from the region of the liver, suggesting some off-tumour expression from those promoters in that organ. In contrast FN1 (fibronectin 1) and CD200R also showed clear signals from the tumour (Figure 12D and E). The control mouse showed a very low level of expression in the region of the liver and intestines, thought to reflect light emission properties of the animal feed. Light emission was quantified and is shown in the histogram.

These data indicate that FN1 and CD200R can be used to regulate expression of transgenes within the tumour, and that this can be successful following intravenous administration of genetically modified human monocytes to tumour-bearing animals.

Example 19 Choice of therapeutic transgene products and demonstration of utility

For successful immunotherapy of immunologically‘cold’ tumours by local expression of a bispecific T cell activator it will likely be beneficial to also express chemokines in order to encourage infiltration of more T cells into tumours. Primary human T cells were isolated from blood cones and then incubated in a transwell experiment, with chemokine- containing medium on the other side of a semipermeable membrane. Several different chemokines were compared for their ability to attract CD8 T cells in vitro. Figure l3A(i) shows the effect of chemokines on unstimulated T cells and the chemokines CCL3, CCL3L1, CCL17 and CXCL9 were most effective. When T cells were pre-activated with CD3/CD28 beads, (Figure l3A(ii) CXCL10 and CXCL11 were most effective.

To ensure that genetically expressed chemokines retained the same chemoattractant properties as the commercial proteins, CXCL10 and CCL3L1 were encoded within plasmids and transfected into 293 cells. The plasmid construction is shown in Figure 13B(i), and the results of a PBMC migration assay are shown in Figure 13B(ii), where commercial chemokines or supernatants from cells transfected with plasmid pMX65 (encodes both chemokines) were evaluated for their chemoattractant activity in a transwell assay. The supernatants showed clear chemoattractant activity after both 48h and 72h, indicating that one or more chemokines did indeed retain their activity when expressed from exogenous DNA. The ability of genetically-modified primary human CD 14 monocytes to produce therapeutic agents was assessed and compared with 293T cells. Figure 13Biii shows that, following transduction with a lentiviral vector containing CXCL10 and CCL3L1 under control of a SFFV promoter, both cell types are capable of producing significant levels of the therapeutic chemokines. The ability of human monocytes to produce encoded gamma interferon was also assessed following their transduction with a lentiviral vectors driven by the SFFV promoter. Figure 13C shows monocytes from three independent donors all produce and secrete high levels of gamma interferon.

The ability of primary human CD 14 cells to produce gamma interferon under control of tumour-inducible promoters was also assessed. As seen in Figure 13D, the promoters MRC1 and VEGFA both showed greater expression of gamma interferon when the macrophages were exposed to M2-inducing cytokines, cancer cell-conditioned media or human peritoneal malignant ascites, compared with unpolarised macrophages.

The ability of genetically-expressed EpCAM bispecific T cell activator (produced in 293T cells) to activate T cells was assessed in the presence of EpCAM positive cancer cell lines. T cells only become activated by clustering of CD3, so unless the bispecific molecule has poor solubility properties, T cell activation only occurs following formation of a pseudosynapse with EpCAM -positive target cells. T cell activation was demonstrated by flow cytometry using the T cell activation markers CD25 and CD69 (Figure 13E first image). As shown in the analysis (Figure 13E later images) there is some slight activation of CD4 cells, but the main effect is activation of CD8 cells to express CD25 as expected. Genetically-expressed EpCAM bispecific T cell activator (produced in 293T cells and using the OKT3 antiCD3 component) also induces T cell degranulation in the presence of EpCAM positive breast cancer cell lines but not without EpCAM -positive target cells, as shown in Figure 13F. Isolated primary CD3+ T cells were cultured in the presence of the SKBR3 cancer cell lines (either 0.5e5 cells or le5 cells) ± the EpCAM bispecific T cell activator at two concentrations of the 293A supernatant (1/1000 and 1/500).

Genetically-expressed EpCAM bispecific T cell activator (produced in 293T cells) can also mediate T cell activation and degranulation when co-cultured with the non-breast cancer cell line SKOV3 (Figure l3Gi), showing the utility of this therapeutic strategy is not restricted to breast cancer.

EpCAM bispecific T cell activator expression, produced by lentivirus-mediated expression in M2 macrophages, can induce T cell activation and cytotoxicity when co cultured with the breast cancer cell line MCF7 (Figure l3Gii). This shows that polarised macrophages are capable of producing and secreting functional quantities of biological active therapeutic proteins.

Regulation of expression can be further enhanced through the use of miRNA sites and can complement the use of target promoters, to achieve tumour-specific expression of a therapeutic agent. As shown in Figure 13H, a promoter (such as CD200RI) can be used to achieve expression of a therapeutic agent (in this case gamma interferon) from a therapeutic monocyte upon differentiation and polarisation into a macrophage or a tumour-associated macrophage.

It can be seen that the selectivity profile for therapeutic expression driven by promoters in therapeutic monocytes can be further enhanced by combination with appropriate microRNA biding sites (Figure 13H). Using three copies of the miR-l55 (5p) site in combination with CD200R1 promoter further enhanced the selectivity of expression by reducing production of therapeutic protein in macrophages of the incorrect polarisation states (M0 and Ml in this case), whilst retaining expression in M2. This further demonstrates that target promoters can be usefully combined with polarisation-induced microRNA sites, in accordance with the invention.

A similar combination of target promoter and polarisation-induced microRNA sites were used to regulate expression of a bispecific T cell activator protein. In this case primary CD 14 human monocytes were infected with lentiviruses expressing the antiEpCAM- OKT3 bispecific T cell activator under transcriptional regulation of the MRC1 promoter and containing three binding sites for microRNA 155 (5p) in its 3’ UTR. The cells were differentiated in human serum and then polarisation cytokines were added to drive them towards Ml or M2 phenotype. Supernatant was harvested from the cells. Separately, 20,000 EpCAM -positive SKOV3 human ovarian carcinoma cells were added to wells of an E-plate 96 (96 well plate with gold electrodes on the bottom) and cell growth was monitored by conductivity using an Xcelligence system (Acea Biosystems). Conditioned medium from the infected macrophages (1 in 50 dilution) and primary T cells (100,000/well) were added and the effects on cell growth and survival was monitored in real time by conductivity (Figure 131).

As shown in Figure 131, supernatant from the Ml -polarised macrophages mediated no cytotoxicity towards the SKOV tumour cells, with the conductivity profile similar to controls. In contrast the supernatant from M2 macrophages showed very potent cytotoxicity. This suggests combining a target promoter and polarisation-inducible microRNA sites can give selective activation of transgene expression within M2 polarised cells in accord with the practice of this invention.

Example 20 Use of tet-off regulated system to amplify transgene expression and demonstration of utility using an adenovirus vector

Drug-controlled switches such as tet-off can be used to allow amplification of expression from relatively weak promoters, and also to enable expression of therapeutic agents to be turned off following drug administration. This could provide a useful strategy to control any unwanted toxicities observed in patients.

Plasmids were constructed containing candidate promoters (IL1B, VEGFA) driving expression of the TETR-VP16 fusion protein transactivator alongside a minimal CMV promoter with upstream three TET-operator sites driving expression of GFP (Figure 14A). When the candidate promoter is expressed the TETR-VP16 is produced, and binds to the TETO sites, whereupon the VP 16 component should activate expression from the minimal CMV promoter. Use of a relatively weak candidate promoter can in this way be amplified to give relatively strong levels of transgene expression (in this case using GFP as a surrogate transgene). Figure 14A shows that the system is successful following transfection of the plasmid into 293 cells (where both VEGF and IL6 promoters are only weakly active) with gearing of up to 5 -fold achieved using the candidate promoters. In contrast the strong CMV promoter showed no enhancement, as expected, because it already operates at high level of expression.

This system can be used to regulate expression of transgenes within a therapeutic vector. Figure 14B (i) shows an adenovirus vector containing (from left to right) a CMV enhancer and CMV promoter driving expression of TETR-VP16, followed by eight TETO sites upstream of a minimal CMV promoter driving expression of an EpCAM bispecific T cell activator with a T2A separating the EpCAM bispecific T cell activator coding region from a CCL31 gene coding region and then P2A site separating from a CXCL10 gene coding region, following by a SFFV promoter driving expression of CD 19. When the CMV promoter/enhancer (which here is a surrogate for a candidate or target promoter) is active it will produce TETR-VP16 that should transactivate the minimal CMV promoters to drive expression of the EPCAM bispecific T cell activator, CCL3L1 and CXCL10 (which will all be expressed as individual proteins due to the presence of the T2A and P2A ribosome-skip sites). Addition of doxycycline (which will bind and inactivate the transactivation properties of the TETR-VP16) will turn this expression off. It can be seen that when the virus is used to infect 293T cells (Figure l4B(ii)) it produces a strong bispecific T cell activator signal (detected by western blot using an anti-his antibody, to detect the his tag present in the bispecific T cell activator), and this expression is silenced by the addition of doxycycline as expected. Similarly the virus produces a strong CXCL10 signal (detected by ELISA) (Figure l4B(iii)) that can be silenced using doxycycline as expected. In this way it is clear that the virus is capable of using a gearing system that allows amplification of weak promoters to give enhanced levels of transgene expression, coupled with silencing of expression following addition of doxycycline.

Example 21 Promoter annotation

Promoters showing particularly selective and/or abundant expression in the M2 and CCM induction systems (also in biopsies and in vivo systems where tested) used in the previous Examples were classified based on shared patterns of gene function. Criteria for classification were based on Uniprot descriptions. Results are shown in Table 5 below. Promoters of the exemplified classes are thus generally illustrated to be of utility for specific induction of therapeutic agents in CD 14+ cells in response to tumour-associated micro-environmental signals.

Table 5

CCL3 and CCL8 are particularly preferred chemokine promoters as they displayed consistently higher expression in CCMs as compared to M2. Their expression levels (MFI) were also amongst the highest of the exemplified promoters (see Figures 4C and D). FN1 and SPP1 are particularly preferred promoters for components of the extracellular matrix, as their levels of cognate gene expression within tumour-associated macrophages and within CD 14 cells following treatment with CCM and ascites, but not in naive CD14 cells (Table II), were among the highest of all. Marco and CD163 are preferred promoters from genes with cargo receptor activity, having consistent selective expression in M2 throughout. CCL3 and FN1 are particularly preferred homodimer promoters as they displayed consistently higher expression in CCMs as compared to M2.