The present invention relates to methods for the expansion of γδ T-cells in particular human Vy9V52T-cells having anti-tumor effector function, as well as to reagents and compositions for use in the methods, and the products of the methods and their use in therapy. In addition, the methods are suitable for enhancing cell expansion efficiency and effector function in some instances.

Background to the Invention

γδ T-cells account for up to 10% of circulating lymphocytes and operate at the interface between innate and adaptive immunity. Four attributes of these versatile cells render them ripe for exploitation in therapies and in particular in cancer immunotherapy. First, γδ T-cells recognise genomic, metabolic and signaling perturbations associated with the transformed state [1, 2] . Second, they possess a diverse network of immune effector activities, overlapping and yet distinct to those deployed by "conventional" οίβ T-cells. γδ T-cells release perforin and granzymes, express both FAS and TRAIL, engage in Fc receptor-dependent effector functions and produce a range of immunomodulatory cytokines, including tumor necrosis factor (TNF) -oi, interferon (IFN)-y and IL-17. Third, γδ T-cells act as efficient antigen-presenting cells, enabling the perpetuation of immune attack through adaptive mechanisms [3] . Finally, since these cells are not HLA-restricted, they do not elicit graft versus host disease. This enhances the prospect of their future use in the allogeneic "off the shelf" setting [4] .

Most circulating γδ T-cells in man display a Vy9V52 receptor that recognises non-peptide phosphoantigens (PAgs), best exemplified by IPP and its stereoisomer DMAPP (Fig. 1) [5] . Since PAgs are intermediates of mevalonate metabolism, Vy9V52 T- cells provide an innate mechanism to detect excess activity of this key metabolic pathway. Such surveillance is justified from an evolutionary standpoint since excess mevalonate pathway flux promotes cellular transformation, acting synergistically with p21Ras [6] . This reflects the fundamental role of this network in the biosynthesis of isoprenoids required for post- translational modification of several GTPases, including p21Ras, Cdc42, Rho, Rab and Rac.

Amino-bisphosphonate (NBP) drugs such as zoledronic (ZA), alendronic (AA) , pamidronic (PA) and ibandronic acid (IA) exert anti-tumor activity through a combination of directly cytotoxic and immunomodulatory mechanisms [7] . A key example of the latter is the ability of these drugs to activate Vy9V52 T-cells. This results from inhibition of FPP synthase within the mevalonate pathway, leading to increased PAg accumulation (Fig. 1) [8] . Tumor cells that have been pulsed with NBPs rapidly acquire large PAg loads and thus become more sensitive to recognition by Vy9V52 T-cells [5, 9]. Exploitation of this principle provides an opportunity to enhance tumor susceptibility to γδ T-cell immunotherapy .

The clinical development of γδ T-cell immunotherapy builds on two established findings. First, in an effort to achieve in- vivo expansion of Vy9V52 T-cells, patients with diverse

malignancies have been treated with ZA and low-dose IL-2. In many cases, these small studies have correlated circulating Vy9V52 T-cell numbers with retarded disease progression [10]. Second, ex-vivo expanded Vy9V52 T-cells have been tested as an autologous adoptive immunotherapy in several early phase clinical trials, involving diverse cancers including epithelial ovarian cancer (EOC) [11-13] . Although these studies have demonstrated the safety of infused γδ T-cells, clinical efficacy has been limited (even when combined with ZA) . This highlights the need for better systems to expand these cells at high efficiency, yielding cells that exhibit improved anti-tumor activity .

Transforming growth factor-β (TGF-β) is a secreted protein that exists in at least three isoforms, called TGF-βΙ, TGF-p2 and TGF- 3. It is a cytokine that has a role in a variety of processes including proliferation and cellular differentiation, but also immunity and cancer. It is generally understood that in this context, it has a regulatory immune effect, and this may explain in part why it is upregulated in certain cancers, which overexpress the cytokine to reduce the host immune response. There are many papers showing that addition of TGF-β to T-cells promotes a regulatory phenotype. For example, two independent groups have shown that culturing human peripheral blood mononuclear cells (PBMCs) in the presence of cytokines that included TGF-β resulted in the production of regulatory γδ T- cells expressing high levels of Foxp3 and CD25 having an immunosuppressive function [14, 15].

The applicants have carried out studies of various protocols for the expansion of γδ T-cells and have found a particular set of conditions which produce high levels of cells with enhanced effector activity.

Summary of the Invention

Surprisingly, the applicants have found that the presence of TGF-β can, under certain culture conditions, produce enhance yields of effector T-cells having an immunostimulatory activity in particular against cancer cells. Furthermore, the anticancer efficacy of the cells produced using this method may be increased .

According to the present invention there is provided a method for expanding a population of γδ T-cells, said method comprising culturing isolated activated peripheral blood mononuclear cells (PBMCs) in a medium comprising transforming growth factor beta (TGF-β) under conditions in which the production of effector γδ T-cells having therapeutic activity against malignant disease is favored.

In particular, the T-cell population produced using the method of the invention is rich in γδ cells and in particular νγ9νδ2 cells, having therapeutic activity against malignant disease. Malignant disease in this case includes in particular proliferative disease such as cancer, including solid tumors, liquid tumors or blood cancers or other cancers of the

circulatory system. Examples of solid tumors include breast cancer, ovarian cancer, cancer of the colon and generally the G ( gastro-intestinal ) tract, cervix cancer, lung cancer, in particular small-cell lung cancer, and non-small-cell lung cancer, head and neck cancer, bladder cancer, cancer of the prostate or Kaposi's sarcoma. Examples of circulatory system cancers include leukemias such as Acute Myeloid leukaemia (AML) , Myelo-dysplastic syndrome (MDS), myelo-proliferative diseases (MPD) , Chronic Myeloid Leukemia (CML) T-cell Acute Lymphoblastic leukaemia (T-ALL) , B-cell Acute Lymphoblastic leukemia (B-ALL) , Non-Hodgkins Lymphoma (NHL) and B-cell lymphoma.

As used herein, the expression 'effector T-cells' refers to T-cells having an anti-tumor or anti-leukemic effect rather than a regulutory or immunosuppressive effect on the immune response.

It appears that by including TGF-β in the culture medium under certain conditions, both the yield and efficacy of the effector T-cells is increased. This runs contrary to the prevailing understanding that this cytokine results in the production of principally regulatory T-cells.

The PBMCs used as the starting material in the process of the invention are suitably primate PBMCs such as human PBMCs . They are suitably isolated from blood samples from humans or other primates such as apes, using conventional methods.

The cells may be obtained from a patient and then

reintroduced into that patient (autologous therapy) . However, in some circumstances, it has been found that cells from patients who have been heavily pre-treated, for example for solid tumors such as triple negative breast cancer, expand poorly or not at all. In such cases, it may be necessary to obtain the PBMCs used as the starting material in the method of the invention from a healthy donor and to adopt an allogeneic approach to the therapy. In this case, it would be advisable to purify γδ T cells from the expanded product, in particular to remove potentially hazardous B-cells (CD19 ⁺ ) and ο T-cells, in order to facilitate the safe allogeneic use of the γδ T cells .

The TGF-β is suitably present in the culture medium at a concentration of from 0.1 - lOOng/mL, for example at a

concentration of about 5ng/mL. However the precise amount of TGF-β added may depend upon the biological activity of the TGF-β used. This may be determined using a suitable bioassay which yields an ED50 value, equivalent to a Unit of activity. For example, an ED50 for TGF-β may be determined by TGF-p's ability to inhibit the mouse IL-4-dependent proliferation of mouse HT-2 cells. Typically a concentration of 5ng/ml equates to a specific activity of 2 x 10 ⁵ units. Thus suitably from 4 x 10 ³ to 4 x 10 ⁶ units of TGF-β are added to the culture medium, where the unit is determined as described above.

The applicants have found that the nature of the medium may be important in this context. In particular, the medium employed by the applicants has been produced under good

manufacturing process (GMP) and does not contain fetal calf serum or fetal bovine serum, which is frequently included in conventional T-cell culture media [14, 15 (personal

communication, Dr Rita Casetti)] . These particular attributes of the culture medium appear to impact on the development of T- cells in the presence of TGF-β, favoring expansion of effector cells with anti-tumor activity in preference to regulatory T- cells .

In particular, the medium comprises a serum-free medium, such as a synthetic medium like TexMACS (Miltenyi) or RPMI and may be conducted in the additional presence of human AB serum. The medium is suitably a GMP grade medium.

Furthermore, the medium used may further comprise

interleukin-2 (IL-2) . Additional cytokines may be present provided they do not change the nature of the product as being predominantly effector type T-cells with anti-tumor and antileukemic activity. However, in a particular embodiment, the medium does not contain any additional cytokines.

Interleukin-2 is suitably present in the medium in an amount of from 1-lOOOU/mL, for example at about lOOU/mL, where the U is units . One Unit of IL-2 in this context may be defined as the amount of IL-2 in 1 ml that will induce IL-2-dependent murine T cells to incorporate ³ H-TdR at 50% of their maximum level after 24 hours of incubation.

The TGF-β as well as the IL-2 where present is suitably added repeatedly at intervals during the culture process, in particular in response to the cell expansion, which is suitably monitored throughout by counting cells . The cells used as a starting material are activated. In a particular embodiment, this may be achieved by adding an activator capable of activating particularly Vy9V52 T-cells .

Suitable activators may include amino-bisphosphonate drugs such as zoledronic (ZA), alendronic (AA) , pamidronic (PA) and ibandronic acid (IA) . In a particular embodiment, the activator is Zoledronic acid or a salt thereof. Alternatively cells may be activated using a phosphoantigen such as BRHPP or IPP.

The activator is suitably added in an effective amount. Addition may take place with the first addition of TGF-β and IL- 2 where present. The concentration of activator added will depend upon factors such as the specific type of activator used, but will typically be in the range of from 0.1-10 g/ml, for example at about l g/ml.

After expansion as described above, γδ T-cells may then be obtained by purification of the expanded product. In

particular, the CD19 and οί β T-cells may be removed from the product by negative selection or by use of suitable isolation techniques or kits. The applicants have found that if γδ T- cells are isolated from PBMCs prior to expansion, the expansion process may be ineffective.

Using the methods described above, the yield of effector T- cells expanded in-vitro can be enhanced, and so application of this method for enhancing T-cell expansion yield forms a further aspect of the invention.

Similarly, as described below, the efficacy and in

particular the anti-cancer efficacy of the T-cells obtained using this method is enhanced. As a result, the invention further provides a method for enhancing the anti-cancer efficacy of T-cells expanded in-vitro by use of the expansion method described above.

Yet a further aspect of the invention provides the use of TGF-β for enhancing expansion of effector T-cells, and in particular human Vy9V52 T-cells which are useful in the treatment of malignant disease as described above. In a further aspect, the invention provides the use of TGFp for enhancing the anti-cancer effector ability of T-cells .

T-cells obtained by a method as described above form a further aspect of the invention. These may be used in therapy and in particular in for the treatment of cancer.

The cells may be used in the treatment of patients in a conventional manner. In particular, the invention also provides a method for treating a patient in need thereof by

administration of T-cells obtained as described above. In particular the T-cells are adoptively transferred into patients in accordance with standard clinical practice.

In particular, the cells may be administered in conjunction with an activator such as those described herein and/or a chemotherapeutic agent. Suitable activators include

bisphosphonate drugs such as zoledronic acid, alendronic acid and pamidronic acid. They may activate the T cells and also sensitize tumor to T-cells.

Certain chemotherapeutic agents have also been found to sensitize tumors to γδ T-cells [18] and thus these may also be pre- or co-administered with the γδ T-cells of the invention. Particular examples of such chemotherapeutic acids include cisplatin, etoposide, anthracyclines and, as illustrated hereinafter, cytarabine .

The applicants are the first to sequentially administer cytarabine followed by γδ T-cells to produce an anti-tumor effect and this novel therapy forms a further aspect of the invention. In this therapy, effective amounts of γδ T-cells and cytarabine are administered to a patient in need thereof. In particular, the γδ T-cells are obtained in accordance with the present invention .

It may be desirable also to co-administer a cytokine such as IL-2, in order to extend survival of the T-cells.

Detailed Description of the Invention

The invention will now be particularly described by way of example with reference to the accompanying Figures in which Figure 1 is a schematic view showing the mevalonate pathway.

Phosphoantigens (PAg) recognized by Vy9V52 T-cells include

DMAPP, IPP and Apppl . Points of inhibition of the pathway by amino-bisphosphonates and statins are indicated by circles where IPP = Isopentenyl diphosphate and DMAPP= Dimethylallyl

diphosphate

Figure 2 shows the results of ex-vivo expansion of Vy9V52 T- cells using a comparative method (Method 1) . After culture using conditions described above, the percentage (A) and absolute number of γδ T-cells (B) per 20ml blood sample was evaluated at initiation of the culture period and after 15 days. (C)

Expression of the expected Vy9 and V52 T-cell receptor subunits was determined by flow cytometry. Pooled (D) and representative (E) immunophenotypic data of γδ T-cells, expanded ex-vivo for 15 days from healthy donors and women with newly diagnosed EOC (donor number indicated in brackets) .

Figure 3 shows the result of the expansion of Vy9V52 T-cells in various media, with and without human AB serum using comparative method 1. Panels show total cell number (A) , % γδ T-cells present (B) and yield of γδ T-cells (C) after 14 days of

culture .

Figure 4 shows the results of cytotoxicity assays using cells expanded using the comparative method 1 in an assay against a range of ovarian cancer cell lines as follows: (A) IGROV-1; (B) KOC7C; (C) PEOl; (D) PEA; (E) SKOV-3; (F) TOV-21G.

Figure 5 shows the results obtained using a method to expand Vy9V52 T-cells ex-vivo in accordance with the invention. The results show enrichment (A) and expansion (B) of Vy9V52 T-cells (mean + SEM, n=13 independent replicates) . Percentage γδ T-cells present at the beginning and end of manufacture are also shown (mean + SD, n=10) . * p = 0.03 by Mann Whitney test.

Figure 6 shows the comparative anti-tumor activity of method 1 and method 2-expanded γδ T-cells. After expansion of γδ T-cells for 2 weeks using either method 1 or 2, cytotoxicity assays were established in triplicate at a 5:1 effector : target ratio in 96 well plates. Tumor cells were cultured with the indicated aminobisphosphonates for 24 hours prior to undertaking the cytotoxicity assay. After overnight co-culture with Vy9V52 T- cells, residual tumor cell viability was measured by MTT or luciferase assay. Data show mean + SEM tumor cell killing from 2-5 independent replicate experiments performed using the indicated ovarian cancer cell lines (A) IGROV-1, (B) SKOV-3, (C) Kuramochi and (D) TOV-21G; myeloid leukemic cell lines (E) U937 and (F) KG-1 and breast cancer cell lines (G)MDA-MB-231, (H) MDA-MB-468 and (I) BT-20.

Figure 7 illustrates cytokine production by method 1 and method 2-expanded γδ T-cells. γδ T-cells were expanded using method 1 or 2 and then co-cultivated with bisphosphonate-pulsed or unpulsed tumor cells as described in Figure 6. Supernatants were then harvested after 24h of co-culture and analysed for interferon-γ (A-I) and interleukin-2 (J-O) by ELISA. Interferon (I FN)-Y production is shown for the following ovarian cancer cell lines (A) Kuramochi, (B) IGROV-1, (C) SKOV-3, (D) TOV-21G; breast cancer cell lines (E) MDA-MB-468; (F) MDA-MB-231; (G) BT- 20; myeloid leukemic cell lines (H) U937, (I) KG-1. Interleukin- 2 production is shown for co-cultivation experiments undertaken with (J) Kuramochi, (K) U937, (L) KG-1, (M) MDA-MB-231, (N) MDA- MB-468 and (O) BT-20 tumor cells. Data show mean + SEM from 3-5 independent replicate experiments .

Figure 8 shows the results of immunophenotypic analysis of method 1 and method 2-expanded Vy9V52 T-cells. (A) Method 2 expanded cells express a distinct immunophenotype with higher levels of memory (CD45RO, CD27) and homing receptors (CCR7, CXCR4, cutaneous leukocyte antigen (CLA) and E-selectin binding receptors (detected using E-selectin-IgG fusion protein - B) . (C) Relative (rel.) to method, 1, the proportion of naive (CD45RA ⁺ CCR7 ⁺ ) and central memory (CD45- CD27 ⁺ ) cells was higher in method 2-expanded cells. NS - not significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Figure 9 shows the results of an evaluation of cell number (A) and percentage of γδ T-cells (B) present in cultures obtained using method 1 and the method of the invention in a different basic medium (RPMI + 10% human AB serum) . Figure 10 shows the in-vivo therapeutic activity of

intravenously administered expanded Vy9V52 T-cells obtained using method 1 and the method of the invention, against an established burden of malignant disease (U937 leukemia) in SCID Beige mice .

Figure 11 shows the in-vivo therapeutic activity of

intravenously administered expanded Vy9V52 T-cells obtained using the method of the invention against an established burden of malignant disease (U937 leukemia) in SCID Beige mice where

(A) shows tumor burden, indicated by bioluminescence; and

(B) shows the weight of mice, providing an indication of toxicity of the treatment.

Figure 12 shows the in-vivo therapeutic activity of

intravenously administered expanded Vy9V52 T-cells obtained using method 2 against an established burden of malignant disease (MDA-MB-231 triple negative breast cancer, implanted in the mammary fat pad of SCID Beige mice) . (A) Tumor burden, indicated by bioluminescence. (B) Survival of mice. (C) Weight of mice, providing an indication of toxicity of the treatment. Figure 13 shows the results of the use of various purification methods, where (A) illustrates how Vy9V52 T-cells were purified from freshly isolated PBMC by negative selection using a CD19 and οί β T-cell microbead isolation kit; (B) shows the results of attempts to expand these cells; (C) shows the % cell type obtained in experiments in which γδ T-cells were expanded from PBMC using the method of the invention prior to subsequent depletion of of CD19 and οί β T-cells by negative selection; (D) shows the results of flow cytometry analysis of these cells following depletion of contaminating CD19 and οί β T-cells; (E) shows the results of a 24 hour cytotoxicity test of the cells (5:1 effector : target ratio) against MDA-MB-231 (231), MDA-MB-468 (468) or BT20 triple negative tumor cells or (F) against U937 or KG-1 myeloid leukemic cells; (G) illustrates cytokine

concentration in supernatants that had been harvested from treated breast cancer co-cultures and (H) illustrates cytokine concentration in supernatants that had been harvested from treated leukemia co-cultures (n=2 ) . Figure 14 shows that the flow cytometry results of genetically engineered γδ T-cells obtained using the method of the

invention, using a technique in which viral vector was preloaded onto a RetroNectin coated solid phase (B) , or by addition of viral supernatant to cells (C) , as compared to untransduced controls (A) .

Figure 15 shows the results of in-vitro cytoxicity assays against tumor cells ((A)U937 cells and (B)KG1 cells) when treated with the chemotherapeutic agent, cytarabine, at various concentrations for 24 hours preceding the addition of γδ T-cells, including some obtained using the method of the invention (M2) . Figure 16 is a set of graphs showing the results of ίη-νίνο tests in which γδ T-cells of the invention are administered in combination with cyatarabine and IL-2, as compared to the use of cyatarabine and IL-2 alone where (A) shows the tumor burden as indicated by bioluminescence from malignant cells on data 4, 11, 19 and 26 after administration and (B) shows the weight of the mice over the period of the test. In each case, the cyatarbine was injected as a single dose 24 hours before infusion of γδ T- cells .

Comparative Example A

In previous studies, the applicants have shown that healthy donors have 19,916 + 29,887 (mean + SD, n=21) circulating γδ T- cells . By comparison, patients with newly diagnosed EOC had 14,240 + 15,215 γδ cells/ml blood (mean + SD, n=13; not

statistically significant (NS) ) [16] .

To enrich these cells, peripheral blood mononuclear cells (PBMC) were activated with ZA and cultured in AB serum- containing RPMI 1640 medium, supplemented with IL-2/ IL-15.

Specifically, PBMC isolated from normal (healthy) donors (n=21 separate donors) and from patients with EOC (n=13 separate donors) were cultured with ZA (l g/ml day 1 only), IL-2

(lOOU/ml) and IL-15 (lOng/ml) . Cytokines and medium were added daily.

The percentage number of γδ T-cells and the absolute number of γδ T-cells per 20ml blood sample was evaluated at initiation of the culture period and after 15 days. The results are shown in Figure 2 (A) and 2 ( B ) respectively. This "research grade" method resulted in an average expansion of γδ T-cells by 97-fold (EOC patients) or 172-fold (healthy donors; NS) (Fig. 2B) .

Expression of the expected Vy9 and V52 T-cell receptor subunits was determined by flow cytometry and the results are shown in Figure 2 (C) . As is clear, expanded γδ T-cells from patients and healthy donors expressed the Vy9V52 T-cell

receptor .

Pooled and representative immunophenotypic data of γδ T- cells, expanded ex-vivo for 15 days from healthy donors and women with newly diagnosed EOC (donor number indicated in brackets) was also obtained and the results are shown in Figures 2 ( D ) and 2 ( E ) respectively. There was a predominance of γδ T- cells, with small numbers of contaminating γδ T-cells and natural killer (CD16+56 ⁺ , CD3 ^~ ) cells. Expanded cells

predominantly exhibit an effector and effector memory phenotype, which is similar in patients and healthy volunteers . We

subsequently found that addition of IL-15 made no significant difference to the yield of cells obtained and this was omitted from subsequent expansion runs (data not shown) .

To adapt manufacture of γδ T-cell products for clinical use, we tested commercially available GMP media for their ability to support the expansion of these cells using ZA + IL-2. The method as described above was repeated using clinical grade serum-free medium. PBMC were cultured in RPMI + 10% human AB serum or two commercially available GMP grade media, with or without 10% human AB serum. In each case, ZA (l g/ml) was added to activate γδ T-cells, which were then expanded by addition of IL-2 (lOOU/ml) . The results are shown in Figure 3. These show that the TexMACS medium in particular enables the expansion of these cells under serum-free conditions in "method 1".

Cytotoxicity assays were established in triplicate at a 5:1 effector : target ratio in 96 well plates and the results are shown in Figure . Where indicated, tumor cells were pulsed for 24h with the indicated concentration of zoledronic (ZA) or pamidronic acid (PA), prior to addition of γδ T-cells. Residual tumor cell viability was measured after overnight co-culture with Vy9V52 T-cells by MTT assay for (A) IGROV-1; ( B ) KOC7C; ( C ) PEOl; (D ) PEA; ( E ) SKOV-3; ( F ) TOV-21G. The results show that Vy9V52 T-cells expanded using method 1 exhibited broad and NBP- enhanced anti-tumor activity against a range of ovarian and other tumor cell lines .

Example 1

Expansion of T-cells in accordance with the invention.

Next, we modified method 1 such that transforming growth factor (TGF) -β was added together with IL-2 at all times. This approach is referred to hereafter as method 2.

In a variation of the method of Example A above, blood was collected from heathy donors or patients, in a tube with citrate anticoagulant. Using Ficoll-Paque (GE), PBMCs were isolated according to previously published methodology [17].

Isolated PBMC cells were then reconstituted in GMP TexMACS

Media (Miltenyi) at 3xl0 ⁶ cell/mL. To the reconstituted cells, l g/mL Zoledronic Acid (Zometa, Novartis) was added as an activator, together with 100 U/mL IL-2 and 5ng/mL TGF-β. The cells were incubated at 37°C in air containing 5% carbon dioxide.

On day 3, cells were fed with 100 U/mL IL-2 and 5ng/mL TGF- β. Thereafter, on days 4, 7, 9, 11, 13, 15, cells were counted by trypan exclusion using a hemocytometer . If the number of T- cells was less than lxlO ⁶ cells/mL, a further 100 U/mL IL-2 and

5ng/mL TGF-β were added. If the number of T-cells was between lxlO ⁶ and 2xl0 ⁶ cells/mL, an equivalent volume of TexMACS medium was added together with 100 U/mL IL-2 and 5ng/mL TGF-β. If the number of T-cells was greater than 2xl0 ⁶ cells/nL, double the volume of TexMACS media was added together with 100 U/mL IL-2 and 5ng/mL TGF-β.

After 15 days, the cells were analyzed by flow cytometry with a pan γδ antibody to confirm the enrichment of γδ T-cells in these cultures. The results are shown in Figure 5. These show that the modified method achieves enrichment (A) and improved expansion ( B ) of Vy9V52 T-cells (mean + SEM, n=13 independent replicates). Additionally, the T-cells were immunophenotypically characterised and subjected to functional tests. The relative ability of the T-cells obtained using method 1 above, or the present method of the invention to mediate cytotoxic destruction of tumor cells was evaluated. After expansion of γδ T-cells for 2 weeks using either method 1 or 2, cytotoxicity assays were established in triplicate at a 5:1 effector : target ratio in 96 well plates. Where indicated, tumor cells were pulsed for 24h with the indicated concentration of zoledronic (ZA) , alendronic acid (AA) or pamidronic acid (PA), prior to addition of γδ T- cells . Residual tumor cell viability was measured after

overnight co-culture with Vy9V52 T-cells by MTT or luciferase assay for (A) IGROV-1, (B) SKOV-3, (C) Kuramochi and (D) TOV- 21G; myeloid leukemic cell lines (E) U937 and (F) KG-1 and breast cancer cell lines: (G) MDA-MB-231, (H) MDA-MB-468 and (I) BT-20. The results are shown in Figure 6.

Activation of γδ T-cells when co-cultivated with tumor cells was assessed by measurement of release of IL-2 and IFN-γ. Ability of these expanded γδ T-cells to control an established burden of malignant disease was also assessed in SCID Beige mice with an established burden of U937 myeloid leukemia.

The original rationale for inclusion of TGF-β in the culture process was to try to improve expression of homing receptors such as CXCR4 on these cells . Completely unexpectedly however, addition of TGF-β resulted in substantially enhanced yields of Vy9V52 T-cells as shown in Figure 5.

Method 2-expanded cell products also demonstrated

equivalent or enhanced anti-tumor activity against EOC (IGROV-1, SKOV-3, Kuramochi, TOV-21G) , breast cancer (MDA-MB-231) and myeloid leukemic cells (U937), even in the absence of NBP exposure (Fig. 6; Fig. 10) . However, anti-tumor activity was consistently enhanced by prior NBP sensitization (Fig. 6) .

After expansion of γδ T-cells for 2 weeks using either method 1 or 2, co-cultures were established in triplicate at a 5:1 effector : target ratio in 96 well plates. Where indicated, tumor cells were pulsed for 24h with the indicated concentration ( g/ml) of zoledronic (ZA) or pamidronic acid (PA), prior to addition of γδ T-cells . After a further 24 hours, supernatants were harvested and analysed for Interferon-γ or Interleukin-2 by ELISA. The results are shown in Figure 7. Interferon (I FN)-Y production is shown for the following tumor cell monolayers: ovarian cancer cell lines (A ) Kuramochi, ( B ) IGROV-1, ( C ) SKOV- 3, ( D ) TOV-21G; breast cancer cell lines ( E ) MDA-MB-468; ( F ) MDA-MB-231; ( G ) BT-20; myeloid leukemic cell lines (H) U937, ( I ) KG-1. In addition, interleukin-2 production is shown for co- cultivation experiments undertaken with ( J) Kuramochi, ( K ) U937, ( L ) KG-1, (M ) MDA-MB-231, (N ) MDA-MB-468 and ( O ) BT-20 tumor cells .

When compared to cells that had been expanded using method 1, method 2-expanded cells produced significantly higher levels of IFN-γ when engaging tumor cell targets. This effect was most pronounced when transformed cells had been pulsed with very low concentrations of NBP agents (Fig. 7A-G ) . Method 2-expanded cells also produced IL-2 under these conditions, a finding that was not observed using method 1-expanded cells (Fig. 7H-K ) .

Finally, the phenotype of method 1 and method 2 cells was investigated using conventional methods and the results are illustrated in Figure 8. Method 2-expanded cells were found to express a distinctive phenotype, with high levels of homing receptors (CXCR4, CLA, E-selectin binding activity) and memory markers (CD27, CD45RO) . In addition, the proportion of naive (CD45RA ⁺ and CCR7 ⁺ ) and central memory (CD45- and CD27 ⁺ ) cells was higher in method 2 expanded cells as compared to method 1 expanded cells . Thus these cells are distinguishable from cells produced using other expansion protocols.

Example 2

Alternative cell expansion process

The methodology of Example 1 above was repeated using a different basic medium, specifically RPMI + human AB serum. In particular, PBMC (3 x 10 ⁶ cells/ml) were cultured in RPMI + 10% human AB serum containing zoledronic acid (l g/ml) + IL-2

(lOOU/ml; method 1) or zoledronic acid (l g/ml) + IL-2 (lOOU/ml) + TGF-β (5ng/ml; method 2) . Cell number was evaluated on day 15 and the results are shown in Figure 9A. The percentage of γδ T- cells present in each culture was evaluated on the day of initiation of the cultures (day 1) and after a further 14 days (day 15) and the results are shown in Figure 9B .

As before, it is clear that the addition of TGF-β has enhanced cell expansion.

Example 3

In-vivo therapeutic activity

In addition, the in-vivo therapeutic activity of expanded Vy9V52 T-cells against an established burden of malignant disease were compared. Twenty SCID Beige mice were inoculated with 1 x 10 ⁶ firefly luciferase-expressing U937 leukemic cells by tail vein injection and were then divided into 4 groups of 5 mice each. After 4 days, mice were treated as follows: Group 1 is a control group that received PBS alone. Group 2 received pamidronic acid (200 g IV) alone. Group 3 received pamidronic acid (200 g IV on day 4) followed by 20 x 10 ⁶ (day 5) and 10 x 10 ⁶ (day 6) Vy9V52 T-cells that had been expanded using method 1 (IV) . Group 4 received pamidronic acid (200 g IV on day 4) followed by 20 x 10 ⁶ (day 5) and 10 x 10 ⁶ (day 6) Vy9V52 T-cells that had been expanded using method 2 (administered IV) .

Leukemic burden was monitored thereafter by serial

bioluminescence imaging.

The results are shown in Figure 10. It is clear that the efficacy of the cells obtained by method 2 of the invention is significantly greater in this assay.

Example 4

In-vivo activity of cells of the invention in conjunction with IL-2

In a separate experiment, the in-vivo therapeutic activity of intravenously administered expanded Vy9V52 T-cells obtained using the method of the invention (M2) against an established burden of malignant disease (U937 leukemia) in SCID Beige mice was measured. Mice were divided into 4 groups of 5 mice and each received 1 million U937 cells IV on day 1. Thereafter, one group received treatment that may be summarised as follows: Group Treatment

1 PBS (control)

2 Zoledronic acid + IL-2

3 M2 + IL-2

4 M2 + IL-2 + Zoledronic acid

Where administered, 20μq Zoledronic acid was administered intravenously 24 hours after treatment with U937 cells. Mice receiving M2 cells were given 2 treatments of 15 million γδ T- cells intravenously, one day later. Those receiving IL-2 were given 10,000U of IL-2 by the intraperitoneal (IP) route at the same time as M2 administration. On the following 2 days, mice received 10,000U IL-2 IP. A control group received phosphate- buffered saline (PBS) alone.

Bioluminescence from the malignant cells was measured on days 7, 15, 21 and 28 as an indicator of tumor burden. The results are shown in Figure 11A. The results show that Vy9V52 T-cells obtained using the method of the invention significantly reduce tumor burden, in particular when administered with an activator .

Mice were weighed over the course of the treatment to provide an indication of the toxicity of the treatment. The results , shown in Figure 11B, indicate that that there is no significant toxicity associated with the treatment.

Example 5

In-vivo therapeutic effect against breast cancer

In this experiment, 20 SCID Beige mice having an

established burden of malignant disease in the form of MDA-MB- 231 triple negative breast cancer, implanted in the mammary fat pad of the mice, were used. Again, mice were divided into four groups for treatment. Mice were treated as follows: Group 1 is a control group that received PBS alone. Group 2 received 20 g Zoledronic acid intravenously. Group 3 received 20 x 10 ⁶ (day 2) and 10 x 10 ⁶ (day 3) Vy9V52 T-cells that had been expanded using method 2 intravenously. Group 4 received 20 g Zoledronic acid intravenously on day 1 followed by 20 x 10 ⁶ (day 2) and 10 x 10 ⁶ (day 2) Vy9V52 T-cells that had been expanded using method 2. The resultant tumor burden as measured by bioluminesence was measured over a period of 28 days. The results are shown in Figure 12. In this case, the cells obtained using the method of the invention produced a significant reduction in tumor burden (Figure 12A) accompanied by prolonged survival (Figure 12B) .

Mice were weighed over the course of the treatment to provide an indication of the toxicity of the treatment. The results, shown in Figure 12C, indicate that that there is no significant toxicity associated with the treatment.

Example 6

Purification of expanded γδ T-cells

In a first experiment, Vy9V52 T-cells were purified from freshly isolated PBMC by negative selection using a CD19 and/or a ο T-cell microbead isolation kit. Where both kits were used, residual contaminating CD19 and οί β T-cells were <0.1% as shown in Figure 13 (A) .

The purified cells were subjected to expansion using method 2 as described in Example 1. However, these cells were not able to expand as illustrated in Figure 13 ( B ) . Thus it appears that the starting material must comprise PBMCs .

In other experiments, γδ T-cells were expanded from PBMCs using method 2 for 15 days. At this point, flow cytometry analysis demonstrated that significant numbers of οί β T-cells remain, accompanied by small numbers of CD19 ⁺ cells (n=4) (Figure 13 ( C ) ) .

The resultant product was then depleted of CD19 and οί β T- cells by negative selection, as described above in relation to Figure 13 (A) . Two representative flow cytometric analyses are shown in Figure 13 (D ) to indicate the efficiency of the

depletion process.

Following purification by negative selection using the MACS beads (Miltenyi), method 2-expanded γδ T-cells were tested in a 24 hour cytotoxicity assay (5:1 effector : target ratio) against MDA-MB-231, MDA-MB-468 or BT20 triple negative tumor cells or U937 or KG-1 myeloid leukemic cells using methodology similar to that described in Example 1. Cells were tested alone, or in combination with zoledronic acid. There was a negative control and a control with activator alone. Tumor cell viability was measured by luciferase assay and/or MTT assay (n=2) . The results are shown in Figure 13(E) and Figure 13(F) respectively. As is clear, the combination of T cells and activator produced a significant reduction in tumor cell viability.

Supernatants were harvested from these breast cancer and leukemia co-cultures, after 24h, and analysed for the presence of IFN-γ and/or IL-2. The results are shown in Figures 13(G) and 13 (H) respectively. Cytokine levels were substantially raised in the case of the combination of T-cells expanded in accordance with the invention and activator.

These experiments show that method 2 expanded γδ T-cells are fully functional if purified by negative selection after expansion, but not before. This purification facilitates the safe allogeneic use of these cells since potentially hazardous B-cells (CD19 ⁺ ) and ο T-cells have been removed.

Example 7

Genetic Engineering of Expanded Cells

To further confirm the functionality of γδ T-cells expanded in accordance with the invention, they were genetically

engineered by retroviral transduction. Cells were either transduced by pre-loading viral vector onto a RetroNectin coated solid phase or by addition of viral supernatant to the expanding cells .

It was clear that in order to preserve the efficient enrichment of these cells during expansion, it is preferable to pre-load viral vector onto a RetroNectin coated solid phase (Figure 14(B) ) , rather than addition of viral supernatant

(Figure 14(C) ) . This is indicated by the greater percentage of transduced cells and the greater percentage of γδ T-cells present when gene transfer is achieved using the pre-loading method .

Example 8

Effects of combination of γδ T-cells with chemotherapeutic agent

Cytotoxicity assays were established in triplicate at a 1:1 effector : target ratio in 96 well plates containing either U937 tumor cells or KG-1 tumor cells. Where indicated, tumor cells were pulsed for 24h with the indicated concentrations of cytarabine, prior to addition of γδ T-cells, produced either using the method of the invention (M2) or the method of the comparative example (Ml) above. There were three donors for the M2 cells and two donors for the Ml cells. A control group received no cytarabine .

Residual tumor cell viability was measured after overnight co-culture with Vy9V52 T-cells by luciferase assay. The results, shown in Figure 15 show that sub-lethal doses of cytarabine potentiated the anti-tumor activity of Vy9V52 T-cells expanded using method 2 against two cell models of AML (three donors for M2 cells and two donors for Ml cells)

In a separate experiment, fifteen SCID Beige mice were inoculated with 1 x 10 ⁶ firefly luciferase-expressing U937 leukemic cells by tail vein injection and were then divided into 3 groups of 5 mice each. After 4 days, mice were treated as follows: Group 1 is a control group that received PBS alone. Group 2 received cytarabine (480mg/Kg IV on day 4) and IL-2 (10000 IP on day 5, 6, 7 and 8) . Group 3 received cytarabine (480mg/Kg IV on day 4) followed by 20 x 10 ⁶ (day 5 and 6) Vy9V52 that had been expanded using method 2 (IV) and IL-2 (10000 IP at days 5 , 6 , 7 and 8 ) .

Leukemic burden was monitored thereafter by serial bioluminescence imaging. Bioluminescence from the malignant cells was measured on days 4, 11, 19 and 26 as an indicator of tumor burden. The results are shown in Figure 16(A) and indicate that Vy9V52 T-cells obtained using the method of the invention reduce tumor burden most effectively when administered with cytarabine. Mice were weighed over the course of the treatment to provide an indication of the toxicity of the treatment. The results shown in Figure 16 (B) , indicate that that there is no significant toxicity associated with the treatment.

References

[1] P. Vantourout et al., Nat Rev Immunol, 13 (2013) 88-100

[2] P. Vantourout et al., Sci Transl Med, 6 (2014) 231ra249

[3] M. Brandes et al . , Science, 309 (2005) 264-268. [4] M. Wilhelm et al., J Transl Med, 12 (2014) 45.

[5] I. Benzaid et al., Cancer Res, 71 (2011) 4562-4572.

[6] J.W. Clendening et al . , Proc Natl Acad Sci U S A, 107 (2010)

15051-15056.

[7] U. Laggner et al . , Clin Immunol, 131 (2009) 367-373.

[8] J.E. Dunford et al . , J Med Chem, 51 (2008) 2187-2195.

[9] I. Benzaid et al., Clin Cancer Res, 18 (2012) 6249-6259.

[10] F. Dieli et al . , Cancer Res, 67 (2007) 7450-7457.

[11] J. Bennouna et al . , Cancer Immunol Immunother, 57 (2008) 1599-1609.

[12] H. Kobayashi et al., Anticancer Res, 30 (2010) 575-579.

[13] A.J. Nicol et al., Br J Cancer, 105 (2011) 778-786.

[14] Y. Gu et al., J Immunol Methods, 402 (2014) 82-87.

[15] R. Casetti et al., J Immunol, 183 (2009) 3574-3577.

[16] A.C. Parente-Pereira et al . , J Immunol, 193 (2014) 5557- 5566.

[17] A.C. Parente-Pereira et al., J Biol Methods, 1 (2014) e7.

[18] S. R. Mattarollo et al . , Cancer Immunol. Immunother. (2007) 56: 1285-1297