Modified NK-92 cells, and therapeutic and diagnostic uses thereof

Abstract

Modified natural killer 92 (NK-92) cells and their use in cancer therapy, in particular for the prevention or treatment of solid tumours such as sarcomas, carcinomas, melanoma and lymphoma, and non-solid tumours such as leukaemia and related disorders. Embodiments of the invention further relate to the use of the modified NK-92 cells for in vitro diagnosis, diagnostics and/or screening, methods for the preparation of a modified NK-92 that is specific for a target antigen of a target cell in a subject, and to an expression vector, comprising the nucleic acid sequences of an antigen-specific functional T cell receptor (TCR), CD3, CD4 and/or CD8.

Claims

1. A modified natural killer 92 (NK-92) cell expressing: (1) an antigen-specific functional T cell receptor (TCR), (2) CD3, (3) CD8 and/or CD4.

2. The modified NK-92 cell according to claim 1, wherein said TCR has the ability of specifically binding to an antigen MHC complex on the surface of a target cell.

3. The modified NK-92 according to claim 1, wherein the target cell is a cancer cell.

4. The modified NK-92 cell according to claim 1, wherein the TCR originates from 5B2, 25F2 or 5H11 cells.

5. The modified NK-92 cell according to claim 4, wherein the 5B2 TCR is composed of an alpha chain encoded by a nucleic acid sequence as defined in SEQ ID NO: 1 and a beta chain encoded by a nucleic acid sequence as defined in SEQ ID NO: 2, or degenerate variants thereof.

6. The modified NK-92 cell according to claim 4, wherein the 25F2 TCR is composed of an alpha chain encoded by a nucleic acid sequence as defined in SEQ ID NO: 3 and a beta chain encoded by a nucleic acid sequence as defined in SEQ ID NO: 4, or degenerate variants thereof.

7. The modified NK-92 cell according to claim 4, wherein the 5H11 TCR is composed of an alpha chain encoded by a nucleic acid sequence as defined in SEQ ID NO: 5 and a beta chain encoded by a nucleic acid sequence as defined in SEQ ID NO: 6, or degenerate variants thereof.

8. The modified NK-92 cell according to claim 1, wherein the TCR is expressed by the NK-92 cell as a fusion protein together with CD3 in conjunction with CD8 and/or CD4.

9. A modified natural killer 92 (NK-92) cell according to claim 1 for use as a medicament.

10. A modified natural killer 92 (NK-92) cell according to claim 1 for use in cancer therapy.

11. The modified NK-92 cell according to claim 10, wherein the cancer is a solid tumour.

12. A modified natural killer 92 (NK-92) cell according to claim 1 for use in the prevention or treatment of leukaemia and related disorders.

13. The modified NK-92 cell according to claim 12, wherein the leukaemia is selected from acute myeloid leukaemia (AML), chronic lymphocyte leukaemia (CLL), acute lymphocytic leukaemia (ALL), chronic myeloid leukaemia (CML), chronic myelomonocytic leukaemia (CMML), eosinophilic leukaemia, hairy cell leukaemia, Hodgkin's lymphoma (HL), multiple myeloma (MM), non-Hodgkin's lymphoma (NHL), myeloproliferative disorders or myelodysplastic syndrome (MDS).

14. Use of a modified natural killer 92 (NK-92) cell according to claim 1 for in vitro diagnosis, diagnostics and/or screening.

15. A pharmaceutical composition, comprising a modified natural killer 92 (NK-92) cell according to claim 1, and at least one pharmaceutically acceptable carrier or excipient.

16. An in vitro method for the preparation of a modified natural killer 92 (NK-92) cell that is specific for a target antigen of a target cell in a subject, (1) determining a target antigen in the target cell that is expressed on the surface of the target cell, (2) identifying the type of the MHC complex in the subject, (3) providing a NK-92 cell that expresses i. an antigen-specific T cell receptor (TCR) that has the ability of specifically binding to the antigen-MHC complex identified in step (1) on the surface of the target cell, ii. CD3, iii. CD4 and/or CD8 to produce a NK-92-TCR-CD3.sup.+CD4.sup.+ cell or NK-92-TCR-CD3.sup.+-CD8.sup.+ cell or NK-92-TCR-CD3.sup.+-CD4.sup.+-CD8.sup.+ cell.

17. The method according to claim 16, wherein the CD3 consists of a CD3 γ chain, a δ chain, two ε chains and two ζ chains.

18. An expression vector, comprising the nucleic acid sequences of an antigen-specific functional T cell receptor (TCR), CD3, CD4 and/or CD8.

19. The expression vector of claim 18, wherein the vector comprises nucleic acid sequences derived from any one of the nucleic acid sequences defined in SEQ ID NO: 1 to 12 for expressing TCR-CD3-CD4 and/or -CD8.

20. The modified NK-92 cell according to claim 11, wherein the solid tumour is a sarcoma, a carcinoma, a melanoma, or a lymphoma.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0065] FIG. 1 shows the gene card of plasmid pMX_IRES_Puro_TCR-25F2

[0066] FIG. 2 shows the gene card of plasmid pMX_IRES_Puro_TCR-562

[0067] FIG. 3 shows the gene card of plasmid pMX_IRES_Puro_TCR-5H11

[0068] FIG. 4 shows the gene card of plasmid pMX_IRES_Neo_CD8α1β2

[0069] FIG. 5 shows the gene card of plasmid pMX_IRES_Neo_CD8α1β5

[0070] FIG. 6 shows the gene card of plasmid pMX_IRES_Neo_CD8α2β3

[0071] FIG. 7 shows the gene card of plasmid pMXs_CD3 complex_Dest. The genes encoding for the CD3δ-, CD3γ-, CD3ε- and CD3ζ-chain, connected to each other by insertion of sequences coding for the F2A, T2A, and P2A self-cleaving peptides were cloned into a pMXs-based destination vector modified for Gateway-Cloning®. The CD3 complex was then finally transferred into the retroviral pMXs_IRES_Puromycin vector for retroviral transduction of parental NK92 and successful selection of NK92CD3.sup.+ cells.

[0072] FIG. 8 shows expression of different TCRs in combination with CD3 on redirected NK92 cells. A CD3_IRES_Puromycin construct was transduced into NK92 cells, and cells were puromycin selected. Successful CD3 gene transfer was monitored upon co-transduction of three different TCRs (5H11, 25F2, 5B2). Expression of the known TCRß-chains was examined using an anti-Vß8-PE mAb reactive to 5H11 (92%), an anti-Vß8-FITC mAb reactive to 25F2 (93.1%), and an anti-Vß21.3-FITC mAb reactive to 5B2 (86.6%). CD3 was monitored using a CD3-Pacific Blue mAb. TCR/CD3 double positive cells appear in the right upper quadrant while naturally no staining is detectable for TCR and CD3 alone shown in the left upper and right lower quadrant.

[0073] FIG. 9 illustrates that CD8 expression can be detected at the surface of NK92CD3.sup.+ cells following retroviral transduction with a CD8_IRES_Neomycin construct and puromycin/neomycin selection. Cells were stained with an anti-CD8-PE mAb and a representative histogram plot shows that 96.8% of NK92CD3.sup.+ cells also express CD8α1β2.

[0074] FIG. 10 demonstrates expression of the TCR 5H11 following retroviral transduction into NK92 cells previously modified to express a combination of CD3 and CD8. The TCRß-chain was detected using a Vß8-specific mAb conjugated with PE. CD3 and CD8 expression were examined with anti-CD3-APC and anti-CD8-APC mAbs, respectively. Retroviral TCR gene transfer in NK92CD3.sup.+CD8.sup.+ cells resulted in 98.8% TCR.sup.+CD3.sup.+ cells (FIG. 10A) and 87.4% TCR.sup.+CD3.sup.+ cells stained for co-expression of CD8 (FIG. 10B) while virtually no (0.1%) CD3 was detectable in NK92CD3.sup.+CD8.sup.+ control cells due to the absence of TCR expression (FIG. 10C).

[0075] FIG. 11 shows expression of the TCR 25F2 following retroviral transduction. The TCRß-chain of TCR 25F2 was detected using a Vß8-specific mAb conjugated with FITC. CD3 and CD8 expression was analyzed with anti-CD3-Pacific Blue and anti-CD8-PE mAbs, respectively. TCR transduction resulted in 86.7% 25F2CD3.sup.+ NK92 cells (FIG. 11A) and 82.2% of 25F2CD3.sup.+CD8.sup.+ cells (FIG. 11B) while 0.1% of TCR/CD3 expression was observed in non TCR transduced NK92CD3.sup.+CD8.sup.+ cells (FIG. 11C).

[0076] FIG. 12 demonstrates expression of the TCR 5B2 following retroviral transduction of NK92CD3.sup.+CD8.sup.+ cells. The TCR ß-chain of TCR 5B2 was detected using a Vß21.3-specific mAb labeled with FITC. CD3 and CD8 expression was examined by staining with anti-CD3-Pacific Blue and anti-CD8-PE mAbs, respectively. FIG. 12A illustrates the percentage of TCR 5B2 transduced NK92CD3.sup.+CD8.sup.+ expressing 5B2 and CD3 (84.9%) while FIG. 12B depicts the 5B2CD3.sup.+CD8.sup.+ NK92 cells (83.0%). In FIG. 12C 0.1% TCR/CD3 expression was observed in non TCR transduced NK92CD3.sup.+CD8.sup.+ cells.

[0077] FIG. 13 depicts the results of an IFN-γ ELISPOT assay following overnight stimulation of 5H11 TCR-redirected NK92CD3.sup.+CD8.sup.+ cells with the HLA-matched, EBV-immortalized B cells (B-LCL) of patient MZ-580 (EBV-B580). Non-HLA-matched B-LCL (EBV-6667) were used as specificity control while K562 cells served as target control for NK cell mediated responses. Non TCR transduced NK92CD3.sup.+ and 5H11CD3.sup.+ NK92 were included as effector controls. Medium; effector cells without target cells (B-LCL, K562).

[0078] FIG. 14 shows the results of an IFN-γ ELISPOT assay following overnight stimulation of 25F2 TCR-redirected NK92CD3.sup.+ and NK92CD3.sup.+CD8.sup.+ cells with HLA-matched AML blasts (MZ-921). AML MZ-667 and K562 cells were used as specificity controls for non HLA-matched and NK-cell mediated responses. Non TCR transduced NK92CD3.sup.+ served as effector controls. Medium; effector cells without target cells (AML, K562).

[0079] FIG. 15 illustrates the results of an IFN-γ ELISPOT assay following overnight stimulation of 562 TCR-redirected NK92CD3.sup.+CD8.sup.+ and NK92CD3.sup.+ cells with HLA-matched AML blasts derived from patient MZ 653. AML 667 and K562 cells served as specificity controls for non-HLA matched controls and NK cell mediated responses. Non TCR transduced NK92CD3.sup.+ served as effector controls. Medium; effector cells without target cells (AML, K562).

[0080] FIG. 16 depicts TCR 5H11 mediated killing of 5H11CD3.sup.+CD8.sup.+ NK92 cells measured by reduction of a luciferin/ATP-dependent bioluminescence signal (BLI) of B-LCL target cells constitutively expressing a firefly luciferase reporter gene (FLuc). TCR 5H11-redirected NK92CD3.sup.+ and 5H11CD3.sup.+CD8.sup.+ NK92 as well as TCR 562-expressing NK92CD3.sup.+CD8.sup.+ and non TCR transduced NK92CD3.sup.+ were cocultured in triplicates at indicated ratios with FLuc expressing HLA-matched, EBV immortalized B cells (B-LCL) from patient MZ 580 in the presence of D-Luciferin. After 18 h of coculture remaining BLI was determined with a BMG Fluostar Omega Reader at 10 s integration time per well and TCR-specific lysis was quantified by deterioration of FLuc-signal compared to residual luminescence of target cells cocultured with non TCR transduced NK92CD3.sup.+. TCR 5B2-redirected NK-92CD3.sup.+CD8.sup.+ served as specificity control.

[0081] FIG. 17 demonstrates NK-cell mediated cytotoxicity of the TCR 5H11 and 5B2 transduced NK92 subsets measured by reduction of luciferin/ATP-dependent BLI in FLuc-K562 target cell transfectants. FLuc-expressing K562 were co-cultured with non TCR transduced NK92CD3.sup.+ as well as with TCR 5H11 or 5B2 expressing NK92CD3.sup.+ and NK92CD3.sup.+CD8.sup.+ at an effector to target ratio of 10:1 in triplicates in the presence of D-Luciferin. Remaining BLI was examined every hour over a 5 h time period and after 24 h of co-culture with BMG Fluostar Omega Reader at 10 s integration time per well. General lysis was quantified as deterioration of FLuc-signal compared to luminescence of target cells cultivated with no effectors.

[0082] FIG. 18 shows expression profiling of different activating and inhibiting markers monitored on wild type (wt) NK92, NK92CD3.sup.+ and NK92CD3.sup.+CD8.sup.+ cells. Modified NK cells have been retrovirally transduced to express the TCR 5H11 and 5B2. The following mAb-conjugates were used to detect expression by flow cytometry: anti-NKp30-APC, anti-NKp44-PE, anti-NKG2C-PE, anti-NKG2D-APC, anti-NKG2A-PE, anti-TIGIT-PE, anti-CD96-PE, anti-TIM3-PE, anti-CTLA4-APC, anti-PD1-APC. While NKp30 expression appears to be clearly reduced in NK92-5B2CD3.sup.+, NK92-5B2CD3.sup.+CD8.sup.+ and NK92-5H11CD3.sup.+CD8.sup.+ cells, no further significant changes in the expression level of other markers could be detected between the NK92 cell populations.

MATERIAL AND METHODS

Donors and Patients

[0083] Healthy donors of T lymphocytes and leukaemia patients participated in the study after informed consent in accordance with the Helsinki protocol. High-resolution genomic HLA typing was performed according to standard procedures.

Primary Cells and Cell Lines

[0084] NK92 cells were cultured in Alpha-MEM medium supplemented with 20% fetal calf serum (FCS) (PAA Laboratories, Pasching/Austria), 1% Penicillin/Streptomycin (Gibco/Thermofisher Scientific), 0.2M Inositol (Applichem, Darmstadt, Germany), 0.02M Folic acid (Applichem), 0.1 mM Mercaptoethanol (Sigma Aldrich, Steinheim, Germany), and 200 IU recombinant human (rh) IL-2/ml (Novartis). Medium was exchanged every 2-3 days. AML blasts (AML 667, 921 and 653) were isolated either from peripheral blood, bone marrow biopsies, or therapeutic leukapheresis products of patients by standard Ficoll separation and cryopreserved until use. All leukaemia samples contained >95% leukaemia blasts. EBV-transformed B-lymphoblastoid cell lines (B-LCL) were generated from patient peripheral blood mononuclear cells (PBMC) according to standard procedures. The chronic myelogenous leukaemia (CML) cell line K562 was cultured using standard protocols.

Generation of AML-Reactive CD8 CTL-Clones 5H11, 25F2 and 5B2

[0085] Naive CD8.sup.+CD45RA.sup.+ T cells of healthy donors were MACS® isolated (Naive CD8.sup.+ T cell Isolation Kit, Miltenyi Biotec, Bergisch Gladbach, Germany) to be stimulated at 1:1 ratio with fully HLA-matched AML blasts (5×10.sup.4 cells/well) and autologous feeder cells (CD45RA.sup.− PBMC) (5×10.sup.4 cells/well) to generate CTL clones. AML blasts and feeder cells were irradiated prior to co-cultures (35 Gy for feeder cells, 60 Gy for AML blasts). Medium was AIM-V (Gibco/Life Technologies, Thermofisher Scientific) supplemented with 10% pooled and heat inactivated human serum and experimentally determined optimal concentrations of 5 ng/mL of each rhIL-7 and rhIL-15 (Peprotech), 1 ng/mL rhIL-12 (R&D Systems), and 10 ng/mL rhIL-21 (Biomol). T-cell cultures were expanded by weekly addition of irradiated AML blasts and cytokines. From d14 onward, IL-12 was replaced by 100 IU/mL rhIL-2 (Novartis). T cells were regularly tested for reactivity using IFN-γ ELISPOT assays. Clonality of T cells was determined by flow cytometry using Vß-profiling monoclonal antibodies (mAbs) (Beckman Coulter).

Cloning of TCR Genes

[0086] Cloning of TCR genes was performed as originally described by Birkholz K et al. (A fast and robust method to clone and functionally validate T-cell receptor, Journal of Immunol. Methods. 2009 (346): 45-54). Briefly, total RNA of a T-cell clone was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturers' instructions. Synthesis and amplification of cDNA was performed as originally described for tumour cell RNA (Harris et al., 2005). In brief, reverse transcription was performed on ≤1.0 μg of RNA. Ten pmol of the 64T-primer (CGATAAAAGCTCCGGGGATAACAGAT63VN, V=A,G,C; N=A,C,G,T) and 10 pmol of the capswitch oligo (AAGCAGTGGTAACAACGCAGAGT ACGCGGG) were added to the RNA, and primer annealing was performed for 2 min at 72° C., followed by 1 min at 4° C. The cDNA synthesis was performed using Supercript II reverse transcriptase (Invitrogen/Thermofisher Scientific), first strand buffer (Invitrogen), 100 mM DTT (Invitrogen), 10 mM dNTPs (Invitrogen) and subsequent incubation for 60 min. at 42° C., followed by 1 min. at 4° C. Amplification of the cDNA was performed by adding Advantage 2 polymerase (Clonetech) in the presence of 64T-primer and T7-Capswitch-primer (TTATACGACTCACTATAGG GAGGAAGCAGTGGTAACAACG CAGAGT) and dNTPs using 20 PCR reaction cycles. The quality of the amplified cDNA was analyzed by standard gel electrophoresis. Next, the complete cDNA of the TCRα- and β-chains were amplified using primers designed according to the sequence results. The amplified TCR chains were then cloned for further use and sequenced. All different regions of the TCRs were determined using the IMGT V-QUEST database (IMG/V-QUEST; www.imgt.orq).

Cloning of CD8 and CD3 Coreceptors

[0087] According to Szymczak, A. L., et al. (2004) (“Correction of multi-gene deficiency in vivo using a single ‘self-cleaving’ 2A peptide-based retroviral vector.” Nature Biotechnology 22: 589.) a polycistronic vector encoding the δ-, ε-, γ-, ζ-subunit of the human CD3 complex was cloned for retroviral transduction. In order to generate retroviral CD8-expression vectors, RNA was isolated from the original CTL clones, the TCRs 5B2 and 25F2 derived from, using the RNeasy Mini Kit (Qiagen, Hilden, Germany). Upon reverse transcription with SuperScript III reverse transcription kit (ThermoFisher Scientific, Waltham, USA) the CD8α- and β-chains expressed by the original CTL were amplified via PCR and In-Fusion cloning into pMX retroviral vector backbone was performed using NEBuilder HIFI DNA-Assembly Kit (New England Biolabs GmbH, Frankfurt, Germany).

Generation of Retroviral Particles for Transduction of NK92 Cells

[0088] For the generation of retrovirus, a second-generation retrovirus producer cell line (Phoenix-Ampho) was utilized that stably expresses gag-pol and the envelope vector pColtGalv. The day before transfection, 2.5×10.sup.6 Phoenix cells were plated in a 100 mm cell culture dish. For transfection, 5 μg of each vector for packaging and virus envelope (pHit60 and pColtGALV) and 10 μg of the retroviral transfer vector were mixed in Opti-MEM medium with polyethylen-eimine-(PEI) and allowed to form PEI:DNA complexes. Then, the mixture was filled up to a total volume of 5 mL with Opti-MEM and applied dropwise onto Phoenix cells prewashed with PBS. After 4 h of incubation, the mixture was replaced by 5 mL fresh medium (complete Alpha-MEM medium as described above) and the cells were cultivated for 48 h. Retroviral supernatant was harvested 48 hours after transfection and sterile filtered using a 0.45 μm pore sized filter.

Retroviral Transduction of NK92 Cells

[0089] For the transduction of NK92 cells the spin infection method using polybrene was applied. Polybrene acts as a polycationic linker molecule to increase infection efficiency. One×10.sup.6 NK92 cells/well were resuspended in 1 ml freshly harvested retroviral particles supernatant and plated into a 24-well plate. After the addition of 5 μg mL.sup.−1 polybrene, the plate was centrifuged at 2,000 rpm for 90 min at 32° C. without deceleration. Thereafter, the cells were incubated for 24 h at 37° C. in the retroviral particles supernatant. Afterwards cells were harvested, counted and washed in order to remove residual retroviral particles. Cells were then resuspended in complete NK92 cell medium including rhIL-2 as described. For selection of successfully transduced NK92 cells neomycin or puromycin was added at a final concentration of 500 μg/mL (G418) and 1 μg/ml (Puromycin) for up to 7 days. Since transduction of CD8 conferred neomycin resistance while NK92CD3.sup.+CD8.sup.+ cells were also resistant to puromycin, additional expression of a given TCR was achieved by 2-3 rounds of enrichment of TCRCD3.sup.+CD8.sup.+ NK92 cells using anti human CD3 mAb-conjugated immunomagnetic beads and magnetic cell sorting (MACS®, Miltenyi Biotec) or fluorescence activated cell sorting (FACS®) on a BD Aria FACS-sorter. Expression of CD3, CD8 and TCRs was regularly monitored by flow cytometry as described. For continuous cultivation, the cells were seeded at 2 to 3×10.sup.5 cells/ml in a small culture flask.

Flow Cytometric Analysis

[0090] NK92 cells were incubated with FITC-, PE-, APC- or Pacific Blue-conjugated monoclonal antibodies (mAbs) specific for the indicated antigens TCR-Vß8 (Biolegend), TCR-Vß21.3 (Beckman Coulter), CD3, CD8, CTLA4, PD1 (all from BD Biosciences) CD96 (Santa Cruz Biotechnol.), NKG2A, NKG2C (both from Miltenyi Biotec), TIGIT, NKG2D, NKp30, NKp44 and TIM-3 (all from Biolegend) for 15 min at 22° C. and washed afterwards. 10.sup.4-10.sup.5 events of viable cells were analyzed on a BD FACSCanto II flow cytometer. eGFP expressing B-LCL transfectants were measured by GFP expression using the FITC channel.

IFN-γ Enzyme-Linked Immunosorbent Spot (ELISpot) Assay

[0091] Multiscreen HTS™ IP plates (Millipore, Bedford, Mass.) were coated with 10 μg/mL mAb anti-hIFN-γ 1-DIK (Mabtech, Stockholm, Sweden). Parental or genetically modified NK 92 cells were seeded at 1×10.sup.5/well and target cells at 1×10.sup.5/well in Alpha-MEM medium supplemented as described above. Modified NK92 and targets seeded alone served as background controls. After overnight incubation at 37° C., plates were washed with PBS including Tween 20 and captured IFN-γ was detected by biotinylated mAb anti-hIFN-γ 7-B6-1 (Mabtech) at 2 μg/mL, a avidin/horseradish peroxidase complex and AEC solution to visualize captured IFN-γ. Spots were developed and counted using a computer-assisted video image analysis system (KS ELISpot 4.9; Zeiss, Jena, Germany). Shown results are means±SD of representative duplicates.

Evaluation of TCR-Mediated Cytotoxicity In Vitro

[0092] To measure the effect of TCR redirection and CD8 co-expression on NK92 cytotoxicity, 1×10.sup.4 FLuc transduced K562, EBV-B-LCL 580 and 667 target cells were cocultured per well in triplicates with NK92 effectors at E:Ts from 40:1 to 0.625:1 in black 96-well plates in the presence of the FLuc-substrate D-Luciferin (Thermo Scientific). After 18 h of incubation relative luminescence units were determined by the FluostarOmega-Reader (BMG LABTECH, Offenburg, Germany) with 10 s integration time per well. Specific lysis of CD8 and or TCR positive cells was quantified by loss of FLuc signal and normalized to untransduced NK92 according to the following equation: specific lysis [%]=(killing by NK92 CD3−killing by TCR.sup.+ NK92)/killing by NK92 CD3×100. General lysis was determined as follows: general lysis [%]=100*(RLU of targets without effector cells−RLU of targets cocultured with TCR.sup.+ NK92)/(RLU of targets without effector cells RLU of maximal lysis control).

Statistics

[0093] Statistical data analysis was conducted with Graph Pad Prism Software using two-way ANOVA or multiple t-test using the Bonferroni-Dunn method. P<0.05 was considered statistically significant. Mean values and standard deviations (SD) were calculated from at least 2 independent experiments.