Expansion of allogeneic NK cells with efficient antibody-dependent cell cytotoxicity against multiple tumors.

Monoclonal antibodies (mAbs) have significantly improved the treatment of certain cancers. However, in general mAbs alone have limited therapeutic activity. One of their main mechanisms of action is to induce antibody-dependent cell-mediated cytotoxicity (ADCC), which is mediated by natural killer (NK) cells. Unfortunately, most cancer patients have severe immune dysfunctions affecting NK activity. This can be circumvented by the injection of allogeneic, expanded NK cells, which is safe. Nevertheless, despite their strong cytolytic potential against different tumors, clinical results have been poor.
Methods: We combined allogeneic NK cells and mAbs to improve cancer treatment. We generated expanded NK cells (e-NK) with strong in vitro and in vivo ADCC responses against different tumors and using different therapeutic mAbs, namely rituximab, obinutuzumab, daratumumab, cetuximab and trastuzumab.
Results: Remarkably, e-NK cells can be stored frozen and, after thawing, armed with mAbs. They mediate ADCC through degranulation-dependent and -independent mechanisms. Furthermore, they overcome certain anti-apoptotic mechanisms found in leukemic cells.
Conclusion: We have established a new protocol for activation/expansion of NK cells with high ADCC activity. The use of mAbs in combination with e-NK cells could potentially improve cancer treatment.

Introduction

Recent progress in cancer treatment is primarily related to the development of novel targeted therapies 1. These require the identification of suitable targets that are mainly expressed by the tumor cell population and/or playing a critical role in neoplastic cell growth. Therapeutic monoclonal antibodies (mAbs) particularly illustrate this concept. Indeed, rituximab (RTX), an IgG1 mAb directed against CD20 antigen, has now become the treatment of choice for most B-cell chronic lymphocytic leukemias (B-CLL) and B-cell non-Hodgkin's lymphomas (B-NHL). The combination of RTX with conventional chemotherapy has shown better efficacy in randomized clinical trials. Similar success has been found with other cytotoxic mAbs, such as trastuzumab in breast cancer or cetuximab in colorectal carcinoma and squamous cell carcinoma of the head and neck 2, 3. Nevertheless, mAbs alone generally have modest clinical activity. For example, the anti-CD20 mAbs RTX and obinutuzumab (OBZ; previously GA101, Roche, Genentech), when used as monotherapy in patients with relapsed follicular lymphoma (FL), have only led to short progression-free survival (PFS) 4. Thus, there is a need to optimize their use in combination therapy.

RTX success is related to its capacity to induce Fc-antibody-dependent cell-mediated cytotoxicity (ADCC). One receptor for human IgG1 is FcγRIIIa (CD16a), which is expressed on natural killer (NK) cells and macrophages. The link between FcγRIIIa-158VF polymorphism and RTX clinical responses strongly suggests that ADCC is critical 5. This polymorphism is located on the extra-cellular domain of FcγRIIIa, and amino-acid 158 is involved in the interaction with CH2 of human IgG1 4. Human IgG1 has a higher affinity for VV-NK cells compared to FF-NK cells 5. Based on these observations, there has been an attempt to produce new anti-CD20 mAbs by either Fc mutations or by glycoengeenering that exhibit higher affinity for FcγRIIIa 4, 6. Lowering the fucose content of the N-glycan is currently under clinical investigation in B-cell malignancies with the mAb OBZ, which shows stronger ADCC in vitro and in a lymphoma xenograft mouse model relative to RTX. It also demonstrated improved clinical activity for treating B-CLL and other B-cell malignancies 4. OBZ is approved for first-line B-CLL in association with chlorambucil, and in combination with bendamustine for the treatment of patients with FL who relapse or are refractory to a RTX-containing regimen 4. Initial results show that lenalidomide, which stimulates NK cell activity 7, activates NK cells in OBZ-treated patients8.

NK cells mediate ADCC but also possess natural cytotoxicity, which is mediated by engagement of their natural cytotoxicity receptors (NCRs). These play a central role in triggering NK activation. In humans, NKp30, NKp46, and NKp80 are constitutively expressed on resting and activated NK cells 9. The NK cell-activating receptor CD16 mediates ADCC. Hematological cancer patients possess antitumor NK cells that are unable to control disease 10, 11. Notably, blood-borne tumor cells use different mechanisms for immune escape 12, 13, e.g., by inducing NK cell dysfunction 7, 14. This mechanism has also been observed in a variety of patients of solid tumors 3. In addition, NK cell differentiation may be inhibited by the presence of tumor cells, e.g., acute myeloid leukemia (AML) cells infiltrating bone marrow 15, 16. Therefore, the failure of mAbs in monotherapy could be related to impaired NK cell function. Hence, there is a clinical interest to reactivate or replace patient NK cells 17. Clinical-grade production of allogeneic NK cells is efficient and NK cell-mediated therapy after hematopoietic stem cell transplantation (HSCT) seems safe 16, 18, 19. Despite the strong cytolytic potential of expanded NK cells against different tumors, clinical results have been very limited 16, 18, 19.

The combination of allogeneic NK cells with mAb could improve cancer treatment by replacing the defective effector immune cells. In addition, mAbs would effectively guide these effectors to their tumor targets. Several groups have tried this combination with varying results that could be due to deficient CD16 expression or lack of proper activation of expanded NK 20-23. In addition, these studies did not include a systematic evaluation of the effect of these cells in combination with several mAbs on different tumors, nor did they include primary tumor cells.

The aim of this work was to generate allogeneic NK cells with strong ADCC response against different tumors and mediated by different therapeutic mAbs. In addition, NK cell production should be easily scaled up and developed with good manufacturing practices (GMP). We have produced umbilical cord blood (UCB)-derived NK cells because UCB are rapidly available, present low risk of viral transmission and have less strict requirements for HLA matching and lower risk of graft-versus-host disease (GvHD) 18. For NK cell expansion we used Epstein-Barr virus (EBV)-transformed lymphoblastoid B cell lines as accessory cells, which induce a unique genetic reprogramming of NK cells 24. This generates effectors that overcome the anti-apoptotic mechanism of leukemic cells 25 and that are able to eliminate tumor cells from patients with poor prognosis 26. We show that NK cells obtained with our protocol are able to perform ADCC in vitro and in vivo. The ADCC response was induced by using different therapeutic antibodies and against multiple target cells.

Methods

Ethics statement

Experimental procedures were conducted according to the European guidelines for animal welfare (2010/63/EU). Protocols were approved by the Animal Care and Use Committee “Languedoc-Roussillon” (approval number: CEEA-LR-12163). The use of human specimens for scientific purposes was approved by the French National Ethics Committee. All methods were carried out in accordance with the approved guidelines and regulations of this committee. Written informed consent was obtained from each patient prior to surgery.

Chemicals

The D1D2 peptide has been previously described 27. IL-2 and IL-15 were obtained from Miltenyi Biotec. To produce deglycosylated cetuximab, a commercial cetuximab solution was treated overnight with PNGAseF (Promega) at 37 °C in 50 mM sodium bicarbonate buffer (pH 7.8) at 125 U/mg of cetuximab. Deglycosilated cetuximab was purified by gel filtration using a Sephadex 75 column in PBS, and sterilized by filtration. The AF 647 Goat F(AB')2 anti human IgG (H+L) min x (BOV, HRS, MS) was from Interchim.

B-CLL patients

Data and samples from patients were collected at the Clinical Hematology Department of the CHU Montpellier, France, after patients' written consent and following French regulations. Patients were enrolled in two independent clinical programs approved by the “Comités de Protection des Personnes Sud Méditerranée I”: ref 1324 and HEMODIAG_2020 (ID-RCB: 2011-A00924-37). Samples were collected at diagnosis and kept by the CHU Montpellier 11, 28. For analysis, peripheral blood mononuclear cells (PBMCs) were obtained by ficoll gradient and stored frozen in liquid nitrogen until use. The percentage of CD19+CD5+ cells was always higher than 90%. Other samples from hematological cancer patients were obtained from the same collections.

Cell lines

The (EBV)-transformed lymphoblastoid B cell line PLH, the hematopoietic cell lines HL60 and MV4-11 (acute myeloid leukemia), Daudi and Raji (Burkitt's lymphoma derived, CD20+), K562 (erythroleukemia), MM.1S and U266 (multiple myeloma), MEC1 (B-chronic lymphocytic leukemia) and its variants MEC-1-BCL-XL and MEC-1-MCL-1 were cultured in 10% FBS RPMI medium. The adherent cell lines Calu-1 (lung cancer), A549 (lung adenocarcinoma), SK-OV-3 (ovarian carcinoma) and SKBR3 (breast adenocarcinoma) were cultured in 10% FBS DMEM medium.

PBMC and UCBMC purification

UCBs were obtained from healthy donors from the CHU Montpellier. Prof. John de Vos is the responsible of the “Collection du Centre de Resources Biologiques du CHU de Montpellier” - http://www.chu-montpellier.fr/fr/plateformes (Identifiant BIOBANQUES - BB-0033-00031). PBMC and UCB mononuclear cells (UCBMC) were respectively collected from peripheral blood samples and UCB units using Histopaque -1077 (Sigma). Briefly, 13 mL Histopaque was added to 50 mL centrifugation tubes and 30 mL of 1/2 diluted blood in RPMI, (Invitrogen) was slowly added at the top. Tubes were centrifuged at 400 rcf for 30 min at 20 °C without brake. Mononuclear cells were collected from the interlayer white ring, washed in RPMI and suspended in RPMI medium supplemented with 10% FBS (Invitrogen).

Isolation and activation of human NK cells

Frozen UCBMCs were depleted of T cells by using EasySepTM CD3 Positive Selection Kit (STEMCELL technologies). Cells were cultured for 10 to 20 days with γ-irradiated PLH cells at 1:1 NK cell:accessory cell ratio in the presence of IL-2 (100 U/mL) and IL-15 (5 ng/mL), or with IL-2 alone (1000 U/mL). PLH cells were added every four days and fresh cytokines every two days. At the end of the process, NK cell purity (CD56+/CD3-) was always higher than 90%.

FACS analysis

For phenotype analysis, cells were stained with 7AAD (Beckman) to identify viable cells and antibodies against the surface markers CD25-FITC, CD45RO-FITC, CD69-PE, CD62L-PE, CD19-PE, CD3-PE, CD19-ECD, CD56-PECy7, CD56-APC, CD3-APC, CD45-APCAlexaFluor750, CD45RA-APCAlexaFluor750, CD16-PacificBlue, CD57-PacificBlue, CD45-KromeOrange, CD16-KromeOrange (Beckman), CD158b-FITC, CD158a-PE, CD107a-HV500 (BD Biosciences), CD158e-Vioblue (Miltenyi). 1×105-3×105 cells were incubated for 20-30 min at 4 °C with different antibodies in PBS containing 2.5% FBS. Cells were then washed and suspended in 200-250 µL of the same media. Staining was analyzed on a Gallios flow cytometer (Beckman) using the Kaluza software. Alive lymphocytes were gated using FSC/SSC and 7AAD staining. B lymphocytes (CD19+), T lymphocytes (CD3+CD56-) and NK cells (CD56+CD3-) were distinguished using, respectively, CD19, CD3 and CD56 antibodies.

NK cell-mediated cytotoxicity

Fresh or frozen (stored in liquid nitrogen) NK cells were labeled with 3 µM of CellTracker™ Violet BMQC Dye (Life Technologies) and incubated overnight with target cells at different E:T ratios. Subsequently, phosphatidylserine (PS) translocation and membrane damage were analyzed in the violet fluorescence-negative target cell population by flow cytometry using Annexin V-FITC (Immunostep) and 7AAD (BD Biosciencies) or propidium iodide (PI) as previously described 25, 29. We consider all cells positive for annexin-V and/or PI (or 7-ADD) as dead (or dying).

In ADCC experiments, we incubated target cells with the relevant antibodies (RTX and OBZ at 10 µg/mL; daratumumab, cetixumab and trastuzumab at 5 µg/mL) for 30 min at 37 °C. To arm NK cells, we incubated them at the same concentration of antibodies before washing and incubation with target cells. EGTA was used at 1 mM to block the granular exocytosis pathway and MgCl2 at 1.5 mM to maintain the osmotic pressure.

We used the tetrazolium dye MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazomium bromide) to determinate cellular viability. We added 10 μL of MTT (5 mg/mL) to the adherent cells (100 μL of medium after 2 washes with PBS) and incubated for 1 h at 37 °C, then added 100 μL of 0.05 M HCl in isopropanol to dissolve the crystals and quantified absorption at 550 nm using a spectrophotometer.

In all experiments, we calculated the basal cell death in the absence and presence of the different mAbs. These values were subtracted from those obtained after NK cell or NK cell+mAb treatments to generate the specific natural cytotoxicity or specific ADCC, respectively. All mAbs gave very low levels (<3%) of cytotoxicity in presence of heat-inactivated serum media.

Evaluation of RTX-armed e-NK

e-NK cells (2×105) were incubated for 1 h with 10 µg/mL RTX, washed and incubated with 1:800 solution of a goat F(ab')2 anti-human IgG (H+L) for 30 min at 4 °C. After incubation, NK cells were washed with PBS and RTX binding was analyzed by FACs. As a control, cells were only stained with the goat F(ab')2 anti-human IgG.

NK degranulation assay

Briefly, 50×103 target cells per well were placed in RPMI, 10% FBS, IL-2 100 U/mL with monensin (BD Biosciences) in a 96-well V-bottom plate. NK and target cells were incubated overnight at 37°C in 5% CO2 and living cells were counted using a Muse cytometer (Millipore) with the count and viability kit (Millipore). As a control, NK cells were incubated without targets. CD107a+ NK cells were analyzed on a Gallios flow cytometer (Beckman Coulter) using 7AAD, CD45RO-FITC, CD19-PE, CD56-PECy7, CD3-APC, CD45RA-APCAlexaFluor750, CD16-KromeOrange and CD107a-HV500 (BD Biosciences). Results were analyzed using Kaluza software.

In vivo experiments

In vivo experiments were carried out using 6-8-week-old male NOD scid gamma (NSG) mice. Mice were bred and housed in pathogen-free conditions in the animal facility of the European Institute of Oncology-Italian Foundation for Cancer Research (FIRC), Institute of Molecular Oncology (Milan, Italy). For engraftment of human cells, mice were subcutaneously engrafted with 5×106 BCL-P2 or 10×106 LNH1 primary tumor cells derived from a B-cell lymphoma (BCL) patient (BCL P2) or a diffuse large B-cell lymphoma (DLBCL) patient (LNH1). At day 4, we engrafted 15 (BCL-P2) or 10 (LNH1) million e-NK cells and at day 6, mice were treated i.p. with RTX (in saline medium) 3 mg/kg once a week for 3 weeks; or with a combination of both treatments e-NK and RTX. Tumor growth was monitored at least once a week using a digital caliper, and tumor volume was calculated according to the formula: L × W2/2 (mm3), where W represents the width and L the length of the tumor mass.

Statistical analysis

Experimental figures and statistical analysis were performed using GraphPad Prism (v6.0). All statistical values are presented as * p<0.05; ** p<0.01; *** p<0.001 and **** p<0.0001. Mean values are expressed as mean plus or minus the standard error of the mean (SEM).

Results

Costimulation with the EBV lymphoblastoid PLH cell line more efficiently expands UCB NK cells for clinical use than IL-2 stimulation

Cytokines and encounter with target cells induce dissimilar gene expression on NK cells 24. We used umbilical cord blood (UCB) cells and compared two NK cell activation/expansion protocols: one using a high dose of the cytokine IL-2 (1000 U/mL) and the other using cell costimulation. The costimulation protocol was performed with the EBV cell line PLH together with low concentrations of IL-2 (100 U/mL) and IL-15 (5 ng/mL) 9. NK cell expansion is jeopardized by T cells, therefore we depleted them from UCB before expansion. NK cell cultures underwent massive cell death at day 12 in IL-2-driven expansion (data not shown). So, we compared different parameters that reflect NK activity at day 10:

Proliferation. Costimulation-driven expansion was considerably more efficient (Figure ).

Cytotoxicity. IL-2-driven expansion led to NK cells with superior natural cytotoxicity against K562 (Figure ), Daudi (Figure ) and primary CD20+ B cell lymphoma cells (BCL P2; Figure ). Moreover, these NK cells also showed higher ADCC activity with RTX (Figure and Figure ). However, natural cytotoxicity gradually increased in the costimulation protocol when cells were activated for longer periods of time (Figure ). This correlated with a notable large-scale expansion of cells (median ± SD, IL-2 10 d (16.8 ± 22.2), costimulation 10 d (140.5 ± 235.8) and costimulation 20 d (260.9 ± 141.2), n=10).

Activation markers. Both protocols increased the expression of the activation marker CD69 (Figure ) and decreased CD45RA expression to that of CD45RAdim cells (Figure ). This was associated with an increase in the activation marker CD45RO, as previously published 10. Costimulation maintained higher CD16 expression (Figure ).

Exhaustion markers. We investigated the expression of two markers that could suppress NK cell-mediated cytotoxicity: TIM-3 30 and PD-1 31. While both protocols did not affect their expression, the mean fluorescence intensity (MFI) of positive populations tended to increase (Figure ). This probably reflects that, after long activation, some NK cells become exhausted.

Maturation and homing markers. UCB NK cells showed a low percentage of CD62L+ cells (18.1% ± 6.7%, n=3) that increased 10 d after IL-2 treatment (68.2% ± 13.5%) and costimulation (56.7% ± 11.9%). However, at day 20 post-costimulation, CD62L expression was lost (1.6% ± 0.4). IL-2-stimulated cells did not survive this long; so, we could not measure CD62L levels. CD62L is a "homing receptor" facilitating naive lymphocytes to enter secondary lymphoid tissues. Mature NK cells express low CD62L, which favors their peripheral trafficking 32.

In agreement with others 33, few naïve UCB NK cells expressed CD57 (1.2% ± 1.3%, n=3). IL-2 stimulation barely increased expression (7.3% ± 3.9%) and costimulation did not change it (0.6% ± 0.6). Longer costimulation, i.e., 20 days, also had no effect (1.6% ± 0.5%). The lack of CD57 expression did not impair NK cell cytolytic activity (see below).

In summary, costimulation led to higher numbers of activated and functional NK cells with higher CD16 expression. This prompted us to only use costimulation for the next experiments. On the other hand, IL-2 induced higher and faster cytotoxicity and could be the best option for autologous NK cell grafts.

Frozen/thawed NK cells keep their cytolytic activity

For clinical purposes, it would be advantageous to have a bank of cryopreserved expanded NK cells ready to use 34. Compared with fresh expanded NK cells, frozen/thawed NK cells lost roughly 35% of CD16 expression and 50% of their cytolytic activity (Figure ). As shown in Figure , 20 day-activation showed higher cytolytic activity than shorter expansions. For the next experiments, we used 20-21 days costimulation-induced expansion of UCB-derived NK cells containing more than 90% of NK cells that were kept frozen until use. Hereafter, we call them e-NK.

e-NK cells mediate ADCC against target cells with diverse CD20 levels

We observed that e-NK performed ADCC similarly on Raji and Daudi cells, which express high CD20 levels (Figure ), as on primary tumor cells, which express low levels (Figure ). Even though P2 cells probably express more CD20 than P148, they were slightly less sensitive to RTX-mediated ADCC. In fact, e-NK performed ADCC even if their natural cytotoxicity against some patient cells was low or absent (Figure and Figure ). Hence, e-NK show strong ADCC with RTX independently of their natural cytotoxicity and with lower variability (Figure ). The glycoengineered mAb OBZ 4, 35 induced higher ADCC than RTX (Figure and Figure ).

e-NK cells can be “armed” with mAbs to facilitate treatment

We next used e-NK cells coated with anti-CD20 mAb (“armed” e-NK cells) as an alternative to antibody-coated target cells. “CD20-armed” e-NK showed similar results to opsonizing tumor cells with anti-CD20 (Figure ). The presence of RTX after e-NK “arming” was visualized by using a fluorescent anti-IgG (Figure ). “OBZ-armed” e-NK also efficiently generated ADCC (Figure ).

Statistical analysis of several e-NK productions on cells from 5 CLL patients did not show any differences between opsonizing targets or “arming” e-NK (Figure ). Moreover, the analysis of these 4 e-NK expansions on the CLL targets showed that all productions could be armed (Figure ). Combining all these results statistically showed that significant ADCC was mediated by e-NK (Figure ).

Cytotoxicity requires degranulation and cell interaction by ICAM

Primary human NK cell cytotoxicity is largely independent of degranulation 36 and resides in death receptor binding on tumor cells by ligands expressed by NK cells. In agreement with this, e-NK natural cytotoxicity was only partially diminished by the degranulation inhibitor EGTA (Figure ), which in contrast, largely blocked ADCC.

The interaction of NK cell-expressed LFA-1 with its target cell ligand ICAM modulates NK cell cytotoxicity 37. Blocking this interaction with the D1D2 peptide 27 partly reduced natural cytotoxicity and almost completely abolished ADCC (Figure ). Therefore, our e-NK use similar mechanisms for eliminating target cells as primary human cells.

e-NK mediate ADCC with daratumumab

Next we tested if e-NK produced ADCC with the anti-CD38 daratumumab 38. We used 3 target cells that express CD38 (MM.1S, MV4-11 and BCL-P2; Figure ) and observed that three different e-NK preparations showed ADCC with daratumumab (Figure ). In contrast, daratumumab failed to induce ADCC against U266 that are negative for CD38 (Figure ). Several e-NK productions efficiently performed ADCC with daratumumab on MM.1S and P2 that was statistically significant (Figure ).

e-NK mediate ADCC with cetuximab

Next, we analyzed if e-NK cells mediate ADCC with other mAbs used to treat solid tumors. We used the cell lines Calu-1 and A549. Calu-1 cells express more epidermal growth factor receptor (EGFR) than A549 (data not shown). Both lines are targets of the anti-EGFR cetuximab, which has been proposed for use with adoptively-transferred expanded allogeneic NK cells in clinical trials for cervical cancer 39. In fact, in vitro and clinical data suggest that cetuximab mediates ADCC through NK cells 23, 34. We observed a relatively large variation in the natural cytotoxicity of the different e-NK donors versus these target cells. The decrease in cell viability, as measured by MTT formation, was low (Figure ). However, the increase in cell death, measured by annexin-V / 7-ADD staining, was higher. This showed that e-NK had induced the initial steps of apoptosis (annexin-V staining), but longer times were required to evaluate cell viability with MTT. Cetuximab increased early apoptosis and accelerated the process of cell death, decreasing viability. Several e-NK productions efficiently performed ADCC with cetuximab on CALU-1 and A549 that was statistically significant (Figure ).

EGTA diminished ADCC but insignificantly (Figure ). This suggested that the mechanism of action only partly involved degranulation, indicating a possible participation of death ligand-induced apoptosis.

e-NK mediate ADCC with trastuzumab

We next tested the anti-HER2 mAb trastuzumab on SK-BR-3 cells, which express high HER2, and A549 cells, which express low levels. Under this condition, e-NK performed ADCC in both cell lines by decreasing viability or increasing apoptosis (Figure ). Natural cytotoxicity, as well as ADCC, heavily depended on degranulation because EGTA largely decreased both (Figure ). Statistical analysis of several e-NK productions on SK-BR-3 and A549 showed that cell viability was significantly reduced and apoptosis tended to increase, although this was not statistically significant. This suggested that e-NK efficiently performed ADCC with trastuzumab (Figure ). However, the increase in apoptosis was not statistically significant. Finally, we extended this study to the ovarian cell line SK-OV-3 that was resistant to both natural cytotoxicity and ADCC during short treatment (data not shown). A 6-day treatment revealed that NK cells, mainly with trastuzumab, efficiently killed these cells (Figure ).

e-NK mediated ADCC in vivo

We next evaluated e-NK activity in vivo by engrafting primary tumor cells from a B-cell lymphoma (BCL) patient (P2) or from a diffuse large B-cell lymphoma (DLBCL) patient (LNH1) into NSG mice. NSG mice have a complete null mutation by knockout of the γ chain of the interleukin 2 receptor (Il2rγ), which is a common component of the cell surface receptors for several cytokines, including IL-2 and IL-15. Therefore, the signaling pathways for these cytokines are blocked in Il2rγ knockout mice. They should lack functional NK cells, which require IL15 signaling to develop. Four days later, mice were engrafted with e-NK and, 2 days later, treated with RTX; the latter decreased tumor growth (Figure ), showing that RTX possesses direct, non-effector functions independently of NK cell-mediated effects. While e-NK also decreased tumor growth (Figure ), co-treatment was more effective, protecting all mice from BCL cells and 4 out of 5 mice from DLBCL cells.

e-NK cells showed ADCC against chemoresistant cells

EBV-activated NK cells overcome anti-apoptotic mechanisms active in chemoresistant cells 25, 26. Overexpression of BCL-XL and MCL1 are common features of several hematological cancers 40. Jurkat cells over-expressing BCL-XL are resistant to doxorubicin and to soluble TRAIL 41. MCL1 over-expression protects them from ibrutinib cytotoxicity 42. e-NK killed the CD20+ B-CLL cell line MEC-1 overexpressing BCL-XL or MCL1, neither of which conferred protection against ADCC compared to wild type cells (Figure ). Several e-NK productions kill chemoresistant cell lines to statistically significant levels (Figure ). However, BCL-XL overexpression significantly decreases ADCC at high E:T ratios. This suggests that chemoresistance could partially protect tumor cells from e-NK-mediated ADCC, but not from e-NK natural cytotoxicity.

Discussion

Clinical mAbs fail to improve prognosis in a large number of patients. This could be due to the impairment of NK cells observed in most cancer patients 3, 16, 18. Therefore, restoring this immune function should improve mAb clinical benefits. We focused on developing a protocol to obtain NK cells in sufficient number and with high ADCC activity together with different therapeutic mAbs. From UCB-derived NK, we produced e-NK that only partially lost ADCC function after cryopreservation (Figure ) and preserved ADCC in vivo (Figure ). e-NK do not require relatively high Ag levels to perform ADCC, since they were effective with different cell lines expressing variable Ag levels.

The coupling of mAbs and e-NK should synergize in several clinical contexts. First, e-NKs should bypass NK dysfunction by increasing mAb-induced ADCC in patients with immune defects. Second, the clinical activity of NK cells is uncertain in solid cancers 16, 18. Probably these effectors scarcely recognize solid tumor targets in vivo and/or fail to infiltrate these tumors—e.g., NK have been detected in the tumor stroma but not within the tumor lesion in some cases 43-45. Moreover, the adoptive transfer of autologous NK cells in patients as single therapy maintained high levels of circulating NK cells but did not mediate tumor regression 3, 18, 46. mAbs should recruit e-NK to the selected targets and can also facilitate target elimination by favoring the recognition of opsonized cells. In fact, haploidentical NK cells combined with anti-GD2 mAb therapy has shown promising antitumor activity in pediatric recurrent/refractory neuroblastoma 47, 48. Third, e-NKs overcome anti-apoptotic mechanisms active in leukemic cells (Figure and 25), allowing elimination of tumor cells from patients with poor prognosis 26.

e-NK could also have anti-tumor activity per se. First, high numbers of tumor-infiltrating NK correlates with a better prognosis in some tumors 16. Second, NK are the first lymphocytes to recover after HSCT including after umbilical cord blood transplantation (UCBT). The speed of recovery correlates with the prognosis 18. In spite of these findings, NK cell adoptive immunotherapy has provided clinical benefit. Perhaps current expansion protocols fail to produce enough NK cells to support clinical success or generate cells with impaired activity 16. An inconvenience of engraftment of allogeneic expanded NK cell is their low survival in vivo 49. The persistence of ex vivo haploidentical IL-2-activated and -expanded NK-DLIs reaches a maximum of 7 days in lymphoma patients 50. This leaves grafted NK cells little time to eliminate their targets. The advantage is that NK will be less likely to generate the clinical problems found with CAR-T cells, which produce some chronic effects related to their long-term persistence (51; http://www.medscape.com/viewarticle/876591). One of the main concerns in using allogeneic immune cells is the incidence of GVHD. Allogeneic NK cells infusion is well tolerated in cancer patients 3, 18 and the severity of aGVHD correlates with impaired reconstitution of the NK cell compartment 52. To our knowledge, engraftment of NK cells has been linked to GVHD only when combined to HLA-matched, T-cell-depleted nonmyeloablative peripheral blood stem cell transplantation 53. NK cells likely contributed to GVHD in this setting by augmenting underlying T-cell alloreactivity 53.

An interesting alternative to allogeneic NK is KIR/KIRL blocking antibodies that activate endogenous NK cells 54. This approach has the inconvenience that cancer patient NK cells are hyporeactive 3, 16, 18, suggesting that they are too inefficient to totally eliminate tumors. Moreover, recent clinical data suggest that such antibodies modify the endogenous NK repertoire. This would make KIR-expressing NK cells, which are those with higher cytolytic activity, hyporeactive 55. This raises concerns about the clinical use of these antibodies. There are other ways to activate endogenous NK cells, such as the use of lenalidomide (LEN) 7, 8. Preliminary results from the Phase Ib/II clinical trial GALEN suggest that LEN could facilitate OBZ-mediated NK cell activation 8, as was observed with RTX 56.

In collaboration with the University Hospital of Montpellier, we wish to test the clinical efficiency of e-NK in lymphoma patients resistant to standard treatments, including RTX.

Supplementary figures.

Click here for additional data file.