Good manufacturing practice production of an off-the-shelf CD34⁺ progenitor-derived NK cell product with preserved anti-tumor functionality post-infusion in NOD/SCID/IL2Rg^null mice

Abstract

The adoptive transfer of natural killer (NK) cells represents a promising cancer therapy due to their intrinsic ability to distinguish between malignant and healthy cells in an allogeneic context, enabling off-the-shelf manufacturing possibilities. On demand availability of cryopreserved advanced therapy medicinal products (ATMPs) could promote enrolment in clinical trials and eventually commercialization with repeated dosing possibilities. However, NK cells are considered highly sensitive to cryopreservation-induced defects, including impaired viability, anti-tumor cytotoxicity, and in vivo expansion capacity. Here, we present the GMP-compliant manufacturing of an off-the-shelf NK cell product (RNK001), derived ex vivo from CD34⁺ hematopoietic stem and progenitor cells (HSPCs). To facilitate scalability and reduce hands-on time, the production process was adapted to a G-rex bioreactor, yielding high numbers of a pure CD56⁺CD3^- NK cell product. Cryopreservation of HSPC-NK cells using an optimized freeze-thaw protocol resulted in a consistently high post-thawing viability of mature and differentiated cells. Surviving HSPC-NK cells post-thawing exhibited enhanced proliferative capacity compared to fresh cells in vitro and their persistence in vivo was similar to that of fresh cells when administrated intravenously or intraperitoneally in NOD/SCID/IL2Rg^null mice. Moreover, cryopreserved HSPC-NK cells had robust anti-tumor functionality, efficiently killing tumor spheroids embedded in a 3D collagen matrix and maintaining degranulation and interferon-γ production capacity comparable to fresh cells following in vivo infusion. Together, these findings show the potential of cryopreserved HSPC-NK cells with potent effector functions, allowing the manufacturing of an off-the-shelf therapeutical NK cell product for hematological and solid malignancies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-025-05727-4.

Introduction

Natural killer (NK) cells are cytotoxic innate immune cells that play a crucial role in the first-line defense against viral infections and malignant cells. Unlike T cells, NK cells operate without prior antigen sensitization, and their activation is tightly controlled by differential signaling through activating and inhibitory receptors [1, 2]. Upon activation, NK cells exert several anti-tumor mechanisms, including the release of degranulating enzymes, receptor-mediated apoptosis, and the secretion of pro-inflammatory cytokines [1, 2]. Given their inherent anti-tumor potential and ability to distinguish healthy cells in an allogeneic context, the adoptive transfer of donor-derived NK cells is an attractive strategy for cancer treatment. Allogeneic NK cell therapies have demonstrated to be safe and potentially effective in controlling various hematological malignancies. Moreover, studies exploiting genetically engineered NK cells to enhance anti-tumor efficacy and counteract the immunosuppressive tumor microenvironment in hematological and solid cancers are emerging [3, 4].

The clinical translation of this promising advanced therapy medicinal product (ATMP) relies on its uniformity and wide-spread availability. This is highly dependent on the consistency and scalability of the manufacturing process, which have long posed major challenges in the field. Traditionally, peripheral blood (PB)-enriched NK cells have been used for NK cell therapy [3, 4]. To increase their numbers, these cells are expanded ex vivo using cytokine mixtures or feeder cell lines. However, PB-NK cell products often exhibit substantial donor-to-donor variability, variable expansion potential, and CD3⁺ T cell contamination, posing a risk for graft-versus-host disease [5, 6]. In this context, NK cells differentiated ex vivo from CD34⁺ hematopoietic stem and progenitor cells (HSPCs) present a promising alternative to PB-NK cells. Previously, we established a good manufacturing practice (GMP)-compliant cytokine-based culture protocol for ex vivo generation of a pure and mature NK cell product from umbilical cord blood (UCB)-derived HSPCs in a closed-system culture bag (named RNK001) [7, 8]. This manufacturing platform achieved a 1,000-fold expansion of HSPC-NK cells with excellent inter-donor consistency that demonstrated high anti-tumor potential against ovarian cancer (OC) and acute myeloid leukemia (AML) cells in vitro and in an OC xenograft mouse model [8–10]. Notably, our HSPC-NK cells have been examined in a phase I clinical trial in patients with recurrent OC (NCT03539406; manuscript in preparation) and are currently under investigation in a phase I/IIa clinical trial in patients with relapsed or refractory AML (NCT04347616), emphasizing their potential for clinical use [8, 11].

Next to the need for product consistency and process scalability, the clinical application of NK cell therapies would greatly benefit from the availability of off-the-shelf cryopreserved NK cell products. On demand availability would highly enhance accessibility for clinical trials, particularly for patients with rapidly progressing disease, facilitate centralized manufacturing and multi-site trials, support repeated dosing regimens, and simplify logistics of the manufacturing process. However, NK cells are considered hard-to-cryopreserve cells with impaired recovery, migrative capacity, and in vitro and in vivo anti-tumor cytotoxicity being reported post-thawing [12–14]. These cryopreservation defects appear to be influenced by the cell source and NK cell activation status [12, 15]; yet, research focusing on the cryopreservation of HSPC-NK cells using GMP-compliant reagents remains limited [12]. Domogala et al. reported a poor post-thawing recovery (mean 29%, range 9.6–62%) of NK cells derived from UCB-HSPCs in a feeder cell-based system [16]. Furthermore, studies on PB-NK cells, both using activated or expanded cells, often demonstrated minimal to no cytotoxicity directly after thawing, which could partly be recovered after resting in IL-2 supplemented media [13, 17–19]. Consequently, clinical trials have incorporated post-thawing re-culture periods prior to NK cell infusion, which undermines the advantages of an off-the-shelf product due to extended handling time, risk for product variation, increased costs, and the necessity for GMP manufacturing facilities at the administration site [20, 21]. Thus, the development of a cryopreservation protocol that ensures high viability and functionality of HSPC-NK cells post-thawing is crucial.

Here, we describe the scalability of our GMP-compliant HSPC-NK cell production platform using a bioreactor system which reduces hands-on time and aseptic handling steps. Additionally, we established a cryopreservation procedure for the high-fold recovery of HSPC-NK cells post-thawing. Surviving cryopreserved cells had a high proliferative capacity and were able to kill spheroids in a 3D tumor environment. Moreover, viable thawed HSPC-NK cells demonstrated high anti-tumor potency and resembled persistence of fresh cells upon in vivo infusion. This work underscores the potential of on-demand available cryopreserved HSPC-NK cells as a highly functional allogeneic cancer therapy.

Methods

Culture of tumor cell lines

The tumor cell lines K562, THP-1, and SKOV-3 (RRID: CVCL_0004, RRID: CVCL_0006, and RRID: CVCL 0532, respectively) were purchased from ATCC. OVCAR-4 (RRID: CVL_1627) was obtained from the DCTD Tumor Repository. K562 and THP-1 cells were cultured in Iscove’s Modified Dulbecco’s medium (IMDM; Gibco, 21980-32) supplemented with 10% heat-inactivated fetal calf serum (HI-FCS; Integro, 5-45900). SKOV-3 and OVCAR-4 were maintained in Roswell Park Memorial Institute 1640 medium (RPMI; Gibco, 21875-034) supplemented with 10% HI-FCS for SKOV-3 or 10% HI-FCS and 2mM L-glutamine (Gibco, 25030-024) for OVCAR-4 and passaged following trypsin (Life Technologies, 25300054) treatment. For formation of SKOV-3 spheroids, 30,000 cells were seeded in 1% agarose-coated round bottom 96-well plates (Corning, 3799) in DMEM/F12 (Invitrogen, 11320-033) supplemented with 10% HI-FCS and 1% penicillin/streptomycin (Life Technologies, 15140130). Three days post-seeding, spheroids were used for 3D killing assays. All cell lines were maintained at 37 °C and 5% CO2 and were tested to ensure they were mycoplasma-free every six months using the MycoAlert^® Mycoplasma Detection Kit (Lonza, LT07-418). Cells were maintained in culture for maximally three months. In functional assays comparing fresh and cryopreserved HSPC-NK cells, the passage number of target cell lines was kept identical.

HSPC-NK cell culture

UCB was collected at delivery in the Radboudumc following written informed consent (CMO 2014 − 226) or obtained from the Radboudumc cord blood bank (informed consent available). CD34⁺ HSPCs from cryopreserved UCB derived from the cord blood bank were isolated using the CliniMACS plus system (Miltenyi Biotec) and directly used for HSPC-NK cell culture afterwards. For UCB units obtained from caesarean sections or spontaneous deliveries, peripheral blood mononuclear cells were separated using Ficoll Paque, after which CD34⁺ HSPCs were magnetically separated using CD34 microbeads (Miltenyi Biotec, 200-070-065). Subsequently, HSPCs were cryopreserved until use. For quality control (QC), HSPC cell numbers and purity were assessed by staining for 7-AAD (1:1000), CD45, CD3, and CD34 by the Radboudumc Laboratory for Hematology Immunophenotyping unit (ISO15189 certified) using the Navios flow cytometer (Beckman Coulter) (Supplementary Table 1). RNK001 HSPC-NK cells were generated from UCB-derived CD34⁺ HSPCs in VueLife cell culture bags (Saint-Gobain, 118/750AC) or six-well plates (Corning, 3506) using NK MACS medium (Miltenyi Biotec, 130-114-429/170-076-356) supplemented with 10% human serum (HS; Sanquin Blood Supply Foundation) and cytokine cocktails as described previously [8]. For culture in scalable G-rex bioreactors, this protocol was adapted involving six standardized medium additions at day 9/10, 11/12, 14/15, 17/18/19, 24/25/26, and 28/29 (Fig. 1A). HSPCs were expanded for nine days in a G-rex 10 M(-CS) (Wilson Wolf, 80110(-CS)), after which the cells were transferred to a G-rex 100 M(-CS) (Wilson Wolf, 81100(-CS)). Fold expansion of the HSPC-NK cell product was assessed by trypan blue-based cell count at day 35. Purity of the product was checked by the Radboudumc Laboratory for Hematology Immunophenotyping unit using 7-AAD, CD3, CD19, and CD56 and was acquired on the Navios flow cytometer (Supplementary Table 1).

Fig. 1.

High post-thawing viability and proliferative capacity of G-rex expanded HSPC-NK cells cryopreserved using an optimized freezing and thawing protocol A Overview of the GMP-compliant generation of NK cells from HSPCs in a G-rex scalable bioreactor using six standardized medium additions. B Fold expansion at day 35. C Absolute number of CD34⁺ HSPCs at day 0 and CD56⁺ HSPC-NK cells at day 35. D Composition of the HSPC-NK cell product. E Trypan blue-based post-thawing viability of HSPC-NK cells frozen in cryobags. F Recovery of viable cells post-thawing compared to pre-freezing dosage of bag-frozen cells. G HSPC-NK cell dosage post-thawing. H Stability of thawed HSPC-NK cells in CryoStor CS10 for a 30-min timespan. I Post-thawing viability following twelve months of storage in the liquid nitrogen vapor phase, assessed using cryovials. Stability study ongoing till five years post-freezing. J Absolute number of viable fresh or cryopreserved HSPC-NK cells cultured up to seven days in 50 ng/ml IL-15. K Proliferation score, defined as the fold change reduction in MFI of Proliferation Dye eFluor450 compared to day zero, of fresh and cryopreserved HSPC-NK cells following seven days of culture in 50 ng/ml IL-15. Donor 4 (▼) involves a non-GMP compliant pre-validation run. For donor 3 (∎), t = 0 represents t = 15 (E-G) and t = 30 represents t = 45 (G) due to technical issues. Bars in B, D-G, K represent the mean. Significance was evaluated using the Wilcoxon test for H, Friedman test with Dunn’s multiple comparison correction for I, and a two-way ANOVA with Šídák multiple comparison correction for K; no statistical significances were found

HSPC-NK cell cryopreservation and thawing

Prior to cryopreservation, HSPC-NK cells manufactured in a GMP-compliant process were washed twice with CliniMACS PBS/EDTA buffer (Miltenyi Biotec, 200-070-025) with 0.5% Human Serum Albumin (HSA, Albuman, Prothya Biosolutions, RVG 103595) and resuspended in 0.9% NaCl (Baxter, RVG 27512) + 5% HSA. Harvested NK cells were centrifuged for 10–15 min at 500x g for 15 ml/50 ml or 250 ml tubes, respectively. Afterwards, cells were cryopreserved in either room temperature (RT) or 4 °C pre-cooled CryoStor CS10 freeze medium (BioLife Solutions, 210102) at a concentration of 10-100 × 10⁶ cells/ml in Cryotube vials (cryovials; ThermoFisher Scientific, 363401 or Greiner Bio-One, 126263) or CryoMACS freezing bags 50 (Miltenyi Biotech, 200-074-400). Afterwards, HSPC-NK cells were incubated for 10 min at RT or 4 °C, respectively, prior to freezing to -150 °C using the controlled-rate Kyro 1060 freezer (Planer, Middlesex, United Kingdom) according to the temperature gradient presented in Table 1. Subsequently, cells were stored in the liquid nitrogen vapor phase until use (stability study ongoing till five years post-freezing). For thawing of cryovials, the ThawSTAR^® CFT2 system (BioLife Solutions, Bothell, USA) was used resembling thawing in a 37 °C water bath until a small clump of ice remains, or cryovials were thawed in a 37 °C water bath for a standardized time of 10 min. Cryobags were thawed in a 37 °C water bath until no ice remained. Subsequently, the cell suspension was dropwise diluted ten-fold in 0.9% NaCl + 5% HSA in a 15 ml/250 ml tube. Afterwards, HSPC-NK cells were centrifuged at varying speed (125-500xg) and time (10–30 min) before use in subsequent assays (maximally within three hours post-thawing).

Table 1.

HSPC-NK cell cryopreservation profile using the Planer controlled-rate freezer

Temperature range (°C)	Temperature gradient (°C/min)
[6, -8]	-1
[-8, -15]	-10, followed by 5-min hold time
[-15, -40]	-2
[-40, -150]	-5

HSPC-NK cell recovery and viability assays

Short-term recovery of viable cells post-thawing or -centrifugation was assessed by trypan blue-based cell counts and compared to the pre-cryopreservation or -centrifugation dosage, respectively. Viability of (CD56⁺) HSPC-NK cells was assessed by trypan blue-based manual counting or flow cytometric staining for 7AAD^- and Apotracker Green^- (200 nM), and cells were acquired on the CytoFLEX flow cytometer (Beckman Coulter) (Supplementary Table 1). Long-term recovery effects were assessed by culturing fresh or cryopreserved HSPC-NK cells for seven days in NK MACS with 10% HS and 50 ng/ml IL-15 (Immunotools, 11340155) or 1,000 IU/ml IL-2 (Chiron Corporation, 53905-991-01), with addition of fresh medium on day four. Recovery compared to t = 0 post-thawing was assessed at day four and seven by comparing the absolute numbers of 7AAD^-CD56⁺ HSPC-NK cells following a no-wash staining for 20 min at 4 °C (Supplementary Table 1). Cells were acquired on the Gallios flow cytometer (Beckman Coulter).

Proliferation assay

Fresh or thawed HSPC-NK cells were labelled with 10 µM Cell Proliferation Dye eFluor450 (eBioScience, 65-0842-85), according to manufacturer’s instructions. Next, 100,000 CD56⁺ HSPC-NK cells/well were cultured in a 96-well round bottom plate in NK MACS with 10% HS and 50 ng/ml IL-15 or 1,000 IU/ml IL-2 for 7 days. Fresh medium was added at day four. At day zero, four, and seven, eFluor450 expression of 7AAD^-CD56⁺ HSPC-NK cells following a no-wash staining was determined using the Gallios flow cytometer (Supplementary Table 1). The proliferation score was defined as the fold change reduction in MFI of eFluor450 on day four/seven compared to day zero.

Microbiology

For GMP HSPC-NK process validation runs, HS was tested for mycoplasma according to Ph. Eur. 2.6.7 by Eurofins MicroSafe Laboratories. Other raw materials were animal- and human-component free or assessed as low risk and therefore not tested for mycoplasma. Microbiological control testing of CD34⁺ isolated HSPCs was performed using the Bactec culture and detection system (BD biosciences) by the Radboudumc department of Microbiology. After formulating the final HSPC-NK cell product on day 35, at least 2% of the total volume was tested for sterility according to Ph. Eur. 2.6.1. by Eurofins Bactimm. Additionally, HSPC-NK cell products were tested for endotoxin content according to Ph. Eur. 2.6.14 by the Radboudumc department of Pharmacy using the chromogenic kinetic method. This test was performed on cells in 0.9% NaCl + 5% HSA right before final formulation in CryoStor CS10 due to incompatibility of the assay with DMSO content. The acceptable endotoxin limit was established at < 1.0 EU/ml (< 5 EU/kg at 50 kg).

Flow cytometric analysis

HSPC-NK cells were phenotyped using the titrated antibodies listed in Supplementary Table 1. Viability dyes 7-AAD, eFluor780, and Sytox Blue were used at a final concentration of 1:1000, 1:4000, and 1:5000, respectively. Briefly, maximally 1 × 10⁶ cells were washed with PBS with 0.5% bovine serum albumin (FACS buffer). Cells were incubated for 20 min with antibodies in FACS buffer in the presence of 50 pg/ml Nanogam (total human IgG to block Fc receptors, Sanquin Bloodbank) except for the QC markers where no Nanogam was included. All stainings were performed at 4 °C except for mixes involving CD183, CD184, and CD62L which were incubated at 37 °C. Afterwards, cells were washed twice with FACS buffer and acquired on the Gallios (in vitro and ex vivo panels) or the Navios flow cytometer (QC panel). Mice blood and spleen samples were lysed with PharmLyse (BD, 555899) prior to staining. NK cells were gated based on CD45⁺CD56⁺CD3^- expression for the in vitro panel and QC panel. For the ex vivo panel, NK cells were gated based on mouse (m) CD45^- human (h) CD45⁺CD56⁺ expression. Marker expression gates were based on isotype (in vitro and ex vivo panels) or FMO (QC panel) controls, or on standard marker expression profiles. For in vivo experiments, isotype analysis was performed on cells pooled from all mice. Data was analyzed using Kaluza V2.2.0.

NK cell potency assay

Reactivity of fresh and cryopreserved HSPC-NK cells (used within three hours post-thawing) was assessed by four-hour stimulation of CD56⁺ NK cells with K562, THP-1, SKOV-3, or OVCAR-4 at an effector-to-target (E: T) ratio of 1.5:1 in the presence of CD107a. As control, HSPC-NK cells were stimulated with 25 ng/ml PMA (Sigma, P1585) and 1 µg/ml ionomycin (Sigma, I0634). Brefeldin A (BD Bioscience, 5550229) was added one hour after stimulation. Next, cells were stained extracellularly for CD56 and eFluor780. Following fixation and permeabilization with the Fixation/Permeabilization kit (eBioscience, 00–5521-00) according to manufacturer’s instructions, cells were stained intracellularly for Perforin and IFNγ in permeabilization buffer (eBioscience, 00–8333). Antibodies used in this assay are listed in Supplementary Table 1. Analysis of CD107a and IFNγ expression was performed within the viable CD56⁺Perforin⁺ cell population with unstimulated HSPC-NK cells as negative control. For ex vivo potency assays with HSPC-NK cells recovered during mouse experiments, the E: T ratio was based on total recovered nucleated cells. Additionally, mCD45 and hCD45 were included in the extracellular staining to identify human CD56⁺ HSPC-NK cells.

2D and 3D cytotoxicity assays

Fresh and cryopreserved HSPC-NK cells (used within three hours post-thawing) were labelled with 10 µM Cell Proliferation Dye eFluor450 following manufacturer’s instructions. For 2D killing assays, 30,000 SKOV-3, OVCAR-4, K562, or THP-1 cells/well were co-cultured in a 96-well flat bottom plate (Corning, 3506) with labelled CD56⁺ HSPC-NK cells at E: T ratios ranging from 0.6:1 to 10:1 for 16 h. SKOV-3 and OVCAR-4 cells were left to adhere for at least four hours prior to NK cell addition. Afterwards, cells were directly harvested in case of K562 and THP-1 cells or washed and trypsinized for SKOV-3 and OVCAR-4 cells. For 3D killing assays, supernatant of SKOV-3 spheroids (formed for three days) was replaced with 1.67 mg/ml bovine collagen solution (Purecol, Advanced BioMatrix, 5005) embedding 0.6 or 1.2 × 10⁶ CD56⁺ HSPC-NK cells/well. Following collagen polymerization at 37 °C for 30 min, IMDM + 10% HI-FCS was added and cells were incubated for 16 h. Afterwards, the collagen matrix was degraded by collagenase I (Sigma, C0130) treatment for 60 min at 37 °C followed by spheroid disruption using TrypLE Express Enzyme (ThermoFisher Scientific, 12605028) for 60 min at 37 °C. Following target cell harvesting in 2D and 3D assays, cells were resuspended in FACS buffer containing 7-AAD and counted using the Gallios flow cytometer (7-AAD^-eFluor450^- cells represent viable target cells). Target cell lysis was calculated as follows: (1 – (number of viable target cells after co-culture with NK cells/number of viable target cells in target alone condition)) x 100%.

Spontaneous migration assay

To remove dead cells and debris, viable fresh or cryopreserved HSPC-NK cells (used within three hours post-thawing) were separated using Ficoll Paque. Afterwards, cells were stained using the live/dead Cell Imaging kit (ThermoFisher Scientific, R37601), following manufacturer’s instructions. Viable cells were stained with Calcein AM and dead cells were labelled with BOBO-3 Iodide. Subsequently, cells were embedded in 1.67 mg/ml bovine collagen solution containing X-VIVO 15 medium (Lonza, BEBP02-061Q) + 10% FCS, suitable for imaging, and plated at 30,000 cells/well. Following collagen polymerization, X-VIVO 15 + 10% FCS was added, and cells were rested for two-three hours at 37 °C. Afterwards, cell migration was monitored for 800 s by fluorescent time-lapse microscopy using the Axiovert 200 M microscope (Zeiss) with Axiocam MRm camera (Zeiss) at 37 °C using the H301-K-FRAME incubation system (Okolab). Images were taken with 40 s intervals at two positions/well for six wells in total. Analyses were performed on Calcein AM⁺ viable HSPC-NK cells with the TrackMate plugin in Fiji (version 1.53t) using the LoG detector to detects spots ≤ 25 μm [22, 23]. Tracks were analyzed using the LAP tracker and filtering of tracks was based on a quality score of 5, ≥ 10 spots per track, a maximum linking distance between spots of 15 μm, and a maximum gap-close distance of 30 μm/2 frames. Additionally, spots not appearing as cells or tracks that could not be validated since they resided on the edge of the frame were manually excluded. Taken together, this resulted in 185–730 cells being analyzed. Based on the Trackmate analysis, the mean track speed and total distance travelled/40 sec, calculated by dividing the total travelled track distance by the cumulative number of spots and gap-close events in a track to correct for differences in timeframes of the tracks, was defined.

In vivo HSPC-NK cell persistence experiments

All animal experiments were approved by the Dutch central committee of animal experiments (DEC 2018-0029). Mice were monitored daily for behavior and general humane endpoints and weight was assessed twice a week. Female NOD-Prkdc^scid-IL2rg^Tm1/Rj mice (NXG, Janvier Labs) aged six to eight weeks were injected intravenously (i.v.) or intraperitoneally (i.p.) with fresh or cryopreserved HSPC-NK cells (infusion within three hours post-thawing) at a dosage of 20 × 10⁶ or 15 × 10⁶ viable cells, respectively. Starting at day zero, 1 µg IL-15 was administered every two days via subcutaneous (s.c.) injection in case of i.v. cell infusion or via i.p. injection in case of i.p. cell administration. For the i.v. experiment, tail vein blood collection was performed at day one and four. After seven days, mice were sacrificed and persistence of HSPC-NK cells in blood, collected via cardiac puncture, bone marrow, and spleen was determined for the i.v. experiment. For the i.p. experiment, peritoneal wash-outs and splenocytes were analyzed. For blood, bone marrow, and spleen samples erythrocyte lysis (PharmLyse) was performed prior to staining, peritoneal washings were stained directly. HSPC-NK cells were defined as 7AAD^-mCD45^-CD56⁺ (Supplementary Table 1). Engraftment was calculated as follows: (number of hCD56⁺ HSPC-NK cells/(total number of hCD56⁺ HSPC-NK cells and mouse CD45⁺ cells) x 100%. Additionally, recovered HSPC-NK cells were phenotyped (Supplementary Table 1) and potency of cells recovered from the peritoneal cavity (i.p. experiment) or spleen upon lympholyte isolation of mononuclear cells (Sanbio, CL5301) (i.v. experiment) was assessed as described previously. Mice effected by technical errors in animal or experimental procedures were excluded from the analysis.

Statistics

Statistical analysis was performed with Prism software version 10.4.0 (GraphPad, San Diego, CA, USA). The Mann-Whitney test, Kruskal-Wallis test, Wilcoxon test, or two-way ANOVA were used for assessing statistically significant differences, as indicated in the figure legends. For experiments with a sample size ≤ 4, the two-way ANOVA was performed in an exploratory manner. Significance was defined as P < 0.05 (*), P < 0.01 (**), or P < 0.001 (***).

Results

HSPC-NK cell manufacturing in a scalable closed-system G-rex bioreactor results in a high yield of a pure NK cell product

Clinical application and eventually commercialization of cell therapies benefit from GMP-compliant, scalable manufacturing processes with minimal hands-on time. Previously, we reported the GMP-compliant, five-week expansion process of RNK001 HSPC-NK cells in closed-system culture bags involving concentration-dependent medium additions every two to three days [8]. Here, we further optimized this manufacturing process by translation to a scalable G-rex bioreactor system with six standardized medium additions (Fig. 1A). In the first nine days, CD34⁺ HSPCs were expanded in a G-rex 10 M, followed by transfer to a G-rex 100 M to allow sufficient nutrient supply. This process was validated using one research-grade development run and three GMP-compliant manufacturing runs. The starting material obtained from different UCB units consistently met our pre-defined release criteria (Table 2). CD34⁺ HSPCs showed high-fold expansion into HSPC-NK cells (mean fold expansion 767, range 364–1014; Fig. 1B), resulting in 1.4–2.4 × 10⁹ CD56⁺CD3^- cells (Fig. 1C; Table 2). Viability of the products was high at the end of the culture (≥ 97%; Table 2). Additionally, pure, and differentiated products were obtained with a mean CD45⁺CD56⁺CD3^- population of 94% (Fig. 1D; Table 2) and a T and B cell content of < 5.0 × 10⁶ and < 15.0 × 10⁶ cells/dose, respectively (Table 2). Finally, test results for microbiological control of all three GMP process validation runs were below the defined limits (Table 2). Collectively, HSPC-NK cell product manufacturing in a G-rex yields GMP-compliant expansion and purity characteristics corresponding to the quality of bag-cultured cells while allowing scalability and reducing aseptic handling [8].

Table 2.

Test results of the cryopreserved RNK001 HSPC-NK cell product

	Starting material			Purity and viability
	#CD34+ cells (x106)	Viability (%)	Microbiology	CD56+CD3−dose (x109)+	Viability (%)	Purity (%)	T cell content/ dose (x106)	B cell content/ dose (x106)
Criteria	≥ 1.0	≥ 50%	No growth	1.0–3.0	≥ 70	≥ 70	< 5.0	< 15.0
Run 1	1.38	90	No growth	2.4	99	94	0.4	0.5
Run 2	3.94	89	No growth	1.4	99	94	0.2	0.2
Run 3	1.71	79	No growth	2.1	97	92	0	0
Research*	2.16	94	NA	2	99	97	0	0

	CD45 + CD56 + CD3 − cell phenotype						Microbiology
	NKG2A (%)	NKG2D (%)	DNAM-1 (%)	NKp46 (%)	NKp44 (%)	NKp30 (%)	Endotoxin (EU/ml)	Sterility
Criteria	> 30	> 30	> 30	> 30	> 30	> 30	< 1.0	Sterile
Run 1	85	100	91	94	81	100	< 1.0	Sterile
Run 2	50	100	90	85	92	99	< 1.0	Sterile
Run 3	70	96	89	72	83	91	< 1.0	Sterile
Research*	65	100	97	97	77	100	NA	NA

Purity (CD45⁺CD56⁺CD3⁻), viability, and microbiology status were assessed pre-cryopreservation; phenotype was defined post-cryopreservation

+ Dose is defined as 75 ml distributed over 3 bags of 25 ml

* Research-grade development run

NA: Not applicable

Development of a HSPC-NK cell cryopreservation protocol for high post-thawing recovery and viability

Next, a cryopreservation procedure for HSPC-NK cells was established to create an off-the-shelf product with high post-thawing recovery and viability. Cells were frozen in cryovials in CryoStor CS10 freezing medium, specialized for fragile cells and available clinical-grade, using a Planer controlled-rate freezer. Within this process, freezing and thawing methods were compared for optimal recovery of viable CD56⁺ cells and post-thawing viability, defined as the fraction of non-apoptotic (Apotracker^-) and non-necrotic (7AAD^-) cells (Supplementary Fig. 1A).

DMSO-containing freezing media are typically pre-cooled to minimize damaging exothermic reactions, but cell penetration of DMSO is faster at RT than 4 °C and may therefore be beneficial [24]. To investigate the optimal procedure, we compared cryopreservation using 4 °C pre-cooled or RT CryoStor CS10 but we found no significant difference in post-thawing recovery (Supplementary Fig. 1B). Moreover, high viability was maintained for up to two hours post-thawing in both conditions (Supplementary Fig. 1C). Next, as NK cell recovery has previously been reported to be influenced by freezing density [16], we evaluated cell concentrations of 10, 30, 50, or 100 million cells/vial. Post-thawing recovery and viability were consistently high across all concentrations (Supplementary Fig. 1D, E). To select an optimal thawing procedure, two approaches were compared; HSPC-NK cells were thawed until a small clump of ice remained using the ThawStar system (4 °C thaw) or for a standardized time of 10 min in a 37 °C water bath (37 °C thaw). Both methods yielded similar recovery and viability outcomes (Supplementary Fig. 1F, G). For clinical application, we have the possibility to directly infuse the thawed product without washing into patients as the 10% DMSO is diluted directly post-infusion. This approach is routinely applied for cryopreserved autologous stem cell transplants and CAR-T cell products [25, 26]. However, in vitro and murine studies require centrifugation to reduce volume and remove DMSO, which is toxic to mice at the concentration used. Importantly, we found stable post-centrifugation recovery of thawed HSPC-NK cells regardless of the centrifugation settings used (Supplementary Fig. 1H). These results demonstrate that cryopreservation of HSPC-NK cells in CryoStor CS10 using a controlled-rate freezer ensures high post-thawing recovery and viability, regardless of evaluated freezing and thawing parameters.

Validation of the GMP manufacturing and cryopreservation process shows high post-thawing viability and proliferative potential of thawed HSPC-NK cells

As the high recovery and viability of our cryopreserved HSPC-NK cell product was independent of the assessed freezing, thawing, and centrifugation parameters, the cryopreservation procedures were selected to ensure translatability to a GMP-compliant setting. We decided to cryopreserve HSPC-NK cells in a standardized volume of 75 ml at a dosage of 1.0–3.0 × 10⁹ nucleated cells, distributed evenly across three cryobags of 25 ml to provide options for repeated dosing. Although we observed no difference in post-thawing viability between pre-cooled and RT CryoStor CS10, pre-cooling was selected as preferred method. This way, we ensure product stability during the freezing procedure, which might potentially take more time for the GMP-compliant process than the previously evaluated research process. Thawing was set to follow standard procedures for thawing cell therapy products in a 37 °C water bath until a small clump of ice is left, thereby eliminating the need for specialized equipment. Finally, for in vitro and murine studies, thawed cells were to be centrifuged at 500x g for 10 min, which was identified as the minimum duration required for optimal cell recovery. To validate this fixed cryopreservation protocol, one development run and three GMP-compliant runs were conducted using G-rex cultured HSPC-NK cells.

The protocol demonstrated high robustness, achieving a mean cell viability of 67% (range 65–68, Fig. 1E) and a post-thawing recovery of 75% (range 67–84%, Fig. 1F) directly after thawing. This resulted in a viable CD56⁺CD3^- dosage of 1.0–1.7 × 10⁹ cells post-thawing (Fig. 1G). Viability was determined by Trypan Blue-based hand count, excluding necrotic events, since no apoptosis, defined by Apotracker⁺ events, was observed within two hours after thawing during the optimization phase (Supplementary Fig. 1A). Similar post-thawing outcomes were observed with cryovials, used for functionality and stability studies (Supplementary Fig. 2A, B). Importantly, cells maintained viable up to 30 min post-thawing in CryoStor CS10, emphasizing the feasibility of direct infusion following bedside thawing (Fig. 1H). Additionally, the product demonstrated long-term stability, as viability directly after thawing remained stable upon storage for twelve months in the liquid nitrogen vapor phase (Fig. 1I). Despite the high viability observed directly after thawing, cell numbers decreased compared to fresh cells when cultured for four days in IL-15 or IL-2 (Fig. 1J, Supplementary Fig. 2C). This delayed decrease in viability has previously been linked to intracellular leakage of granzyme B (GZMB) in PB-NK cells. Pre-treatment with IL-18 and IL-15 before cryopreservation was reported to lower intracellular GZMB levels, thereby preventing viability loss post-thawing [13]. However, we found no decrease in the GZMB percentage or median fluorescent intensity upon stimulation of fresh HSPC-NK cells with IL-18 or additionally IL-15 at the end of the culture (Supplementary Fig. 2D, E). Importantly, after four days of culture in IL-15 or IL-2, surviving cryopreserved HSPC-NK cells started to proliferate again (Fig. 1J, K Supplementary Fig. 2C, F, G). Although cell numbers of freshly cultured cells were not reached after seven days of culture (Fig. 1J, Supplementary Fig. 2C), surviving cryopreserved HSPC-NK cells exhibited enhanced proliferative capacity compared to freshly cultured cells, particularly when stimulated with IL-15 (Fig. 1K, Supplementary Fig. 2F, G). This suggests that the established cryopreservation protocol yields high post-thawing recovery of HSPC-NK cell with potent proliferation capacity.

Cryopreserved HSPC-NK cells maintain an activated phenotype and function and efficiently lyse tumor spheroids in a 3D environment

We next evaluated whether the cryopreservation procedure had impact on the phenotype and effector functions of surviving viable HSPC-NK cells directly after thawing (defined as a maximum of three hours post-thawing). Cryopreserved HSPC-NK cells exhibited a highly mature and differentiated phenotype, with no significant differences compared to freshly cultured cells (Fig. 2A, B, Supplementary Fig. 3). Moreover, GMP-release criteria of critical phenotypical markers were met for all GMP-compliant cultures (Table 2). Nevertheless, the activation markers CD25 and CD69 and the Fc receptor CD16, required for antibody-dependent cellular cytotoxicity, showed a trend towards decreased expression for the thawed HSPC-NK cell product (Fig. 2A, B). Notably, expression of the homing markers CD183 and CD184 remained stable, although CD62L exhibited a trend towards reduced expression after thawing (Fig. 2C). Perforin expression, crucial for NK cell anti-tumor potential, was comparable between fresh and cryopreserved cells (Fig. 2D). In line with the trend for decreased CD25 and CD69 expression, expression of the degranulation marker CD107a was slightly reduced on thawed HSPC-NK cells stimulated with K562 or the OC cell lines OVCAR-4 and SKOV-3 (not significantly) (Fig. 2E, F). Importantly, thawed cells had similar degranulation capacity compared to fresh cells upon challenge with the AML cell line THP-1. Furthermore, IFNγ production capacity retained relatively stable following cryopreservation (Fig. 2E, G), demonstrating the potential of cryopreserved HSPC-NK cells to exert their anti-tumor functions.

Fig. 2.

HSPC-NK cells maintain mature, activated, and potent against cancer cell lines upon cryopreservation Flow cytometric expression of A activation, B maturation, C homing, and D effector markers in fresh HSPC-NK cells or directly after thawing. E Representative example of CD107a and IFNγ expression of cryopreserved HSPC-NK cells without stimuli or upon four-hour stimulation with K562. F Degranulation capacity, defined by CD107a, and G IFNγ expression of fresh or cryopreserved CD56⁺ conventional (Perforin⁺) HSPC-NK cells upon four-hour stimulation with cancer cell lines at an E: T ratio of 1.5:1 or with PMA/ionomycin. All experiments were initiated within three hours after thawing. Bars in A-D, F, G represent the mean. Significance testing was performed using the Wilcoxon test for A-D (no significance was found) and a two-way ANNOVA with Šidák multiple comparison correction for F-G to evaluate the difference between fresh and cryopreserved cells. * P < 0.05, ** P < 0.01

The effector functions of the viable cryopreserved HSPC-NK cell product were further assessed by evaluating the lytic capacity against various tumor cell lines (Supplementary Fig. 4). After 16 h of co-culture with K562 and OVCAR-4, target lysis was lower for thawed HSPC-NK cells compared to fresh cells (Fig. 3A). Killing of SKOV-3 and THP-1 was retained by thawed HSPC-NK cells in one of three donors. Notably, lower viable HSPC-NK cell counts were observed at the end of the assay for the cryopreserved condition, indicating that the reduced killing capacity may, at least partially, be due to HSPC-NK cell death occurring during the assay (Fig. 3B). Previously, it has been reported that reduced post-thawing cytotoxicity of NK cells could also be attributed to reduced motility [14]. To explore this, we next evaluated the spontaneous migration potential of HSPC-NK cells embedded in a collagen matrix for 800 s using fluorescent live-cell imaging. Indeed, surviving cryopreserved cells exhibited a significantly reduced mean speed and travelled a shorter distance compared to freshly cultured cells (Fig. 3C, D, Supplementary Fig. 5). To investigate both migrative capacity and cytotoxicity in a complex 3D tumor environment and over a longer timeframe, SKOV-3 spheroids were embedded in a collagen matrix with either fresh or cryopreserved HSPC-NK cells and co-cultured for 16 h (Fig. 3E). Interestingly, despite significantly lower viable HSPC-NK cell counts at the end of the experiment in the cryopreserved condition, thawed cells exhibited comparable spheroid lysis to freshly cultured cells (Fig. 3F, G). Together, these findings indicate that HSPC-NK cells that remain viable post-thawing have a high cytotoxic potential in a 3D environment.

Fig. 3.

Cryopreserved HSPC-NK cells show impaired 2D tumor lysis and spontaneous migration capacity but maintain their killing capacity in a complex collagen-embedded SKOV-3 spheroid model A Lysis of target cell lines by fresh or cryopreserved HSPC-NK cells following 16 h co-culture in a 2D setting at varying E: T ratios. B Absolute number of NK cells at varying E: T ratios at the end of the 2D cytotoxicity assays. C, D Representative example of mean speed (C) and mean distance of fresh or cryopreserved HSPC-NK cells travelled/40 sec (D) in a 3D collagen matrix assessed using live-cell imaging for 800 s. E Schematic representation of the 3D cytotoxicity assay involving SKOV-3 spheroids embedded in a collagen matrix containing fresh or thawed HSPC-NK cells. F Lysis of SKOV-3 spheroids in a 3D 16 h-cytotoxicity assay using a dosage of 0.6 or 1.2 × 10⁶ fresh or cryopreserved HSPC-NK cells. G Absolute number of NK cells found at the end of the 3D cytotoxicity assay. All experiments were initiated within three hours after thawing. Bars in A, B, F, G represent the mean. The boxplot in C, D represents the median with errors bars showing the min and max values. A two-way ANOVA with Šídák multiple comparison correction was used to test significance between fresh and cryopreserved conditions in A, B, F, and G and a Mann-Whitney test was used for C and D. * P < 0.05, ** P < 0.01, *** P < 0.001

Cryopreserved HSPC-NK cells demonstrate high persistence in NXG mice upon i.p. and i.v. injection

The in vitro study provides promising evidence regarding the effector functions of surviving HSPC-NK cells after cryopreservation. However, as these results are obtained directly after thawing, understanding the behavior of cryopreserved cells upon in vivo infusion is crucial to better characterize the product as infused in patients. To address this, NXG mice were injected i.p. with the same dosage of fresh or thawed viable HSPC-NK cells, accompanied by IL-15 support every other day (Fig. 4A). Seven days post-infusion, cell persistence was evaluated in the peritoneal cavity and spleen. Notably, the number of viable HSPC-NK cells residing in the peritoneal cavity as well as the engraftment of hCD56⁺ cells was comparable between freshly cultured and thawed cells (Fig. 4B, C). Although only a small number of HSPC-NK cells migrated to the spleen following i.p. injection, similar cell numbers and engraftment were observed for both fresh and cryopreserved cells (Fig. 4B, C).

Fig. 4.

High persistence of cryopreserved HSPC-NK cells in NXG mice upon i.v. and i.p. infusion Persistence of fresh versus thawed HSPC-NK cells was evaluated after A i.p. or D i.v. infusion in non-tumor bearing NXG mice with two-daily IL-15 support. For the i.v. experiment (D) blood was collected at day one, four, and seven. At day seven, HSPC-NK cell numbers in the peritoneal cavity and spleen (A) or the bone marrow and spleen (D) were evaluated. B Absolute cell numbers and C engraftment, defined as the fraction of human CD56⁺ HSPC-NK cells compared to total CD56⁺ and mouse cells, in the peritoneal cavity and spleen seven days post i.p. infusion. E Absolute cell numbers and F engraftment of human CD56⁺ HSPC-NK cells in the blood, spleen, and bone marrow seven days after i.v. injection. B, C, E, F N = 10–12. Spleen analysis in B and C was performed on pooled mice samples due to minimal hCD56⁺ events/mice. Mice showing < 20 hCD56 + events were marked grey (ν). Bars in B, C, E, F represent the mean. Statistical differences between fresh and cryopreserved conditions were evaluated using the Mann-Whitney test for B, C and spleen and bone marrow results in E and F. Blood (E and F) was analyzed using a two-way ANOVA with Šidák multiple comparison. * P < 0.05, *** P < 0.001

In addition to i.p. infusion, which is relevant for treating abdominal malignancies, the persistence of cryopreserved HSPC-NK cells following intravenous (i.v.) administration was explored. This route allows for the potential application of the cryopreserved cell product in for instance hematological malignancies as well as metastatic disease of solid cancers. Persistence was assessed in the blood, spleen, and bone marrow seven days post-infusion (Fig. 4D). Additionally, blood was collected at day one and four to gain insight into time kinetics. Freshly cultured HSPC-NK cells showed increased viable cell numbers and engraftment in the blood from day one to four, followed by a decline between day four and seven (Fig. 4E, F). Cryopreserved HSPC-NK cells exhibited significantly reduced viable cell numbers and engraftment in the blood on day one and four compared to fresh cells. Interestingly, by day seven, persistence in the blood was comparable between fresh and cryopreserved HSPC-NK cells, which aligns the delayed cell death post-thawing and enhanced proliferative capacity of surviving thawed cells found in vitro. Finally, cryopreserved cells showed lower viable cell numbers and engraftment in the spleen and bone marrow (Fig. 4E, F). Overall, these findings demonstrate that cryopreserved HSPC-NK cells exhibit in vivo persistence following both i.p. and i.v. administration.

Cryopreserved HSPC-NK cells residing in vivo resemble phenotype and anti-tumor potency of fresh cells

In addition to persistence, the clinical application of a cryopreserved NK cell product relies on its anti-tumor potential after in vivo infusion. We first analyzed the phenotypical profile of viable HSPC-NK cells recovered from the peritoneal cavity seven days post-i.p. infusion. The expression of maturation and differentiation markers as well as the functional marker Perforin was comparable between fresh and cryopreserved cells (Fig. 5A). A significant but minimal reduction in NKp30 expression was found in the cryopreserved condition. Importantly, CD16 expression, which showed a trend towards decreased levels directly after thawing, was comparable to that of fresh cells in vivo. Additionally, the potency of the recovered HSPC-NK cells against K562 was evaluated. Notably, while CD107a expression upon K562 stimulation was reduced directly after thawing, cryopreserved cells demonstrated a significant increase in degranulation capacity compared to fresh cells in vivo (Fig. 5B). Furthermore, the IFNγ production potential of fresh and cryopreserved persistent cells was similar.

Fig. 5.

Thawed HSPC-NK cells recovered following infusion in NXG mice demonstrate a comparable phenotype and functionality to fresh HSPC-NK cells. A Phenotype of HSPC-NK cells recovered from the peritoneal cavity seven days after i.p. injection in NXG mice. B Fresh or cryopreserved conventional (Perforin⁺) HSPC-NK cells recovered from the peritoneal cavity were stimulated for four hours with K562 cells after which CD107a and IFNγ expression was determined. C Flow cytometric expression of activating, maturation, and effector markers on fresh or cryopreserved hCD56⁺ HSPC-NK cells recovered from blood, spleen, or bone marrow seven days after i.v. injection. D CD107a and IFNγ expression levels of conventional HSPC-NK cells recovered from the spleen seven days post i.v. injection after four-hour stimulation with K562. For B and D, expression levels are corrected for background activation in the NK alone controls, specifically showing target reactive cells. N = 8–12, with mice showing < 20 positive events or additionally < 50 positive CD56⁺ cells being marked grey (). Bars represent the mean. A two-way ANOVA with Šidák multiple comparison was used to test significance between the fresh and cryopreserved conditions, except for perforin (C) which was evaluated using the Mann-Whitney test. * P < 0.05, ** P < 0.01, *** P < 0.001

Consistent with the profile of HSPC-NK cells persisting after i.p. infusion, the phenotype of fresh and cryopreserved cells residing in the blood, spleen, and bone marrow seven days post-i.v. injection was highly comparable (Fig. 5C, Supplementary Fig. 6A). Interestingly, cryopreserved cells exhibited significantly higher expression of CD25 in the spleen, but significantly lower expression in the bone marrow compared to fresh cells, while CD25 levels in the blood were similar between both groups. As we previously observed trends of decreased CD16, CD25, CD69, and CD62L expression directly after thawing, we further investigated the expression dynamics of these markers on HSPC-NK cells residing in the blood. Notably, CD16 and CD62L expression was already comparable between fresh and cryopreserved cells one day post-infusion (Supplementary Fig. 6B). Moreover, expression of CD69 was significantly increased in fresh versus cryopreserved cells one day post-infusion and CD25 showed a significant increase both at day one and four post-infusion. In addition, similar to the i.p. mouse experiment, cryopreserved HSPC-NK cells recovered from the spleen demonstrated significantly enhanced degranulation capacity and maintained IFNγ production potential compared to fresh cells (Fig. 5D). Collectively, these results demonstrate that cryopreserved HSPC-NK cells persisting and expanding in vivo upon i.p. or i.v. infusion exhibit anti-tumor functionality comparable to fresh cells.

Discussion

The adoptive transfer of allogeneic NK cells has gained interest as a safe and effective anti-cancer immunotherapy with the potential for off-the-shelf production, enabling scalable and readily available treatment options. However, to date, clinical trials using cryopreserved NK cells remain limited as NK cells are sensitivity to freeze-thawing, which reduces cell viability and impairs functionality and expansion capacity in vivo [12, 18, 27]. Additionally, most research is restricted to PB-NK cells, and standardized protocols for the cryopreservation of HSPC-NK cells are lacking. Here, we present the development of a robust, GMP-compliant manufacturing process of cryopreserved RNK001 NK cells cultured ex vivo from UCB-derived HSPCs in a scalable G-rex bioreactor. This protocol achieved consistently high recovery rates of mature and potent HSPC-NK cells post-thawing. Moreover, surviving cryopreserved cells retained their ability to migrate towards and lyse tumor spheroids in a complex 3D environment and demonstrated in vivo persistence comparable to that of fresh cells, underscoring their potential for clinical application.

We optimized our GMP-compliant HSPC-NK cell production process by transitioning from a closed- system bag culture to a closed-system G-rex bioreactor [8]. This approach resulted in a high fold-expansion of pure HSPC-NK cell products equivalent to bag-generated products. However, the gas-permeable membrane of the G-rex enables the addition of larger medium volumes in a standardized manner, reducing the need for frequent additions every two to three days in bags to only six additions for the entire culture period. This improvement not only decreases the labor-intensity of the production process, but also reduces contamination risks associated with aseptic handling. Furthermore, the G-rex is scalable up to five-liter systems, potentially greatly enhancing HSPC-NK cell yields. To achieve optimal expansion in this system, our ongoing research is focused on determining optimal HSPC seeding densities, as lower starting densities appear to promote proliferation capacity (Fig. 1C). Additionally, since NK cells do not require HLA matching, pooling UCB units could further increase HSPC-NK cell yields. However, HSPC-NK cells not only offer advantages over PB-NK cells regarding expansion capacity and purity, but are also more amenable to genetic engineering since HSPCs are generally more easily manipulatable than NK cells. Thus, this platform can facilitate the generation of high numbers of genetically modified NK cells with enhanced anti-tumor potential and tumor reactivity.

The success of this scalable production platform depends on the ability to cryopreserve HSPC-NK cells with high viable cell numbers and functionality after thawing. Using our established cryopreservation protocol, we achieved high recovery and viability of cells directly after thawing. However, cell numbers declined upon resting in IL-15 or IL-2, which was also demonstrated in a NXG mouse model where a significant reduction in circulating cryopreserved cells compared to fresh cells was observed within one day after i.v. infusion. These findings correspond to previous research showing reduced viability of PB-NK cells within 12 to 24 h post-thawing in the presence of IL-2 [13, 15, 19, 28]. Berjis et al. attributed this decline in viability to apoptosis caused by intracellular GZMB leakage, which could be mitigated by 24 h treatment with IL-18 and IL-15 to decrease intracellular GZMB levels pre-freezing [13]. Nevertheless, IL-18 and IL-15 pre-treatment failed to lower pre-freezing GZMB cell content in our system, possibly due to the already high cytokine stimulation, including IL-15, during our HSPC-NK cell process.

Despite extensive efforts to counteract cryopreservation-induced cell death, it may in fact have a favorable effect on the behavior of surviving NK cells in our product. We observed enhanced proliferative capacity of the surviving cells upon thawing compared to fresh cells when stimulated with IL-15 or IL-2 in vitro. Previous studies have reported impaired in vivo expansion of cryopreserved allogeneic NK cells in mice and patients [18, 27]. Miller et al. showed that ex vivo-expanded or activated PB-NK cells infused i.v. in NSG mice directly after thawing or following overnight IL-2 resting had impaired persistence compared to fresh cells and only minimally expanded [18]. Similarly, a clinical trial in relapsed multiple myeloma patients showed that cryopreserved PB-NK cells failed to proliferate in vivo after i.v. injection [27]. In contrast, we observed enhanced numbers of surviving thawed HSPC-NK cells in the circulation between four and seven days post i.v. infusion, consistent with their in vitro proliferation potential. Although the enhanced proliferative capacity of viable cryopreserved cells did not result in similar yields to fresh cells after seven days of culture in IL-15 or IL-2 in vitro, comparable persistence of fresh and cryopreserved cells was found in the circulation upon i.v. infusion. The number of fresh HSPC-NK cells in the circulation decreased from day four to day seven, which likely indicates infiltration in tissues. The fraction of thawed HSPC-NK cells in the spleen and bone marrow was reduced compared to fresh cells. This might be caused by delayed infiltration due to initial loss of circulating cells or migration towards non-analyzed tissues such as the liver and lungs, as we previously showed for freshly cultured HSPC-derived NK cells [29]. Further tissue infiltration might be enhanced by the presence of a tumor, as we also observed in our 3D cytotoxicity model where thawed HSPC-NK cells showed comparable killing of spheroids as fresh cells in a setting where migration is required. Importantly, we also demonstrated comparable persistence of fresh and cryopreserved HSPC-NK cells following i.p. infusion, highlighting the potential application of this cryopreserved product for both hematological and solid malignancies.

Next to cell persistence, a potential clinical benefit of a cryopreserved NK cell product depends on its anti-tumor functionality. Surviving cryopreserved HSPC-NK cells retained their mature and differentiated phenotype. While not statistically significant, a trend for decreased expression of the ADCC receptor CD16, activation markers CD25 and CD69, and the bone marrow-directed homing marker CD62L was observed directly after thawing. This corresponds to previous literature showing decreased CD16 and CD69 levels immediately post-thawing [14–16, 21, 28]. Mark et al. suggested that reduced CD16 levels were due to activation-induced cleavage by ADAM17 matrix metalloproteases [14]. Recently, we engineered our HSPC-NK cell product with a non-cleavage CD16 variant with enhanced affinity, which could potentially further improve ADCC activity of cryopreserved cells [30]. Like CD16, shedding of CD62L by ADAM17 has been reported upon activation, potentially explaining its downregulation [31]. Consistent with decreased CD25 and CD69 levels, a slight reduction of degranulation capacity but not intracellular IFNγ production potential was observed directly after thawing. Importantly, one day post-infusion in NXG mice, expression of all markers was similar between fresh and cryopreserved HSPC-NK cells. Moreover, cryopreserved cells exhibited comparable IFNγ production capacity and significantly higher CD107a levels than fresh cells following both i.v. and i.p. infusion. This corresponds to previous work on PB-NK cells demonstrating enhanced CD107a expression after 24 h of resting in IL-2 supplemented medium [13]. Combined with the high proliferative capacity of cryopreserved cells, two potential mechanisms could explain these observations; either cryopreservation selectively preserves proliferative, active, and potent NK cells, or the cryopreservation process itself induces NK cell activation promoting proliferation and degranulation capacity.

Even in studies reporting high initial viability and anti-tumor potency of PB-NK cells post-thawing, the cytolytic potential of cryopreserved cells was impaired directly after thawing or following resting in the presence of IL-2 [13, 14]. Although tumor lysis by our cryopreserved HSPC-NK cell product was reduced in a 2D environment, this outcome was highly biased by thawing-induced NK cell death occurring during the 16-hour assay. In fact, one batch exhibited post-thaw cytotoxicity comparable to its donor-matched fresh product despite lower NK cell numbers. The retained effector function of surviving cryopreserved HSPC-NK cells was further validated using a complex 3D model, where spheroids where embedded in a collagen matrix, requiring NK cells to efficiently migrate towards, infiltrate, and eventually eradicate the spheroid. Although we observed reduced spontaneous migration potential of viable cryopreserved cells in a collagen matrix, which is in line with literature, cryopreserved HSPC-NK cells exhibited comparable lysis of collagen-embedded spheroids to fresh cells, even at a reduced viable NK cell dosage [14]. This discrepancy may be explained by spheroid-induced chemoattraction, which could enhance cell motility. We have previously demonstrated that SKOV-3 spheroids, secrete, amongst others, CXCL10, which stimulates HSPC-NK cell infiltration [32]. Moreover, Mark et al. showed that cryopreserved cells with remained motility retain their cytolytic function in a tumor-containing 3D environment comparable to fresh cells [14]. Collectively, these models demonstrate that the anti-tumor effector function of surviving cryopreserved HSPC-NK cells is maintained post-thawing.

Overall, this study presents the optimization of our GMP-compliant RNK001 HSPC-NK cell production process by promoting scalability and establishing an off-the-shelf cryopreserved product. This procedure was successfully validated and documented in an investigational medicinal product dossier for application in our ongoing phase I/IIa clinical trial in patients with relapsed AML (NCT04347616) [8]. Although cryopreservation reduced the viable HSPC-NK cell dosage, surviving cells had a high proliferative capacity and retained anti-tumor functionality and in vivo persistence compared to fresh cells. This optimized manufacturing pipeline facilitates the on-demand availability of HSPC-NK cells, and eventually genetically engineered NK cells, offering an effective and scalable therapeutic option for hematological and solid malignancies.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: LH, MVM, BS, AW, PJ, HD. Data acquisition: LH, MVM, BS, VCR, PJ. Data analysis: LH Interpretation of data: LH, MVM, BS, DVE, VCR, WH, AG, SVD, JJ, AW PJ, HD. Drafting and revising the manuscript: LH, MVM, BS, DVE, VCR, WH, AG, SVD, JJ, AW PJ, HD. Supervision: HD, PJ, AW. All authors have approved the final article.

Funding

This research is financed by the Dutch Cancer Society under project number 11564, 13535 (Postbus 75508, 1070 AM Amsterdam).

Data availability

Further details on data obtained in this study are available from the corresponding author on request.

Declarations

Ethics approval

UCB was collected following written informed consent from delivery at the Radboudumc (CMO 2014 − 226) or from the Radboudumc cord blood bank. Ethical approval for animal studies was granted by the Dutch central committee of animal experiments (DEC 2018-0029).

Competing interests

The authors declare that they have no financial or non-financial interests to disclose.