Immune normalization strategy against suboptimal health status: safe and efficacious therapy using mixed-natural killer cells

Generation and characterization of NKM cells

To achieve safe and efficacious immunotherapy for patients, we established an in vitro culture system for the production of immune cells without separation or biological additives. The medium used in this culture system was designed specifically for the expansion of NK cells and T cells. Autologous plasma was prepared and inactivated, then used for cell culture. Autologous immune cells were manufactured as shown in Figure 1A and in the Methods and Materials section: these were NKM cells. After the culture was complete, we could obtain ~2.5 billion immune cells, the main components of which were NK cells and T cells.

We analyzed the immune-cell subsets of NKM cells. The proportion of monocytes (CD14⁺), B cells (CD19⁺) and dendritic cells (CD11c⁺) was extremely low. In total, 369 NKM cells were analyzed for the distribution of NK cells (CD45⁺CD3⁻CD16/56⁺) and T cells (CD45⁺CD3⁺). NKM cells comprised ~20% NK cells (CD3⁻CD16/56⁺), ~30% NK T-like cells (CD3⁺CD16/56⁺), and other T cells. Of the T cells within NKM cells, ~30% were T helper (T_h) cells (CD3⁺CD4⁺) and 50% were cytotoxic T cells (CD3⁺CD8⁺) (Figure 1B). In PBMCs, NK cells comprised mainly CD3^-CD16⁺CD56^neg NK cells and CD3⁻CD16⁺CD56^dim NK cells (Supplementary Figure 1A). However, the NK cells from NKM cells mainly comprised CD3⁻CD16⁺CD56⁺ NK cells, of which the main subtype was CD3⁻CD16^brightCD56^bright NK cells (Figure 1C and Supplementary Figure 1B). Furthermore, we analyzed the subpopulations of NK cells from donor-L at different stages of cell culture, including at the primary-culture stage (PBMCs, T75 flask, and T225 flask) and expansion stage (CO₂ Bag). The main subpopulation of NK cells in PBMCs was CD16^bright NK cells (Figure 1D). The latter transformed into CD16^dim NK cells upon primary culture (Figure 1E, 1F), and finally became CD16^brightCD56^bright NK cells after expanded proliferation (Figure 1G).

In vitro cytotoxicity and its correlations with the subpopulation of NKM cells

We tested the in vitro cytotoxicity of NKM cells on a cancer cell line (K562), and compared its cytotoxicity with that of PBMCs and NK92 cells. In PBMCs (from healthy donors), the mean cytotoxicity was ~7.4% (effector/target (E/T) ratio = 10:1, 2-h incubation; n = 9) and 16.8% (E/T ratio = 10:1, 4-h incubation; n = 28). The mean cytotoxicity of NKM cells (n = 198) that we manufactured was ~65.6% (E/T ratio = 10:1, 2-h incubation), which was almost tenfold higher than that of PBMCs (Figure 2A and Supplementary Table 1). Approximately 84% of NKM cells showed >40% cytotoxicity. Even with more effector cells and longer incubation times, the cytotoxicity of PBMCs remained far inferior to that of NKM cells (Figure 2B, 2C). Then, the cytotoxicity of NK92 cells was investigated with different E/T ratios (Supplementary Figure 2A). We discovered that the cytotoxicity of NKM cells (E/T ratio = 10:1) was comparable with that of NK92 cells, with an E/T ratio of 5:1 (2-h incubation) (Supplementary Figure 2B). Taken together, these results suggested that, although NKM cells had only half the cytotoxicity of that observed in “pure” NK cells (NK92), their cytotoxicity was much higher than that of PBMCs.

NKM cells are a mixed population of cells, so we wondered which subsets of these cells were responsible for in vitro cytotoxicity. We analyzed the subpopulations of NKM cells, and undertook correlation analyses of their in vitro cytotoxicity. We separated NKM cells into NK cells (CD3⁻CD16⁺ or CD56⁺), NK T-like cells (CD3⁺CD16⁺ or CD56⁺), CD4 T cells (T_h cells, CD3⁺CD4⁺) or CD8 T cells (cytotoxic T cells, CD3⁺CD8⁺). A strong positive correlation was observed between the cytotoxicity of NKM cells and the proportion of NK cells (r = 0.58, p < 0.0001), but a negative correlation with the proportion of T_h cells was noted (r = −0.32, p < 0.0001) (Figure 3A, 3C). There was no correlation with the proportion of NK T-like cells and cytotoxic T cells (Figure 3B, 3D).

Furthermore, we divided NK cells into three subpopulations: CD56⁺CD16⁻ NK cells, CD56⁺CD16⁺ NK cells, and CD56^-CD16⁺ NK cells. The cytotoxicity of NKM cells showed a significant correlation only with CD56⁺CD16⁺ NK cells (r = 0.56, p < 0.0001), and no correlation was found in CD56⁺CD16⁻ NK cells or CD56−CD16⁺ NK cells (Figure 3E–3G). Then, we used sorting of immunomagnetic beads to separate different cell subpopulations in NKM cells, and verified them with flow cytometry. We discovered that NK cells isolated from NKM cells showed the highest cytotoxicity (Figure 3H), CD3⁻ cells, CD4⁻ cells and CD8⁻ cells showed moderate cytotoxicity, and NK-negative cells, CD3⁺ cells, CD4⁺ cells, and CD8⁺ cells showed the lowest cytotoxicity (Supplementary Figure 3A–3C).

In addition, we studied the secretion of cytokines in these NKM cells: interleukin (IL)-4, IL-6, IL-10, tumor necrosis factor (TNF)-α, and interferon (INF)-ɣ. The media from 10 cultured NKM cells were collected and analyzed. The concentration of IL-4, IL-6, IL-10 and TNF-α was very low (<100 pg/mL), whereas the INF-ɣ concentration was very high (~300 pg/mL on average) (Supplementary Figure 4).

Taken together, these data suggested that the in vitro cytotoxicity of NKM cells was caused mainly by the cytotoxicity of CD56⁺CD16⁺ NK cells, and was suppressed by T_h cells.

Potential function of NKM cells for adjuvant cancer treatment (in vivo cytotoxicity)

After the characterization of NKM cells, the potential in vivo anti-tumor activities of NKM cells were investigated in cancer patients for adjuvant cancer therapy. We recruited 11 patients with malignant lymphoma (diffuse large B-cell lymphoma [seven patients], peripheral T-cell lymphoma [two patients], anaplastic large-cell lymphoma [one patient], and Hodgkin’s lymphoma [one patient]) who suffered primary relapse/were refractory to treatment (one patient), relapsed after salvage chemotherapy (eight patients), or relapsed after autologous stem-cell therapy (two patients). The treatment process is shown in Figure 4A. Each patient received an intravenous injection of 20–30 ×10⁸ NKM cells each time (treatment once every 2 weeks). The response to NKM treatment was not robust (a complete response in one case and stable disease in one case) (Supplementary Table 2), but quality of life was improved greatly. In the patient who had a complete response (case #2), the number of cancer foci was reduced significantly after treatment with NKM cells (Figure 4B). During and after intravenous infusion of NKM cells, the body temperature, blood pressure, and heartbeat remained stable, and adverse reactions were not found. The patients tolerated the treatment and clinical side-effects were absent.

Systematic reviews of therapy with NK cells have indicated that immunotherapy with allogeneic NK cells has better clinical efficacy than that using autologous cells [29–31]. Thus, we recruited another patient with pancreatic ductal carcinoma (T4N1M1c) whose cancer cells had metastasized, and he could not sustain chemotherapy. Then, we treated this patient with a human leukocyte antigen-matched infusion of haploidentical NKM cells (donor was the patient’s son). The patient received an intravenous injection of 20–30×10⁸ NKM cells every week. After six treatments, the patient showed good toleration of treatment, and the number of metastatic foci in the renal pelvis and abdominal pelvic peritoneum was reduced significantly (Figure 4C). The patient’s vital signs and physical status were improved significantly.

Taken together, these results suggested that NKM cells displayed efficacious anti-tumor activity in cancer patients, and suggested that NKM cells could be used for adjuvant cancer therapy.

Identification and recharacterization of sub-healthy individuals on the basis of PD-1 expression in PBMCs

When analyzing the PBMCs of cancer patients, researchers found that PD-1 expression was significantly induced in peripheral immune cells [18, 19, 32–34]. Previously, we found that among adolescents (10–20 years), the proportion of PD-1⁺ PBMCs (CD3⁺CD4⁺PD-1⁺ or CD3⁺CD8⁺PD-1⁺) was very low (<2%) (data not shown). However, the proportion of PD-1⁺ PBMCs in adults varies greatly, and the proportion in individuals is stable if external intervention is not applied. Here, to identify sub-healthy individuals, we measured expression of PD-1 and cytotoxic T lymphocyte-associated protein (CTLA)-4 in the NK cells and T cells of PBMCs in 95 individuals aged >50 years (Supplementary Table 3). PD-1 was expressed only on T cells, including T_h cells and cytotoxic T cells, but not on NK cells, whereas CTLA-4 was not expressed in any PBMCs (Figure 5A, 5B). High expression of PD-1 in T cells may indicate that a large number of T cells are exhausted [13, 14] which, in turn implies an abnormal immune system.

The proportion of CD3⁺PD-1⁺ cells in PBMCs was usually <4%, and only 16.8% of individuals had >4% of CD3⁺PD-1⁺ cells in PBMCs (Figure 6A). The proportion of CD3⁺CD4⁺PD-1⁺ cells in CD3⁺CD4⁺ T cells was usually <10%, and only 9.5% of individuals had >10% of CD3⁺CD4⁺PD-1⁺ cells in CD3⁺CD4⁺ T cells (Figure 6B). The proportion of CD3⁺CD8⁺PD-1⁺ cells in CD3⁺CD8⁺ T cells was usually <10%, and 37.9% of individuals had >10% of CD3⁺CD8⁺PD-1⁺ cells in CD3⁺CD8⁺ T cells (Figure 6C). To define sub-healthy individuals, we set the threshold of the ratio of CD3⁺PD-1⁺ cells/PBMC cells to >4% and the threshold of the ratio of CD3⁺CD8⁺PD-1⁺ cells/CD3⁺CD8⁺ cells to >10%. Furthermore, we measured PD-1 expression in NKM cells. The proportion of CD3⁺PD-1⁺ cells and CD3⁺CD4⁺PD-1⁺ cells was increased in NKM cells, whereas the proportion of CD3⁺CD8⁺PD-1⁺ cells was reduced significantly in NKM cells (p < 0.0001) (Supplementary Figure 5). About 4.2% of individuals in our cohort showed these characteristics, so the immune system of these people may be abnormal, and we defined these people as “sub-healthy individuals”.

Potential immune normalization of autologous NKM cells in sub-healthy individuals

NKM cells are a mixture of immune-cell types, including T_h cells (which inhibit cytotoxicity) and NK cells (which elicit cytotoxicity). We wondered if these NKM cells may display more efficacious functions in the immune normalization of sub-healthy individuals.

We recruited four potential sub-healthy individuals and the treatment schedule is shown as Figure 7A. Approximately 20×10⁸ autologous NKM cells were used for each treatment. Measurement of PD-1 expression in T cells was done before treatment, after the first injection, and after the third injection. In Patient-M and Patient-Z, upon a single injection of NKM cells, PD-1 expression in CD3⁺, CD3⁺CD8⁺, and CD3⁺CD4⁺ cells could be reduced significantly to a normal level (CD3⁺PD-1⁺ cells/CD3⁺ cells was <4%; CD3⁺CD4⁺PD-1⁺ cells/CD3⁺CD4⁺cells was <10%; CD3⁺CD8⁺PD-1⁺ cells/CD3⁺CD8⁺ cells was <10%) (Figure 7B, 7C). In Patient-J, the effect of one injection of NKM cells was not obvious. After three injections of NKM cells, PD-1 expression in CD3⁺, CD3⁺CD4⁺, and CD3⁺CD8⁺ cells decreased eventually to a normal level (Figure 7D). In Patient-C, three injections of NKM cells reduced PD-1 expression in CD3⁺ cells and CD3⁺CD8⁺ cells, but not in CD3⁺CD4⁺ cells (Figure 7E). All sub-healthy individuals showed good tolerance to treatment, and side-effects were not observed.

In summary, we found that with three therapies with autologous NKM cells, we could reduce PD-1 expression in PBMCs, and achieve renormalization of the immune system in sub-healthy individuals.

Culture of NKM cells

NKM cells were manufactured in an ISO Class 5 Good Manufacturing Practice Laboratory within the International Research Center for Regenerative Medicine in BOAO International Hospital (Qionghai, China). Primary PBMCs were separated from whole blood (30–60 mL). A T75 flask was pre-coated with OKM25 medium (Fukoku, Ageo, Japan). PBMCs from individuals were cultured for 3–5 days in T75 and T225 flasks with OKM100 medium containing 1750 JRU/mL of IL-2 (Fukoku) and 10% autologous inactivated plasma. Then, NKM cells were transferred to a CO₂ Bag containing 1 L of OKM200 medium (containing 175 JRU/mL of IL-2) (Fukoku) and 1% autologous inactivated plasma, followed by culture for 3–5 days. NKM cells were incubated at 37° C in a humidified incubator in an atmosphere of and 5% CO₂.

Ethical approval and consent to participate

The study protocol was approved by the Medical Ethics Committee of BOAO International Hospital. Sample collection, written informed consent from and recruitment of healthy donors and patients followed Ethics Review Board protocols from BOAO International Hospital (QiongHai, China; approval number of BIH-2018-1001 for adjuvant cancer therapy, and BIH-2018-1002 for care of sub-healthy individuals). The study was conducted according to the principles of the Declaration of Helsinki 1964 and its later amendments.

Flow cytometry and antibodies

We wished to analyze the T-cell subpopulations of PBMCs and NKM cells. BD Tritest™ CD3 fluorescein isothiocyanate (FITC)/CD16⁺CD56 phycoerythrin (PE)/CD45 peridinin chlorophyll protein (PerCP) and BD Tritest CD4 FITC/CD8 PE/CD3 PerCP (BD Biosciences, San Jose, CA, USA) were used for flow cytometry. BD Pharmingen™ PE Mouse Anti-Human CD56 and BD Pharmingen FITC Mouse Anti-Human CD16 (BD Biosciences) antibodies were used for analyses of the NK cell-subpopulations of PBMCs and NKM cells. BD Pharmingen APC Mouse anti-Human CD279, BD Pharmingen APC Mouse Anti-Human CD152, and BD Pharmingen APC Mouse IgG1, κ Isotype Control antibodies (BD Biosciences, Shanghai, China) were used for flow cytometry of the PD-1⁺ cell and CTLA-4⁺ cell subpopulations of PBMCs and NKM cells. Flow cytometry was done according to manufacturer (BD Biosciences) instructions.

Assay to measure cell cytotoxicity

The cytotoxicity of PBMCs and NKM cells was assessed by measuring the mean fluorescence intensity of Calcein AM (catalog number, C3099; Thermo Fisher Scientific, Waltham, MA, USA) in intact cancer cells using Terascan VPC (Minerva Tech, Tokyo, Japan). Briefly, target cells (cancer cell line K562) were resuspended with Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, and mixed with 20 μL of Calcein AM at 37° C for 30 min for staining. PBMCs or NKM cells were co-cultured with pre-stained target cells at a particular E:T ratio for 2 h or 4 h. The mean fluorescence intensity of living target cells was measured. The percent cytotoxicity of cells was calculated using the following formula [35]: % cytotoxicity = (1 − [(average fluorescence of the sample wells − average fluorescence of the maximal-release control wells)/(average fluorescence of the minimal-release control wells − average fluorescence of the maximal-release control wells)]) ×100.

Sorting of immunological magnetic beads

Cell subsets of NKM cells were sorted using an NK Cell Isolation Kit (130-092-657) for NK cells, CD3 MicroBeads (130-050-101) for CD3⁺ T cells, CD4 MicroBeads (130-045-101) for CD4⁺ T cells, and CD8 MicroBeads (130-045-201) for CD8⁺ T cells according to manufacturer (Miltenyi Biotec, Bergisch Gladbach, Germany) instructions. Isolated cells (negative and positive) were sorted further with a flow cytometer.

Cytokine analyses with a cytometric bead array (CBA) kit

The cytokines secreted by NKM cells were analyzed by Human Th1/Th2 Cytokine Kit II (551809) from BD Biosciences. NKM cells (~2×10⁹ cells) were harvested from 1.5 L of culture medium. The cytokines in these media were dyed and analyzed by flow cytometry according to the instructions of the CBA-kit supplier. Then, the collected data were analyzed with FCAP Array v3.0 (BD Biosciences).

Statistical analyses

Experiments were carried out independently in triplicate. Data are the mean ± SD. Graphs were plotted and analyzed with Prism 7.1 (GraphPad, La Jolla, CA, USA). P < 0.05 was considered significant.

Availability of data and material

All data have been added in the Supplementary Tables.