- Generation and Characterization of NKM cells
In order to achieve safe and effective immunotherapy for patients, we established an in vitro culture system for the production of immune cells without any separation or any biological additives. The autologous plasma was prepared and inactivated, then used in the cell culture process. The autologous immune cells are manufactured as shown in figure 1A and the method section, and we call these cells mixed-NK cells (abbreviated as NKM). After the culture is completed, we can obtain about 2.5 billion immune cells, including the NK cells and T cells.
We further analyzed the immune cell subsets of these NKM. Totally 369 NKM cells were analyzed for the distribution of NK cells and T cells. The results showed that the NKM were composed by approximately 20% NK cells (CD3-CD16+ or CD56+), approximately 30% NKT-like cells (CD3+CD16+ or CD56+), and other T cells. The T cells of NKM were composed by the approximately 30% regulatory T cells (Treg, CD3+CD4+) and 50% cytotoxic T cells (CD3+CD8+) (Figure 1B). In PBMC, the NK cells were mainly composed by the CD3-CD16+CD56neg NK cells and CD3-CD16+CD56dim NK cells (Supplementary Figure S1A). However, the NK cells from NKM were mainly composed by the CD3-CD16+CD56+ NK cells, of which the main subtype is CD3-CD16brightCD56bright NK cells (Figure 1C and Supplementary Figure S1B). Furthermore, we analyzed the NK subpopulations from donor-L at different stages of cell culture, including the primary culture stage (PBMC, T75 flask and T225 flask) and expansion stage (CO2Bag). The results showed that main NK subpopulations in PBMC were CD16bright NK cells (Figure 1D). The CD16bright NK cells transformed into CD16dim NK cells in primary culture stage (Figure 1E and 1F), and finally became CD16brightCD56bright NK cells after expanded proliferation (Figure 1G).
- In vitro cytotoxicity and its correlations with the subpopulation of NKM cells
For the immune cells, the cytotoxicity on cancer cells is one of the most important functions. Thus, we conducted the in vitro cytotoxicity of NKM cells on K562 cancer cells, and compared them with PBMC and NK92 cells. In PBMC (from health donor), the average cytotoxicity is around 7.4% (E/T = 10:1, 2h incubation; n=9) and 16.8% (E/T = 10:1, 4h incubation; n=28). The average cytotoxicity of NKM cells (n=198) that we manufactured is around 65.6% (E/T = 10:1, 2h incubation), which is almost ten times higher than the PBMC (Figure 2A and Supplementary Table S1). Approximately 84% of NKM cells show more than 40% cytotoxicity. Even with more effector cells and more incubation times, the cytotoxicity of PBMC is still far inferior to NKM cells (Figure 2B and 2C). Then the cytotoxicity of NK92 cells was investigated with different E/T ratio (Supplementary Figure S2A), the result showed that the cytotoxicity of NKM cells (E/T = 10:1) is comparable to that of NK92 with an E/T ratio of 5:1 (2h incubation) (Supplementary Figure S2B). In summary, these results indicate that although the NKM cells have only half the cytotoxic capacity of pure NK cells (NK92), their cytotoxicity is much higher than that of PBMC.
As the NKM cells are a mixed cell population, we wonder which subsets of these cells are responsible for their in vitro cytotoxicity. Then we analyzed the subpopulation of the NKM cells, and performed correlation analysis with their in vitro cytotoxicity. We separated the NKM cells into NK cells (CD3-CD16+ or CD56+), NKT-like cells (CD3+CD16+ or CD56+), CD4T cells (Treg cells, CD3+CD4+), CD8T cells (cytotoxic T cells, CD3+CD8+). Strong positive correlation was observed between the cytotoxicity of NKM cells and the ratio of NK cells (r=0.58, p<0.0001), but a negative correlation with the ratio of Treg cells (r=-0.32, p<0.0001) (Figure 3A and 3C). There is no correlation with the ratio of NKT-like cells and cytotoxic T cells (Figure 3B and 3D). Furthermore, we divide NK cells into three subpopulations: CD56+CD16- NK cells, CD56+CD16+ NK cells and CD56-CD16+ NK cells. The cytotoxicity of NKM cells only showed a significant correlated with CD56+CD16+ NK cells (r=0.56, p<0.0001), while no correlation was found in CD56+CD16- NK cells and CD56-CD16+ NK cells (Figure 3E-G). Thus, these data indicated that the in vitro cytotoxicity of NKM cells is mainly caused by the cytotoxicity of CD56+CD16+ NK cells, and is suppressed by Treg cells.
- Potential therapeutic function of NKM cells in a pancreatic cancer patient (In vivo cytotoxicity)
After the characterization of NKM cells, the potential in vivo anti-tumor activity was also need to be investigated. Previous systematic reviews of immune cell therapy with NK cells indicated that allogeneic NK cells immunotherapy has better clinical efficacy than autogeneic therapy[30-32]. Thus, we recruited a patient with pancreatic cancer (T4N1M1c) whose cancer cells had metastasized and cannot sustain chemotherapy. Then we treated this patient with HLA-matched haploidentical NKM cell infusion (Donor: patient’s son). The treatment process was shown in the figure 4A. The patient received the intravenous injection of 20~30 ×108 NKM cells each time. After six treatments, the patient showed well tolerated, and cancer metastatic foci of the renal pelvis and abdominal pelvic peritoneum were significantly reduced (Figure 4B and 4C). The patient's vital signs and physical condition were significantly improved. These results indicated that the NKM cells displayed high effective anti-tumor activity in the cancer patient, and suggested that the NKM could potentially be used for adjuvant cancer therapy.
- Identify and recharacterize the sub-healthy individuals on the basis of PD-1 level on PBMCs
Previously, when analyzing the PBMC of many cancer patients, the researchers found that PD-1 expression is usually significantly induced in peripheral immune cells[18, 19, 33-35]. And we also found that in adolescents (10 to 20 years old), PD-1 expression of PBMC is very low (<2%) (Data not shown). Therefore, in order to identify sub-healthy individuals, we investigated PD-1 and CTLA-4 expression in NK and T cells of PBMC in 95 individuals over 50 years of age (Supplementary Table S2). The results showed that PD-1 was only expressed on T cells, including the Treg cells and cytotoxic T cells, but not NK cells, while CTLA-4 was not expressed in any PBMC (Figure 5A and 5B). High proportion of PD-1 expression in T cells may indicate that a large number of T cells were exhausted[13, 14], which in turn implies an abnormal immune system.
Further statistical analysis found that the proportion of CD3+PD-1+ cells in PBMC was usually less than 4%, and only 16.8% of individuals were higher than 4% (Figure 6A). The proportion of CD3+CD4+PD-1+ cells in CD3+CD4+ T cells was usually less than 10% and only 9.5% of individuals were higher than 10% (Figure 6B). The proportion of CD3+CD8+PD-1+ cells in CD3+CD8+ T cells was usually less than 10% and 37.9% of individuals were higher than 10% (Figure 6C). Here, in order to define sub-healthy individuals, we set the threshold of CD3+PD-1+/PBMC to >4% and the threshold of CD3+CD8+PD-1+/CD3+CD8+ to >10%. In our cohort, about 4.2% of individuals showed these characteristics, so their immune system of these people may be abnormal, and we define these people as sub-healthy individuals.
- Potential immune normalization of autologous NKM cells in sub-healthy individual
Since the NKM cells are mixed immune cells, including Treg cells that inhibit cytotoxicity and NK cells that produce cytotoxicity, we wonder that these NKM cells may display more effective functions in the immune normalization of sub-healthy individuals. Then we recruited four potential sub-healthy individuals to participate in this study. The treatment schedule was shown in the figure 7A. Approximately 20 ×108 autologous NKM cells were used for each treatment. The examination of PD-1 expression in T cells were performed before the treatment (pre-), after the 1st injection (mid-) and after the 3rd injection (post-). In Patient-M and Patient-Z, with a single injection of NKM cells, the PD-1 expression of PBMC, CD3+CD8+ and CD3+CD4+ could be significantly reduced to normal levels (Figure 7B and 7C). While in Patient-J, the effect of one injection of NKM cells is not obvious. After three injections of NKM cells, the PD-1 expression of PBMC, CD3+CD8+ and CD3+CD4+ eventually reduce to normal levels (Figure 7D). In Patient-C, three injections of NKM cells could only reduce the PD-1 expression of PBMC and CD3+CD8+, but not the PD-1 expression of CD3+CD4+ cells (Figure 7E). All sub-healthy individuals showed good tolerance and no side effects were found clinically. In summary, we found that with three autologous NKM cell therapies in general, we could successfully reduce the PD-1 expression in PBMC, and achieve the renormalization of the immune system in sub-healthy individuals.