Since G-CSF increases neutrophil mobilization and maturation, it was initially used in clinical practice to prevent and treat neutropenia [
45]. However, nowadays its ability to mobilize HSC from the bone marrow [
46] has positioned this molecule as the most commonly used factor for this purpose, being the gold standard and changing the paradigm of stem cell transplantation [
47]. That great performance could be related to some advantages such as the increase in the number of different peripheral blood white cells such as lymphocytes, monocytes, and obviously HSC [
48]; the reduced time for the restoration of neutrophils and platelets post-transplant [
49], major safety, and importantly, according to clinical trials, normal donors prefer donation of HSPCs from blood, instead of donation from pelvic marrow, which is more invasive and dangerous [
50].
In short, this hypothesis suggests that, on one hand, G-CSF prompts neutrophil activation, ending up in a degranulation, releasing some neutrophil’s protease enzymes such as neutrophil elastase, cathepsin G, dipeptidyl peptidase 1I and matrix metalloprotease-9 [
54] that can ultimately accumulate in BM and lead to the degradation of cell adhesion-related molecules such as VCAM-1[
55], SDF-1 and C-kit [
53]. Moreover, G-CSF increases CD26, a serine exopeptidase on the surface of endothelial cells that causes internalization and degradation of VE-cadherin, opening endothelial boundaries. G-CSF also triggers erythroblasts to secrete fibroblast growth factor-23 (FGF-23), which counteracts the function of CXCR-4. On the other hand, following G-CSF administration it has been registered a reduce in the number of monocytes, macrophages and suppression of osteoblasts in bone marrow [
56], which could enhance the mobilization process [
57,
58]. In addition to that, it seems that G-CSF induces sympathetic neurons to release noradrenaline which promotes a suppression of osteoblast [
59], and macrophages [
60] to release unknown factors that suppress SDF-1 expression on the surface of niche cells.
Therefore, it is fairly noticeable that nowadays, there is a large number of G-CSF mobilized and cryopreserved apheresis products both from donors and patients, stored in biobanks for HSCTs main purpose, which could be also used to obtain T lymphocytes and NK cells for subsequent generation of CAR products.
3.1. CAR-T Immunotherapy
Since CAR-T therapy has been proposed as an ultimate procedure after standard cancer treatment, it seems reasonable to think that T lymphocytes for CAR-T therapy could be isolated and expanded from the same cryopreserved product obtained for HSCT, in order to combine both treatments. The paradigmatic case of this potential use would be multiple myeloma or NHL, where many patients undergo CAR-T therapy after relapse following autologous transplantation [
62]. In the case of LAL, successful cases have been described with the infusion of allogeneic CAR-T cells from the donor after early relapse following allogeneic HSCT [
63,
64,
65]. On the other hand, the hematopoietic toxicities of autologous CAR-T cell therapies are a major concern, and there are many cases reported that have required the infusion of hematopoietic progenitors as rescue [
66,
67].
In all these circumstances it would be beneficial to have a single procedure for apheresis, but all of them require the administration of G-CSF as a mobilization agent, and this growth factor has not been considered for the manufacture of the CAR-T products among the different parameters that have been previously well determined [
4].
Several studies on the functionality of T lymphocytes after mobilization provide reasons not to use this cell source for CAR-T cell production. In the late 1990s, a series of publications studied mobilized stem cell samples and concluded that G-CSF has a pleiotropic effect in different cell populations, including monocytes which inhibit T cell proliferation and function [
48,
68]. First, Young et al. informed the induction of immunosuppressive cells associated with myelopoiesis and stimulated by GM-CSF and IL-3 [
69]. Subsequently, Ino K. et al analyzed the mobilized apheresis product of 21 patients with malignancies. They carried out different cellular inhibition assays and showed that the low-density fraction enriched with CD14+ significantly inhibited the functions of T cells and led to activation-induced apoptosis of these cells [
70]. These findings suggest that the augmented monocyte fraction is responsible for the impaired function and inhibition of T-cells after mobilization
This could be an important reason to not utilize G-CSF mobilized PBMNC for immunotherapeutic purposes. However, one feasible way to mitigate this T-cell misfunction is to isolate and expand the T-cell CD3+ fraction specifically [
71]. Nowadays there are a variety of kits widely used that are designed to enrich solely this CD3+ fraction [
72,
73]. Studies such as that by Ji et al. demonstrate that G-CSF priming did not change the total number of CD3+ cells in marrow grafts but decreased CD4+ cells and increased CD8+ cells, resulting in a significant reduction in CD4:CD8 ratio [
74].
Besides the interactions with other non-lymphocytic cells, several publications have concluded that G-CSF promotes the generation of Treg phenotype in T cells [
75], which produces IL-10 [
76] and transforming growth factor-β promoting an increase in lymphocyte T helper Th2 differentiation while suppressing Th1 differentiation [
77]. This information was also supported by other publications in which transcriptomic analysis of G-CSF mobilized peripheral blood from donors, revealed an upregulation of Th2 genes, Tregs; and a downregulation of Th17 and Th1 genes [
78]. Overall, these findings suggest that G-CSF negatively affects antigen-specific T cells, and T-cell banking before mobilization might be the best option to optimize T-cell production [
79] because the efficiency of CAR-T cells that are generated from T cells exposed to G-CSF could be reduced
In sharp contrast to all these results questioning the functionality of T lymphocytes in PBMNCs obtained after mobilization are the clinical results of thousands of allogeneic transplants of mobilized PB in which neither graft versus host disease (GVHD) nor graft-versus leukemia (GVL) effect are reduced by mobilization. Moreover, frozen T-lymphocytes from the original mobilized PBMC graft product are used clinically for immunotherapeutic purposes for the generation of functional virus-specific T cells with anti-viral and tumorigenic functions [
80,
81,
82], and more routinely for donor lymphocyte infusions (DLI) as a treatment for recurrent malignant neoplasms. DLI is capable of eradicating minimal residual disease and rescuing hematological decline, being able to induce lasting remissions [
83]. It mainly works well to treat mixed chimerism in which there is a persistent or increasing number of malignant host cells after allo-HSCT, which is a predictor of disease relapse. DLI [
84] has the potential to improve the GVL effect, reducing the risk of relapse in patients with mixed chimerism. With the increased use of unrelated donors for hematopoietic cell transplantation, there is renewed interest in the use of large volumes of frozen mobilized apheresis products that could be the source for DLI as a complementary treatment to prevent or treat the appearance of mixed chimeras.
Considering the benefits previously mentioned and with the knowledge of the clinical success of T lymphocytes after mobilization [
80,
81,
82] researchers have recently opened the debate about its utilization for purposes such as CAR-T cell production. In fact, it is not unreasonable to think that with the current sample processing and enrichment techniques, one could use a determined volume of the mobilized product to isolate solely the T-cell fraction without altering the functionality of these cells[
72]. Some examples using a variety of cytokines and activation protocols for T-cells [
85] can obtain different subpopulations. In addition, other techniques such as Cell Sorting could solve the problems resulting from the effect that other cells like monocytes have on the suppression of T lymphocytes by choosing the optimal CD3+ subpopulation [
86].
Recently, innovative studies have analyzed the effect of the administration of G-CSF to obtain CAR-T cells such as the study conducted by Cummins et al [
11]. They have extensively evaluated this effect through the analysis of cryopreserved apheresis samples derived from 4 healthy donors, both mobilized with G-CSFrh and unmobilized. The study included the cellular culture where T cells were activated ex vivo using CD3/CD28 beads, transduced with a lentiviral vector, and expanded in a specific cell culture media enriched with human serum and IL-7/IL-15. Subsequent analysis by flow cytometry, CyTOF, single-cell RNA sequencing (RNA-Seq), and metabolomics by mass spectrometry showed, in a non-significant manner, a higher expression of the CAR construct in non-mobilized cells. Remarkably, no significant differences were found in terms of the CD4/CD8 ratio obtained in both groups, nor differences resulted in terms of degranulation, cytokine production, and in vitro tumor cytotoxicity assays. In vivo analysis of xenografts in an acute myeloid leukemia mouse model showed no statistical differences in mouse body weight, toxicity, or survival in both analyzed groups. Finally, the RNAseq analysis showed a similar expression transcriptomic profile for both groups. The research indicates that the antitumor efficacy and in vivo toxicity of these products are comparable, with no significant differences observed in the product exposed to G-CSF, as determined through multi-omics analyses
In another recent comprehensive study, Canesin et al [
87] determined the impact on apheresis product of mobilization or not with G-CSF and plerixafor from 30 healthy donors by assessing immune cell composition, T cell phenotype, and T cell functionality in controlling AML tumor growth following anti-CD33 CAR transduction. The resulting in vivo immunophenotypic analysis yields remarkably interesting results showing that mobilization decreases the overall percentage of CD3+ T cells but increases naive (CD45RA+/CCR7+) T cells and decreases the T cell population of effector memory (CD45RA-/CCR7-) and central memory (CD45RA-/CCR7+). In vitro functional cytotoxic assays demonstrated that mobilized-antiCD33-CAR T cells were as effective as non-mobilized-antiCD33-CAR T cells in killing CD33+ AML cells.
In the realm of multiple myeloma treatment, Battram et al [
88] propose utilizing G-CSF-mobilized leukapheresis products to obtain CAR-T cells targeting BCMA. Their study reveals the minimal impact of G-CSF on T cell phenotype, both in vitro and in patients. Pre-treatment with G-CSF does not affect T cell survival or apoptosis during culturing, activation, transduction, and expansion. CAR T production is unaffected, with no impact on cell growth, differentiation, or anti-tumor killing capacity. Exhaustion markers like PD-1, LAG-3, TIM-3, and TIGIT show no significant increase with G-CSF, except for reduced TIM-3 expression in CD8+ cells, this being the only significant difference indicating less exhaustion.
In PBMCs obtained from MM patients through G-CSF mobilized and non-mobilized apheresis [
88], no significant differences exist in the CD4:CD8 ratio or Treg population. G-CSF reduces TSCM cell frequency, specifically CD8+ TSCM cells, without affecting other memory and effector T cell populations. Although CAR transduction is statistically lower in mobilized cells, the high expansion rates and increased T-cell numbers compensate for this reduction. In vivo functionality of CAR-T cells, assessed in a mouse model with MM xenografts, shows similar disease development in CAR-T-treated mice from mobilized and non-mobilized cells. In summary, the study supports the feasibility of using CAR-T cells obtained from G-CSF-mobilized leukapheresis products for treating multiple myeloma. Despite a reduction in CAR transduction, the high cell expansion compensates for this effect, suggesting that these CAR-T cells could be an effective therapeutic option for multiple myeloma.
Another recent retrospective study by Künkele et al [
10] has been carried out using cryopreserved PBMCs mobilized apheresis products derived from 8 patients diagnosed with neuroblastoma. Samples were obtained early in the treatment protocol. To activate and expand the T cells, a monocyte depletion of thawed PBSC units was required due to the outgrowth of monocytes. Furthermore, the CAR transduction efficiency in CD4+ and CD8+ mobilized products to generate the CAR-T cells product ranged from 65-75%. Besides that, flow cytometric analysis showed that cryopreserved G-CSF-stimulated apheresis products contain sufficient numbers of CD4+ and CD8+ T cell precursors with a naïve and central memory phenotype that showed increased replicating potential and high capacity to generate large numbers of effector T cells after tumor stimulation.
Lastly, other research works related to human γδ T Cells such as the study performed by Otto et al [
89], have shown that the PBMCs G-CSF mobilized derived from healthy donors succeeded in retaining their cytotoxicity and in the production of a variety of cytokines. These cells produced considerable amounts of IL-6, IFNγ, TNFα, and GM-CSF, cytokines that are important in mediating cytotoxicity and supporting inflammation. Other cytokines released were IL-4, IL-5, IL-8, IL-13, G-CSF, MCP-1, and MIP-1β, relevant factors that help attract other effector cells. As for the cell phenotype, significant differences were observed in the phenotype of isolated γδ-T cells compared to non-mobilized γδ-peripheral donor T cells. There was increased expression of CD8, CD56, CD28, and CD11b/CD18 (MAC-1) in these isolated cells. This evidence suggests that these cells would have the ability to modulate the immune responses and play a significant role in adoptive immunotherapy in the same way the non-mobilized cells would do.
Overall, these experimental publications have some things in common. As Battram et al mentioned, the way they proceed is different from the late 90’s past experimental approaches due to the reduced exposure time of the G-CSF and the starting material for the ex vivo culture of the CAR-T. This source of starting material should be highly purified T cells and not the whole PBMC fraction [
88]. This would lead to a major advantage thanks to the removal of T cell inhibitory cells such as monocytes CD14+. This particular principle is not against the Advanced Therapy Medicinal Products (ATMP) guidelines of CAR-T therapy. Guidance [
90] of ATMPs indicates that the CAR T cell manufacturing process usually requires 12 days, and starts, in short, with the isolation of T cells from the leukapheresis product of a patient, followed by activation and genetic modification of the cells to express the respective CAR. Thus, given that these cells have to be isolated, cultured, usually sorted by cell phenotype and treated appropriately, it seems reasonable to assume that prior mobilization of G-CSF would not significantly affect the protocol, with these lymphocytes being fully functional as explained throughout this section.
Besides that, as we mentioned previously, G-CSF mobilization products may not be the first indicated option to obtain T-cell fraction, however, it can be seen that cryopreserved stored mobilized PBMCs could perform as well as non-mobilized do in terms of CAR-T production.
3.2. CAR-NK Immunotherapy
One of the main limitations in the development of CAR-NK is obtaining NK cells in massive quantities since these cells are found in a lower proportion (from 5% to 15% of blood circulating lymphocytes) than T or B lymphocytes under normal circumstances [
38]. Source and clinical scale-up of NK cells with long-lasting cytotoxicity activity are the main challenges and the strategies differ depending on whether the sample is freshly isolated, activated, or in vitro expanded because of the variety of phenotypic and functional differences found. Nowadays, the principal source of obtaining NK-CAR for clinical trials is the NK92 [
91,
92] cell line due to the lack of variability that this cell line offers, together with its unlimited proliferation ability and modest cost. Nevertheless, there are some drawbacks including the lack of CD16 expression and higher tumorigenicity risk because is a tumoral cell line[
93]. To circumvent these problems, it seems necessary to find new tools to broaden the source of these cells in order to obtain an “Off-The-Shelf” therapy. Other NK cell sources have been studied, for instance, isolated from PBMCs, UCBs, CD34+ hematopoietic progenitor cells [
8] and induced Pluripotent Stem cells (iPSCs) [
9].
One source being recently quite used is PBMC-derived CAR-NK. There are some fairly good advantages such as the uncomplicated way of collection and the lack of GVHD that allows being isolated from matched and mismatched HLA donors [
94]. Besides that, these collected cells using isolation kits such as CD3+ depletion and CD56+ enrichment, are CD56dimCD16+ which is a more cytotoxic and mature phenotype than NK92 line [
95]. Also, even though they have a reduced capacity for proliferation when compared to other phenotypes, with proper isolation and cell culture these cells can expand quite nicely. Another option is to obtain NK from UCB in the same manner although it can be seen reduced numbers of UCB NK due to the limited volume of UCB units [
96].
Since PBMCs seem to be a good option to obtain NK cells for CAR manufacture, it is not far-fetched to think that, similarly to CAR-T, cryopreserved G-CSF mobilized apheresis products could be used to obtain these cells. Nevertheless, as with CAR-T, the apheresis product mobilized using G-CSF is not a common protocol to obtain the NK cell isolate, since the apheresis mobilization process does not aim to increase the number of lymphocytes or NK cells [
97]. Some studies indicate that the mobilization process can negatively affect the functionality of NK cells [
98], due to the over-proliferation of polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) subpopulations [
99], and the inhibition of IFN-Y secretion by them [
100].
However, there is a debate well reviewed by Gazitt Y. [
101] in which different publications related to cell populations from mobilized PBMCs such as CD34+ cells, T-cells, NK cells and dendritic cells are analyzed including his own experimental results. In this paper is highlighted that there are other studies that do not find differences between the mobilized and non-mobilized products in terms of NK, noticing a lack of consensus. In addition, a great heterogeneity is observed in terms of NK cell mobilization between patients, perceiving that those patients with relatively low NK cell or poor NK activity before mobilization had poor PBSC collection post-mobilization in terms of quantity and functionality. Thus, it is pointed out that these variables, together with the small number of patients included in some of the experimental publications analyzed, should be taken into consideration and the mobilization process itself could not be an issue
A recent publication by Xiong et al [
98], showed significant differences between non-mobilized and mobilized-derived PBSC NK. In this case, mobilized NK fraction showed a CD56bright+ CD16- phenotype population suggesting that G-CSF favors the accumulation of less mature NK cell subsets. Besides that, they observed a decreased expansion rate in mobilized NK compared to non-mobilized NK. However, when IL-15 is added to the culture, it can be observed a progression and restoration in cytokine secretion profile in vitro, suggesting that these cells can experience a maturation. Notwithstanding the potential perception of this aspect as a drawback, expeditious dismissal of its optimization is unwarranted. It is imperative to recognize that presently, one of the preeminent sources of NK-CAR resides in NK-92, a cell line characterized by CD56+CD16- which exhibits low cytotoxicity and in vivo persistence. Besides that, as we mentioned before, the manufacture of CAR requires an optimal culture of the highly purified cells, thus it is reasonable to think that to design a CAR NK cell product, a lower maturation stage and an ex vivo culture modulation through cytokines should be a promising alternative source to certain NK-CAR phenotype [
102,
103].
Otherwise, as NK cells originate from CD34+ hematopoietic stem cells, one feasible way to obtain NK cells is to focus on the source of this population, which encompasses bone marrow or umbilical cord blood. As shown by Oberoi et al. [
104] in their study, the effect of G-CSF can be used to obtain the mobilized CD34+ population also from peripheral blood in massive quantities and in a simple manner, to subsequently differentiate into NK cells. These cells are then expanded and differentiated into mature NK cells using a cocktail of cytokines in a culture system. There have been many advances in the way of culturing, expanding and differentiating functional NKs, for example, by co-culture with K562 feeder cells[
105] that co-express the 4-1BB ligand and membrane-anchored IL-15 and IL-21, giving excellent results and solving the problem found in variability between donors. The resulting CD56+CD3− NK cells are mostly similar to PB NK cells, express NK cell-activating receptors, and exhibit potent cytotoxicity against leukemic cells in vitro and in vivo.
Some studies, such as the one published by Patel et al [
97], have designed and evaluated a CAR-NK development protocol using NK derived from PBSC mobilized with G-CSF. The results derived from these studies achieved the differentiation of NK cells in the culture using cytokines such as IL-7 and IL-15 obtaining cell proliferation and expansion rates equivalent to those obtained from UCB, with a CD56+/CD16+/CD94+ phenotype showed by 10 to 40% of the CD56+ cells. They concluded that large-scale generation of CAR-NK products from PBSC source is feasible and compatible with good manufacturing practice (GMP) which is mandatory for ATMP products, and it is defined as potent products manufactured safely according to standardized methods under controlled, reproducible, and auditable conditions [
39].
Finally, it is remarkable the success of Zhu et al [
106] in the invariant natural killer (iNKT) cell generation, through TCR genetic engineering of peripheral CD34+ HSCs samples mobilized by G-CSF. iNKT are immune system cells with a strong potential to fight cancer. However, its clinical use has been limited due to its scarcity in cancer patients. In this study, they developed a proof of concept for cell therapy using HSC-iNKT, to provide sustained therapeutic levels of iNKT throughout treatment. Using a mouse model with human hematopoietic stem cell grafts and human T cell receptor genetic engineering for iNKT, these researchers demonstrated the efficient and long-term generation of HSC-iNKT cells in vivo.
These HSC-iNKT cells showed similarity to natural human iNKT cells and exhibited multiple mechanisms to attack tumor cells, effectively suppressing tumor growth in mouse models with human tumor xenografts. In addition, preclinical safety studies that were performed revealed no toxicity or tumor formation associated with HSC-iNKT cell therapy. Mobilized HSCs source presents a greater number of iNKT than non-mobilized samples, emphasizing its potential and safety in hematologic malignancy cell therapy.
Overall, similarly to CAR-T Immunotherapy, as CAR-NK manufacture is yet to be explored, it seems necessary to increase the experimental work related to NK cells and to exploit other potential sources available.