Immunophenotype and antitumor activity of cytokine-induced killer cells from patients with hepatocellular carcinoma

Background Cytokine-induced killer (CIK) cells are heterogeneous lymphocytes from human peripheral blood mononucleated cells (PBMCs) co-cultured with several cytokines. The main purpose of this study is to evaluate the functional characteristics and anticancer ability of CIK cells from hepatocarcinoma (HCC) patients. Methods CIK cells were activated ex-vivo and expanded from PBMCs from HCC patients. The immunophenotype and the ex-vivo killing ability of CIK cells were evaluated. Human CIK cells were intravenously injected into NOD/SCID mice to evaluate the in vivo anticancer ability. Results More than 70% of CIK cells were CD3+CD8+, and 15%–30% were CD3+CD56+. These cells expressed an increased number of activated natural killer (NK) receptors, such as DNAM1 and NKG2D, and expressed low-immune checkpoint molecules, including PD-1, CTLA-4, and LAG-3. Among the chemokine receptors expressed by CIKs, CXCR3 and CD62L were elevated in CD8+ T cells, representing the trafficking ability to inflamed tumor sites. CIK cells possess the ex-vivo anticancer activity to different cell lines. To demonstrate in vivo antitumor ability, human CIK cells could significantly suppress the tumor of J7 bearing NOD/SCID mice. Furthermore, human immune cells could be detected in the peripheral blood and on the tumors after CIK injection. Conclusions This study revealed that CIK cells from HCC patients possess cytotoxic properties, and express increased levels of effector NK receptors and chemokine molecules and lower levels of suppressive checkpoint receptors. CIK cells can suppress human HCC ex-vivo and in vivo. Future clinical trials of human CIK cell therapy for HCC are warranted.

Introduction Hepatocellular carcinoma (HCC) is the sixth most common cancer and the third leading cause of cancer-related death with a particularly high prevalence in Asia. Surgical resection and other local therapeutic strategies are considered curative therapies for localized disease, however, more than half of these would recur even with this type of curative treatment [1]. The results of other treatments for palliative purpose remain a huge unmet need despite the recent success of targeted and immunotherapy in a small fraction of patients [2]. HCC patients are often found to have functional deficiencies in host adaptive and innate immunity responses [3]. Most of the patients affected by HCC in Taiwan are associated with viral infections, such as the HBV or HCV. Evidence from PD-1 blockade therapy has suggested that viral-associated HCC is also associated with better responses [4,5]. This implies that the immune systems of these patients may harbor certain antitumor responses [5,6]. Adoptive immunotherapy that expands and rejuvenate autologous immune cells may be potentially used as an efficient adjuvant anticancer therapy for patients with HCC [7].
Cytokine-induced killer (CIK) cells are a heterogeneous subset of ex-vivo expanded T lymphocytes which express both the T-cell marker CD3 and NK cell receptors [8]. Patient peripheral blood mononuclear cells (PBMCs) are stimulated and cultured with a cytokine cocktail comprising interferon-γ (IFN-γ), interleukin-2 (IL-2), anti-CD3, and/or IL-1 for 2-4 weeks. These cells are able to expand up to 1000-fold. This expansion allows an easy adoptive transfer of large amounts of cells compared with conventional adoptive cellular therapy [8,9]. The final products of these cells are usually CD3+CD8+ T cells, among which 15%-30% are also CD56+ [10]. Interestingly, most of these cells will express NK cell surface receptors, such as NKG2D and DNAM-1 [11][12][13]. Prior studies have confirmed that CIK cells may kill tumor cells according to a nonmajor histocompatibility-complex (MHC)-restricted mechanism, thus suggesting the importance of NK surface receptors in mediating the killing process [11,[13][14][15].
Preclinical data and initial clinical studies support CIK cells as a promising cell population for adoptive immunotherapy [16][17][18]. The clinical efficacy has a broad coverage, especially on hematogenous malignancy [19,20], melanoma [21], lung [16], and HCC, which are among the most reported types of cancers [12,22]. A few clinical trials have shown that CIK cell therapy could prolong survivals in HCC in metastatic or adjuvant settings [23,24]. It is interesting that treatments with CIK cells have also been associated with decreases in HBV viral loads [25]. Recently, a Korean, randomized, phase-3 clinical trial showed adjuvant adoptive CIK cell therapy for patients who underwent curative treatment for HCC could increase recurrencefree and overall survival [24]. A renowned worldwide cancer treatment guideline has adopted this result, and this CIK therapy has become a recommended treatment for HCC adjuvant therapy [26].
The clinical benefit of CIK therapy has been documented, however, there is uncertainty about common applications of this strategy to routine practice. Key issues for clinical application of adoptive immunotherapy include the identification of methodologies to scale-up the expansion process to meet quantity and quality requirements for immunotherapy; the expressed markers that could represent non-MHC restricted cytotoxicity include the CD56 molecule, other candidate markers and the ability for trafficking into tumor sites. In this study, we evaluated the immunophenotype and ex-vivo and in vivo killing ability of CIK cells based on a modified protocol established in our laboratory. This work sets up a reference laboratory standard and can be applied as a quality index for future clinical level manufacturing.

Study subjects
We collected PBMCs samples form adult HCC patients for the expansion of CIK cells. All of the HCCs were diagnosed by cytological or pathological evaluations, or according to the criteria of the American Association for the Study of Liver Disease [27]. We excluded patients with severe alcoholic hepatitis or acute or chronic liver failure, non-HBV, HCV, or alcohol-related HCC, and patients with concurrent evidence of sepsis.

Ethics statement
This study was approved by the Institutional Review Board of Chang Gung Medical Foundation. Approval from the institutional review board was obtained for the analysis of this series (IRB: 201600271B0). All participants provided their written informed consent to participate in this study. All animal procedures followed the Guide for the Care and Use of Laboratory Animals as promulgated by the Institute of Laboratory Animal Resources, National Research Council, National Academy of Science (United States), and were approved by the Animal Care and Use Committee of the Chang Gung Memorial Hospital at Linkou Medical Center. In all mice, tissue collection procedures were initiated after mice had either been euthanized or were under deep anesthesia and unresponsive to all stimuli, and all precautions were taken to minimize suffering. Mice were placed into a chamber filled with vapor of the anesthetic isoflurane until respiration ceased followed by 100% carbon dioxide into a bedding-free cage initially containing room air with the lid closed at a rate sufficient to induce rapid anesthesia, with death occurring within 2.5 minutes.

Isolation and culture of CIK cells
PBMCs were isolated from HCC patients based on the Ficoll-Paque plus density gradient. The cells were washed twice with RPMI-1640. PBMCs were cultured with RPMI-1640+10%FBS medium with anti-CD3 antibody (50 ng/ml), IL-1α (100 U/ml), and IFN-γ (1,000 U/ml) at 37˚C with 5% CO 2 for 24 h. From days 2 and 3, IL-2 (300 U/ml) was added to the medium. We then changed the IFN-γ and IL-2 containing medium every 5 days. On day 14, CIKs were harvested (viability: 90%), and the phenotypes and killing abilities of CIK cells were evaluated. The preparation of CIK cells is illustrated in

Cytotoxicity assessment
The cytotoxicity of CIK cells was tested from the HCC patients. Cell lines K562, HepG2, and J7 from the American Type Culture Collection were selected as target cell lines. For K562 and HepG2 cytotoxicity assessments, we used flow cytometry to determine the expression Annexin V and 7-AAD with different effector-to-target ratios (from 1:10 to 20:1) based on different killing capacity. For the J7, target cells (1 × 10 5 cells/mL) and effector cells were incubated for 4 h at a ratio from 5:1 based on the cytotoxic ability and cells harvesting. Subsequently, 50 μl samples of culture supernatant were loaded into a 96-well plate with 50 μl of LDH substrate mixture, and incubated in the dark for 30 min at room temperature (detection for cell lysis). Subsequently, 50 μl of stop solution was added to each sample to stop the reaction, and the absorbance was recorded at 490 nm.

Blocking assay
Antibody-mediated blocking of K562 cell death was performed. Solitary and synergistic use of anti-DNAM-1(BD Bioscience) and anti-NKG2D (R&D Systems) antibodies were explored to evaluate whether they could decrease the cytolytic capacity of CIK cells towards K562 cells. The target cells were incubated with blocking antibodies at a concentration of 20μg/ml for 20 minutes before co-culture with effector cells. K562-only control samples were incubated with isotype control antibodies (BD Bioscience, R&D Systems). These samples were then co-cultured with the effector cells for 4 hours and evaluated by the Annexin V apoptosis kit (BD Bioscience) assisted by flow-cytometry analysis.

In vivo cytotoxicity in animal experiments
NOD/SCID mice were purchased from the National Laboratory Animal Center and were bred in SPF conditions. Mice at the 6 th week of age were used for experiment. In total, 1.33 × 10 6 human J7 cells were diluted in 100 μl HBSS and injected into the flank's subcutaneous tissues of these mice. When the tumor grew up to at least 0.2 cm, mice were injected either with human CIK cells (1 × 10 7 in 400 μl PBS through the tail vein) or with 400 μl PBS every 4 days (N = 5 each). Seven cycles of injection were performed at most. The sizes of the tumors were measured at each injection. After 4 weeks, mice were sacrificed. The tumors infiltrating lymphocytes were analyzed for the presence of human CIK cells, and paraffin embedded specimens were used to detect the presence of human CD4 and CD8 T cells. All experiments were approved by the Animal Care and Use Committee of the Chang Gung Memorial Hospital, Linkou.

Statistical analysis
Continuous variables are expressed as median [interquartile range (IQR)]. The nonparametric Mann-Whitney U-test was used to compare continuous variables between unmatched groups. When more than 20% of data presented an expected frequency < 5, the Fisher's exact test was substituted for the Chi-square test. The Wilcoxon signed-rank t-test was used when two related samples, matched samples, or repeated measurements on a single sample were compared to assess whether their population means ranked differently. Calculations were made using GraphPad Prism (version 5.0, GraphPad Software, Inc). p-values < 0.05 were considered statistically significant.

Phenotypical characteristics of human CIK
CIK cells generated from freshly isolated PBMCs of 10 HCC patients were successfully expanded ex-vivo for 14 days per standard protocol with additions of IFN-γ, anti-CD3, and IL-2. The CIK cells from 5 patients were studied for the immunophenotype and the ex-vivo killing ability. The CIK cells from other 5 patients were tested for in vivo mice model.

Cytotoxicity of human CIK cells
The ex-vivo cytotoxicity of CIK cells was evaluated using different human cancer cell lines, including the human myelogenous leukemia cell line K562, hepatoma cell line HepG2, and J7 as the target cell lines. As shown in Fig 4A, these CIK possessed cytotoxicity properties to the MHC-deficiency, NK-sensitive K562, and MHC expression HepG2 in terms of induced cell death assessed based on the expressions of Annexin 5 and AAD7 in a dose-dependent E:T ratio manner. By contrast, the cytotoxicity of J7 ATP methods used for assessments yielded a modest cytotoxicity ability at 5:1 E:T ratio compared with positive controls. Furthermore, we found that synergic use of anti-DNAM-1 and anti-NKG2D antibodies decreased the cytolytic ability of CIK cells obtained from HCC patients, suggesting that CIK cells may display cytotoxicity in a DNAM-1 and NKG2D-dependent manner (S4 Fig). We also evaluated the in vivo antitumor ability, wherein human CIK cells were intravenously infused into NOD/SCID mice (N = 5), which were subcutaneously implanted with the J7 HCC cell line (Fig 4B). In the control group, the tumor grew rapidly from a volume of 113 mm 3 to 969 mm 3 after it was treated with seven wash cycles with PBS. However, the tumor in the CIK group grew slowly from 59 mm 3 to 149 mm 3 after 7 cycles of CIK cells (Fig 4B). By contrast, the adoptive transfer of human CIK cells every 4 days suppressed the J7 tumor growth on NOD/SCID mice compared with the control (PBS) group (p < 0.05, Fig 4B). In addition, human immune cells could be detected in the peripheral blood (1.6% of murine cells) and in the tumor microenvironment (15.8% of mouse CD45+CD3+ cells) of NOD/SCID mice after CIK treatment, as evaluated by flow cytometry (Fig 4C). Furthermore, human CD8 and CD4 cells could be detected in the tumor microenvironment by immunohistochemical staining (Fig 4C). All these results demonstrated that these CIK cells execute cytotoxicity against human tumor cells in vitro, and also migrate to the tumor microenvironment to inhibit the growth of the tumors in vivo.

Discussion
This study focused on the assessment of the immunophenotype and cytotoxicity ability of CIK cells from patients with HCC. The CIK preparation protocol we adopted produced a significant number of cells up to the 14 th day of culture. We characterized a deeper effector functional profile of CIK cells from patients with HCC by exploring additional receptor molecules potentially involved in the interaction with tumor targets and the microenvironment. These cells could express high levels of activated NK receptors, including NKG2D and DNAM1. Concerning the immune checkpoint receptors that would dampen the cytotoxic immune response, CIK cells expressed low levels of CTLA-4, PD-1, and LAG-3. Among the chemokine receptors, CD62L and CXCR3 were enhanced in the CD8+ T cells of CIK, thus justifying the trafficking potential to inflamed sites. We also demonstrated the ex-vivo and in vivo cytotoxic abilities of CIK to suppress tumor cells. In addition, human immune cells could be detected in the peripheral blood and on the tumors of NOD/SCID mice after CIK treatment. This is rare in the literature and provides animal model data prior to the initiation of the human CIK study or clinical use [28]. The preclinical data support the potential of adoptive CIK therapy for HCC patients.

PLOS ONE
soluble proliferating factors or conducts contact-dependent signals to boost T-cell proliferation [30][31][32]. The variation might have a negative impact on clinical use especially for patients with low-expansion rates [33]. Our data indicated that there were still multifold variations from patients with HCC. However, most PBMCs from patients could be easily expanded at small scales within a laboratory facility. We anticipate that for future clinical level expansion with Good Manufacture Product (GMP) facilities, most HCC patients could still provide a reasonable amount of CIK cells for treatment.
CIK cells were expanded from PBMCs by a nonspecific protocol to enrich CD3 positive Tcell proliferation and activation. This indicates that the CIK cells are a pool of T cells with naive and memory subpopulations without antigen specificity. Previous studies have suggested that the killing mechanism of CIK is nonrestricted MHC [11]. This is particular important as there is increasing evidence that the presence of bystander T cells in the tumor microenvironment may play an important role in mediating cellular immune responses against tumor [34][35][36]. The concept of cooperation with the adoptive immune cells (whether specific or bystander cells) with endogenous antigen-specific cytotoxic cells that turn the tumor microenvironment to a controlled tumor growth microenvironment is probably more realistic, particular in the era of immune checkpoint blockade therapy [37]. Many have suggested that the bystander killing of CD8+ T cells is driven by the cytokine stimulation in conjunction with NK surface receptor-mediated killing. The K562 cell line does not express the MHC complex, thus further supporting the possible killing mechanism via NK receptors [38]. In the past, the CIK population was categorized by a certain CD56 percentage of expression, a hall marker for NK cells [39], however, our data suggest and are in agreement with others that the presence of NKG2D or DNAM-1 is probably more relevant to the killing ability [13,40]. The characterization of the expression profile of NK receptors should be essential for future clinical applications other than those related to CD56.
As T-cell exhaustion is a hall marker for tumor immune suppression after protracted T-cell activation. In addition, upregulation of immune checkpoint molecules indicates T-cell exhaustion, and particularly PD-1 expression [41]. Even though with anti-CD3 activation and IL-2 expansion, the CIK cells from our laboratory express very low levels of PD-1, CTLA4, and Tim-3, outcomes are compatible with literature publications. The significantly increased expression of LAG3 is uncertain in CIK, while the role of LAG-3 on CD8+ T cells may comprise suppressive functions [42]. However, whether anti-LAG3 Ab could potentiate the effect of CIK deserves additional studies. Our data suggest that CIK cells do not possess an exhausted phenotype that could potentially maintain their activation in vivo. This was further confirmed in our in vivo animal study which confirmed the presence of human CIK cells in both PBMCs and the tumor microenvironment.
CIK cells are from PBMCs and the circulating cells are usually cells with central memory phenotype that are not ready to enter the inflamed site. Our analysis found that CIK cells express higher levels of CD62L and CXCR3, and lower CCR7 levels. The presence of CD62L suggested that many of these cells remain in naïve status, but the presence of CXCR3 indicates that these cells are able to migrate to inflamed sites [43]. Lower CCR7 expressions suggest there is not central memory [44]. In combination with the animal experiment, our data support the fact that CIK cells are inflamed tissue-homing T cells with effector functions.
The role of CIK for HCC was promising. Hui et al. reported that CIK therapy significantly improved disease-free survival, but not the overall survival in a randomized trial [23]. Weng et al. used intrahepatic artery CIK infusions in a randomized study as adjuvant therapy, and demonstrated a benefit on recurrence free survival [45]. While a prospective randomized trial by Yu et al. reported significant improvement in OS and DFS by CIK cells in HCC patients who were not candidates for surgical intervention [46]. A meta-analysis summarized these data and validated the benefits on DFS for HCC CIK cells after local treatment [47]. However, some clinicians would criticize the methodology, study design, or heterogenous patients that may diminish the value of these results. The Korean study enrolled only stage I and II HCC patients who had already completed local treatments. The results showed that in the 5-year follow up, that there was a 37% death rate reduction in favor of adoptive immunotherapy with CIK on 230 patients. These encouraging results suggested that HCC should be one of the most valuable cancer types for CIK clinical utilities.
The mice model is an important issue for human anti-cancer study. NOD/SCID mice contain impaired T cell, B cell lymphocyte and NK cell function, and lack macrophage and dendritic cell activity as well as reduced complement activity. Therefore, NOD/SCID mice exhibit a sufficient ability to support engraftment with human cancer tissues including liver cancer [48][49][50][51][52], and with hematopoietic cells [53][54][55]. It provides a reliable model of anti-cancer therapy screening. Although, many cancers still fail to engraft efficiently in NOD/SCID mice, largely due to remaining NK cell activity and other residual innate immune function. This leakiness may be eliminated in the RAG-1 and RAG-2 mice, or especially the NSG (NOD-SCI-D-IL2γ −/− ) mice, produced by the targeted mutation of the IL-2 receptor γ-chain locus in a previously bred NOD-SCID strain, lacking T cells, B cells, macrophages, NK and NKT cells, may also be platforms to overcome the human engraft failure. Beyond the endogenous immune cells, xenograft mouse models affect the tumor growth includes the influences on immune cell infiltration and angiogenesis. The immune cell activation patterns or antigen expression between the mouse and the human immune system were other hurdles. Advances in humanized mouse models are remarkable but are not yet fully established in HCC research [48]. Nevertheless, immunosuppressive cells such as myeloid-derived suppressor cells (MDSC), T-regulatory cells and M2 macrophages could also play an important role in the interplay between a patient's tumor and the immune system. One previous study showed human CIK cells could be affected by MDSC [56]. Taken together, there are our limitations to establish an artificial environment to define the exact anticancer effect of human CIK cells.
The limitation of our study was that we did not conduct this protocol at a large-scale level and used the GMP facility. This pivotal study is a proof-of-concept, and the proof of feasibility aims to define the immunophenotype of CIK cells and validate their cytotoxic properties. The large-scale, GMP facility production protocol uses a stricter procedure, high-quality medium, and cytokines for human use purposes. The culture density will be much lower than the current laboratory that allows an increased expansion capacity for these cells. Therefore, we expect that our next preclinical preparation would produce folds of CIK cells with identical immune phenotypes. Additionally, there is still a concern on whether the patient characteristics will affect the quality and quantity of CIK cells given the observed variation in the CIK cell population. We found a variation in the expansion of fold changes among different patients. However, the number of patients was still limited to allow us to conclude any potential factors associated with the variation. For the immunophenotype, it seems that the variation was much lower, thus suggesting that the quality of CIK cells could be maintained among different individuals. We only enrolled HCC patients. Whether the CIK population is limited to HCC patients remains elusive, despite prior report findings that supported the fact that the CIK treatment series is not limited to HCC. However, we still need to enroll patients affected by other cancer types to validate the applicability of this protocol to other cancer types. By using a mathematical approach, one study suggested that calculation of CD8+ or CD4+ T-cell percentages on early days could predict the CIK expansion quality [57]. Other strategies were used to improve the expansion and function of CIK cells, such as the addition or replacement of cytokines, like IL-7, IL-15 or thymoglobulin, instead of IL-2 [58][59][60][61], addition of allogeneic APCs, depletion of regulatory T cells, or pharmacological interventions [15,62].

Conclusions
In summary, we characterized the immunophenotype of CIK cells and proved its cytotoxic ability against human cell lines. We set up the quality control parameters for the CIK cells for future clinical applications other than the classical CD56 surface marker. This study further supports the possible mechanism of the killing effects of CIK cells as a group of highly activated, inflamed site-trafficking bystander T cells. The cooperated potential with antigen-specific T cells highlights an important role of adoptive CIK therapy for the future clinical application for HCC.