Expression Characteristics of Surface Markers of Memory T cells, CD45RO, CCR7 and CD62L, in Tumor-infiltrating Lymphocytes in Liver Cancer Tissues of Patients with Hepatocellular Carcinomas

Methods: TILs from liver tumor and paracancerous tissue (PT) of 30 HCC patients and 30 subjects who were performed hepatectomy due to benign pathogenic lesion (control) were stained with anti-CD3, CD4, CD8, CD45RO, CD62L and CCR7 antibodies. The distribution of four T cell subsets, CD45RO+CD62L+CCR7+T (central memory T cells, TCM), CD45RO +CD62L-CCR7-T (effector memory T cells, TEM), CD45RO + CD62L+CCR7-T (CCR7-T) and CD45RO+ CD62L-CCR7+T (CCR7+T), among TILs was measured with polychromatic flow cytometry and analyzed statistically.


Introduction
After contacting with antigens for the first time, the naïve T cells (Tn) within body recognize antigen presenting cells (APC). After presenting antigens, they are activated and differentiate into effector T cells (T EC ) and under the action of chemokines, they migrate to the antigen presenting sites where they play a role in mediating immune responses. After differentiation from Tn cells to T EC , great changes take place in the expanding capability, chemotactic properties and functions. When the antigens are removed, a large number of T EC undergo apoptosis via Fas-FasL and TNF-α-TNF-R and a part of T EC differentiate into memory T cells (Tmem) [1]. Tmem cells are different from Tn in term of phenotype characteristics, tissue distribution, secretion of cytokines, and effector functions. Traditionally, CD45RA+T cells, also known as Tn cells, i.e. the naive T cells, are thought to have the immunosuppressive functions whereas CD45RO + T cells known as memory T cells (Tmem) have the accessory inductive functions [2]. Based on the differences in homing characteristics and effector functions, CD45RO + Tm can be divided into four subsets: CD45RO + CD62L + CCR7 + T cells, CD45RO + CD62L -CCR7 -T cells, CD45RO + CD62L + CCR7 -Tcells and CD45RO + CD62L -CCR7 + T cells. The first two subsets are the currently more studied, so called central memory T cells (T CM ) and effector memory T cells (T EM ) [3], which are also the focal points of the concern and discussion in this study. However, other two subsets are poorly known and their functions are practically unknown. When Tmem cells face the same antigen again, they are more easily activated to generate antigen-specific T helper cells and cytotoxic T lymphocytes (CTL) and induce rapid, strong, specific and effective immune responses. Thus, they play an important role in protective immunity within the body [4]. At the present, little is known about the roles of different subsets of T cells in anti-cancer immunity. Thus, in the present study, based on different surface markers of different subsets of T cells, we applied polychromatic flow cytometry (PFC) to analyze the expression characteristics of various subsets of tumor-infiltrating memory T cells of patients with HCC, aiming to gain deeper insights into immune state within the tumor microenvironment of liver cancer.

Clinical data
The specimens of liver tissues were all derived from surgically operated patients who were performed hepatectomy from October 2011 to December 2012. A total of 30 patients with liver cancer were pathologically confirmed to be HCC post-operationally. Among them, 23 patients were males and 7 patients were females with the mean age of 45.4 ± 12.6 years, ranging from 35 to 58 years. They had not received any related therapies such as radiotherapy and chemotherapy. Their tumors were completely resected during the surgical operation. 30 patients with liver cancer all have the case history of hepatitis B virus (HBV) infection and were diagnosed with the diagnosis criteria for HBV set by the Ministry of Health of People's Republic of China. The mixed infection with other hepatitis virus was eliminated. Among 30 patients with HCC, the liver tumor tissues of 4 patients had fewer than 5000 infiltrating lymphocytes and thus, they were not included in the statistical analysis (n=26). There were 30 cases in the control group. These patients were performed hepatectomy because of the benign pathological changes but they were not infected by HBV (n=30). Among them, 13 patients were males and 17 patients were females with a mean age of 48.1 ± 12.4 years, ranging from 35 to 64 years. All the investigated patients had no history of autoimmune disease such as diabetes and hyperthyroidism and had no history of immunotherapy. The clinical data included sex, age, HBV infection history, serum alpha-fetoprotein (AFP), hepatitis B two half-and-half check and postoperational pathological data.

Ethic and sampling
The research protocols were approved by The Ethic Committee of Peking University Shenzhen Hospital and all the patients and the volunteer participants signed the Informed consents. Fresh liver tissues were obtained from the patients who were performed hepatectomy and liver tissues were treated within 6 hours. For the patients with HCC, specimens were obtained from the tumor tissues (inside the tumor 1 cm from the edge of tumor tissues) and the paracancerous tissue (more than 5 cm from the tumor tissues) and included in the HBV-HCC group and paracancerous tissue group, respectively. The specimens in the control group were obtained from the normal liver tissues of patients who were performed hepatectomy due to the benign pathological changes.

Measurement methods:
Fresh specimens of liver tissues were put to the culture dish and washed with physiological saline solution to remove the surface blood residues. The bile duct and necrosis tissues were removed with scissors. The fresh specimens of liver tissues were chopped into 1-2 mm 3 pieces and ground or digested with enzymes, and washed repeatedly with PBS solution to collect the cell suspension, filtered with 100/300 mesh screening filter and centrifuged at low speed to remove the cell aggregate and cell debris. The suspension solution was centrifuged at 2000 rpm for 5 min. The supernatant was discarded and the cell pellet was re-suspended with PBS solution. The lymphocytes were isolated with the method of Ficoll density gradient centrifugation. The resulting infiltrating-lymphocyte suspension of liver tissue was washed with 5 ml of PBS twice and centrifuged at 2000 rpm and 1500 rpm successively at room (26°C) for 5 min. The lymphocytes were suspended to 500 µl with PBS. After being counted with hemocytometer, the lymphocyte number in the tissue infiltrating lymphocyte suspension was re-adjusted to more than 1×10 6 /ml. The specimen tube and the control tube were set. 200 µl of infiltrated lymphocyte suspension solution was added to each tube. 5 µl of each of 6 fluorescently labeled antibodies PE/Cy7-CD3, FITC-CD4, PE-CD8, Percp/Cy5.5-CD45RO, APC/Cy7-CD62L and APC-CCR7 was added to the specimen tube. 5 µl of each of PE/Cy7-CD3 and FITC-CD4 and the corresponding isotype control antibodies, APC/Cy7-IgG1, APC-IgG2a, and Percp/Cy5.5-IgG2a, was added to the control tube. The mixtures were mixed gently and incubated at dark for 30 min, then, 1-2 ml of 10% hemolysin optilyse C was added and incubated in dark at room temperature for 5-10 min. The mixtures were washed with the same volume of PBS twice and centrifuged at 1500 rpm for 5 min. The supernatant was discarded and the antibody residues were washed. After being re-suspended with 200 µl of PBS, the lymphocytes were measured with flowcytometry. More than 5000 lymphocytes were acquired and measured with FACSCanto flowcytometry and the results were analyzed with Flow jo software.

Statistical methods
SPSS13.0 software was used for performing statistical analysis of the experimental data and the results of each group were expressed as x ± SD %. The differences of the mean values between groups were tested by t-test. When p value was <0.05, the difference between groups was regarded as statistically significant.

Results
Among 30 patients with HCC, infiltrating lymphocytes isolated from their liver tumor tissues of 4 patients were fewer than 5000 and thus, they were not included in the statistical analysis. When infiltrating lymphocytes of liver tissues were analyzed, they were gated on subsets through forward scatter (FSC) and side scatter (SSC) and then were gated on the subsets of CD3 + CD4 + T cells ( Figure 1A) and CD3+CD8+T cells ( Figure 1B). The CD45RO+Tm cells were gated on CD3+CD4+T cells and CD3+CD8+T cells, respectively ( Figures 1C and 1D). Finally, CD45RO+Tm cells were further sub-grouped according to the expression of CD62L and CCR7 ( Figures 1E and 1F). The upperright corner (where Q2 was located) was CD45RO + Tm CD62L + CCR7 + subset, i.e. T CM subset; The lower-left corner (where Q4 was located) was the CD45RO + Tm CD62L -CCR7subset, i.e. T EM subset; The lower-right corner (where Q3 was located) was CD45RO + Tm CD62L + CCR7subset; and the upper-left corner (where Q1 was located) was CD45RO + Tm CD62L -CCR7 + subset.

Analysis of T CM and T EM subsets of CD4 + CD45RO + Tm cells
The analyzed results of T CM and T EM subsets of CD4 + CD45RO + Tm cells were presented in Table 1, which indicated that CD4 + T CM subset accounted for 0.25 ± 0.17%, 0.29 ± 0.22% and 0.37 ± 0.33% of total CD4 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group and control group, respectively. Multiple comparisons revealed that the differences between HBV-HCC group, paracancerous tissue group, between HBV-HCC group and control group; and between paracancerous tissue group and control group were not statistically significant (P=0.624, 0.089, 0.229; P>0.05, respectively). CD4 + T EM subset counted for 89.73 ± 7.41%, 98.68 ± 0.55% and 95.45 ± 3.66% of CD4 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group and control group, respectively. The multiple comparisons among three groups (P=0.000, 0.000, 0.013, respectively) indicated that the ratio of CD4 + T EM relative to CD4 + CD45RO + Tm in HBV-HCC group was significantly lower than that in paracancerous tissue group and control group (P<0.05). The CD4 + T EM subset is the predominant subset among CD4 + CD45RO + Tm in HBV-HCC, paracancerous tissue group and control groups whereas the ratio of CD4 + T CM relative to CD4 + CD45RO + Tm was very small in all three groups.

Analysis of T CM and T EM subsets among CD8 + CD45RO + Tm
The analyzed results of T CM and T EM subsets among CD8 + CD45RO + Tm were presented in Table 3, which indicated that the ratio of CD8 + T CM subset relative to CD8 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group and control group was 0.41 ± 0.31%, 0.55 ± 0.60% and 0.26 ± 0.25%, respectively. The multiple comparisons between HBV-HCC group and paracancerous tissue group, between HBV-HCC and control group and between paracancerous tissue group and control group (P=0.221, 0.176, 0.010) revealed that the ratio of CD8 + T CM /CD8 + CD45RO + Tm in HBV-HCC group was significantly higher than that in control group (P=0.010, P<0.05). The ratio of CD8 + T EM relative to CD8 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group and control group was 92.39 ± 4.30%, 98.14 ± 0.74% and 98.16 ± 1.05%, respectively. The multiple comparisons between HBV-HCC group and paracancerous tissue group, between HBV-HCC group and control group and between paracancerous tissue group and control group (P=0.000, 0.000, 0.974) indicated that the ratio of CD8 + T EM /CD8 + CD45RO + Tm in HBV-HCC group was significantly lower than that in paracancerous tissue group and control group (P=0.000, 0.000; P<0.05). The CD8 + T EM subset was the predominant subset among CD8 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group and control group whereas the ratio of CD8 + T CM among CD8 + CD45RO + Tm was very small in all three groups. These results were similar to the distribution of T EM and T CM subsets among CD4 + CD45RO + Tm described above.

Analysis of CD62L + CCR7 and CD62L -CCR7 + T subsets of CD8 + CD45RO + Tm
The analyzed results of the ratios of CD62L + CCR7 and CD62L -CCR7 + T subsets relative to CD8 + CD45RO + Tm were presented in Table  4, which showed that the ratio of CD8 + CD45RO + CD62L + CCR7 -T cell subset relative to CD8 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group, and control group was 0.76 ± 0.47%, 0.28 ± 0.24% and 0.83 ± 0.66%, respectively. The multiple comparisons between HBV-HCC group and paracancerous tissue group, between HBV-HCC group and control group and between paracancerous tissue group and control group (P=0.001, 0.578, 0.000; P>0.05) indicated that the ratio of CD8+CD45RO+CD62L+CCR7-T/CD8+CD45RO+Tm in paracancerous tissue group was significantly lower than that in HBV-HCC and control groups and the differences were statistically significant. The ratio of CD8 + CD45RO + CD62L -CCR7 + T cell subset/ CD8 + CD45RO + Tm in HBV-HCC group, paracancerous tissue group and control group was 6.44 ± 4.20%, 1.03 ± 0.45% and 0.75 ± 0.72%, respectively. The multiple comparisons between HBV-HCC group and paracancerous tissue group, between HBV-HCC and control group and between paracancerous tissue group and control group (P=0.000, 0.000, 0.658; P>0.05) revealed that the ratio of CD8 + CD45RO + CD62L + CCR7 -T/ CD8 + CD45RO + Tm in HBV-HCC group was significantly higher than that in paracancerous tissue group and control group. Note: (a) compared to that of paracancerous tissue and control group P<0.05; (b) compared to that of the control group, P<0.05.

been stimulated by antigens, T cells can be divided into naïve T cells, effector T cells and memory T cells. The naïve T cells are those T cells that have just left thymus but have not been stimulated by antigens.
These T cells play a role in immunologic surveillance through re-circulation between blood and the secondary lymphatic organs. It is generally thought that CD45RA is the surface marker of naïve T cells, which is mainly circulated within the lymphatic systems whereas CD45RO is the surface marker of memory T cells, which is evenly distributed among lymphatic and non-lymphatic tissues [5]. Co-stimulatory signals are necessary for the activation of naive T cells whereas the dependence of memory T cells on the co-stimulatory signals such as CD28 and CD154 is significantly reduced and the activation threshold for memory T cells is very low. The second stimulation can occur in the outside of the secondary lymphatic organs (or tissues) such as lymphatic nodes but they do not need the assistant of specific MHC. After in vitro stimulation, the naïve T cells do not generate IFN-γ, IL-4 and IL-5 but only generate lower level of IL-2, indicating that these cells are different from effector T cells and memory T cells not only in phenotypes but also in tissue distribution and functions.
CD45 molecule, also known as leukocyte common antigen (LCA), is encoded by a single gene located in chromosome 1q32. Alternative splicing of the first three consecutive exons (A, B and C) nearby the amino acid terminal of its full length mRNA leads to generation of a CD45 isoform called CD45RO that lacks these three exons. Another CD45 isoform called CD45RA only expresses exon A. Both CD45RA and CD45RO are two isoforms of CD45, which are widely present on the surface of leukocytes and belong to type I trans-membrane proteins. In cytoplasm, they also possess tyrosine phosphatase activity. They play important roles in the development and maturation of lymphocytes, signal transduction and functional regulation. Based on the expression of CD45RA and CD45RO, CD4 + and CD8 + T cells can be divided into naïve T cells and memory T cells. Based on the homing of the receptor of CD62L and the chemokine receptor 7 (CCR7), the memory T cells can be further divided into T CM and T EM subsets. The surface marker of T CM is CD45RO + CD62L + CCR7 + whereas the surface marker of T EM is CD45RO + CD62L -CCR7 -. CCR-7 is a member of the chemokine receptor family. CD62L, i.e. L-selectin, is a member of selectin family. Both CCR-7 and CD62L are involved in the homing of lymphocytes to peripheral lymphatic nodes. T EM is present in the non-lymphatic tissues of the infected sites and plays a role in immunologic surveillance. When they contact an antigen for the first time, under the stimulation of TCR, T EM can rapidly generate IFN-γ, IL-4 and IL-5 and fulfills its responding function, whereas T CM is mainly located in the secondary lymphatic organs and can rapidly proliferate and chemotaxi to the effector sites where they supplement the effector T cells of the surrounding organs to play a role in removing antigens. Even under the stimulation of TCR, T CM does not immediately generate prototypic cytokines such as IFN-γ etc. of the effector cells The mechanisms that control the conversion of effector T cells to the phenotype of memory T cells are still not clearly understood and even very little is known about how T CM and T EM subsets are formed. The prerequisite for the conversion of effector T cells into memory T cells is that they must escape the programmed cell death (apoptosis). Acquiring the survival advantage in time is extremely important for memory T cells. IL-7 is the key factors for the conversion of effector cells into memory T cells. It enables the CD4 + and CD8 + effector T cells to acquire long-tern survival and steady proliferation [6,7]. There are data indicating that CD8+ memory T cells are directly originated from CD8 + effector T cells [8]. These CD8 + memory T cells express effector protein molecules such as perforin, granzyme-B and cytokines and, at the same time, highly express IL-17 receptor and a large amount of anti-apoptotic molecules [9]. In addition to cytokines, the generation of CD8 + memory T cells also requires the assistant of CD4 + T cells. If the assistant of CD4 + T cells is lacking, CD8 + T cells would not be converted to immune memory T cells [10]. Transcriptional factors play important roles in the formation of CD8 + memory T cells. Sullivan et al. [11] found that lacking of transcriptional factor FOXO3 can increase the number of CD8+ memory T cells but does not influence the phenotypes and functions of CD8+T memory T cells, implying that during the clonal expanding period, apoptosis of effector CD8 + T cells can be controlled by regulating the expression level of transcriptional factor FOXO3 to influence the proliferation or vanish of memory T cells and finally influence of the formation of memory T cells.
The mechanisms by which the effector CD4 + T cells are converted into CD4 + memory T cells require to be further investigated. CD4 + T cells are a complex immune regulatory network and the main members include helper T cells 1 (Th1), Th2, Th17, follicular helper T cells (Tfh) and regulatory T cells (Treg). It has been demonstrated that Th1-like effector cells can generate Th1 memory cells and these Th1 memory cells possess the "traditional characteristics" of T EM [12]. Similarly, there is evidence demonstrating that after the contraction phase of immune responding reactions, Th2 effector cells can form T EM [13]. While it is not clear how Th17 effector cells enter the memory cell pool, there are many lines of evidence supporting that the in vitro derived Th17 cells could become IFN-γ-secreting cells after being introduced into mice [14,15]. These studies have indicated that Th1 effector cells, Th2 effector cells and Th17 effector cells can be possibly converted into CD4 + T EM . Currently, there is very rare evidence for the entrance of the transcriptional factor FOXP3-expressing Treg cells to the memory T cell pool, because there is experimental evidence confirming that FOXP3 + cells are not present in the memory T cell populations after infection [16]. However, another study reported that Treg cells had lost their expression of FOXP3 before they became memory T cells [17]. Regarding to the origin of T CM , several investigators believed that Tfh cells that have the characteristics of expressing chemokine receptor CXCR5 was one of the origins for T CM [18,19] whereas other investigators believed that there was the existence of precursor cells of T CM [20,21]. Additionally, B cells also play similarly important roles in the generation of memory T cells and in "determining" T CM -T EM [22][23][24].
In this study, we found that among the tumor infiltrating CD45RO + Tm cells, T EM subset was the predominant subset in both CD4 + -CD45RO + Tm cells and CD8 + -CD45RO + Tm cells whereas the proportion of T CM subset was very small. This may be related to the heterogeneity of the generation processes and tissue distribution of T EM and T CM . After leaving the thymus, naïve T cells express CCR7 and CD62L and are re-circulated in blood and the secondary lymphatic organs to perform the immune surveillance function. After being activated by the antigen-presenting dendritic cells (DC), naïve T cells, CD28 enhances the activation of T cells and the activated T cells migrate to the target organs where they play the effector functions. After the antigens have been removed, more than 90% of effector T cells are removed by the mean of apoptosis and only 5% to 10% of the effector T cells differentiate into resting memory T cells. Among which, a part of memory T cells also express CCR7 and CD62L, i.e. T CM subset, and are homing to lymphatic nodes. Another part of memory T cells, i.e. T EM subset, are selectively remained in the tissues [25]. T EM cells can also express high levels of α1-and α2-integrin. Integrin plays a role in mediating the adherence of T cells to extracellular matrixes and is a cell surface receptor required for the entrance of T cells into the inflammatory tissues. The CCR7-Tm including T EM cells highly express the tissue-specific homing receptorsCD103 and CLA that help T cells to enter the normal or inflammatory tissues and also express inflammatory cytokine receptors that have the activity of chemotactic attraction. These receptors stimulate T EM to enter non-lymphatic tissues [26]. In this study, we observed that the ratios of CD4 + -and CD8 + T EM subsets in CD45RO + Tm of liver tissues were very high, indicating that they play the major effector roles. In contrast, T CM express two lymphatic node-homing receptors, CD62L and CCR7. Thus, T CM are selectively homing to lymphatic nodes and thus, their distribution in liver tissues is much smaller.
Our study has also demonstrated that among the tumor infiltrating CD45RO + Tm in HCC patients, the ratios of CD4 + T EM and CD8 + TEM relative to CD45RO + Tm in liver tumor tissues were significantly lower than those in non-tumor tissues whereas the ratios of CD4 + T CM and CD8 + T CM relative to CD45RO + Tm in liver tumor tissues were not different from those in non-tumor tissues. T CM are homing to the secondary lymphatic organs. They play a major role in mediating reactive memory [27]. When they contact the antigen for the second time, they proliferate again and differentiate into effector T cells but play little effector role and some T CM do not have effector function at all. Different from T CM , T EM cells migrate to the surrounding tissues and are mainly involved in mediating protective memory [27]. When they are stimulated by the antigen again, they rapidly generate effector factors and immediately confer effector functions. For example, CD4 + T EM and CD8 + T EM can rapidly generate IFN-γ, IL-4 and IL-5. CD8 + T EM can also immediately secret perforin to play a killing role. In the HBV-liver cancer tissues, the ratios of CD4 + T EM and CD8 + T EM relative to CD45RO + Tm were significantly smaller than those in non-tumor tissues, indicating that under the influence of tumor micro-environment, the number of T EM was significantly reduced and their ability to immediately confer effector function is reduced when they face secondary response and are unable to play a role in killing tumor cells. In patients with HCC, the reduced number and the decreased functions of T EM cells are related with several factors. The up-regulation of FoxP3 + Treg cells in cancerous tissue during the progression of HCC significantly was found to inhibit the tumor infiltratingCD8 + T cells [28]. The functions of cancerinfiltrating CD8 + T cells are also influenced and they express lower levels of IFN-γ and perforin [29]. These effects are related to the action of the cytokine, transforming growth factor-β1 (TGF-β1), produced by HCC cells because TGF-β1 up-regulates the expression of FOXP3 gene [30]. While the mechanisms involved are currently unknown, they may be related the tumor immune suppression.
Immune memory is one of the important characteristics of acquired immunity in higher animals. When the animals contact the same antigen again, they rapidly generate effective immune responses, which are accomplished by a series specific memory T cells. Thus, the memory T cells have received much concern from the researcher circles and drawn increasing attentions. Compared to the naïve T cells, the memory T cells have lower activation threshold and stronger effector functions. During the process of preventing the reinfection of pathogens, memory T cells respond rapidly, functioning strongly and play an important role. Thus, the current studies on memory T cells have mainly focused in the fields of infectious diseases such as viral hepatitis, AIDS and tuberculosis etc. [31]. Tumor infiltrating lymphocytes (TIL) isolated from tumor tissues are thought to have stronger capability of killing tumor cells. It has been found that the prognosis for the HCC patients whose tumor tissues have significant infiltrating lymphocytes is more superior over that of those HCC patients who had almost no infiltrating lymphocytes in liver tumor tissues [32]. This observation indicates that TILs play an important role in regional immune microenvironment within tumor tissues. Analysis of the distribution characteristics of the tumor infiltrating memory T cell subsets is helpful for a better understanding of the state of tumor immune microenvironment and for judging prognosis. After being activated, adoptive infusion of the tumor-specific memory T cells isolated from TILs into the tumor-carrying patients will increase the effectiveness of the adoptive immunotherapy because these cells have already been sensitized and activated by autologous tumor cells. Vaccinations against HBV have already been one of the most effective measures for preventing the invasion of HBV. However, for patients who have already been infected with HBV, the risk of developing HCC is still very high. Perhaps, in the near future, we are able to prepare the HCC-related antigens, use them to vaccinate the susceptible populations and generate T EM cells existing in the liver tissues for a long run, preventing the occurrence of liver cancer. In the cancerous tissue of patients with HCC, T EM cells are in the immune suppressive state. Further study to understand the mechanisms underlying the suppressive state of T EM cells will be helpful for us to change this state and to enhance the anti-cancer immune effects of T EM cells. Additionally, the in vitro culturing the functionally competent T EM cells and transplanting them to the patients with HCC may also be a promising immune therapy method for HCC in the future. For instance, Klebanoff et al. [33] performed adoptive cell transfer (ACT) with T CM , T EM and T SCM (stem cell memory cells) to mouse B16 melanoma and obtained relatively satisfactory results, i.e. ACT of these cells caused a significant regression of tumors. Their further study [34] also has indicated that it is the CD8 + T CM rather than CD8 + T EM cells that have more potent anti-tumor effects. Of courses, only a few studies have been conducted on memory T cells in the cancer fields, many unknown mechanisms remain to be further elucidated by the cancer researcher circles.