|Home | About | Journals | Submit | Contact Us | Français|
Human Vγ2Vδ2 T cells play important role in immunity to infection and cancer by monitoring self and foreign isoprenoid metabolites with their γδ T cell antigen receptors. Like CD4 and CD8 αβ T cells, adult peripheral Vγ2Vδ2 T cells represent a pool of heterogeneous cells with distinct functional capabilities.
The aim of this study was to characterize the phenotypes and functions of various Vγ2Vδ2 T cell subsets in patients with nasopharyngeal carcinoma (NPC). We sought to develop a better understanding of the role of these cells during the course of disease and to facilitate the development of immunotherapeutic strategies against NPC.
Although similar total percentages of peripheral blood Vγ2Vδ2 T cells were found in both NPC patients and normal donors, Vγ2Vδ2 T cells from NPC patients showed decreased cytotoxicity against tumor cells whereas Vγ2Vδ2 T cells from normal donors showed potent cytotoxicity. To investigate further, we compared the phenotypic characteristics of Vγ2Vδ2 T cells from 96 patients with NPC and 54 healthy controls. The fraction of late effector memory Vγ2Vδ2 T cells (TEM RA) was significantly increased in NPC patients with corresponding decreases in the fraction of early memory Vγ2Vδ2 T cells (TCM) compared with those in healthy controls. Moreover, TEM RA and TCM Vγ2Vδ2 cells from NPC patients produced significantly less IFN-γ and TNF-α, potentially contributing to their impaired cytotoxicity. Radiotherapy or concurrent chemo-radiotherapy further increased the TEM RA Vγ2Vδ2 T cell population but did not correct the impaired production of IFN-γ and TNF-α observed for TEM RA Vγ2Vδ2 T cells.
We have identified distinct alterations in the Vγ2Vδ2 T cell subsets of patients with NPC. Moreover, the overall cellular effector function of γδ T cells is compromised in these patients. Our data suggest that the contribution of Vγ2Vδ2 T cells to control NPC may depend on the activation state and differentiation of these cells.
Human γδ T cells are a unique subset of T cells that constitute 1–5% of peripheral blood T cells. The Vδ2 chain of the T cell receptor (TCR) usually pairs with the Vγ2 chain to form the Vγ2Vδ2 TCR (also termed Vγ9Vδ2 TCR), which is expressed by the majority of circulating γδ T cells [31, 32]. Vγ2Vδ2 T cells are present in low numbers at birth but expand during infancy due to environmental stimuli . Preferential expansion of Vγ2Vδ2 T cells has been observed during a number of microbial infections . The stimulating microbial antigens have been shown to be small nonpeptide metabolites (commonly termed phosphoantigens) essential for isoprenoid biosynthesis. We and others have demonstrated that the most potent natural microbial antigen for Vγ2Vδ2 T cells is (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP) [21, 37, 39]. HMBPP is a microbial metabolite in the methyl erythritol phosphate pathway responsible for isoprenoid biosynthesis that is used by most Eubacteria and Apicomplexan parasites . The major human antigen reported for Vγ2Vδ2 T cells is isopentenyl pyrophosphate (IPP) , which is increased in some malignant cells and in almost all cells upon pharmacological treatment with bisphosphonates [16, 52] or alkylamines.
Unlike CD4 and CD8 αβ T cells, the majority of adult peripheral blood Vγ2Vδ2 T cells express various phenotypic markers generally found on memory cells . Similar to CD4 and CD8 αβ T cells, these memory Vγ2Vδ2 T cells are heterogeneous and multiple cell subsets can be distinguished based on their maturation surface markers and functional activities . Various cell surface receptors have been used to divide memory cells into subsets. One criterion primarily used for human CD4 αβ T cells has been the expression of the CCR7 chemokine receptor, which is required for lymph node homing [41, 42]. CD4 central memory T cells (TCM) express CCR7, whereas effector memory cells (TEM) do not. In addition to CCR7 , we and others have found that the expression of CD27, CD28, CD45RA, and CD45RO molecules also serves to distinguish memory γδ T cell and CD8 αβ T cells subsets [1, 2, 19, 50]. Central memory CD8 αβ T cells (TCM, also termed early memory cells by Appay) were identified as CD28+ CD27+ CD45RA− cells and also included early effector cells  that secrete IFN-γ and TNF-α but have little or no cytolytic activity due to low perforin expression [19, 50]. On rare occasions, CD8 αβ T cells in the central memory pool lose CD27, but retain CD28 expression as well as other central memory cell properties (TCM 27-) . In contrast, effector memory T cells (TEM, also termed intermediate memory cells) lack CCR7 and CD28, but continue to express CD27. Additionally, TEM also produce low levels of cytokines, but intermediate levels of perforin . Late effector memory T cells (TEM RA, also termed CD45RA+ late memory cells) lack both CD28 and CD27 expression and represent terminally differentiated effector T cells that express a high level of perforin, but demonstrate limited proliferative capacity.
Upon activation, Vγ2Vδ2 T cells kill a broad range of tumor cells, including lymphoma cells, leukemia cells, and other malignant hematological cells. Vγ2Vδ2 T cells can also kill a variety of solid tumors, including nasopharyngeal carcinoma , breast carcinoma , hepatocellular carcinoma , lung carcinoma , renal cell carcinoma [6, 24, 53], pancreatic adenocarcinoma , prostate carcinoma , and neuroblastoma . Moreover, adoptively transferred Vγ2Vδ2 T cells exhibit anti-tumor activity in various xenograft mouse models [22, 30, 57]. There is increasing evidence for the use of Vγ2Vδ2 T cells in cancer immunotherapy. In one study, stimulation of Vγ2Vδ2 T cells with the bisphosphonate drug pamidronate in combination with IL-2 led to partial remission and stable disease in some patients with refractory or relapsing B cell malignancies . In a second study, the bisphosphonate, zoledronate, together with IL-2 was used to treat patients with metastatic prostate carcinoma . A higher proportion of the treated patients demonstrated partial remission or stable disease and those patients with the highest numbers of Vγ2Vδ2 T cells exhibited the lowest levels of prostate-specific antigen, a cancer cell marker . These findings suggest that Vγ2Vδ2 T cells may play a role in tumor immunosurveillance, as well as the possibility of employing these γδ T cells in cancer immunotherapy.
Venous blood samples were obtained from patients with nasopharyngeal carcinoma diagnosed at the National Cancer Center of Singapore. Healthy controls (NOR) were recruited from the Center for Transfusion Medicine, Health Sciences Authority of Singapore. Approval from the Institutional Review Board was obtained for the present protocol prior to the collection of blood samples from healthy individuals and NPC patients. A total of 96 NPC patients and 54 healthy donors were included in this study. The diagnosis of NPC was determined based on histopathological analysis. Clinical characteristics including gender, age, and tumor size and staging of cancer patients at the time of diagnosis were obtained from medical records (Table 1). Of the 96 NPC patients, 69 were male and 27 were female with a median age of 50 (range: 29–77); whereas the healthy donors comprised 45 males and 9 females with a median age of 37.5 (range: 22–59).
PBMC were isolated by Ficoll-Paque density centrifugation. The following monoclonal antibodies were used: FITC- or PE-labeled V δ2 (B6.1), FITC-labeled CD8 (RPA-T8), PE-labeled CD27 (M-T271), APC-labeled CD27 (0323), PE-Cy5-labeled CD28 (28.2), PE-Cy7-labeled CD28 (28.2), R-PE-labeled CD3 (HIT3a), and PE-Cy7-labeled CD3 (SK7). All Abs used in the present study were purchased from BD Biosciences (San Jose, CA, USA). Three- and four-color flow cytometry were performed and analyzed using a FACSCalibur or a FACSCantoII flow cytometer (BD Biosciences).
PBMC were incubated with FITC anti- V δ2 TCR mAb on ice for 30 min. Cells were washed once with chilled 2% FCS to remove excess unbound mAb. Anti-FITC MicroBeads were added and the cells were incubated on ice for another 30 min. The magnetic isolation of V δ2 T cells was performed as described in the manufacturer’s instructions (Miltenyi Biotech). These cells were subsequently passed through magnetic columns that separate non- V δ2 T cells from V δ2 T cells. The purity of the V δ2 T cells was determined by flow cytometry.
CTL assays were performed using DELFIA EuTDA Cytotoxicity Reagents (Perkin-Elmer) according to the manufacturer’s instructions. Briefly, 5×103 K562 cells were labeled with a fluorescence enhancing ligand, (bis(acetoxymethyl) 2,2′:6′,2″-terpyridine-6,6″-dicarboxylate (BATDA), in 1 ml of culture medium at 37°C for 30 min. The cells were washed 4 times with PBS and adjusted to 5 ×104 cells/ml. One hundred microliter aliquots containing 5×103 cells each were then plated in 96-well round bottom plates. Purified Vγ2Vδ2 T cells were added at the indicated E:T ratio and centrifuged at 100 x g for 2 min. The plates were incubated for 2 h at 37°C. After incubation, the plates were centrifuged at 100 × g for 2 min. Twenty microliters of supernatant was transferred to 96-well flat bottom plates, followed by the addition of 180 μl of the provided europium solution. The plates were then shaken for 15 min at RT. The fluorescence intensity was measured using a time-resolved fluorometer (VICTOR3). The percentage of specific lysis was calculated as: % specific lysis = 100 × (experimental release - spontaneous release)/(maximum release - spontaneous release). The background signal was subtracted from the maximum and spontaneous release values before the values were used to calculate the percentage of specific lysis.
PBMC were plated at 1×105 cells/well in 96-well round-bottom tissue culture plates in complete RPMI 1640 medium with or without HMBPP (0.316 μM) for 4 hr in the presence of Brefeldin A (GolgiPlug, BD Biosciences). To determine the full cytokine capability of CD8 T cells, PBMC were stimulated with PMA (10 ng/ml) and ionomycin (1 μM) under the same conditions. After stimulation, the cells were harvested and stained with FITC-labeled anti- V δ2 or FITC-labeled CD8, APC-labeled anti-CD27, and PE-Cy7-labeled anti-CD28. Cells were washed twice in staining buffer and subsequently fixed in 100 μl of 1X BD Cytofix/Cytoperm for 20 min on ice. After fixation, the cells were washed twice with 1X permeabilization buffer, followed by cytoplasmic staining with PE-conjugated anti-TNF-α (MAB11) or PE-conjugated anti-IFN-γ (4S.B3), both from BD Biosciences. The percentage of cytokine-producing V δ2 or CD8 T cells among the four Vγ2Vδ2 T cell subsets was determined by four-color cytometry. A total of 100,000 to 150,000 events were acquired for each sample and analyzed using CellQuest Pro (BD Biosciences) and FlowJo 8.5.2 (TreeStar).
Data obtained for different Vγ2Vδ2 T cell subsets were analyzed by one-way ANOVA using GraphPad Prism 4.0 software. Results are presented as scatter plots in which each dot represents either a healthy donor or a NPC patient. Comparisons between groups were performed using a non-parametric Mann-Whitney U test. A p value < 0.05 was considered significant.
To compare the cytolytic activity of Vγ2Vδ2 T cells from NPC patients and healthy controls (NOR), highly purified γδ T cells were incubated with K562 (a γδ T cell-sensitive cell line) or HK-1 (a human NPC cell line). Due to their low frequencies in peripheral blood, Vγ2Vδ2 T cells were isolated from 6 NPC patients and 5 NOR and pooled into their respective groups. The γδ T cells from NPC patients demonstrated lower cytolytic activity against both K562 and HK-1 compared with those from NOR (Figure 1a and 1b). The cytotoxicity assay represents a single experiment that was performed using pooled Vγ2Vδ2 T cells from NPC patients and NOR.
A comparison of the frequency of Vγ2Vδ2 T cells from 96 patients with active NPC disease and 54 NOR revealed no significant difference in the total percentage of Vγ2Vδ2 T cells between NPC patients and NOR (Figure 1c). The median frequency of Vγ2Vδ2 T cells detected in NPC patients and NOR was 1.8% (range: 0.03%–16.2%) and 2.0% (range: 0.3%–10.7%) respectively. Although the incidence of NPC is 2–3 times more common in males than in females, we could not detect any sex-related difference in the total number of Vγ2Vδ2 T cells between the two genders (Figure 1d).
Using CD27 and CD28, Vγ2Vδ2 T cells can be delineated into four distinct subsets: TCM (CD27+ CD28+), TCM 27- (CD27− CD28+), TEM (CD27+ CD28−), and TEM RA (CD27− CD28−) (Figure 2). Similar to CD8+ αβ T memory cells, we and others have previously suggested that TCM cells predominantly represent early memory Vγ2Vδ2 T cells, whereas TEM RA cells are predominantly late memory Vγ2Vδ2 T cells . In support of this, TEM RA cells have been shown to possess the highest level of perforin expression among the four γδ T cell subsets, suggesting that TEM RA cells have higher cytolytic activity towards tumor cells than the other T cell subsets. Circulating Vγ2Vδ2 T cells are very heterogeneous, and the ability to identify γδ cellular subsets based on CD27 and CD28 expression (Figure 2a) may offer an opportunity to “fine-tune” the effector immune functions of γδ T cells during immunotherapy for NPC.
We examined the proportions of the various Vγ2Vδ2 T cell subsets in NPC patients. The major Vγ2Vδ2 cell subset present in the peripheral blood of both NPC patients and NOR was TCM cells. The median frequency of TCM Vγ2Vδ2 cells in NOR was 61.8% (range: 12.1%–91.7%). The median frequency of TCM Vγ2Vδ2 cells in NPC patients was 42.7% (range: 0.7%–86.8%), a significantly lower frequency compared to NOR (Figure 2b). In contrast, the median frequency of late effector memory cells (TEM RA) was 25.2% (range: 0%–98.4%), which was significantly higher in comparison to NOR, because the median frequency of TEM RA in NOR was 10.7% (range: 0.3%–69.4%). The increased TEM RA cells and decreased TCM cells in NPC patients were not related to age (Figure 2c) because linear regression analyses showed no correlation between the frequencies of these two subsets and age in both NPC patients and NOR. Interestingly, the median frequency of TCM cells in NOR was marginally lower (p=0.0447) in males than in females (Figure S1). However, this was not observed in NPC patients. Overall, there was no observable difference in the total circulating Vγ2Vδ2 T cells between NPC patients and NOR; however, a specific and significant reduction in the TCM cells and an increase in the TEM RA cells could be detected in NPC patients.
Undifferentiated carcinoma (UNC, WHO type III) and non-keratinizing squamous cell carcinoma (SCC, WHO type II) are the two most common types of NPC in southeast Asia, including Singapore. Among the NPC specimens studied, 73 were UNC type and 23 were SCC type. When compared individually, there was no significant difference in the total percentage of Vγ2Vδ2 T cells among patients with UNC or SCC compared with normal controls (Figure S2a). However, when the various Vγ2Vδ2 T cell subsets were studied, TCM cells were significantly reduced in patients with UNC or SCC compared with normal controls (Figure S2b). The median frequency of TCM cells in patients with UNC was 43.5% (range: 0.7%–86.8%), whereas patients with SCC exhibited a median frequency of 41.3% (range: 10.5%–85.1%). The median frequency of TEM RA cells in UNC patients was 23.1% (range: 0%–98.4%), a significant increase compared to NOR (Figure 4b). A similar increase was noted in TEM RA cells in SCC patients (median frequency of 29.5% (range: 1.2%–71%)). This difference was slightly below statistical significance (p=0.0759) (Figure S2b).
Similar comparisons were performed after segregating the NPC specimens according to their clinical, nodal, and overall disease staging. Although a similar trend of decreased TCM cells and increased TEM RA cells was observed (Figures S3), these differences were not statistically significant due to the limited number of specimens available for each of the clinical sub-groups.
Vγ2Vδ2 T cell-mediated killing of tumor targets is mediated by the release of stored perforin upon contact with tumor cells. To determine if there were differences in perforin expression in Vγ2Vδ2 T cells between patients with NPC and normal controls, we performed intracellular perforin staining on blood Vγ2Vδ2 T cells. In Vγ2Vδ2 T cells from normals, perforin expression was higher in TEM RA cells than TCM cells (Figure 3a, left panel). The median proportions of TEM and TEM RA Vγ2Vδ2 T cells that expressed perforin were 77.2% (range: 31.1%–94.0%) and 89.7% (range: 40.0%–96.7%), respectively. These proportions were significantly higher than those of early memory TCM cells (median: 31.3%, range: 5.1%–83.4%) and TCM 27− cells (median: 37.5%, range: 11.9%–89.4%) (Figure 3a, right panel). Similar to healthy controls (NOR), a greater proportion of TEM and TEM RA Vγ2Vδ2 T cells expressed perforin compared with TCM and TCM 27− cells from NPC patients (n=10) (Figure 3b). However, a significantly greater proportion of TCM cells from NPC patients expressed perforin compared with TCM cells from healthy controls (p=0.029). No statistically significant differences were detected in the other γδ T cell subsets (Figure 3b).
To determine the functional capabilities of γδ T cells from NPC patients compared with healthy controls, Vγ2Vδ2 T cells were stimulated with HMBPP and IFN-γ and TNF-α production measured by intracellular staining (Figure 4a, left panel). The proportion of Vγ2Vδ2 T cells producing IFN-γ and TNF-α in NPC patients was significantly lower than in healthy controls (Figure 4a, right panel). For NPC patients, the median percentage of IFN-γ- and TNF-α-positive Vγ2Vδ2 T cells was 17.4% (range: 3.4%–46.4%) and 17.0% (range: 3.4%–45.2%), respectively (Figure 4a, right panel). In contrast, the median percentage of IFN-γ- and TNF-α-positive Vγ2Vδ2 T cells in NOR was 26.9% (range: 10.5%–46.4%) and 23.4% (range: 11.7%–56.2%), respectively (Figure 4a, right panel).
To further delineate whether a specific T cell subset within the Vγ2Vδ2 T cells contributed to the overall impairment of cytokine production, IFN-γ and TNF-α production by the various γδ T cell subsets was evaluated in NPC patients and NOR. In NPC patients, the proportion of IFN-γ-positive Vγ2Vδ2 T cells was significantly reduced in all four subsets, with the TCM and TEM RA subsets demonstrating the most severe reductions (Figure 4b). Whereas the median proportion of IFN-γ-positive TCM cells in NPC patients was 18.6% (range: 5.4%–57.1%), it was 32.7% (range: 17.3%–58.4 %) in NOR. The median proportion of IFN-γ-positive TEM RA cells in NPC patients was 9.6% (range: 5.4%–22.5%) compared with 19.8% (range: 4.8%–47.4%) in NOR (Figure 4b). When compared to NOR, the proportion of TNF-α-positive TCM and TEM RA cells in NPC patients was also significantly reduced (Figure 4b). The median percentage of TNF-α-positive TCM cells detected in NPC patients was 19.2% (range: 3.5%–49.4%) compared with 28.7% (range: 13.2%–66.6%) in NOR. Similarly, the median percentage of TNF-α-positive TEM RA cells detected in NPC patients was 7.18% (range: 0%–67.1%), in contrast to 18.8% (range: 0%–40.6%) in NOR (Figure 4b).
Linear regression analyses showed no correlation between the number of cytokine-producing γδ T cells and the age of the NPC patients or healthy controls (Figure 4c). This shows that the reduction in cytokine-producing γδ T cells in NPC patients was not due to differences in the ages of the NPC patients and the healthy controls.
To determine the maximum cytokine-producing capability of γδ T cells and CD8 T cells in response to mitogenic stimuli, PBMC from NPC patients and NOR were stimulated with PMA and ionomycin. Similar proportions of IFN-γ and TNF-α-positive CD8 T cells were detected in NPC patients and NOR (Figure 4d). There was also no significant difference in the proportion of cytokine-positive Vγ2Vδ2 T cells between NPC patients and NOR. These results suggest that the reduced cytokine production of γδ T cells from NPC patients in response to the HMBPP antigen is caused by a defect in TCR signaling.
NPC patients with low-grade tumors respond well to radiotherapy. However, the response rate for NPC patients with tumor grades of 3–4 is poor and concurrent chemo-radiotherapy is usually introduced to improve the clinical outcome for these patients. To determine whether these treatments could normalize the γδ T cell subset distributions and γδ T cell function in NPC patients to those observed in healthy controls, PBMC were obtained from NPC patients 3–6 months after receiving their last treatment and the various Vγ2Vδ2 T cell subsets assessed. Instead of normalizing, the percentage of Vγ2Vδ2 TCM cells dropped even further to very low levels with a corresponding rise in the percentage of TEM RA cells (Figure 5a).
When the total proportion of IFN-γ- and TNF-α-positive Vγ2Vδ2 T cells was analyzed in healthy controls, NPC patients prior to therapy, and NPC patients post-therapy, the proportion of IFN-γ- and TNF-α-positive Vγ2Vδ2 T cells continued lower post-therapy than healthy controls with no evidence of improvement when compared with patients prior to therapy (Figure 5b). Further analysis of the various γδ T cell subsets showed slight restoration of IFN-γ-positive TCM cells in NPC patients following treatment (p=0.0478). However, the impaired IFN-γ production observed for TEM RA cells and the production of TNF-α by both TCM and TEM RA cells were not corrected following treatment (Figure 5c).
Vγ2Vδ2 T cells have recently been exploited to provide protective immunity against intracellular pathogens and cancer . In the present study, we show that Vγ2Vδ2 T cells from NPC patients are functionally impaired with respect to both their ability to kill tumor cells and their capacity to produce IFN-γ and TNF-α. These functional defects are associated with alterations in the distribution of the various Vγ2Vδ2 T cell subsets. Immunophenotypic analyses of PBMC from NPC patients and healthy controls demonstrated that, although there was no difference in the total percentage of Vγ2Vδ2 T cells, the distribution of Vγ2Vδ2 T cell subsets was altered in NPC patients compared with healthy controls.
Differences in the expression of the costimulatory receptors CD28 and CD27, define 4 subsets of Vγ2Vδ2 T cells in NPC patients and healthy controls. Both naïve and central memory CD8 αβ T cells express CD28 and CD27 . Naïve T cells can be further discriminated from central memory cells because naïve T cells express both the CD45RA isoform and the lymph node homing receptor, CCR7, whereas central memory cells retain CCR7 but lack CD45RA  . In most healthy controls and some NPC patients, CD28+ CD27+ Vγ2Vδ2 T cells represent the majority of γδ T cells. However, unlike CD8 αβ T cells, very few naïve Vγ2Vδ2 T cells are present in the peripheral blood of adult individuals. De Rosa et al. showed that naïve Vγ2Vδ2 T cells were CD45RO−, CD11alow, and CD27high and the percentage of CD27high cells falls below 10% after 1 year of age . Because it is technically difficult to consistently discriminate CD27high from CD27low based on CD28 and CD27 staining, we are unable to determine the exact proportion of naïve Vγ2Vδ2 T cells. We determined CD28+ CD27+ Vγ2Vδ2 T cells to represent predominantly memory cells. This is in agreement with an earlier study showing that the predominant population of Vγ2Vδ2 T cells in the peripheral blood was CD45RO+ CD45RA− CD27+ . Similar to CD8 αβ central memory T cells, CD45RA− CD27+ V δ2 T cells also express CD62L and CCR7 . Our results showed that in NPC patients the CD28+ CD27+ (TCM) Vγ2Vδ2 T cell subset was markedly reduced whereas the CD28− CD27− (TEM RA) Vγ2Vδ2 T cell subset was significantly increased.
Similar CD8 αβ T cells differentiation has been shown to correlate with disease progression in patients with viral infections . For instance, HIV-infected long-term non-progressors are characterized by an increase of highly proliferative CD27+ CD28+ HIV-specific memory CD8 αβ TCM cells . In contrast, other persistent viral infections deplete CD8 αβ TCM cells by driving their differentiation to the CD27− CD28− terminally differentiated phenotype . The observed reductions in TCM cells and increases in TEM RA Vγ2Vδ2 T cells in NPC patients is likely due to a similar differentiation of TCM cells. However, it is unclear whether this differentiation is due to a concurrent viral infection or is driven by the presence of the NPC only.
EBV infection is one possible viral cause of terminal differentiation of Vγ2Vδ2 T cells in NPC patients. The development of NPC in EBV-endemic regions is highly associated with latent EBV infections. Primary EBV infection in childhood usually manifests as an asymptomatic infection. In developed countries, however, not all individuals are infected in childhood. Primary EBV infection in adolescents and adults can then result in acute infectious mononucleosis. Acute infectious mononucleosis is associated with γδ T cell expansions [7, 20]. Moreover, Vγ1 T cells proliferate in response to stimulation by EBV-transformed B lymphocytes [18, 34]. However, when we compared the percentage of Vγ1+ cells in NPC patients and healthy controls, we did not observe a significant expansion in the Vγ1+ population in NPC patients (p=0.2291). It is therefore unlikely that the decline in TCM Vγ2Vδ2 T cells in NPC patients is caused by changes in Vγ1 T cells.
High frequencies of CD8+ αβ TEM RA cells (CD28−  or CD45RA+ CD27− ) are found in individuals who have been infected by CMV , likely due to continued stimulation of CD8 αβ T cells by chronic CMV infection. Increasing evidence suggests that this type of chronic immune activation can alter the overall composition of T cells, culminating in dysfunctional T cell responses. Our data demonstrate an increase in the TEM RA subset of Vγ2Vδ2 T cells in NPC patients that parallels the increase in late-differentiated CMV-specific CD8 αβ T cells observed in patients with CMV infection. Thus, CMV infection is another possible cause for the terminal differentiation of Vγ2Vδ2 T cells in NPC patients. Consistent with this hypothesis, CMV infection is endemic in Singapore. For example, when 120 antenatal Singaporean women were screened for CMV infection, 87% tested seropositive . Moreover, γδ T cells expand in kidney transplant recipients developing CMV infection. However, against this hypothesis, these expansions are restricted to Vγ1 and Vγ3 T cells, not Vγ2Vδ2 T cells . In our study, NPC patients were not tested for CMV seropositivity. Therefore, we were unable to determine if CMV seropositivity correlated with the observed changes in the V δ2 T cell subset distribution.
The increased levels of TEM RA cells in NPC patients could be due to increased apoptosis of TCM V δ2 T cells. However, this explanation seems less likely because TCM cells express higher levels of the anti-apoptotic protein bcl–2 compared with TEM and TEM RA cells . This argues against the suggestion that the increased sensitivity of TCM cells to apoptosis could account for the loss of TCM cells in NPC patients. Furthermore, gag-specific CD27+, but not CD27−, CD8 T cells persist in HIV-infected patients following adoptive transfer suggesting that the expression of CD27 on memory T cells may confer a survival advantage in vivo . Therefore, the increased number of TEM RA cells in NPC patients may result from the differentiation of TCM to TEM RA cells with the loss of CD27 and CD28 expression due to viral or tumor effects. Alternatively, the increased levels of TEM RA cells could simply be a consequence of the preferential expansion of TEM RA cells in NPC patients although this has never been described.
Previously, an age-related decrease in γδ T cells  and an increase in the CD27− CCR7− effector/memory Vγ2Vδ2 cells was reported for healthy donors . Furthermore, the increased number of CD27− V δ2 T cells negatively correlated with the ability to expand in response to IPP and IL-2 . In the present study, there was no correlation between the levels of TCM or TEM RA V δ2 T cells and the age of the NPC patients or healthy controls. Similarly, there was no correlation between the proportion of cytokine-positive Vγ2Vδ2 T cells and the age of NPC patients or healthy controls. Thus, the reduced number of cytokine-producing TCM and TEM RA in NPC patients compared with healthy controls is not due to differences in their ages.
Vγ2Vδ2 T cells can be activated to produce large amounts of TNF-α and IFN-γ after stimulation by phosphoantigens [26, 27, 45]. Here we demonstrated that fewer Vγ2Vδ2 T cells from NPC patients produced IFN-γ and TNF-α upon stimulation with the HMBPP antigen compared with Vγ2Vδ2 T cells obtained from healthy controls. However, mitogenic stimulation using PMA and ionomycin did not show the same decline in the numbers of IFN-γ producing cells suggesting that the decrease in cytokine production by Vγ2Vδ2 T cells is due to alterations in T cell receptor signaling following antigen presentation. The impaired production of cytokines by TCM and TEM RA γδ T cells from NPC patients may reflect general Vγ2Vδ2 T cell dysfunction. There also may be an increased susceptibility to tumorogenesis since IFN-γ deficiency in mice makes them more susceptible to the development of lymphomas  and epithelial malignancies .
IFN-γ secretion does not always correlate with cytotoxicity . Using a single-cell assay that simultaneously measures cytotoxicity and IFN-γ secretion, Snyder et al. found that these two effector functions could be regulated independently . In the present study, however, the reduced frequency of IFN-γ- and TNF-α-producing Vγ2Vδ2 T cells in NPC patients was also associated with lower cytotoxicity against tumor cells. Thus, the impairments that lead to the decreased production of IFN-γ and TNF-α by Vγ2Vδ2 T cells from NPC patients may also affect Vγ2Vδ2 T cell cytotoxicity. These functional defects could potentially limit the ability of V δ2 T cells to mount effective responses against tumor cells. However, it remains unclear whether the loss of function and altered subset distribution of Vγ2Vδ2 T cells in NPC patients reflect a general alteration in all circulating T cells including αβ T cells. Further studies are needed to address this possibility.
NPC are relatively radiosensitive and, therefore, radiation is a standard modality of treatment for NPC. For NPC patients with advanced disease, combined radiotherapy and chemotherapy is often administered to improve clinical outcome and survival rates. Although these are useful modalities for NPC treatment, neither treatment restored Vγ2Vδ2 T cell function or returned Vγ2Vδ2 T cell subset distributions to normal.
Recent advances in immunotherapy specifically targeting γδ T cells have shown promise for the treatment of multiple myeloma , renal cell carcinoma , and prostate cell carcinoma . Large-scale ex vivo expansion of Vγ2Vδ2 T cells to provide an adequate source of cells for adoptive immunotherapy is now feasible . The present study shows that alterations in the proportions of the Vγ2Vδ2 T cell subsets and their functions could have profound effects on overall Vγ2Vδ2 T cell responses. Thus, monitoring Vγ2Vδ2 T cell subset distributions and functions may allow predictions about the effectiveness of Vγ2Vδ2 T cell immunotherapy. Such monitoring offers a rational basis for assessing the potential effectiveness of immunotherapy strategies using Vγ2Vδ2 T cells to optimize for the best clinical outcome.
Figure S1. No differences in the percentages of Vγ2Vδ2 T cell subsets between male and female NPC patients or healthy controls. NPC patients and healthy controls (NOR) were segregated into males (M) and females (F) and differences in the percentages of various V δ2 T cell subsets between sexes were determined. Significance was assessed using the non-parametric Mann-Whitney U test. In NOR, the median percentage of TCM V δ2 T cells was marginally reduced in males compared with females, but there was no statistical difference between both sexes among NPC patients.
Figure S2. Decreased frequency of TCM and increased frequency of TEM RA Vγ2Vδ2 T cells are observed in undifferentiated nasopharyngeal carcinoma and non-keratinizing squamous cell nasopharyngeal carcinoma patients. (a) The total percentage of V δ2 T cells and (b) the percentage of V δ2 T cell subsets were compared for 73 undifferentiated carcinoma (UNC), 23 non-keratinizing squamous cell carcinoma (SCC) NPC patients, and 54 healthy donors (NOR). Note that there was no statistically significant difference in the total percentage of V δ2 T cells between UNC and SCC or between either NPC type and NOR. However, TCM V δ2 T cells from UNC and SCC patients were significantly reduced, whereas TEM RA V δ2 T cells were increased in both histological types of nasopharyngeal carcinoma.
Figure S3. No correlation between the frequencies of Vγ2Vδ2 T cells or TCM and TEM RA subsets and disease prognosis. (a) The total percentage of V δ2 T cells and (b) TCM and TEM RA V δ2 T cells obtained from NOR and NPC patients with different tumor, nodal, and disease staging were analyzed by one-way ANOVA. The percentage of TCM V δ2 T cells in NPC patients diagnosed with stage II tumors was lower than the percentage of TCM V δ2 T cells in NOR. Similarly, the percentage of TCM V δ2 T cells in NPC patients diagnosed with stage 0 nodal was lower than the percentage of TCM V δ2 T cells detected in NOR.
We thank Florencia Goh, Jolin Ning Ning, and Lan Yin Wang of the Division of Clinical Trials and Epidemiological Sciences, National Cancer Center, for their efforts in collecting the blood samples and the patients’ clinical data. This work was supported by the Biomedical Research Council of Singapore, National Medical Research Council of Singapore, Singapore Millennium Foundation Fund to K.M.H.; a Singapore Millennium Foundation Scholarship to K.J.P.; and the NIH National Institute of Arthritis and Musculoskeletal and Skin Disease (AR45504), the National Institute of Allergy and Infectious Diseases (Midwest Regional Center of Excellence for Biodefense and Emerging Infectious Diseases Research, AI057160), and the National Cancer Institute (CA113874) to C.T.M.
Conflict of interest statement: None of the authors has a financial or other relationship that might lead to a conflict of interest.
Electronic supplementary online material: This article contains supplementary figure legends and figures that are available online only.
The original publication is available at springerlink.com. The link is: http://www.springerlink.com/content/w6842226665518w0/?p=d81a013938b74286af6c985c2e4fa537&pi=0.