Natural killer T (NKT) cells are CD1d-restricted lymphoid cells characterized by an invariant T-cell receptor, which consists of a Vα24 chain paired with a Vβ11 chain and by coexpression of NK cell receptors.1 NKT cells can rapidly produce large amounts of IFN-γ and IL-4 upon activation with α-galactosylceramide (α-GalCer).2, 3 IFN-γ is mainly produced by CD4−CD8− and CD8+ NKT cells while IL-4 and IL-13 are predominantly produced by CD4+ NKT cells.4 NKT cells have been demonstrated to express granzyme B and are able to kill tumor cells in vitro.5, 6, 7 Moreover, they can activate NK cells, CD4+ and CD8+ T cells and dendritic cells via cell-cell contact and the production of cytokines.8, 9 It is therefore not surprising that many studies have shown an immunoregulatory role of NKT cells in several immune responses, including antitumor immune responses.10, 11
Treatment of mice with the NKT cell ligand α-GalCer has been shown to have an antitumor effect in tumor metastasis models of the liver, lung and lymph nodes.12, 13, 14, 15 This effect is most likely caused by subsequent activation of NK cells, CTL and antigen presenting cells.16 In 2 clinical phase 1 trials in advanced cancer patients injected with α-GalCer17 or α-GalCer-loaded immature dendritic cells,18 a distinct activation of the immune system was observed, but only in cancer patients with detectable Vα24+Vβ11+ NKT cell numbers.17 Although no long-term tumor regression was observed, stable disease was recorded in several patients without any toxicity17, 18 and some patients even showed a (transient) reduction of serum tumor markers or necrosis of the tumor.18
Thus far, data on the putative decrease of NKT cell numbers in cancer patients have been obtained from relatively small patient cohorts,17, 19, 20 whereas actually normal Vα24+Vβ11+ NKT cell numbers in cancer patients were reported in another study.21 Importantly, the potential influence on circulating NKT cell numbers of both age, as described by DelaRose et al.,22 and gender, as described by Sandberg et al.,23 was not taken into account.
Animal models suggested a critical role of NKT cell-derived IFN-γ in tumor immunosurveillance and antitumor immunity.24, 25, 26 However, conflicting data exist on IFN-γ production by NKT cells in cancer patients. Clinical studies in lung cancer patients, using quantitative single-cell RT-PCR,19 showed no reduction in IFN-γ-producing Vα24+Vβ11+ NKT cells in cancer patients. On the other hand, Tahir et al.20 found a significantly decreased IFN-γ secretion by in vitro cultured NK T-cell lines of prostate cancer patients.
Here, we first investigated the effect of age and gender on circulating Vα24+Vβ11+ NKT cell numbers (further referred to as NKT cells) in a group of 69 healthy controls. Subsequently, we determined circulating NKT cell numbers in a large cohort of 120 cancer patients and evaluated the effect of tumor type and tumor load on these numbers using age- and gender-matched healthy controls. In addition, the effect of tumor reduction by surgery or radiotherapy was studied in groups of breast and head and neck cancer patients, respectively. With CD4 and CD8 expression as putative markers of the cytokine profile of NKT cells,4 we also analyzed the expression of these molecules on circulating NKT cells in a cohort of breast cancer patients. Finally, we investigated the frequencies of IFN-γ-secreting NKT cells in both cancer patients and healthy controls using an ex vivo ELISPOT assay.
Material and methods
To investigate the number of circulating Vα24+Vβ11+ NKT cells, heparinized blood samples were taken from 69 healthy subjects and 120 advanced cancer patients. Study group characteristics are depicted in Table I. Patients were designated to the “no evidence of disease” (NED) group if radiologic and physical examination did not reveal any signs of cancer. In all other cases, patients were designated to the “evidence of disease” group (ED). None of the cancer patients received chemotherapy or radiotherapy at the time of analysis.
Table I. Study Group Characteristics
Mean age in years (range)
(N)ED, (No) Evidence of Disease.
All cancer patients tested
Head and Neck cancer
Renal cell cancer
The effect of tumor resection on circulating NKT cell numbers was investigated in 18 breast cancer patients, sampled prior to (day −1) and after (day 6) tumor resection. To evaluate the effect of surgery on NKT cell numbers, we analyzed noncancer patients sampled prior to (day −1) and after (day 4) orthopedic surgery. The effect of radiotherapy on circulating NKT cell numbers was studied in 14 head and neck carcinoma patients sampled prior to (week 0) and after curative radiotherapy (week 18).
CD4 and CD8 expression on NKT cells was determined in 9 randomly chosen breast cancer patients prior to surgery (mean age, 52 years; range, 43–64 years) and in 9 age-matched healthy controls (mean age, 53 years; range, 45–57 years). ELISPOT assay was performed on 11 additional cancer patients [6 colon carcinoma patients (ED/NED = 5/1) and 5 breast cancer patients (ED/NED = 4/1; mean age, 61 years; range, 37–64 years] and 8 age-matched healthy controls (mean age, 50 years; range, 35–63 years). The medical ethics committee of the VU Medical Center, Amsterdam, The Netherlands, approved the study. All patients gave informed consent.
Antibodies and reagents
The following monoclonal antibodies and reagents were used: FITC-labeled antihuman Vα24 (clone C15), PE- or biotin-labeled antihuman Vβ11 (clone C21; Immunotech, Marseille, France), PE-labeled antihuman CD8β (clone 2ST8.5H7), FITC- or APC-labeled IgG1 (clone X40), PE-labeled IgG2a (clone X39), APC-labeled IgG1 (clone X40), FITC-, RPE-Cy5-, or APC-labeled antihuman CD3 (clone SK7), PE-labeled antihuman CD16 (clone B7.3.1), PE-labeled antihuman CD56 (clone MY31) and APC-labeled antihuman CD4 (clone SK3; Becton-Dickinson Biosciences, San Jose, CA), RPE-Cy5-labeled IgG1 (clone DAK-GO1), RPE-Cy5-labeled streptavidin (Dako, Glostrup, Denmark), Ficoll-Isopaque density gradient (Axis-Shield, Oslo, Norway), lysing solution (Becton-Dickinson Biosciences), FlowCount fluorospheres (Beckman-Coulter, Miami, FL), IFN-γ ELISPOT assay (Mabtech, Nacka, Sweden), PHA (Murex Biotech, Dartford, U.K.) and α-galactosylceramide (Kirin Brewery, Tokyo, Japan).
Lymphocyte numbers were determined by adding fixed volumes of FlowCount fluorospheres to the leukocytes after erythrocyte lysis (Becton-Dickinson Biosciences lysing solution) just before flow cytometric evaluation. Lymphocytes were characterized by monoclonal antibodies or relevant isotype controls: NKT cells were defined as Vα24+Vβ11+CD3+, NK cells as CD16/56+CD3− and T cells as CD3+. NKT cells were defined by coexpression of CD3, Vα24 and Vβ11 since this combination has been shown to be highly specific for α-GalCer-reactive CD1d-restricted NKT cells.27 For analysis of CD4 and CD8 expression, peripheral blood mononuclear cells (PBMCs) were prepared from heparinized blood using Ficoll-Isopaque density gradient centrifugation. Cells were subsequently stained with antibodies against human Vα24, Vβ11 and CD3, human Vα24, Vβ11, CD8β and CD4, or with the relevant isotype controls. A minimum of 50,000 T cells was analyzed per sample. Flow cytometric analysis was performed on a FACS Calibur using CELL Quest software (Becton-Dickinson Biosciences).
Quantification of freshly isolated IFN-γ-secreting NKT cells by single-cell enzyme-linked immunospot (ELISPOT)
The relative numbers of α-GalCer-induced IFN-γ-secreting NKT cells were determined using an ELISPOT. PBMCs were prepared from heparinized blood using Ficoll-Isopaque density gradient centrifugation. Serially diluted PBMCs, starting at 5 × 105 cells per well (with a minimum of 50 NKT cells per well), were placed into wells (Millipore, Bedford, MA) coated with anti-IFN-γ antibody. The PBMCs were subsequently incubated with α-GalCer (100 ng/ml) or vehicle for 16 hr at 37°C. Incubation of the cells with PHA (1 μg/ml) was used as a positive control for the assay. After incubation, the wells were rinsed several times with PBS and developed according to the manufacturer's instruction. Spot-forming cells were quantified using an ELISPOT reader. The total number of α-GalCer-reactive IFN-γsecreting NKT cells was calculated by subtracting the number of spots detected in the vehicle condition (background). The percentage of IFN-γ-secreting NKT cells was calculated from the number of spots (minus background) and the individual FACS data (% Vα24+Vβ11+ NKT cells from total PBMCs). For some experiments, PBMCs were first depleted from Vα24+ cells using magnetic cell sorting (MACS; Miltenyi Biotec, Bergisch-Gladbach, Germany).
Linear regression was performed to analyze NKT cell numbers/million T cells in healthy controls and to analyze the influence of age on the percentage of IFN-γ-secreting NKT cells in healthy controls (ELISPOT). The influence of age, gender, cancer, tumor type and tumor load on circulating NKT cell numbers was evaluated using multiple regression analysis. A paired-samples Student's t-test was used to investigate the influence of surgery and radiotherapy on NKT cell numbers. ELISPOT data were statistically analyzed using an independent samples t-test. A p-value of < 0.05 was considered significant. All data were analyzed using the SPSS 9.0 software.
Circulating Vα24+Vβ11+ NKT cell numbers are dependent on age and gender
First we addressed the question whether age affected peripheral blood NKT cell levels in healthy controls (Fig. 1, Table II). Despite the broad range of normal NKT cell levels, linear regression analysis revealed a clear age-dependent decrease of these cells both in absolute levels (per ml blood) and within the T-cell compartment (per million T cells; p = 0.001 in both cases). No age-related decline was observed for Vβ11 single positive T cells or Vα24 single positive T cells (graph not shown). Also, the total number of T cells and circulating lymphocytes were not affected by age. NK cell numbers per ml blood were found to increase slightly with age (1.4% increase/increasing year of age; p < 0.001). No significant difference in the total number of T cells could be observed between cancer patients and healthy individuals. An age-related decline in NKT cells was also observed in cancer patients (Fig. 2).
Table II. Effects of Age, Gender and Disease on NKT Cell Numbers in Healthy Donors and Cancer Patients
Mean variation of immune parameter
Tumor types investigated: head and neck cancer, colon carinoma, breast cancer, renal cell cancer, melanoma.
Effects of age in healthy controls (linear regression)
NKT cells/ml blood
3.4 % reduction/increasing year of age
NKT cells/106 T cells
3.4 % reduction/increasing year of age
NK cells/ml blood
1.4 % increase/increasing year of age
T cells/ml blood
No age effect
Vα24− Vβ11+ T cells/106 T cells
No age effect
No age effect
Effects of age, gender and disease (multiple regression), NKT Cells/106 T cells
Males vs. females
Males: 4.1 % reduction/increasing year of age
Females: 3.3 % reduction/increasing year of age
Healthy vs. cancer
47% reduction in addition to age and gender effects
ED vs. NED
In addition to the observed age effects, circulating NKT cell numbers (per million T cells) were lower in males than in females and decreased faster with age in males than in females (p = 0.008; Fig. 2, Table II). To exclude effects of overall T-cell fluctuations, NKT cell numbers were routinely assessed per million T cells in all further analyses.
Circulating Vα24+Vβ11+ NKT cell numbers are reduced in cancer patients
When the effects of age and gender were not taken into account, cancer patients showed a 69% reduction in NKT cell numbers compared to healthy controls (p < 0.001; data not shown). After correction for age and gender, a 47% reduction in NKT cells was observed (p = 0.013). This reduction was of the same magnitude in males and females of all ages (Fig. 2, Table II). Figure 3 shows a representative flow cytometric dot plot illustrating the expression of Vα24 and Vβ11 on CD3+ cells in a male cancer patient and healthy control of similar age.
Tumor type and tumor load have no effect on circulating Vα24+Vβ11+ NKT cell numbers
Cancer patients who were included in the study had different types of tumors, i.e., colon carcinoma, head and neck carcinoma, renal cell carcinoma, breast cancer and melanoma, and had different tumor loads (ED vs. NED; Table I). Interestingly, neither the tumor type nor the absence or presence of detectable tumor load had a significant effect on circulating NKT cell numbers (Table II).
A potential effect of tumor load on NKT cell numbers was additionally evaluated after resection of the tumor. The number of circulating NKT cells did not change after tumor removal by means of surgery in breast cancer patients (Fig. 4a) or by means of radiotherapy in head and neck cancer patients (Fig. 4c). Surgery alone did not affect NKT cell numbers in noncancer orthopedic patients (Fig. 4b). Taken together, the low number of NKT cells in cancer patients is not restored to normal (healthy) levels after macroscopic removal of the tumor within the time period tested.
CD4 and CD8 expression on circulating Vα24+Vβ11+ NKT cells is similar for cancer patients and healthy controls
It was shown by Lee et al.4 that CD4+ NKT cells produce relatively high levels of IL-4 after α-GalCer activation in contrast to CD4−CD8− (double negative, or DN) and CD8+ NKT cells, which produce relatively more IFN-γ.
We investigated whether cancer patients have relatively less proinflammatory DN and CD8+ NKT cells in comparison to CD4+ NKT cells. The percentage of CD4+ NKT cells was larger than the percentage of DN NKT cells in both healthy individuals and cancer patients. Hardly any CD8+ NKT cells could be detected in the circulation of healthy controls and cancer patients (Fig. 5). Importantly, the ratio of CD4+vs. DN NKT cells did not differ significantly between cancer patients and age-matched healthy controls.
IFN-γ-secreting Vα24+Vβ11+ NKT cell numbers are reduced in cancer patients
Stimulation of PBMCs with α-GalCer resulted in the formation of spots in an IFN-γ ELISPOT assay. Since about 10% of Vα24+Vβ11+ NKT cells secrete IFN-γ upon α-GalCer stimulation and a maximum of 500,000 PBMCs could be stimulated per well (monolayer), we performed the ELISPOT only when the PBMCs contained at least 0.01% NKT cells (= 50 NKT cells per well). Specific α-GalCer-induced spot formation and a lack of spots when Vα24+ cells were removed or when NKT cell levels were below detection level (flow cytometry) confirmed the specificity of the assay for IFN-γ-secreting Vα24+ (NK)T cells.
Interestingly, an age-dependent decrease (1.25% decrease/increasing year of age) in the percentage of IFN-γ-secreting NKT cells was observed in healthy controls (p = 0.032). No significant difference could be observed in the percentage of IFN-γ-secreting NKT cells between cancer patients (8.84% ± 1.65%, mean ± SEM) and age-matched healthy controls (8.75% ± 1.62%; Fig. 6). However, as a consequence of the reduced number of NKT cells in cancer patients, the absolute number of IFN-γ-secreting NKT cells in the circulation is reduced in cancer patients.
Numbers of NKT cells vary widely between healthy individuals: Vα24+Vβ11+ NKT cells constitute 0.01–2% of the T-cell population in the peripheral blood. Earlier studies suggested that NKT cell numbers may be influenced by age22 or gender.23 However, results from those studies were unclear since they also included a large population of non-NKT Vα24+Vβ11− T cells. Nevertheless, recently we also observed a marginal, yet nonsignificant age-related reduction in circulating Vα24+Vβ11+ NKT cells in healthy controls28 and with an extended group of healthy elderly controls we now show that circulating NKT cell numbers significantly decrease with advancing age and that this decrease is more prominent in males than in females. Therefore, when evaluating Vα24+Vβ11+ NKT cell numbers in future studies, study populations certainly need to be matched for age as well as gender.
Using these criteria, circulating NKT cell numbers were found to be significantly decreased (47%) in cancer patients compared to age- and gender-matched healthy controls. Earlier, in a small-scale study, Yanagisawa et al.21 had not observed a reduced number of circulating Vα24+Vβ11+ NKT cells in cancer patients (n = 21) compared to age- and gender-matched healthy controls (n = 22). Nevertheless, our present data are in line with results obtained in our phase 1 study on α-GalCer17 and with results from small-scale studies of Motohashi et al.19 and Tahir et al.,20 who reported that Vα24+Vβ11+ NKT cell levels are reduced in patients with primary lung cancer and advanced prostate cancer, respectively.
Reduced NKT cell numbers in cancer patients were independent of tumor type or tumor load (ED or NED). Even after effective removal of the tumor by surgery or radiotherapy, NKT cell numbers were not restored to a normal (healthy) level. Though it cannot be excluded that eventually NKT cell numbers would restore after tumor removal, the present data suggest that a low number of circulating NKT cells may precede the development of cancer and thus represent a risk factor for the development of malignancies rather than a result of the tumor presence.
A reduced number of peripheral blood NKT cells in cancer patients may result from NKT cell death, impaired NKT cell proliferation, or an accumulation of NKT cells in the tumor tissue. Whether NKT cells do accumulate at tumor sites is still not very clear: some investigators found only low numbers of CD56+CD3+ NKT cells29 or Vα24+Vβ11+ NKT cells30 in tumor-bearing livers. In contrast, Motohashi et al.19 reported on increased Vα24+Vβ11+ NKT cell levels in lung tumors. Differences in NKT cell recruitment in these studies might result from variations in chemokine production by tumors of different origin.31
Not only NKT cell numbers might be of importance in determining the outcome of cancer or therapeutic strategies, but also their cytokine secretion patterns. Some investigators showed that NKT cells in cancer patients are functionally impaired and produce less IFN-γ20 while others demonstrated that NKT cells are functionally normal32 and have a preserved production of IFN-γ.19 Since NKT cells producing type 1 (IFN-γ) or type 2 (IL-4, IL-13) cytokines were reported to be distinguishable by their expression of CD44, 33 and CD8,4 we investigated whether cancer patients have relatively less proinflammatory DN and CD8+ NKT cells in comparison to CD4+ NKT cells. No significant difference in the percentage of CD4+ and DN or CD8+ NKT cells was observed between breast cancer patients and healthy controls. These results were confirmed by our IFN-γ-ELISPOT data, which show that the percentage of IFN-γ-secreting cells within the NKT cell pool (about 9%) is similar for cancer patients and healthy controls. These data confirm the results of Motohashi et al.,19 who showed, using a single-cell RT-PCR method, that the percentage of IFN-γ-producing NKT cells (about 57%) is similar for primary lung cancer patients and healthy controls. Our relatively low percentage of IFN-γ-secreting NKT cells (about 9%) probably relates to the fact that we measured the actual secretion of IFN-γ protein while Motohashi et al.19 based their findings on mRNA evaluation. In any case, even though the percentage of IFN-γ-secreting NKT cells is normal, we conclude that, due to reduced absolute numbers of NKT cells, cancer patients have decreased numbers of IFN-γ-secreting NKT cells.
Since malignancies are apparently associated with reduced numbers of circulating IFN-γ-producing NKT cells and these cells are considered to be important for immunosurveillance, cancer patients might benefit from immunotherapies aiming at increasing these NKT cells. In a clinical phase 1 trial, circulating Vα24+Vβ11+ NKT cell numbers could be increased by injecting advanced cancer patients with α-GalCer-pulsed immature dendritic cells. This increase, however, was only transient and NKT cell levels returned to baseline again after 7–14 days.18 This return to homeostatic levels has also been shown for murine NKT cells after activation-induced proliferation.34, 35 Another therapeutic option might be the in vitro expansion of NKT cells and subsequent adoptive transfer of these cells. However, the impaired proliferative response of NKT cells derived from cancer patients6, 20, 21 should then be overcome first by addition of G-CSF21 or by repeated stimulation with α-GalCer-pulsed dendritic cells.6
In conclusion, we have shown that circulating IFN-γ-secreting Vα24+Vβ11+ NKT cell numbers are significantly decreased in cancer patients as compared to age- and gender-matched healthy controls. This reduction is independent of tumor type and is not restored after removal of the tumor by surgery or radiotherapy. A reduced number of circulating IFN-γ-secreting Vα24+Vβ11+ NKT cells might therefore rather represent a risk factor for the development of malignancies than a result of the tumor presence. These results suggest that future immunotherapeutic strategies should aim at increasing these IFN-γ-secreting NKT cells (e.g., by adoptive transfer or in vivo expansion) in cancer patients.
The authors are grateful to Martine Reijm, Taco Waaijman, Ingeborg Ouwehand and Corina Pronk for their technical assistance. They also thank Kirin Brewery for providing the α-GalCer. Supported by grant VU2002-2607 (to B.M.E.v.B., R.J.S. and A.J.M.v.d.E.) from the Dutch Cancer Society and by grant 920-03-142 (to H.J.J.v.d.V.) from the Netherlands Organization for Scientific Research.