**5. The development of a new index: non-incubating natural killer (NINK) score**

The findings concerning NK cells described above led our research group to surmise that calculation of a prediction score based on several factors that play a role in NK cells might be utilized as an effective index to determine the performance of NK cell activity of PBMCs in an individual without the need to prepare target cells. Therefore, we statistically analyzed factors that may correlate with the NK cell activity of human PBMCs using multiple regression analysis with linear regression model, and in doing so attempted to arrive at a formula to calculate the prediction score of NK cell activity (**Figure 4**) (manuscript of an original article under preparation). In that analysis, the following parameters were used as independent variables for NK cell activity: the percentage of CD3− CD56<sup>+</sup> NK cells (NK%) in PBMCs and mRNA levels of NKp46, granzyme B, FasL, TNF-α and IFN-γ relative to GAPDH mRNA levels in isolated NK cells, which were measured by flow cytometry and real-time PCR, respectively.

#### **Figure 4.** Scheme of strategy for construction of a new index to determine natural cytotoxicity without the use of target cells. The upper box shows six candidate parameters which we selected to explore the new index in an effort to determine natural cytotoxicity. The percentage of NK cells in PBMCs represents the amount of NK cells as a quantitative parameter, while mRNA expression levels of NKp46, granzyme B, FasL, TNF-α and IFN-γ are thought to function or interfere with the strength of lytic activity of NK cells as qualitative parameters. We analyzed the relationship of those parameters with respect to NK cell activity using multiple regression analysis.

The value of the mRNA level, being ΔCq obtained from real-time PCR, was log-transformed (base-10) and used for the multiple regression analysis. The NK% and other parameters were examined as quantitative and qualitative parameters, respectively, and related to the performance of NK cell activity, the reason why those parameters were chosen for that analysis. The conventional index of NK cell activity of PBMCs was assayed using the K562 cell line as the target cells. The results of the multiple regression analysis showed a significant correlation between NK cell activity and NK%, NKp46 mRNA and IFN-γ mRNA, and the prediction formula obtained from the statistical analysis comprises the aforementioned three correlation factors as shown below.

cells is insufficient to estimate the natural cytotoxicity of those cells and that determination of multiple parameters related to NK cells is necessary in order to effectively evaluate the performance of natural cytotoxicity in an indirect manner. All of this information indicates that the performance of NK cells is reflected by the strength of stimulation through cell surface activating receptors as well as the subsequent production of functional molecules as

The findings concerning NK cells described above led our research group to surmise that calculation of a prediction score based on several factors that play a role in NK cells might be utilized as an effective index to determine the performance of NK cell activity of PBMCs in an individual without the need to prepare target cells. Therefore, we statistically analyzed factors that may correlate with the NK cell activity of human PBMCs using multiple regression analysis with linear regression model, and in doing so attempted to arrive at a formula to calculate the prediction score of NK cell activity (**Figure 4**) (manuscript of an original article under preparation). In that analysis, the following parameters were used as independent vari-

CD56<sup>+</sup>

levels of NKp46, granzyme B, FasL, TNF-α and IFN-γ relative to GAPDH mRNA levels in isolated NK cells, which were measured by flow cytometry and real-time PCR, respectively.

**Figure 4.** Scheme of strategy for construction of a new index to determine natural cytotoxicity without the use of target cells. The upper box shows six candidate parameters which we selected to explore the new index in an effort to determine natural cytotoxicity. The percentage of NK cells in PBMCs represents the amount of NK cells as a quantitative parameter, while mRNA expression levels of NKp46, granzyme B, FasL, TNF-α and IFN-γ are thought to function or interfere with the strength of lytic activity of NK cells as qualitative parameters. We analyzed the relationship of those parameters with

NK cells (NK%) in PBMCs and mRNA

**5. The development of a new index: non-incubating natural** 

described above.

188 Natural Killer Cells

**killer (NINK) score**

ables for NK cell activity: the percentage of CD3−

respect to NK cell activity using multiple regression analysis.

$$\begin{array}{l} \text{NINK score} = \text{a} + \text{b}[\text{NK\%}] + \text{c}[\text{NKp46 mRNA in NK cells}] \\ \text{+d}[\text{IFN-gmRNA in NK cells}] \end{array} \tag{2}$$

The score calculated using the prediction formula with values of each parameter derived from each individual showed a better Pearson's correlation coefficient with NK cell activity than using either NK%, NKp46 mRNA or IFN-γ mRNA levels alone. These results indicate that this prediction score, named the non-incubating natural killer (NINK) score, can reflect the performance of natural cytotoxicity without the use of target cells to measure NK cell activity (patent pending). Finally, we confirmed the feasibility of the NINK score using another group of individuals. In this experiment, blood in collection tubes was stored overnight in a

**Figure 5.** The positive correlation of the NINK score with NK cell activity. The NINK score was calculated using the prediction formula comprising NK% in PBMCs and mRNA levels of NKp46 and IFN-γ in isolated NK cells derived from PBMCs from each individual, and then examined for a correlation with NK cell activity using Pearson's correlation test. The graph shows a significantly positive correlation of the NINK score with NK cell activity. The correlation coefficient and statistical significance of the p value are shown in the graph.

container box at 22°C prior to executing the assays outlined below, since the actual procedures involving PBMC preparation, NK% measurement and isolation of CD56<sup>+</sup> NK cells, followed by subsequent measurement of mRNA levels, may need to be performed on the following day after the blood is collected at a distant clinic and then transported to the institute where the subsequent procedures are performed. The results of that experiment clearly demonstrated that the NINK score calculated with values comprising NK% and mRNA levels of NKp46 and IFN-γ in NK cells obtained even from blood stored for 1 day show good correlation with NK cell activity (**Figure 5**). When individuals were divided into groups comprising low and high NK cell activity or groups comprising low and high NINK score using the averages as cut-off values, most of the low NINK score group (87.5%) showed low NK cell activity, while most of the high NINK score group (85.7%) showed high NK cell activity. Taken together, these findings indicate that the NINK score is an effective measure of the natural cytotoxicity of specimens and obviates the need to assay for NK cell activity using target cells.

## **6. Discussion**

Our demonstration indicates that the NINK score can be employed as a new index to determine the performance of natural cytotoxicity of PBMCs without the use of target cells or cell culture equipment for incubation. **Figure 6** shows the difference between the conventional index of NK cell activity and our new index of the NINK score. The conventional index of NK cell activity is useful since it can be determined by only using fluorescence or radioisotope-labeled target cells. However, it is often difficult to maintain good conditions of NK sensitivity in target cells. Technicians need to pre-culture cells under the same cell density and culture period (days) conditions in an effort to maintain good NK sensitivity of the cells. If daily measurement of NK activity is required, many pre-culture lines need to be prepared, which is unrealistic in a small institute. Additionally, since measurement of NK cell activity is based on a "bioassay", this traditional index is prone to variation between assays. This problem often troubles researchers since the altered sensitivity of target cells creates difficulties when combining results obtained from multiple assays. Similar to the measurement of NK cell activity by examining the release of 51Cr or fluorescence labeling, the lactate dehydrogenase (LDH) release assay has a problem in terms of the bioassay. In the LDH release assay, the activity of LDH derived from lysed target cells in media is measured as an absorbance following incubation of effector cells with target cells [47–49]. As an alternative approach, the level of degranulation induced by stimulation with activating receptors was assayed to assess the lytic activity of NK cells *in lieu* of using target cells [50]. In this method, the increase in cell surface expression of LAMP1/CD107a is examined by flow cytometry following stimulation. The expression of LAMP1/CD107a is high on the membrane of lytic granules but increases on the plasma membrane as a result of degranulation. However, even this method is unable to avoid the use of incubated effector cells to assay the lytic activity, which can lead to difficulties resulting from the "bioassay". In contrast, determination of the NINK score does not require the use of target cells or cell culture equipment, and is therefore free of the potential complications associated

container box at 22°C prior to executing the assays outlined below, since the actual procedures

by subsequent measurement of mRNA levels, may need to be performed on the following day after the blood is collected at a distant clinic and then transported to the institute where the subsequent procedures are performed. The results of that experiment clearly demonstrated that the NINK score calculated with values comprising NK% and mRNA levels of NKp46 and IFN-γ in NK cells obtained even from blood stored for 1 day show good correlation with NK cell activity (**Figure 5**). When individuals were divided into groups comprising low and high NK cell activity or groups comprising low and high NINK score using the averages as cut-off values, most of the low NINK score group (87.5%) showed low NK cell activity, while most of the high NINK score group (85.7%) showed high NK cell activity. Taken together, these findings indicate that the NINK score is an effective measure of the natural cytotoxicity of

Our demonstration indicates that the NINK score can be employed as a new index to determine the performance of natural cytotoxicity of PBMCs without the use of target cells or cell culture equipment for incubation. **Figure 6** shows the difference between the conventional index of NK cell activity and our new index of the NINK score. The conventional index of NK cell activity is useful since it can be determined by only using fluorescence or radioisotope-labeled target cells. However, it is often difficult to maintain good conditions of NK sensitivity in target cells. Technicians need to pre-culture cells under the same cell density and culture period (days) conditions in an effort to maintain good NK sensitivity of the cells. If daily measurement of NK activity is required, many pre-culture lines need to be prepared, which is unrealistic in a small institute. Additionally, since measurement of NK cell activity is based on a "bioassay", this traditional index is prone to variation between assays. This problem often troubles researchers since the altered sensitivity of target cells creates difficulties when combining results obtained from multiple assays. Similar to the measurement of NK cell activity by examining the release of 51Cr or fluorescence labeling, the lactate dehydrogenase (LDH) release assay has a problem in terms of the bioassay. In the LDH release assay, the activity of LDH derived from lysed target cells in media is measured as an absorbance following incubation of effector cells with target cells [47–49]. As an alternative approach, the level of degranulation induced by stimulation with activating receptors was assayed to assess the lytic activity of NK cells *in lieu* of using target cells [50]. In this method, the increase in cell surface expression of LAMP1/CD107a is examined by flow cytometry following stimulation. The expression of LAMP1/CD107a is high on the membrane of lytic granules but increases on the plasma membrane as a result of degranulation. However, even this method is unable to avoid the use of incubated effector cells to assay the lytic activity, which can lead to difficulties resulting from the "bioassay". In contrast, determination of the NINK score does not require the use of target cells or cell culture equipment, and is therefore free of the potential complications associated

NK cells, followed

involving PBMC preparation, NK% measurement and isolation of CD56<sup>+</sup>

specimens and obviates the need to assay for NK cell activity using target cells.

**6. Discussion**

190 Natural Killer Cells

**Figure 6.** Illustration of differences between the conventional NK cell activity index and new NINK score index. To measure the NK cell activity, NK-sensitive target cells such as K562 cells need to be prepared and appropriately maintained prior to use in the assay, and then PBMCs need to be incubated with the target cells for 4 h, which is often troublesome and can cause large variation between assays. In contrast, a determination of the NINK score does not require the use of target cells or cell culture equipment for incubation. Technicians simply have to measure NK% and isolate NK cells from PBMCs. Messenger RNA levels of the two genes in the NK cells (retrieved from frozen storage) can then be determined and the NINK score calculated using the prediction formula with the measure values of the three parameters.

with the inappropriate preparation of those cells. Additionally, on any given day, technicians will measure NK%, isolate NK cells from PBMCs by flow cytometry or magnetically, and then store the NK cells frozen before subsequent assays are performed. At some later time, mRNA levels of NKp46 and IFN-γ can be determined in the frozen cells by real-time PCR and the NINK score calculated using the prediction formula with the measured values of the three parameters. Thus, employing our new approach based on the NINK score can free the technician from continuous pre-culture of target cells required in the traditional NK activity assay, and provides a stable tool to measure the performance of NK cells in each individual and independent bioassay. The NINK score is beneficial as a measure of NK cell performance in each individual, and is a requirement for clinicians working in the field of cancer therapy as well as for the assessment of healthy individuals interested in preventive medicine and health promotion. Since NK cells are linked to the CTL response in acquired immunity as mentioned above, it may be valuable for clinicians to assess the progress of the functional status of NK cell activity in a patient following cancer therapy in a steady manner by using the NINK score. It would be better if the NK cell activity could be checked in terms of a solid parameter such as bone density or body fat percentage, which can be achieved by measuring the NINK score. A vast amount of data concerning the relationship between NK cell activity and various kinds of factors related to lifestyle can be collected using the NINK score, the results of which might lead to new insights concerning healthier and disease-preventive lifestyles from an immunological perspective. The ability to store isolated NK cell specimens temporarily in a frozen state provides a benefit in that the following measurements of mRNA levels can be performed simultaneously, which allows for an accurate comparison of NINK scores among specimens. Additionally, the NINK score may be useful as a screening device to identify NK cell-activating natural compounds from within a compound library using *in vitro* experiments of PBMC cultures. An actual effect of some compound on NK cell activity can then be examined using animal experiments (mice), even where a modified method of the NINK score may be utilized to measure the performance of NK cell activity *in lieu* of using the method with NK cellsensitive murine targets such as YAC-1 cells [51, 52]. In contrast, the traditional method involved in the NK assay using target cells is unsuitable for these purposes given the lower stability resulting from the use of bio-assays as mentioned above. The NINK score might contribute to an assessment of the efficacy of a drug at different stages of cancer treatment in patients, as well as an assessment of the effectiveness of different lifestyles, exercise and food consumption in maintaining good health in individuals. We hope that the NINK score would be utilized in various fields to facilitate the promotion of general health in the population.
