Functional role of natural killer T cells in non-obese pre-diabetes model mice

Abstract

Pre-diabetic patients have a high risk of developing diabetes as well as other associated diseases. From the viewpoint of risk assessment and to assist the development of protective therapies, we focused on the functional role of natural killer T (NKT) cells in pre-diabetes. We found that the expression of an NKT cell marker gene, Va14-Ja18, was significantly lower in specific tissues/organs such as adipose tissue and pancreas in non-obese pre-diabetes model mice than in their normal littermates. Subsequently, in the pre-diabetes model mice, Va14-Ja18 was activated with α-galactosylceramide (α-GalCer) and its effect on glucose tolerance was estimated. The simultaneous injection of α-GalCer and lymphocytes improved glucose tolerance with its maximum effect on the 3rd day. An analysis of circulating cytokine levels revealed that interferon-γ, which is a pro-inflammatory cytokine, was secreted only on the 1st day after treatment with α-GalCer and that interleukin (IL)-4, which is an anti-inflammatory cytokine, was secreted from the 1st to the 4th day. The prolonged secretion of IL-4 was thought to substantially contribute to the improvement of glucose tolerance. Based on these results, the functional role of NKT cells in pre-diabetes is to improve metabolic dysfunctions.

Electronic supplementary material

The online version of this article (10.1007/s10616-017-0157-5) contains supplementary material, which is available to authorized users.

Introduction

The number of pre-diabetic patients has recently been dramatically increasing worldwide (Rydén et al. 2014; Seino et al. 2010). Pre-diabetic patients have a high risk of developing not only diabetes but also various associated conditions such as high blood pressure, dyslipidemia, and obesity. Hyperglycemia might be a predominant-factor involved in the development of those conditions, although the severity of hyperglycemia in pre-diabetes is mild. Therefore, it is important to protect pre-diabetes patients against the development of such conditions.

From the viewpoint of risk assessment and to assist future studies on the development of protective therapies, it is essential to define the pre-diabetes state quantitatively. Until recently, however, the term pre-diabetes has been referred to only as a state of not-yet-diabetes. In fact, there has been no report on a complete theoretical definition of pre-diabetes.

According to the method described by Bali et al. (1995) we formerly established a novel mouse strain with impaired glucose tolerance by systemic heterozygous knockout of the glucokinase gene (Gk) (Saito et al. 2015). To characterize this strain, the glucose tolerance test was conducted with a series of strains with backcross numbers of 5–10. The threshold values between the normal and pre-diabetic conditions and the pre-diabetic and diabetic states could be determined based on the accumulated results of the glucose tolerance tests. Thus, the determined threshold value between the pre-diabetic and diabetic states was thought to be applicable as a quantitative standard in the risk assessment of diabetes (Saito et al. 2016).

Here, we have focused on investigating immunological dysfunctions in pre-diabetes because such dysfunctions have been thought to underlie various diseases. To date, a large number of studies have revealed the involvement of natural killer T (NKT) cells in diabetes (Ohmura et al. 2010; Brooks-Worrell et al. 2012; Ji et al. 2012; Schipper et al. 2012). Therefore, we suspected that their involvement could be observed even in the pre-diabetic state. In fact we found that expression of an NKT cell marker in the non-obese pre-diabetic model mice was significantly lower than that in their normal littermates (see “Results” section).

According to former reports, however, the details of the involvement of NKT cells in diabetes are complicated. For instance, the activation of NKT cells worsened inflammation and glucose tolerance in adipose tissue in high-fat diet-fed obese mice (Ohmura et al. 2010). On the contrary, there was a contradictory report that NKT cells could improve glucose tolerance in obese mice (Ji et al. 2012). Furthermore, it was reported that the activation of NKT cells could prevent autoimmune diabetes in non-obese diabetic mice (Hong et al. 2001). Although obesity does not necessarily indicate a diabetic state, it is still unclear whether NKT cells may act to improve or worsen the dysfunctions associated with diabetes.

The apparent functional bipolarity of NKT cells is understood to be associated with the function of macrophages, which also show functional bipolarity (Takatsu et al. 2009; Wentworth et al. 2010). When activated NKT cells secrete inflammatory cytokines, macrophages are converted to the inflammation-inducing M1 type. In contrast, when activated NKT cells secrete anti-inflammatory cytokines, macrophages are converted to the inflammation-reducing M2 type. These known aspects of immunology suggest that the role of NKT cells as well as the effects of their activation should depend upon obesity, blood glucose, insulin secretion, and various other metabolic factors. Therefore the functional role of NKT cells in pre-diabetes cannot be assumed to simply be an average of their roles in the normal and diabetic states.

The purpose of this study is to analyze the functional role of NKT cells in non-obese pre-diabetic model mice. We hypothesized that the role of NKT cells in the pre-diabetic state might be to improve the metabolic dysfunctions and protect against disease progression. To test this hypothesis, we initially analyzed the net effects of the non-obese pre-diabetic state on the expression of an NKT cell marker. Then NKT cells were activated by α-galactosylceramide (α-GalCer) in the non-obese pre-diabetic mice, (Kawano et al. 1997; Bendelac et al. 2007) and the secretion of inflammatory and anti-inflammatory cytokines was analyzed. Finally, the effect of α-GalCer on glucose tolerance was investigated to support the protective role of NKT cells.

Materials and methods

Mouse breeding

Ten-week-old male C57BL/6 mice and the pre-diabetes model mice were used in this study. The pre-diabetes model mice were those previously developed by systemic heterozygous knockout of Gk and registered as B6;129-Gckt ^m1Tms (http://ilarlabcode.nas.edu/search_codes_full.php?labcode_id=9447&user_id=57616) (Saito et al. 2015), and here designated as G-mice. Normal littermates of the pre-diabetes model mice were used as the control wild mice (W-mice). Breeding was conducted in a specific pathogen free room under conditions of 12 h illumination and 12 h darkness each day. Every mouse was fed with control diet (CD) (solid diet MF, Oriental Yeast Co., Ltd., Tokyo, Japan) for 10 weeks from birth. Then, the diet was either changed to a high-fat diet (HFD) (HFD 32, CLEA Japan, Inc., Tokyo, Japan) or kept unchanged for the control group to prepare 4 groups: HFD fed G-mice (G-HFD), CD fed G-mice (G-CD), HFD fed W-mice (W-HFD), and CD fed W-mice (W-CD). All animal experimental procedures were conducted according to the guidelines of the “guide for the Care and use of the Laboratory animals” of Tokyo University of Agriculture and Technology, Japan.

Quantitative real-time PCR

Va14-Ja18 was selected as an NKT cell specific gene marker and its expression intensity was measured by quantitative reverse transcription polymerase chain reaction (qRT-PCR) in the liver, spleen, adipose tissue, and pancreas of each mouse. Nos2, Tnfa, Il6, and Mcp1 were selected as gene markers specific for inflammatory M1 type macrophages. Arg1, Fizz1, and Ym1 were selected as gene markers specific for anti-inflammatory M2 type macrophages. Genes expression intensity was measured by qRT-PCR in the adipose tissue. The samples were homogenized in ISOGNE II reagent (Nippon gene, Tokyo, Japan), and the total RNA was prepared from each sample according to the manufacturer’s instructions. Oligo (dT)-primed cDNA was prepared from total RNA using SuperScript II (Clontech Laboratories, Inc., Tokyo, Japan). The expression intensities were analyzed by qRT-PCR using primers listed in Supplementary Table 1. The expression of each gene was normalized to the level of Hprt or Gapdh. SYBR Premix Ex Taq II (Clontech Laboratories, Inc.) was used to amplify and monitor cDNA.

Glucose tolerance test

Mice were challenged with an oral administration of 2 mg/g body weight of glucose following an overnight fast. Blood samples were drawn from the tail vein at 0, 15, 30, 60, 90, 120, and 180 min post-treatment for the measurement of the blood glucose level at each time point. Glucose concentration was measured with a G-sensor (ARKRAY Inc., Kyoto, Japan).

Administration of α-GalCer and lymphocytes

An α-GalCer (KRN7000, Kirin Company, Ltd., Tokyo, Japan) solution (1 mg/ml, 100 ng/g body weight) and 200 μl of a phosphate buffer solution containing 2 × 10⁶ lymphocytes including NKT cells were administered to a test mouse simultaneously by intraperitoneal injection. The lymphocytes were separated from the spleen of a wild mouse and purified using a cell sorter (FACSAria™ II, BD Biosciences, San Jose, CA, USA) beforehand.

Cytokine analysis

Blood was sampled using a heparinized hematocrit tube and 200 μl of the sample was transferred into a 0.6 ml-microtube. The microtube was centrifuged at 1200 × g for 20 min at 4 °C to isolate plasma as the supernatant. A 50 μl aliquot of the plasma was assayed for IFN-γ and another 50 μl aliquot of the plasma was assayed for IL-4 using a Mouse IFN-γ ELISA kit (RayBiotech, Inc., Norcross, GA, USA) and a Mouse IL-4 ELISA kit (RayBiotech, Inc.), respectively.

Analyses of CD1d mRNA expression and free fatty acids

The expression of CD1d in adipose tissue was analyzed by qRT-PCR using primers listed in Supplementary Table 1. The concentrations of free fatty acids in plasma were analyzed with a LabAssay™ NEFA kit (Wako Pure Chemical Industries, Ltd., Osaka, Japan) and expressed as milli-equivalent (mEq)/l.

Results

Net effects of non-obese pre-diabetic state on the expression of an NKT cell marker

A comparative study of the Vα14-Jα18 expression intensity was conducted using G-HFD, G-CD, W-HFD, and W-CD. After initial feeding with CD for 10 weeks and successive feeding with HFD or CD for 8 weeks, the liver, spleen, adipose tissue, and pancreas were isolated from each mouse and assayed for the expression of an NKT gene marker Va14-Ja18. This process was repeated to collect 5 and more data per condition.

The intensities of Va14-Ja18 expression in W-HFD mice relative to those in W-CD mice were 0.4 (liver), 0.8 (spleen), 0.4 (adipose tissue), and 0.5 (pancreas) (Fig. 1). These results indicated that the HFD caused a decrease in Va14-Ja18 expression. The W-HFD mice were obese and therefore this result should reflect the effect of obesity. On the other hand, the intensities of Va14-Ja18 expression in G-HFD mice relative to those in W-CD mice were 0.5 (liver), 0.7 (spleen), 0.75 (adipose tissue), and 0.4 (pancreas). These results should reflect the effect of the diabetic state, i.e., a mixture of the effects of hyperglycemia and obesity. These results of a decrease of Va14-Ja18 expression in the diabetic and obese states confirm the results previously reported elsewhere (Ji et al. 2012).

Fig. 1.

Expression profiles of an NKT cell marker to HFD feeding and/or non-obese pre-diabetic condition. A Liver, B spleen, C adipose tissue, D pancreas. Data were normalized against the results for wild-type control diet (W-CD) mice and expressed as mean ± standard error of the mean (SEM) for n = 9 (liver, spleen), 8 (adipose tissue), or 6 (pancreas). *p < 0.1

In contrast, the intensity of Va14-Ja18 expression in G-CD mice was significantly lower than that in W-CD mice, though it was observed only in adipose tissue and pancreas. The G-CD mice exhibited moderate hyperglycemia but were not obese. Such a decrease of Va14-Ja18 expression in non-obese mice is a novel finding. This suggests that the decrease of expression of an NKT cell marker in adipose tissue might be a dysfunction-worsening factor in the pre-diabetic state. Following this finding, we focused on the phenomena observed in adipose tissue.

Effects of pre-diabetic and diabetic states on the macrophage types

The worsening of the pre-diabetes-associated hyperglycemia in the G-CD mice implied the conversion of macrophages type to the pro-inflammatory M1 type rather than the anti-inflammatory M2 type. The expression profiles of M1-specific and M2-specific marker genes, however, indicated no clear pattern (Fig. 2). Against our expectation, none of the 4 inflammatory marker genes (Nos2, Tnfa, Il6, and Mcp1) showed higher expression levels in G-CD mice than those in W-CD mice (Fig. 2A–D). On the other hand, among 3 anti-inflammatory marker genes, only Ym1 showed a lower level of expression in G-CD mice than in W-CD mice (Fig. 2E–G). Therefore, it should be difficult to evaluate the functional role of NKT cells in pre-diabetes based only on the dynamic balance between M1-type and M2-type macrophages.

Fig. 2.

Expression profiles of gene markers specific to M1 or M2 type macrophages. A Nos2, B Il6, C Tnfa, D Mcp1, E Arg1, F Fizz1, G Ym1. Gene markers of M1: A–D, Gene markers of M2: E–G. Data were normalized against the results for W-CD mice and expressed as mean ± SEM for n = 6. *p < 0.1, **p < 0.05, ***p < 0.005

Effects of the introduction of α-GalCer and NKT cells on the glucose tolerance of pre-diabetic model mice

We investigated the direct effects of NKT cell activation on glucose tolerance. After the administration of NKT cells and α-GalCer into G-CD mice, glucose tolerance tests were conducted on the 3rd and the 6th day. Consequently, an improvement of glucose tolerance was observed until the 6th day (Fig. 3A–C). This effect was strongest on the 3rd day and then slightly decreased. To investigate why the improving effect decreased after the 3rd day, the concentrations of circulating cytokines were analyzed. One pro-inflammatory cytokine, IFN-γ, was secreted markedly on the 1st day (Fig. 3D). In contrast, one anti-inflammatory cytokine, IL-4, continued to be secreted for 4 days (Fig. 3E). In total, the anti-inflammatory effect should have surpassed the inflammatory effect, and this may explain why an improvement of glucose tolerance could be observed on the 3rd day. After that, however, the secretion of IL-4 stopped, which may explain why the improving effect decreased after the 3rd day.

Fig. 3.

Effects of NKT cell activation on glucose tolerance and dynamic excretion of cytokines. A Results of glucose tolerance test before administration of NKT cells and α-GalCer. B Results of glucose tolerance test on the 3rd day after administration of NKT cells and α-GalCer. Solid line is the control (no administration). Dotted line indicates the post-ministration. C Results of glucose tolerance test on the 6th day after administration of NKT cells and α-GalCer. Solid line is the control (no administration). Dotted line indicates the post-ministration. Data were expressed as mean ± SEM for n = 5. D IFN-γ secretion, E IL-4 secretion. Data of D and E: mean ± SEM for n = 5. *p < 0.1, **p < 0.05, ***p < 0.005

Discussion

NKT cells are understood to be activated via antigen-presenting cells (APCs). One of the major histocompatibility complex antigens, CD1d, is expressed by APCs. CD1d can also be expressed by adipocytes; this causes adipocytes to become APCs, and consequently they can activate NKT cells directly.

In support of this, the expression of CD1d by adipocytes was previously well discussed (Blumberg et al. 1995). In the human Simpson–Golabi–Behmel syndrome (SGBS) cell line, CD1d expression was markedly low (Satoh et al. 2016). However, its expression increased sufficiently after differentiation of the SGBS cells to adipocytes. We also confirmed that CD1d expression in the mouse 3T3-L1 pre-adipocyte cell line increased markedly after differentiation of the cells into adipocytes (data not shown). According to this assumption that NKT cells are activated by CD1d-expressing adipocytes, the activity of NKT cells could be directly correlated with CD1d expression.

On the other hand, insulin receptors are expressed on the cell surface of adipocytes. The binding of insulin to these receptors inhibits the intracellular lipase that would otherwise hydrolyze neutral lipids to generate free fatty acids that would then be secreted into blood vessels. The circulating concentration of insulin was lower in G-CD mice than in W-CD mice. Therefore, the lipase activity was thought to be higher in G-CD mice than in W-CD mice. Consequently, the concentration of fatty acids in blood should be higher in G-CD mice than that in W-CD mice.

To confirm this speculation, we analyzed the expression level of CD1d in adipose tissue and also the concentration of free fatty acids in blood. As expected, CD1d expression was lower in G-CD mice than in W-CD mice, and the free fatty acids concentration was higher in G-CD mice than in W-CD mice (Fig. 4).

Fig. 4.

Expression of CD1d and excretion of free fatty acids. A CD1d expression intensity, B free fatty acids concentration. Data of CD1d were normalized against the results for W-CD mice. Data were expressed as mean ± SEM for n = 6. **p < 0.05, ***p < 0.005

In conclusion, the functional role of NKT cells in non-obese pre-diabetes is to improve metabolic dysfunctions.

Electronic supplementary material

Below is the link to the electronic supplementary material.