Type-2 diabetes (T2D) is a complex disorder characterised by hyperglycaemia, insulin resistance (IR) and chronic inflammation of insulin target tissues. T2D is now also considered an inflammatory disease, affecting both innate and acquired immune systems, skewing them towards a pro-inflammatory phenotype [reviewed in (1)]. There is also growing evidence for autoimmune involvement in T2D overlapping with the pathophysiology of T1D (2). Inclusion of cellular autoimmunity with traditional diabetic parameters is now reflected by a number of diabetes sub-types that do not fit into T2D or T1D including type 1.5 diabetes mellitus or latent autoimmune diabetes of the adult (LADA) or the young (LADY) and double diabetes (mixed symptoms of T1D and T2D) (3, 4). This recalibration of diabetic phenotypes led us to re-examine the use of obese pre-clinical diabetic models, as these no longer adequately reflect the human clinical context (5). Most of the studies conducted on inflammation in T2D involved overweight or obese subjects and obesity is clearly associated with a low-grade chronic inflammation (6, 7). Since T2D is not always associated with obesity, it is essential to dissect the exact contributions of obesity and diabetes in the immune phenotype.

Although there are many well-known obese models of T2D, there are few non-obese models (8). Goto-Kakizaki (GK) rats are a non-obese model that present a pre-diabetic phase before becoming spontaneously diabetic, similar to the human T2D pathophysiology (9). However, despite the growing immune literature in human T2D, there is little data available on the GK rat immune system. Circulating white blood cell (WBC) levels have been reported to be unchanged in GK rats, although they are known to be biased towards a T_h2 phenotype (10), have fewer B-cells, a higher IgM production and reduced monocyte-phagocytic activity (5, 11). Surprisingly, this phenotype is somewhat contrary to that observed in the Type 1 model “Diabetes-prone Biobreeding (DP-BB) rats” (5) and high-fat diet induced obesity in mice (12). In contrast, previous works have demonstrated a marked proinflammatory profile in pancreatic islets in GK animals, with macrophage infiltrations and upregulation of mRNA expression of several proinflammatory cytokines in this tissue (13, 14). Subsequently fibrosis develops within the islets which further alter their normal secretory function (14).

Given the potential for the GK rat model to represent non-obese and emerging diabetes subtypes, we used a wide-ranging rat immune system profiling panel (15) and multiplex cytokine panel to characterise adult, diabetic GK rats. By investigating levels of principal circulating immune cell subsets and plasma cytokines, we aim to provide a baseline immune profile to further study the role of immune system in the aetiology of T2D. Furthermore, we reanalysed previous transcriptional datasets (16) for the consequences of exposure to this cytokine milieu in the liver, adipose tissue and muscles.

T2D has been commonly associated with low grade inflammation, but it is not clear whether this inflammatory profile is related to diabetes or to obesity. The aim of the present study was to examine peripheral inflammatory status in an animal model of T2D without obesity. Surprisingly, our results suggest that GK rats do not show a low-grade inflammatory profile as reported in obese models, but instead exhibit a marked decrease of the examined immune markers. Beta cell dysfunction in the GK rats is the primary defect that leads to overt hyperglycaemia and marked islet inflammation (13, 14, 35). Although, systemic inflammation is often associated with IR, it is not the sole contributor to the development of IR. Ycaza et al. reported that adipose insulin resistance did not correlate with markers of subcutaneous adipose inflammation (36). GK rats develop hepatic insulin resistance at a young age (37), while whole-body insulin resistance (which is not associated with liver, adipose or muscle inflammation) appears later in life (14). Pitasi et al. recently showed increased levels of GSK3B protein in the GK pancreas (14), which has been reported to be elevated in insulin target tissues in humans with T2D, and (38) is correlated with decreased insulin sensitivity (39). Thus, we postulate that a possible over expression of GSK3 in insulin target tissues in the GK rat, could participate to the development of insulin resistance in this model.

Here, we provide a detailed description of the GK rat immune system, which shows limited peripheral inflammation, in addition to a lower percentage of phagocytes (macrophages, monocytes, dendritic cells, granulocytes and B-cells). This partially confirms impaired phagocytosis and antigen presentation as previously described in GK rats (5, 10, 11), however Zhai et al. reported a higher percentage of activated B cells in obese diabetics, which does not fit our observation (40), suggesting a major role of obesity in the inflammatory profile. The observed low percentage of T_h cells in the GK rats may also have an effect on the activation status of B- and cytotoxic T-cells, and may further explain their lower numbers in GK rats. Diabetes has also been known to disrupt monocyte recruitment to sites of injury, thereby impairing phagocytosis and impeding the switch from pro- to an anti-inflammatory state (41). Interestingly, wound healing in diabetics is also known to be hindered as diabetes prompts neutrophils to cause tissue damage by NETosis (42), which further contributes to the inflammation causing a vicious pathological cycle (43).

We have seen a decrease in CD8+ T_c cells in our GK rats, which are normally responsible for eliminating viral-infected host cells and thus is also in line with the predisposition of T2D patients towards severe infections (44). This agrees with the reported decrease in naive T_c cells, and accumulation of EMRA immunosenescent cells in man (45).

A higher CD4/CD8 ratio conventionally corresponds to a healthier immune system and estimates the likelihood of development of infections. In rats, this ratio is strain specific (46). GK rats show a significantly higher CD4/CD8 ratio (Supplementary Figure 8), which can be attributed to the low number of T_c cells and may not necessarily represent a healthier immune system compared to Wistar rats. Interestingly, a recent study also reported a higher CD4/CD8 ratio after a glucose bolus in both diabetics and non-diabetics (47), indicating a lymphocyte redistribution, which may be the case for our GK rats as well. We also examined T_h1/T_h2, T_c1/T_c2 and T_h17/Treg ratio, which have been reported/reviewed to be impaired in T2D with comorbid obesity (48), inflammatory diseases such as Behcet’s disease (49) and metabolic disorders (50) respectively. We did not see significant differences for any of these ratios in our GK rats (p>0.05, Mann-Whiney test, data not shown).

We interpret the overexpression of CD161a in GK rats as most likely to be an already activated defence system as a compensatory mechanism for their compromised immune system to combat an invading or existing pathogen/infection. It is logical to conclude that this is achieved by the activation of monocytes and NK cells to target infected cells and by secretion of cytokines to further activate these cells or this can be indicative of healthy cell/tissue destruction. Altered immune functionality in both human B- and NK-cells has been reported to be mediated by global hypermethylation in obese (BMI>30) T2D patients (51). The data on human NK cells is somewhat unclear. Meta-analysis of 13 NK cell studies identified 3 with decreased levels/activity, 2 reported elevated levels and 8 of them reported no differences between circulating NK cells in T2D patients and healthy controls (52).

RORγT+CD8+ lymphocytes are a somewhat underexplored population, but they are very pro-inflammatory (53). GK rats show pancreatic islet inflammation (35) and in our data, peripheral inflammation appears to be limited to the observed increase in T_c17 cell population. T_c17 cells are produced under inflammatory conditions in a manner similar to their CD4+ counterparts, T_h17 cells (54), skewing the immune system towards a pro-inflammatory phenotype, although, as outlined above IL-17 levels are lower in GK rats. Interestingly, CD8+T_c17 cells have been reported to show a strong repression of SOCS3 expression (55), a pivotal gene regulating insulin sensitivity (56) in addition to driving inflammation [as reviewed (57)]. T_c17 cells have been shown to degrade pancreatic beta islet cells and induce hyperglycaemia in mice (57). Thus, it is quite likely that this cell population contributes to the vicious diabetic and (peripheral) inflammatory aetiopathology of the GK rats.

Similarly to the T_c17 cells, T_c2 cells have not been highly investigated in the light of diabetes but have been reported to be more diabetogenic compared to naive T_c cells (58). However, in T2D with comorbid tuberculosis, there appear to be a clear role for both T_c17 and T_c2 cells (59). Although, there is a preponderance of literature on immune changes in complicated diabetes and unfortunately, the immune status in uncomplicated human T2D is still ambiguous.

Circulating baseline cytokine levels have been studied in T2D for many decades (60). In this study, we present the first account of multiplexing a panel of 27 cytokines simultaneously from the same samples to get a clearer, detailed overall picture of the GK immune system (Figure 2). The reduced levels of the cytokines in GK rats is most likely due to low number of circulating CD11b/c+ cells, principally monocytes/macrophages in addition to low percentages of CD45RA+ (B-cells) and CD8+ (T_c cells) cell populations. Interestingly, despite an overall decrease of both anti- and pro-inflammatory cytokines plasma levels, our results reveal a significant increase of interferon gamma-induced protein 10 (IP-10), a pro-inflammatory chemokine which plays an important role in the aetiology of inflammatory diseases. Previous studies suggest that IP-10 is associated with metabolic disorders. Accordingly, IP-10 levels were increased in the early stage of type 1 diabetes (61) and were also predictive of insulin resistance and diabetes in patients suffering from nonalcoholic fatty liver disease (62). Previous studies suggest a positive association between leptin levels and IP-10 in type 2 diabetic patients (63), which also fits our GK rat data (Supplementary Figure 5V).

We also saw an increase in leptin levels in our GK rats. However, this does not necessarily indicate an increased secretion of proinflammatory cytokines from GK adipocytes due to tissue specific gene expression. Previously, we have shown that the expression of cytokines such as IL-6, IL-1β and TNF-α were increased in the GK islets, but not in the adipose tissues (14). Thus, while we observe elevated levels of adipokines associated with the increased adiposity in GK rats, we do not see an increased pattern secreted proinflammatory cytokines.

Thus, it is clear from our study that the GK cytokine profile alone can be quite misleading, unless amalgamated with cognate immune date. However, there is no clear pro-inflammatory cytokine profile as previously thought. Our data suggests that the inflammatory milieu in GK rats is most likely due to T_c17 cells and activated monocytes, NK cells as represented by elevated CD161a+ cells. In the human context, T2D patients have been shown to have higher IL-17 levels compared to healthy controls, which contribute to the pro-inflammatory phenotype commonly observed in T2D (64, 65), thereby justifying our hypothesis.

This disproof of the proinflammatory profile in GK rats led us to investigate the gene expression pattern in liver, muscle and adipose tissues: all key tissues that not only play important roles in glucose homeostasis affecting glycaemic status but are also exposed to the circulating immune cells and cytokines. Therefore, we re-analysed previously published GK rat micro-array expression data for these three key diabetes related tissues. The extracted biological pathways from the gene expression profiles confirmed a downregulated pattern of most involved genes in the GK rats. The most prominent gene pathway was diabetic cardiomyopathy, fitting previous reports of ventricular hypertrophy, impaired diastolic function, and cardiomyopathy in GK rats (66, 67) that may be mediated by CRP (68). This peripheral gene expression profile contrasts sharply with the inflammatory profile of the GK rat pancreatic islets, where similar to diabetic human islets (69), we consistently show increased pro-inflammatory cytokine levels, associated with increased pro-inflammatory immune cell numbers around and within the pancreatic islets (13, 14, 35). This profound islet inflammation is a contributor to the impairment of beta cell growth and function, which constitutes the main characteristics of T2D in this model. GK rats serve as an excellent model to study the development of T2D, and the transition from prediabetic to a fully diabetic phenotype may yield insight into the development of T2D.

In conclusion, our study reports 3 lines of evidence proving that, outside the pancreatic islets, GK rats present a limited pro-inflammatory profile unlike previously thought. Additionally, non-obese GK rats exhibit a marked immune dysfunction as indicated by the overall blunted levels of cytokines and changes in the distribution of immune cells. Thus, it is also clear that there is a knowledge gap between uncomplicated human non-obese T2D, and the immune system that needs to be addressed in detail and that GK rats can serve as a good model for this.

The raw data supporting the conclusions of this article are available from the corresponding author upon reasonable request.

All experimental procedures were carried out in accordance with the European Union guidelines for the use of animals for experimental purposes (Council Directive 2010/63/EU) and with the French ones (Directive 87/148, Ministère de l’Agriculture et de la Pêche – Apafis #16924).

Conceptualization, SS, MD, and JT. Data collection, MH, CP, SS, DB, JM, and MD. Data analysis, SS, CH, MH, CP, MD, JT. Writing and editing, all authors. All authors contributed to the article and approved the submitted version.

This work was funded by French National Research Agency (ANR-17-CE37-0020, Maternal diabetes and neuropsychiatric vulnerability in offspring: role of DNA methylation -MADAM) and by Fonds National de Recherche Luxembourg (INTER/ANR/16/11568350 ‘MADAM’) and (Age Acceleration through the life course (C19/SC/13650569 ‘ALAC’). This work was supported by the University of Bordeaux and by the AlimH department of the Institut National de la Recherche Agronomique (INRAE).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.