Background
Low-grade inflammation in adipose tissue is a hallmark of type 2 diabetes and adiposity [
1]. Adipose tissue inflammation is attributed to aberrant proportions of resident macrophages: an increase in inflammation-promoting M1 macrophages and a decrease in anti-inflammatory M2 macrophages [
2,
3]. The imbalance of adipose tissue macrophages is triggered by cytokines released by enlarged adipocytes, as well as an accumulation of immune cells [
4]. On the other hand, type 2 innate lymphoid cells and their target eosinophils repress M1 macrophage activation in adipose tissue, and subsequently attenuate metabolic disorders [
5,
6]. Moreover, regulatory T cells inhibit chronic inflammation by the secretion of anti-inflammatory cytokines, such as interleukin (IL)-10 and transforming growth factor (TGF)-β1 [
7,
8]. Myeloid derived-suppressor cells (MDSCs) are another cell type that has the potential to modulate inflammation [
9‐
11]. MDSCs, as their name represents, are known to be derived from myeloid lineage progenitors, and intensely express both CD11b and Gr-1 as cell surface markers in mouse [
9,
10]. MDSCs produce vast amounts of IL-10 and arginase, resulting in strong attenuation of immune activation in tumors and septic regions [
9,
10]. Previously, CD11b
+, Gr-1
+ MDSCs were shown to inhibit the progression of type 2 diabetes in obese mice [
12]. Although manipulation of MDSCs is beneficial in anti-diabetic treatment, an induction procedure for MDSCs has not yet been established.
Propolis is a resinous mixture that honey bees produce by mixing their waxes and saliva with exudate from botanical substances. Thus, the composition of propolis is determined by geographic location and the bee’s genetics [
13,
14]. Brazilian green propolis is produced by Africanized bees, which are hybrids between
Apis mellifera scutellata (African honey bee) and
Apis mellifera (European honey bee), and is rich in organic substances mainly from
Baccharis dracunculifolia [
15,
16]
. Brazilian propolis ethanol extract (PEE) exerts various biological effects, including tumoricidal, immunomodulatory, and tissue repair effects [
17‐
20]. Some of these effects are considered to be attributable to cinnamic acid derivatives and flavonoids, including artepillin C, p-coumaric acid, caffeic acid phenethyl ester (CAPE), and kaempferol [
21‐
24]. Traditionally, supplements containing propolis are taken to prevent the progression of metabolic disorders. In agreement with this type of propolis use, orally administration of propolis was evinced to have potential to restore type 2 diabetes in human [
25,
26]. Previously, we found that intraperitoneal injection of PEE ameliorates type 2 diabetes in
ob/ob mice, although it did not influence the body weight gain of the mice [
27]. In these mice, PEE modulated the number of eosinophils and M1 macrophages in mesenteric adipose tissue where adiposity is closely associate with insulin sensitivity. However, it is still elusive which cells are direct targets of propolis in the adipose tissue. In this study, we found that PEE induces MDSCs in adipose tissue. We attempted to identify cells which are source of PEE-elicited MDSCs. We also tried to identify chemicals which increase MDSCs
in vitro and
in vivo.
Discussion
PEE represses the progression of type 2 diabetes in mouse and rat models [
27,
34‐
37]. This beneficial effect of PEE on type 2 diabetes has been observed in several ways. For example, PEE suppresses feeding through the secretion of leptin from adipocytes [
37]. PEE also stimulates glucose uptake in skeletal muscle by recruiting glucose transporter 4 to the plasma membrane [
38]. Previously, we demonstrated that intraperitoneal injection of PEE attenuated adipose tissue inflammation in leptin-deficient
ob/ob mice resulted in a dramatic repression of metabolic deterioration [
27]. In this study, we demonstrate that PEE causes a rapid accumulation of MDSCs in the visceral adipose tissue of mice. MDSCs have the potential to repress inflammation in adipose tissues [
12], therefore it is likely that PEE alleviates inflammation through the induction of MDSCs. MDSCs accumulate in the adipose tissue of obese and lean mice, implying that PEE has the potential to prevent both the onset and progression of type 2 diabetes.
In this study, MDSCs accumulated in visceral adipose tissue and in peritoneal fluid. The source of the PEE-elicited MDSCs is still unclear; they are possibly recruited from other tissue such as bone marrow or differentiated from resident macrophages or circulating monocytes. We demonstrated that PEE directly induces MDSCs from cultured peritoneal macrophages. Peritoneal macrophages can be divided into two populations: F4/80
high MHC-II
− predominant in intact mice, and F4/80
mid MHC-II
+ predominant in thyoglycolate-elicited mice [
39,
40]. Since
in vivo F4/80
high MHC-II
− and
in vitro F4/80
mid MHC-II
+ macrophages can produce nitric oxide in response to LPS stimulation, both cell types display characteristics of M1 macrophages [
39]. On the other hand, PEE also increases the expression of the MDSC markers
Ly6g and
Il10 in the cloned cell line J774.1, which constantly express the M1 macrophage markers tumor necrosis factor (TNF) -α and inducible nitric oxide synthase [
41]. PEE induced MDSC markers in GM-CSF-treated bone marrow cells (M1 macrophages), but not in M-CSF-treated cells (M2 macrophages). These results collectively suggest that the PEE-elicited MDSCs originate from M1 macrophages in the visceral organs or blood.
Of the twelve resinous compounds of PEE that were assessed, kaempferol increased
Ly6g and
Il10 mRNA in cultured macrophages when used at a dosage comparable to that found in PEE. Moreover, repeated injections of kaempferol increased the number of CD11b
+, Gr-1
+ MDSCs in the adipose tissue of mice. These results suggest that kaempferol is the compound participating in the PEE-elicited MDSC induction. Kaempferol has been shown to display anti-inflammatory effects mediated by PPARγ activation [
42], while PPARγ has also been shown to participate in the induction of MDSCs in the peritoneum after marijuana cannabidiol treatment [
43]. A previous paper suggested that PPARγ is involved in the restoration of metabolic disorder by propolis or kaempferol treatment [
44‐
46]. Therefore, PPARγ may also be critical for MDSC induction in visceral adipose tissue after PEE or kaempferol treatment.
In addition to
Ly6g and
Il10, MDSCs are known to express arginase-1 (
Arg1), which attenuates nitric oxide synthesis [
47]. In our preliminary experiments, PEE increased the expression of
Arg1 in J774.1 cells, as well as in bone marrow-derived M1 macrophages. Conversely, kaempferol failed to elevate
Arg1 expression in J774.1 cells, although it did increase
Arg1 mRNA in isolated macrophages. The differing effects of PEE and kaempferol on
Arg1 mRNA might be attributed to other chemical compounds present in PEE. In support of this idea, naringenin was shown to potentiate
Arg1 expression in macrophages [
48]. Further studies are required to assess the combinatory effects of the chemical compounds in PEE on the expression of each molecule predominantly expressed in MDSCs, as well as on mice with different metabolic states.
In
in vitro culture systems, a combined treatment with IL-6 and GM-CSF induced MDSCs from bone marrow myeloid progenitor cells [
49]. Our results indicate that MDSCs can also be obtained from differentiated M1 macrophages
in vitro. M1 macrophages are major participants in the pathogenesis of adipose tissue inflammation through the secretion of proinflammatory molecules, such as TNF-α and IL-6 [
50,
51]. Thus, the PEE- and kaempferol-elicited M1 macrophage reduction potentially contributes to the repression of adipose tissue inflammation. Together with the strong immune-suppressive phenotype of MDSCs [
9], PEE and kaempferol provide two additional advantages: the decrement of proinflammatory cells and the increment of anti-inflammatory cells.
Although the intraperitoneal injection of PEE strongly induced MDSCs in the visceral tissues of mice, there are several preventive and therapeutic issues to address regarding type 2 diabetes in humans. There are some differences in the detailed characteristics of MDSCs between human and mice, such as surface markers [
9], and it is recommended that further studies evaluate if PEE induces MDSCs in humans. It is also recommended that further studies examine whether oral administration, rather than intraperitoneal injection, causes accumulation of MDSCs in adipose tissue, in consideration of the safety and accessibility of medication.
In this research, we conducted animal experiments to evaluate the
in vivo effects of PEE. The original aim of this study was to identify the immunosuppressive cellular population in adipose tissue, and
in vitro models could not replace mice. All efforts were made to reduce the number of animals used compared to previous studies [
27,
37] and we concluded that four animals per experiment appears to be sufficient to examine the effects of PEE on adipose tissue inflammation in C57BL/6 mice.