Background
According to the latest statistics, females are more likely to develop breast cancer (BC) than any other cancer. It is the most commonly diagnosed cancer in more and less developed regions, and the second most common cause of cancer death in women worldwide [
1]. BC initiation and progression are related to many factors including inflammatory factors which may be implicated in the development of therapy resistance [
2].
During tumor development, the immune system can either recognize and destroy tumors or promote their growth. This process is called immunoediting [
3]. Many studies have shown that the immune system is a major player in the cancer cell/tumor microenvironment crosstalk. Tumor-infiltrating immune cells are frequently observed and associated with cancer prognosis [
4‐
7]. Several clinical studies have evaluated the prognostic significance of tumor-infiltrating lymphocytes (TILs) and tumor-associated macrophages (TAMs) in BC [
8,
9]. Furthermore, inflammatory cytokines, such as interleukin 6 (IL-6) and tumor necrosis factor alpha (TNFα), have been shown to play important roles in the progression of BC [
10].
Allograft inflammatory factor-1 (AIF1) was first identified in rat cardiac allografts undergoing chronic rejection [
11]. In humans, the
AIF1 phylogenetically conserved gene is encoded within the major histocompatibility complex class III region on chromosome 6p21.3, which is known to harbor clusters of genes involved in inflammatory responses such as TNFα and nuclear factor-kappa B (NF-κB) [
12]. Three splice isoforms have been identified including the AIF1 splice variant 3 (AIF1v3) considered to be the “wild-type” and the largest isoform encoding a 143–amino acid hydrophilic polypeptide of 17 kDa.
The function of AIF1 is not entirely known, but it has been found to be mainly expressed by immunocytes and closely associated with inflammatory diseases [
13], obesity [
14,
15], diabetes [
16‐
18] and cancers [
19,
20]. It is a well-known central mediator of inflammation by regulating the expression of inflammatory mediators such as cytokines, chemokines and inducible nitric oxide synthase [
21,
22].
Indeed, numerous studies have demonstrated that AIF1 is involved in inflammatory responses, auto-immune diseases, reproductive immunity as well as immune activation and macrophage function [
13]. AIF1 can increase IL-6, IL-10, and IL-12 production in the RAW 264.7 macrophage cell line stimulated with lipopolysaccharides [
22]. In addition to immunomodulating functions, a recent report indicates that AIF1 may regulate several important cell adhesion molecules [
23].
Previous studies have reported an increase in
AIF1 expression in malignancies and suggest that it may have a significant role in cancer progression [
20,
24]. Furthermore, AIF1v3 may promote BC proliferation through activation of the NF-κB/cyclin D1 pathway [
25]. Additional studies have shown that AIF1v3 may promote BC cell migration via the upregulation of TNFα-mediated activation of the p38-MAPK signaling pathway [
26] and may increase the resistance of BC cells to cisplatin [
27].
However, whether other AIF1 isoforms are also involved in BC development and progression has not yet been reported. The role and expression of AIF1 isoforms in the tumor microenvironment are also not known.
The present study aims to explore potential functions of two AIF1 isoforms (AIFv1 and AIF1v3) in breast tumors of varying severity and breast adipose tissue by evaluating their expression and relationship to metabolic and clinical parameters of BC patients. To better understand the association linking inflammation, AIF1 and BC progression, the relationship between components of the tumor inflammatory cell infiltrate and AIF1 expression in breast adipose tissue was examined in histopathological breast tumor sections. The effect of omega-3 fatty acids (FA) on AIF1 isoform expression was evaluated in lymphoblastoid cell lines (LCLs) cell lines to determine their potential functions.
Although AIF1 has been studied previously in other health conditions such as inflammatory diseases, to our knowledge, this is the first evaluating the AIF1v1 isoform in the context of BC.
Discussion
AIF1 is a cytoplasmic, EF-hand calcium-binding, inflammation-responsive scaffold protein that is implicated in various disease processes. Originally identified in chronically rejecting cardiac allografts, it was initially demonstrated that AIF1 was a modulator of the immune response [
43]. Previous reports have described a range of AIF1-related splice variants, and three splice isoforms have been identified: AIF1v1, AIF1v2 and AIF1v3 (GenBank accession nos. NM_032955, NM_004847 and NM_001623, respectively) but to date, there are no studies published on AIF1v1. In the present study, we investigated, for the first time, the functional and structural differences of AIF1v1 and AIF1v3, and their expression in breast tumors and breast tumor microenvironment. We identified AIF1 isoforms in a cohort of BRCAX individuals issued from families with high BC risk. In these families,
AIF1v1 and
AIF1v3 were significantly and differentially expressed between BC affected and unaffected sisters within the same families, which led us to investigate their implication in BC development.
Cells transfected with AIF1v1 showed reduced cell viability, in agreement with a previous study with AIF1 in pancreatic cells [
17]. In addition, both
AIF1 isoforms, but mainly
AIF1v1, were highly expressed in the less severe BC tumors (DCIS and luminal subtype) suggesting their involvement in tumor initiation and progression. Two previous studies demonstrated that AIF1v3 in transfected human BC cell lines could promote BC cell proliferation via the NF-κB pathway and enhance cell migration by activation of p38-MAPK pathway suggesting a possible role in BC progression [
25,
26]. However, roles of both isoforms and interactions with other proteins in the tumor microenvironment have not been elucidated.
Our data revealed that while AIF1 isoforms are not present in epithelial cells, they are highly expressed in breast adipose tissue. Furthermore, we observed large differences in expression rates and isoform sources. While activated macrophages are the major source of
AIF1v3 in breast adipose tissue, which is consistent with previous findings in human white adipose tissue [
15],
AIF1v1 appears to be significantly less expressed by macrophages, suggesting that its high expression is due to another immune cell type. Since BRCAX LCLs displayed high
AIF1v1 mRNA expression levels, we hypothesized that lymphocytes might be a major source. The latter result is supported by a previous study where several T-cell lines were screened by real-time PCR to compare
AIF1 isoform expression levels with those of peripheral blood mononuclear cells and showed a higher expression of the
AIF1v1 isoform in all immortalized cell lines screened [
44].
Furthermore, bioinformatics analyses showed a significant structural difference between both isoforms. AIF1v1 appears to lack an entire region including specific conserved motifs and binding sites that are present in AIF1v3. This observation is of substantial relevance and could explain why the two isoforms may be differentially expressed by various cell types regulated by cytokines and growth factors of the environment and thus, behave differently. Previous studies demonstrated that
AIF1v3 is strongly expressed in macrophages and activated T-cells [
19,
44,
45], but no data exists on
AIF1v1. TILs and TAMs have essential roles in mediating tumor progression in all BC subtypes. TAMs have been shown to possess features of the pro-inflammatory M1 phenotype in the early stages of tumorigenesis, but switch to an anti-inflammatory M2-like phenotype, with the acquisition of proangiogenic capability [
46]. Furthermore, IL-6 and other cytokines secreted by M1-polarized macrophages have been shown to be involved in a wide range of tumorigenic processes [
47].
This is concordant with our results showing that
AIF1v3 is more highly expressed in M1 than M2 macrophages, its expression being the highest in DCIS and luminal, and decreasing as BC prognostic severity increases. A recent study showed that RAW264.7 cells overexpressing
AIF1v3 increased markers associated with M2 polarization and decreased those associated with M1 polarization [
48]. However, these results were carried out in a specific transfected subset of colony-stimulating factor (CSF1)-induced macrophages in the context of hepatocellular carcinoma and not BC. In addition, M2 polarization is a complex process that involves multiple factors other than CSF1, such as monocyte chemoattractant protein-1 [
49]. Nevertheless, in our case,
AIF1v3 was expressed in both M1 and M2 macrophages with significantly higher expression in the M1 phenotype.
In one retrospective study of 53 mastectomy samples, increased B-cell and T-cell immune infiltrate was identified in benign ductal hyperplasia and was increased in DCIS and highest in invasive BC [
50]. This suggests that a particular class of lymphocytes is responsible for its expression in these particular BC subtypes. Two TILs phenotypes have been described: type 1, which is assumed to have anti-tumor properties and type 2, which may promote an anti-inflammatory immune response that could enhance tumor growth [
51,
52]. The distribution of these different types of lymphocytes in each BC subtype needs to be further investigated. Furthermore, whether
AIF1v1 is only expressed by a particular class of lymphocytes requires further study.
Previous studies have established that cancer development and progression depend on complex interactions between the tumor and the local inflammatory response [
53], and a number of immune cell types implicated in this response have been described [
4,
5,
7,
54]. Our assessment of the tumor inflammatory cells infiltrate in breast tumors showed that adipose AIF1v1 was associated with the number of lymphocytes infiltrating breast tumors in both the peri and intra-tumoral regions and total number of plasma cells, which allow us to confirm our previous hypothesis. It has been reported that the prognostic significance of tumor-infiltrating T cells in breast carcinoma depends on their relative density and tissue location (peri or intra-tumoral) [
55]. Given the functional heterogeneity of TILs, the link between the tissue location of TILs infiltrate and
AIF1v1 expression would need to be further investigated.
It is well known that estrogens, which are expressed in many immune cells, modulate inflammatory cytokine gene expression [
56‐
59]. A previous study reported that E2 increased
AIF1v3 expression in a RAW264.7 murine macrophage cell line [
60]. However, our analysis showed that AIF1v3, as well as AIF1v1, do not interfere with the biological activity of estrogens in MCF7 cells thus the effect of estrogens on AIF1 is likely due, in part, to another mechanism.
As for omega-3 FA, it has been reported that they decrease cell proliferation and induce apoptotic cell death in human BC cells through the NF-κB cell pathway [
61], and it is established that AIF1 promotes BC proliferation via activation of the NF-κB/cyclin D1 pathway [
25]. This may explain our results showing the ability of DHA to modulate
AIF1v1 and
AIF1v3 expression in BRCAX LCLs in a dose-dependent manner. This is of interest because it shows for the first time that omega-3 FA, namely DHA, may potentially work as adjuvants and safe complementary therapies to standard cancer treatment [
62‐
64], and prevent tumor growth and progression by reducing
AIF1v1 and
AIF1v3 expression in BC patients, particularly those exhibiting less aggressive tumors.
Concerning clinical and metabolic phenotypes in BC patients, AIF1v1 was significantly correlated with age and menopausal status in breast tumors. Since this is the first time these relations have been observed, the influence of menopause on AIF1v1 deserves further investigation. AIF1 was also positively correlated with weight, WHR and adipose breast area in mammograms. These significant correlations with adipose
AIF1 expression are in agreement with previous findings suggesting AIF1 is an adipokine associated with clinical parameters related to obesity [
15]. The fact that
AIF1v1 expression was strongly associated with CYP19A1, leptin and ERα shows that it is involved in this pathway and plays a significant role in adipose-inflammation induced BC. Indeed, CYP19A1 provides instructions for making an enzyme called aromatase. Aromatase expression and activity in the breast adipose tissue is upregulated by leptin and inflammatory mediators and is associated with increased tissue levels of COX2 and prostaglandin E2 (PGE2) [
65]. The upregulation and associated effects can drive aberrant estrogen production within the mammary tissue, thereby promoting BC tumorigenesis. Finally, its association with inflammatory factors such as COX2, IL-6 and TNFα provides further evidence that AIF1v1 is a key regulator of inflammation in the breast tumor microenvironment and interacts with a wide variety of cytokines and adipokines.
Taken together, these results imply that AIF1v1 can potentially regulate the recruitment and activation of inflammatory cells, particularly lymphocytes, and redirect the immune response to promote the construction of a microenvironment that is more suitable for breast cancer cell progression. The underlying mechanism is yet to be elucidated. However, we can hypothesize it implies the production of TNFα by AIF1v1-activated lymphocytes, which will lead to the activation of NF-κB, thereby promoting the production of IL-6 and other cytokines, and growth-factor signals.
Authors’ contributions
FD and CD designed the research protocol. FAS and GO conceived, designed and performed the experiments. SJ contributed samples. FAS and KEI performed the analysis on inflammation. FAS, FD and CD drafted the manuscript. Critical revision was done by all authors. All authors read and approved the final manuscript.