Background

Since Yoshinori Ohsumi, the Nobelist in Physiology or Medicine, received his prize for elucidating the mechanisms of autophagy, more advances and highlights in the understanding of autophagy have been noted and shown to improve clinical outcomes in multiple areas, such as cancer [1], cardiovascular disease [2], obesity [3] and synapses [4]. Recent available data indicate that autophagy is a highly dynamic, multistep process that can be modulated at several steps, both positively and negatively. One critical point is that autophagy is commonly a conserved process in eukaryotes, involving the catabolism of multiple cytoplasmic components to maintain energy homeostasis and to protect cells against stress. Moreover, a critical process, known as selective autophagy, was more recently shown to selectively eliminate unwanted, potentially harmful cytosolic material, for example, damaged mitochondria or protein aggregates, thereby acting as a major cytoprotective system [5].

Recently, studies found that the expression of autophagy-related proteins (ATGs) at CAFs or cancer cells, such as microtubule-associated protein light chain 3 (MAP1LC3/LC3), Beclin-1 (BECN1) and sequestosome 1 (SQSTM1 /p62) et al., might be promising indicators of for tumor recurrence and prognosis [6, 7]. During autophagy, cells recycle whole organelles and macro-molecules by three stages: 1) forming a phagophore by the initiation complex, including Unc-51 like autophagy activating kinase 1 (ULK1) complex and phosphatidylinositol 3-kinase catalytic subunit type 3 (PIK3C3)/vacuolar protein sorting protein 34 (Vps34) complex; 2) forming a compartment called autophagosome by two ubiquitin-like conjugating systems, Atg12-Atg5 and LC3 complex; 3) the maturation of autolysosome through the fusion between autophagosome and lysosomes (Fig. 1) [8, 9]. Phosphorylation of BECN1 and Vps34 triggers the activation of the PIK3C3 complex, resulting in local phosphatidylinositol-3-phosphate (PI3P) production [10]. Expansion of nascent precursor vesicles relies on the autophagosome protein LC3. Critical for this process is the phosphatidylethanolamine (PE) conjugated LC3-I and form of LC3-II [11]. In autophagy, p62 stably binds to the LC3-II protein acting a role as proteins trafficking in the context of assembling autophagosomes [12].

Fig. 1
figure 1

Autophagy-related proteins in the process of cell autophagy. The initial steps in autophagy include the nucleation, elongation, and maturation of an isolated membrane, usually called a phagophore. The formed phagophore then unites to form the autophagosome, and the fusion with a lysosome follows to form an autolysosome, where the captured materials and eliminated. Molecules which generally act as the markers of autophagy in the researches of CAFs and tumor, including LC3, ATGs, BECN1 and p62 were participated in the process of forming phagophore, autophagosome and autolysosome

Full size image

The tumor microenvironment has recently gained much attention as a critical determinant of tumor heterogeneity, initiation, progression, metastasis, and resistance to systemic therapies. The tumor microenvironment consists of immune cells (lymphocytes, natural killer cells, and antigen-presenting cells), stromal cells (including myofibroblasts), vasculature endothelial cells and adipocytes. In particular, cancer-associated fibroblasts (CAFs) are myofibroblast-like cells that induce the formation of a desmoplastic “reactive stroma”, compared with normal fibroblasts (FIBs), and promote tumor growth and aggressiveness [13]. The origin of CAFs is yet not well defined, but it is suggested that they arise from progenitors, in general mesenchymal stem cells (MSCs) or from other differentiated cells, such as smooth muscle cells (SMCs), cells of epithelial origin, endothelial cells, perivascular cells, and adipose tissue-derived stem cells [14]. Based on experimental evidence, CAFs markers, either CAF specific or CAF derived, have demonstrated an independent association with survival. This includes members of the platelet-derived growth factor receptor (PDGFR) family, transforming growth factor beta 1 (TGFB1) signaling, CAF-markers, such as podoplanin and fibroblast activation protein (FAP), as well as transcription factors (FoxF1) and secreted factors (matrix metalloproteinases (MMPs)) [15,16,17]. To identify specific markers to CAF subsets, Shicheng S et al. recently found two cell-surface molecules, CD10 and GPR77, which define a specific CAF subset that sustains cancer stemness and promotes tumor formation and chemoresistance [22]. Another study illustrated that the CAFs in autophagy, stimulated by tumor cells, cause alanine secretion, which actually outcompetes glucose- and glutamine-derived carbon, in turn providing fuel for the tricarboxylic acid (TCA) cycle and thus for nonessential amino acids and lipid biosynthesis in its low-glucose microenvironment [23]. Obviously, this description gives a selective view and may exclude unknown factors and mechanisms.

The role of CAFs autophagy

The role of the autophagy in CAF biology is complex, and it is shown to play critical roles that differ depending on the chemical treatment and biological context [24].The tumor mass reaction, including fibrotic stromal, also impairs the vasculature, leading to a highly hypoxic environment, due to the inhibition of the Notch/Hedgehog pathway, and a nutrient-poor environment, which suggests a link to the resulting “angiogenic switch” independent hypovascularity and perfusion impairment for tumor progression [25]. CAFs of tumor microenvironment positively influenced the proliferation and metabolism of cancer cells, through oxidative stress induced autophagy pathway which were initially induced by neighboring tumor cells [26]. Secretory autophagy is involved in the export of a variety of cellular cargoes. This includes leader less cytosolic proteins and inflammatory mediators, such as interleukin 1β (IL-1β), IL-6, IL-8 and IL-18 [27]. New J et al. showed that mitigating autophagy significantly reduced CAF-induced progression through IL-6, IL-8 and bFGF in neck squamous cell carcinoma. Treatment with autophagy target Vps34 inhibitor, SAR405, attenuated xenograft growth and inhibited the effects of standard therapy [28]. Similar results were found both in vitro (co-cultured model) or in vivo (xenografted model and clinical tissue) in cancer researches such as breast cancer, ovarian cancer, liver cancer, colorectal cancer and pancreatic adenocarcinoma [29, 30]. Additionally, there is extensive evidence in literature demonstrating that both radiation and chemotherapeutic drugs promote cytoprotective autophagy in tumor cells. Stress-related inducers triggered CAFs autophagy participate actively in tumor growth, invasiveness, and resistance to chemotherapy [31,32,33].

The mechanism of CAFs autophagy

Hypoxic-autophagy positive feedback cycle in CAFs

Hypoxia is not only a prominent stressor in the microenvironment but also a crucial contributor to the heterogeneity of tumors to drive adaptations to support tumor growth and resistance to systemic therapies [34]. To understand the contribution of the microenvironment in promoting tumor growth and metabolic mechanisms, two Nobel Laureates formulated a hypothesis to explain the “fundamental basis” of cancer. In the “Warburg Effect” hypothesis, cancer cells produce energy via the conversion of glucose into lactate, despite the presence of oxygen, a process known as aerobic glycolysis. The “Pasteur effect” further described the inhibiting effect of glycolysis upon oxygen, confirming aerobic glycolysis as a hallmark of the cancer phenotype. This glycolytic pathway is accentuated under hypoxia, which acts pleiotropically to upregulate glucose transporters and multiple enzymes by independently increasing the levels of the HIF1a and HIF2a transcription factors [35]. Lisanti MP et al. proposed the “tumor-stroma co-evolution” model, indicating that breast cancer cells induced oxidative stress in adjacent stromal fibroblasts and upregulated autophagy and mitophagy in the tumor microenvironment [36]. This reliance on the autophagy and oxidative stress pathways demonstrated in CAFs is consistent within many tumors.

By the overwhelming intercellular change of oxidative stress, such a positive-feedback cycle turns a condition of autophagy in CAFs or cancer cells (Fig. 3). Data on the stroma-tumor crosstalk indicated that caveolin-1 (CAV1) and breast cancer type 1 susceptibility protein (BRCA1) involved in oxidative stress pathway in CAFs. The membrane protein CAV1 possesses tumor-suppressor properties within stromal cells, whereas downregulation of stromal CAV1 correlated with cancer progression, invasion and metastasis and thus, a worse clinical outcome [37]. A coculture system, by which a loss of stromal fibroblast CAV1 induces a “lethal tumor microenvironment,” demonstrated that MCF7 epithelial cancer cells induce oxidative stress in adjacent CAFs, resulting in the autophagic/lysosomal degradation of stromal CAV1 [38, 39]. Tumor cell-derived reactive oxygen species (ROS) decrease the expression of CAV1 in CAFs [40]. Consistent with the metabolic findings, certain miRs are found to be associated with oxidative stress (miR-34c) or activation of the hypoxic response/HIF1a (miR-31), which is sufficient to drive autophagy/mitophagy. Upregulating oxidative stress in CAFs is sufficient to induce genomic instability in adjacent cancer cells, via a bystander effect, potentially increasing their aggressive behavior [41]. Studies demonstrated that BRCA1 is mutated in 45% of hereditary breast cancers, which has been described recently as an autophagy inhibitor [42, 43] Salem AF et al. reported that BRCA1 induces several antioxidant genes that are responsible for ROS inhibition, and knockdown of BRCA1 in CAFs is able to significantly promote tumor growth [44]. Preclinical studies using xenografts demonstrate that shBRCA1 fibroblasts, with an increase in markers of autophagy and mitophagy, induced an ~ 2.2-fold increase in tumor growth when coinjected with MDA-MB-231 cells into nude mice via ketone production [45].

Fig. 3
figure 3

Hypoxic-autophagy positive feedback cycle in CAFs. In a coculture system of different cancers, tumor cells induce oxidative stress in adjacent stromal CAFs and upregulate autophagy and mitophagy in the tumor microenvironment

Full size image

Additionally, oxidative metabolism of CAFs provide nutrients (such as ketone) and cytokines to stimulate mitochondrial biogenesis and autophagy, by a reverse Warburg effect, in adjacent cancer cells. Cancer cells escape oxidative mitochondrial damage and apoptosis by the upregulation of antioxidant enzymes, such as peroxiredoxin-1 [36]. In a noncontact coculture system of human colorectal/ ovarian fibroblasts and cancer cells, CAFs positively influence the metabolism of cancer cells, through the autophagy and oxidative stress pathways, which are initially induced by neighboring tumor cells [26, 46]. Further research indicates that CAFs promote irradiated cancer cell recovery and tumor regrowth postradiation. In in vitro and in vivo xenograft models of lung cancer and melanoma cells, CAFs produce IGF1/2, CXCL12 and β-hydroxybutyrate and increase the level of ROS postradiation, which enhances protein phosphatase 2A (PP2A) activity, resulting in repressing mTOR activation and increasing autophagy in cancer cells postradiation. A further point mutation result suggests that the oxidation of PP2Ac Cys251 could enhance PP2A activity, validating the IGF2 function through above-mentioned mechanism [32]. This postradiation result indicates that CAFs play key roles in irradiating cancer cell recovery, which is in accordance with the previous observations that preexisting CAFs enhance the radiation resistance of tumor cells [47]. Although the role of hypoxic stress in the crosstalk among CAFs and tumor cells is not fully elucidated, it is widely appreciated that the hypoxic zone in CAFs induces autophagy in themselves or by a paracrine pathway of secreting cytokines in tumor cells.

Metabolic cross-talk between CAFs and the tumor induced by autophagy

With the novel concepts of the “reverse Warburg effect” and the “autophagic tumor stroma model of cancer metabolism” that Michael P et al. proposed, an in vitro study demonstrates that the enhanced aerobic glycolysis and/or autophagy in the CAFs supports epithelial cancer cell growth and aggressive behavior via the secretion of high-energy metabolites by the tumor stroma [38, 48] (Fig. 4). These nutrients include ketones and lactate, as well as chemical building blocks such as amino acids (glutamine) and nucleotides. Lactate and ketones serve as fuel for cancer cell oxidative metabolism, and building blocks sustain the anabolic needs of the rapidly proliferating cancer cells. A further in vivo study of a xenograft model shows that the recombinant overexpression of pyruvate kinase M (PKM1 and PKM2), a key enzyme in the glycolytic pathway, is sufficient to promote the growth of breast cancer cells in human fibroblasts, increasing tumor mass and tumor volume, without an increase in tumor angiogenesis [49]. The expression of PKM1 enhances the glycolytic power of stromal cells, with an increased output of lactate, and induces tumor inflammation. PKM2 increases the output of the ketone body 3-hydroxybutyrate, triggering a “pseudo-starvation” response and the induction of an NFκB-dependent autophagic program in stromal cells. A similar result was found in Paola Avena’s study. CAFs with activated peroxisome proliferator-activated receptor γ (PPARγ), display metabolic features with increased autophagy, glycolysis and senescence [50]. Overexpressing PPARγ in the tumor stroma reveals a 70% increase in L-lactate accumulation, relative to control fibroblasts. The controversial role of PPARγ, showing either an autophagy-induced protumorigenic effect in the CAFs or antineoplastic effects in epithelial cancer cells, suggests that the activation of an autophagic program has both pro- or antitumorigenic effects depending on the cell compartment in which it occurs [51, 52]. Recently, a small signaling phospholipid suffices to activate mTORC1 and suppress autophagy [53], Lysophosphatidic acid (LPA), was identified a role for LPA-HIF1α signaling-hub in the maintenance of the glycolytic-phenotype in CAFs [54]. The signaling locus for CAF-phenotype targeted inhibition of LPA-mediated metabolic reprogramming in CAFs may represent an adjuvant therapy in ovarian cancer. Pancreatic stellate cells (PSCs) are the precursors of CAFs, which potentiate pancreatic tumor growth and progression [55]. A previous study illustrates that an intratumoral metabolic cross-talk occurs between different populations of cells in a tumor [56]. Recently, the alterations in lactate and alanine were reported using imaging studies during pancreatic cancer progression in mouse models [57]. Consistent with the previous studies, Cristovão M et al. showed that PSCs are critical for pancreatic ductal adenocarcinoma (PDAC) metabolism through the secretion of nonessential amino acids (NEAA). Autophagic alanine secretion outcompetes glucose and glutamine-derived carbon in PDAC to fuel the tricarboxylic acid (TCA) cycle, and thus, NEAA and lipid biosynthesis, decreasing the tumor’s dependence on glucose and serum-derived nutrients in microenvironment. Within this shift in fuel source, the alanine secretion by PSCs is dependent on PSC autophagy, a process that is stimulated by cancer cells [23]. This finding indicates a novel metabolic interaction of the tumor stroma, in which alanine acts as an alternative carbon source, an effect that is not even recapitulated with exogenous lactate.

Fig. 4
figure 4

Metabolic cross-talk between CAFs and tumors induced by autophagy. Via the secretion of high-energy metabolites by the tumor stroma, such as lactate, ketone and alanine, aerobic glycolysis and/or autophagy is enhanced in the CAFs to support cancer cell growth and an aggressive behavior

Full size image

In addition, the metabolic alterations in lung cancer-associated fibroblasts were determined by mass spectrometry-based profiling of the abundances of 203 biochemicals of 46 metabolic pathways/groups to compare primary human lung tumor CAFs to “normal” fibroblasts (NFs) [58]. Although the results showed no differences in the individual metabolites distinguishing CAFs from NFs, significant differences were found between CAFs and NFs in the steady-state abundances of the metabolites of select metabolic pathways. The distinct roles of CAFs are related to the tumor’s glycolytic capacity, which is reflected by the metabolic differences between the CAFs from high and low glycolytic tumors. Dipeptide levels are significantly increased in CAFs, which is a general characteristic of CAFs that is reported to be correlated with the glycolytic activity of the tumor [59]. Most interestingly, an increase in basal macroautophagy was found, which might account for the increase in dipeptide levels. The difference between CAFs and NFs is demonstrated in the induction of autophagy promoted by reduced glucose, which, taken together, suggests that the increased autophagy may account for metabolic differences between CAFs and NFs.

CAFs secreted cytokines promote cancer survival by secretory autophagy

The tumor microenvironment is a specialized niche that impacts malignant cells directly and indirectly through stromal cells that support tumor growth [60]. Several cytokines were identified, including IL-6, IL-8, IGF1, IGF2, and CXCL12, all of which promoted survival of cancer cells [30, 61]. Autophagy is closely intertwined with inflammatory and immune responses. Proinflammatory cytokines such as IFN-γ, TNF-α, IL-17, and cytokines of the IL-1 family, regulate or be regulated through autophagy mediated this interaction [62, 63]. A concept discovered less than half a decade ago, secretory autophagy, refer to CAFs secrete soluble factors through autophagy, has a multifaceted impact on the cancer microenvironment [64]. The mechanisms of CAF-tumor cell interaction have been showed including paracrine signaling and exosomal transfer mediated by cytokines such as IL-6 and GM-CSF [65]. Lu H et al. reported that IL-6 could induce autophagy by expressing NS5ATP9, while NS5ATP9 upregulated IL-6 levels in turn, which further induced autophagy [66]. TGFβ1 small latent complex could select by golgi reassembly stacking protein 2 and secret via MAP1LC3/LC3-positive secretory autophagosomes through an unconventional pathway in fibroblasts and macrophages [67].

Recently, Thongchot S et al. firstly demonstrated that CAFs secretory products directly affect the regulation of autophagy and consequently the behavior of cholangiocarcinoma cells. Specifically, resveratrol has the potential to abrogate the effects of IL-6 mediated motility by CAFs and reverted the N-to E-cadherin switch in migrating cells [68]. Ferraresi A et al. showed that polyphenol resveratrol can oppose the stimulatory effect of IL-6 on cell migration through epigenetic up-regulation of autophagy of ovarian cancer cells [69]. Proteomic studies demonstrated that bortezomib could triggers CAFs to produce high levels of IL-6, IL-8, IGF-1, and TGFβ, then activate oxidative stress and pro-survival autophagy in multiple myeloma [70]. Li WL et al. found that IL-13 regulates BECN1 and LC3B expression through IKKβ/NFκBp65 in fbroblasts cocultured with breast cancer cells [111]. Importantly, investigators need to determine whether they are evaluating levels of early or late autophagic compartments, or autophagic flux, which is a form involved in CAF intervention. Many cancer types have a high stromal content, such as pancreatic cancer, non-small cell lung cancer (NSCLC), TNBC and sarcoma et al., contributing to low response rates to current therapies and a poor long-term survival [112, 113]. Emerging evidence suggests that the stromal compartment impedes the effective uptake of chemotherapeutics such as Letrozole and polygonatum, sha** the antitumor immunity and responsiveness to immunotherapy by autophagy [114,115,116,117,118]. All of these results indicate a disruption in the CAFs to improve drug efficiency, and this is a potential strategy that must be further pursued in the future both preclinically and in early clinical trials in stroma-rich tumors. Further original research and clinical trials are needed to make a significant impact for patients, with potential therapeutic strategies using a CAF-focused anticancer intervention. In development or currently underway, we believe that CAFs can help to answer the question of tumor heterogeneity and the inconsistent results of in vivo/in vitro models. Thus, CAFs could be a new model to minimize the gap between experiment and reality.