Lipolysis by pancreatic cancer‐derived extracellular vesicles in cancer‐associated cachexia via specific integrins

Abstract

Here, we report that pancreatic cancer-derived extracellular vesicles (EVs) carry adipocyte-targeting integrins and induce lipolysis, constituting the underlying mechanism of cancer-associated cachexia (CAC). CAC is a life-threatening condition recognised as a paraneoplastic syndrome with body weight loss, skeletal mass wasting and adipose tissue atrophy.1 While CAC occurs in most patients with cancer,1 it typically manifests earlier in pancreatic cancer.2 Fat loss is a feature of CAC, where lipolysis is activated in adipocytes, reducing their size.1, 3 Because CAC phenotypes occur systemically, humoral factors are possibly pathogenic. However, the mechanisms underlying lipolysis remain unclear. EVs, nanoparticles released from cells into bloodstream, contain various bioactive factors that mediate intercellular communication.4 Cancer cells actively release EVs5; we also hypothesised that EVs from pancreatic cancer cells contribute to lipolysis. Mature adipocytes were prepared from human adipose-derived mesenchymal stem cells (Figure S1) and treated with EVs. A higher level of glycerol, a marker for lipolysis, was released from adipocytes treated with EVs from Panc-1 and Miapaca-2 cells, but not from Capan-2 and Human Pancreatic Nestin-Expressing cells (HPNE cells) (Figures 1 and S2). While cyclic adenosine monophosphate (cAMP) levels, protein kinase A (PKA) activity and the subsequent phosphorylation of hormone-sensitive lipase (HSL) in adipocytes play key roles during lipolysis,6 these levels were upregulated in lipolysis-induced adipocytes (Figure 1B,C). A reduction in lipolysis by H89, orlistat and cay10499 suggests that lipolysis, in EV-treated adipocytes, was induced by the cAMP–PKA–HSL pathway (Figure 1D). However, cAMP levels inside EVs were similar irrespective of cell types (Figure 1E), suggesting that differential cAMP levels were not responsible for distinct lipolysis levels. Next, adipocytes were treated with EVs from pooled sera of healthy controls or pancreatic cancer patients. Compared with the controls, EVs from patients induced greater lipolysis (Figure 1F), and HSL phosphorylation and intracellular cAMP levels were remarkably higher (Figure 1G,H). However, the cAMP levels in EVs from controls and patients did not differ (Figure 1I). The tendency of lipolysis levels was similar even after adjustment for EV number and protein weight (Figure 1J,K). Furthermore, pancreatic cancer EVs (Table S1) induced more lipolysis (Figure 1L). However, cAMP levels did not differ between EVs from controls and patients (Figure 1M). Additionally, cAMP levels in EVs were not correlated with lipolysis levels (Figure 1N), suggesting that pancreatic cancer EVs induce greater lipolysis, possibly through a mechanism that depends on other factors. To confirm these results, mice were intravenously injected with EVs (Figure 2A). Panc-1-derived EV-treated mice exhibited 63.6% less weight gain (Figure 2B). Gonadal white adipose tissue (gWAT) was smaller in weight and size (Figure 2C,D); in addition, smaller lipid droplets were observed (Figure 2E). Consistently, EV-treated mice exhibited increased HSL phosphorylation in the gWAT (Figure 2F). These results indicate that cancer-derived EVs induce lipolysis via an HSL-mediated pathway. cAMP levels in EVs were not proportional to lipolysis levels (Figure 1E,I,N). Therefore, we examined whether newly synthesised cAMP was involved in EV-induced lipolysis. β-Stimulators, including isoproterenol, and activate adenylyl cyclases, led to cAMP synthesis. Thus, isoproterenol induced lipolysis, which was inhibited by SQ22536, an adenylyl cyclase inhibitor. However, SQ22536 did not inhibit lipolysis induced by Panc-1 EVs (Figure S3A,B). After 4 h, more Panc-1-derived EVs were taken up by EV-treated adipocytes compared with Capan-2 EVs (Figure S3C). Thus, EV tropism and cAMP uptake levels were linked to differential lipolysis induction. To track EV tropism, we established reporter cells by HiBiT peptide sequence knock-in at the ACTB locus in Panc-1 cells (Panc-1-ACTB-HiBiT cells) because most EVs carry β-actin7 (Figure S4). EVs were taken up by adipocytes and lung tissue, not muscle or liver tissues (Figure 3A). As integrin patterns may be associated with EV tropism,8 integrin expression profiles in Panc-1 cell-derived EVs were evaluated (Figure 3B). Integrins α6 (ITGA6), β1 (ITGB1) and αV (ITGAV) are dominantly expressed in Panc-1- and Miapaca-2-derived EVs. Remarkably, ITGB1 was not expressed in EVs derived from Capan-2, which did not induce lipolysis (Figure 3C). Thus, we hypothesised that ITGB1-related integrin heterodimers play a key role in EV-induced lipolysis. ITGB1 pairs with ITGA6 (integrin α6β1) to generate classical ‘laminin-specific’ integrins, laminin α5 (LAMA5) and α4 (LAMA4), which are expressed in adipose and lung tissues, not skeletal muscle or liver tissues in mice (Figure 3D,E). Similarly, laminin subfamilies were highly expressed in human adipose and lung tissues (Figure 3F). Thus, interactions between integrins and laminin may determine the tropism of pancreatic cancer-derived EVs to adipocytes. We isolated ITGB1-knockout, ITGA6-knockout and ITGAV-knockout EVs to evaluate the roles of ITGB1 and ITGA6 in lipolysis, through establishing Panc-1 and Miapaca-2 cells lacking these molecules or an unrelated ITGAV as a control (Figures 3G and S5A). After confirming that laminin was expressed in human adipocytes (Figure 3H), we determined the level of interaction between adipocytes and EVs. ITGB1-knockout and ITGA6-knockout EVs demonstrated reduced interactions with adipocytes (Figures 3I and S5B); thus, ITGB1 and ITGA6 are crucial for adipocyte–EV interactions. Many EVs were present on the adipocyte surface after treatment (Figure 3J). Several wild-type EVs derived from Panc-1 cells are ingested by adipocytes. Double knockout of ITGA6 and ITGB1 on EVs or LAMA4-knockout adipocytes showed reduced interactions between EVs and adipocytes (Figure 3K–N) and lipolysis was suppressed proportionally (Figures 3O,P and S5C), suggesting that ITGB1 and ITGA6 on pancreatic cancer-derived EVs determine EV tropism to adipocytes. As reported,9 the number of EVs was higher in patients’ sera, and their size was smaller (Figure 4A,B), confirming TSG101 and CD63 expression (Figure 4C). Cancer-derived CA19-9-positive EVs or CD63-positive were specifically concentrated (Figure 4D–F). ITGB1 and ITGA6 expression levels in bulk sera or CD63-positive EVs were not significantly different between controls and patients. However, positivity rates for ITGB1 and ITGA6, but not ITGAV, were significantly higher in concentrated cancer-derived EVs (Figure 4G). In conclusion, we propose the importance of ITGB1 and ITGA6 expression in cancer-derived EVs for lipolysis induction during CAC (Figure 4H) in addition to the importance of examining tissue-specific EVs in a heterogeneous EV population in sera.10 This work was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan (#20J20625, #22H02828 and #22K15390 to Chikako Shibata, Motoyuki Otsuka and Takahiro Seimiya), JST CREST (#JPMJCR19H5 to Motoyuki Otsuka) and the Japan Agency for Medical Research and Development, AMED (JP 22fk0210092 and JP22ck0106557 to Motoyuki Otsuka). The authors declare they have no conflicts of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts