The senescence-associated secretory phenotype in ovarian cancer dissemination

Abstract

Ovarian cancer is a highly aggressive disease with poor survival rates in part due to diagnosis after dissemination throughout the peritoneal cavity. It is well-known that inflammatory signals affect ovarian cancer dissemination. Inflammation is a hallmark of cellular senescence, a stable cell cycle arrest induced by a variety of stimuli including many of the therapies used to treat patients with ovarian cancer. Indeed, recent work has illustrated that ovarian cancer cells in vitro, mouse models, and patient tumors undergo senescence in response to platinum-based or poly(ADP-ribose) polymerase (PARP) inhibitor therapies, standard-of-care therapies for ovarian cancer. This inflammatory response, termed the senescence-associated secretory phenotype (SASP), is highly dynamic and has pleiotropic roles that can be both beneficial and detrimental in cell-intrinsic and cell-extrinsic ways. Recent data on other cancer types suggest that the SASP promotes metastasis. Here, we outline what is known about the SASP in ovarian cancer and discuss both how the SASP may promote ovarian cancer dissemination and strategies to mitigate the effects of the SASP.

INTRODUCTION

Ovarian cancer is the deadliest gynecological cancer for women worldwide (1). Approximately 19,880 new ovarian cancer diagnoses and 12,810 deaths are estimated to occur in 2022 in the United States alone (2). Ovarian cancer incidence and mortality have decreased by ∼30% since the 1970s due to improved understanding of the disease and new therapies (1). Due to a lack of screening methods and nonspecific symptoms, over 70% of ovarian cancers are diagnosed at stage III/IV when the tumors have already disseminated throughout the peritoneal cavity, and these patients have a 5-year survival rate of less than 30%. The low survival rates of metastatic ovarian cancer highlight the importance of understanding underlying mechanisms of dissemination.

Ovarian cancer therapies such as platinum-based agents and poly(ADP) ribose polymerase (PARP) inhibitors induce cellular senescence (3, 4), a stable cell cycle arrest (5). Although initially thought to be tumor suppressive, recent work has demonstrated that senescence is more context-dependent and may be protumorigenic in part through the paracrine effects of the senescence-associated secretory phenotype (SASP) (6). The SASP is composed of proinflammatory cytokines and chemokines, proteases, growth factors, bioactive lipids, and extracellular vesicles that contain a variety of cargo. Acutely, the SASP may be beneficial by promoting immunosurveillance, clearance of senescent cells, and immunotherapy sensitization (4, 7, 8). However, the long-term chronic effects of senescence and the SASP may be detrimental. For instance, the SASP can promote metastasis and invasion, chemoresistance and relapse, epithelial-to-mesenchymal transition, angiogenesis, and stemness, which are all directly linked to cancer initiation, progress, and aggressiveness (6). In addition, age correlates with ovarian cancer dissemination (9), and senescent cells accumulate in aged tissues (10). However, in the context of ovarian cancer, little is known about the mechanisms underlying how the SASP promotes dissemination. Here, we will discuss SASP regulators and effectors that have been implicated in ovarian cancer dissemination and further speculate how modulation of the SASP and SASP factors may have implications in ovarian cancer therapy.

OVARIAN CANCER DISSEMINATION

Ovarian cancer is distinct from other cancers in that its primary mechanism of metastasis is through transcoelomic metastasis, where cells are shed from the primary tumor site into the peritoneal cavity, rather than by the hematogenous route (11, 12). In transcoelomic metastasis of ovarian cancer, which we will simply call dissemination, cells overcome anoikis and undergo an epithelial-to-mesenchymal transition (EMT) to release from the primary ovarian tumor. Metastatic ascites, a fluid that accumulates in the peritoneal cavity, aids in the dissemination of ovarian cancer cells either as single cells or spheroids, which have cancer stem cell-like properties. These cells are highly chemoresistant and able to evade immune detection. Ultimately, cells seed on organs within the peritoneum, with the omentum being the most common site of dissemination. Dissemination is a complex process that involves dynamic changes in cell-intrinsic and cell-extrinsic signaling, suggesting that the microenvironment is critical toward this process. One underappreciated aspect of ovarian cancer dissemination is the contribution of the senescence-associated secretory phenotype (SASP). Herein, we will discuss the SASP, what is known about its regulation, how it may promote ovarian cancer dissemination, and potential therapeutic strategies to target this process.

SENESCENCE-ASSOCIATED SECRETORY PHENOTYPE

The SASP is a complex secretome of senescent cells that presumably evolved to trigger tissue remodeling during development or in response to DNA damage to promote clearance of these cells (13). It is a highly dynamic process whose composition and function vary in a context-dependent manner based on factors such as senescence inducer, cell type, stage of senescence, and tissue context. The SASP is typically composed of proinflammatory cytokines and chemokines, proteases, growth factors, bioactive lipids, and extracellular vesicles that can contain a variety of factors (6). Studies in other cancers have found that the SASP can promote metastasis (6). While there is currently little knowledge on how the SASP promotes dissemination in ovarian cancer, individual SASP-like components have been shown to directly influence tumor dissemination by multiple mechanisms such as resistance to anoikis, promotion of cancer stem cell phenotypes, chemotaxis and invasion, and extracellular matrix (ECM) remodeling (11). We will discuss each in more detail in the following section.

The SASP is regulated at multiple levels. The most well-described upstream regulator of the SASP is cytoplasmic double-stranded DNA [dsDNA; also termed cytoplasmic chromatin fragments (CCFs)] (14). Cytoplasmic dsDNA is sensed by cyclic GMP-AMP synthase (cGAS), which activates the stimulator of interferon genes (STING) pathway. Downstream of cytoplasmic dsDNA, regulation of the SASP occurs at the level of transcription, translation, and secretion. cGAS-STING activates NF-κB and C/EBPβ, two of the main transcription factors of SASP genes (6). In addition, high mobility group box (HMGB) proteins such as HMGB1 and HMGB2 can regulate the expression of SASP genes by modifying chromatin architecture (15). Other important transcriptional regulators of the SASP are p53, p16, and RB (16–18). Furthermore, epigenetic regulators such as enhancers and multiple active histone marks have been shown to transcriptionally upregulate SASP factors (15). At the translational level, mTOR translationally controls the SASP in multiple ways, and inhibition of mTOR blunts the SASP (19). Here we will highlight the relevance of these regulators in ovarian cancer dissemination and discuss targeting these pathways.

OVARIAN CANCER DISSEMINATION IS AFFECTED BY SASP-LIKE FACTORS

Senescence in Ovarian Cancer

Platinum-based therapies have been the standard-of-care for patients with ovarian cancer for decades (20); the emerging therapies for patients with ovarian cancer are poly(ADP-ribose) polymerase (PARP) inhibitors (21). Recent reports in both ovarian cancer cell lines and patient tumors demonstrate that senescence and upregulation of the SASP occur in response to these therapies (3, 4, 22, 23). In addition, senescent cells accumulate during aging (10), and aging is an independent prognostic factor for ovarian cancer (24). Together, these data suggest that senescence induction in ovarian tumors is likely due to two sources: therapies used to treat ovarian cancer and aging.

Interestingly, senescence has been shown to be both beneficial and detrimental in ovarian cancer. Induction of senescence and presence of the SASP increases ovarian cancer survival (23) and correlates with immune cell infiltration and response to immunotherapy (4, 7). On the other hand, higher senescent cell burden corresponds to increased malignant ascites of patients with ovarian cancer (25), and age accelerates tumorigenesis and dissemination in mouse models of ovarian cancer (26), although this study did not directly assess senescence. Together, these data suggest that senescence may play an important role in ovarian cancer initiation, progression, and response therapy. However, experimental evidence is just beginning to emerge about how the SASP promotes dissemination and other phenotypes in ovarian cancer. Many papers on other cancer types have demonstrated that aging, cellular senescence, and systemic SASP may contribute to metastasis and other protumorigenic phenotypes (6), which are relevant to ovarian cancer. Here, we will discuss SASP regulators and effectors that have been implicated in ovarian cancer dissemination (Fig. 1 and Table 1). One caveat to note is that while it is possible that SASP regulators and effectors are important for ovarian cancer, they are not necessarily related to the SASP or senescence. This will need to be empirically explored. We will also further speculate how modulation of the SASP and SASP factors may have implications in ovarian cancer therapy.

Figure 1.

Ovarian cancer-relevant SASP regulators. Cytoplasmic dsDNA or cytoplasmic chromatic fragments (CCFs) activate the cGAS-STING pathway, which in turn promotes SASP gene transcription via two main transcription factors NF-κB and C/EBP-β. The SASP is also transcriptionally regulated by p53, p16/RB, DOT1L, HMGB2, METTL3/14 and others. mTORC1 promotes translation of SASP mRNAs. SASP, senescence-associated secretory phenotype.

Table 1.

Known roles of SASP regulators in ovarian cancer dissemination and therapeutic targeting

SASP Regulator	Role in Promoting SASP	Potential Role in Ovarian Cancer Dissemination	Therapeutic Target in Ovarian Cancer?	Therapeutic Agents (*in Clinical Trials; #in Ovarian Cancer)
cGAS-STING	Sensor of cytoplasmic dsDNA and promotion of downstream transcription via NF-κB (14)	Decreased ascites (27)	Upregulation increases response to immunotherapy (7, 28)	Antagonists:
				C-170
				C-171
				C-176
				H-151
				SN-011
				Agonists:
				GSK3745417*
				E7766*
				MIW815*
				MSA-2
				SNX281#
				TAK-500*
				TAK-676*
p53	Transcriptional repressor (17)	Mutant p53 increases cytokine expression (29) and has more dissemination potential (30)
p16/RB1	Loss leads to transcriptional repression (16, 18)	Unknown, but patients with RB low ovarian cancer are extremely long-lived (31)
NF-κB	Transcriptional activator (6)	Multiple roles in EMT, stemness, angiogenesis (32)	Inhibition affects multiple pathways (32)	Many agents [reviewed in (33)]
C/EBPβ	Transcriptional activator (6)	Linked to a transcriptional hub including inflammatory genes that are a part of the SASP (34) and promotes migration through DAXX (35)		Helenalin acetate
DOT1L	Transcriptional upregulation through H3K79 methylation (36)	Downstream of C/EBP-β (40)		EPZ-5676*
				SCG 0946
				SYC-522
Brd4	Transcriptional upregulation through enhancers (8)	Promotes EMT (37)	Inhibition decreases dissemination (38) and affect multiple other pathways (37)	ABBV-075
				AZD5153#
				BMS-986158*
				BMS-986378*
				Bromosporine
				CPI-203
				GSK525762*
				I-BET 151
				I-BET 762
				JQ1
				LY 303511
				MS 436
				OTX 015
				OXF BD 02
				PFI 1
				PLX51107*
				SYHA1801*
				XMD 8-92
				ZEN-3694#
HMGB2	Transcriptional upregulation through insulation from heterochromatin marks (39)	Promotes dissemination downstream of CENPU (40)		Inflachromene
METLL3	Enhancer-promoter loop formation (15)	Promotes invasion and migration through inhibition of CCNG2 (41)		Cpd-564
				STM2457
				UZH1a
mTORC1	Translational regulation (19)	Decreased mTORC1 activity decreases migration, invasion, dissemination (42–45)	Inhibition decreases dissemination (42–45)	ABI-009#
				ATG-008#
				AZD2014#
				CC-115*
				CC-223*
				DS-3078a*
				Everolimus*
				HEC68498*
				MLN0128#
				PF-05212384#
				PQR309*
				Rapamycin
				SAR245409#
				SF1126*
				TAK-228#
				Temsirolimus
				Torin 1
				WXFL10030390*

Clinical trial information was obtained from clinicaltrials.gov. SASP, senescence-associated secretory phenotype.

SASP Regulators in Ovarian Cancer

cGAS-STING.

The cGAS-STING pathway is activated by cytoplasmic dsDNA that is enriched in senescent cells due to induction of DNA damage (14). While this is still an emerging field in ovarian cancer, studies have found that the cGAS-STING pathway is activated by therapies that induce senescence. A recent publication in a mouse model of HGSOC found that cGAS-STING signaling is activated by cisplatin-induced senescence, which correlated with increased NK and T cell infiltration and response to immune checkpoint blockade specifically in Brca1 null tumors (4). Excitingly, knockdown of cGAS in this model suppressed the SASP and combinatory therapy response, although disseminated tumor burden was not assessed. In addition, cGAS-STING signaling is upstream of the transcription factor nuclear factor-κB (NF-κB)-mediated SASP regulation (46), which is well-known to affect ovarian cancer dissemination and is discussed in more detail below. In other cancer models, cGAS-STING activation has been shown to drive metastasis (47). However, this may not be the case in ovarian cancer. For instance, a recent study found that inhibition of the STING deubiquitinase USP35 decreases ascites volume in an ovarian cancer mouse model (27). More work is needed to fully understand the role of the cGAS-STING axis in ovarian cancer dissemination.

p53.

p53 is mutated in >90% of high-grade serous ovarian cancers (20), and p53 is a critical regulator of cellular senescence and SASP expression (17). Mounting evidence indicates that p53 is a negative regulator of the SASP, suggesting that cells with mutant p53 likely have higher SASP-like effectors. Indeed, p53-null ovarian cancers secrete higher levels of tumor-promoting chemokines, and this response is attenuated by restoring p53 function (29). Therefore, p53 loss or mutation may promote SASP in a way that is detrimental to patients with ovarian cancer outcomes. Indeed, although it has not been linked to the SASP, compared with the wild-type form, mutant p53 was found to have more metastatic potential in a mouse model of ovarian cancer (30). In addition, p53 affects other phenotypes known to be critical for metastasis, such as anoikis, EMT, angiogenesis, and stemness (48). Future work will need to be done to delineate the roles of mutant p53 in cell cycle control versus SASP in the context of ovarian cancer dissemination.

p16^INK4A and RB.

The p16^INK4A (p16)-retinoblastoma (RB) axis is a critical tumor suppressor pathway that is dysregulated in ∼40% of ovarian cancers (49). When altered, the p16-RB axis allows for E2F-mediated transcription of cell proliferation genes. Recent work from our laboratory and others has shown that a decrease in either p16 or RB suppresses the SASP (4, 16), although neither study was performed in ovarian cancer. Our work found that the suppression of SASP upon loss of p16 was not coupled to the cell cycle (16), suggesting that other unknown factors are at play. Interestingly, patients with RB low ovarian cancer are extremely long-lived (31), and it is intriguing to consider that this may be due in part to decreased SASP, thereby limiting dissemination.

NF-κB and C/EBP-β.

Nuclear factor-κB (NF-κB) activity is required for full transcriptional activation of the SASP (50) and is known to be downstream of cGAS-STING (46). NF-κB is well-known to have pleiotropic roles in ovarian cancer in part through promoting SASP-like actions such as stemness, angiogenesis, and inflammation (32) that affect dissemination (11). A recent report found that patients with high NF-κB pathway activity and low PI3K pathway activity have a more favorable prognosis (51), suggesting elevated NF-κB and its promoting of SASP-like factors may provide better immunosurveillance to limit dissemination, although this remains to be determined. Another important regulator of SASP activation is CCAAT/enhancer-binding protein β (C/EBP-β) (6). Interestingly, C/EBP-β has recently been linked to a transcriptional hub including inflammatory genes that are a part of the SASP (34). C/EBP-β also promotes migration of ovarian cancer cells through cooperation with death-domain-associated protein (DAXX) (35). Increased expression of both NF-κB and C/EBP-β have been linked to poor ovarian cancer survival (32, 52), but whether this is due to their contribution to SASP gene expression is currently unclear.

Epigenetic and chromatin regulators.

Multiple epigenetic and chromatin regulators have been shown to affect SASP transcription, including histone methyltransferases MLL1 and DOT1L, RNA methyltransferases METTL3/14, the epigenetic reader Brd4, and chromatin architectural proteins HMGB1/2 (15). Although not all of these have been linked to ovarian cancer dissemination, it is possible that they contribute to the disease phenotype. DOT1L and its methyl mark H3K79 are positively associated with ovarian cancer stage (53), and it may affect dissemination downstream of C/EBP-β activity or SASP expression as we have recently published (36, 52). Brd4 is also linked to ovarian cancer pathogenesis and dissemination, potentially by affecting EMT (37). HMGB2 is associated with ovarian cancer prognosis (54), and it indeed promotes dissemination downstream of the centromere protein CENPU (40). Finally, METTL3 promotes ovarian cancer invasion and migration through indirect inhibition of CCNG2 (41). Again, whether these effects are through SASP-dependent mechanisms needs to be empirically tested.

mTORC1.

mTORC1 has been shown to promote the SASP through its critical role in translation (19). Like many cancers, ovarian cancers have high mTORC1 activity, and a recent publication found that mTORC1 activity through increased p-4EBP1 correlated with ovarian cancer stage (55). Multiple reports have found that suppression of mTOR activity, either through genetic or pharmacological approaches, decreased ovarian cancer cell migration, invasion, and dissemination (42–45). What is not yet clear is whether this observation is due to reduced translation of SASP and SASP-like factors or the other signaling pathways downstream of mTORC1. As mTORC1 activity is important for a multitude of cellular processes, it is likely that it, at least in part, contributes to dissemination through regulation of SASP mRNAs.

SASP Effectors in Ovarian Cancer

The SASP is composed of multiple factors, including cytokines, chemokines, proteases, and metabolites. Previous work in other models suggests that>100 different factors are secreted as a part of the SASP (56). This has not been extensively studied in ovarian cancer. Since aging promotes ovarian cancer dissemination (9), and cancer therapy induces senescence and the SASP (3, 4, 22), it is likely that SASP factors are a key component of ovarian cancer biology. Here, we will briefly describe SASP factors that are highly relevant in ovarian cancer dissemination, namely, IL6/IL8 and MMPs. Clearly, more work is needed to shed light on other SASP factors and their roles in ovarian cancer.

IL6 and IL8.

Many of the cytokines that are secreted as a part of the SASP have been previously linked to chemotaxis, invasion, and metastasis (6). In addition, malignant ascites derived from patients with late-stage ovarian cancer induces cellular senescence in normal peritoneal mesothelial cells to enhance cell adhesion, proliferation, and migration (57). Two of the most well-described SASP factors are IL6 and IL8, and both are known to affect ovarian cancer dissemination, progression, and survival (58–60). IL6 enriches ovarian CSCs (61), which are often found in the ascites fluid and promote dissemination through multiple mechanisms (62). Interestingly, suppression of the paracrine SASP decreases CSCs and promotes therapeutic response (22), suggesting the SASP is an important contributing factor to stemness in ovarian cancer. Moreover, IL6 facilitates EMT to promote invasion and dissemination of ovarian cancer (63). IL6 is also an immunosuppressive cytokine (64), which may decrease immunosurveillance to allow ovarian cancer cells to more easily spread throughout the peritoneal cavity. High levels of the chemokine IL8 and its receptors CXCR1/2 are also associated with ovarian cancer dissemination and chemotaxis (65). Indeed, IL8 expression in ascites fluid is higher in patients with ovarian cancer with stage III/IV tumors than those with stage I/II tumors (66). IL8 also promotes a mesenchymal phenotype, allowing for increased ovarian cancer migration (65). Together, SASP factors IL6 and IL8 are clearly two contributing factors to the dissemination observed in ovarian cancer, although it is less clear what cells are producing these factors in a complex in-vivo setting.

MMPs.

Interestingly, cytokines/chemokines are not the only factors that influence dissemination. Extracellular matrix (ECM) remodelers such as MMPs also play a role. Not all MMPs are characteristic of the SASP. SASP-related MMPs are MMP1, 3, 10, 12, 13, and 14 (67). In ovarian cancer, MMP3 is significantly associated with a higher stage (68), suggesting it plays a role in dissemination and/or invasion. MMP1/3 have also been shown to affect intraperitoneal dissemination of ovarian cancer (69). In addition, MMP1 mRNA was recently shown to be in extracellular vesicles within ascites fluid, which promoted ovarian cancer dissemination in mouse models (70).

Therapeutic Strategies

One of the major difficulties in ovarian cancer treatment is that it is often undetected until the cancer has disseminated to distant sites (1). As discussed earlier, many SASP regulators and effectors are well-known to affect dissemination. Therefore, it is intriguing to speculate that inhibition of the SASP may decrease dissemination and improve patient outcomes. Moreover, many SASP factors also promote chemoresistance (6), suggesting a dual-pronged approach to treating ovarian cancers. Because of these detrimental effects, there is increasing interest in using therapies that kill senescent cells (senolytics) or modulate the SASP (senomorphics). Recent work from the Campisi laboratory demonstrated that clearance of senescent cells using an elegantly engineered mouse model decreases both chemoresistance and metastasis in a mouse breast cancer model (71), showing proof-of-principle of this concept. Senolytic drugs are thought to work by targeting antiapoptotic pathways. However, emerging evidence suggests that senolytic effects are primarily due to elimination of the SASP (72). The senolytic and BCL2 inhibitor navitoclax have been shown to clear senescent ovarian cancer cells after treatment with PARP inhibitors and prevent cancer relapse in mice treated with doxorubicin (3, 71). No work has yet been done in ovarian cancer to investigate whether targeting the SASP using senolytics decreases dissemination.

Other regulators of the SASP may also be targetable in ovarian cancer. For instance, there has been significant excitement about bromo- and extra-terminal domain (BET) inhibitors in ovarian cancer (37), and the BET family protein Brd4 is known to affect the SASP through enhancers (8). Indeed, BET inhibitors suppress ovarian cancer dissemination in part through inhibition of STAT3 (38), a known regulator of the SASP (6). Other pathways known to regulate the SASP such as NF-κB and mTOR have been targeted in ovarian cancer, but with little success thus far (32, 73). It is possible that induction of senescence with platinum or PARP inhibitors before these therapies would provide a better therapeutic window to suppress the SASP and thereby limit dissemination.

On the other hand, recent work has demonstrated that induction of senescence within the tumor promotes immune cell infiltration and increases response to immune checkpoint blockade (4, 7). Indeed, boosting the SASP or increasing the cGAS-STING axis has multiple therapeutic benefits in mouse models of ovarian cancer (7, 28). Collectively, these data indicate that induction of the SASP may be beneficial at least in the context of the immune response to make these tumors “hot,” thereby sensitizing them to immunotherapy. More work needs to be done to assess whether this leads to a long-term, sustained therapeutic response in humans.

SUMMARY

There has been an explosion of new information on regulation of the SASP and its contents at a mechanistic level. As we have discussed, the SASP is extremely diverse and not only modulates inflammation but also ECM remodeling, stemness, chemoresistance, and many other tumor-suppressive and tumor-promoting phenotypes that may contribute to ovarian cancer dissemination. However, many questions remain to be answered, especially related to the causal role of senescence and the SASP in ovarian cancer dissemination. For instance, we still need to investigate the exact components of the SASP from ovarian cancer cells in response to different stimuli and whether those diverse profiles have similar effects on dissemination. In addition, we are just starting to understand the types of immune cells that are recruited and activated by the SASP from ovarian cancer cells. We will need to further delineate whether this is dependent on the site of the tumor (primary versus disseminated) and how that affects ovarian cancer biology. Finally, we need to be able to better understand how to fine-tune the SASP to harness its potential benefits while suppressing its negative effects.

GRANTS

This work was supported by grants from the National Institutes of Health (R01CA259111), the American Cancer Society (RSG-19-113-01), and the Department of Defense Ovarian Cancer Research Program (W81XWH-21-1-0840) (to K.M.A.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.M.A. prepared figures; J.P.V., L.F.F.B., and K.M.A. drafted manuscript; J.P.V., L.F.F.B., and K.M.A. edited and revised manuscript; J.P.V., L.F.F.B., and K.M.A. approved final version of manuscript.