Sweet dreams: Glycosylation controls tumor cell dormancy

Abstract

In a recent study in Cancer Cell, Sreekumar et al. used therapy-associated mouse models as well as in vitro dormancy models to identify extracellular matrix (ECM)-related tumor cell autonomous mechanisms of dormancy in residual tumor cells (RTCs). The study reveals an important role of glycosylation of proteoglycans in sustaining dormancy and opens the door to leverage this biology to eliminate RTCs and prevent recurrence.

Tumor recurrence in breast cancer arises from residual tumor cells (RTCs) that reside in a quiescent state, named cellular dormancy, before they are awakened and restore proliferation. Tumor cells disseminate early during breast cancer progression, seeding different organs such as the lung, brain, or liver, and give rise to recurrences years or even decades after the primary tumor has been treated [1]. Recent work has shown tumor cell intrinsic mechanisms as well as microenvironment-related factors regulate the entrance into and escape from dormancy of these RTCs [2].

An important regulator of tumor cell dormancy is the extracellular matrix (ECM). Building an ECM niche by using tumor-derived ECM has been proposed as a mechanism to sustain tumor cell dormancy [3–4]. ECM formation is a complex process that requires not only the efficient assembly of its constituents but also post-translational modification of the ECM molecules to assure their proper organization and ability to bind to receptors and other molecules. How dormant tumor cells regulate these steps is not completely understood.

To identify cell-autonomous regulators of RTC dormancy, a study in Cancer Cell by Sreekumar et al. used the HER2 downregulation mouse model (MMTV-rtTa;TetO-Her2/Neu) [5], which mimics targeted therapy by depletion of the HER2 receptor upon doxycycline withdrawal. Following primary tumor removal and HER2 downregulation, RTCs enter quiescence in vivo, and give rise to recurrences months later [6]. Analysis of the transcriptome in the RTC population in vitro revealed an upregulation of genes related to ECM organization, validating previous studies showing the relevance of tumor-derived ECM in tumor dormancy maintenance [7]. Interestingly, this RTC gene signature has significant overlap with genes upregulated in RTCs in vivo, and can also be found in early-stage breast cancer patients that have lower risk of recurrence, suggesting these genes exhibit features of cellular dormancy.

Sreekumar et al. performed a CRISPR-CAS9 loss-of-function screening to identify genes involved in RTC survival and identified B3galt6 as a pro-survival gene. B3GALT6 is an ECM regulator that catalyzes the addition of glycosaminoglycans (GAGs) to proteins. B3GALT6 adds the second Gal residue in the GlcA-Gal-Gal-Xyl-O-tetrasaccharide linker, decorating proteoglycan core proteins with GAGs. The addition of GAGs to proteoglycans facilitates the formation of ligand–receptor complexes and cell-matrix interactions, and plays an important role in the tumor microenvironment [8]. In the same screening, the authors showed that DDR1, a collagen receptor previously shown to be important for dormancy [7], regulates cell cycle arrest of RTC and prevents their expansion. While ECM molecules have been shown to be involved in tumor cell dormancy, the implications of posttranslational modifications, and their regulation and impact on tumor cell dormancy, is not understood.

In vivo experiments revealed that depletion of B3GALT6 impaired fitness of RTCs upon HER2 downregulation, through increasing apoptosis. B3GALT6 is also required for efficiency of tumor recurrences from RTCs post-therapy. By using an in vitro dormancy model (the D2.OR: dormant; D2A1: proliferative pair cellular model grown in a 3D matrix [9]) the authors showed that B3GALT6 is also required for the viability and the outgrowth of dormant D2.OR, validating the role of B3GALT6 in a microenvironment-induced dormancy model.

By using liquid chromatography-mass spectrometry (LC/MS) and immunofluorescence studies the authors found that heparin sulfates (HS) were the most abundant type of GAGs compared to chondroitin sulfate in dormant cells. In a series of elegant experiments, the authors showed that the addition of heparin could rescue impaired cell survival observed in B3GALT6-depleted cells. Modifications of HS are crucial for ligand-binding properties of GAGs, and LC/MS analysis revealed an upregulation of sulfated GAGs in dormant cells, in particular 6-O-sulfation.

Finally, the authors determined the downstream signaling effects of sulfated GAGs and the subsequent impact on dormancy. Since heparin 6-O-sulfation promotes FGF signaling by increasing the binding of FGF ligands, the authors analyzed the levels of FGF ligands in dormant cells and found that FGF1 was dramatically upregulated during dormancy. HS6T1, the enzyme mediating 6-O-sulfation, was identified as a pro-survival gene in the CRISPR screen. Indeed, 6-O-glycosylation potentiates FGF signaling as revealed by experiments where Hs6st was depleted. sgHs6st attenuates FGF signaling and cell viability in response to FGF1 stimulation. FGF1 regulates ERK1/2 signaling in dormant cells through FGFR2. Overall, the data revealed a new pathway regulating dormancy where the enzyme HS6T1 mediates 6-O-sulfation of B3GALT6-mediated heparan sulfate GAGs to potentiate FGF1 signaling through FGFR2 (Figure 1).

Figure 1. B3galt6/Hs6st1/Fgf1/Fgfr2 regulates dormant residual tumor cell (RTC) survival.

a) In dormant residual tumor cells B3GALT6 mediates the linkage of heparan sulfate to proteoglycans. HS6ST1 mediates the 6-O-sulfation of heparin sulfate. This step is required to potentiate FGF1 signaling though FGFR2. b) Interfering with B3GALt6, HS6ST1, or the FGFR2 receptor attenuated ERK1/2 signaling and compromised the survival of dormant RTCs, inducing apoptosis. The figure was created with BioRender.

Earlier work showed the effects of targeting glycosylation in preventing metastatic dissemination and metastasis in mouse models [10]. The data presented in this study by Sreekumar et al. revealed an important aspect of glycosylation in regulating a key step of the metastatic cascade: the dormancy phase. Will preventing glycosylation also eliminate dormant cells in patients, as shown in murine models? Other components of the pathway, like FGFR2, could also be targets that could eliminate dormant cells. Interestingly, FGFR2 inhibitors are in clinical trials for intrahepatic cholangiocarcinoma, which could be used as a potential strategy to prevent breast cancer recurrence.