Softening redox homeostasis in cancer cells

Abstract

Extracellular matrix (ECM) rigidity increases during tumour progression. In a recent study, Romani et al. delineated a connection between ECM stiffness and the redox response of disseminated tumour cells. Their results suggest that soft ECM promotes DRP1-mediated mitochondrial fission and an NRF2-dependent antioxidant response.

The tumour microenvironment (TME) exerts a powerful influence over cancer progression, particularly during the process of metastatic dissemination, when primary tumour cells spread to and colonize distant organs. How newly seeded tumour cells progress into an unequivocal metastatic lesion remains poorly understood. Although some newly seeded tumour cells may readily proliferate in their new environment, others may enter a quiescent state of dormancy. Indeed, in breast, prostate and other cancers, these cells can remain dormant for decades before developing into metastatic cancer. Although the clinical importance of tumour dormancy is apparent, its underlying mechanisms remain unclear. In this issue, Romani and colleagues¹ report that a mechanotransduction signalling pathway that promotes mitochondrial fission is activated when cancer cells infiltrate into the ‘soft matrix’ (i.e., tissues of low stiffness) of distant organs. This pathway, which is dependent on dynamin-related protein 1 (DRP1), induces changes in mitochondrial morphology and increased levels of mitochondrial reactive oxygen species (ROS), as well as activation of the transcription factor nuclear factor erythroid 2-related factor 2 (NRF2), the master regulator of a vast array of cytoprotective genes (Fig. 1). By showing that genetic inactivation of this pathway enhances chemosensitivity and curbs cancer awakening in mouse models, this study suggests that mechanotransduction, through its impact on redox metabolic reprogramming, is a key determinant of metastatic spread.

Fig. 1 ∣. Mechanosignaling through changes in cell adhesion and matrix stiffness regulates the dynamin-related protein 1 (DRP1)-nuclear factor erythroid 2-related factor 2 (NRF2) axis in disseminated breast cancer cells.

Romani et al. identified increased incidence of shortened mitochondria, due to DRP1-dependent fission, in breast cancer cells exposed to soft matrix or pharmacological reduction of cell contractility. This morphological change is associated with an increase in mitochondrial reactive oxygen species (mito-ROS) production and activation of the NRF2 antioxidant response. Activation of the DRP1-NRF2 axis leads to increased resistance to oxidative stress in dormant cancer cells in soft tissue.

In living tissues, cells sense and respond to a wide range of mechanical tensions, including extracellular matrix (ECM) rigidity, fluid shear stress and intercellular compression. Through a variety of mechanotransduction pathways, these mechanical signals can be converted into biochemical signals that trigger the desired cellular response. The tumour stroma is known to increase in stiffness over time as it coevolves with cancer cells². This is well exemplified by breast tumours, which are often first detected by direct palpation or mammographic density, both of which reflect the increased stiffness of the tumour compared to surrounding tissues³. Moreover, tumour stiffness, as measured by shear wave elastography ultrasound imaging, is highly predictive of patient response to neoadjuvant chemotherapy⁴ and diminishes significantly upon therapy⁵. Interestingly, the effects of matrix rigidity on cancer cell behaviour have also been seen in vitro, where cells grown in three-dimensional cultures of greater ECM stiffness display a more invasive phenotype⁶. Nonetheless, the impact of adhesion-mediated mechanosignaling on freshly disseminated tumour cells had not been explored. Romani et al. have now employed atomic force microscopy to show that the stiffness of early metastatic sites is four times lower than that of late metastatic outgrowths. The authors also propose that the response of a disseminated cancer cell to the soft stroma of a distant organ promotes the chemoresistance of dormant cancer cells¹. Indeed, emerging evidence suggests that acellular factors, such as changes in the local ECM composition, may trigger dormant cancer cell awakening and metastasis in experimental models⁷. However, the molecular processes by which mechanotransduction regulates the properties of dormant cancer cells remain elusive.

In this regard, it is intriguing that matrix rigidity has recently emerged as a player in metabolic reprogramming. Of note, Tharp et al. reported that untransformed mammary epithelial MCF10A cells display increased mitochondrial fragmentation when exposed to stiff substrates or engineered to express a gain-of-function β1-integrin mutant that promotes focal adhesion assembly; either treatment results in elevated ROS production and activation of the HSF1-dependent oxidative stress response⁸. Similarly, another study performed in U2OS and COS7 cells demonstrated that mechanical forces, applied either directly as external pressure using an atomic force microscope or indirectly by culturing cells on uneven microsurfaces, result in recruitment of the mitochondrial fission machinery and subsequent mitochondrial division⁹. Paradoxically, in RAS-transformed MCF10A cells and D2.0R metastatic breast cancer cells, mitochondrial fission and ROS are instead induced by soft matrix or by treatment with inhibitors of rho-associated kinase (ROCK) or myosin light chain kinase (MLCK)¹. Although at first glance these observations appear to be conflicting, they may reflect the outcomes of distinct mechanosignaling cascades triggered by mechanical stimuli of different natures. For example, reduced cell contractility modelled by ROCK or MLCK inhibition is often viewed as equivalent to growth in soft matrix, but it remains to be established whether the signals conveyed by impaired adhesion and soft matrix are qualitatively the same. This is an important consideration, especially in the context of metastasis, as disseminated cancer cells first experience complete loss of matrix adhesion while in circulation, followed by soft matrix adhesion upon seeding distant organs and, finally, stiff matrix stimulation during metastatic outgrowth (Fig. 1). Meanwhile, recent work has established that cancer cells undergo metabolic reprogramming to overcome the ROS-induced cellular stresses that they experience during dissemination¹⁰. However, the sources of oxidative stress in metastasizing cancer cells remain to be fully determined, especially as the ranges of oxygen tension in blood (e.g., ~40 mmHg for venous and ~100 for arterial) and normal human tissues (10–72 mmHg) overlap¹¹. Collectively these observations raise provocative questions regarding the contributions of mechanosignaling to redox and metabolic states at different stages of metastasis and the extent to which these signals influence properties of dormant cancer cells.

Mitochondria undergo dynamic fission and fusion, the balance of which constitutes an important response to changing physiological conditions. Increased fusion generates interconnected mitochondria, which are favourable for metabolically active cells. On the other hand, quiescent cells favour morphologically and functionally distinct mitochondrial fragments that are generated by increased fission. Dynamin-related protein 1 (DRP1) is essential for mitochondrial division, which in turn increases mitochondrial ROS production in culture¹². In their study, Romani et al. observed that DRP1-mediated mitochondrial fission induced upon soft-matrix culture is associated with activation of NRF2¹, a key regulator of antioxidant and cytoprotective responses. NRF2 is normally sequestered in the cytosol by Kelch-like ECH-associated protein 1 (KEAP1), an adaptor of a ubiquitin ligase complex that constitutively targets NRF2 for proteasomal degradation¹³. The authors found that soft ECM induced dissociation of NRF2 from KEAP1, leading to NRF2 stabilization and transactivation of NRF2-target genes involved in antioxidant and cytotoxic responses. Because activation of NRF2 can in turn promote the degradation of DRP1¹⁴, soft-matrix-induced NRF2 activation may constitute a negative feedback mechanism that maintains mitochondrial homeostasis and provides cytoprotective effects to disseminating cancer cells.

Several important questions arise from the above observations. First, how does mitochondrial ROS induced by soft ECM activate NRF2 in the cytosol? A recent study demonstrated that disruption of mitochondrial thiol homeostasis, but not increased mitochondrial superoxide levels, is required for cytosolic sensing through the KEAP1–NRF2 axis¹⁵. This raises the possibility that mitochondrial fission and mitochondrial thiol homeostasis are co-regulated. Because glutathione is the major thiol-containing compound in the mitochondria, it would be interesting to ascertain whether mechanosignaling influences glutathione import into the mitochondria. Also, given that the authors reported a significant increase in endoplasmic reticulum (ER) ROS upon soft matrix culture¹, it will be important to determine whether ER stress contributes to the observed changes in mitochondrial morphology. Another interesting question is the molecular link between ECM mechanics and mitochondrial fission. Whereas Romani et al. propose a relationship between soft-matrix-induced mitochondrial fission and the remodelling of perimitochondrial F-actin¹, Helle et al. have reported that mitochondrial fission can occur independently of actin dynamics⁹. Finally, as mitochondria are not free-floating but are instead anchored to cellular structures through cytoskeletal components, we should also entertain the possibility that forces generated by alterations in cell adhesion may be directly transferred to mechanical sensors on the mitochondrial surface.

The existence, and potential awakening, of dormant tumour cells represents a formidable challenge to cancer therapy, especially as it exposes patients to the risk of disease recurrence after intended curative surgery. This calls into question whether dormant cancer cells can be targeted as part of adjuvant treatment strategies. The current study suggests that cancer mechanosignaling through the DRP1–NRF2 axis may represent a therapeutically actionable node for curbing the awakening of dormant cancer cells. Understanding the metabolic responses to ECM signals will add a valuable dimension to the ongoing research into cellular and acellular mechanisms that govern dormancy.