The tumor microenvironment (TME) impacts different phases of tumor progression and therapy resistance. Zhang et al. show that senescent stromal cells activate an epigenetic program that controls the senescence-associated secretory phenotype and can be targeted to boost responses to chemotherapy.
Cellular senescence is a physiological process occurring in normal and tumor cells consisting of proliferation arrest coupled to the activation of a peculiar secretory phenotype, known as senescence-associated secretory phenotype (SASP). Senescence can be elicited by different stimuli such as replicative stress, oncogene induction or DNA damage activation1,2. Its establishment and maintenance are accompanied by widespread changes in gene expression, a dynamic process largely regulated by epigenetic alterations. Indeed, senescent cells are generally characterized by changes in chromatin remodeling, histone-associated epigenetic processes, DNA methylation and non-coding RNA expression3. At the global level, the balance between activating and repressive marks in senescent cells is shifted towards an open chromatin structure3. In this issue of Nature Aging, Zhang et al. analyzed histone modifications in tumor-associated fibroblasts after therapy-induced senescence (TIS). They found that KDM4, a family of demethylases that preferentially target histone H3 on lysine 9 and 36 sites (H3K9 and H3K36), is involved in the epigenetic regulation of the SASP in senescent tumor stromal cells4. Their results lead to the proposal of a new therapeutic approach based on the targeting of the SASP of stromal cells in cancer.
Alterations of senescence-associated histone modifications are thought to be the most direct epigenetic response to senescence induction3. They include the reduced expression of histone proteins, the redistribution of histone variants and histone posttranslational modifications. Several histone modifications, correlating with both open (H3K4me3, H3K36me3, H3K27Ac), or closed (H3K27me3, H3K9me3) chromatin states, have been observed in different genomic regions of senescent cells. However, these modifications may vary according to the trigger of senescence induction and to the tissue of origin of senescent cells, precluding the definition of a ‘histone code’ for senescent cells5. Overall, the regulatory networks of the epigenetic mechanisms defining senescence induction and maintenance remain poorly understood.
In their study, Zhang et al.4 observed that senescent fibroblasts present a general downregulation of histone 3.2 (H3.2) lysine methylation paralleled by the upregulation of demethylase enzymes KDM4A and KDM4B. In stromal cell lines of prostate and lung, KDM4A/B upregulation and the consequent H3.2 lysine demethylation were triggered by several chemotherapeutic agents and were also observed in primary prostate and lung fibroblasts undergoing replicative exhaustion and oncogene-induced senescence.
Among the genes actively transcribed by senescent cells, several codify for the plethora of secreted factors determining the SASP. The SASP consists of a large variety of molecules including cytokines, chemokines, growth factors and matrix remodeling factors. Depending on the context, the SASP can be responsible for detrimental consequences, including aging-related disorders or resistance to cancer therapy1,6. Since the SASP and cell cycle arrest are independently regulated7, several groups have tried to avoid SASP-related undesirable effects by interfering with specific signaling pathways (such as nuclear factor-κB (NF-κB) and CCAAT enhancer-binding protein β (C/EBPβ)) by targeting individual secreted factors or by epigenetic interventions. For example, the chromatin shaping factors BRD4 (ref. 8) and MLL1 (ref. 9) have been shown to regulate SASP expression independently of their role in proliferation arrest.
By genetically modulating KDM4A/B enzymes, Zhang et al. demonstrated that H3.2 demethylation in stromal senescent cells is required for the transcriptional activation of the secretome, without exerting any effect on the cell cycle arrest. In their study, the SASP could also be successfully disabled pharmacologically, taking advantage of the small molecule ML324 that is capable of inhibiting KDM4 factors. ML324 treatment of stromal cells phenocopied KDM4A/B silencing and specifically modified chromatin accessibility and transcriptional expression in loci associated with SASP factors, as confirmed by assay for transposase-accessible chromatin using sequencing (ATAC-seq) and chromatin immunoprecipitation with sequencing (ChIP-seq). KDM4A has been previously implicated in senescence as a negative regulator of the p53 pathway and was shown to be required for the expression of the SASP factors interleukin-6 (IL-6) and IL-8 (ref. 10). The work by Zhang et al. expands these findings, showing that KDM4 factors orchestrate a global epigenetic program through demethylation of histone H3.2 sites4. Other epigenetic modulators, like some histone deacetylase (HDAC) and DNA methyltransferase inhibitors, have been approved for cancer therapy. Of note, HDAC inhibitors show antiproliferative or proapoptotic effects in the tumor while activating the SASP in the stromal compartment11. Thus, a better understanding of the chromatin state around SASP genes and how their expression is regulated by histone modifications in each cell compartment is fundamental, since interfering with epigenetic states in the tumor microenvironment might have an opposing effect.
While most studies analyzing cellular senescence usually consider one cellular compartment at a time, Zhang et al. analyzed chromatin changes both in the tumor and in the stroma (Fig. 1). The stroma plays a key role in the paracrine stimulation of tumor growth12,13, and most standard chemotherapies induce senescence in tumor and non-tumoral surrounding cells. Intriguingly, Zhang et al. describe a senescence-associated epigenetic program peculiar to stromal cells: indeed, senescent cancer cells derived from prostate and lung malignancies did not display consistent upregulation of KDM4 factors, nor a general reduction in H3.2 methylation4. Zhang et al. also corroborated this observation in clinical settings by analyzing biopsies from prostate cancer patients before and after chemotherapy. Importantly, they observed a prominent induction of KDM4A/B after chemotherapy in the stromal, but not in the epithelial, compartment, which correlates with a negative prognosis4. Therapy-induced KDM4A/B overexpression correlated with the upregulation of senescence markers and the reduction of H3 methylation, supporting the link between senescence establishment and methylation regulation in the tumor stroma.
Of particular note, KDM4A/B inhibition through ML324 was sufficient to revert the pro-tumoral effect elicited by the senescent stroma in vitro and in vivo (Fig. 1). By exploiting a xenograft model mixing both stromal and epithelial cancer cells, Zhang et al. were able to prove the efficacy of a therapeutic regimen combining a genotoxic agent (mitoxantrone) and the KDM4 inhibitor ML324. The combination therapy reduced tumor growth to a greater extent than the single-agent therapy. Importantly, the positive effects obtained by KDM4A/B inhibition were due to the presence of stromal cells: indeed, in their absence, ML324 did not increase mitoxantrone efficacy.
These data open up the intriguing possibility that KDM4 inhibitors could be effective in blocking tumor progression in patients affected by prostate cancer. While mitoxantrone is not commonly used in the clinic for the therapy of prostate cancer, other agents such as enzalutamide or docetaxel may elicit cellular senescence not only in tumors, but also in stromal cells14,15. However, further data are needed in order to support the clinical development of KDM4 inhibitors in combination with these standard therapies for prostate cancer.
In summary, the study of Zhang et al. identifies a new layer of epigenetic regulation of the secretome of senescent cells through histone modification by KDM4 enzymes. This epigenetic program is dynamically activated in damaged stromal cells and provides a promising target for cancer and, potentially, aging-related disorders.
Gorgoulis, V. et al. Cell 179, 813–827 (2019).
Calcinotto, A. et al. Physiol. Rev. 99, 1047–1078 (2019).
Cheng, L.-Q., Zhang, Z.-Q., Chen, H.-Z. & Liu, D.-P. J. Mol. Med. 95, 1257–1268 (2017).
Zhang, B. et al. Nat. Aging https://doi.org/10.1038/s43587-021-00063-1 (2021).
Paluvai, H., Di Giorgio, E. & Brancolini, C. Cells 9, 466 (2020).
Guccini, I. et al. Cancer Cell 39, 68–82 (2021).
Coppé, J.-P. et al. J. Biol. Chem. 286, 36396–36403 (2011).
Tasdemir, N. et al. Cancer Discov. 6, 612–629 (2016).
Capell, B. C. et al. Genes Dev. 30, 321–336 (2016).
Johmura, Y. et al. Nat Commun. 7, 10574 (2016).
Pazolli, E. et al. Cancer Res. 72, 2251–2261 (2012).
Valkenburg, K. C., de Groot, A. E. & Pienta, K. J. Nat. Rev. Clin. Oncol. 15, 366–381 (2018).
Krtolica, A., Parrinello, S., Lockett, S., Desprez, P.-Y. & Campisi, J. Proc. Natl Acad. Sci. USA 98, 12072–12077 (2001).
Toso, A. et al. Cell Rep. 9, 75–89 (2014).
Ewald, J. A. et al. Prostate 73, 337–345 (2013).
The authors declare no competing interests.
About this article
Cite this article
Desbats, M.A., Zumerle, S. & Alimonti, A. Epiregulation of the SASP makes good neighbors. Nat Aging 1, 420–421 (2021). https://doi.org/10.1038/s43587-021-00068-w