EccDNA, STING activation, and their potential roles in DLBCL prognosis and therapy

Diffuse large B-cell lymphoma (DLBCL), the most common lymphoma in adults, poses a significant clinical challenge due to its marked heterogeneity and high incidence of patients suffering relapse or becoming refractory after the first-line immunochemotherapy. DLBCL cells may originate from different stages of lymphoid differentiation, and thus their gene expression profiling can delineate two distinct molecular subtypes of DLBCL: germinal center B-cell-like (GCB) and activated B-cell-like (ABC) subtypes, with a minority of cases remaining unclassifiable. Alternatively, immunohistochemistry-based algorithms can also dichotomize DLBCL into GCB and non-GCB subtypes. In addition to these classifications related to cell-of-origin, recent in-depth genomic analyses have also revealed recurrent genomic aberrations, prompting the proposal of new classification systems for DLBCL.^{1, 2} However, as yet the correlation between prognosis and molecular subtypes remains elusive.³ As an effort to further explore the heterogeneity of DLBCL, in an article recently published in Clinical and Translational Medicine, Wu et al. reported the first profiling of extrachromosomal circular DNA (eccDNA) in DLBCL cells and identified the oncogenic role and prognostic significance of the eccDNAs in DLBCL. Meanwhile, their multi-omic and mechanistic studies uncovered an activation of the stimulator of interferon genes (STING) signalling pathway by eccDNAs in DLBCL cells, thus suggesting a potential therapeutic approach combining chemotherapy with STING inhibition (Figure 1).⁴

EccDNAs are circular double-strand DNA molecules that are originally derived from, but physically independent of, chromosomal DNA. They are heterogeneous in size ranging from a few hundred to millions of base pairs and appear to be derived from either repetitive sequences or unique genomic sequences, and many eccDNAs can be self-replicated. While the biogenesis and functions of eccDNAs remain to be fully understood, an increase in eccDNA has been observed in several types of human cancers. Importantly, because the eccDNAs provide a significant source of oncogene amplification and can be segregated unequally to daughter cells, they are believed to contribute enormously to tumour evolution and intratumor heterogeneity,⁵ and the presence of eccDNAs in patient tumour samples has been found to be associated with poor outcomes across many types of cancers.⁶ Furthermore, studies have shown that eccDNAs can function as potent innate immunostimulants, which require the STING signalling pathways to serve as intracellular DNA sensors.⁷

By integrating circular DNA sequencing (circle-seq), atomic force microscopy, whole exome sequencing (WES) and single-cell RNA sequencing (scRNA-seq) techniques, Wu et al. performed a comprehensive eccDNA profiling of 18 DLBCL cell lines. Based on the eccDNA abundance, they classified these DLBCL cell lines into eccDNA high (H) and low (L) groups. While the eccDNA abundance was observed to be associated with the genome instability index of the cells (indicated by WES results), it appeared independent of the GCB versus non-GCB classification of DLBCL, therefore implying a potentially distinct classification system for DLBCL cases. To further explore the biological significance of the eccDNA abundance in DLBCL, Wu et al. analyzed the scRNA-seq date of the DLBCL cells, constructed a trajectory of cellular differentiation and identified differentially expressed genes that were associated with eccDNA abundance. Importantly, expression patterns of these genes in DLBCL patient samples were found to be well associated with the prognosis of the patients, thus providing a useful biomarker for prognostic stratification of DLBCL cases (Figure 1, left panel).

Wu et al. then transfected U2932 cells (a non-GCB DLBCL cell line belonging to the L group) with eccDNAs and observed promoted cell proliferation. scRNA-seq analysis showed upregulation of genes associated with cell proliferation and interferon response in the eccDNA-transfected cells. Considering the activation of interferon response and the known functions of the STING signalling pathway to sense intracellular DNA,⁸ Wu et al. went on to determine whether the STING signalling pathway was activated. Indeed, they found that the eccDNA transfected cells showed increased phosphorylated STING (p-STING) levels, as well as activation of several downstream effectors. Meanwhile, they also made efforts to address the mechanism of eccDNA biosynthesis in DLBCL cells and concluded that the chemotherapeutic drug-induced DNA damage might play an important role in eccDNA generation and STING activation (Figure 1, middle panel).

Notably, an interesting point proposed in this article is that the eccDNA-stimulated STING activation is independent of cyclic GMP-AMP synthase (cGAS), which is reminiscent of previous observations of a non-canonical, cGAS-independent STING activation induced by DNA damage.⁹ Herein presented supporting evidence for this point includes the Western blot results showing no significantly different levels of cGAS protein in the H and L group cell lines, relative inefficiency of cGAS inhibition (either by a cGAS inhibitor or knockdown of cGAS) on cell proliferation, and variable levels of immunofluorescent staining of cGAS in patient-derived DLBCL cells regardless of p-STING levels. However, to firmly establish this point, more experiments are still required, and it remains to be clarified whether this cGAS independency is cell type-specific or represents a universal mechanism for eccDNA-stimulated STING activation.

In a therapeutic aspect, Wu et al. evaluated several small molecule drugs targeting the cGAS-STING pathway. The results showed that STING inhibitors, but not STING agonists or cGAS inhibitors, effectively inhibited the proliferation of the H group DLBCL cells and tumour progression in xenograft mouse models. Based on the logic that DNA-damaging chemotherapy can enhance eccDNA generation and STING activation in DLBCL, Wu et al. further invented a combinative therapeutic approach by combining chemotherapy drugs with STING inhibitors. Experimental results indicated synthetic effects in both in vitro and in vivo settings, thus implying a promising new therapeutic strategy for DLBCL, especially for the relapsed and refractory cases which contain higher eccDNA abundance (Figure 1, right panel).

The present study opens new perspectives to evaluate the potential role of eccDNA and downstream STING activation in DLBCL prognosis and therapy; however, there are several interesting questions remaining to be addressed. First, although this study has developed transcriptomic signatures to reflect the eccDNA abundance in DLBCL cells, the eccDNAs themselves, especially considering their relative stability and readily detectability not only in cells but also in plasma and urine,¹⁰ could also be considered as potential biomarkers for DLBCL prognosis. Second, if the eccDNA-stimulated STING activation is truly cGAS-independent, it would be important to clarify what is the exact mechanism for eccDNA sensing. Third, although the STING inhibitors have shown considerable efficacy in treating DLBCL in this study, given the diverse functions of the STING signalling pathway in different cells/tissues, especially in the immune system, more studies are required to evaluate and manage the benefits and risk before their application in clinics. Lastly, besides DLBCL, it would be interesting to investigate whether the herein proposed mechanisms and therapeutic approach could be applicable to other types of cancers.

All authors wrote and edited the manuscript.

The authors declare no conflict of interest.

Not applicable.