**8. cfDNA and STING-cGAS Pathway Activation: Stairway to Heaven or Highway to Hell?**

The cfDNA-mediated activation of the STING-cGAS pathway seems, therefore, to be a promising way to optimize immune checkpoint inhibitor therapy; however, this pathway is a double-edged sword [93]. Beyond the antitumoral effect of STING-cGAS, it also plays a major role in carcinogenesis, particularly via inflammation activation, autoimmune response, and direct tissue toxicity [71]. In carcinogenesis, MYD 88 (molecule myeloid differentiation factor 88) signaling seems to play a major role in cancer development via cytokine, chemokine, and growth factor production. In functional STING mice exposed to DMBA (a polyaromatic hydrocarbon known as carcinogenic agent), cutaneous skin tumors with pro-inflammatory cytokine production and phagocytic infiltration were observed [94]. Conversely, deficient STING mice did not develop such skin tumors post DMBA exposure, suggesting the role STING could play in skin cancer development [95]. Conversely, in some cancer types such as melanoma or colon cancer, the STING-cGAS pathway can be impaired by loss-of-function mutation or epigenetic silencing of the STING-cGAS promoter regions [96]. In this case, quantification of cell-free DNA could not be associated to an upregulation of STING-cGAS. Furthermore, in tongue squamous cells induced by human papillomavirus models, STING-cGAS activation eases T Regs infiltration and could, therefore, have a negative impact [97].

The cfDNA, itself, can be a confounding factor. As previously underlined, cfDNA is more important in cancer patients, and is not always synonymous of inflammatory situations. The cfDNA reflects genome plasticity, leading to a physiological and autonomous process of cells for its elimination [59]. Half-life of cfDNA should also be considered; it varies between 15 min and few hours [98], thus raising questions about appropriate sampling time for decision-making.

#### **9. Discussion: Shine a Light**

Immunotherapy and more specifically, immune checkpoint inhibitors, have fueled huge hope in oncology. However, the clinical results in terms of survival have failed to meet the initial expectations because only a minority of patients show long term survival, in a minority of cancer disease such as melanoma, NSCLC, head, neck, and kidney cancers. Consequently, combinatorial regimens now turned to as the rule, and not anymore, the exception with immunotherapy, in an attempt to harness tumor immunity prior to administrating immune checkpoint inhibitors. This strategy is expected to transform promising and breakthrough pharmaceutical innovations into meaningful survival in patients. The main difficulty when using immune checkpoint inhibitors is the complexity of their mechanisms of action which cannot be reduced to PD-L1, PD1, or CTLA4 inhibition anymore as once thought, but require as well, an adequate tumor micro-environment enriched with activated lymphocytes with little T Regs or MDSC activity. In this respect, using canonical cytotoxics is an appealing strategy to increase immunogenic cell death while down-regulating immunosuppressive cells [99]. However, achieving such immunomodulating features requires fine tuning since they are much probably drug- and dose-dependent [100]. For instance, in several non-clinical studies combining anti-PD1 or anti-PDL1 with anti CDK 4/6 or anti-OX40 drugs, it has been demonstrated that even slight changes in scheduling are likely to dramatically change the response to the combinations [101]. The same phenomenon has been observed as well when combining immune checkpoint inhibitors with radiation therapy [102]. All these studies call for a comprehensive understanding of the exact dynamics of tumor reengineering when drugs are used to harness immunity and expected to yield synergistic effect with immunotherapy. Indeed, capturing this exact dynamic is a tricky but critical issue, because it could provide valuable information on the best dosing, timing, and sequencing, especially when combining cytotoxics with immune checkpoint inhibitors. Unfortunately, as of today current combinatorial strategies remain empirical and concomitant dosing is frequent. Consequently, many clinical trials have failed to yield meaningful results in terms of prolonged survival [103]. Notably, as previously explained, baseline PDL1 expression levels cannot help determining the optimal modality of combination between immune checkpoint inhibitors and associated treatments such as chemotherapy. The same observation can be made with tumor burden, medical records, microbiota features, or any of the numerous parameters tested so far as putative predictive markers with immunotherapy. This calls for using more dynamic markers better reflecting real-time changes in the tumor micro-environment such as immunogenic cell death. Because as seen before, STING-cGAS is a critical signaling pathway associated with response to immunotherapy and that release of cfDNA, especially the shorter ones, activates this pathway, blood quantification of cfDNA could be a convenient surrogate to monitor STING-cGAS activation. This could help predicting the best timing to use immune checkpoint inhibitors next. As previously stated, cfDNA seems to be the only biomarker whose kinetics would allow to define this optimal window, especially with a combinatorial regimen. Indeed, peak of cell-free DNA is dependent upon the tumor itself but also by the concomitant use of other anticancer therapies such as chemotherapy, radiation therapy, but also oral targeted therapies. By discriminating large (i.e., >10,000 bp) cfDNA associated with necrosis from the short (i.e., 180 bp) ones associated with tumor apoptosis with immunogenic response via STING-cGAS, it paves the way for a qualitative and quantitative monitoring of cancer patients. *Pharmaceutics* **2020**, *12*, x 10 of 16 markers better reflecting real-time changes in the tumor micro-environment such as immunogenic cell death. Because as seen before, STING-cGAS is a critical signaling pathway associated with response to immunotherapy and that release of cfDNA, especially the shorter ones, activates this pathway, blood quantification of cfDNA could be a convenient surrogate to monitor STING-cGAS activation. This could help predicting the best timing to use immune checkpoint inhibitors next. As previously stated, cfDNA seems to be the only biomarker whose kinetics would allow to define this optimal window, especially with a combinatorial regimen. Indeed, peak of cell-free DNA is dependent upon the tumor itself but also by the concomitant use of other anticancer therapies such as chemotherapy, radiation therapy, but also oral targeted therapies. By discriminating large (i.e., >10,000 bp) cfDNA associated with necrosis from the short (i.e., 180 bp) ones associated with tumor apoptosis with immunogenic response via STING-cGAS, it paves the way for a qualitative and quantitative monitoring of cancer patients.

In addition, quantification of cfDNA is achieved by liquid biopsy, thus facilitating its implementation and allowing repeated and longitudinal measures throughout time. As a comparison, immunomonitoring provides valuable information on activated T lymphocytes or T Regs (i.e., to assess the impact of chemotherapy on the immune system,) but this can only be done after tumor biopsy, thus limiting its repeated use in routine patients. In addition, quantification of cfDNA is achieved by liquid biopsy, thus facilitating its implementation and allowing repeated and longitudinal measures throughout time. As a comparison, immunomonitoring provides valuable information on activated T lymphocytes or T Regs (i.e., to assess the impact of chemotherapy on the immune system,) but this can only be done after tumor biopsy, thus limiting its repeated use in routine patients.

To date, current biomarkers such as TMB, MSIH, or PD-L1 expression levels are mostly used prior to therapy in a binary, Go/No-Go, fashion. Indeed, despite the current vagueness for defining positivity threshold, PDL1 expression is a green light for using immune checkpoint inhibitors, as MSIH stats for pembrolizumab (Figure 3). Once treatment has started, monitoring these markers does not allow tuning, nor dosing, scheduling, or sequencing should combinatorial regimen be administrated, and evaluation for response is performed several cycles later. To date, current biomarkers such as TMB, MSIH, or PD-L1 expression levels are mostly used prior to therapy in a binary, Go/No-Go, fashion. Indeed, despite the current vagueness for defining positivity threshold, PDL1 expression is a green light for using immune checkpoint inhibitors, as MSIH stats for pembrolizumab (Figure 3). Once treatment has started, monitoring these markers does not allow tuning, nor dosing, scheduling, or sequencing should combinatorial regimen be administrated, and evaluation for response is performed several cycles later.

**Figure 3.** Current use of biomarker prior to setting up combinatorial immunotherapy. Upfront testing helps to determine the Go/No Go by predicting the odds of success. However, no longitudinal monitoring is currently feasible and basal levels have to be considered as granted. **Figure 3.** Current use of biomarker prior to setting up combinatorial immunotherapy. Upfront testing helps to determine the Go/No Go by predicting the odds of success. However, no longitudinal monitoring is currently feasible and basal levels have to be considered as granted.

Conversely, monitoring cfDNA is a dynamic strategy. The cfDNA can also be a Go to start immunotherapy. In case of absence of a DNA peak, this calls for starting for treatments such as neoadjuvant chemotherapy or radiation therapy expected to trigger immunogenic cell death monitoring cfDNA could thus help to determine the best time-window to start immunotherapy with

developing novel strategies to optimize treatment efficacy and treatment cost-effectiveness is now

Conversely, monitoring cfDNA is a dynamic strategy. The cfDNA can also be a Go to start immunotherapy. In case of absence of a DNA peak, this calls for starting for treatments such as neo-adjuvant chemotherapy or radiation therapy expected to trigger immunogenic cell death—monitoring cfDNA could thus help to determine the best time-window to start immunotherapy with respect to changes in tumor immunity (Figure 4). By doing so, the current concomitance of treatments should not be longer the rule as immune checkpoint inhibitors administration would be more wisely guided by a real-time biomarker. With respect to the ever-increasing drug-costs in oncology, developing novel strategies to optimize treatment efficacy and treatment cost-effectiveness is now critical. Rather than relying on a basal biomarker, longitudinal monitoring would allow fine tuning of the therapy, i.e., by stopping a therapy which is doomed to fail, before imaging reveals treatment failure. Unlike costly pan-genomic analysis of tumors or microbiota, cfDNA could be a cheap, rapid, non-invasive, and convenient way to check whether immunotherapy is likely to yield clinical benefit or not. *Pharmaceutics* **2020**, *12*, x 11 of 16 critical. Rather than relying on a basal biomarker, longitudinal monitoring would allow fine tuning of the therapy, i.e., by stopping a therapy which is doomed to fail, before imaging reveals treatment failure. Unlike costly pan-genomic analysis of tumors or microbiota, cfDNA could be a cheap, rapid, non-invasive, and convenient way to check whether immunotherapy is likely to yield clinical benefit or not.

**Figure 4.** Proposed strategy for refining combinatorial immunotherapy. After that standard treatment is given, longitudinal monitoring of cell-free DNA helps to determine the best timing for further administering immune checkpoint inhibitors. Rather than pre-defined dosing, longitudinal monitoring allows customized treatment throughout time. **Figure 4.** Proposed strategy for refining combinatorial immunotherapy. After that standard treatment is given, longitudinal monitoring of cell-free DNA helps to determine the best timing for further administering immune checkpoint inhibitors. Rather than pre-defined dosing, longitudinal monitoring allows customized treatment throughout time.

**Author Contributions:** G.S., F.F., R.F., F.B. and J.C. collected data and wrote the manuscript. G.S. and J.C. prepared the figures. **Author Contributions:** G.S., F.F., R.F., F.B. and J.C. collected data and wrote the manuscript. G.S. and J.C. prepared the figures. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by A\*MIDEX grant "SMART project" (Aix Marseille Univ France) P.I.'ed by Fabrice Barlesi. **Funding:** This work was funded by A\*MIDEX grant "SMART project" (Aix Marseille Univ France) P.I.'ed by Fabrice Barlesi.

1. Barlési, F.; Scherpereel, A.; Rittmeyer, A.; Pazzola, A.; Tur, N.F.; Kim, J.-H.; Ahn, M.-J.; Aerts, J.G.; Gorbunova, V.; Vikström, A.; et al. Randomized Phase III Trial of Maintenance Bevacizumab With or Without Pemetrexed After First-Line Induction With Bevacizumab, Cisplatin, and Pemetrexed in Advanced Nonsquamous Non–Small-Cell Lung Cancer: AVAPERL (MO22089). *J. Clin. Oncol.* **2013**, *31*,

**Conflicts of Interest:** The authors declare no conflict of interest. **Conflicts of Interest:** The authors declare no conflict of interest.

4. Ledford, H. The perfect blend. *Nature* **2016**, *532*, 162–165.

3004–3011, doi:10.1200/jco.2012.42.3749.

**References** 

cancer treatment. *J. Immunother. Cancer* **2019**, *7*, 129, doi:10.1186/s40425-019-0594-0.

Ferris, A.E.; Schilsky, R.L.; et al. The promise of Immuno-oncology: Implications for defining the value of
