(P 326) CELL-FREE DNA AS A BIOMARKER FOR CELL DEATH AND TREATMENT RESPONSE IN ON-CHIP THREE-DIMENSIONAL SARCOMA TUMOR MODELS

Objective: On-chip 3D tumor models have the potential to predict treatment efficacy and guide clinical management because of their high fidelity and throughput efficiency. However, a major bottleneck to their widespread adoption is the lack of reproducible and versatile real-time measurement of cell death. Cell-free DNA (cfDNA) is a promising non-invasive biomarker for monitoring cell death and treatment response in cancer. Human plasma cfDNA is a novel biomarker being assessed for its ability to diagnose and follow patients affected by a variety of cancers, including sarcoma. The objective of this study is to evaluate cfDNA’s potential in predicting cell viability and treatment response for on-chip 3D tumor models. To achieve this, we quantified cfDNA release from treated and untreated spheroids and micro-dissected tumor tissues (MDTs, i.e., ex vivo 3D models with radius of ~300-400µm) derived from mouse sarcoma xenografts and human primary sarcoma tissues cultured on microfluidic chips.

Methods: We utilized four different sarcoma models: spheroids from SK-LMS-1 leiomyosarcoma cell line, MDTs derived from SK-LMS-1 and MCA-205 fibrosarcoma mouse xenografts, and primary sarcoma MDTs from two patients (MDT22: low-grade leiomyosarcoma, MDT23: high-grade undifferentiated soft tissue sarcoma). All models were cultured on specific microfluidic chips designed to maintain and manipulate them. Treatments for spheroids and mouse xenografts included staurosporine (cytotoxic alkaloid chemotherapy), radiotherapy, and/or AZD4320 anti-Bcl2/xL senolytic to assess the correlation between cfDNA levels and cell death (Figure 1.D and 2A). Chip’s supernatant was collected daily and cfDNA was quantified using quantitative PCR (qPCR) with long interspersed nuclear elements (LINE) primers. For primary sarcoma samples, cfDNA release was also correlated with cell viability determined by immunohistochemistry (IHC).

Results: Higher seeding densities in the SK-LMS-1 leiomyosarcoma spheroids resulted in larger spheroids and increased cfDNA release (Figure 1.A-C). Similarly, cfDNA release was notably elevated in channels containing more xenograft derived MDTs, particularly during the initial days (Figure 2.B-C). In other words, daily media cfDNA contents correlate with on-chip tumor burden. Both spheroids and MDTs demonstrated dose-dependent cfDNA release in response to staurosporine, peaking three days post-treatment (Figure 1.E-F, 2.D-E). Radiotherapy had varied effects: in SK-LMS-1 MDTs, cfDNA release remained unaffected after radiotherapy (data not shown), whereas a delayed and slight elevation was observed in MCA-205 MDTs (Figure 2.F-G). This could be due to radiotherapy induced senescence, i.e., a state of durable cell cycle arrest. Sequential addition of AZD4320 five days after radiotherapy, which targets radiotherapy-induced senescent cells, significantly increased cfDNA release in MDTs from both cell lines. (Figure 2.F-G).
Finally, the level of cfDNA in the supernatant of the chips containing human sarcoma MDTs (MDT22 and MDT23) correlated well with cell death progression as assessed by IHC (Figure 3); MDT22 survived well in the chip in the first week with small release of cfDNA. On the other hand, MDT23 showed a significant release of cfDNA in the first few days, signifying considerable amount of cell death which correlated with low viability of MDTs even during the first week, as is apparent in IHC images.

Conclusion: Our study demonstrated a strong correlation between cfDNA release and cell viability, highlighting cfDNA as a reliable biomarker for monitoring cell death and treatment response in various 3D tumor models. The variability in cfDNA response to different treatments underscores the importance of context-specific analyses for accurate biomarker assessment. Future studies should explore the mechanisms behind cfDNA release discrepancies and differentiate between tumor and non-tumor cfDNA to estimate therapeutic windows.