A 96-well valved microfluidic device for testing of live intact tumor cuboids

Introduction:: Drug discovery and development are vastly inefficient. For more than a century, biologists have performed pre-clinical drug tests in Petri dishes and animal models which lack recapitulation of the tumor microenvironment (TME).¹ As a result, drugs that pass pre-clinical tests are not necessarily effective in the context of the human TME. Annually, a dismal < 4% of cancer drugs in clinical trials pass safety and efficacy tests,² each requiring >10 years and >$1 billion for market entry.³ Functional drug testing technologies such as patient-derived organoids⁴ have emerged as promising new approaches, but these ex vivo models fail to preserve the patient’s original, unique TME and suffer from in vitro cultivation bias. A different paradigm for cancer drug development that preserves the human TME is critically needed to keep pace with emerging therapies and realize more affordable treatments. Microdissected tumors preserve the TME and maintain similar manipulation simplicity as small organoids in microsystems, but experimental attempts to date have employed low throughput methods for analysis.^5,6 To address the above limitations, we developed a microfluidic platform for drug development and precision medicine based on regularly-sized, cuboidal-shaped microdissected tumors (referred to as “cuboids”) that are mechanically cut.⁷ Here we present a multi-well platform that hydrodynamically traps arrays of cuboids in 96 wells (384 cuboid traps, with 4 traps per well) for selective treatment and prevents cross-contamination between wells via pressure-controlled thermoplastic polyurethane (TPU) valves.


Materials and Methods:: Our 96-well microfluidic platform consisted of 5 thermoplastic layers bonded with a combination of solvent-bonding,⁸ lamination, and thermal vacuum bonding⁹ (Fig. 1a-b). The primary 0.5 mm-thick polymethylmethacrylate (PMMA, Astra Products Inc.) microfluidic layer was engraved using CNC milling. This design comprised 4 identical quadrants of bifurcating channels which propagated negative pressure from single inlets to 384 traps arrayed in a 96-well arrangement (Fig. 1c-e), creating hydrodynamic suction that drew cuboids to trap positions (Fig. 1f-h). To prevent well-to-well crosstalk through microchannels, we incorporated microvalves using flexible thermoplastic polyurethane (TPU) film (PT9200US NAT, Covestro LLC, USA) sandwiched between two PMMA layers (Fig. 2a-d). TPU covered PMMA half-cylindrical valve seats via lamination (Sky, 325R6). Then, a thermal vacuum process permanently bonded the sandwich after alignment with the pneumatic control layer.⁹ Last, we attached a 6.40 mm-thick PMMA loading frame via adhesive (300LSE, 3M).
Live cuboids of (400 mm)³ size were prepared from U87 glioblastoma multiforme patient-derived xenograft mouse models via mechanical tissue chopper as previously described.⁷ We filled the frame with a suspension of cuboids in 20% polyethylene glycol (8k M.W.) in PBS (PEG/PBS) then connected pneumatic valve and fluidic suction control lines. PEG/PBS had sufficient viscosity to transmit fluid momentum to floating cuboids, enabling hydrodynamic trapping. After sequentially arraying cuboids in each quadrant, we removed PEG/PBS, dispensed rat tail collagen inside each trap, and adhered a bottomless 96-well plate. +4psi valve pressure isolated the wells before we fluorescently stained cuboids with Cell Tracker Orange (CTO), Cell Tracker Green (CTG), and Hoechst.


Results, Conclusions, and Discussions:: To evaluate whether the valves prevented crosstalk between well conditions, we performed fluorescent staining of live cuboids in a full device. We applied a repeating quad pattern of Cell Tracker Orange (CTO), Cell Tracker Green (CTG), Hoechst, or dual CTO+CTG to 4 adjacent wells in close fluidic connection with each other (Fig. 3). After application of +4 psi to close the valves, we applied staining solution to each well and placed the device in a cell culture incubator for 1.5 hrs, maintaining positive pressure to the valve control inlet. Next, we imaged the cuboids after replacing the staining solutions with Live Cell Imaging Solution to stop labeling. A brightfield image of the complete 96-well device with an overlay of the fluorescent cuboids (Fig. 3a) indicates successful staining of living cuboids. The same fluorescent image without brightfield highlights the differential labeling (Figure 3b). Had the valves leaked, we would have expected to see unintended staining in adjacent wells. However, cuboids in each well maintained the expected pattern of fluorescence despite sharing fluidic channels. These results indicated adequate valve closure for the duration of the assay, successfully preventing crosstalk between well conditions. These results support the potential of our 96-well-based device to perform multi-treatment assays on live tumor tissue with high throughput.
Microfluidic approaches can empower the application of functional drug testing to microdissected tumor cuboids that preserve the tumor microenvironment (TME). Our 96-well valved microfluidic device simplified the organization of hundreds of submillimeter-sized samples of tumor tissue in defined arrays for application of multiple different treatments of proof-of-concept live-cell labeling. We also showed the scalability of our manufacturing process to produce 96 completely thermoplastic microvalves that closed using low positive pressures (< 4 psi), a design which may prove useful for other applications. Unlike our platform, present approaches for intact tissue testing have been limited thus far by low-throughput methods. Creating tools that simplify preclinical assays on microdissected tumors which retain the human TME (e.g., our cuboids) could have an enormous impact on the success rates and overall costs of cancer drug development.




Acknowledgements (Optional): : This work was partially supported by NCI R21CA251952, a pilot grant from Juno Therapeutics, and a Gap Innovation Fund grant from UW CoMotion.

References (Optional): :

1. 
   Hanahan, D. & Coussens, L. M. Accessories to the Crime: Functions of Cells Recruited to the Tumor Microenvironment. Cancer Cell vol. 21 Preprint at https://doi.org/10.1016/j.ccr.2012.02.022 (2012).
   
   


2. 
   Wong, C. H., Siah, K. W. & Lo, A. W. Estimation of clinical trial success rates and related parameters. Biostatistics 20, (2019).
   
   


3. 
   Wouters, O. J., McKee, M. & Luyten, J. Estimated Research and Development Investment Needed to Bring a New Medicine to Market, 2009-2018. JAMA vol. 323 844–853 Preprint at https://doi.org/10.1001/jama.2020.1166 (2020).
   
   


4. 
   Jin, M. Z. et al. Organoids: An intermediate modeling platform in precision oncology. Cancer Lett 414, 174–180 (2018).
   
   


5. 
   Aref, A. R. et al. 3D microfluidic: Ex vivo culture of organotypic tumor spheroids to model immune checkpoint blockade. Lab Chip 18, (2018).
   
   


6. 
   Astolfi, M. et al. Micro-dissected tumor tissues on chip: an ex vivo method for drug testing and personalized therapy. Lab Chip 16, 312–325 (2016).
   
   


7. 
   Horowitz, L. F. et al. Microdissected “cuboids” for microfluidic drug testing of intact tissues. Lab Chip 21, 122–142 (2021).
   
   


8. 
   Rodriguez, A. D. et al. A microfluidic platform for functional testing of cancer drugs on intact tumor slices. Lab Chip 20, 1658–1675 (2020).
   
   


9. 
   Pourmand, A. et al. Fabrication of whole-thermoplastic normally closed microvalve, micro check valve, and micropump. Sens Actuators B Chem 262, 625–636 (2018).