Randomly placed pillars on a chip make tumour cell clusters break apart more often

The experiment turns on one variable. Cancer cells were placed at the centre of a pillar forest and tracked with fluorescence imaging and light microscopy for up to eight days ^[7]. In some designs the pillars sat on a square grid, repeating every 18 micrometres from centre to centre ^[16]. In the others, each pillar was shifted off that grid by a random angle and distance ^[4]. The pillars stand in for the tissue fibres that invading cells have to squeeze between ^[3].

Building the disorder took new tools. "There were no suitable tools for designing mathematically defined disorder," said Saren Tasciyan, who designed the devices during his PhD in Michael Sixt's group at ISTA ^[6]. He wrote software to define the patterns and turn them into designs manufactured with standard semiconductor lithography ^[6].

The clusters were more likely to break apart in the disordered forests than on the regular grid ^[8]. In those conditions, Sixt said, "we found that the first cells detach, exactly as happens in a cancer metastasis" ^[9]. The pillars roughened the advancing edge of the cluster into finger-like protrusions, and cells were more likely to leave from the tips ^[10]. "When the cell interface moves through this landscape the effects accumulate over time," Sixt said ^[11].

The Physics World account does not say how much more often the clusters split, or which cancer cells were used. The full study appeared in Science Advances ^[5].

Cells on the chip lasted eight days, so Edouard Hannezo's group at ISTA simulated the process over longer periods than the cells could survive ^{[1] [12]}. "Each cell was modelled as a bead that moves through an environment," said Zuzana Dunajová, the Hannezo group postdoc who built the model ^[13]. The beads attract one another and move through geometries that mimic the devices ^[13]. I think this is the most informative part of the work. As described, the model has cohesion and movement and nothing about genes, so whatever detachment it shows comes from geometry ^[17]. It still shed beads more often among disordered pillars: "the beads detached from the collective more often in the disordered environment compared to the ordered," Dunajová said ^[14].

She went further. "The roughness of the cancer-cell invasion actually follows the same kind of universal behaviour as we see in other systems, such as combustion of paper, spreading of fires or drying coffee drops," she said ^[18]. The claim concerns the statistics of a moving front. Physics World did not report the measurements behind it.

The chip cannot show whether a tumour in a body behaves the same way. The local environment that governs detachment includes blood vessels, immune cells and signalling molecules as well as connective tissue ^[2]. The chip isolates one item on that list, the geometry of the fibres, with pillars standing in for them ^[3].