Cells keep a record of the stiffness they came from, and NUS names the switch: NFATC2

In some confined settings the soft-primed cells were more than 60% faster ^[5], which is a rate. Put it in the units an experimentalist actually watches and a stiff-primed cell needs roughly 1.6 times as long to clear the same channel ^[12].

The timing is the part worth holding on to. Priming ran for several days on soft or stiff hydrogels, and the migration measurement happened afterwards, once the cells had been lifted off and placed on a common surface or into microchannel devices that mimic tight tissue spaces ^[4]. The gel was no longer there to be sensed. "Cells are usually studied based on where they are at a given moment, but our findings show that where they have been is also important," said Andrew Holle, the NUS biomedical engineering academic who led the work ^[9].

The direction of the effect is the useful surprise. Stiff surroundings are usually linked to stronger pulling forces, which help a cell haul itself across a flat surface; getting through a narrow gap instead rewards deformability and looser anchoring ^[6]. So the sign of a cell's mechanical history depends on the test you run it in: the same stiff-primed population that looks capable in a flat traction assay is the slower one in a channel ^[14].

The three cell types did not agree. Fibroblasts and fibrosarcoma cells carried the priming forward ^[5]. The highly invasive breast cancer cells registered stiffness while they were still on the original surface and then did not keep the stiffness-dependent behaviour after transfer ^[7]; the researchers read this as different strategies, with some cells retaining a memory and the most invasive ones prioritising flexibility in surroundings that change fast ^[7].

RNA sequencing after priming is where the molecular handle appears. NFATC2, a member of the NFAT family of transcription factors, stood out in exactly the cells that retained memory: strongly activated after soft priming, and relocated into the nucleus where such factors act on gene expression ^[8][2]. Nuclear localisation plus a matching transcriptional signature is a correlation, and the summary account we have breaks off at the sentence describing what happened when the team disrupted NFATC2 activity ^[13]. That result is the load-bearing one, and it sits in the Cell Reports paper rather than in the release ^[3].

Doctoral student Nicole Lee Jia Wen, the first author, framed the design plainly: the microchannels asked what the cell still remembers once its original surface is taken away ^[10]. The answer, for two of three cell types, is enough to change how fast it gets through a gap.