Fabrication of Three-Dimensionally Addressable Microfluidic Plumbing by Stereolithographic 3D Printing

Abstract

Cell cultures thicker than a few hundred micrometers are hard to feed, because nothing circulates inside them. They are also hard to analyze without destroying them, and hard to correct once the cells are inside. Our group previously answered those three problems for flat cultures with an automated microfluidic device in which a computer opens any one of a grid of ports on its own, an ability we call addressable, to deliver or withdraw cells and chemicals at chosen locations inside one living culture. Here we extend that addressable plumbing to a three-dimensional scaffold, a lattice of culture chambers, designed to be built in one piece by 3D printing. Every chamber has an inlet and an outlet channel on each of the three axes, and those channels are the chamber's ports. Each inlet-outlet pair carries a valve at the chamber's mouth on either side and a bypass channel that joins the two channels around the chamber. Closing the two valves on an axis sends flow along that axis through the bypass, and opening them admits it into the chamber. Fluid carrying cells or chemicals can therefore travel along a row of chambers to any one of them without entering, or disturbing, the chambers before it. Prior addressable layouts either confine such transfers to one row or leave every other column of chambers empty to serve as the detour. We also place valves on the bypasses themselves. With a bypass closed and a chamber opened toward its neighbor, the two chambers exchange molecules and cells by diffusion while the rest of the lattice stays isolated, so one device can run as a single interconnected culture, as isolated wells, or as isolated sub-volumes. A combinatorial multiplexer, a valve matrix that opens one of many lines from a few control lines, is designed to drive the lattice's valve lines from a handful of external pneumatic solenoids. We then printed the building block of the lattice, one port with its valves, as a single part on a commercial stereolithographic printer, and then one complete unit module as a single cube. Two of the block's valves stand in orientations that, to our knowledge, no one had printed before: a valve whose membrane is printed standing vertically rather than flat on the build plate, and a valve whose control pressure arrives perpendicular to the membrane rather than along it. We expected each to fail, the first by splitting along the seams between print layers and the second by tearing at the center of the membrane, and flow simulations confirmed the second cause. We removed both causes by geometry: the channel that brings the control pressure now enters at the sides of the membrane rather than its center, and a reinforcement whose layer seams are offset from the membrane's backs the vertical membrane. Valves printed with those changes sealed and opened under pneumatic control. The scaffold is built for patterned co-cultures of osteoblasts and chondrocytes monitored through their own ports, and we will report that application in a later version of this preprint.