A device for separated and reversible co-culture of cardiomyocytes

Abstract

A novel technique is introduced for patterning and controllably merging two cultures of adherent cells on a microelectrode array (MEA) by separation with a removable physical barrier. The device was first demonstrated by separating two cardiomyocyte populations, which upon merging synchronized electrical activity. Next, two applications of this co-culture device are presented that demonstrate its flexibility as well as outline different metrics to analyze co-cultures. In a differential assay, the device contained two distinct cell cultures of neonatal wild-type and beta-adrenergic receptor (beta-AR) knockout cardiomyocytes and simultaneously exposed them with the beta-AR agonist isoproterenol. The beat rate and action potential amplitude from each cell type displayed different characteristic responses in both unmerged and merged states. This technique can be used to study the role of beta-receptor signaling and how the corresponding cellular response can be modulated by neighboring cells. In the second application, action potential propagation between modeled host and graft cell cultures was shown through the analysis of conduction velocity across the MEA. A co-culture of murine cardiomyocytes (host) and murine skeletal myoblasts (graft) demonstrated functional integration at the boundary, as shown by the progression of synchronous electrical activity propagating from the host into the graft cell populations. However, conduction velocity significantly decreased as the depolarization waves reached the graft region due to a mismatch of inherent cell properties that influence conduction.

Figure 1. Co-culture apparatus

(A) MEA designed for use with a standard 35-mm Petri dish and (B) the center well with the recording electrode array and larger auxiliary electrodes used for stimulation or additional recording electrodes over a larger area. (C) A co-culture wall divided the center recording array into 6 × 3 subarrays and allowed analysis of the boundary between cultures. (D) The reusable acrylic barrier bisects the ring and defines two chambers. (E) The ring is held face down in place with an accompanying support structure, consisting of a base clamped around the Petri dish, and an overhanging arm contacting the ring through a 20 gauge needle.

Figure 2

(A) HL-1 cardiomyocytes were seeded on both sides of the barrier in a standard 35-mm Petri dish and allowed to merge. Initial cell contact was observed at 25 h. (B) The same experiment was performed on a MEA, where two asynchronous sets of electrical signal were initially observed, but synchronized after merging also 25 h after removing the barrier.

Figure 3

(A) Following removal of the co-culture device, cells were divided over the surface of the MEA and exposed to 10 μM ISO. (B) A differential response between WT and DKO cultures was observed in action potential rate and amplitude as shown from two representative electrodes. WT activity displayed a bursting rhythm of contractions, whereas DKO cells did not respond due to a lack of appropriate receptors. The spaces in between WT data points indicate no activity. (C) Extracellular action potentials traces during the time indicated by the dashed boxes in (B) are shown during ISO exposure, revealing the dramatic impact of the bursting contraction behavior on the action potential amplitude.

Figure 4

(A) The beat rate response of WT and DKO co-cultures in a heterogeneous population behaved as a single syncitium with synchronous rates of contraction across six representative electrodes. (B) WT cells responded with an increase in signal amplitude to 10 μM ISO, whereas DKO cells revealed a much smaller response that was likely due to the change of beat rate in both populations. (C) Analysis across electrodes within each region were consistent, and revealed a significant (*P < 0.05, represented as a bar between significant groups) increase in WT culture signal amplitudes due to ISO exposure. Repeated trials (N = 4) displayed similar significant increases in WT culture signal amplitudes without any bursting behavior as seen in Figure 3.

Figure 5. HL-1 cardiomyocytes (host) were co-cultured with C2C12 skeletal myoblasts (graft)

(A) A representation of the MEA displays electrodes after cultures have merged on Day 2. Electrodes displaying electrical activity on Day 2 were assumed to originate from the host, and are represented by solid circles. Additional electrodes on the graft side that previously did not display activity began exhibiting action potentials in subsequent days and are represented by triangles. (B) Activity from the highlighted electrodes are displayed for Day 4, showing a difference in amplitude between cultures, but still synchronous behavior (N = 5). (C) Conduction analysis on both sides on Day 4 indicated that electrical activity originated from the host, and experienced a significant (P < 0.05) decrease in conduction velocity on the graft side, as averaged across five beats.