To fabricate a light-responsive bioactuator, we used D. melanogaster (wild-type strain, Canton-S third instar) larvae material. We accessed the GAL4/UAS system [28] with the targeted genetic property on the cells of the D. melanogaster larvae as shown in Fig. 2. Adult male D. melanogaster, which have myocyte enhancer factor-2 (Mef2) on the GAL4 line (GAL4-Mef2) and adult virgin female D. melanogaster, which have ChR2 on the UAS tag (UAS-H134R-ChR2 [29]) were mated in a glass bottle that contained food including the photosensitizer, all-trans-retinal (ATR, Sigma-Aldrich). To obtain virgin female D. melanogaster, male and female flies must be separated within 7 h of their emergence. ATR is indispensable in the utilization of ChR2 because the insects do not generate ATR in their bodies. After mating, the females laid eggs within two days. The eggs hatched within a day; after five days, we obtained full-sized larvae just before they became pupae. This entire process was carried out at room temperature (25 °C). ChR2 expression on the larvae muscle cells was verified by observing red fluorescence from mCherry, which expresses at the same location as ChR2 (Fig. 3), using a fluorescence microscope (BZ-X710, Keyence). The success rate of the expression was 100% (n = 18).

The micro tube for the bioactuator was excised from the larva expressing ChR2 by means of dissection (Fig. 4). We cut off the head and tail with a stainless steel disposable surgical blade (Kai Medical) and drew out the tube contents. This tube consisted of cuticle and muscle tissues attached to the cuticle. Excised tubes were kept in Schneider’s Insect Medium (S0146, Sigma-Aldrich). The main nervous system of the larva was cut off, and no spontaneous contraction occurred on it. These micro tubes and their cross-sections were observed directly using a fluorescence microscope, as described in Section 2.1. The photoresponsivity of the tubes was verified by observing the contraction on the surface with blue light stimulation. The success rate of the expression was 60% (n = 5).

The D. melanogaster larvae muscle is forced to contract in response to blue light stimulation (λ = 466 nm). As the contraction of the larvae varies according to light intensity, a uniform field of illumination is important. We built our stimulation setup from several pieces of equipment. An inverted microscope (Ti-U, Nikon) with an objective lens (4× magnification, numerical aperture 0.13, Nikon), a digital mirror device (DMD) projector (DLP Light Commander, Texas Instruments), a plano-convex lens (diameter = 25 mm, f = 50 mm, Sigma Koki) and an achromatic lens (diameter = 25 mm, f = 100 mm, Sigma Koki) were assembled as shown in Fig. 5. The larvae tissues were stimulated on the microscope stage. The power density of the blue light was 2.0 mW·mm⁻² on the stage surface. The propagation speed of the stimulation area was controlled as shown in Fig. 6. Each stimulation pattern was prepared using Paint software (Microsoft Paint, Microsoft). Each image was output as a bitmap graphic. The set of images was input into control software in order to control the projector. The speed of light propagation was controlled by changing the frequency. Our system had a temporal resolution shorter than 1 ms and a spatial resolution of up to 20 μm. A movie camera (FASTCAM SA3, Photron) was attached to record the moving images.

To observe the internal flow, microbeads (i.e., polymer microspheres with the mean diameter of 99.5 cm ± 1.5 μm in water, 4310A, Duke Scientific Corporation, Lot No. 30911) were employed. The microbeads were diffused in Schneider’s Insect Medium and injected into the micro peristaltic tube with a micropipette. The peristaltic pumping action was controlled and observed with the experimental setup shown in Fig. 5.

The micro peristaltic tubes were observed using fluorescence. The average length of the tube structure was 1342 μm (n = 6). The inner diameter was 423 μm and the outer diameter was 790 μm (n = 6). The original length of the larvae was greater than 3 mm, but dissection removed the inner pressure and the entire tissue shrank. The sizes were measured under standard atmospheric pressure; however, since the tubes were soft, they could be stretched by pulling.

We captured moving images with the setup described in Section 2.3. Each stimulation pattern was composed of a combination of a black rectangle (648 × 864 pixels) and a white rectangle (42 × 864 pixels), as shown in Fig. 6. We used 36 images for a stimulation pattern. With 36 images, a frequency of 36 Hz resulted in each cycle being one second long. With this stimulation, we controlled the peristalsis of muscle tissue, as shown by the time-lapse images in Fig. 7 and in the movie provided in the Supplementary data (Movie S1). The micro tube exhibited a back-and-forth motion of contractions when the propagation direction of the stimulation was changed (Movie S2 in the Supplementary data). Without such control, the peristaltic cycles of larvae are unidirectional in nature. Therefore, this result indicated that the contraction was indeed induced by the light stimulation. By changing the speed of the propagating light stimulation while fixing the width of stimulation area at 400 μm, we examined the peristaltic pumping action by observing surface displacement (Δx), as shown in Fig. 8. The propagation speeds of the stimulation areas were adjusted from 0 to 500 μm·s⁻¹, where 0 μm·s⁻¹ meant stimulation without propagation. As the propagation speed increased, the contractile displacement became smaller, as shown in Fig. 9(a). This change probably occurs because the muscle contraction is caused by cation flow, which is induced by ChR2. In this case, ChR2 did not receive enough energy to contract the whole tissue. To obtain larger displacement on the muscle with a high propagation speed, a more powerful resource is needed.

We calculated the induced flow volume as the volume of the surface contraction. The surface displacement volume was calculated using Eq. (1), where V is the flow volume, I is the inner radius of the tube, and D is the displacement rate. The displacement rate on the tissue surface was equal to the displacement (Δx in Fig. 8) divided by the radius of the body. P is the propagation speed of the light stimulation (Fig. 8(b)).

The dimension represented by ΔV is volume change per unit time. The propagation speed that gave the largest ΔV was selected as the optimal speed.

Next, using Eq. (1), we calculated the flow rate. As shown in Fig. 9(b), the highest flow rate was obtained when the propagation speed of the light stimulation was 400 μm·s⁻¹.

We captured moving images of the tube. To control the peristaltic pumping action, we used 12 images for the stimulation wave, as shown in Fig. 10. Each stimulation pattern was composed of a combination of a black rectangle (1024 × 768 pixels) for the background and two white rectangles (75 × 768 pixels) for the light area, and was prepared using Paint software. Each image was output as a bitmap graphic. The set of images was input to control software in order to control the projector. By controlling the peristaltic pumping action of the tube, we succeeded in transporting the microbeads, as shown in Fig. 11 and the movie provided in the Supplementary data (Movie S3). When the width of the stimulation area was fixed at 400 μm and the propagation speed was fixed at 400 μm·s⁻¹, the microbeads were transported at an average speed of 120 μm·s⁻¹.

The experimental flow rate was calculated using Eq. (2), where V, I, and D are the same terms as in Eq. (1), and B is the speed of the transported microbeads (Fig. 8(b)).

For a B of 120 μm·s⁻¹, we calculated the experimental flow rate to be 6 nL·s⁻¹ using Eq. (2). This value was smaller than the value of 12 nL·s⁻¹ that was obtained from Eq. (1). One reason for this difference may be that the tube surface was not smooth.

We then compared the estimated performance with previously reported cell- or tissue-based bio-micropumps. Compared with another reported pump using a cardiomyocyte sheet (2 nL·min⁻¹) [1], our present flow rate was 180 times higher. Compared with a polydimethylsiloxane (PDMS) membrane-based pump (0.226 nL·min⁻¹) [30], our flow rate was 1800 times higher. The sizes of these pumps were different, and the power of the larvae-derived muscle tissue was significant. We obtained stable flow without the use of any artificial valves or other soft materials. Since our tube was composed of only biomaterial, there was less wasted energy consumption.