1.1.

Distinct Routes of BL-OG Activation Provide Distinct Activation Profiles for ChP Study

While LMO3 has been effectively activated in vivo in a variety of neural cell types and at different developmental time points, this construct has not been tested in ChP epithelium, and differences resulting from the local milieu (e.g., pH) or cell-type specific expression, could render its bioluminescent recruitment inactive. To test its functionality, and determine the time course and relative amplitude of LMO3 bioluminescent activation depending on the method of in vivo CTZ injection, we compared three different routes of delivery: intraperitoneal (IP), intravenous (IV), and intracerebroventricular (ICV) injection. Coelenterazine (CTZ) was administered to FoxJ1-Cre:LSL-LMO3-eYFP mice, which show strong expression in ChP epithelial cells.^3,29,34 In all testing, mice were implanted with an imaging cannula³ that allows optical access to the left lateral ventricle and ChP. We recorded ChP bioluminescent output through the canula in anesthetized mice (isofluorane .5-1%) using an Andor iXon Ultra 888 EMCCD with a 5X, 0.14 NA objective (Mitutoyo). Because the imaging cannula precluded direct vertical placement of the ICV delivery cannula into the same lateral ventricle, we targeted the injection cannula to the contralateral lateral ventricle by an oblique trajectory through the dorsal posterior cortical surface [diagrammed in Fig. 1(a) and described in the Appendix (Sec. 3)].

Fig. 1

BL-OG in ChP enables multiple time courses of chemogenetic control. (a) Diagram of visualization area. (b) Example of BioLuminescent photon output with ICV luciferin delivery. (c) Example of distinct photon output in the same mouse with two different luciferin injections. (d) Individual time courses for IV (fast), IP (sustained), and ICV BioLuminescent activation. (e) Mean BioLuminescent activation in all conditions. (F) Comparison of IV and IP activation metrics in 3 mice.

The three modes of administration, ICV, IV, and IP, each provided a distinct pattern and timing of response onset. With ICV administration, bioluminescent photon output began 10 to 30 s following injection onset and progressed from center outward, in a pattern spatially consistent with a route of CTZ diffusion from the contralateral ventricle, through the third ventricle, to the ChP in the imaging field. Activated tissue continued to peak bioluminescence over tens of seconds and then faded slowly to the minimal observed effect [Fig. 1(b)]. To test the impact of rate and volume of CTZ injections on ChP bioluminescence, in one mouse we injected CTZ, $25\mu \mathrm{l}$ over 60 s, and $10\mu \mathrm{l}$ over 30 s, into the lateral ventricle. A fourfold difference in magnitude between peak bioluminescent response was observed between the two conditions [Fig. 1(c), light versus dark red curves]. A bioluminescent response of similar amplitude and time constant to that following the $10\mu \mathrm{l}$ to 30 s injection was produced in a second animal from a trial with a $5\mu \mathrm{L}$ to 20 s ICV injection [Fig. 1(d), bottom green trace].

Prior studies have shown that IV luciferin injections drive a faster rise time and higher peak photon production than IP injections.^35–37 We next tested whether these principles hold for BL-OG in the unique circulatory and chemical environment of the ventricle.³⁸ We imaged anesthetized LMO3 in ChP mice following IV or IP CTZ injection ($N=3$ mice). Figure 1(d) (left) shows responses normalized as percent of maximal emission for trials from each animal in IP or IV injection conditions and a plot from an additional animal in an ICV trial for comparison (bottom). Initial bioluminescent output onsets $\sim 10\mathrm{s}$ following the start of IV injection [Fig. 1(d), gold]. In contrast, initial activity after IP injection was significantly slower in onset and rise time, and varied widely in ascent, on the scale of hundreds of seconds, as observed in prior studies (purple).³⁷ Also, unlike ICV or IV injections, where response curves comprised a rapid initial climb to peak and then a smooth exponential decay, bioluminescent output curves resulting from IP injections demonstrated local sub-peaks and troughs in photon production in all mice [see Figs. 1(d) and 1(e), purple].

Figure 1(e) shows the difference in temporal progression of the response between IP, ICV, and IV trials, plotted as the average percent of maximal response across subjects for each condition. The blue and red dashed expansions show the relative time course onset for the three injection types, showing IV onset within seconds of injection, an intermediate rise time for ICV injection, and the slow onset and rise time with IP injection. While the IP response begins later, it persists much longer than the other administration routes, on the scale of tens of thousands of seconds: we did not observe a return to baseline within any IP imaging session. To quantify these temporal response differences, we calculated the time-to-peak for each animal across trial types. In all cases, responses to IV injection peaked $<25\mathrm{s}$ after onset, whereas IP injection responses peaked from 2000 to $>4000\mathrm{s}$ after onset [Fig. 1(f), left]. Considering the post-peak half-maximal response, all IV trials decayed to half-maximal $<20\mathrm{s}$ after maximal effect, whereas IP trials decayed to half-maximal from 2000 to 6000 s.

To compare response magnitude, we normalized data across trial-types for imaging exposure (0.5 Hz IV, 0.33 Hz IP frame rate), and across individuals for small differences in mg/Kg of CTZ introduced by our standard injection in an individual. We found $\sim 100$-fold higher peak emission on IV versus IP trials. To describe the differences in total light emission across trial types, we integrated normalized bioluminescence responses from injection to peak-effect timepoints. Although IP responses to CTZ injection peaked and persisted considerably longer than IV responses, integrated responses from IV injections showed one to two orders of magnitude greater total output [Fig. 1(F)].

1.2.

Repeated CTZ Administration Does Not Impact Basic Health Indices

As a further enabling step in the use of BL-OG for ChP research, we tested whether repeated CTZ administration or LMO3 expression in the ChP impacts basic indices of health or behavior. To obtain systematic video for analysis, we designed and laser cut de novo plexiglass cage tops for Thoren split cages, using novel three-dimensional (3D) printed clamps. Video cameras were held on adjacent tripods, one split cage/camera. Data were taken with Raspberry Pi 4B 2G MakerFocus Night Vision Cameras with Adjustable Focus (5MP HD Wide Angle Fisheye Lens OV5647 1080p; two cameras had a mechanically shuttered IR filter). Mice were kept on a reverse light:dark cycle. Head, body, and tail identification points were trained in DeepLabCut to analyze mouse location, movement, and orientation and then placed into real space through a custom-written Cartesian transformation.³⁹ Figure 2 shows a sample image obtained in the dark cycle (ruler: 6 in./15 cm).

Fig. 2

Example frame from video obtained for multi-mouse automated assessment of the impact of 1 week of CTZ injections. Please see text for further details.

Mice ($N=12$) were monitored for seven days following daily intraperitoneal (IP) injection of CTZ ($75\mu \mathrm{l}$ on d1-4, $100\mu \mathrm{l}$ on d5-6, and $110\mu \mathrm{l}$ on d7 at 2.36 mM in saline; prolume coelenterazine in vivo) or vehicle control. Double-transgenic mice were generated by crossing the FoxJ1-Cre and LSL-LMO3 (“LMO3 in ChP” mice: $N=6$; Jax Laboratories Strain #: 034853; see Ref. 21 for details). The remaining $N=6$ were C57/Bl6 mice without LMO3 expression (non-LMO3). For each group, half ($N=3$) were assigned to receive either CTZ or vehicle (LMO3 in ChP given CTZ, one female; control CTZ, 1F; LMO3 in ChP vehicle, 2F; control vehicle, 2F). Each mouse was weighed and checked for signs of discomfort or lethargy daily, video monitored in their home cages and housed under a 12/12 LD cycle to record estimates of wakefulness, movement, and drinking behaviors.

Across d1-7 both mice in the CTZ and vehicle group showed robust health. No mice showed signs of discomfort (e.g., porphyrin secretions), diminished grooming, or failed startle responses. Across days, mice in the CTZ and vehicle groups maintained stable body weights (Friedman rank sum tests; CTZ ${\chi }^2=10.21$, $p>0.05$; vehicle ${\chi }^2=13.16$, $p<0.05$, all Bonferroni corrected pairwise comparisons $p>0.05$; Fig. 3). No significant differences in total weight change, day 1 to day 7, were found between luciferin groups (CTZ versus vehicle) nor between genotype groups (ChP-LMO3 versus non-LMO3; Wilcoxon Mann-Whitney test, $p>0.05$; Fig. 3).

Fig. 3

Repeated CTZ administration and transgenic expression of LMO3 in ChP did not impact weight across days. Across several metrics, mice did not show any adverse effect of CTZ administration or expression of LMO3 in the ChP. Please see accompanying text for details.

Mice in all groups displayed normal phasic home cage behaviors, as assessed by observation and quantified by videography obtained on days 4 to 7. During the light (resting) phase of the circadian cycle, mice in all groups initiated more sleep bouts per monitored hour, and slept for more total time, than in the dark (waking) phase (light $M=9.35$, $\mathrm{SD}=0.85$, dark $M=5.60$, $\mathrm{SD}=0.58$, Wilcoxon rank sum test $n=12$, $p<0.05$). Mice also initiated more movement bouts in the dark cycle (light $M=526.05$, $\mathrm{SD}=53.85$, dark $M=754.30$, $\mathrm{SD}=48.31$, Wilcoxon rank sum test $n=12$, $p<0.05$), moved at greater average velocity per bout (light $M=1.21$, $\mathrm{SD}=0.06$, dark $M=1.54$, $\mathrm{SD}=0.23\mathrm{cm}$ per second, Wilcoxon rank sum test $n=12$, $p<0.05$). All these behaviors were stable and did not drift significantly day-to-day.

No effects were observed between CTZ and vehicle in sleep or movement metrics in the lights on or in the lights off epochs (Wilcoxon rank sum test, $p>0.05$ for all comparisons). For LMO3 in ChP versus non-ChP mice, no difference was found in sleeping bouts in the lights off epoch (Wilcoxon rank sum test, $p>0.05$). We did observe a modest, if significant, difference in sleep bouts in the lights on condition, with LMO3 in ChP showing a greater number ($p=0.026$; LMO3 in ChP $9.85\pm 0.83$ versus non-ChP $8.86\pm 0.53$, mean ± standard deviation). However, this finding was not observed in the other metrics of sleep persistence, as there were no difference observed in sleep time or movement bouts between these groups in these epochs (Wilcoxon rank sum test, $p>0.05$ for all comparisons).

3.1.

In Vivo Bioluminescent Imaging

Mice were initially anesthetized with 2% isoflurane in oxygen delivered at $1.5\mathrm{l}/\min$ and then moved into a light-tight enclosure for imaging. Animals were secured in the imaging environment by clamping their head-bar to steel posts (Thor Labs) and isoflurane concentration was reduced to 0.5% to 1% to maintain anesthesia for the course of the experiment. An Andor iXon Ultra EM-CCD camera with a 5×, 0.14 NA objective (Mitutoyo) was focused on the ChP through the animal’s imaging cannula and window. Images were viewed and recorded with Andor Solis software. For all experiments, we performed injections of water-soluble CTZ (Prolume) diluted in deionized water (2.36 mM). Injections of CTZ were delivered by a Quintessential Stereotaxic Injector syringe pump (Stoeling). For ICV injections, CTZ was loaded into a syringe attached to catheter tubing terminating in a 26-gauge cannula (WPI). ICV caps were removed from the head-restrained, anesthetized mouse, and the 26-gauge cannula was gently threaded into the ICV cannula. For all IP and IV injections, CTZ-loaded syringes were attached to catheters terminating in a 29 gauge needle (Instech). The needle was used to cannulate either the lateral tail-vein (IV) or peritoneum (IP) and was then held in place with surgical tape.

3.2.

Imaging Bioluminescence During ICV Injection of CTZ

Imaging data were recorded from two mice with ICV implants for CTZ administration. Baseline images were acquired for a duration of 2 min. Images were recorded at a rate of 0.1 or 0.5 Hz in the first and second mouse, respectively. After baseline recording, controlled injection of CTZ was initiated. Post-injection, images were recorded for an additional 33 min. Following recording, anesthesia was discontinued, mice regained consciousness and were returned to their home cage. See text above for other relevant information.

3.3.

Imaging Bioluminescence During IV and IP Injection of CTZ

Imaging data were recorded from three mice, each of which underwent IP and IV injections of CTZ on separate days. Images were acquired at a rate of 0.5 Hz for IV trials and 0.33 Hz for IP trials, respectively. Baseline images were acquired for 2 min (IV trials) or 3 min (IV trials). Following baseline recording, controlled injection of CTZ was initiated. $75\mu \mathrm{L}$ of CTZ was injected over a 10 s duration in all IV trials. For IP trials, injections of 310, 350, or $400\mu \mathrm{L}$ were delivered over a 10 s duration. Following injection, images were recorded for an additional 4.03 h. Following recording, anesthesia was discontinued, mice regained consciousness, and were returned to their home cage.

3.4.

Analyses of Bioluminescence Imaging Data

Images were analyzed using Matlab (Mathworks). For all trials, regions of interest (ROIs) were created to assess overall bioluminescence emanating from the ChP by tracing a mean projection image [Fig. 1(b), left panel]. Pixel values recorded before exposure to CTZ (zero bioluminescence): Those in the ChP were selected as in the top 93% of intensity to form a mask baseline ROI. Difference images were created by subtracting the ROI image at each time point from this baseline ROI image. Pixels from each difference image were averaged to create a timeseries of changes in bioluminescence from baseline for each trial, in arbitrary units.