Optimizing anesthesia and delivery approaches for dosing into lungs of mice

Microbes, toxins, therapeutics and cells are often instilled into lungs of mice to model diseases and test experimental interventions. Consistent pulmonary delivery is critical for experimental power and reproducibility, but we observed variation in outcomes between handlers using different anesthetic approaches for intranasal dosing into mice. We therefore used a radiotracer to quantify lung delivery after intranasal dosing under inhalational (isoflurane) versus injectable (ketamine/xylazine) anesthesia in C57BL/6 mice. We found that ketamine/xylazine anesthesia resulted in delivery of a greater proportion (52±9%) of an intranasal dose to lungs relative to isoflurane anesthesia (30±15%). This difference in pulmonary dose delivery altered key outcomes in a model of viral pneumonia, with mice anesthetized with ketamine/xylazine for intranasal infection with influenza A virus developing worse lung pathology and more consistently losing body weight relative to control animals randomized to isoflurane anesthesia. Pulmonary dosing efficiency through oropharyngeal aspiration was not affected by anesthetic method and resulted in delivery of 63±8% of dose to lungs, and a non-surgical intratracheal dosing approach further increased lung delivery to 92±6% of dose. We conclude that anesthetic approach and dosing route can impact pulmonary dosing efficiency. These factors should be considered when planning and reporting studies involving delivery of fluids to lungs of mice.

resulted in delivery of a greater proportion (52±9%) of an intranasal dose to lungs relative to 23 isoflurane anesthesia (30±15%). This difference in pulmonary dose delivery altered key 24 outcomes in a model of viral pneumonia, with mice anesthetized with ketamine/xylazine for 25 intranasal infection with influenza A virus developing worse lung pathology and more 26 consistently losing body weight relative to control animals randomized to isoflurane anesthesia. 27 Pulmonary dosing efficiency through oropharyngeal aspiration was not affected by anesthetic 28 method and resulted in delivery of 63±8% of dose to lungs, and a non-surgical intratracheal 29 dosing approach further increased lung delivery to 92±6% of dose. We conclude that anesthetic 30 approach and dosing route can impact pulmonary dosing efficiency. These factors should be 31 considered when planning and reporting studies involving delivery of fluids to lungs of mice. 32

Introduction 33
Studies investigating lung infections, lung injury, allergic airway inflammation, lung fibrosis, lung 34 cancer, and lung stem cell biology often require delivery of experimental agents to lungs of 35 mice. Administration routes for bolus dosing of fluids into lungs include intranasal (i.n.) dosing, 36 intratracheal (i.t.) dosing and dosing through oropharyngeal aspiration (o.a.). Choice of dosing 37 route is an important decision in study design as experimental outcomes can be altered by the 38 quantity of dose delivered to lungs or to extrapulmonary tissues. Different dosing routes also 39 vary in anesthetic requirements, invasiveness, and technical difficulty. 40 To guide experimental approach, a previous study assessed the effect of various factors 41 including type of anesthetic on the distribution of i.n. doses into BALB/c mice. This study 42 concluded that either injectable (Avertin) or inhaled (isoflurane, halothane) anesthetics resulted 43 in similar delivery to lungs (1). Since this influential report, several factors have changed. Safety 44 concerns have led to a decline in use of both Avertin and halothane (2, 3). Increased availability 45 of knockouts and transgenics on the C57BL/6 (B6) background has led to B6 mice becoming 46 the most widely used laboratory strain. Additionally, minimally invasive approaches for dosing 47 via o.a. and i.t. routes have been developed which can more efficiently deliver fluids to lungs 48 relative to i.n. dosing (4-6). 49 In our previous work we noticed that i.n. doses passed more readily into the nasal sinuses in 50 studies where B6 background mice were anesthetized with ketamine/xylazine compared to 51 experiments in which mice were anesthetized with isoflurane (7, 8), but we did not know 52 whether pulmonary deposition of dose differed depending on anesthetic used during dosing. To 53 guide future studies using these mice and anesthetics, we therefore measured the effect of 54 anesthetic approach on i.n. delivery of fluid to lungs of B6 mice. As we have used and refined 55 o.a. and i.t. methods, we also measured dose distribution using these administration routes. 56 We found striking effects of both anesthetic approach and dosing route on the efficiency of 57 pulmonary delivery. Our results will be useful to guide design of experiments with improved 58 reproducibility, with potential to reduce the number of mice needed to produce clear results from 59 experiments involving dosing of fluids into lungs. 60 C57BL/6 background mice (Jax #000664) were housed at the UCSF Parnassus Laboratory 63 Animal Resource Center specific pathogen-free facility. Both male and female mice were used 64 at ages 6-14 weeks. Mice were kept on a 12-hour light-dark cycle. Protocols were approved by 65 the UCSF Institutional Animal Care and Usage Committee. 66 Anesthesia 67 Mice were anesthetized either by inhalation of isoflurane (4% in oxygen) or by intraperitoneal 68 (i.p.) injection with ketamine (70 mg/kg) and xylazine (15 mg/kg) in normal saline. For fluid 69 dosing to lungs, it is important not to overdose ketamine/xylazine as higher doses can cause 70 asphyxiation from aspirated fluid. For terminal anesthesia prior to collection of lung samples, 71 mice were euthanized with ketamine (100 mg/kg) and xylazine (40 mg/kg) prior to 72 exsanguination. 73

Intranasal dosing 74
Gel-loading pipette tips (Sorenson #13810) were used to introduce 50 µl of dose dropwise into 75 the posterior opening of one nare. Mice were held upright for 20 seconds after dosing to allow 76 aspiration of dose. 77

Tracking radiolabeled albumin doses 78
For quantitative and visual tracking of inoculum we used 50 µl of phosphate-buffered saline 79 containing 125 I-albumin (0.25 mg/ml, ~2.5 KBq/ml, Jeanatope, Iso-Tex Diagnostics, Inc.) and 80 Evans blue dye (1 mg/ml). Organ samples were collected 10 minutes after mice were dosed. 81 Dose distribution was measured using a gamma counter (Packard 5000 series) against three 82 standards containing 100% of injected dose. 83 penicillin/streptomycin in a humidified incubator at 37ºC and 5% CO2 were infected and cultured 88 for 72h. The supernatant containing the virus was then collected and stored at -80ºC. Infectious 89 virus was quantified by culturing dilutions of the viral stock with MDCK cells in a 6-well plate for 90 1 hour, followed by addition of an overlay of 1.2% Avicel RC-581 in MEM, culture for 72 hours, 91 formalin fixation and staining with crystal violet for p.f.u. determination (9). Viral stocks were 92 diluted in sterile PBS at 4ºC prior to inoculation. Mice were dosed at zeitgeber time (ZT) 3-5 and 93 handled under biosafety level 2 conditions. 94

Bronchoalveolar lavage analysis 95
After terminal anesthesia and exsanguination, lungs were collapsed by opening the diaphragm 96 and tracheal insertion of 20G stub needles. A 1 ml syringe containing 1 ml of phosphate-97 buffered saline was then washed in and out of the lungs three times to recover bronchoalveolar 98 lavage (BAL) fluid. BAL cells were counted using a LUNA-II automated cell counter (Logos 99 Biosystems) and BAL supernatant total protein was measured using a Pierce total protein assay 100 (Thermo Scientific, #23225). 101 Dosing by oropharyngeal aspiration 102 As previously described, anesthetized mice were placed on an intubation platform suspended 103 by their upper incisors with the tongue gently pulled out of the mouth (5, 10). The fluid dose was 104 then pipetted directly onto the distal oropharynx at 50 µl volume with both nares covered to 105 obligate breathing through the mouth. After ~30 seconds, mice were removed from the platform 106 and placed supine until sample collection. 107

Non-surgical intratracheal dosing 108
Mice were positioned as with o.a. dosing, with transillumination and adjustment of body position 109 used to visually identify the larynx ( Figure 4B). Mice were orotracheally intubated with a 22G 1" 110 Safelet IV Catheter (Nipro, #CI+22225-2C) customized into an endotracheal tube and stylet with 111 addition of cushions (see Figure 4A). Tracheal positioning was confirmed by attaching a 112 manometer to the endotracheal tube and checking for oscillation of water column with breathing 113 movements ( Figure 4C) (11). Dose at 50 µl volume was then injected through the endotracheal 114 tube using a customized 28G ½" insulin syringe (BD, #329461) with added PE-10 tubing and 115 cushion for placement within catheter (see Figure 4A), followed by 120 µl air. 116

Experimental design and statistical analysis 117
Mice were randomly assigned to groups with blocking by cage, and samples were collected and 118 quantified with investigators blinded to groups. For the influenza infection study, the handler 119 dosing mice was also blinded during dosing, with a second unblinded handler in control of 120 anesthesia. Group n was set prior to study initiation and analysis. Where necessary, data were 121 transformed prior to statistical testing according to distribution. Statistical analyses used 122 InVivoStat 4.4 (body weight and power analysis) or GraphPad Prism 9 (other comparisons). The 123 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 3, 2023. ; https://doi.org/10.1101/2023.02.01.526706 doi: bioRxiv preprint tests used for each analysis are stated in figure legends with P=0.05 as α threshold. Data are 124 reported as means ± standard deviation. 125

Results 126
In previous experiments we noticed that B6 mice anesthetized with ketamine/xylazine smoothly 127 aspirated i.n. doses, whereas i.n. doses sometimes bubbled back out of the nares of isoflurane-128 anesthetized mice (7, 8). As previous studies assessing effects of anesthesia on i.n. delivery to 129 lungs used anesthetics or mouse strains not used in our protocols (1, 12, 13), we aimed to 130 determine whether use of isoflurane or ketamine/xylazine anesthesia during i.n. dosing affected 131 delivery of dose to lungs of B6 mice. 132 We found that relative to isoflurane anesthesia, use of ketamine/xylazine anesthesia during i.n. 133 dosing resulted in delivery of dose to more distal regions of lung ( Figure 1A) and increased 134 pulmonary dosing efficiency ( Figure 1B). 135 The i.n. dosing route is in widespread use in respiratory virus infection models. Current 136 protocols suggest that handlers can use either isoflurane or ketamine/xylazine anesthesia 137 during i.n. infection with influenza A virus (14). We therefore formally tested whether the effect of 138 anesthetic approach on intranasal dosing to lungs could be a factor altering outcomes and 139 reproducibility of studies of respiratory viral infection. 140 With one handler delivering anesthesia, and a second handler blinded to anesthetic approach 141 dosing and assessing mice, we gave B6 mice randomized to isoflurane or ketamine/xylazine 142 anesthesia prior to i.n. doses containing 50 p.f.u. of PR8 influenza A virus. 143 We observed bubbling of dose back out of nares and down the philtrum in mice in our 144 biodistribution study. During infection with PR8, the handler blinded to anesthesia approach 145 therefore recorded whether dose reflux was observed. We found that isoflurane-anesthetized 146 mice consistently refluxed some dose back out of their nares, whereas mice anesthetized with 147 ketamine/xylazine smoothly aspirated doses without visible reflux (Figure 2A,B). 148 We also monitored body weight daily as an index of general health status. All mice anesthetized 149 with ketamine/xylazine at time of infection had lost weight at day 9, but weight loss was 150 significantly lower in the isoflurane-anesthetized group from 5 to 9 days post infection, with 151 some mice in the isoflurane group gaining weight after inoculation ( Figure 2C). 152 At 9 days post infection we collected bronchoalveolar lavage (BAL) fluid from infected mice to 153 measure vascular leak and leukocyte recruitment into lung airspaces as indices of lung injury 154 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 3, 2023. ; https://doi.org/10.1101/2023.02.01.526706 doi: bioRxiv preprint and inflammation. Both supernatant protein concentration and leukocyte counts were higher in 155 BAL fluid from ketamine/xylazine-anesthetized mice compared to isoflurane-anesthetized mice 156 ( Figure 2D,E). 157 Using the BAL protein data in Figure 2D we ran a power analysis to determine the group size 158 needed for future experiments aimed at detection of a 30% change in BAL fluid protein 159 concentration using unpaired two-tailed t-tests. We found that 9 mice per group would be 160 needed to run such an experiment with 95% power with ketamine/xylazine anesthesia (Figure  161 2F). In comparison, the isoflurane anesthesia approach would likely not be feasible for 162 experimental use as an experiment with 25 mice per group would still have less than 50% 163 power ( Figure 2F). Tracking dose delivery visually and quantitatively, we did not detect any effect of anesthesia 175 approach on o.a. dose delivery to the lungs (Figure 3A,B). Breath-holding responses were 176 observed in some isoflurane-anesthetized mice after the dose was dropped onto the 177 oropharynx, but doses were eventually aspirated with reflux prevented by retraction of the 178 tongue and covering the nares. 179 We conclude from this study that anesthetic approach is therefore unlikely to have a major 180 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 3, 2023. ; https://doi.org/10.1101/2023.02.01.526706 doi: bioRxiv preprint involving direct visualization of the larynx, orotracheal intubation with customized catheter, 186 confirmation of airway placement using a manometer, and then injection using a customized 187 syringe (5, 6, 10, 11) ( Figure 4A-C), We therefore sought to measure pulmonary dosing 188 efficiency using our non-surgical approach for i.t. dosing, comparing to a control group dosed 189 with the o.a. approach, using ketamine/xylazine anesthesia. 190 We found that i.t. dosing yielded increased pulmonary dose deposition relative to o.a. dosing 191 (Figure 4D,E). This result is indicative that although o.a. and i.t. routes deliver the majority of 192 injected dose to the lungs, i.t. dosing might be desirable in situations where precise dosing to 193 lungs is needed. 194

Discussion 195
In this study, we found that anesthetic approach and administration route can affect the 196 efficiency of fluid dosing to lungs of mice. 197 We conclude from our results that ketamine/xylazine anesthesia is preferable where consistent 198 i.n. dosing to lungs of B6 mice is needed. Our results in the influenza infection model 199 demonstrate that optimal sedation for i.n. dosing has potential to reduce the number of mice 200 needed in experiments, and that the anesthesia approach used during i.n. dosing should be 201 reported for experimental reproducibility. The effects we found also suggest that tracking 202 biodistribution of i.n.-dosed radiotracers might be useful for quantitative in vivo studies of the 203 interaction of anesthetics with airway-protective glottic reflexes. 204 A likely consequence of poor delivery to lungs using i.n. dosing under isoflurane anesthesia is 205 that inoculating doses containing greater quantities of virus will be used to produce infections 206 that consistently result in robust lung inflammation, exposing extrapulmonary tissues to higher 207 quantities of virus. This is not desirable as exposure of the nasal sinuses to high quantities of 208 viral particles could cause serious adverse effects, as recently demonstrated in a study showing 209 lethal SARS-CoV-2 neuroinvasion when K18-hACE2 mice were infected using intranasal dosing 210 but not when mice were infected using aerosolized virus (19). 211 Injectable ketamine/xylazine anesthesia may not always be preferable as recovery time can be 212 longer than with isoflurane, and use of needles is discouraged where possible due to safety 213 risks. In our study, we found that it was feasible to give mice i.p. ketamine/xylazine injections in 214 a biosafety cabinet separate from that used for handling virus to minimize infection risk to 215 handlers. Limiting dose reflux using ketamine/xylazine anesthesia might also reduce risk of 216 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 3, 2023. ; https://doi.org/10.1101/2023.02.01.526706 doi: bioRxiv preprint aerosolization and surface contamination by inoculum. In studies where severe viral pneumonia 217 is not of interest, the limited pneumonia and patchy pathology we observed using isoflurane 218 anesthesia for i.n. infection with PR8 may be useful for studying infected lung lesions with 219 minimal animal suffering. 220 Our study provides quantification of pulmonary dosing efficiency comparing i.n., o.a. and i.t. 221 methods using anesthetics and mouse strains in current widespread usage. The percentage 222 lung delivery values that we measured are largely consistent with those from previous studies 223 which used a range of mouse strains, anesthetic approaches and methods for measuring dose 224 distribution (1, 4, 12, 13). Our conclusion that anesthetic type can alter lung deposition of i.n. 225 doses differs from that of a previous study which found no effect of different anesthetics 226 (isoflurane, halothane, Avertin) during i.n. dosing on pulmonary dosing efficiency in BALB/c 227 mice, although ketamine/xylazine was not examined in this previous report (1). Curiously, 228 another study using BALB/c mice found increased bacterial content of lungs after intranasal 229 dosing with Francisella tularensis under isoflurane compared to ketamine/xylazine anesthesia 230 (20). Effects of anesthetics may therefore vary depending on mouse strain and pathogen 231 biology. 232 In summary, we recommend the use of ketamine/xylazine anesthesia over isoflurane anesthesia 233

337
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

347
F. Output of power analysis using total protein data in Figure 2D to estimate number of mice needed per group to detect a 30% 348 change in BAL protein concentration using an unpaired two-tailed t-test.

352
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

360
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 3, 2023. ; https://doi.org/10.1101/2023.02.01.526706 doi: bioRxiv preprint

371
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 3, 2023. ; https://doi.org/10.1101/2023.02.01.526706 doi: bioRxiv preprint