key: cord-0321781-agd047mn authors: Mountford, Paul. A.; Leiphrakpam, Premila. D.; Weber, Hannah. R.; McCain, Andrea; Scribner, Robert. M.; Scribner, Robert. T.; Duarte, Ernesto M.; Chen, Jie; Borden, Mark. A.; Buesing, Keely. L. title: Colonic Oxygen Microbubbles Augment Systemic Oxygenation and CO2 Removal in a Porcine Smoke Inhalation Model of Severe Hypoxia date: 2021-12-09 journal: bioRxiv DOI: 10.1101/2021.12.08.466665 sha: 6c97253708801495fbc5bc76205e9368e62ac500 doc_id: 321781 cord_uid: agd047mn Inhalation injury can lead to pulmonary complications resulting in the development of respiratory distress and severe hypoxia. Respiratory distress is one of the major causes of death in critically ill patients with a reported mortality rate of up to 45%. The present study focuses on the effect of oxygen microbubble (OMB) infusion via the colon in a porcine model of smoke inhalation-induced lung injury. Juvenile female Duroc pigs (n=6 colonic OMB, n=6 no treatment) ranging from 39-51 kg in weight were exposed to smoke under general anesthesia for 2 h. Animals developed severe hypoxia 48 h after smoke inhalation as reflected by reduction in SpO2 to 66.3 % ± 13.1% and PaO2 to 45.3 ± 7.6 mmHg, as well as bilateral diffuse infiltrates demonstrated on chest x-ray. Colonic OMB infusion (75 – 100 mL/kg dose) resulted in significant improvements in systemic oxygenation as demonstrated by an increase in PaO2 of 13.2 ± 4.7 mmHg and SpO2 of 15.2% ± 10.0% out to 2.5 h, compared to no-treatment control animals that experienced a decline in PaO2 of 8.2 ± 7.9 mmHg and SpO2 of 12.9% ± 18.7% over the same timeframe. Likewise, colonic OMB decreased PaCO2 and PmvCO2 by 19.7 ± 7.6 mmHg and 7.6 ± 6.7 mmHg, respectively, compared to controls that experienced increases in PaCO2 and PmvCO2 of 17.9 ± 11.7 mmHg and 18.3 ± 11.2 mmHg. We conclude that colonic OMB therapy has potential to treat patients experiencing severe hypoxemic respiratory failure. One Sentence Summary Enteral oxygen microbubbles increase systemic oxygen and decrease carbon dioxide levels in acutely hypoxic pigs after smoke inhalation-induced respiratory failure. Prior to the spread of SARS-CoV-2, acute respiratory distress syndrome (ARDS) historically occurred in ~10% of patients entering the intensive care unit -affecting nearly 190,000 patients per year in the US alone -with a reported mortality rate ranging from 35% to 46% (1, 2) . As of October 20, 2021, a total of 730,368 COVID-19 deaths have been reported in the US (3) . The primary symptom of COVID-19 infection requiring hospitalization is hypoxemic respiratory failure (4) . Regardless of underlying pathology, mechanical ventilation remains the mainstay of oxygenation and ventilatory support for severe respiratory failure; however, complications such as ventilator-induced lung injury, ventilator-associated pneumonia, barotrauma, and progressive deconditioning leading to ventilator dependence remain unacceptably high (5) . Despite decades of research, therapeutic options for patients with severe hypoxemia that fail mechanical ventilatory support are limited. Extracorporeal membrane oxygenation (ECMO) -a temporary, artificial extracorporeal support of the respiratory and/or cardiac system -is one of the last resorts for treating refractory respiratory failure (6) . On the spectrum of ARDS treatment, ECMO is invasive and complex. When used for respiratory failure, it serves as a pulmonary bypass, circulating the patient's blood through an external circuit to exchange oxygen and carbon dioxide (7) . While usually effective at correcting acute hypoxemia, ECMO has a substantial contraindication list and risk profile leading to unacceptably high 30-and 60-day mortality rates of 39-50% (6, 8) . The risks of ECMO are significant, and include hemorrhage, thrombocytopenia, circuit failure, embolism, hemolysis, and limb ischemia among other potential complications (8) . Alternative therapies that can provide meaningful systemic oxygen augmentation and allow the lungs to rest without the risks of ECMO need to be actively pursued and investigated. Oxygen microbubble (OMB) therapy is a novel technology that shows promise as a method of extrapulmonary oxygenation that is relatively simple and safe to administer, does not require the use of anticoagulants and does not have the risk profile associated with ECMO. OMB is comprised of a high concentration of micron-scale (1-20 um diameter) bubbles that contain an oxygen "core" and are encapsulated by a lipid monolayer shell (9) , similar to the pulmonary alveolus. When administered as a bolus dose into the abdominal cavity (akin to peritoneal dialysis), OMBs have been reported to augment systemic oxygenation and improve outcome in small animal pilot studies involving unilateral pneumothorax, tracheal occlusion and LPSmediated severe ARDS (9) (10) (11) . Here, we introduce a novel delivery pathway for OMB therapy -the colon -as an improved translational candidate for minimally invasive, nonsurgical oxygenation and carbon dioxide removal for the treatment of severe hypoxia. The colonic mucosa is associated with a rich capillary matrix. Oxygen tension in the mucosal layer has been studied in small animal models of hyperbaric oxygen therapy, where investigators found that oxygen diffused from intestinal tissue and established a radial gradient from the tissue interface to the colonic lumen (12) . It naturally follows that if systemic hyperoxia can augment luminal oxygen content via the capillary gradient, establishing an elevated oxygen content in the colonic lumen would lead to diffusion across the capillary bed, augmenting systemic oxygenation in states of hypoxemia. A similar argument holds for carbon dioxide exhaust from the same capillary gradient. Moreover, the colon provides an ability to deliver a clinically relevant volume of OMB without the need to place a surgical port, as required by alternative enteral routes. This study examines the hypothesis that colonic OMB therapy can significantly increase systemic oxygen levels, and reduce systemic carbon dioxide levels, in a large-animal model of severe hypoxia. Prior to OMB treatment at 48 h after smoke inhalation, lung injury was assessed by chest x-ray (CXR) and carotid, femoral and pulmonary arterial catheter blood gas sampling. CXR confirmed the presence of diffuse bilateral infiltrates indicative of ARDS (Fig. 1A 1C ). There was an increase in IL-6 inflammation within the lungs (Fig. 1H , BAL). Additionally, we observed a significant increase in overall lung injury score (Fig. 1I ) and average wet-dry (W/D) weight ratio of lung tissues 48 h after smoke exposure compared to the control animals ( Fig. 1J ). Upon achieving severe hypoxia due to smoke inhalation injury, OMB was administered to the colon (Fig. 2D ) in the form of three bolus injections at a rate of 500 mL/min for a total dose size of 75 to 100 mL/kg (3.6-4.5 L total for pigs ranging from 40-50 kg). The OMB had a number-weighted average microbubble diameter of 1-10 um with most of the oxygen gas volume existing in bubbles 1-20 um in diameter (Fig. 2B) . Within 120 min after the start of OMB treatment, all blood and non-invasive oxygen vitals showed statistically higher oxygen content for animals receiving OMB treatment compared to no-treatment control animals. Specifically, P a O 2 rose significantly for OMB-treated animals within the first 15 min to 52.5 ± 6.9 mmHg (OMB Δ P a O 2 = 9.3 ± 6.4 mmHg, no treatment (NT) Δ P a O 2 = -2.3 ± 6.7 mmHg) and continued rising to 56.4 ± 7.9 mmHg (OMB Δ P a O 2 = 13.2 ± 4.7 mmHg, NT Δ P a O 2 = -8.2 ± 7.9 mmHg) after 150 min (Fig. 3A,C) . Additionally, P mv O 2 increased significantly for OMB treatment animals to 34.8 ± 5.8 mmHg (5.0 ± 5.9 mmHg over -7.3 ± 6.9 mmHg as seen by the NT animals) after 150 min (Fig. 3A,D) . SpO 2 also rose significantly by 14.2% ± 9.5% after 60 min as Moreover, arterial measurements showed blood gas CO 2 declining from treatment until the end of the study for animals receiving OMB. Specifically, P a CO 2 (Fig. 3B ,F) was significantly lower for animals receiving OMB after 15 min at 61.20 ± 6.39 mmHg (OMB Δ P a CO 2 = -10.0 ± 6.4 mmHg, NT Δ P a CO 2 = 5.4 ± 4.6 mmHg) and remained lower out to 150 min at 53.6 ± 5.6 mmHg (OMB Δ P a CO 2 = -19.7 ± 7.6 mmHg, NT Δ P a CO 2 = 17.9 ± 11.7 mmHg). No significant changes were observed in lung injury score and wet/dry weight ratio at the 3-h time point (Fig. 4A-D) . However, we observed a significant increase in IL-1β levels in immunoblotting of lung tissue lysates at 48 h post smoke inhalation in smoke injury (SI) animals compared to the control animals (Fig. 4E,F) . Expression status of both cytokines was reversed at 3 h post OMB treatment (Fig. 4F) , indicating an anti-inflammatory effect of OMB. There were lower IL-6 levels for the OMB group (BAL = 1.7 ± 1.5 pg/mL, Plasma = 5.3 ± 5.5 pg/mL) as compared to the NT group (BAL = 32.9 ± 24.0 pg/mL, Plasma = 14.0 ± 13.9 pg/mL) for both the BAL and Plasma samples (Fig. 4G ). Fresh frozen lung tissues of control, SI and SI with OMB groups were compared for their global protein expression by proteomic analysis. Several proteins were differentially expressed between these groups. Of these proteins, 320 proteins were significantly upregulated or downregulated at 48 h post smoke exposure compared to control (Fig. S5C) . Interestingly, at 3 h post OMB treatment, the expression status of these 320 proteins was partially reversed, and we observed a significant differential expression of 68 proteins compared to SI animal tissue samples as shown by the Venn diagrams ( Fig. 4H and 4I) and heat map analysis (Fig. 4J ). The primary driving force behind systemic oxygenation with OMB treatment is diffusion. Upon delivery to the colon, the OMBs comprising 95-98% oxygen gas deliver oxygen to hypoxic tissue with a diffusivity of ~2.4 × 10 -6 cm 2 /s (13) . OMBs are superior for delivering oxygen as compared to a macro, non-shelled oxygen gas bubble due to their ability to intimately spread and mix throughout the colon submucosa and increase the interfacial gas-liquid surface area. This enhancement in transport for microbubbles has led to their utilization in industrial fermentation (14) (15) (16) , and other applications requiring rapid gas absorption. The colon walls are highly vascularized and absorptive (removing ~ 2 L/day of water from chyme and stool (17) ) and thereby allow oxygen and carbon dioxide to diffuse across the submucosa from the lumen to adjacent tissue. Within the colon tissue, oxygen diffuses into capillary vessels, binds to deoxygenated blood, and circulates resulting in augmentation of systemic oxygen levels. The statistically significant rises in oxygen blood gas sampled at the carotid and pulmonary arteries demonstrate that OMBs ability can deliver systemic oxygen to a patient via the splanchnic circuit. The diffusion mechanism responsible for delivering systemic oxygen to an OMB-treated patient is also responsible for removing systemic CO 2 . The permeability of CO 2 in tissue is ~20 fold higher than for oxygen (18) . Thus, CO 2 counter-diffuses into the OMB microfoam as O 2 diffuses into tissue, and the OMB structure is stabilized during this gas-exchange process by the structural integrity of the lipid shell (19) . As hypercapnic venous blood passes through the splanchnic circuit, it sheds CO 2 into the OMB matrix. This lowering of CO 2 blood gas in the mixed venous return, verified by both P mv CO 2 and ETCO 2 ( Fig. 3B ,F,H), reduces the alveolar CO 2 that would otherwise dilute alveolar O 2 , thereby increasing P A O 2 (Fig. S3) . Conversely, all NT animals experienced an increase in CO 2 levels throughout the study. In this study, we showed that colonic OMB can improve systemic O 2 delivery and CO 2 removal for up to 150 min. This time duration is relevant for acute, severe hypoxia as a possible bridge to alternate therapy, where resources/access to ECMO is limited, or when the risks of pulmonary bypass via established methods outweigh the benefit. Eventually the OMB bolus will deplete its oxygen supply and saturate itself with CO 2 . In such cases, multiple bolus administrations can be administered, where the expired bolus can be flushed naturally or with the help of conventional pro-motility enema. This capability favors colonic administration as the preferred enteral route for OMB therapy. In prior work, we focused on the intraperitoneal (IP) route in small animal lung injury models (9) (10) (11) . We also examined the IP route here in our SI porcine model and found similar results to the colon route (Figs. S2, Table S2 ). IP-OMB significantly elevated arterial and mixed venous oxygen tension and simultaneously reduced carbon dioxide over 120 min. IP-OMB increased systemic oxygenation over the NT group (Δ P a O 2 = 17.2 ± 13.9 mmHg within 30 min and Δ P a O 2 = 11.2 ± 4.6 mmHg at 120 min). IP-OMB also decreased systemic CO 2 levels compared to the NT group (Δ P a CO 2 = -6.4 ± 10.8 mmHg at 30 min and Δ P a CO 2 = -7.9 ± 7.8 mmHg at 120 min). It is plausible that the peritoneal cavity may serve as an additional route to the colon to increase oxygenation further beyond what can be obtained by the colon alone. However, the colon route remains favorable because, for the same OMB dose range, the increase in systemic O 2 and decrease in CO 2 was similar for both routes, and the colon route has additional advantages of not requiring surgery to establish a port, the colon can be easily flushed naturally or by enema and re-administered, and it requires only foodgrade sterility since it is contained within the gastrointestinal tract. The main limit of this study is that we examined effects out to a pre-determined 150 min endpoint. Severe hypoxia from respiratory distress may occur for a much longer period, over days and even weeks. Therefore, future work will focus on understanding not only the therapeutic duration of a single OMB bolus, but also the therapeutic effects of multiple doses. Additionally, colonic OMB therapy should be tested in alternative large-animal ARDS models, such as lipopolysaccharide and oleic acid, with varying degrees of severity of lung insult and hypoxia, as well as the interaction between OMB therapy and mechanical ventilation. Colonic OMB therapy should also be investigated as a bridge to ECMO for rapidly deteriorating patients who require additional time for transport and to setup the circuit, or for austere environments with limited resources. All animal experiments were approved by the University of Nebraska Lincoln (UNL) Institutional Animal Care and Use Committee (IACUC). Female pigs (31-51 kg, n=6 treatment, n=6 no treatment) were housed and cared for according to USDA (United States Department of Agriculture) guidelines. Animals were acclimated to the facility for 4-7 days and received food reward training to ease handling and blood draws. Study animals were fasted overnight and given free access to water for procedures on the following day. temperature were monitored throughout the study. Blood samples were drawn from the CA, FA and PA catheters for the measurement of baseline blood gas prior to smoke inhalation and at predetermined time intervals throughout the study period (ABL80 FLEX CO-OX, Radiometer, Brea, CA). To maintain patency, catheters were flushed throughout the experiment with 3-5 mL of sterile saline, and a heparin solution (1:500 dilution in 50% dextrose solution) was infused to fill the volume of the port chosen as a "lock" solution. Sedated/anesthetized animals from survival surgeries were continuously monitored until they regained sternal recumbency. All catheters were removed after smoke inhalation was completed. The surgical procedures were repeated at 48 hr after smoke inhalation prior to treatment with OMBs. Upon completion of all surgical procedures, animals (n=12) were exposed to oak wood smoke from a custom-made smoke generator connected in parallel to the endotracheal tube (Fig S1A) . The duration of the smoke exposure was 2 hr, and the volume of smoke was approximately 1000 L (as estimated from TV, RR, and total time of smoke exposure). Invasive and noninvasive vital signs were monitored continuously during the experiment. Following smoke exposure, blood samples were collected from CA, FA and PA ports for blood gas analysis. Smoke exposure was stopped immediately if the animal developed hemodynamic instability, which was determined by hypotension (systolic blood pressure less than 60) and irreversible desaturation (SpO 2 less than 70% despite rescue maneuvers such as increasing FiO 2 ). OMBs were generated via sonication as described by Feshitan et al. 8 The resulting OMB solution, which contained 15% oxygen gas by total volume (void fraction (VF) = 0.15), was then centrifuged (Sorvall Legend T, Thermo Scientific, Waltham, MA) as described by Swanson et al. 12 in batches of four 140 mL syringes (Covidien Monoject 140, Medtronic, Minneapolis, MN) at 300 relative centrifugal force (RCF) for 1 min to achieve a final oxygen gas content of VF ≥ 0.7 (+70% oxygen gas by total volume). The resulting high-concentration oxygen microbubble foam was collected in 2 L gas-tight syringes (S2000, Hamilton, Reno, NV) and stored at ~ 4 ºC. OMB size (Figure 2A) was measured using the electrozone sensing method (Coulter Multisizer III, Beckman Coulter, Opa Locka, FL). OMB VF was calculated by subtracting the weight of a fixed volume of OMB from the weight of an equivalent volume of aqueous solution (lipid-PBS mixture) and dividing by the weight of the aqueous solution at the same fixed volume. Finally, oxygen gas content by total gas volume (%) was measured with an oxygen needle sensor (OX-NP, Unisense, Aarhus, Denmark). At 48 h after smoke inhalation, CXR was obtained and CA, FA, and PA catheters were again placed for serial blood sampling and monitoring as described above. Lung injury from smoke inhalation was confirmed by presence of bilateral diffuse infiltrates on CXR ( Figure 1D & E) . After completion of catheter placement, FiO 2 was lowered to 21% and maintained throughout the remainder of the experiment. Other ventilator parameters were set to TV = 6-8 mL/kg, PEEP = 0-1 to maintain normal driving pressure, and RR was adjusted to maintain eucapnia. Baseline Lumos™ coupled with UltiMate 3000 HPLC system (Thermo Scientific). Quantification parameters were set: Peptides to use: unique+razor, Normalization mode: total peptide amount. All oxygen and CO2 blood gas data are reported as mean ± standard deviation. Statistical significance was based on an unpaired parametric t-test with Welch correction (non-equal standard deviations) between the delta of the no treatment group (n=6) and the delta of the OMB treatment group at the 15, 30, 45, 60, 90, 120 and 150 min time points (Table S1 ) and out to 120 min for the intraperitoneal OMB treatment group (Table S2 ). The colonic OMB treatment group fell to an n=5 after 60 min due to a loss in data collection for one of the animals. The same animal had a loss in collection of PaCO2 data within 15 min of OMB treatment. An additional animal dataset did not contain PmvCO2 data due to equipment error bringing the PmvCO2 sample size from n=5 to n=4 after 60 min. The intraperitoneal OMB treatment group had a n=5 due to the study switching to the colonic route of administration after the 5 th animal. One-way ANOVA for multiple comparisons was used to compare control (n=4) vs. SI (n=5) vs. SI+OMB Fig. 1. Porcine Smoke Inhalation Injury. A and B ) Before and after (D and E) chest x-ray images confirming the presence of diffuse bilateral infiltrates indicative of ARDS due to smoke inhalation injury. C) P a O 2 (red, p=0.000147), P mv O 2 (blue, p<0.0001), SpO 2 (violet, p<0.0001), P a CO 2 (gold, p<0.0001), P mv CO 2 (green, p=0.000123) and ETCO 2 (teal, p<0.0001) measurements taken both before (t=-48 h) and after (t=-0.5 h) smoke inhalation injury. Hematoxylin and eosin (H&E) staining of paraffin embedded lung tissue sections of baseline (t = -48 h) (F) and SI + 48 h (t = -0.5 h) (G) animals (scale bar = 100 um). H) IL-6 marker analysis for baseline and smoke injury (SI) + 2 h (t = -46 h) for BAL and Plasma samples. Comparison of lung injury score (I) and lung wet/dry ratios (J) showing a significant difference between control and SI + 48 h (t = -0.5 h) samples (p<0.0001 and p=0.0188, respectively) Past and Present ARDS Mortality Rates: A Systematic Review Acute respiratory distress syndrome Silent Hypoxemia in Patients with COVID-19 Pneumonia: A Review The future of mechanical ventilation: lessons from the present and the past Clinical review: Extracorporeal membrane oxygenation Venovenous Versus Venoarterial Extracorporeal Membrane Oxygenation for Adult Patients With Acute Respiratory Distress Syndrome Requiring Precannulation Hemodynamic Support: A Review of the ELSO Registry A meta-analysis of complications and mortality of extracorporeal membrane oxygenation Systemic oxygen delivery by peritoneal perfusion of oxygen microbubbles Evaluation of Peritoneal Microbubble Oxygenation Therapy in a Rabbit Model of Hypoxemia Treatment of a Rat Model of LPS-Induced ARDS via Peritoneal Perfusion of Oxygen Microbubbles Correlation Between Intraluminal Oxygen Gradient and Radial Partitioning of Intestinal Microbiota Oxygen diffusivity of biologic and synthetic scaffold materials for tissue engineering Mass-transfer properties of microbubbles. 1. Experimental studies Microbubbles for Enhancement of Oxygen Transfer in Xanthan Gum Fermentation Enhancement of oxygen gas-liquid mass transfer with colloidal gas aphron dispersions Biomedical Engineering: Bridging Medicine and Technology THE DIFFUSION OF CARBON DIOXIDE IN TISSUES Lipid monolayer dilatational mechanics during microbubble gas exchange E and F) Immunoblot analysis of IL-1β expression levels in fresh frozen lung tissues of baseline (t = -48 h), NT and OMB samples (t = 3 h). Difference between baseline and NT groups was statistically significant (p=0.0092). G) IL-6 marker analysis for NT and OMB (t = 3 h) for BAL and Plasma samples. H, I and J) Proteomic analysis of control (t = -48 h), SI (NT, t = 3 h) and SI+OMB (t = 3 h) groups for their global protein expression as described in Material and Methods section. Venn diagram showed 320 proteins with significant differential expression between SI and control groups (H and I) We would also like to acknowledge Andrew Kingsbury for generating the schematics presented in this manuscript. This study was also supported by the National Institutes of Health award R01HL151151 awarded to Mark A. Borden.