Author contributions: J.D.B., D.G., H.B., L.E.O., and G.T. conceived the idea. C.S., D.R.S., M.T., R.L., L.E.O., and G.T. supervised this project. J.D.B., H.B., A.L., A.J.W., J.W., A.A., and K.W. created the pressurized devices to generate the GEMs. H.B., K.M.K., and A.T.C. performed the CO quantification in the materials and samples. V.R.F. and A.L. evaluated the physical properties of the materials. J.D.B., D.G., H.B., S.L.B., J.S.L., H.K., S.S., G.R.L., J.B., E.W., K.I., A.H., J.L.P.K., and J.J. performed small and large animal studies. E.C., M.S.L., W.R.J., D.E.B., and M.C.C. analyzed the samples from the small and large animal studies. J.D.B., D.G., H.B., L.E.O., and G.T. wrote the manuscript and revised according to the comments from other coauthors.
Carbon monoxide (CO) has long been considered a toxic gas but is now a recognized bioactive gasotransmitter with potent immunomodulatory effects. Although inhaled CO is currently under investigation for use in patients with lung disease, this mode of administration can present clinical challenges. The capacity to deliver CO directly and safely to the gastrointestinal (GI) tract could transform the management of diseases affecting the GI mucosa such as inflammatory bowel disease or radiation injury. To address this unmet need, inspired by molecular gastronomy techniques, we have developed a family of gas-entrapping materials (GEMs) for delivery of CO to the GI tract. We show highly tunable and potent delivery of CO, achieving clinically relevant CO concentrations in vivo in rodent and swine models. To support the potential range of applications of foam GEMs, we evaluated the system in three distinct disease models. We show that a GEM containing CO dose-dependently reduced acetaminophen-induced hepatocellular injury, dampened colitis-associated inflammation and oxidative tissue injury, and mitigated radiation-induced gut epithelial damage in rodents. Collectively, foam GEMs have potential paradigm-shifting implications for the safe therapeutic use of CO across a range of indications.
Here, inspired by molecular gastronomy, we present the development and preclinical evaluation of gas-entrapping materials (GEMs) for the delivery of CO through the GI tract. These simple systems were designed using Food and Drug Administration (FDA)–classified generally regarded as safe (GRAS) components to support rapid clinical translation. We tested our formulations in three small animal models associated with inflammation and oxidative stress–induced tissue injury: acetaminophen (APAP)–induced acute liver injury, experimental colitis, and radiation-induced proctitis ( 14 – 16 ). We also selected models based on reported efficacy of inhaled CO ( 17 , 18 ). Foam GEMs delivered high therapeutic amounts of carbon monoxide locally and systemically and reduced inflammation-associated damage in each animal model. These foam GEMs offer alternative modalities for the delivery of CO, enabling a spectrum of safe, effective, and potent delivery methods for enhanced translatability.
The discipline of molecular gastronomy has inspired delectable gas-filled materials over the past 40 years ( 13 ). Chefs across the world, including F. Adriá, have used foams and meringues to capture unique tastes and textures that appeal to the senses. Using whipping siphons, hand blenders, and mixers, these culinary artists create a deluge of different froths including shaving cream–like billows or light delicate aerated foams ( 9 , 13 ). The techniques from molecular gastronomy have seldom been translated for pharmacologic intent and provide a unique opportunity to enable delivery of CO and other gasotransmitters across the epithelium of the GI tract.
Given the demonstrated preclinical benefit, therapeutic CO delivery by the inhalation route has been evaluated in clinical trials ( 7 , 8 ). Inhalational delivery, however, presents profound challenges because of the variability in patient ventilation, environmental safety concerns for patients and healthcare workers, and the need for large amounts of compressed CO gas in cylinders that pose a health hazard due to the potential for cylinder leak or rapid depressurization ( 9 ). Therefore, other methods of administration have been developed including CO-releasing molecules (CORMs) and COHb infusions ( 10 , 11 ). These alternatives are limited either because of toxicity of transition metals or potency. An oral liquid with dissolved CO is also being developed to deliver CO primarily through the stomach and upper gastrointestinal (GI) tract, but the tunability of the formulation is unclear ( 12 ). Delivery through the GI tract is particularly promising because of the high diffusivity of CO across the epithelial barrier of the stomach and intestines ( 12 ). Moreover, the potential for local anti-inflammatory effects could enhance the application of CO for diseases affecting the GI mucosa.
Carbon monoxide (CO), an odorless and colorless gas, has long been recognized as a silent killer because of its strong affinity to hemoglobin (Hb). Carbon monoxide competitively displaces oxygen to form carboxyhemoglobin (COHb), thereby decreasing the body’s oxygen-carrying capacity. In general, if COHb reaches 50%, then it may result in coma, convulsions, depressed respiration, and cardiovascular status compromise or even fatal consequences. Conversely, at lower concentrations, CO acts as a gasotransmitter with beneficial properties akin to nitric oxide and hydrogen sulfide and has been implicated in a range of diverse physiological and pathological processes. Carbon monoxide has well-established immunomodulatory effects exerted through the heme oxygenase-1 (Hmox1; HO-1) pathway, which is implicated in adaptive cellular responses to stressful stimuli and injury ( 1 – 5 ). Exogenous administration of CO has been beneficial for the treatment of many diseases in preclinical models, including cardiovascular disorders, sepsis and shock, acute lung, kidney and liver injury, infection, and cancer ( 6 ).
The efficacy of CO-GEMs in dampening inflammation and tissue damage was next evaluated in a radiation-induced proctitis model. Rats were administered a single 18-Gy dose of radiation previously shown to induce acute proctitis within 2 weeks ( 28 ). Doses of this magnitude are routinely used for cancer therapy within the pelvis for definitive or palliative intent ( 29 , 30 ). Compared to air GEM and no-treatment controls, CO-GEMs administered rectally before and after radiation resulted in a reduction in intestinal crypt injury ( ). Moreover, there was a twofold increase in the number of normal crypts in rats treated with CO-GEMs compared to both air-GEMs and no treatment. Weight gain was similar among all groups ( fig. S16 ).
GEMs were next evaluated in the well-described experimental model of dextran sodium sulfate (DSS)–induced colitis in mice. In this model, mice were exposed to DSS for 3 days to induce colitis before initiation of CO foam treatment. Animals were maintained on DSS-containing water, and GEMs (air or CO) were administered once a day for an additional 4 days. DSS water was then stopped, and daily administration of foam GEM, air GEM, or no treatment was continued for an additional 3 days ( ). Foam GEM–treated mice showed reduced inflammation and tissue injury as evidenced by inhibition of DSS-induced colonic length shortening and reduced histological scores ( ). Further analyses showed reduced crypt injury, edema, and infiltration of polymorphonuclear neutrophils seen by histological staining ( ). Staining for 3-nitrotyrosine (3NT) adduct as a marker of oxidative damage to proteins mediated by peroxynitrite (ONOO − ) and glutathione-adducted proteins, common markers of oxidative stress, showed increased oxidation of proteins in the large intestine in animals treated with air GEM or DSS controls. This increase in expression of oxidative stress markers was significantly suppressed in foam GEM–treated mice ( ) ( 25 , 26 ). In addition, analyses of the intestinal microbiome using 16S sequencing showed increased presence of Romboutsia and Muricomes spp. in CO-GEM–treated animals compared to controls ( fig. S15 ) and may represent a mechanism by which CO limits intestinal injury and promote healing as reported by ( 27 ). These data support the therapeutic benefits of CO in limiting tissue damage in the colon after injury.
To characterize the efficacy of foam GEMs on dampening inflammatory responses in a disease model, we tested the foams in a well-characterized APAP overdose model ( 18 , 24 ). In this model, fasted mice administered APAP intraperitoneally showed an expected radical-mediated hepatocyte cell death and a rapid proinflammatory response, resulting in acute liver failure. Mice were dosed with foam GEMs 1 hour after APAP and then hourly for a total of three doses ( ). These animals were compared to control mice that received air GEM or no treatment. Animals that received foam GEM were found to have a dose-dependent reduction in hepatocellular injury as measured by serum alanine aminotransferase (ALT) concentration compared to controls that received air GEM or no treatment ( and fig. S14 ). Reduced ALT concentrations were associated with less caspase 3 staining, and hematoxylin and eosin (H&E) staining showed reduced apoptosis, liver necrosis, and congestion compared to controls ( ). Liver from foam GEM–treated mice appeared histologically normal compared to controls. Collectively, these data demonstrate the protective effects of a foam GEM in preventing APAP-induced centrilobular congestion, hepatocellular degeneration, and coagulative necrosis.
Tissue amounts of CO were determined in mice that were rectally administered foam GEMs ( , and ). Greater amounts of CO were detected in multiple tissues 15 min after administration, particularly blood-filled organs that included the liver, heart, spleen, lungs, and kidney. The small and large intestines were also found to have higher CO amounts compared to naïve controls.
The pharmacokinetics of rectal administration of CO-GEMs (5 g/kg) were characterized in both small and large animals ( , and ). COHb percentages in mice and rats reached a maximum directly after foam administration and rapidly decreased over 2 hours, paralleling the half-life observed with inhaled CO ( 23 ). Similarly, COHb percentages increased in nonventilated-anesthetized pigs over 2 hours after a single dose was administered intrarectally. Intestinal amounts of CO were higher for rectal foams (5 g/kg) compared to inhaled CO [250 parts per million (ppm) for 1 hour] ( fig. S13 ).
We next evaluated the behavior of the foams under flow conditions. The foams behaved similar to viscoelastic solids, with storage moduli (G′) increasing with xanthan gum concentration and exceeding loss moduli (G″) for all formulations ( fig. S12A ). In addition, all formulations were highly shear thinning, indicative of their ease of deployment with spraying or injection ( fig. S12 , B and C ). The 0.5 wt % xanthan gum foam was able to rapidly alternate between flowable and solid-like behavior at high and low shear strains, respectively ( fig. S12D ), and was therefore chosen for further testing in small and large animals.
Excipient concentrations, dosing, and pressure within the whipping siphon were evaluated to optimize CO delivery from foams. To study these parameters, we assessed visual differences in the macroscopic and microscopic appearance of the foams, volumetric foam stability, and CO release kinetics, as well as maximum COHb percentages in mice in vivo ( and figs. S4 to S6 ). Among the excipients within the foams, xanthan gum was found to have the greatest influence on foam stability ( , to , and fig. S5 ). The higher the xanthan gum concentration, the more stable the foam and the more prolonged the COHb maxima; the 0.25 weight % (wt %) xanthan gum formulation was the least stable and showed the fastest COHb peak among the different formulations tested ( , and , and fig. S6 ). Xanthan gum concentration had no influence on the amount of CO encapsulation ( fig. S7 ). The COHb percentages were found to be directly correlated with dosing of the foams ( ). Furthermore, the amount of CO encapsulated in the foam GEMs increased with higher pressures within the whipping siphon ( fig. S8 ). COHb percentages correlated with increasing pressures in the whipping siphon after a single rectal administration of equivalent dosing (5 g/kg; ). Benefits of the foams are long shelf life and the ability to formulate with other pharmaceutical agents ( figs. S9 and S10 ). Last, rectal dosing of foams over multiple dosing schedules spanning hours and days show stability of COHb with no compounding effect observed with these regimens ( fig. S11 ).
These materials were then evaluated using gas chromatography to quantify the amount of CO entrapped within the GEMs. The foam, solid, and hydrogel GEMs encapsulated about 25 times more CO than CO-enriched lactated Ringer’s ( ) ( 19 ). Next, we evaluated the peak COHb percentages after a single GI administration of each GEM in mice; foam GEMs were given rectally, whereas the solid and hydrogel GEMs were placed in the stomach through a laparotomy at equivalent doses (5 g/kg). A single dose of these materials resulted in higher peak COHb percentages than those previously reported for GI formulations of CO ( and table S1 ) ( 10 , 11 , 17 , 20 – 22 ). The pharmacodynamics of the GEMs demonstrated different CO exposures by each formulation ( fig. S4 ). Given the maximum COHb and ease of rectal administration of the foams, the foams were further optimized and subsequently evaluated in experimental animal models of disease.
Pressurized vessels were used to physically entrap CO in GRAS materials that can be easily administered to the GI tract through the oral cavity or rectum ( ). Specifically, whipping siphons were adapted to generate robust foams and hydrogels ( and figs. S1 and S2 ). The whipping siphons were rated for pressures near 500 psi, which provided a sufficiently large pressure range for operation. A custom-made high-pressure stirring reactor enabled the creation of solid gas–filled materials ( and figs. S1 and S3 ) in a process similar to that used for the candy, Pop Rocks. These pressurized vessels generated GEMs that can be rapidly dissolved based on the GEM composition and diluent ( figs. S2 and S3 ). In particular, the solid GEMs rapidly dissolved in aqueous media, whereas the hydrogel GEMs used an EDTA solution to dissolve the ionic cross-linking.
The highly permeable nature of the GI tract epithelium allows for rapid absorption of gases, thereby positioning it as an attractive delivery pathway for the therapeutic use of CO. However, the translation of therapeutic gases from basic research into everyday treatments has remained a challenge due to numerous safety and dosing constraints (31). To address this problem, we physically entrapped CO gas in GRAS materials for delivery across the GI epithelium and termed them as GEMs. The physical methods of entrapment were inspired by existing technologies currently used in the culinary arts (32). Because of their ease of use, the GEM systems may be adapted to a variety of gasotransmitters, such as oxygen, nitric oxide, and hydrogen sulfide, for various disease-related applications.
Here, we showed that administration of CO through foam, solid, and hydrogel GEMs can deliver titratable amounts of CO locally and systemically. We demonstrated that formulations enabling the administration of CO through noninhaled routes are tunable and not limited by delivery materials, toxicity, or potency. Our preclinical results suggest that delivery of CO through these materials is dose dependent and tunable and very amenable to rectal administration. Our results confirm the benefits of CO in part through direct modulation of tissue free radicals and secondary oxidative species that not only contribute to tissue injury but also elucidate a new potential mechanism involving effects on intestinal commensal bacteria, including CO-induced increases in Romboutsia spp., a taxon known to be decreased in patients with Crohn’s disease (33). Although CO-specific global microbiome diversity changes were not detected, alterations among select taxa may provide additional benefit of the CO foams in DSS-induced colitis and is the focus of ongoing experimentation. In addition, our systems are amenable to coadministration with other therapies to improve treatment efficacy, such as in concert with powdered drug formulations.
Carbon monoxide has well-established cytoprotective effects first identified through studies of HO-1, which is implicated in adaptive cellular responses to stressful stimuli and injury (1–5). Administration of CO can mimic the benefits of inducing HO-1 activity and thus endogenous CO generation during heme catalysis. This enzymatic function of HO-1 naturally produces CO endogenously that regulates a variety of intra- and extracellular cytoprotective and homeostatic effects (6, 34). Although CO has been shown to be beneficial for multiple disease states, it may be particularly well suited for use in regulating GI inflammation because the dense capillary network of the intestinal mucosa allows for rapid, direct delivery and uptake of CO for remote organ effects such as the liver (35). In addition, HO-1 has been shown to play a key role in modulating intestinal inflammation and innate immunity, and activation of this pathway through delivery of CO has already demonstrated benefit in the treatment of intestinal disease in animal models of colitis (17, 36). Because of our ability to locally deliver CO, our GEMs could expand treatment options for other inflammatory pathologies of the GI tract that involve chronic oxidative stress and tissue damage, such as inflammatory bowel diseases, radiation-induced injury, colon cancer, and gastroparesis (37–40). Moreover, GI administration per os or per rectum can provide simpler and potentially more effective modalities of administering CO for the management of non-GI disorders given the high diffusivity of CO. Ingestion of CORMs or CO-saturated solutions have been shown to effectively treat sickle cell anemia, ischemia-reperfusion injury of the kidney, and prevention of gastric ulcer formation (12, 41–44). The challenges associated with these treatment strategies are potency and safety.
Although we achieved improvement in multiple small animal models of inflammation-mediated disease with these GEMs, the efficacy of these materials may be further improved by manipulating the materials to increase CO content or expand the utility of this approach with other gases such as nitric oxide or hydrogen sulfide. The pressure in the whipping siphons used to create foam and hydrogel GEMs was less than half of the pressure rating. Thus, increasing the pressure may further enhance entrapment of CO, thereby maximizing loading efficiency and reducing the volume of material needed to be delivered to achieve a therapeutic benefit. Alternative materials beyond those tested here may enhance the volume fraction of the gas, which would also reduce the total dosing volume. Although rectal delivery of the foams was evaluated, oral delivery of the foam, solid, and hydrogel GEMs may increase the translatability of the materials given the ease and comfort associated with oral administration. Moreover, the materials enable tunable release of CO that could broaden the number of disease indications for which GEMs may be effective.
Further research is required to better understand the safety and tolerability of GI delivery of CO. The FDA placed a maximum safe COHb of 14% for inhalational CO; however, it is possible that GI delivery may enable a different therapeutic index for CO delivery based on differing pharmacodynamics and pharmacokinetics (10). There are reports that route of administration leads to marked differences in toxicology. Insufflation of the abdomen of dogs with CO to achieve COHb percentages comparable to CO administered by inhalation showed no toxicity as otherwise seen in dogs that inhaled CO (45). This strongly supports investigating alternative modes and routes of CO administration. Understanding person-to-person variability of COHb achieved between delivery modalities will be important to ensure patient safety and to maximize therapeutic benefit (31). Last, toxicity evaluation of the different materials and CO release kinetics will help determine safe formulation candidates to advance clinically.
For clinical translation of the GEMs, testing of these materials in large animal disease models will help determine the effectiveness in animals that more closely approximate the human condition. Our swine data confirm similar pharmacokinetics and feasibility for further translation into humans. Upon successful GEM optimization as a means to dampen the inflammatory response in large animal models, we will next evaluate these materials in healthy individuals to determine appropriate dosing, safety, and stability. Manufacturing of individual delivery devices that are recyclable and scalable will be required to facilitate safe handling for use in hospitals, clinics, in the field in ambulances, and helicopters as well as for in-home use.
Given the broad therapeutic benefits of CO, the GEMs described may be adopted for not only many different clinical applications, particularly inflammation-mediated GI conditions, but also pathologies as diverse as cancer, ileus, trauma, and organ transplant, among others (10). Furthermore, these materials could be extended to concomitant delivery of other pharmacologic agents for synergistic benefit, such as analgesics or antibiotics. In summary, our innovative approach using safe materials will offer modalities for the administration of therapeutic gases and treatment of acute and chronic inflammatory disorders.