Hexametaphosphate Sodium

Extracellular pyrophosphate: the ‘early years’

Pyrophosphate and polyphosphates are good complexing agents for metal ions (e.g. calcium and transition metals) giving them many uses in industrial chemistry. In particular, polyphosphates have long been used to prevent calcification; for example, sodium polyphosphate was first used in the Calgon® water softener in the 1930s. Pyrophosphate and related polyphosphates, such as hexametaphosphate, have also been extensively used as toothpaste additives to prevent dental calculus formation and as food additives. However, it was the pioneering work of Fleisch and colleagues in the 1960s that identified the ability of pyrophosphate to inhibit biomineralisation [4–7]. They discovered that pyrophosphate potently antagonises the ability of calcium to crystallise with phosphate to form hydroxyapatite (Ca₁₀(PO₄)(OH₂)) [5,7]. Pyrophosphate also binds stongly to the surface of hydroxyapatite crystals and blocks their ability to act as a nucleator for mineralisation therefore preventing further crystal growth [8].

This initial work helped to establish the concept that pyrophosphate is the body's own ‘water softener’ which acts to prevent harmful soft tissue calcification and regulate bone mineralisation [8,9]. Subsequent studies using ³²P-labelled pyrophosphate in dogs enabled the kinetics of pyrophosphate production and elimination to be examined (Figure 2) [10]. This work suggested that the daily turnover of extracellular pyrophosphate in an adult human might be in the range 100 mg/day, a very small amount compared with the many grams likely to be generated intracellularly during biosynthetic reactions. Early studies also revealed that the pyrophosphate in human bodily fluids, including urine, is endogenous and does not come from dietary sources [3]. Indeed feeding large amounts of pyrophosphate did not increase levels any more than giving the same amount of inorganic phosphate. This is because pyrophosphate, like other phosphate compounds, seems to be completely hydrolysed within the intestinal tract by enzymes including alkaline phosphatase located on the brush borders of intestinal villous cells. In the 1960–1970s it was thought that feeding phosphate might be effective in reducing kidney stone formation in patients; although this seemed counterintuitive it increased urinary pyrophosphate, by a mechanism that appeared to involve inhibition of its intra-renal hydrolysis [11]. Reduced levels of pyrophosphate are also found in some groups of stone formers [12].

Figure 2. Systemic extracellular metabolism of pyrophosphate. Studies using ³²P-pyrophosphate injected into dogs showed how extracellular pyrophosphate is produced and eliminated systemically.

Source: Figure is adapted from Jung et al. [10].

Pyrophosphate is found in mineralised tissues (e.g. teeth and bone) at concentrations representing approximately 0.5% of the total phosphate content [13,14]. The intracellular concentrations have been difficult to determine, not least because of compartmentalisation, but are likely to be at least tenfold lower than that of inorganic phosphate. Interestingly in platelets, pyrophosphate is found in dense granules which are released during blood clotting [1^••]. This is important because serum levels of pyrophosphate produced in vitro can be several-fold higher than plasma concentrations, and this has previously led to misinterpretation of circulating levels of pyrophosphate in human diseases.

Deposits of pyrophosphate as calcium salts occur in humans, such as in the disease chondrocalcinosis, but also in nature. For example, deposits of amorphous calcium pyrophosphate mixed with calcium phosphates are found in the hepatopancreas of snails where they are thought to selectively accumulate metal irons, and have been used as monitors of toxic metals like cadmium, zinc and mercury in the environment [15].

Much remains to be learnt about the role of pyrophosphate in biology and mineralisation. For example, high pyrophosphate levels (>100 μM) are found in milk where it may help to keep the extremely high concentrations of calcium and phosphate in a colloidal state and prevent them from precipitating out (RGG Russell, unpublished observation). This article is dedicated to the memory of Herbert R Fleisch and William F Neuman, whose discoveries laid the foundations for understanding the role of pyrophosphate in mineralisation. It will summarise our current understanding of how this simple molecule regulates mineralisation.

Phosphate Classification and Nomenclature

Inorganic phosphates are classified by the number of phosphorus atoms in the phosphate molecule. The types of phosphates that are important to the meat industry include the orthophosphates, the pyrophosphates, and the straight-chain phosphates.

Orthophosphates contain only one phosphorus atom per molecule. The pyrophosphate molecule consists of two phosphorus atoms linked by a shared oxygen atom. An inorganic phosphate of such structure is referred to as a condensed phosphate. The two pyrophosphates approved by the US Department of Agriculture's Food Safety Inspection Service (USDA FSIS) are sodium acid pyrophosphate and tetrasodium pyrophosphate.

Inorganic phosphates of three or more phosphorus atoms are referred to as polyphosphates. Sodium or potassium tripolyphosphates consist of three linked phosphorus atoms. Sodium hexametaphosphate is a misnomer for a long straight-chain polyphosphate. The ‘meta’ designation is correctly given to cyclic polyphosphates. Hexametaphosphate contains an average of 10–15 phosphorus atoms. The name ‘sodium hexametaphosphate’ has been deleted from USDA regulations and replaced with ‘sodium polyphosphate, glassy’ (Graham's salt) and ‘sodium metaphosphate, insoluble’ (Maddrell's salt).

1 Introduction

Mislocalization of DNA into the cell cytosol activates the human innate immune system. The enzyme cyclic GMP–AMP synthase (cGAS) recognizes cytosolic DNA and cGAS activity is critical for the cellular response to pathogen replication, stress, and aberrant DNA replication in cancer (Chen, Sun, & Chen, 2016; Khoo & Chen, 2018; Sun, Wu, Du, Chen, & Chen, 2013). cGAS binds directly to double-stranded DNA and undergoes a conformational change that activates the enzyme to catalyze synthesis of the nucleotide second messenger 2′–5′/3′–5′ cyclic GMP–AMP (2′3′ cGAMP) (Ablasser et al., 2013; Civril et al., 2013; Diner et al., 2013; Gao, Ascano, Wu, et al., 2013; Kranzusch, Lee, Berger, & Doudna, 2013; Li et al., 2013; Sun et al., 2013; Zhang et al., 2013). 2′3′ cGAMP diffuses throughout the cell, binds to the receptor Stimulator of Interferon Genes (STING), and initiates a signaling cascade that induces downstream type I interferon and NF-κB signaling (Burdette et al., 2011; Gao, Ascano, Zillinger, et al., 2013; Ishikawa & Barber, 2008; Jin et al., 2008; Sun et al., 2009, 2013; Zhang et al., 2013; Zhong et al., 2008).

Following the initial discovery of cGAS by Sun et al. (2013), numerous techniques have been developed to investigate regulation of cytosolic DNA recognition. Initial structural and biochemical studies revealed the basic mechanism of enzyme activation and 2′3′ cGAMP synthesis, but relied primarily on mouse cGAS and other mammalian cGAS homologs that exhibit increased activity and in vitro stability (Civril et al., 2013; Gao, Ascano, Wu, et al., 2013; Kato et al., 2013; Kranzusch et al., 2013; Li et al., 2013; Zhang et al., 2014). These pioneering studies are essential to our understanding of cGAS biology, but key limitations of these approaches have prevented fully understanding human cGAS function.

The CGAS gene has undergone strong positive-selection and is one of the most divergent genes shared between the human and mouse genomes (George et al., 2011; Hancks, Hartley, Hagan, Clark, & Elde, 2015). The high level of species-specific variation in human cGAS directly impacts enzyme function and intrinsic regulation, and human cGAS exhibits a markedly reduced rate of 2′3′ cGAMP synthesis compared to mouse cGAS (Zhou et al., 2018). Although promising small molecule inhibitors of cGAS have recently been identified (Hall et al., 2017; Vincent et al., 2017), the reduced catalytic efficiency of human cGAS has limited development of screens designed to directly identify potent human cGAS agonist and antagonists. An additional species-specific phenomenon recognized by many groups in the field is that human cGAS aggregates in vitro in the presence of DNA. Unlike mouse or other mammalian cGAS homologs, human cGAS does not appear to form a stable, minimal protein–DNA complex. This challenge has limited the application of common in vitro techniques to monitor human cGAS interaction with activating DNA and prevented determination of high-resolution structures of human cGAS in an activated state.

Here we present detailed protocols that overcome challenges specific to human cGAS and allow determination of the structure of the human cGAS–DNA complex (Zhou et al., 2018). Through a combination of bacterial genetics and biochemical assembly of stable human cGAS–DNA complexes in vitro, our approach mapped a key species-specific determinant of cGAS regulation and provided a new template to explain the structural basis of cGAS small molecule inhibitor specificity. We have used these experimental advances to dissect the mechanism of human cGAS activation and explain how species-specific cGAS variations impact DNA recognition in human cells. Importantly, we anticipate that adaptation of these techniques will enable continued advances in understanding the biology and therapeutic potential of cGAS-STING signaling.

3.16.8.6 The Transduction Channel

Mechanosensation is ubiquitous; organisms as diverse as bacteria and blue whales utilize mechanosensation. The first mechanosensitive ion channel to be cloned was the MscL channel in Escherichia coli, which is gated by tension within the lipid bilayer and functions in osmoregulation. MscL can be mechanically gated when it is reconstituted in a lipid bilayer (Martinac, B., 2001). In contrast, most vertebrate and invertebrate mechanotransduction channels are thought to be gated by cytoskeletal and extracellular elements. Therefore, the mechanosensitivity of putative vertebrate channels often cannot be reconstituted in heterologous expression systems that lack these elements, making it much more difficult to directly demonstrate that individual proteins function as mechanotransducers.

Two families of ion channel proteins have been found to be necessary for mechanosensation in a variety of cell types: the degenerin/epithelial sodium channel (DEG/ENaC) family of ion channels and the transient receptor potential (TRP) family. In the worm Caenorhabditis elegans, a random mutagenesis screen was used to identify genes involved in mechanosensation, the mec genes. MEC-4 and MEC-10 are members of the DEG/ENac family of ion channels and are expressed in touch-sensitive neurons. They contain two transmembrane-domains, a large extracellular domain and a pore-forming loop. Certain gain-of function mutations in MEC-4 or MEC-10 cause the neurons to degenerate, hence the name degenerin (Huang, M. and Chalfie,M., 1994). MEC-2 and MEC-6 associate with MEC-4 and MEC-10, and can regulate channel activity (O’Hagan, R. et al., 2005). In a heroic series of experiments, O’Hagan, R. et al. (2005) applied forces to the body wall of the worm using a flexible glass probe while simultaneously recording from the touch neurons in vivo. Null mutations in MEC-4, MEC-2, and MEC-6 did not affect voltage-gated currents in the neurons, but completely abolished mechanotransduction currents. The mechanotransduction currents are carried by Na⁺ and can be blocked by amiloride, consistent with the properties of DEG/ENaC channels. Finally, more subtle mutations in specific glycine residues of both MEC-4 and MEC-10 reduced mechanotransduction current amplitudes and shifted the reversal potential of the current. O’Hagan R. et al. (2005) concluded that MEC-4 and MEC-10 must carry the mechanotransduction current. By directly demonstrating that the C. elegans DEG/ENaC complex composed of MEC-4, MEC-10, MEC-2, and MEC-6 forms a functional sensory mechanotransduction channel, O’Hagan R. et al. (2005) were the first to show that specific metazoan proteins are mechanotransducers. Based on similarities with C. elegans DEG/ENaC proteins, mammalian mechanosensitive BNaC channels were identified which are expressed in touch receptor cells (Garcia-Anoveros, J. et al., 2001). Targeted gene deletion of BNaC1 caused a loss of touch sensitivity in mice (Price, M. P. et al., 2000) but no deficits in auditory or vestibular function. Although the DEG/ENaC family of ion channels is of considerable interest to inner ear biologists, none of the members of the family have emerged as compelling candidates for the hair cell transduction channel.

Instead, attention has shifted to the TRP family of ion channels as several members were recently cloned from lower organisms, such as fruit flies and zebrafish, where they were found to be important for mechanotransduction (Walker, R. G. et al., 2000; Kim, J. et al., 2003; Sidi, S. et al., 2003). TRP channels were first described in flies, where mutations in the trp gene expressed by photoreceptor cells cause transient voltage responses to a continuous light stimulus (Clapham, D. E., 2003; Moran, M. M. et al., 2004).

The TRP superfamily is extremely diverse, and channels are identified as belonging to the family based on homology, rather than on their selectivity or ligand sensitivity. In mammals, the TRP superfamily contains six subfamilies: the classical TRPs (TRPCs), vanilloid receptor TRPs (TRPVs), melastatin TRPs (TRPMs), the mucolipins (TRPMLs), the polycystins (TRPPs), and TRPA1 (also called ANKTM1, for ankyrin transmembrane protein 1) (Moran, M. M. et al., 2004). The amino acid sequence identity among mammalian TRP channels is as low as 20%. TRP channels are involved in a diverse array of sensory processes, including hot and cold temperature sensation, the signaling of osmotic pressure, taste, alleged pheromone sensation, and mechanosensation (Clapham, D. E. 2003). TRP channels, with the exception of some TRPPs, are all predicted to form subunits consisting of six transmembrane (TM) domains with a pore-loop between TM5 and TM6, and may assemble into oligomers. They are generally nonselective, weakly voltage-sensitive, cation channels. Their properties are quite varied, and single TRP channels are often activated and modulated by multiple types of stimuli, second messengers, or other ligands (Clapham, D. E. 2003).

The fruitfly sensory bristle mechanotransduction channel, nompC or no mechanoreceptor potential C (later named TRPN1), was the first mechanosensitive TRP channel cloned (Walker, R. G. et al., 2000). Fly sensory bristles consist of a hair shaft whose base is attached to the dendritic tip of a bipolar sensory neuron. Deflecting the hair shaft compresses the dendritic tip, and gates the mechanotransduction channel. The fly bristle mechanotransduction process is remarkably similar to the mammalian hair cell transduction process, in that it exhibits directional sensitivity, a submillisecond response latency, sensitivity to nanometer-scale deflections, a steep I(X) relationship, and similar adaptation. Flies carrying loss-of-function mutations in the nompC gene exhibit transduction currents that are ∼10% of the wild-type response. Another nompC allele exhibits nearly normal transduction, but abnormally fast adaptation. The nompC gene was cloned, and found to encode a previously unidentified ion channel with 29 ankyrin (ANK) repeats on its NH₂-terminus, and six predicted transmembrane domains with a pore region between TM5 and TM6 (Walker, R. G. et al., 2000).

The first putative vertebrate mechanotransduction channel to be cloned was the zebrafish ortholog of nompC. Zebrafish nompC, like fly nompC, belongs to the TRPN family within the TRP superfamily. It also contains 29 ANK repeats in its NH₂-terminus. Zebrafish injected with anti-sense morpholinos to disrupt nompC expression lacked an acoustic startle response. Hair cells within the ear of injected zebrafish did not take up the styryl dye FM-1-43, which permeates the transduction channel. Microphonic potentials, which are generated by current flowing through the transduction channels in a population of hair cells, were abolished in morpholino-injected zebrafish. The amount of dye uptake and the reduction in microphonic potential amplitudes were correlated with the severity of the acoustic startle phenotype. Finally, all three phenotypes were reversible; 24 hours after the morpholinos washed out nompC expression and all of the phenotypes returned to normal. Sidi S. et al. (2003) searched the mouse and human genomes but nompC homologs were not found.

The remarkable similarities between fruitfly mechanosensory bristle transduction, zebrafish hair cell transduction, and vertebrate hair cell mechanotransduction suggest that they may all use similar mechanotransduction channels. Yet, since nompC is not found in mammals, what is the mammalian transduction channel? The properties of the channel, namely its large single-channel conductance and permeability to cations, as well as the blocking effect of extracellular calcium, all suggest that it could be a member of the TRP family of ion channels.

In 2004 TRPA1 was presented as a mammalian hair cell transduction channel candidate (Corey, D. P. et al., 2004, Nagata, K. et al., 2005), but the suggestion remains controversial. TRPA1 was previously identified as a nonselective cation channel activated by noxious cold (∼<17 °C) in a subset of nociceptive dorsal root ganglia neurons (Story, G. M. et al., 2003; Bandell, M. et al., 2004). TRPA1 is also activated by isothiocyanate compounds, such as mustard oil and wasabi, Delta(9)-tetrahydrocannabinol (THC) (Jordt, S. E. et al., 2004), allicin, an unstable component of fresh garlic (Macpherson, L. J. et al., 2005), and other pungent compounds such as in cinnamon oil (cinnamaldehyde), wintergreen oil (methyl salicylate), clove oil (eugenol), and ginger (gingerol) (Bandell, M. et al., 2004).

Corey D. P. et al. (2004) presented TRPA1 as a candidate for the mammalian hair cell transduction channel based on several lines of indirect evidence. Corey D. P. et al. designed in situ hybridization probes for 33 TRP channels within the mouse genome, and looked for expression in the mouse ear. Only TRPA1 labeled hair cells above background levels. Hair cells of the mouse utricle first become sensitive to mechanical stimuli around embryonic day 17 (Géléoc, G. S. and Holt, J. R., 2003a). Quantitative RT-PCR was used to assess mRNA levels of various TRP channels during development of the mouse utricle, and TRPA1 levels were found to steadily increase to a prominent peak at E17. An antibody to the carboxy terminus of TRPA1 labeled the upper half of bullfrog stereocilia, along the length of the kinocilium, and the pericuticular zone around the cuticular plate. Light labeling was seen in the cell body. In mouse vestibular hair cells, light stereociliary labeling could be seen, along with stronger labeling within the kinocilium and pericuticular zone. In mouse auditory hair cells, labeling was the strongest in the stereocilia, but could be seen in the cell bodies (Figure 8(c)). Labeling at stereociliary tips could be seen in auditory hair cells from animals older than postnatal day 4. Interestingly, TRPA1 labeling in the stereocilia, like that of cadherin-23, was greatly diminished after treating cells with BAPTA or La³⁺ to break the tip links (Corey, D. P. et al., 2004).

Corey D. P. et al. (2004) first inhibited TRPA1 expression in zebrafish hair cells. In zebrafish embryos injected with antisense morpholinos designed to inhibit TRPA1 protein expression, FM1-43 dye uptake by hair cells was significantly reduced, as compared to controls. Microphonic potentials, which result from current flow through transduction channels from multiple hair cells, were also reduced. TRPA1 morpholinos did not appear to simply delay hair cell development, as control and TRPA1 morpholino-injected fish had a similar number of hair cells. Finally, RT-PCR of TRPA1 message and immunoblots of extracted protein confirmed that TRPA1 levels were reduced in TRPA1 morpholino-injected animals as compared to control-injected animals. If TRPA1 and TPRN1 (nompC) are both required for hair cell transduction in zebrafish, it may be because they form heteromultimers together, or possibly because individual hair cells use one or the other (Corey, D. P. et al., 2004).

Next, the functional consequences of inhibiting TRPA1 expression in mouse hair cells were investigated. Two hairpin small interfering RNAs (siRNAs) that targeted TRPA1 were packaged into adenoviral vectors along with the gene for green fluorescent protein. Mouse utricular hair cells were infected with the TRPA1 adenoviruses at embryonic day 16, before the onset of mechanotransduction. The authors recorded transduction currents at embryonic day 17–20 from siRNA treated cells, cells exposed to a control vector that expressed GFP alone and uninfected control cells. Both siRNA constructs yielded cells with reduced FM1-43 uptake and transduction current amplitudes relative to uninfected control cells (Figure 8(d)). From these data, the authors concluded that expression of siRNAs directed against TRPA1 resulted in a loss of hair cell mechanotransduction (Corey, D. P. et al., 2004).

Based on the immunolocalization data, expression analysis, knock-down, and functional data, TRPA1 seemed a reasonable candidate for the hair cell transduction channel. However, recent data call that suggestion into question. Two groups generated mice that carried a targeted gene deletion of TRPA1, and both found that hearing and balance function in the knockout animals was identical to wild-type controls (Bautista, D. M. et al., 2006; Kwan, K. Y. et al., 2006). Furthermore, Kwan K. Y. et al. (2006) found normal FM1-43 uptake and mechanotranduction currents in TRPA1^−/− animals. Several possibilities exist that may reconcile the observations of Corey D. P. et al. (2004) and Nagata K. et al. (2005) with those of Kwan K. Y. et al. (2006) and Bautista D. M. et al. (2006). First, TRPA1 could be an integral member of the transduction apparatus, coupling (either directly or indirectly) the actual transduction channel to the cytoskeleton using its ankyrin repeats (see Section 3.16.8.7). TRPA1 localizes along the kinocilium, and, like cadherin-23, at kinocilial links between the tallest stereocilia and the kinocilium. With its long ankyrin domain, TRPA1 could simply play a structural role in this location, as the kinocilium itself is not mechanosensitive. It is possible that TRPA1 is a nonessential structural component of the mechanotransduction apparatus. Secondly, it could be argued that TRPA1 plays a purely developmental role. Disruption of TRPA1 expression with siRNA- adenoviruses was performed between embryonic day 15 and 18 (Corey, D. P. et al., 2004) whereas TRPA1-deficient mice were examined at later postnatal and adult stages (Kwan, K. Y. et al., 2006). Third, the siRNAs designed to inhibit TRPA1 expression may have had unintended, nonspecific effects on the expression of other molecules required for transduction. Lastly, it remains a possibility that expression of some other transduction channel subunit compensated for the lack of TRPA1 expression in the knockout mice. Whatever the reason, two important questions remain. If TRPA1 is not the transduction channel, what is? And furthermore, what is the function of TRPA1 in hair cells?

Another TRP channel expressed in hair cells is TRPML3 (Di Palma, F. et al., 2002). Mutations in TRPML3 are responsible for hair bundle disorganization, circling behavior, and hearing loss in the varitint-waddler mouse. TRPML3 localizes primarily to organelles within the cell body, suggesting that deafness results from abnormal organelle trafficking, rather than from defects in transduction. However, TRPML3 has not been ruled out as candidate for the hair cell transduction channel. A third candidate is TRPV4 (previously named VR-OAC), a mechanically sensitive channel that is activated by osmotic changes (cell swelling), fluid flow, and by heat. TRPV4 is expressed by multiple cells types within the inner ear, including the hair cells and the marginal cells within the stria vascularis. TRPV4 knockout mice have hearing deficits, but the deficits only appear in older mice and only shift hearing thresholds by about 20 dB. The mechanical activation of TRPV4 is also relatively slow, suggesting that a second messenger is involved, which is inconsistent with the rapid kinetics of the hair cell transduction channel (Corey, D. P., 2006). Although multiple candidates have been identified, the molecular identity (or identities) of the hair cell mechanotransduction channel is still unknown.

Alfaxalone

Alfaxalone, solubilized in 2-hydroxypropyl β-cyclodextrin (Alfaxan®; henceforth termed alfaxalone) has been developed, initially in Australia by Jurox Pty Ltd, and (2012) is licensed as an anaesthetic drug for use in cats and dogs in Australia, New Zealand, South Africa and Europe. It is presented as an aqueous solution of 10 mg/mL, in bottles. There is no preservative, and it is recommended that an opened bottle vial be used within a day.

Alfaxalone is relatively short acting, non-irritant and is non-cumulative, making it suitable both for IV induction of anaesthesia and TIVA by bolus injection or by infusion. The dose recommended by Jurox for induction of anaesthesia in non-premedicated dogs is 2 mg/kg IV given slowly over 60 seconds, although the current European data sheet suggests that up to 3 mg/kg might be needed on occasions. In premedicated dogs, reduced doses are used depending on the degree of sedation. For maintenance of anaesthesia, recommended doses are, in non-premedicated dogs, 0.13–0.16 mg/kg/minute or boluses of 1.3–1.5 mg/kg and, in premedicated dogs, 0.1–0.12 mg/kg/minute or boluses of 1–1.2 mg/kg. Cats appear to need higher doses, 5 mg/kg, again given slowly IV, being recommended for induction of anaesthesia. Suggested doses for maintenance in cats are, for non-premedicated 0.16–0.18 mg/kg/minute or boluses of 1.6– 1.8 mg/kg and for premedicated cats, 1.1–1.3 mg/kg/minute or boluses of 1.1–1.3 mg/kg. There appear no contraindications to use of alfaxalone with any of the usually used premedicant sedative, analgesic or anticholinergic agents, but administration together with other injectable anaesthetic agents is specifically contraindicated. Some of the experimental work supporting these claims has yet (2012) to be published in peer-reviewed journals but abstracts of presentations are available (www.Alfaxan.co.uk).

Alfaxalone is rapidly metabolized in the liver. The pharmacokinetics in dogs have been investigated (Ferre et al., 2006) at both the usual IV induction dose (2 mg/kg) and at a supramaximal dose (10 mg/kg). At 2 mg/kg and 10 mg/kg respectively, the mean duration of anaesthesia until the dog responded to the presence of an endotracheal tube was 6.4 and 26.2 minutes; plasma clearance 59.4 and 52.9 mL/kg/minute, plasma terminal half-life (t_1/2) 24.0 and 37.4 minutes, and for both, volume of distribution between 2 and 3 L/kg. The authors concluded there was no clinically significant difference in pharmacokinetic parameters between the doses. In greyhounds (Pasloske et al., 2009), following 2 mg/kg alfaxalone IV, t_1/2 was 34.3 minutes and plasma clearance 48.4 mL/kg/minute, suggesting alfaxalone is slightly longer acting in this breed, but the difference is unlikely to be of major clinical significance. A similar study in cats (Whittem et al., 2008) examined the pharmacokinetics of 5 mg/kg and 25 mg/kg IV; results for these two doses were, respectively; plasma clearance 25.1 and 14.8 mL/kg/minute and t_1/2 45.2 and 76.6 minutes. The cat study also investigated continuing anaesthesia by four incremental doses; the duration of subsequent anaesthesia resulting from each increment did not change, suggesting that at least over this dosage schedule there were no clinically evident cumulative effects.

Safety studies, giving the intended dose, and three and ten times its multiple in both dogs and cats have been published (Muir et al., 2008, 2009). In dogs, this dosing schedule was 0 (placebo), 2, 6 and 20 mg/kg; in cats 0, 5, 15 and 50 mg/kg of alfaxalone. Rescue IPPV was allowed if necessary.

In dogs, induction of anaesthesia was smooth and rapid; for 2, 6 and 20 mg/kg respectively, the duration of lack of response to noxious stimuli was 9.3, 32 and 69.7 minutes and duration of anaesthesia (to extubation) 9.8, 31.4 and 75.1 minutes. Quality of anaesthesia appeared excellent and, from the results, antinociception appeared to last throughout the whole period. Recovery was not described, but scores did not differ between doses. Alfaxalone caused an increase in heart rate, and a fall in arterial blood pressure, cardiac output and systemic vascular resistance, but the changes were minimal except at 20 mg/kg and, even at this dose, were clinically acceptable. Respiratory depression occurred at all doses; apnoea was common particularly at the higher doses.

In cats, induction of anaesthesia was smooth and rapid; for 5, 15 and 50 mg/kg respectively, the duration of lack of response to noxious stimuli was 15.3, 48.4 and 143.7 minutes and duration of anaesthesia 26, 75.4 and 172 minutes. Thus, in contrast to dogs, antinociception did not appear to last for the whole period of anaesthesia. Recovery quality was scored as excellent after doses of 5 and 15 mg/kg, but 5 hours after receiving 50 mg/kg, five (of eight) cats had not fully recovered and were euthanized. Alfaxalone caused a dose dependent fall in heart rate, arterial blood pressure, cardiac output and systemic vascular resistance, but the changes at 5 and 15 mg/kg were rarely statistically and unlikely to be clinically significant. At 50 mg/kg (i.e. a 10× overdose), cardiovascular depression was very marked. Apnoea was common at all doses. Even at 5 mg/kg, PaO₂ fell and 5/8 cats required IPPV at some period; by 15 mg/kg all cats received IPPV. In summary, in cats, doses of up to 15 mg/kg produced minimal or clinically acceptable cardiovascular changes, but respiratory depression was a potential problem even at 5 mg/kg. The dose of 50 mg/kg was obviously an unacceptably high overdose but, even then, no cat died at the time of drug administration.

The peer-reviewed publication of the pharmacokinetic and safety studies described above demonstrate that IV alfaxalone produces good quality anaesthesia in dogs and cats, is non-cumulative, at clinically applicable doses cardiovascular effects are minimal but (in contrast to Saffan in cats), dose-dependent apnoea and respiratory depression may occur. In clinical practice, however, it is usual to use ancillary sedative and analgesic agents. There are now a number of publications, both clinical and experimental, investigating such combinations for induction of anaesthesia, and often comparing alfaxalone with other anaesthetic regimens. In dogs, alfaxalone did not cause pain on injection (Michou et al., 2012). To compare the incidence of apnoea induced by alfaxalone or propofol in dogs (Keates & Whittem, 2012), a dose-escalation study found the median dose of alfaxalone which induced apnoea was 10 mg/kg (i.e. 5× recommended doses) and for propofol was 13 mg/kg (i.e. twice the normal induction dose). Amengual et al. (2013), however, found that, in premedicated dogs, the incidence of post-induction apnoea was high (around 50%) with either alfaxalone (1.5 mg/kg) or propofol (3 g/kg) given by fast IV injection, reinforcing the data sheet advice for Alfaxan® that it should be given slowly. An experimental study in non-premedicated dogs (Rodriguez et al., 2012), compared the cardiorespiratory effects of a mean of 4.15 mg/kg alfaxalone (i.e. above the recommended dose) with a mean of 2.91 mg/kg etomidate; at these doses alfaxalone increased (statistically but not clinically significantly) heart rate and cardiac index, with a non-significant fall in arterial blood pressure and systemic vascular resistance. Quality of induction and recovery were better than those of etomidate. Maddern et al. (2010) showed that, in dogs, the mean dose of alfaxalone needed for anaesthetic induction was 1.2 mg/kg after premedication with medetomidine 4 µg/kg or butorphanol 0.1 mg/kg but, in dogs receiving both premedicants, the alfaxalone dose was significantly reduced to 0.8 mg/kg. Psatha et al. (2011) compared the use of alfaxalone to a diazepam/ketamine/propofol combination for methadone-premedicated dogs in ASA grade 3 or worse; they found that 1–2 mg/kg alfaxalone gave satisfactory induction of anaesthesia, systolic arterial blood pressure did not fall from pre-induction values and that cardiovascular parameters throughout the subsequent gaseous maintenance of anaesthesia did not differ between treatments. Alfaxalone can be used satisfactorily for induction of anaesthesia in premedicated (acepromazine, morphine and atropine) puppies aged 6–12 weeks (O'Hagan et al., 2012a); dose rates were similar to those seen in adult dogs.

In cats premedicated with acepromazine and meloxycam, mean doses of alfaxalone required for induction of anaesthesia and endotracheal intubation were 4.7 mg/kg (Taboada & Murison, 2010). In young cats (>12 months), the mean dose needed to induce anaesthesia in non-premedicated cats was 4.2 mg/kg but, in cats premedicated with acepromazine and butorphanol, was 2.7 mg/kg (Zaki et al., 2009). Recovery quality was better in premedicated cats. O'Hagan et al. (2012b) found that the mean dose of alfaxalone required for induction for endotracheal intubation in heavily premedicated kittens (acepromazine, morphine and atropine) under 12 weeks of age was 4.7 mg/kg.

The studies described above all support the information supplied by the manufacturers that IV alfaxalone administered over a minute leads to a smooth induction of anaesthesia, cardiovascular changes are minimal (increased heart rate and sometimes a small fall in arterial blood pressure) even in dogs with a compromised circulation; there may be transient apnoea but this is no more, and probably less, likely to occur than after propofol. However, in many of the above mentioned studies, although recovery was usually classified as excellent or acceptable, there is mention of occasional incidences of twitching and paddling in recovery, even following relatively prolonged anaesthesia with inhalant agents (Jimenez et al., 2012; Mathis et al., 2012).

Alfaxalone for TIVA is of particular interest as this was one of the major uses of Saffan in cats. In an experimental study in dogs sedated with acepromazine and hydromorphone, Ambros et al. (2008) compared alfaxalone and propofol for TIVA. Cardiovascular parameters (including cardiac output) did not differ between treatments. Respiratory depression was greater with propofol. Recovery times and quality did not differ between treatments. Two studies have examined alfaxalone TIVA for ovariohysterectomy in dogs, in each case the dogs received oxygen supplementation. Suarez et al. (2012) premedicated the bitches with acepromazine (0.01 mg/kg) and morphine (0.4 mg/kg) then used alfaxalone or propofol for induction and for maintenance of anaesthesia. Doses required for maintenance were 0.1–0.02 mg/kg/minute for alfaxalone and 0.3–0.5 mg/kg/minute for propofol. Both agents provided adequate anaesthesia for the purpose. There were no significant differences in cardiopulmonary parameters or recovery parameters between treatments, but both caused marked respiratory depression. Herbert et al. (2013) compared premedication with buprenorphine (0.02 mg/kg) and either acepromazine (0.05 mg/kg) or dexmedetomidine (approximately 0.01 mg/kg) prior to induction and TIVA with alfaxalone. Alfaxalone infusion rate was significantly lower after dexmedetomidine (median 0.08 range 0.06–0.19 mg/kg/minute) than after acepromazine (median 0.11 range 0.07–0.33). Beths et al. (2012) premedicated cats with medetomidine (0.02 mg/kg) and morphine (0.3–0.5 mg/kg); mean induction and infusion rates were 1.8 mg/kg and 0.18 mg/kg/minute and apnoea and hypoventilation were common. Quality of anaesthesia was better, but quality of recovery worse in the medetomidine group. In cats, alfaxalone as a total intravenous anaesthetic was used in eight very young (<12 weeks) premedicated kittens (O'Hagan et al., 2012b) to enable ovariohysterectomy; following induction dose the first increment was required at 5.5 minutes; bolus administrations to maintain anaesthesia amounted to 0.18 mg/kg/minute. Mean recovery time (in some cases after atipamezole) was 40 minutes. Apnoea did not occur. Noise induced twitching was noted in three of 37 cats. In all these studies, premedication used moderately high doses of opioid, and/or α₂-agonists; these were considered necessary to contribute to perioperative analgesia but may well have influenced both respiratory depression and time and quality of recovery. There is still need to elucidate the best combination with alfaxalone to enable surgery.

Alfaxalone has been used in rabbits, doses of 2–3 mg/kg IV providing satisfactory conditions for endotracheal intubation prior to administration of inhalation agents (Grint et al., 2008). Alfaxalone can be used by IM injection. Saffan was used by this route routinely in cats. For alfaxalone, it is probable that combinations with sedatives will be needed to ensure a practicable volume. Alfaxalone has already been used by a variety of routes in exotic animals. IM administration of 20 mg/kg to red-eared sliders (turtles) resulted in a smooth induction of sedation/anaesthesia of short duration (Kischinovsky et al., 2013). Alfaxalone by immersion followed by branchial/transcutaneous irrigation proved adequate for surgery on an axolotl (McMillan & Leece, 2011). In the green iguana, 10 mg/kg alfaxalone IM provides light sedation, 20 mg/kg short duration anaesthesia and 30 mg/kg anaesthesia for 40 minutes (Bertelsen & Sauer, 2011). Approximately 10 mg/kg IM in marmosets provides adequate sedation to allow mask induction or IV injection (Thomas et al., 2012). Combinations with medetomidine have been used IM to immobilize wallabies (Bouts et al., 2011). These early studies suggest that alfaxalone will have a major role to play in the immobilization of exotic animals.

In larger animals, IV alfaxalone, in combination with heavy sedation, has been used successfully to anaesthetize horses, both as an induction agent and as ‘top ups’ to enable castration (Leece et al., 2009; Kloppel & Leece, 2011; Keates et al., 2012). The pharmacokinetics have been investigated in adult horses and in foals (Goodwin et al., 2011, 2012). Alfaxalone has also been used in pigs (Keates, 2003) and sheep (Andaluz et al., 2012; Torres et al., 2012; Walsh et al., 2012), in all cases providing satisfactory short duration anaesthesia.

Alfaxalone in cyclodextrin has been available in veterinary practice in Australia for some years and, by August 2010 it was licensed in Australia, New Zealand, Thailand and in six European countries. Its use in small animals as an induction agent has become very popular; studies in use as TIVA are less well developed. Its potential for causing respiratory depression is accepted but, to date, there have been no published reports of any unexpected side effects. It is anticipated that this agent will prove very useful in small animal and other areas of veterinary anaesthetic practice.

5.3 Other experimental parameters

In addition to the experimental parameters discussed above, other considerations may affect the results. Christenson et al. (2016) found that MnO₂ formed on the surface of permanganate candles and impaired its longevity [28]. While it is feasible to periodically scrape off the oxide from the surface of candles installed in permanent wells, this will require on-going maintenance and the labour cost could be high. Subsequently, the researchers added an anti-scaling polyphosphate agent, sodium hexametaphosphate (SHMP) to their CRMs [34]. The type of pollutant considered is another important experimental parameter. In the study by Kambhu et al., (2012) benzoic acid was used as a surrogate for BTEX because it is less volatile and, therefore, does not require zero headspace apparatus to preserve the mass balance [39]. This rendered the experiments easier to perform but less representative. When colloidal silica was used for making CRM gel, the gelation process was affected by the pH, solid content, and ionic strength of the solution [77]. Xiong et al. (2016) found that pH, temperature, and dissolved oxygen all had effects on the release characteristics of a wax-based KMnO₄ CRM [55].

4.3.3 Chitosan

Chitosan (CS), a natural polymer derived from deacetylation of chitin, and comprised of arbitrarily distributed β-(1-4)-linked D-glucosamine (deacetylated unit) and N-acetyl-D-glucosamine (acetylated unit), is one of the few positively charged polysaccharides that can be used to form hydrogel (Bhattarai et al., 2010). Due to the abundance of amino groups in its chain, CS exhibits pH-sensitive behavior and contains weak polybases. The pH-sensitive swelling process is based on the protonation of its amine group at low pH. Effects of protonation include chain repulsion, diffusion of protons and counterions with water in the gel, and dissociation of secondary interaction (Yao et al., 1994). The presence of ionizable amino groups on chitosan resulted in its cationic charge, which had a substantial impact on its ability to create hydrogels (Abreu et al., 2012; Luo et al., 2011). The electrostatic interaction between positively charged amine (pK_a = 6.5) in chitosan polymer chain and negatively charged sialic acid residue on mucoglycoprotein affects the mucoadhesive properties of chitosan (Menchicchi et al., 2014). It is highly possible to design adhesive hydrogels as future drug delivery vehicles by forming disulfide bonds between thiomers and mucus glycoproteins (mucins), protecting therapeutic drugs from unfavorable environments of the upper GIT, and having colonic flora break down chitosan's glycosidic bonds to specifically release the drugs encapsulated there (Bhattarai et al., 2010).

CS hydrogels are often created through complexation with anionic biopolymers or multivalent ions like tripolyphosphate (Yang et al., 2021). compared to sodium hexametaphosphate, the hydrogel that was crosslinked by sodium tripolyphosphate (STPP) shown a higher capacity for expansion and a more steady release spectrum (no more than 10% in the upper GIT), which might transfer bioactive ingredients to the colon. Electrostatic contact causes the surface of the cs-based hydrogel to become hydrophilic, which promotes Caco2 cell growth. Cell viability for the cs-based hydrogel ranges from 118.86% to 147. 22%. Additionally, variable hydrogel release characteristics and safety were produced by using different crosslinker concentrations during the crosslinking process.

3.2.3.2 Solubility

The alginic acid and its polyvalent metal salts exert the insolubility in water, but the rest of the alginic acid salts along with the alkali metal (e.g., K⁺, Na⁺) and quaternary ammonium and ammonium compounds exhibit solubility. Alginate shows solubility in ketones and alcohols, but insolubility in milk and hard water because these contain Ca²⁺ [52]. If sodium alginate is required to be added in such solution, a chelating agent such as sodium hexametaphosphate or EDTA (ethylene diamine tetraacetic acid) must be used to sequester Ca²⁺. Propylene glycol alginate (80%–85% esterified) can be used in milk as it is usually less affected by Ca²⁺ ions. It remains soluble and can be used in milk when pH is lowered to about pH 2 [53].

16.13.2.3 Preparation of Porous Ceramic Support

Some experience is needed to perform the last step efficiently. From our previous work, the strength of the ceramic support will increase with a decrease in the amount of dispersion solution.

The advantage of sol–gel technology is the ability to produce a highly pure γ-alumina and zirconia membrane at medium temperatures, about 700 °C, with a uniform pore size distribution in a thin film. However, the membrane is sensitive to heat treatment, resulting in cracking on the film layer. A successful crack-free product was produced, but it needed special care and time for suitable heat curing. Only γ-alumina membrane have the disadvantage of poor chemical and thermal stability.