Keywords

1 Introduction

Perna canaliculus belongs to the class Bivalvia , the phylum Mollusca and family Mytilidae . The genus Perna contains species of both green and brown mussels located predominantly in the Southern Hemisphere but also found in North Africa and the northern coasts of South America, with Paleontological data dating the genus back to the Eocene period (60 million years ago) (Wood et al. 2007). Three well-defined species are recognised in the Perna genus that includes P. viridis (Asian green mussel ) found through Indo-Pacific regions, P. perna (brown or rock mussel) found through Atlantic regions and P. canaliculus which is endemic to New Zealand waters only and has been commercially and sustainably farmed since the early 1970s (Wood et al. 2007). P. canaliculus is distinguished from other mussel species by the bright green stripe around the posterior ventral margin and inside the lip of its shell (see Fig. 1) (Wolyniak et al. 2005). Numerous bioactive compounds have been identified in both the Mytilus and Perna genera of mussels, but it is P. canaliculus that has been most comprehensively studied for medicinal purposes. It has supported the development of commercial therapeutic products to treat arthralgia in humans and animals. It has also been assessed as an adjunct therapy for rheumatoid arthritis (RA), asthma and gastrointestinal tract (GIT) complaints (Gibson et al. 1980; Gibson and Gibson 1998; Coulson et al. 2012; Mickleborough et al. 2013). P. canaliculus is manufactured in New Zealand as unadulterated freeze-dried whole (i.e. without shell) extract of the mussel meat; as whole with the lipid fractions removed and as a concentrated lipid extract only.

Fig. 1
figure 1

Perna canaliculus a mussel shell b mussel in shell, note green lip on inside and outer posterior ventral margin (used with permission from Aroma New Zealand)

Full size image

1.1 Growing and Harvesting Perna canaliculus

Interest in the application of P. canaliculus for arthritic conditions began in the 1960s, when research was undertaken to discover new natural compounds from marine organisms, which included P. canaliculus to treat cancer. P. canaliculus extract did not provide significant results for cancer outcome measures, but it was found that the study participants who also suffered from arthritis reported less pain and stiffness and improved mobility when taking the extract (Kendall 2000). It was also observed that coastal Maoris of New Zealand, whose staple diet consisted of P. canaliculus had a lower incidence of arthritis than Maoris residing in land (Halpern and Georges 2000; Brien et al. 2008a). Research has therefore focused on the anti-inflammatory capabilities of P. canaliculus extract and its fractions. Mussels are farmed in New Zealand using long-line technology around sheltered, in-shore areas such as the Marlborough Sounds . The spat, or seed, is collected from farmers suspending the spat catching lines or by collection of seaweed that spat (<5 mm in size) have naturally adhered to. The collected spats are then resettled onto nursery lines and grown for 3–6 months (10–30 mm) after which the juveniles are resettled onto thicker ropes and grown to maturity (90–120 mm) for another 12–18 months depending on the growing conditions. Once harvested, mussels are quickly transported to the processing plant where they are chilled (<10°C) in holding tanks. Mussels are then placed in a continuous centrifuge that separates the meat from the shell, which is then placed in refrigerated tanks and a natural anti-oxidant is added to improve stability. Mussel meat is then freeze-dried (lyophilised) at −20°C for 20–22 h. The freeze-dried product is then milled into a fine powder (FAO 2014; Aroma 2014). The lipid extract from the stabilised freeze-dried mussel powder is obtained by a supercritical fluid extraction process (SFE) using liquefied carbon dioxide (CO2).

1.2 Nutritional Content of Perna canaliculus

Whole P. canaliculus extract consists of a complex mixture of compounds being predominantly 55–60 % protein, 5–15 % carbohydrates, 5–15 % glycosaminoglycans (including chondroitin sulphate and heparin), 3–5 % lipids, 5 % minerals and 0.5–4 % water (Ulbricht et al. 2009). Vitamins A, D3, E and B12 are also present. The concentrated lipid extract contains a complex profile of fatty acids classes including sterol esters of cholesterol and desmosterol/brassicasterol , triglycerides , free fatty acids (FFAs) , sterols and phospholipids (Ulbricht et al. 2009; Whitehouse et al. 1997; Murphy et al. 2003). Analytical assessment of aqueous and lipid metabolomes by nuclear magnetic resonance (NMR) and gas chromatography-mass spectrometry (GC-MS) can clearly demonstrate constitutive differences in mussel species, for example, between P. canaliculus and Mytilus galloprovincialis (Australian Blue mussel ). There are distinguishing patterns of amino acids, several metabolites, glucose and lipids between the two species, although some of these differences could in part be due to location rather the species (Rochfort et al. 2013). Heavy metals that accumulate in water such as arsenic, mercury, cadmium and lead are also present in the whole mussel meat due to their filter-feeding behaviour; however, heavy metal limits are rigorously monitored (see Table 1). Furthermore, the growing waters from which the mussels are harvested are monitored weekly for biotoxins . If levels exceed the legislational limit in New Zealand, no harvesting of the mussels can take place.

Table 1 Whole P. canaliculus extract: typical nutritional evaluation (Source Aroma NZ Ltd and Biolane®)
Full size table

There are no standardisation procedures in place for P. canaliculus raw material suppliers. This may result in potential variability between the nutrient profiles and stability of the raw materials. Temperature and season both affect the nutrient profile of the mussels during harvest. As the key constituent responsible for the observed therapeutic activity is not definitively known, lack of standardisation practises may influence therapeutic efficacy between marketed products (Whitehouse et al. 1997, 1999). Early human clinical studies for both osteoarthritis (OA), RA and animal studies have resulted in variable assessments of the efficacy of P. canaliculus whole extract powders, reporting both positive (Gibson et al. 1980; Audeval and Bouchacourt 1986) and negative outcomes (Highton and McArthur 1975; Huskisson et al. 1981; Caughey et al. 1983; Larkin et al. 1985). It was only in 1986 when New Zealand manufacturers began stabilising the whole mussel extracts with 3 % tartaric acid (a metal chelator and anti-oxidant) immediately after removing the flesh from the shell, preventing auto-oxidation , when the activity of P. canaliculus whole extract powders began to demonstrate more potent activity (Whitehouse et al. 1997).

2 Clinical Therapeutic Activity

Current opinion concerning the therapeutic efficacy of P. canaliculus whole extract and/or the lipid fraction is that, the existing evidence is rather inconclusive for the treatment of OA symptoms, with overall evidence for RA suggesting inefficacy (Brien et al. 2008b; Cobb and Ernst 2006; Ulbricht et al. 2009). Individual studies to assess P. canaliculus (whole and lipid extract) for treating joint symptoms of OA have all reported a clinically relevant reduction in joint pain and stiffness. Clinical studies assessing P. canaliculus for OA, RA and asthma are presented in Table 2. Systematic reports of little or no conclusive evidence from the available studies are generally due to poor methodological rigour, variations in product stability and dosing, lack of raw material standardisation and use of inappropriate placebos, such as dried fish powder. The majority of studies assessing P. canaliculus for RA were conducted in the mid-1970s and early 1980s; the results of which may have been influenced by the lack of product stability during this period. Assessment of efficacy is also difficult due to the variable prescribed dosing patterns (dose and duration) used for both OA and RA symptoms. Importantly, the use of rescue pain medication , in the form of either acetaminophen (paracetamol) or non-steroidal anti-inflammatory (NSAID) medications were inconsistent and poorly reported in these earlier studies and may have further influenced the interpretation of the results. It is now recognised; however, that P. canaliculus may have credible pharmacoactivity as demonstrated in animals and in vitro studies, which requires further rigorous scientific investigations to assess efficacy and optimal dosage in humans (Rainsford and Whitehouse 1980; Brien et al. 2008b).

Table 2 Human clinical studies assessing the therapeutic activity of P. canaliculus whole extract and SFE lipid-rich fraction
Full size table

The lipid fraction is obtained by supercritical fluid extraction (CO2-SFE) from the stabilised, freeze-dried mussel powder that is then combined with olive oil and vitamin E as an anti-oxidant. By using CO2 as an extracting medium, high temperatures and solvents are not used for extraction, thus maintaining therapeutic activity of the extract. Processing with high temperature activates degrading enzymes within the mussel, namely phospholipases and lipoxygenases that degrades the lipid components (Grienke et al. 2014; Wakimoto et al. 2011). Further, fractionation and analysis of the active components in the whole lipid fraction is difficult due to their instability and concomitant decomposition during the bioassay process (Wakimoto et al. 2011). It is reported that the stabilised SFE lipid fraction significantly improves asthmatic symptoms (hyperpnea-induced bronchoconstriction and mild to moderate atopic asthma ) when compared to placebo in humans (see Table 2) (Mickleborough et al. 2013; Emelyanov et al. 2002). The lipid extract has also been assessed for its anti-inflammatory activity with the assessment of serum inflammatory markers such as TNF-α and IL-1β , but with non-significant results (Murphy et al. 2006). The cohort recruited, however, were healthy individuals that did not demonstrate any signs of inflammation, with serum cytokine markers within normal reference ranges before intervention. Such a design is unlikely to answer the question of whether the lipid extract reduces markers of inflammation.

2.1 Dosing

The optimal therapeutic dose of either the whole or lipid extract has not been clearly ascertained and can only be estimated from previous clinical research. It is clear that carefully designed Phase I dose-ranging studies are required to ascertain what the effective prescribed dose should be for both extracts and in various disease contexts. Clinical studies have used a dose range between 1,050 and 3,000 mg/day for the whole extract and 210–1,200 mg/day for the lipid extract in OA patients with dose duration between 8 and 24 weeks. A dose between 900 and 1,180 mg/day of the whole extract and 140 mg/day of the lipid extract has been assessed in RA patients with a dose duration between 8 and 24 weeks and a dose between 300 and 1,200 mg/day of the lipid extract has been assessed for asthma patients with a dose duration of 8 weeks (see Table 2). The recommended dose from the manufacturer for the P. canaliculus whole extract is typically 1,500 mg/day and 200 mg/day for the lipid extract. It is unclear, however, if these are the effective therapeutic doses for each extract based on the presented research dose variations.

2.2 Adverse Effects and Toxicity

Adverse events reported in clinical studies assessing both the whole and lipid extract have included mild gastrointestinal events such as reflux, flatulence, epigastric discomfort, fluid retention, nausea and altered bowel habits, headaches and a transient increase in knee symptoms. Apart from these minor events, the intake of P. canaliculus is not associated with any serious adverse events and is generally well tolerated. The use of P. canaliculus should, however, be avoided by people with allergies to shellfish . Heavy metal poisoning is unlikely to occur (Ulbricht et al. 2009); however, biotoxins may be found in shellfish due to their filter-feeding behaviour and ingestion of large amounts of algae. The majority of the mussel’s diet consists of nutrient-rich eukaryotic microalgae , typically diatoms and dinoflagellates, but it is mainly when they ingest harmful algal blooms from the surrounding water that toxins become a serious threat to both the health of the consumer and the mussel industry (Grienke et al. 2014). In New Zealand, the growing waters from which the mussels are harvested are tested weekly for biotoxins and if the levels exceed the National limit harvesting is prohibited. The very strict guidelines now in place ensure that there is very low risk of the mussel products containing any biotoxins. To date, there are no reports of either GLM extract or the lipid extract interacting adversely with pharmaceutical or nutraceutical medications, but rather they may enhance their therapeutic effect (Rainsford and Whitehouse 1980; Whitehouse and Butters 2003).

2.3 Gastroprotective Activity

The SFE lipid-rich fraction has exhibited synergistic anti-inflammatory therapy when combined orally with NSAIDs and analgesic pharmaceutical medications such as prednisone, pentoxifylline or meloxicam in rat models. When administered as tandem therapies to reduce paw swelling in adjuvant-induced arthritis and zymosan-induced paw inflammation in rats, the result was more effective than using therapy alone (Whitehouse 2004; Whitehouse and Butters 2003). The whole stabilised extract powder also demonstrates equally synergistic anti-inflammatory activity when combined with prednisone or meloxicam. Both extracts have NSAID and steroid-sparing effects when administered concomitantly, reducing the effective dose required for the drug and also protecting the gastrointestinal tract (GIT) from the adverse effect of such medications. The stabilised whole extract powder reduced the occurrence of gut lesions in rats, more than the lipid fraction, when combined with NSAIDs in an animal model (Whitehouse and Butters 2003). Lipid and whole extracts reinforce the anti-inflammatory therapeutic activity of acetylsalicylic acid and indomethacin while concomitantly exhibiting gastroprotective activity (Rainsford and Whitehouse 1980). Supportive data indicates that supplementation with whole P. canaliculus extract may support GIT function and even show gastroprotective activity when administered with anti-inflammatory and analgesic medications in patients with OA (Coulson et al. 2012, 2013). Further, preliminary evidence has demonstrated that the SFE lipid-rich fraction significantly reduced colonic damage in an inflammatory bowel disease (IBD) animal model (Tenikoff et al.2005) and partially improved selected indicators of intestinal inflammation and intestinal morphology in an animal model of chemotherapy-induced mucositis.

3 Bioactive Metabolites of Perna canaliculus

Studies to identify bioactive metabolites within P. canaliculus products has led to the evaluation of extracts, hydrolysates and purified components from the fractionated lipids, carbohydrates and proteins present in the mussel meat. A review of these fractions is discussed in depth by Grienke 2014 (Grienke et al. 2014). The fraction(s) responsible for the therapeutic efficacy demonstrated in the OA disease model, both human and animal, is not yet fully defined. Previous claims suggest that the lipid fraction represents the dominant anti-inflammatory component of P. canaliculus (Ulbricht et al. 2009; Halliday 2008). Early clinical trials reported mixed results with whole extract powder. Stabilisation of the mussel meat with 3 % tartaric acid in the 1980s resulted in a more active product (Whitehouse et al. 1997). Furthermore, while Gibson and Gibson (1998) comparing the lipid extract to the whole stabilised powder extract in treating joint symptoms of both RA and OA, both mussel preparations demonstrated significant therapeutic activity with no substantial difference found between either treatment (Gibson and Gibson 1998). The major classes of compounds found in mussel meat (peptides, carbohydrates and lipids) have demonstrated various anti-microbial, anti-inflammatory, anti-oxidant, bioadhesive and anti-hypertensive activities (Grienke et al. 2014).

The content of bioactive metabolites in mussel meat is influenced by the season, life cycle, diet and habitat in which the mussels are grown and can therefore vary between harvests (Fearman et al. 2009; Narvaez et al. 2008). Furthermore, there are evident metabolic differences between mussel species and also within the same species when collected from different locations. For example, metabolomic assessment of the Australian Blue mussels (Mytilus galloprovincialis) and P. canaliculus found taurine, glycine, lactate, succinate, homarine, ATP, ADP, valine and leucine were elevated in P. canaliculus while betaine, isoleucine, acetoacetate and glucose were elevated in M. galloprovincialis. Also, analysis of lipid methyl ester derivatives indicted a clear separation between the species, with significantly higher levels of palmitic acid methyl ester (C16:0), cis-5,8,11,14,17 eicosapentaenoic acid methyl ester (C20:5n3) and palmitoleic acid methyl esters (C16:1) obtained from P. canaliculus, which overall contained a higher lipid level. These differences are likely due in part to the different environments that each species are grown in lower water temperatures correlating with higher degrees of unsaturated lipids (Rochfort et al. 2013). Experimental studies of bioactive carbohydrate compounds from P. canaliculus, is limited, with one report of a glycogen isolate demonstrating anti-inflammatory activity (after i.v. administration) against carrageenan-induced arthritis in the footpad of rats (Miller et al. 1993). The authors, however, confirmed the anti-inflammatory activity was lost when the glycogen isolate was treated with either potassium hydroxide (KOH) or proteinase-K, proposing that the anti-inflammatory activity was actually due to the protein moieties associated with the glycogen macromolecule.

3.1 Bioactive Proteins, Peptides and Amino Acids

Approximately 70 % of whole mussel meat is protein. The anti-inflammatory and immunomodulating activity of the fractionated extracts of whole extract powder in animal and in vitro models have suggested that the predominant active agent is associated with a protein moiety or is itself a protein macromolecule; however, supportive research for a bioactive high molecular weight protein is currently limited (Couch et al. 1982; Miller et al. 1993; Mani and Lawson 2006; Grienke et al. 2014). Current research has reported anti-bacterial , anti-fungal , anti-inflammatory , anti-hypertensive , anti-oxidant , anti-thrombin and anti-coagulant bioactive proteins, peptides and amino acids from various mussel species. The only bioactive protein identified from P. canaliculus is pernin from the cell-free haemolymph . It is an aggregating, non-pigmented, glycosylated protein extract composed of 497 amino acids with a particularly high content of histidine and aspartic acid residues. Pernin can act as a serine protease inhibitor but only demonstrates weak anti-thrombin activity. The pernin content from homogenise whole mussel meat averages 0.2 mg per mussel (Scotti et al. 2001).

Anti-microbial peptides (AMPs) are also present in the mussel haemolymph and are a vital part of the mussel’s innate immunodefense system, protecting it from bacterial, fungal and viral attack. AMPs have been a focus in marine mussel research, particularly in the Blue mussel (Mytilus edulis) and the Mediterranean (or Blue) mussel (Mytilis galloprovincialis) species. Several cysteine-rich peptides from M. edulis were reported to be potent bactericides (i.e. against both Gram-positive organisms, e.g. Enterococcus faecalis, Staphylococcus aureus and Gram-negative bacteria, e.g. Escherichia coli bacteria) and anti-fungal (i.e. Neurospora crassa and Fusarium culmorum) (Charlet et al. 1996). The AMPs were identified as isoforms from the peptide families of defensins, mytimycin and mytilins with big-defensins (and mytimacins) also being described (Charlet et al. 1996; Grienke et al. 2014). Crustacean haemolymph, particularly from the crab, contains a multitude of AMPs that participate not only as endogenous antibiotics but may also have a role in inflammation, wound repair and regulation of the adaptive immune system. This has generated some interest in using marine peptides for pharmaceutical developments (Fredrick and Ravichandran 2012). Furthermore, fermented M. edulis is reported to contain peptides that inhibit angiotensin I converting enzyme (ACE) with the anti-hypertensive activity confirmed in vivo rat models (Je et al. 2005). An anti-coagulant peptide has also been identified in M. edulis (Jung and Kim 2009). Anti-inflammatory activity of proteinaceous fractions in carrageenan-induced footpad swelling in rats was expressed only when following i.p. or i.v. injections (see Table 3). One study (Miller and Ormrod 1980) compared i.p. to orally administered whole extract powder which may or may not have been a stabilised extract. P. canaliculus extract powder may also inhibit the production of prostaglandins in rats (Miller and Wu 1984). The therapeutic activity of mussel protein and peptide fractions has not yet been investigated in humans.

Table 3 Pharmacological activity of protein fractions from P. canaliculus in animal and in vitro models
Full size table

3.2 Bioactive Lipid Fractions

The concentrated lipid extract of P. canaliculus contains a complex profile of five main lipid classes that include sterol esters (cholesterol and desmosterol/brassicasterol), triglycerides, free fatty acids (FFAs), sterols and phospholipids (Ulbricht et al. 2009; Whitehouse et al. 1997; Murphy et al. 2003). The fatty acid and sterol composition of the mussel lipid is influenced by water temperatures in which the mussel is grown and also the mussels’ diet which includes marine phytoplankton , dinoflagellates and zooplankton (Rochfort et al. 2013; Murphy et al. 2003). Approximately 90 fatty acids are present in the concentrated lipid extract with the omega-3 (Ω-3) FFAs making up approximately 40 % of the fatty acid content in the lipid extract with docosahaexanoic acid (DHA) and eicosapentaenoic acid (EPA) accounting for 84 % of the Ω-3 PUFA (Murphy et al. 2003; Wolyniak et al. 2005; Lee et al. 2009). The FFA fractions and to a lesser extent the triglyceride fractions, are reported to be the only lipids inhibiting cyclooxygenase (COX) isoforms (McPhee et al. 2007; Wakimoto et al. 2011; Whitehouse et al. 1997; Macrides 1997). Furan fatty acids (F-acids) have been identified as minor components of the fatty acids in the lipid extract and have shown anti-inflammatory activity in rat models of adjuvant-induced arthritis, more than that of EPA (Wakimoto et al. 2011). Further, another single phospholipid compound was isolated from freeze-dried whole powder and identified as lysolecithin which demonstrated anti-histamine activity in an ex vivo experiment (Kosuge et al. 1986). The investigation of single lipid components is extremely difficult due to the instability of the extract during the purification process; analytical studies have tendered to focus in characterising lipid extracts rather than identifying single lipid compounds (Grienke et al. 2014).

3.2.1 Anti-inflammatory Activity

The anti-inflammatory activity of the CO2-SFE lipid-rich fraction isolated from stabilised freeze-dried mussel meat has been demonstrated using in vitro analysis, particularly decreasing leukotriene (LTB4 and 5-HETE) and COX metabolite (COX-1 and COX-2 ) synthesis using variable doses of the SFE lipid-rich extract or further fractionations from this preparation dosed orally (Whitehouse et al. 1997; Dugas 2000; McPhee et al. 2007; Treschow et al. 2007; Macrides 1997). Adjuvant-induced arthritis rat studies have further analysed the anti-inflammatory and pain reducing effects of whole CO2-SFE lipid-rich fractions and some fractions administered orally, demonstrating the SFE lipid-rich fractions reduce paw swelling and pain when compared to control groups being equal or superior to oral NSAIDs such as naproxen (see Table 4) (Whitehouse et al. 1997; Lee et al. 2009; Whitehouse 2004; Wakimoto et al. 2011; Butters and Whitehouse 2003). Further, it is reported that a fractionated FFA class exhibited greater anti-inflammatory activity at a lower dosage (30 mg/kg) and for a shorter duration (5 days) when compared to the significant anti-inflammatory activity demonstrated for the crude lipid component, a SFE lipid-rich fraction (100 mg/kg) administered for 15 days to adjuvant-induced arthritis rats via subcutaneous injection (Singh et al. 2008).

Table 4 Therapeutic activity of SFE lipid-rich fractions from P. canaliculus in animal and in vitro analysis
Full size table

A further hypothesis for the anti-inflammatory activity demonstrated in in vitro and animal models, beyond that of reducing inflammatory cytokines (LTB4, 5-HETE, COX-1), is that the lipid-rich fraction may beneficially influence change in protein expression related to arthritis (Lee et al. 2008). Lee et al. (2008) conducted a proteomic study examining the effect that the lipid-rich fraction had on protein expression in splenocytes from adjuvant-induced arthritis rats . They found that in rats administered the lipid-rich fraction, six particular proteins (related to metabolism) were decreased while malate dehydrogenase (MDH), which is specifically related to glucose, metabolism was increased, possibly equating to a decline in glucose levels available for the activation of major histocompatibility complex class I (MHC-I). MHC-1 gene expression contributes to autoimmune diseases . Elevated levels of MDH can possibly decrease free glucose in the cytoplasm by converting pyruvate into malate (Lee et al. 2008).

Rat studies reporting anti-inflammatory activity with the SFE lipid-rich fraction have typically orally administered 20 mg/kg of body weight per day (highest 100 mg/kg/day) to rats with adjuvant-induced arthritis. The current dosage recommendation for humans, however, is 100–200 mg per day of the SFE lipid-rich fraction (administered with olive oil as a carrier). When comparing the dosages , using a 70 kg person for example, the doses used in rat studies are typically 7–14 times higher than that which is recommended for relief from arthritis-induced inflammation and pain in humans. How the low dose for human consumption was recommended, when rat studies demonstrated anti-inflammatory and pain relief at high doses, is not clear. It may be due to variations of metabolic capacity between rats and humans.

3.2.2 Anti-asthmatic Activity

The inhibition of leukotriene and prostaglandin E series production by the lipid-rich fractions has led to its assessment in both animal and human asthma models. Significant reductions in the development of allergic inflammation and airway hyperresponsiveness (rat model) (Wood et al. 2010) and asthma symptoms were attained in humans (Mickleborough et al. 2013; Emelyanov et al. 2002). Mickleborough et al. (2013) conducted a placebo controlled, double-blind randomised crossover study in patients with mild to moderate asthma (n = 20) who were given either eight capsules per day of a stabilised SFE lipid-rich extract (providing 72 mg EPA and 48 mg DHA) or placebo (olive oil) for 3 weeks duration, then followed by a 2 week ‘washout’ period before treatments were crossed over. The study showed that the lipid-rich fraction significantly reduced airway inflammation and the bronchoconstrictor response to dry air hyperpnea . The lipid-extract group also benefited by reduced asthma symptom scores and their lesser use of rescue medication compared to the placebo group (Mickleborough et al. 2013). Emelyanov et al. (2002) also demonstrated in a double-blind, parallel group, randomised, placebo-controlled study (n = 46) that the stabilised SFE lipid-rich fraction when supplemented at four capsules/day for 8 weeks duration, significantly decreased day time wheeze, concentration of exhaled H2O2 and an increase in morning peak expiratory flow in patients with atopic asthma compared to the placebo group (olive oil) (see Table 2) (Emelyanov et al. 2002).

3.3 Gastrointestinal Protection

The therapeutic efficacy of the CO2-SFE lipid-rich fraction has also been assessed in gastrointestinal disorders, with significant efficacy in a dextran sodium sulphate-induced inflammatory bowel disease (IBD) model in rats (Tenikoff et al. 2005). The lipid fraction significantly limited body weight loss, reduced disease activity indices and overall morphology of the inflamed intestinal tissue reducing crypt area loss preventing cecum and colon weight loss, all providing potential evidence for successful management of IBD. The lipid fraction was also assessed as a potential treatment in chemotherapy-induced intestinal mucositis in a rat model, using 5-fluorouracil as the toxin; however, the lipid fraction demonstrated only limited efficacy in reducing the symptoms (Torres et al. 2008) (see Table 4).

4 The Role of Intestinal Microbiota in the Therapeutic Activity of Perna canaliculus for Inflammatory Conditions

Humans and commensal bacteria coexist in a usually symbiotic relationship with a host to microbe cell ratio of 10:90 %, respectively. It is estimated that the GIT microorganisms collectively make up to 100 trillion cells, tenfold the number of human cells (Lederberg 2000). The collective microbial community is termed the microbiota or the microbiome and it populates specific human environments (e.g. the skin, mouth, nasal cavity, GIT and the urogenital tract). The GIT, skin, urogenital and respiratory systems are extensively colonised by symbiotic microorganisms (Singh et al. 2013). In the human GIT, there is a gradual increase (proximally to distally) in the density and diversity of the microbiota, with the large bowel microbiota representing the most dense, diverse and complex microbial ecosystem known (Tremaroli and Backhed 2012). The genomic content of the GIT microbiome is reported to encode 3.3 million unique bacterial genes, out-numbering the human genome by a factor of approximately 150 (Qin et al. 2010). The human genome, together with its associated microbiome, shares a mutually symbiotic relationship. The microbiota that colonise the GIT regulate normal development and function of the mucosal barriers; assist with maturation of immunological tissues, such as gut-associated lymphoid tissues, promoting immunological tolerance to antigens (foods, environment, pathogens); induce chemical communication to target tissues such as the liver, brain, muscle, adipose tissue, heart and GIT; prevent propagation of pathogenic microorganisms as well as control nutrient uptake and metabolism (Shen et al. 2013).

The GIT microbiota contributes to the metabolism of ingested compounds during the digestive process, including both foods and pharmaceutical drugs, to produce numerous metabolic products. Such metabolites function as signalling molecules between the bacteria and host cells. Metabolites that regulate host–microbiota dialogue include short-chain fatty acids (SCFA) , bile acids (e.g. choline) and lipids (i.e. LPS and peptidoglycan). The genetic richness of the GIT microbiota allows the expression of specific metabolic activities that are not encoded by human DNA (Gill et al. 2006; Egert et al. 2006; Laparra and Sanz 2010), including the hydrolysis and fermentation of dietary polysaccharides (Tremaroli and Backhed 2012). Therefore, the commensal GIT microbiota plays a critical role in human GIT metabolism. The metabolism of P. canaliculus by commensal microbial species has not been well explored. However, in vitro analyses have indicated that certain commensal bacteria ferment and metabolise the popular anti-arthritic medication D- glucosamine (Foley et al. 2008; Koser et al. 1961; Wolfe and Nakada 1956; Lutwak-Mann 1941; Faulkner and Quastel 1956). The metabolic capacity of intestinal microbiota can modify bioactive food components altering the hosts’ exposure to these components and potentially enhancing or diminishing their health effects. Furthermore, a number of microbiota-based interventions have shown to contribute to human health through maintaining normal microbial composition, improving metabolism and immunity of the gut and by enhancing mucosal integrity and barrier function (Turnbaugh et al. 2006; Gigante et al. 2011). Functional food components such as inulin, are known to influence the growth and metabolic activity of the GIT microbiota and thus its composition and subsequent metabolic capacity (Laparra and Sanz 2010; Campbell et al. 1997; Gibson et al. 2005). The intestinal microbiota is a target of nutritional interventions such as P. canaliculus, influencing bacterial viability, growth and metabolic activity (Coulson et al. 2013). Bacterial microbiota may consequently influence biological activity of nutritional supplements. It is proposed, therefore, that the therapeutic activity of P. canaliculus is potentially, or in part, due to its interaction with gut bacteria and consequential influence on the host immune system.

The implication of the GIT microbiota in rheumatic diseases has been recognised in in vivo studies. The discovery of a wide variety of bacterial species and bacteria-derived peptidoglycan-polysaccharides (PG-PS) present in synovial fluid from not only reactive arthritis, but also chronic forms of arthritis including RA and OA, indicates that arthritic joints are not sterile as previously thought (Kempsell et al. 2000b). The presence of bacterial antigens within the synovial fluid may play a role in the pathogenesis of several forms of arthritis, other than septic arthritis, such as triggering or exacerbating joint inflammation (Gerard et al. 2001; Kempsell et al. 2000b; van der Heijden et al. 2000; Siala et al. 2009a; Olmez et al. 2001; Carter et al. 2009). Analysis of synovial fluid from OA patients using various polymerase chain reaction (PCR) methods including reverse-transcriptase PCR, broad-range PCR and 16S rRNA PCR, has detected DNA from various bacterial strains. Bacterial DNA that has been detected in OA patients includes Pseudomonas sp., Shigella sp., Escherichia coli, Chlamydia trachomatis and Chlamydia pneumonia (Olmez et al. 2001; Carter et al. 2009; Gerard et al. 2001). From OA synovial fluid, immunoglobulin G (IgG) antibodies have been detected against Porphyromonas gingivalis, Prevotella intermedia and Bacteroides forsythus using enzyme-linked immunosorbent assays (ELISA), while IgA antibodies against B. forsythus have also been detected (Moen 2003). Additionally, there is evidence suggesting that host MHC genes may affect the microbiological milieu of the gut (Vaahtovuo et al. 2003; De Palma et al. 2010). High levels of antibodies directed against antigens of certain gut bacteria in RA patients propose a pathogenic relationship between these bacteria and RA (Scher and Abramson 2011). The exposure to bacterial cell walls may increase the susceptibility to develop arthritis as shown in animal studies (van den Broek et al. 1988; Jonsson et al. 2003). Evidence that bacterial DNA can be detected in OA joints, albeit not as frequently as in RA joints, highlights the importance of patient genetic variability and tolerance. Furthermore, in the case of Chlamydia trachomatis , there are several different serotypes that may predict various pathogenic outcomes (Carter et al. 2009). The source of bacteria detected in synovial fluids is not known, but it is suggested they may be derived from environmental sources or from the enteric microbiota (Siala et al. 2009). Periodontal pathogens may also be implicated in arthritic joint inflammation with antibodies to Gram-negative , anaerobic periodontal pathogens such as Porphyromonas gingivalis, Prevotella intermedia, Prevotella melaninogenica and Tannerella forsythia detected in the serum and synovial fluid of RA patients (Ogrendik 2012). These investigations support the hypothesis that in genetically susceptible subjects, exposure to degraded products of the gut bacteria locally in synovial fluids may cause inflammation. Consequently, it is proposed that bacteria may cause or influence joint disease in a number of ways such as creating persistent infection, inducing autoimmune pathophysiology, producing bacterial antigens or the induction of immune dysfunction (Kempsell et al. 2000a). However, the exact role and the full clinical implications of finding bacterial DNA in arthritic joints are still unknown.

RA patients have altered intestinal microbial profiles that may be relevant to the aetiopathogenesis of RA (Gul’neva and Noskov 2011; Toivanen 2003; Vaahtovuo et al. 2008). RA patients demonstrate decreased Bifidobacterium , Lactobacillus and Bacteroides-Porphyromonas-Prevotella species, elevated opportunistic Enterobacteria and Staphylococci species and variable reports of high or low Clostridium profiles. Evaluation of gut microbial profile compositions are limited in OA patients, however elevated Clostridium and Staphylococcus profiles have been found (Coulson et al. 2013). Pain relief medications used by rheumatic patients (i.e. acetaminophen and NSAIDs) may also contribute to altered bacterial profiles in conjunction with their known gastrotoxic effects (Upreti et al. 2010; Cuzzolin et al. 1994; Al-Janabi 2010). In genetically predisposed individuals, environmental factors such as diet, infections and smoking can cause dysbiosis in the GIT microbiota recognised as a microbial imbalance in which one or more bacterial phyla, genus or species overgrow and negatively impact on other beneficial bacteria (Taneja 2014). This dysbiosis may be related to the production of some metabolites as well as the activation of Nuclear Factor-kappa B (NF-kB) pathway that mediates the release of anti-inflammatory cytokines , compromising the integrity of the colonic epithelial cells, increasing gut permeability and consequently affecting health (Chen and Kasper 2014). Therefore, modification of inflammatory conditions such as OA, RA, asthma and IBD may be achieved in part through the refinement of GIT bacterial profiles to reflect a more homeostatic status (Coulson et al. 2012, 2013).

5 Discussion/Conclusion

The therapeutic efficacy of P. canaliculus for the treatment of OA, RA and asthma has been a contentious issue with a lack of conclusive evidence-based research. Further scientific investigations are required to evaluate product stability, optimal dosage, novel bioactive compounds and GIT microbiota profiles when assessing the efficacy of P. canaliculus. The predominant compound of P. canaliculus is protein, which has shown anti-inflammatory and immunomodulating activity in in vitro studies; however, its efficacy has not been investigated in humans yet. Understanding the interaction of the bioactive compounds in P. canaliculus with commensal and pathogenic bacterial may facilitate the development of novel interventions to control intestinal and extraintestinal inflammation.

While only two bacterial divisions ( Bacteroidetes and Firmicutes ) have been reported to dominate the gut microbiome, thousands of bacterial genera and species inhabit the human gastrointestinal tract. Hence the administration of compounds such as P. canaliculus and glucosamine to ameliorate the symptoms of OA and perhaps also RA may involve the actions of the gut microbial cohort to down regulate gut mucosal inflammatory sequalae. Recent clinical data (Coulson et al. 2012, 2013) plausibly suggests that these nutraceuticals may act as prebiotics in the gut, attenuating musculoskeletal inflammatory pain via interactions with the gastrointestinal microbiome.

Advanced sequencing tools/methodologies and experimental approaches have brought novel insights into the mechanisms that promote and maintain gut inflammatory processes that also include auto-inflammatory processes such as in RA. Indeed, it is now possible to locate the site and identity of thousands of bacteria (as well as their functions). This understanding has provided a previously unmatched level of bacterial communities and species detail. For example, animal models studying RA have shown the capacity of specific commensal bacteria to activate pro-inflammatory signalling, which in turn initiate and progress deleterious effects in the joints. The clinical implications of these findings, in parallel with reports that demonstrate humans harbour distinct enterotypes, strongly suggest that musculoskeletal diseases such as RA and the perpetuation of OA may originate in the gut. Certainly this can be plausibly pre-empted for RA from well-characterised studies utilising DNA-parallel sequencing in animal models elucidating possible dysbiosic states (Scher 2010).

If a distinct microbiota profile or pathogen promoted enterotype can be identified, it would then be possible to speculate whether a particular microbiome triggers or drives autoimmunity in genetically predisposed individuals or progresses pro-inflammatory sequale from the gut to the systemic circulation and the musculoskeletal joints. The identification of gut pathogenic commensal profiles could provide insights into the environmental triggers of musculoskeletal diseases and lead to a new understanding of disease pathogenesis, perhaps leading to novel approaches for adjunctive thereby.