The potential of selected Australian medicinal plants with anti-Proteus activity for the treatment and prevention of rheumatoid arthritis

INTRODUCTION

Autoimmune inflammatory disorders (AIID's) are a group of debilitating conditions including rheumatoid arthritis (RA), ankylosing spondylitis, lupus and multiple sclerosis, which afflict genetically susceptible individuals. There is no common susceptibility profile for these disorders. RA, for example, is most prevalent in middle-aged to older women, whereas the onset of ankylosing spondylitis occurs most frequently in younger males.[1] There are currently no cures for any of these conditions and current treatment strategies aim to alleviate the symptoms (particularly pain and swelling) via the use of analgesics and anti-inflammatory agents, and/or to modify the disease process through the use of disease-modifying anti-rheumatic drugs (DMARDs). None of these treatments are ideal as prolonged usage of these drugs can result in unwanted side effects and toxicity.[2] There is a need to develop safer, more effective drugs for the treatment of inflammatory diseases which will not only alleviate the symptoms, but which may also cure or prevent the disease.

Eradication of the cause of an inflammatory disease is an attractive target for drug design as this would not only block/decrease the late-phase inflammatory symptoms, but would also block the immune response and subsequent tissue damage associated with AIID's. While the causes of RA are not comprehensively understood, it is generally accepted that it is an autoimmune disorder which is triggered by specific microbial infections in genetically susceptible individuals (individuals with the MHC class 2 allele HLA-DR4).[3] Proteusmirabilis infections have been proposed to trigger RA based on several lines of evidence:

• Elevated serum levels of P. mirabilis specific cross-reactive antibodies have frequently been reported in individuals suffering from RA[456789]

• P. mirabilis antibodies from RA patients have cytopathic effects on joint tissue possessing P. mirabilis cross-reactive antibodies[10]

• P. mirabilis infections have been frequently reported in urine samples from patients with RA[11]

• Sera from rabbits immunized with HLA-DR4 positive lymphocytes bind specifically to Proteus[12]

• Amino acid sequence homologies have been identified between the “EQ/KRRAA” motif present in RA HLA-susceptibility antigens and the “ESRRAL” amino acid sequence present in P. mirabilis hemolysins[13]

• A further sequence homology between the “LRREI” sequence of type XI collagen (present in joint cartilage) and the “IRRET” motif present in P. mirabilis urease enzyme has also been reported.[14]

Based on the evidence linking Proteus bacterial infections with pathogenesis, a mechanism of RA disease progression has been proposed [Figure 1]: Gastrointestinal P. mirabilis acts as a trigger for RA.[10] Thus, limiting the levels of gastrointestinal P. mirabilis (1) would prevent RA initiation and minimize its downstream effects. Gastrointestinal P. mirabilis will not initiate the autoimmune events associated with RA unless it is able to interact with the immune system. This most often results from urinary tract infections (2) or when epithelial lesions (or other epithelial interruptions) allow for the production of anti-P. mirabilis antibodies.[1015] Inhibition of the causative agents of gut lesion forming conditions (e.g. Crohn's disease) would also be expected to decrease RA initiation events. Furthermore, prevention and early detection of urinary tract infections (UTIs) (the major pathway for interaction of P. mirabilis with the immune system) and/or the colonization of the bladder (2 and 3) would block the onset of RA and lessen downstream effects. Blocking the immune response by blocking the interaction of P. mirabilis with immunological cells (4) or by immunomodulation (5) also prevents the production of self-reactive antibodies (6) or the cross-reactivity with self-tissue, (9) thereby diminishing the later phase events of RA and thus the disease symptoms (10). However, immunomodulatory therapy should be used with caution as inhibiting the patient's immune capability would also expose the patient to a variety of other infections. Most current RA therapies target the later phase events by (7) blocking the inflammatory cascades or (8 and 10) by decreasing the symptoms of RA (e.g. pain, swelling, heat). While drugs targeting the late events are effective in easing patient discomfort, they still allow the tissue damage (which is associated with the self-reactive antibody action) to occur. Targeting earlier events prior to the induction of the immune response would not only alleviate the symptoms and discomfort of RA, but would also lessen/prevent the joint damage associated with chronic inflammation.

Figure 1

A proposed schematic representation of the main events in rheumatoid arthritis (RA) disease etiology and progression. Only major events are shown. Numbers refer to current and/or proposed targets for the prevention and treatment of RA

Targeting P. mirabilis infections may provide a new therapeutic approach for preventing and treating RA. One strategy is the development of anti-Proteus vaccines. While the production of a vaccine may block Proteus pathogenesis, it is also a problematic approach as anti-Proteus antibodies would also be likely to cross-react with the host connective tissue in susceptible individuals and thus exacerbate the symptoms of RA. If antibodies lacking cross-reactive epitopes are developed in the future, this approach may be effective as it would block susceptible individuals from acquiring a Proteus infection, thus effectively blocking RA progression. However, the development and usage of Proteus sensitive antibiotics may prove a more effective means of treating the RA bacterial trigger and thus blocking this disease. By destroying the Proteus bacteria, this treatment modality would be expected to greatly reduce the impact of the bacteria and the production of anti-Proteus antibodies and thus decrease disease progression.

Many antibiotics are already known to inhibit Proteus growth and/or have bactericidal effects toward Proteus spp. However, the development of super-resistant bacterial strains has resulted in currently used antibiotic agents failing to end many bacterial infections.[1617] For this reason, the development of new anti-P. mirabilis chemotherapeutic agents for the prevention and treatment of RA has received recent attention. Recent studies have examined the anti-P. mirabilis activity of conventional antimicrobials such as carbapenems[18] and of complementary and alternative therapies including nano-metallic preparations[19] and traditional South African medicinal plants.[20] A re-examination of traditional medicines for the treatment of inflammation and rheumatic conditions is an attractive prospect as the antiseptic qualities of medicinal plants have been long recognized and recorded. Furthermore, there has recently been a revival of interest in herbal medications due to a perception that there is a lower incidence of adverse reactions to plant preparations compared to synthetic pharmaceuticals.

In this study, a selection of Australian plants was identified which were traditionally used for the treatment of RA as well as other inflammatory conditions [Table 1]. While the ethnobotanical uses of these plants in traditional Aboriginal medicine systems have been recorded, rigorous scientific studies are lacking for many species. We were unable to find studies examining any anti-inflammatory properties of Aleurites moluccanus, A. excelsa, A. caerulea, and Duboisia leichhardtii despite their usage in traditional healing systems.[2122] As high levels of antioxidants have therapeutic effects against many diseases and medical conditions including inflammation and bacterial diseases, several high antioxidant plants were also included in this study. Tasmannia spp. and Syzygium spp. in particular, have been shown to have high antioxidant contents,[23] as well as documented medicinal uses.[2122] Other species with high antioxidant contents (Davidsonia pruriens, Elaeocarpus angustifolius) were also included in the study despite a lack of documented ethnobotanical usage or reports of anti-inflammatory activity.

Table 1

The ethnobotanical usage and common names of the plant species tested in this study

Several of the plant species identified as having anti-inflammatory properties have been studied previously for their antiseptic properties, although we were unable to find studies aimed at testing these plant species against RA. For example, recent studies have shown Tasmannia lanceolata to have potent inhibitory activity towards a wide range of bacteria.[24] Similarly, Australian Acacia spp.,[25] Callistemon spp.,[2627] Eucalyptus spp.,[2728] Leptospermum spp.,[29] Melaleuca spp.,[27] Syzygium spp.,[2127303132] P. augustifolia,[33] and Petalostigma spp.[34] have well-documented antibacterial activity against a wide variety of bacteria. Scaevola spinescens extracts have not only been shown to have broad-spectrum antimicrobial activity,[35] but also have documented uses in the treatment of UTI's and urinary disorders.[2122]

Several other plant species examined in this study are closely related taxonomically with species with known anti-inflammatory and antimicrobial activities. While we were unable to find reports of the ethnobotanical usage of Kunzea flavescens, extracts prepared from related Kunzea species are used in the treatment of rheumatism, swellings and inflammation.[2122] Other species examined in this study (e.g. T. insipidia and T. stipitata) have not been extensively studied for their anti-inflammatory or antimicrobial properties and are included in this study due to their taxonomic relationship with T. lanceolata. The selected plant species were extracted, and the extracts were tested against P. mirabilis to screen their potential to block the microbial triggers of RA.

MATERIALS AND METHODS

RESULTS

Liquid extraction yields

Extraction of 1 g of the dried plant materials with methanol and water yielded dried plant extracts ranging from 19.9 mg (P. pubescens leaf water extract) to 477 mg (T. lanceolata peppercorn methanolic extract) [Table 2]. In general, methanol was more efficient at extracting material from the plant samples than water. Methanol extracted a greater amount of material for 45 of the 53 plant samples tested (84.9%), whereas water extracted the greater amount of material for 6 of the 53 plant samples (11.3%). Two samples extracted approximately equal masses in both water and methanol. A correlation was also noted between the plant part tested and the amount of extracted material. For 12 plant species (A. caerulea, C. citrinus, C. salignus, D. pruriens, Leptospermum bracteata, L. juniperium, M. integriflora, P. pubescens, P. trilocularae, S. australae, S. leuhmannii and T. lanceolata), multiple plant materials were extracted. For those species where both leaf and flower were tested (C. citrinus, C. salignus, L. bracteata, L. juniperium), extraction of the leaf with either water or methanol generally resulted in higher yields of extracted material than from the flower extracts. Conversely, when both leaf and fruit extracts from the same plant were tested (A. caerulea, D. pruriens, M. integriflora, P. pubescens, P. triloculorae, S. australe, S. leuhmannii and T. lanceolata), extraction of the fruit with either solvent generally resulted in higher extraction yields than the corresponding leaf extracts. S. australe extracts were the exception to this trend, with both methanol and water extracting greater masses from the leaves than from the fruit. The dried extracts were resuspended in 10 ml of deionized water resulting in the extract concentrations shown in Table 2.

Table 2

Extract yields, MIC (μg/ml) of plant extracts which displayed inhibitory activity in the screening experiment against P. mirabilis and 24 h LC₅₀ values (μg/ml) of the plant extracts in the Artemia nauplii bioassay

Antibacterial activity

As decoctions and tinctures are the main forms in which plants were traditionally used in ethnobotanical medicinal systems, the zones of inhibition of the methanolic and aqueous extracts were tested undiluted to provide an approximate measure of the efficacy of the form in which the traditional medications would be used. Aliquots (10 μl) of each extract were screened against P. mirabilis [Figure 2]. P. mirabilis was inhibited by 65 of the 106 (61.3%) extracts tested. Similar numbers of methanolic (33 of 53 methanolic extracts; 62.3%) and aqueous extracts (32 of 53 aqueous extracts; 60.4%) inhibited P. mirabilis growth.

Figure 2

Antibacterial activity of plant water and methanolic extracts measured as zones of inhibition (mm) against Proteus mirabilis. Inhibition zones are represented as the means of at least triplicate experiments ± standard error of the mean. M and W refer to methanolic and water extracts respectively. (a) 1 = A. auriculiformis leaf; 2 = A. diasparima leaf; 3 = A. leptoloba leaf; 4 = A. moluccanus nut; 5 = A. excelsa leaf; 6 = A. caerulea fruit; 7 = A. caerulea leaf; 8 = B. citriodora leaf; 9 = B. myrtiflora leaf; 10 = Callistemon citrinus leaf; 11 = C. citrinus flower; 12 = C. formosus leaf; 13 = C. salignus leaf; 14 = C. salignus flower; 15 = C. oliveri leaf; 16 = Davidsonia pruriens leaf; 17 = D. pruriens fruit; 18 = D. leichhardtii leaf; 19 = Elaeocarpus angustifolius fruit; 20 = E. baileyana leaf; 21 = E. major leaf; 22 = Kunzea flavescens leaf; 23 = Leptospermum bracteata leaf; 24 = L. bracteata flower; 25 = L. juniperium leaf; 26 = L. juniperium flower; 27 = L. longifolium leaf; 28 = L. petersoni leaf; 29 = Macadamia integriflora nut; 30 = M. integriflora leaf; (b) 31 = M. alternifolia leaf; 32 = M. quinquenervia leaf; 33 = Petalostigma pubescens leaf; 34 = P. pubescens fruit; 35 = P. triloculorae leaf; 36 = P. triloculorae fruit; 37 = P. augustifolium leaf; 38 = Scaevola spinescens leaf; 39 = Syzygium anisatum leaf; 40 = S. australe leaf; 41 = S. australe fruit; 42 = S. forte leaf; 43 = S. francisii leaf; 44 = S. moorei leaf; 45 = S. puberculum leaf; 46 = S. wilsonii leaf; 47 = S. leuhmannii leaf; 48 = S. leuhmannii fruit; 49 = Tasmannia insipida leaf; 50 = T. lanceolata leaf; 51 = T. lanceolata fruit; 52 = T. lanceolata peppercorn; 53 = T. stipitata leaf; Amp = Ampicillin control (2 μg); Chl = Chloramphenicol control (10 μg)

The most effective inhibitors of P. mirabilis growth were the T. lanceolata and Datura leichardtii extracts based on the zones of inhibition (24.4 mm for the T. lanceolata leaf methanolic extract). Indeed, these extracts displayed substantially more potent P. mirabilis growth inhibition than either of the control antibiotics (ampicillin and chloramphenicol). The A. moluccanus nut, A. caerulea leaf, C. formosus leaf, C. oliveri, Eucalyptus major, L. longifolium, M. alternofolia, P. pubescens leaf, P. triloculare leaf, P. augustifolium, S. australe fruit, S. leuhmannii fruit, and T. stipitata leaf extracts also displayed strong P. mirabilis growth inhibition, with zones of inhibition >10 mm and thus were also considered promising anti-P. mirabilis extracts.

The relative level of antibacterial activity was further quantified by determining the MIC values for each extract [Table 2]. The A. moluccanus, D. leichardtii, E. major, L. bracteata, L. juniperium, M. integriflora nut, M. alternifolia, M. quinquenervia, P. pubescens, P. triloculorae, P. augustifolium, S. spinescens, S. australe, and T. lanceolata extracts were determined to be the most effective inhibitors of P. mirabilis growth, with MIC values generally significantly below 1000 μg/ml. T. lanceolata fruit extracts were the most potent growth inhibitors, with MICs of 11 and 126 μg/ml for the methanolic and aqueous extracts, respectively. Furthermore, noteworthy was the low MIC values seen for M. integriflora nut methanolic extract (15 μg/ml) and the S. spinescens leaf water extract (25.4 μg/ml).

Quantification of toxicity

The plant extracts were serially diluted in artificial seawater for toxicity testing in the Artemia nauplii lethality bioassay. For comparison, the reference toxin potassium dichromate was also tested in the bioassay. Figure 3 shows the percentage mortality induced in the Artemia nauplii following 24 of exposure. Potassium dichromate (reference toxin) was rapid in its induction of mortality, inducing the onset of mortality within the first 3 h of exposure (results not shown). By 24 h of exposure, potassium dichromate had induced 100% mortality in the Artemia nauplii. At 24 h, several extracts induced mortality significantly above that of the seawater control. The E. baileyana methanol extract, S. australe leaf and fruit methanol and water extracts, S. leuhmannii leaf and fruit methanol and water extracts and all extracts of Tasmannia spp. induced 100% mortality at 24 h. While appearing less toxic, C. oliveri methanol extract, both D. pruriens fruit extracts, both E. angustifolius extracts, E. major methanolic extract and both S. forte extracts all induced significant (>50%) mortality in the Artemia nauplii bioassay. The levels of mortality for all other extracts following 24 h and exposure was below 50% and thus these extracts were considered nontoxic.

Figure 3

The lethality of Australian plant extracts (2000 μg/ml) towards Artemia franciscana nauplii after 24 hours exposure. Results are expressed as mean ± SEM of at least triplicate determinations. M and W refer to methanolic and water extracts respectively. (a) 1 = Acacia auriculiformis leaf; 2 = A. diasparima leaf; 3 = A. leptoloba leaf; 4 = A. moluccanus nut; 5 = A. excelsa leaf; 6 = A. caerulea fruit; 7 = A. caerulea leaf; 8 = Backhousia citriodora leaf; 9 = B. myrtiflora leaf; 10 = C. citrinus leaf; 11 = Callistemon citrinus flower; 12 = C. formosus leaf; 13 = C. salignus leaf; 14 = C. salignus flower; 15 = C. oliveri leaf; 16 = Davidsonia pruriens leaf; 17 = D. pruriens fruit; 18 = D. leichhardtii leaf; 19 = Elaeocarpus angustifolius fruit; 20 = E. baileyana leaf; 21 = E. major leaf; 22 = Kunzea flavescens leaf; 23 = Leptospermum bracteata leaf; 24 = L. bracteata flower; 25 = L. juniperium leaf; 26 = L. juniperium flower; 27 = L. longifolium leaf; 28 = L. petersoni leaf; 29 = Macadamia integriflora nut; 30 = M. integriflora leaf; (b) 31 = M. alternifolia leaf; 32 = M. quinquenervia leaf; 33 = Petalostigma pubescens leaf; 34 = P. pubescens fruit; 35 = P. triloculorae leaf; 36 = P. triloculorae fruit; 37 = P. augustifolium leaf; 38 = Scaevola spinescens leaf; 39 = Syzygium anisatum leaf; 40 = S. australe leaf; 41 = S. australe fruit; 42 = S. forte leaf; 43 = S. francisii leaf; 44 = S. moorei leaf; 45 = S. puberculum leaf; 46 = S. wilsonii leaf; 47 = S. leuhmannii leaf; 48 = S. leuhmannii fruit; 49 = Tasmannia insipida leaf; 50 = T. lanceolata leaf; 51 = T. lanceolata fruit; 52 = T. lanceolata peppercorn; 53 = T. stipitata leaf. PC = Potassium dichromate positive control (1000 μg/ml); NC = seawater negative control

To benchmark the toxicity of the extracts that induced significant mortality for comparison with other toxins, the LC₅₀ values of the extracts was determined by testing across the concentration range 2000 μg/ml to 125 μg/ml in the Artemia nauplii bioassay. For comparison, potassium dichromate was tested across the same concentration range. Table 2 shows the LC₅₀ values of the extracts which had shown toxicity toward A. franciscana. While several extracts appeared toxic in the extract mortality screening study [Figure 3], determination of their LC₅₀ values demonstrates that most were nontoxic as LC₅₀ values of ≥1000 μg/ml have previously been defined as nontoxic in the Artemia nauplii bioassay.[41] While the C. oliveri methanol extract, both D. pruriens fruit extracts, both E. angustifolius extracts, S. australe leaf extracts and all Tasmannia spp. extracts appeared toxic in the screening study [Figure 3], all had LC₅₀ values of ≥1000 μg/ml and are therefore considered to be nontoxic. In contrast, all Eucalyptus spp. extracts, the S. australe fruit extracts and all S. forte and S. leuhmannii extracts had LC₅₀ values ≤1000 μg/ml and were therefore considered to be toxic. While these extracts displayed significant toxicity, potassium dichromate (the positive control) displayed substantially higher toxicity at 24 h (24 h LC₅₀ 92.3 μg/ml).

High performance liquid chromatography-MS/MS analysis of extracts displaying antimicrobial activity

The Proteus growth inhibition and MIC studies indicated that T. lanceolata fruit extracts had the most promise for further phytochemical characterization. These were examined for the presence of the trans–3, 4 5-trihydroxy-trans-stilbene (resveratrol) to examine whether resveratrol may be responsible for the anti-Proteus activity observed in these studies. Resveratrol is a potent anti-inflammatory compound produced as a phytoalexin (plant antibiotic) by many diverse plant species including grapes, berries, peanuts[42] as well as some species of pine trees, Acacia, Terminalias,[43] orchids, and lilies (Polunin et al., 2002).[44]

Optimized HPLC-MS/MS parameters were developed and used to profile the compounds from the methanolic [Figure 4] and aqueous [Figure 5] extractions of T. lanceolata fruit. Although the negative ion chromatograms yielded significantly higher signals than those observed for the positive ion (an order of magnitude higher), the negative ion chromatogram had significantly higher background levels, due to ionization of the reference ions in this mode. The T. lanceolata fruit methanol extract chromatograms in both positive and negative ion modes [Figure 4] revealed numerous overlapping peaks, particularly in the early and middle stages of the chromatogram corresponding to the elution of polar compounds. Nearly, all of the methanol extract compounds had eluted by 12 min (corresponding to approximately 27% acetonitrile). Indeed, numerous peaks eluted in the first 5 min with 5% acetonitrile. However, a prominent peak eluting later in the positive ion chromatogram [Figure 4a] at approximately 16 min (corresponding to approximately 44% acetonitrile) indicates the broad spread of polarities of the compounds in this extract. This later eluting peak was not apparent in the T. lanceolata fruit methanol extract negative ion chromatogram [Figure 4b].

Figure 4

Reversed-phase high performance liquid chromatography total peak chromatogram of 10 μl injections of the methanolic extracts of Tasmannia lanceolata fruit (a) in positive ionisation mode and (b) in negative ionisation mode. Extracts were dried and resuspended in deionised water. Arrows indicate the stilbene peaks detected in each chromatogram. Chromatography conditions were as described in the methods section

Figure 5

Reversed-phase high performance liquid chromatography total peak chromatogram of 10 μl injections of the aqueous extracts of Tasmannia lanceolata fruit (a) in positive ionisation mode and (b) in negative ionisation mode. Extracts were dried and resuspended in deionised water. Arrows indicate the stilbene peaks detected in each chromatogram. Chromatography conditions were as described in the methods section

The T. lanceolata fruit water extract chromatograms [Figure 5] also showed large amounts of polar material eluting early in the chromatogram at similar elution volumes as many of the compounds in the methanol extract. The aqueous extracts also had many of the same later eluting peaks seen in the methanolic extracts, including the peak at approximately 16 min. Interestingly, none of the extracts tested displayed peaks corresponding to the molecular mass signal/empirical formula of a pure resveratrol standard in either positive or negative ionisation mode. However, both the methanolic and aqueous T. lanceolata fruit extracts displayed small peaks, with molecular weights consistent with picied and combretastatins A–1 and A–4, indicating the presence of only low levels of stilbenes in these extracts. A comparison with the accurate mass databases putatively identified these compounds as Piceid (molecular weight 404.1465 Da), combretastatin A-1 (molecular weight 332.1319 Da) and combretastatin A-4 (molecular weight 316.1287 Da).

An examination of the fragmentation ions in positive mode provided further evidence of the putative identity of these stilbenes in the T. lanceolata fruit extracts [Table 3]. Fragmentation of the compound identified as Piceid (identified by comparison to molecular mass databases) yielded peaks corresponding to the molecular ion mass (404.1465 Da) as well as fragments corresponding to 286.1453, 270.0472, 228.1447, 124.1348, and 108.1212 Da. The empirical formulae corresponding to these fragments were determined using the Find by Molecular Feature algorithm in the Masshunter Qualitative analysis software package (Agilent Technologies). The molecular formulae C₁₃H₁₈O₇, C₆H₁₂O₆, C₁₄H₁₂O₃, C₇H₈O₂, and C₇H₈O were assigned for the fragments, respectively. Comparison with the molecular mass databases indicated the putative structures of these fragments [Figure 6]. It is likely that fragments b1 and b5 arise from a cleavage of the glucose moiety from the resveratrol moiety of Piceid. Thus, these fragments were putatively identified as resveratrol (228.1447 Da, C₁₄H₁₂O₃) and glucose (108.1212 Da, C₆H₁₂O₆) respectively. Similarly, fragments b2 (108.1212 Da, C₇H₈O) and b4 (124.1348 Da, C₇H₈O₂) were assigned to the products arising from the cleavage of resveratrol (b1, 228.1447 Da, C₁₄H₁₂O₃). These structural assignments were supported by the presence of fragment b3 (C₁₃H₁₈O₇), which would result from glucose (b5) combining with fragment b4.

Figure 6

Chemical structures of (a) resveratrol and the stilbenes and stilbene glycosides identified in Tasmannia lanceolata extracts: (b) piceid, (c) combretastatin A-1, (d) combretastatin A-4. (b1-b5) represent the molecular weight fragments of piceid detected by ESI-MS in positive mode; (c1-c2) represent the molecular weight fragments of combretastatin A-1 detected by ESI-MS in positive mode; (d1-d2) represent the molecular weight fragments of combretastatin A-4 detected by ESI-MS in positive mode

Table 3

Compound MWs (calculated and measured) and fragment MWs of prominent peaks detected in electron spray ionisation mass spectroscopy. Mass spectral analysis was run in positive ion mode

A similar strategy was used to confirm the structures of the combretastatins. Fragmentation of the compound identified as combretastatin A-1 yielded peaks corresponding to the molecular ion mass (332.1319 Da) as well as fragments corresponding to 182.1872 and 152.1445 Da. The molecular formulae C₁₀H₁₄O₃ and C₈H₈O₃ were assigned for these fragments, respectively. Comparison with the molecular mass databases indicated the putative structures of these fragments [Figure 6c1 and 6c2]. These fragments were assigned to the cleavage of combretastatin A-1 (332.1319 Da, C₁₈H₂₀O₆) to 182.1872 Da (C₁₀H₁₄O₃) and 152.1445 Da (C₈H₈O₃) fragments.

The compound putatively identified as combretastatin A-4 [Figure 6d; 316.1287 Da, C₁₈H₂₀O₅] also fragmented, giving 182.1788 Da (C₁₀H₁₄O₃), providing further evidence of the structural similarity of the compounds putatively identified as combretastatins A-1 and A-4. A further fragment (138.1463 Da) was detected in the fragmentation of combretastatin A-4. The molecular formula C₈H₁₀O₂ was assigned to this fragment by comparison with the molecular mass databases. From a comparison of the proposed fragmentation patterns with the putative molecular ion structure, it is evident that the combretastatin A-4 structure is consistent with these results.

DISCUSSION

Plant remedies are becoming increasingly sought after in the treatment of a myriad of diseases and disorders due to their perception of greater safety than synthetic drugs, as well as the failure of current drug regimes to effectively treat many diseases. This is especially true for RA. The current treatments utilizing DMARDs to alleviate the symptoms of RA and/or alter the disease progression are not entirely effective and have been associated with numerous adverse effects.[2] Furthermore, many of the current treatments are aimed at treating the symptoms of RA without addressing the underlying causes and disease mechanisms. An understanding of the initiation and progression of RA is important to allow for the development of new drugs to target specific processes and thus more effectively treat the disease.

The role of P. mirabilis infection as a trigger of RA is now widely accepted due to a wealth of supporting evidence.[15] A treatment regime targeting the microbial trigger of RA is an attractive prospect as it would be expected to block/decrease production of self-reactive antibodies and thus effectively block disease progression. The results presented here demonstrate the ability of a wide variety of plants traditionally used to treat RA and inflammatory complaints in Australian Aboriginal medicinal systems to inhibit P. mirabilis growth when tested in the forms that they would have been used in traditional medicine (decoctions and tinctures). Further quantification by MIC determination demonstrated the efficacy of these extracts as inhibitors of the microbial trigger of RA. The T. lanceolata extracts were of particular interest, with MIC's as low as 11 μg/ml (0.11 μg in the disc) for the methanolic fruit extracts. Indeed, these extracts proved to be substantially stronger than the control antibiotics (ampicillin and chloramphenicol) and are thus worthy of further mechanistic studies. The current study was limited to the inhibitory activity toward the microbial trigger of RA. It is possible that many of the plant extracts studied here may also have effects on other inflammatory processes (e.g. cytokine release) and, therefore, may have pluripotent anti-RA mechanisms.

The major anti-inflammatory components of many medicinal plants are known and are already available in a pure form as anti-inflammatory agents. These compounds include a variety of polyphenolic compounds, of which 3, 5 4’- trihydroxy-trans-stilbene (resveratrol) has received much recent attention.[45] Most studies examining the role of resveratrol in the treatment of RA concern its ability to act as a potent specific inhibitor of nuclear factor (NF)-κB activation through its induction by tumor necrosis factor alpha and interleukin-1 beta (IL-1 β).[45] Thus, resveratrol treatment is known to block cytokine production and inflammation via its inhibition of NF-κB activation. However, it is possible that resveratrol may also have other anti-rheumatic effects as it also has antibacterial activity against a variety of bacterial species.[46] It was therefore thought likely that it may also function to block the microbial trigger of RA and the plant extracts examined in this study were screened for its presence.

Interestingly, while resveratrol has been reported to have inhibitory activity against other bacterial species, it was unable to inhibit the growth of P. mirabilis in our studies, even at high concentrations (unreported results). Furthermore, we were unable to find any reports of in vitro anti-Proteus growth inhibition in the published literature. Several studies have reported on the ability of resveratrol to inhibit P. mirabilis swarming and virulence factor expression in vivo, so it is likely that resveratrol does affect P. mirabilis colonization and infection of the urinary tract,[47] albeit possibly by mechanisms other than bactericidal or growth inhibition mechanisms. Any extracts that were capable of inhibiting Proteus spp. growth and also contained the resveratrol would be likely to have anti-RA activity via several mechanisms (growth inhibition, colonization blocking, cytokine production inhibition) and, therefore, would be particularly effective in treating RA.

High accuracy HPLC-MS/MS analysis of the T. lanceolata fruit extracts failed to detect resveratrol in either extract. However, it is possible that resveratrol may still contribute to the anti-inflammatory activities of these extracts. The resveratrol glycoside Piceid (2-[3-hydroxy-5[(E)-2-(4-hydroxyphenyl) ethenyl] phenoxy]-6-(hydroxymethyl) oxane-3, 4 5-triol) was detected in both T. lanceolata fruit extracts. Piceid may be hydrolysed in vivo to remove glucose, thus releasing the resveratrol moiety. Furthermore, while other resveratrol glycosides like polydatin were not detected by matching with the compound databases used in these studies, it is possible that they may still be present in small quantities. Piceid and other glycosylated stilbenes (if present) may also contribute to the anti-inflammatory activity of T. lanceolata fruit extracts via mechanisms other than Proteus growth inhibition. Several glycosylated stilbenes (including Piceid) have both been shown to block inflammation by decreasing IL-17 production in stimulated human mononuclear cells.[48] Another study[49] determined that the antibacterial activity of resveratrol is due to a protein tyrosine kinase activity. This same study also reported that resveratrol glycosides do not have the same bacterial inhibitory activity, indicating that free hydroxyl groups on both phenyl groups are required for antibacterial activity.

In addition to Piceid, other stilbenes were also detected in the T. lanceolata fruit extracts. Combretastatins A -1 and A -4 were detected in both T. lanceolata fruit extracts, although the relatively low peak sizes indicate that both compounds are present in low abundance. Combretastatins are well –known for their potent ability to block cancer cell progression and induce apoptosis by binding intracellular tublin, thereby disrupting microtubule formation.[50] Accounts of direct anti –inflammatory activity of combretastatins are lacking. However, it is believed that they act by a similar mechanism to that of colchicine (N -[(7S) -1, 2, 3 10 -tetramethoxy -9 -oxo -5, 6, 7 9 -tetrahydrobenzo[a] heptlen –7 -yl] acetamide) by binding the colchicine binding site on the tubulin peptide and inhibiting polymerization.[51] It is likely that they may have a similar anti –inflammatory activity and mechanism to colchicine.

We were also unable to find accounts of the antibacterial activities of natural combretastatins in the literature. However, recent studies have examined the growth inhibitory activity of several synthetic combretastatin and resveratrol structural analogues.[52] These studies reported potent growth inhibition toward a panel of bacteria including Staphylococcus aureus, Streptococcus pneumoniae, Escherichia coli, and Neisseria gonorrhoeae. Especially interesting was the low MIC values of some analogues against N. gonorrhoeae, although several other species were also highly susceptible to the modified stilbenes. The same study also reported potent inhibitory effects of the panel of modified stilbenes against the fungal species Candida albicans and Cryptococcus neoformans.

A number of other inflammatory stilbenes have also previously been reported in other plant species. For example, 2, 3, 4 5-tetrahydroxystilbene-2-O-β-D-glucoside (TSG) inhibits inflammation by suppressing the induction of pro-inflammatory mediators by reducing NF-κB binding to DNA.[53] The same study detected TSG in numerous herbs used to treat inflammation in Chinese traditional medicine. Furthermore, nine stilbene and stilbene derivatives isolated from the roots of Cicer spp. (chickpeas) were shown to inhibit bacterial and fungal growth.[54] While these compounds were not detected in the compound databases used in our studies, it is possible that they may still be present in small quantities.

It is likely that other phytochemical classes may also contribute to the anti-inflammatory properties of these extracts. Alkaloids, anthraquinones, flavonoids, polyphenolics, phytosterols, saponins, tannins, and terpenes have also been linked with anti-bacterial activity in different plant species and thus may be responsible (at least in part) for the anti-anti-Proteus activities reported here. Several terpenoids previously reported in T. lanceolata fruit extracts have been reported to suppress NF-κB signaling (the major regulator of inflammatory diseases).[55] The monoterpenes limonene[5657] and α-pinene[58] have been reported to inhibit NF-κB signaling pathways. α-Pinene affects inflammation by inhibiting p65 translocation into the nucleus in Lipopolysacccharide-induced NF-κB signaling.[58] Furthermore, many other sesquiterpenes and sesquiterpene lactones also have well-established anti-inflammatory activities.[55] While much work is still needed to characterize the mechanisms of action of these compounds, it appears that NF-κB inhibitory activities may be responsible.

The antimicrobial activity of Drimys winteri (a species closely related to T. lanceolata) essential oils has been well-documented against a variety of bacterial species, and it has been established that terpenoids contribute to this activity.[59] Drimys winteri essential oils contain many of the same monoterpenoid constituents as T. lanceolata essential oils (including polygoidal, α-pinene, β-pinene, sabinene, mycrene, terpinene, limonene, and β-phellandrene). That study demonstrated good antibacterial activities for all of these compounds. Further studies have also shown that the monoterpene piperitone reduces the resistance of several strains of Enterobacteriaceae to the antibacterial agent nitrofurantoin.[60] Other studies have reported similar antibacterial activities for the sesquiterpenoids α-cubebene, copaene, and caryophyllene isolated from Pilgerodendron uviferum.[61]

This study was limited to an examination of the phytochemistry of the most potent P. mirabilis growth inhibitor. Further studies are needed to examine the compounds present in the other highly potent extracts. Specifically, the A. moluccanus, D. leichardtii, E. major, L. bracteata, L. juniperium, M. integriflora nut, M. alternifolia, M. quinquenervia, P. pubescens, P. triloculorae, P. augustifolium, S. spinescens and S. australe extracts should also be screened for the presence of stilbenes.

The findings reported here also demonstrate that the majority of plant extracts tested did not display significant toxicity in the Artemia nauplii bioassay. Indeed, with the exception of some Eucalyptus and Syzygium extracts, all other extracts exhibiting Proteus inhibitory activity were also shown to be nontoxic, or of low toxicity in the Artemia nauplii bioassay. Only extracts from these species displayed LC₅₀ values below 1000 μg/ml and therefore all other plant extracts that inhibit P. mirabilis growth are defined as nontoxic as compounds with an LC₅₀ of >1000 μg/ml toward Artemia nauplii have been previously defined as being nontoxic.[41] However, even the Eucalyptus spp. extracts (with LC50 values generally >650 μg/ml) are considered of only moderate toxicity.

CONCLUSIONS

The results of this study partially validate the usage of these plants in traditional Australian Aboriginal medicinal systems to treat RA and other inflammatory conditions and indicate that the phytochemistry and mechanism of action are worthy of further study. Similarly, it is also of interest to screen these extracts for further activities associated with the treatment of RA (e.g. inhibition of cytokine production) to further evaluate their potential as anti-RA drugs. While the extracts examined in this report are promising as anti-RA agents, caution is needed before these compounds can be applied to medicinal purposes. In particular, further toxicity studies using human cell lines are needed to verify the suitability of these extracts for these purposes.