Artificial oral fluid characterisation: Potential for use as a reference matrix in drug testing

Quality assurance schemes for drug-screening programmes require access to large quantities of biological matrices for reference or control samples. This presents problems when the availability of a matrix, such as oral fluid (OF) for screening or for confirmatory purposes, limits the collection of large volumes. In such cases, synthetic alternatives of OF may provide a solution. The preparation of an artificial (synthetic) oral fluid (AOF) was conducted by dissolving its components (salts, surfactant, antimicrobial agent and mucin) in water. We characterised the physical properties of AOF to determine its suitability as a matrix for quality assurance purposes. The evaluation of pH, specific gravity (SG), conductivity (mS cm − 1 ), freezing point depression ( (cid:1) C), light-scattering and kinematic viscosity (mm 2 s − 1 ) showed AOF to be a stable, reliable matrix. Synthetic OF was prepared using components (mucin, surfactants and so on) obtained from different suppliers and a comparison was performed. Our results sug-gest that AOF is a feasible matrix for the preparation of quality assurance samples for confirmatory or drug screening programmes.


| INTRODUCTION
Oral fluid (OF) has gained a considerable hold as a biological matrix for drug-testing purposes in recent years [1][2][3][4] and is a convenient alternative to blood because it does not require invasive collection procedures or complex professional skills for sample collection. There is a preponderance of evidence that commends the usefulness of OF for the detection of both illicit drugs 5 and prescribed medication, 6 the basic premise being that a drug present in blood may be detected when it passes through the oral mucosa into the oral cavity. There has also been considerable interest in the efficacy of devices for the collection, transportation, handling and storage of OF. 7 Particular interest has been shown in the development of POCT (point of collection drug-screening devices) for OF, recently reviewed by Wolff et al. 8,9 OF has been identified as the most accessible matrix for road-side screening of drugs in drivers apprehended and suspected of driving under the influence of psychoactive drugs because OF is usually easy to obtain and may be indicative of recent consumption 10 and, for example, has been used in Australia for confirmatory testing.
However, human OF is a complex mixture of fluids originating from different salivary glands and gingival crevicular fluid. Secretions from the parotid, submandibular, sublingual and minor mucous glands are the primary contributors to this mixture 11 along with glycoproteins known as mucins, which form mucus when dissolved. It has been reported that mucins contribute to the rheological properties of OF as a result of their unique chemical and structural characteristics. 12 Mucins not only demonstrate antibacterial activity but also act as a protective surfactant to cover biomaterials to suppress an immunological response. 13 As lubricants, mucins can also produce oral surface covering substitutes in the absence of saliva. 14 Mucus itself can cause problems in handling when transferring small volumes by pipette. Ease of handling may also be related to the elasticity and viscosity (spinnbarkeit) of OF. 15 Handling human OF in the laboratory can be complicated by the presence of cellular debris, nasal secretions, bacteria and residues of ingested fluids and foodstuffs, materials that need to be removed to prevent interference with assay procedures.
The preparation of quality assurance samples for OF drugscreening programmes requires laboratory assay and handling, storage and preservation studies under carefully controlled circumstances.
Organisations such as the U.S. Food and Drug Administration (FDA) require rigorous approval processes whilst routine assays require periodic testing for quality assurance. 16 All such testing requires large quantities of OF with consistent composition. Human OF is difficult to obtain in large volumes, and the composition of unstimulated whole human OF is not consistent showing differences between the sexes, 17 in the concentrations of sodium and chloride, in the circadian rhythm 18 and hence flow rate, 19 and in relation to the donor's state of health.
In the general population, collecting human OF for evaluation and/or quality control purposes may be difficult because of the variability in 'normal' salivary secretion rates, obviously volunteers with 'dry mouth' syndrome (xerostomia) or hyposalivation would be unsuitable. [20][21][22][23] While it is theoretically possible to collect large volumes of pooled human OF from a large number of volunteers in order to make a uniform reference sample and indeed has been used by NIST for the testing of devices in the US, 24 no reference material is currently available from any of the organisations (such as NIST or NIBSC) who prepare uniform international reference samples and materials. Exogenous factors such as other drug use may impact on salivation because both prescribed medication and illicit drug use have been reported to inhibit salivation, for example, cocaine 25 and morphine. 26 Indeed, many prescription and nonprescription drugs list dry mouth as a major side effect including benzodiazepines and narcotic analgesics, 27 glucocorticoids and antipsychotics among others. 28 A high prevalence of buccal dryness has also been reported with elderly populations (aged 65 to 95 years), because the aging process is associated with reduced salivary flow. 29 In addition, the variability of collection devices used in vivo and the need for evaluation of analytical processes predicates the need for a more dependable source of this matrix.
Preparation of artificial OF (AOF) is not new, and a patent was filed for 'synthetic oral fluid standards' with the United States Patent Office (Patent No 5,736,322) in 1998 by Goldstein for testing, calibration and standardisation of devices using OF; the OF standard contained a mucin and a protease inhibitor. In 2012, as part of the 'Guides to Type Approval' document for screening devices, the U.K. Home Office Centre for Applied Science and Technology (CAST) created an AOF as a matrix for drug driving tests. 30 Artificial saliva has been used during numerous studies in odontology 31 and has been manufactured as a saliva substitute to serve as mouth and throat lubricants in many instances. [32][33][34][35] The complex nature of mucin has prevented the chemical synthesis of truly traceable material, and hence, this is a potentially important consideration in the preparation of synthetic OF. Mucin is however available from wellknown sources and have traceable batch references, which can be used instead, but the material needs to be characterised as suggested by Pillai et al. 36 We used the Guides to Type Approval formulation (Home Office, 2012) 30 and sought to characterise AOF for use as a suitable reference material and investigate its usefulness as a quality control matrix for approval of devices to be used at the road-side to screen for drugdriving. Synthetic OF could also be used for method development of confirmatory tests. Physical measurements selected as relevant to the intended purpose of the AOF were: pH, kinematic viscosity (mm 2 s −1 ), specific gravity (SG), conductivity (mS cm −1 ), freezing point depression ( C) and light-scattering (average particle size [z average]), the distribution of particle sizes (polydispersity index-PDI) and the number of particles present (derived count rate-DCR). The intention was to select a range of physicochemical measurements to determine which, if any, might be affected by changes in the composition of the AOF including any batch to batch variability and storage conditions. This paper indicates the value of the measurements selected some of which, for example, pH and perhaps also SG and conductivity, can be readily reproduced (at low cost) in most analytical laboratories. We wished however also to evaluate some of the more sophisticated approaches such as light scattering to evaluate any lack of stability of the AOF and, especially, its tendency to flocculate.

| Preparation of AOF
Bovine mucin from submaxillary glands was obtained from two suppliers Sigma-Aldrich (Dorset, UK) and Merck Millipore (Watford, UK).
Potassium thiocyanate, calcium chloride dihydrate, magnesium chloride hexahydrate, ProClin™ 300, Tetronic® 90R4 and Tween® 20 were purchased from Sigma-Aldrich (Dorset, UK). Potassium chloride, potassium dihydrogen phosphate, sodium chloride, sodium azide, sodium hydrogen carbonate and urea were purchased from Fisher Scientific (Loughborough, UK). All reagents were of analytical reagent grade apart from Tween® 20, which was BioXtra grade. The AOF was synthesised according to the composition given in Table 1

| Choice of mucin
Two sources of bovine mucin from submaxillary glands were investigated: one from Sigma-Aldrich (Dorset, UK) produced by recombinant DNA technology and one batch from Merck Millipore, Nottingham, UK (lyophilised bovine mucin isolated from cattle). The Sigma-Aldrich material had a declared molecular weight (MW) of 400 kDa and was a white-yellowish powder with ≤2.5% free and 9%-24% bound sialic acid and was prepared using a modified method described by Tettamanti and Pigman. 37 The material purchased from Merck Millipore was a flaky, red-pinkish, lyophilised hygroscopic powder stated to contain 74% protein, 12.8% sialic acid, 8.1% galactosamine and 3.4% glucosamine (product information sheet) prepared using the method of Nisizawa and Pigman. 38,39 Four batches of 100 ml of AOF containing surfactant and bovine mucin (from each supplier) were prepared (solution 1-Sigma-Aldrich mucin plus Tetronic® 90R4; solution 2-Sigma-Aldrich mucin plus Tween® 20; solution 3-Merck Millipore mucin plus Tetronic® 90R4; solution 4-Merck Millipore mucin plus Tween® 20) (see Table 2).  Table 3).

| Choice of antimicrobial agent
The efficacy of two antimicrobial agents, sodium azide, primarily active against Gram-negative bacterial species, 40 and ProClin™ 300 a broad spectrum microbial inhibitor 41 and potentially safer antibacterial

| Characterisation of AOF
The following physical properties of each batch of AOF were determined in singlicate: Appearance, SG and pH measurement: a 10-ml volume of AOF was withdrawn from fresh 100-ml stock into a glass vial. The appearance (e.g., colour) was noted against a white background and SG, and pH recorded using an automated instrument (Mettler Toledo SC30, Leicester, UK). Light scattering, measurement of the average particle size (z average), the distribution of particle sizes (PDI) and the DCR, which is a measure of particle size and number, was performed using a  Data from the temperature logging device confirmed that temperatures within the laboratory remained at 22.5 C ± 2.5 C.

| Choice of mucin
Human OF can be weakly alkaline to weakly acid, the pH ranging from approximately 6.0 to 8.0, 44  Although data are limited, one report records the specific conductivity for human saliva, on six volunteers, as 1.14 mS cm −1. 35 The specific conductivity of AOF designed for orthodontic research varied between 3 and 27 mS cm −1 . 31 The specific conductivity of our AOF had an average value of 6.74 mS cm −1 within the reported range for orthodontic AOF and showed general consistency across the solutions (see Table 2).

| Choice of surfactant
Tween® 20 was the preferred surfactant over Tetronic® 90R4. It was found that Tween® 20 provided more consistent data than Tetronic® 90R4 when drugs spiked into AOF were analysed by LC-MS, which will be described in more detail in our next publication.

| Choice of antimicrobial agent
In general, a smaller number of colonies was observed with AOF containing ProClin™ 300 than with the AOF containing sodium azide. ProClin™ 300 was found to be more effective in preventing the growth of E. coli and S. aureus, but not P. aeruginosa. This matches documented literature. 53 However, P. aeruginosa is not normally found in the oral cavity, but was chosen as an example of a representative Gram negative bacterium. The antimicrobial agents were less successful in limiting growth of the fungi C. albicans and A. brasiliensis.  Table 3  Similarly, there were differences in light scattering measurements.

| Characterisation of AOF
The average particle size (z average) was larger on day 7 in both solutions containing surfactants (solutions A and B) increasing by approximately, 50% after day 7 (see z average values), although this was less marked in the solution with no surfactant. The DCR decreased over time in all three solutions suggesting particle flocculation. The increase in particle size (aggregates) and decrease in the number of particles after 7 days supported these results (Table 3).
In unstimulated whole saliva, kinematic viscosity has been reported to be 1.40 ± 0.39 mm 2 s −1 (RSD % = 27.81) and in healthy volunteers viscosity decreased exponentially as a function of time after sampling, reaching a plateau around 1.12 mm 2 s −1 (1.12 cSt). 47 Our solutions responded in a similar way, although solution B behaved most like human saliva. We deduced that the preferred combination of surfactant and mucin for our AOF was solution B containing Tween® 20 (Table 3).
3.6 | Stability testing 3.6.1 | Stability of the physical properties of AOF over time using Sigma-Aldrich mucin and Tween® 20 surfactant with ProClin™ 300 antimicrobial agent and at three different temperatures The physical properties of AOF prepared using the surfactant Tween® 20 at a concentration of 900 mg/L with the antimicrobial agent ProClin™ 300 at 20 mg/L and stored at three different temperatures for 28 days are shown in Table 4. Light scattering data also demonstrated small changes over time; the average particle size (z average) increased after 28 days compared with day 0 at all temperatures. Over time, the distribution of particle sizes (PDI) changed less and the number of particles present (DCR) in the AOF was most stable when stored at −20 C ( 3.6.3 | Stability of the physical properties of AOF over time using Sigma-Aldrich mucin and Tween® 20 surfactant with ProClin™ 300 following four freeze-thaw cycles at −20 C A separate batch of AOF stock stored at −20 C underwent four freeze-thaw cycles and the physical properties were compared with the physical properties of AOF at day 0 ( Table 5). The freeze-thaw process did not appear to affect SG, specific conductivity or kinematic viscosity. The freeze-thaw process however had a noticeable effect on the light scattering measurements ( 50%) after the first cycle and then remained stable during future freeze-thaw cycles. The DCR decreased to about a quarter of the original number of particles suggesting particle flocculation due to the freeze-thaw cycles (Table 5). Whembolua et al. 54 suggests that human OF should be refrigerated immediately at 4 C if freezing is not possible to minimise degradation of any unstable analytes and to prevent bacterial growth but that storage at this temperature should be for no longer than necessary before freezing at or below −20 C. Although in our study the freeze-thaw cycles induced the particle flocculation, a subsequent centrifugation step afterwards may allow easier handling.
3.6.4 | The physical properties of AOF over time (storage for 3 months at −20 C) using Sigma-Aldrich mucin and Tween® 20 surfactant with ProClin™ 300 In order to investigate whether long-term storage of prepared material was possible, we analysed a batch of AOF stored for 3 months at −20 C (Table 6). There was little difference in the pH, SG and specific conductivity after 3 months. We observed a slight decrease in kinematic viscosity and changes in light scattering measurements (Table 6).

| LC-MS/MS analysis
The results of the LC-MS/MS analysis of the AOF spiked at 50 ng/ml with THC and stored at different temperatures and time points are shown in Figure 3 relative to the initial measured concentration.
These initial data show that THC appears to be stable in the AOF for 7 days at 4 C, but perhaps less so at RT or at −20 C. Further evaluation of the stability of THC and other drugs of abuse in AOF at different storage temperatures and times is the subject of a further communication.

| CONCLUSIONS
We investigated the physical properties of AOF to determine the usefulness of AOF as a traceable matrix for drug testing purposes. Evaluation of pH, SG, specific conductivity (mS cm −1 ), freezing point depression ( C), light-scattering and kinematic viscosity (mm 2 s −1 ) showed our AOF to be a stable and reliable matrix. Two different types of mucin were investigated and mucin (produced by recombinant DNA technology) from Sigma-Aldrich better reflected the characteristics of human OF in terms of viscosity (spinnbarkeit) and particulate distribution when compared with Merck Millipore mucin; the user should take care that different batches give consistent results. The use of a surfactant was important to stabilise the viscosity of the AOF. Tween® 20 maintained particle size and uniformity more consistently than Tetronic® 90R4.
Tween® 20 was found to provide more consistent data than Tetronic® 90R4 when drugs spiked into AOF were analysed by F I G U R E 3 Liquid chromatography with tandem mass spectrometry (LC-MS/MS) normalised (relative to day 0) data for artificial oral fluid (AOF) spiked with THC at 50 ng/ml and stored at room temperature (RT), 4 C or −20 C for 7 days [Colour figure can be viewed at wileyonlinelibrary.com] LC-MS. ProClin™ 300 was the preferred choice over the more toxic antibacterial sodium azide. We have established that freezing AOF at or below −20 C immediately after preparation helped preserve sample integrity. However, the flocculation of the particles was noticed after storage at −20 C.
This study provides evidence of the suitability of AOF for quality assurance purposes using Sigma-Aldrich mucin and Tween® 20 such as would be needed when large quantities of OF were required for widescale drug testing purposes such as for a national road-side drug testing scheme. No comparison (with reproducibility in mind) with human OF has been made, which would in any case be difficult given the variability of human OF. This is clearly a scientific weakness that needs to be kept in mind when considering the use of human OF in analytical toxicology irrespective of the use of AOF. However, we believe that our AOF being formulated to be similar to human OF benefits in consistency and is easy to prepare. An initial investigation seems to show the stability, at least of THC, at 4 C for 7 days. A further investigation of the application of the AOF described in this paper, in terms of storage conditions, for the LC-MS measurement of a range of drugs commonly controlled and tested for in both workplace drug testing programmes and roadside drug tests will be carried out.