TECHNICAL APPENDICES
Appendix 1: Creatinine analysis
Frozen samples were allowed to thaw slowly overnight at 2-8°C. A 10ml aliquot was removed and the pH adjusted to 1.5 – 2.0 with 5M HCl. Samples were clarified by passing through a 5.0µm filter and a sample was taken for creatinine analysis.
Creatinine analysis was carried out using Sigma Diagnostics creatinine kit (555-A, Lot 020K 6045), according to the published methodology with the following modifications: All volumes used were 50% of those stated in the published protocol to permit analysis in a 96 well plate format. Tubes received either 150µl of water (blank), or 150µl of a 3µg/dl standard or 150µl of thawed unprocessed urine sample diluted 1:10 in HPLC grade water. 1.5ml of Alkaline Picrate solution (APS) was added to every tube, the tubes were mixed and allowed to stand. 250µl was then transferred from each tube to a 96 well plate and absorbance was determined for each well 8-12 minutes after the addition of APS. This plate was then discarded. To the original tube was then added 50µl of acid reagent, and then 250µl was transferred from each tube into a fresh 96 well plate. Absorbance was determined for each well after a further 5 minutes. Creatinine values were calculated as follows:
Initial Abstest – Final Abstest
Initial Absstd – Final AbsStd
Appendix 2: Chromatography run conditions
All chromatography was conducted using a reversed phase C18 column (ACE 5 C18 250 x 4.6mm [cat. no. ACE-121-2546]) with a Phenomenex pre-column guard (AJO-6073).[1] comprising cooled auto sampler (A-905), fraction collector (Frac-950), column heater/cooler (Jones Chromatography 7995), in-line solvent de-gasser (Merck L-7614) and a triple wavelength UV detector
UV absorbance measurements (215nm) were automatically taken every 0.8 seconds over the first 118 minutes of the run, generating an ASCII file listing approximately 8800 data points.
The chromatography run conditions were a modification of those provided by K.L. Reichelt (personal communication.) and were based upon those published by Ek et al.(1)
Buffer A: 0.1% trifluoroacetic acid (TFA) in 5% acetonitrile/94.9% dH2O. Buffer B: 0.1% TFA in 94.9% acetonitrile/5% dH2O. Flow rate was set at 1 ml/min. Run conditions were as follows:- Step1, column pre-equilibration; -16 to 0 mins 100% A: 0% B. Step 2, sample injection at 0 mins. Step 3, 0-18 mins 100% A: 0% B (to wash through amino acids, urea and salts). Step 4, between 18-78 mins a linear gradient of 100% A: 0% B to 60% A: 40% B (elution stage for metabolites of interest) . Step 5, between 78-88 mins a constant 60% A: 40% B. Step 6, between 88 and 93 mins a linear gradient of 60% A: 40% B to 40% A: 60% B. Step 7, between 93-109 mins 40% A: 60% B. Step 8, from 109-125 mins 100% A: 0% B.
References
(1) Ek J, Stensrud M, Reichelt KL. Gluten-free diet decreases urinary peptide levels in children with celiac disease. Journal of Pediatric Gastroenterology & Nutrition 29(3):282-5, 1999 September.
Appendix 3: Determination and stability of putative opiod peptides in urine
a) Determination of putative opioid peptides in urine
Synthetic opioid peptides (Table A1) were purchased from Bachem Ltd, St. Helens UK. The HPLC retention time for each of the peptides in urine was established and visually compared against all urinary chromatograms. Any urine samples exhibiting peaks at the appropriate retention times were re-fractionated and fractions were collected every minute. Fractions immediately surrounding the expected retention times of any opioid peptides were individually concentrated tenfold using a Speed VacTM and 0.5µl of concentrated sample was applied on two individual spots onto the MALDI-TOF analysis plate, followed by 0.5µl of matrix per spot (α-cyano-4-hydroxycinnamic acid; Ciphergen). Sample spots were allowed to air dry and then analysed using a PerSeptive Biosystems Voyager DE PRO Biospectrometry work station linked to a Tektronix TDS 540C digitizing oscilloscope. Mass spectra were acquired using an automated programme and calibrated using an external calibrant applied to an adjacent spot on the plate (SequazymeTM Calibrant 1 mass standards, Applied BioSystems). Mass spectra were collected in reflector mode and the traces were manually examined for the presence of masses (m/z) corresponding to the expected opioid peptide in the particular HPLC fractions.
Table A1
Peptide / Product / Type / Retention time min / Monoisotopic mol. weight / Stability in urine 20 hrsA / B
Met-enkephalin 1-5
H-Tyr-Gly-Gly-Phe-Met-OH / H-2785 / Endorphin / 54.1 / 573.23 / NT / NT
Neuropeptide-FF
H-Phe-Leu-Phe-Gln-Pro-Gln-Arg-Phe-NH2 / H-5655 / Endorphin / 63.4 / 1080.59 / 0% / 93%
Dermorphin
H-Tyr-D-Ala-Phe- Gly-Tyr-Pro-Ser-NH2 / H-2565 / Exorphin / 55.8 / 802.37 / NT / NT
β-Casomorphine 7 (bovine)
H-Tyr-Pro-Phe-Pro-Gly-Pro-Ile-OH / H-2280 / Exorphin / 62.8 / 789.41 / 40% / 100%
Gluten exorphin B5
H-Tyr-Gly-Gly-Trp-Leu-OH / H-1666 / Exorphin / 62.2 / 594.28 / 95% / 100%
Gluten exorphin C
H-Tyr-Pro-Ile-Ser-Leu-OH / H-1412 / Exorphin / 58.1 / 591.33 / NT / NT
b) Stability of synthetic opioid peptides in urine.
The stabilities of exorphins β-casomorphin 7 and gluten exorphin B5, together with the endorphin neuropeptide FF were examined in detail under the collection, storage and processing conditions employed in this study to ensure that such materials were not lost prior to analysis. Some peptide degradation was apparent in urine which was not pH-adjusted, this being more noticeable at elevated temperatures. For example, only 5% of the endorphin neuropeptide FF remained after a 12 hour period at 20oC, while the amounts remaining for exorphins β-casomorphine 7 and gluten exorphin B5 were 60% and 100% respectively. However, when the pH was lowered to 2.0, and the temperature to 10oC, no apparent degradation occurred over a 20 hour period (Table A1, columns 6A and 6B). Since these were the assay conditions used in this study, if these 3 opioid peptides had been present in the urine samples at the time of loading, they should not have been lost prior to HPLC analysis. However, this does not discount the possibility of peptide degradation while the urine was stored in the bladder. The freeze/thaw stability of β-casomorphin 7, gluten exorphin B5, and neuropeptide FF in urine, and the unidentified urinary peaks 33.8, 39.5, 46.9, 52.5 and 55.7 (Table 2) were also examined by subjecting urine samples to 3 freeze/thaw cycles. No loss of peak height was observed following this procedure (data not shown). These stability studies suggested that such compounds were not lost from the urine during the collection, pre-processing steps and analysis procedures.
Appendix 4: Lower limit of detection of synthetic peptides in urine
HPLC: A single urine sample was adjusted to pH 2.0 and filtered through a 0.22μM syringe filter (Sartorius). Pure opioid peptides were then added to a final concentration of 50, 25, 12.5 and 5 μg/ml and each dilution was then run in triplicate on the HPLC. Each peptide was analysed in a separate urine sample to prevent possible interference effects. MALDI-TOF: A single urine sample was adjusted to pH 2.0 and filtered through a 0.22μM syringe filter and pure opioid peptides were individually added to separate urine aliquots to a final concentration of between 15μg/ml – 0.2μg/ml. Urine samples cannot be analysed directly with MALDI-TOF MS and were de-salted by solid phase extraction using a Sep-Pak light cartridge (Waters WAT023501). The cartridge was first conditioned with 3ml of 100% methanol, followed by 3ml of 0.1% TFA. The urine sample (3ml) was then loaded, followed by 3 ml of washing solution (5% methanol, 0.1% TFA). Sample was eluted with 1.5ml of 50% acetonitrile, 15% methanol in 0.1% TFA. 0.5 μl of eluate was then spotted directly onto the MALDI-TOF plate prior to analysis. The lower limit of detection was taken as the lowest concentration which gave a signal which was 2.5x that of baseline noise.
The lower limits of detection (LLOD) for various opioid peptides were investigated by MALDI-TOF MS and HPLC (Table A2) by spiking urine with decreasing amounts of pure peptide and then following the same analysis procedures undertaken for the urine samples used in this study. The LLOD was defined as representing the lowest concentration of peptide which reproducibly gave a signal 2.5 x greater than the surrounding background signal. Using HPLC, the LLOD of all 3 peptides tested were fairly similar, at approximately 5 μg/ml in urine (equivalent to 6-9 μM depending upon the peptide). MALDI-TOF MS was observed to be more sensitive than HPLC, although the difference was relatively small (2-4 fold increase in sensitivity). This was most likely due to the presence of salts and other contaminants in the urine suppressing the signal intensity or a loss of peptides during pre-analysis procedures. The MALDI-TOF MS peptide signal intensity was found to be semi quantitative in response to concentration changes, while the HPLC absorbance measurement (both peak area and peak height) was highly quantitative (R2 > 0.9; data not shown).
Table A2 The reproducible lower limit of detection (LLOD) of synthetic opioid peptides spiked into urine and tested by either MALDI-TOF mass spectrometry or HPLC. NT = not tested
Peptide / MALDI-TOF MS / HPLCGluten exorphin B5 / 2 μg/ml / 5 μg/ml (8.4 μM)
β-casomorphin 7 / 0.75 μg/ml / 5 μg/ml (6.3 μM)
Dermorphin / NT / 5 μg/ml (6.2 μM)
Appendix 5: System reproducibility & statistical analysis
The percentage co-efficient of variation for system reproducibility was calculated as the standard deviation (SD) divided by the mean multiplied by 100%. Statistical analyses of the mean responses from different groups were carried out using Minitab 13 and Genstat 8.1. These were comparisons of total peak area (TPA) per sample for the full spectrum (0 – 78 minutes retention time) and for the truncated spectrum (40 – 78 minutes), using either raw uncorrected data, or the data which had been corrected for urinary creatinine concentration, and a comparison of the data summarising the urinary creatinine levels. Initial investigation of the data suggested that each variate required transformation to meet the distributional assumptions required for ANOVA methods. The TPA dataset of the full and truncated spectra and the creatinine variates were transformed using a square-root transformation, since a log-transform was found to be too extreme. The creatinine-corrected TPA variates for both full and truncated spectra were even more skewed, and required a repeated log transformation to make the residual distribution acceptably symmetrical, although the tails of the distribution were not consistent with a normal distribution (p<0.001, using the Anderson-Darling test). In each case, analysis was carried out using oneway ANOVA in Minitab 13, and p-values were controlled to allow for multiple comparisons using both Fishers and Dunnett’s methods, where the individual and family error rates respectively were 5%, and the Normal group was selected as the baseline comparison for the Dunnett procedure. In the case of creatinine corrected TPA variates for full and truncated spectra, to confirm that the conclusions of the analysis were not compromised by the lack of normality in the fit of the model to the data, the data were reanalysed using a randomisation test in Genstat 8.1. The p-values for the differences between the observed means (evaluated by considering the difference between the maximum and minimum of the four means) under the null hypothesis of no differences between groups were equivalent to those estimated using ANOVA, confirming the validity of these results.
Validation of assay techniques
Reproducibility. HPLC system reproducibility for peak retention time, peak height and peak area was established (Table A3). Firstly, identical amounts (100 μg/ml) of synthetic opioid peptides β-casomorphin 7 and gluten exorphin B5 were ‘spiked’ into 10 different urine samples and individual chromatographs generated. Over the 10 runs the coefficients of variation for the retention time, peak height, and peak area respectively were 0.05%, 2.7% and 3.6% (β-casomorphin 7) or 0.18%, 1.8% and 7.4% (gluten exorphin B5). Secondly, peaks representing unknown urinary components within a single urine sample were analysed over multiple runs; 3 runs per day run over 5 days (i.e. 15 runs in total). Five well-defined peaks with retention times from 33.8 – 55.7 minutes were chosen for analysis. System reproducibility was very accurate, with peak retention times exhibiting only 0.2 – 0.4% variability over the 15 runs. This allowed a degree of confidence when comparing apparently identical peaks (based upon retention times) present in different urine samples. The variability of urinary peak heights and areas were greater compared to that observed for the synthetic opioid peptides, at between 0.9-16.5% dependent upon peak.
Table A3 HPLC system reproducibility for synthetic opioid peptides and for unidentified urinary peaks over 10(synthetic peptides) or 15(unidentified urinary peaks) traces, expressed as co-efficient of variation.
Co-efficient of variation over 10a or 15b HPLC runs for:Retention time / Peak area / Peak height
Synthetic peptidesa
β-casomorphin 7 / 0.05% / 3.6% / 2.7%
Gluten exorphin B5 / 0.18% / 7.4% / 1.8%
Unidentified urine peaksb (retention time in minutes)
33.8 / 0.4% / 7.7% / 4.4%
39.5 / 0.4% / 3.8% / 0.9%
46.9 / 0.3% / 4.6% / 2.5%
52.5 / 0.3% / 16.5% / 2.6%
55.7 / 0.2% / 5.4% / 10.2%
aComparing peaks representing synthetic peptides spiked into 10 different urine samples. bComparing peaks within a single urine sample analysed on 15 different occasions.
Specificity of MALDI-TOF
A control experiment was performed to demonstrate the accuracy and specificity of the system. A mixture of opioid peptides met-enkephalin (m/z +1 = 574.23, retention time 54.1 mins) and gluten exorphin B5 (m/z + 1 = 595.29, retention time 62.2 mins) was run through the HPLC system, and individual fractions surrounding the expected elution time were collected and analysed by MALDI-TOF MS (Figure A1). The met-enkephalin peak was spread over Fractions A4 and A5 and absent in Fraction A6 on the HPLC; this is exactly mirrored by the MALDI-TOF MS traces, where m/z peaks of 574.18 and 574.27 are observed in Fractions A4 and A5 respectively, being completely absent in Fraction A6. Similarly with peptide gluten exorphin B5, both the HPLC peak and the expected ion (595.31) are contained within Fraction A12 only.
Figure A1 HPLC trace of pure peptides met-enkephalin and gluten exorphin B5 showing fractions taken for MALDI-TOF MS analysis (upper graph, x-axis shows retention time in minutes). The mass to charge (m/z) ratios of respective peptides are shown on the MALDI traces (lower graph; y-axis shows relative peak intensity with largest peak being 100%).
Appendix 6: Comparison of uncorrected average urinary trace from ASD and control samples over the full 78 minute run.
Appendix 7
Comparison of uncorrected average urinary trace from ASD subgroups and control samples.
1
[1]Technical Appendix 2