Abstract Objectives: The HLA-B*27 genotype is strongly associated with ankylosing spondylitis (AS) and also other spondyloarthropathies (SpA) such as psoriatic SpA. Our objective was to develop and validate a real time PCR for HLA-B*27 with greater capacity than our current allele specific approach without loss of fidelity.

Methods: We describe a real-time PCR method for the detection of HLA-B*27 using allele specific primers combined with TaqMan fluorescent minor groove binder (MGB) probes. One probe is specific for human leucocyte antigen (HLA) B*27 conjugated with the fluorochrome FAM and the other specific for human growth hormone (HGH) conjugated with the fluorochrome VIC which acts as endogenous control.

Results: 441 consecutive samples were tested of which 139 tested HLA-B*27 positive and 302 tested negative. We tested a further 12 external quality control samples of which four were positive and eight negative for HLA-B*27.

Conclusions: RT-PCR is suitable for HLA-B*27 genotyping and has the advantages of lesser hands-on time, does not require any post amplification processing, and does not use the toxic DNA intercalating dye ethidium bromide.

Keywords: HLA-B*27, TaqMan, PCR, MGB, minor groove binder, NFQ, non-fluorescent quencher, real time PCR, ankylosing spondylitis.

N Z J Med Lab Sci 2014; 68:

Introduction The human leucocyte antigen (HLA) B*27 is strongly associated with the spondyloarthropathies (SpA’s). HLA-B*27 is found in 90-95% of ankylosing spondylitis (AS) patients and also lesser frequency in forms of psoriatic arthritis and reactive arthritis (1,2). HLA-B*27 may have a direct role in the pathogenesis of SpA which is supported by epidemiological studies and transgenic rat studies (3). Different hypotheses exist, such as the arthritogenic peptide hypothesis, the misfolding hypothesis or molecular mimicry (3-5). Less than 5% of those members of the general population that are HLA-B*27+ develop AS, and HLA-B*27 has been estimated to account for only 20-50% of the overall genetic susceptibility to AS (6). In addition to HLA-B*27, a recent genome wide association study has confirmed 25 loci that confer increased genetic risk for AS, including genes involved in peptide processing prior to MHC class I presentation and alterations in the IL-23 pathway (7).

There are more than 50 subtypes of HLA-B27 identified which mostly differ by only a small number of nucleotides. (IMGT/HLA database: http:\\www.ebi.ac.uk/imgt/hla/). The subtypes HLA-B*27:02, B*27:04, B*27:05, B*27:07, B*2:708 are associated with AS although B*27:07 is suggested to be associated but not in Greek Cypriots (8,9). Important conformational differences exist in subtypes not associated with AS (4,10-14).

Many laboratories use an allele specific PCR based on that of Sayer et al (17), which offers excellent specificity for HLA-B*27 although they report a lack of sensitivity for subtype HLA-B*27:12. This lack of sensitivity is also likely for the subtypes, HLA-B*27:16, HLA-B*27:18, HLA-B*27:23, HLA-B*27:29.

The aim of our study was to adapt our conventional allele specific HLA-B27 protocol to a fluorogenic real time (RT-PCR) method. The allele specific PCR detects the presence of the B*27 genotype by amplifying a region between primer sets that recognise only B*27 specific sequences. The PCR reaction amplifies this region to sufficient quantity so that it may be visualised after electrophoresis in agarose gel. Visualisation is achieved by staining with ethidium bromide (a fluorescent DNA intercalating dye) and viewing the gel under UV light (Figure 2). Both allele specific and RT-PCR methods detect the presence or absence of B*27 and do not discriminate between heterozygous and homozygous individuals. The RT-PCR method involves the addition of specific dual labelled hydrolysis probes to the PCR. These TaqMan probes have a fluorescent reporter (ie. FAM or VIC) attached to the 5 end of the probe sequence and a non-fluorescent quencher (NFQ) attached to the 3 end. Addition of a minor groove binder (MGB) to the probe sequence at the 3 end of the probe permits it to anneal at a higher melting temperature (Tm), allowing the use of shorter more specific probes (2,18). The probe in free form is normally quenched as random coiling brings the probe and fluorophore close together (19). The MGB works together with the quencher to improve quenching. During the extension step of the PCR, the probe binds to its complementary sequence at a position that sits between the forward and reverse primers.(Fig 1) During extension Taq polymerase cleaves the probe, because of intrinsic 5 to 3 nuclease activity, releasing the fluorescent reporter (FAM or VIC) into the mix thus causing an irreversible increase in fluorescence(19). As the cycles continue fluorescence increases rapidly until the probe is exhausted. The advantages of this method are that primers and probes provide specificity to the reaction products and allow immediate visualisation. It is possible to use several different probes in each PCR to simultaneously detect specific products. In this case both the HLA-B*27 and HGH (internal control) products are detected independently. The analytical capacity of this technique is much greater than the allele specific PCR and allows for near complete automation.

Figure 1A. Target DNA, primers and probes are mixed together.

Figure 1B. After melting, primers and probes are allowed to bind to their complementary sequences.

HLA-B27 real-time PCR using TaqMan-MGB sequence specific probes Stewart M Smith, Andrew D Laurie, Howard C Potter and Ben D McGettigan

NZ J Med Lab Science 20144

Figure 1C. Taq Polymerase extends primers and in the process its nuclease activity removes probes causing an irreversible increase in fluorescence in the process. MGB: minor groove binder. NFQ: non-fluorescent quencher. Pol: polymerase. FAM: fluorescent dye (6-Carboxyfluoroscein).

Materials and methods Sample DNA DNA was isolated from EDTA whole blood samples from 441 consecutive sample sent for HLA-B*27 testing at our laboratory with a TECAN Genesis RSP 100 liquid handling system, using the AGOWA® mag Maxi DNA Isolation Kit.

An additional 12 external control samples from the Royal College of Pathologists of Australasia Quality Assurance Programs Pty Ltd (RCPA QAP) HLA-B*27 program were prepared. The average DNA concentration as measured by NanoDrop ND-1000 ver 3.1.2 software, National Instrument Corporation, was 20ng/μl with the 5th to 95th percentile being 11 – 34 ng/μl. A 5μl aliquot of DNA was used for each PCR reaction, and therefore approximately 100ng of each sample was used, although there was considerable variation.

B7 positive control samples Four of the HLA-B*27 external quality assurance samples provided by RCPA Immunology QAP were reported as heterozygote for HLA-B*7 and negative for HLA-B*27. These samples were included as part of the validation to provide assurance of primer specificity.

Allele specific PCR Allele specific PCR was carried out according to the method of Sayer et al (17). Briefly, sequence-specific primers amplify a 141bp band from exon 2 of the HLA-B*27 family. Internal PCR amplification control primers specific for the human growth hormone (HGH) are used for each sample to ensure there is sufficient sample DNA for amplification to proceed or to show amplification failures for any other reason. This internal control produces a 437bp band which must be present to validate the assay. The PCR was performed in a total volume of 25μl containing 2.5μl of 10x Roche FastStart PCR Buffer with 2mM MgCl2, 2.5μl of 2mM dNTP’s, 5.0μl of 5x FastStart GC rich buffer, 1U FastStart Taq polymerase. The reactions contained 560nM B*27 specific primers and 280nM HGH specific primers along with 5.0μl of 1/5 diluted DNA from each sample.

Figure 2. Conventional allele specific PCR. Reverse black/white image of typical 2% agarose gel showing 437bp human growth hormone specific band and positive samples also displaying the 141bp B*27 specific band. Samples amplified with conventional allele specific PCR.

One cycle of denaturation at 96°C for 4 minutes was followed by 30 cycles of 96°C for 30 seconds, 66°C for 2 minutes, 72°C for 30 seconds followed by a final extension at 72°C for 5 minutes. Amplified products were visualised under UV light after electrophoresis on a 2% agarose gel and staining with ethidium bromide.

Real time PCR Allele specific PCR primers were used as above, according to the method of Sayer et al (17). The Alt HGH internal control primers / HGHVIC1 probe used were different from the ones above and they were designed using ABI Primer Express v3.0 software (Table 1). The same B27 primers above were used combined with a 22 nt. fluorescent TaqMan probe B27FAM1, which was designed manually in a configuration allowing efficient binding between the B27EX294F / B27EX2199RC primers (Table 1). Both TaqMan probes were purchased from Applied Biosystems while primers were purchased from Sigma. Probe concentrations of 0.25μM and Primer concentrations of 0.5μM were used in PCR reactions. Samples were then amplified using the Roche Lightcycler 480 in 15μl reaction volumes using 10μl of PCR buffer (Roche Lightcycler 480 Probes master mix), primer/probe mix and 5μl of undiluted DNA. One cycle of denaturation at 95°C for 5 minutes was followed by 40 cycles of 95°C for 15 seconds, 66°C for 15 seconds, and 72°C for 30 seconds. The intensity of fluorescence of the two different fluorophores was measured each cycle as the PCR proceeded in real time. Samples are considered positive when fluorescence intensity rose above the threshold value, which usually occurs from cycle 26 to 30.

Table 1. Primers and TaqMan probes used in the RT-PCR HLA -B*27 genotyping procedure.

PCR and the signal from each probe was separated by its distinct emission spectra collected using appropriate filter sets. Curves were then constructed with Cycle number (x-axis) vs log fluorescence intensity (y-axis) for each well. Fluorescence intensity for the B*27 positive samples was usually about 15 times that of negative controls and blanks by cycles 35 to 40. When the fluorescence level rose above the discriminator (horizontal dashed line) the sample was scored as positive for that probe.

NZ J Med Lab Science 20145

Figure 3. A curve for each well is constructed for HGH (Figure 3A) and HLA-B*27 (Figure 3B) specific probes with cycle number (x-axis) vs. log fluorescence intensity (y-axis). The discriminator (horizontal dashed line) is set manually and results calculated for each well. Positivity is expressed as a red square for those wells where fluorescence levels are above the threshold.

Results Of the 453 comparative samples tested, 100% concordance with the conventional allele specific PCR was recorded.

Table 2. Comparison of results between allele specific and RT-PCR.

Allele SpecificPCR RT-PCR

HLA-B*27 positive 139 139 HLA-B*27 negative 302 302 RCPA QAP HLA-B*27 positive 4 4 RCPA QAP HLA-B*27 negative 8* 8*

Total 453 453

*Four of the HLA-B*27 external quality assurance samplesprovided by the RCPA Immunology QAP program were reported as heterozygote for HLA-B*7.

Effects of DNA Dilution Five HLA-B*27 positive and two HLA-B*27 negative samples were chosen to perform dilution experiments. These samples were all extracted as described previously and had DNA concentrations ranging from 15 – 65ng/μl. Each sample was serially diluted eight times, with dH2O so as to obtain a halving of DNA in each successive dilution and these samples were then analysed by the RT-PCR assay. This gave a range of DNA concentrations over which the sensitivity of this test could be measured, from 0.12 – 65ng/μl. Ideally for each doubling dilution of DNA one would expect to see a one cycle increase in the cycle number at which fluorescence levels increase above an arbitrary threshold. This would hold true so long as the threshold is set in the exponential phase of amplification.

The question of sensitivity of detection can be answered by observing the lowest concentration at which fluorescence is clearly distinguishable from negative samples and blank reactions containing no DNA. Figures 4A and 4B below show how reactions proceeded for both the HGH and B*27 specific fluorescent probes. Fluorescence increased exponentially until buffer constituents were exhausted whereupon the fluorescence plateaus. The figures show that positive signals were obtained over the full range of concentrations used. Of note, Figure 4B shows that when DNA concentration increased

from 123ng to 246ng there was no difference in the cycle number threshold suggesting saturation of amplification at higher concentrations. The sensitivity of detection was therefore at least 0.12ng/μl. This is well below the lowest concentration of DNA isolated by our DNA isolation system by a factor of at least 10.

Figure 4. VIC fluorescence from HGH (Figure 4A) and FAM fluorescence from B*27 (Figure 4B) specific probes. Fluorescence levels cross the threshold on average one cycle later for each halving of input DNA .

Fluorescence intensity for the B*27 specific probe reached the plateau level in all dilutions for all five samples tested (data not shown). Fluorescence intensity for the HGH specific probe did not reach plateau levels in all dilutions and Figure 4A is typical of the seven samples tested.

By graphing the cycle number value against DNA concentration, all the samples tested can be represented with clarity that would be lost if every fluorescence curve was represented as in Figure 4. Figures 5A and 5B show a relatively linear decrease in cycle number (average 1.0 – 1.1 cycles) required to reach the exponential phase of amplification as a function of halving DNA concentration. Samples 2, 3, and 4 departed from this linearity at higher DNA concentrations for the B*27 probe. Sample 2 also departed significantly at lower concentrations and was observed to cross the threshold line 2.2 cycles later than the next highest dilution for the B*27 probe rather than 1. This demonstrates that wide ranges of DNA concentrations can be used in the RT-PCR, however, at the extremes of DNA concentration measured the reaction efficiency is not optimal.

NZ J Med Lab Science 20146

Discussion We have shown that a RT-PCR for the detection of HLA-B*27 can be performed that gives 100% concordance with our current allele specific PCR. Furthermore it also showed 100% agreement with external control samples. The method showed no cross reactivity with the four HLA-B*7 positive samples available. DNA concentration did not affect interpretation of this test when DNA is prepared by our automated DNA extraction method. The method sensitivity to DNA concentration was well within the range isolated routinely in our laboratory. There were signs of departure from linearity for some samples at higher DNA concentrations and for one sample at low concentrations which did not affect the ability to discriminate whether the samples were positive or negative.

Other investigators have shown applications using a Lightcycler PCR thermocycler such as Sylvain et al (2) and Bon et al ( 20). Bon et al (20) used SYBR green I and melting curve analysis for genotyping of HLA-B27 and Sylvain et al (2) used a TaqMan-PCR assay using sequence-specific primers and a MGB fluorescent probe with the use of an artificial exogenous internal positive control (IPC) DNA added to each PCR reaction. This was used to distinguish true negatives from false negatives. Our assay is advantageous in that it does not use artificial exogenous control material, but instead the HGH gene as the internal control. This is an important distinction as both the quantity and quality of DNA isolated, including the presence of any PCR inhibitors, is reflected by the HGH internal control. This is something that exogenous control material cannot do.

Of note Sayer DC et al (17) reported a lack of sensitivity for HLA-B*27:12 because of a different B pocket configuration in its peptide binding groove and differs from HLA-B*27:05 AS associated F pocket configuration. For these reasons it is unlikely to be associated with AS(17,21), although it may be associated with reactive arthritis (22).

HLA-B*27:18, HLA-B*27:23 and HLA-B*27:29 have altered B pockets because of a substitution at residue 67 (Ser, Phe and Ser respectively instead of Cys (6,23,24). This mutation in transgenic rats leads to gut disease but markedly less susceptibility to arthritis (25). HLA-B*27:16, HLA-B*27:18, HLA-B*27:23 and HLA-B*27:29 as well as HLA-B*27:12 all share Thr/Asn/Thr at residues 69/70/71 instead of Ala/Lys/Ala shared by other B*27 (23,24). This substitution is within the B pocket and for these rare subtypes disease association with AS is not established.

A number of subtypes contain single nucleotide mismatches with one of the primers used in this study, B27EX2199RC. The effect of primer-template mismatches has been studied by others such as Kwok et al (26). They found “… that a single mismatch 3 residues from the 3 terminal base of a primer can be efficiently extended without modification of amplification reactions.” Wu et al (27) however “… observed that the single mismatches present at the last five positions from the primer 3 end could totally block the primer extension.” The one base pair mismatches in the other rarer subtypes HLA-B*27:26, HLA-B*27:31 and HLA-B*27:39 are unlikely to prevent amplification as they occur at positions 9 to 16 from the 3 end of the primer. However, the one base pair mismatch for the rarer subtype HLA-B*27:0509 occurs at position 5 from the 3 end of the primer, which may interfere with binding according to Wu et al (27). Primer B27EX2199RC has been observed to correctly amplify up to an annealing temperature of 69°C (data not shown) and as our RT-PCR assay routinely anneals 3°C lower, at 66°C, this may give the primer some mismatch tolerance.

In conclusion, this real time PCR reliably and accurately types whole blood samples for HLA-B27. It is very easy to automate using the AGOWA® mag Maxi DNA isolation kit for extraction of DNA through to end result as there is no post PCR gel electrophoresis and photographic interpretation. This minimises contamination risk of post amplification PCR products and overcomes the use of the DNA intercalating dye and known carcinogen ethidium bromide. This technique is suitable for HLA-B*27 genotyping in the routine laboratory.

Acknowledgements This work was supported by the Immunology and Molecular Pathology sections at Canterbury Health Laboratories.

Author information Stewart Smith, DipMLT, Medical Laboratory Scientist1 Andrew D Laurie, PhD, Scientific Officer2

Howard C Potter, MSc FNZIMLS MHGSA, Acting Section Head2

Ben D McGettigan, MB ChB, Registrar1

1Immunology Laboratory and 2Molecular Pathology Laboratory, Canterbury Health Laboratories, Canterbury District Health Board, Christchurch, New Zealand.

Corresponding author: Stewart Smith, Immunology, Canterbury Health Laboratories, PO Box 151, Christchurch, New Zealand. Email: stewart.smith@cdhb.health.nz

References 1. Seipp MT, Erali M, Wies RL, Wittwer C. HLA-B27 typing:

evaluation of an allele-specific PCR melting assay and twoflow cytometric antigen assays. Cytometry B Clin Cytom2005; 63: 10–15.

2. Sylvain K, Aurélie H, Marc M, Christophe R. Rapidscreening for HLA-B27 by a TaqMan-PCR assay usingsequence-specific primers and a minor groove binder probe,a novel type of TaqMan probe. J Immunol Methods 2004;287: 179– 186.

3. Marcilla M, López de Castro JA. Peptides: the cornerstoneof HLA-B27 biology and pathogenetic role inspondyloarthritis. Tissue Antigens 2008; 71:4 95–506.

Figure 5. Total DNA per PCR reaction (x-axis) vs. cycle number (y-axis) for HGH specific probe (A) and B*27 specific probe (B). Each line represents a single sample and is constructed by plotting the cycle at which fluorescence first crosses the threshold line at each DNA dilution. The average crossing time (Ct) increase is between 1.0 and 1.1 cycles.

NZ J Med Lab Science 20147

5. Reveille JD. Major histocompatibility genes and ankylosingspondylitis. Best Pract Res Clin Rheumatol 2006; 20: 601–609.

6. Végvári A, Szabó Z, Szántó S, Glant TT, Mikecz K,Szekanecz Z. The genetic background of ankylosingspondylitis. Joint Bone Spine 2009; 76: 623-628.

7. International Genetics of Ankylosing Spondylitis Consortium(IGAS), Cortes A, Hadler J, Pointon JP, Robinson PC,Karaderi T, et al. Identification of multiple risk variants forankylosing spondylitis through high-density genotyping ofimmune-related loci. Nat Genet 201345: 730-738.

8. Akkoc N, Khan MA. Etiopathogenic role of HLA-B27 allelesin ankylosing spondylitis. APLAR J Rheumatol 2005; 8: 146–153.

9. Gómez P, Montserrat V, Marcilla M, Paradela A, de CastroJA. B*2707 differs in peptide specificity from B*2705 andB*2704 as much as from HLA-B27 subtypes not associatedto spondyloarthritis. Eur J Immunol 2006 36 : 1867–1881.

10. Fabian H, Huser H, Narzi D, Misselwitz R, Loll B, Ziegler A,et al. HLA-B27 subtypes differentially associated withdisease exhibit conformational differences in solution. J MolBiol 2008; 376: 798–810.

11. Hülsmeyer M, Hillig RC, Volz A, Rühl M, Schröder W,Saenger W, et al. HLA-B27 subtypes differentiallyassociated with disease exhibit subtle structural alterations.J Biol Chem 2002; 277: 47844–47853.

12. Hülsmeyer M, Welfle K, Pöhlmann T, Misselwitz R, AlexievU, Welfle H, et al. Thermodynamic and structuralequivalence of two HLA-B27 subtypes complexed with a self-peptide. J Mol Biol 2005; 346:1 367–1379.

13. Sesma L, Montserrat V, Lamas JR, Marina A, Vásquez J,López de Castro JA. The peptide repertoires of HLA-B27subtypes differentially associated to spondyloarthropathy(B*2704 and B*2706) differ by specific changes at threeanchor positions. J Biol Chem 2002 ; 277: 16744-16749.

14. Sidney J, Peters B, Frahm N, Brander C, Sette A. HLA classI supertypes: a revised and updated classification. BMCImmunol 2008; 9:1.

15. Ward MM, Hendrey MR, Malley JD, Learch TJ, Davis JC Jr,Reveille JD, et al. Clinical and immunogenetic prognosticfactors for radiographic severity in ankylosing spondylitis.Arthritis Rheum 2009; 61: 859–866.

16. Zochling J, Braun J. Management and treatment ofankylosing spondylitis. Curr Opin Rheumatol 2005; 17: 418-425.

17. Sayer DC, Cassell HS, Christiansen FT. HLA-B*27 typing bysequence specific amplification without DNA extraction. MolPathol 1999; 52: 300–301.

18. Kutyavin IV, Afonina IA, Mills A, Gorn W, Lukhtanov EA,Belousov ES, et al. 3 -minor groove binder-DNA probesincrease sequence specificity at PCR extensiontemperatures. Nucleic Acids Res 2000; 28: 655-661.

19. Lukhtanov EA, Lokhov SG, Gorn W, Podyminogin MA,Mahoney W. Novel DNA probes with low background andhigh hybridization-triggered fluorescence. Nucleic Acids Res2007; 35: e30.

20. Bon MA, van Oeveren-Dybicz A, van den Bergh FA.Genotyping of HLA-B27 by real-time PCR withouthybridization probes. Clin Chem 2000; 46: 1000-1002.

21. Balas A, Santos S, García-Sánchez F. Complete codingsequence of HLA-B*2712: a serologic B27-negative antigenassociated to Bw6. Tissue Antigens 1998; 51: 394-397.

22. Siala M, Mahfoudh N, Gdoura R, Younes M, Fourati H,Kammoun A, et al. Distribution of HLA-B27 and its alleles inpatients with reactive arthritis and with ankylosing sponylitisin Tunisia. Rheumatol Int 2009; 29: 1193-1196.

23. Ball EJ, Khan MA. HLA-B27 polymorphism. Joint BoneSpine 2001; 68: 378-382.

26. Kwok S, Kellog DE, McKinney N, Spasic D, Goda L,Levenson C, et al. Effects of primer – template mismatcheson the polymerase chain reaction: human immunodeficiencyvirus type 1 model studies. Nucleic Acids Res 1990; 18: 999-1005.

27. Wu JH, Hong PY, Liu WT. Quantitative effects of positionand type of single mismatch on single base primerextension. J Microbiol Methods 2009; 77: 267-275.

© 201 The authors. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are credited.

4. Khan MA, Mathieu A, Sorrentino R, Akkoc N. Thepathogenetic role of HLA-B27 and its subtypes. AutoimmunRev 2007; 6: 183–189.

24. Blanco-Gelaz MA, López-Vázquez A, García-Fernández S,Martínez-Borra J, González S, López-Larrea C. Geneticvariability, molecular evolution, and geographic diversity ofHLA-B27. Hum Immunol 2001; 62: 1042–1050.

25. McLean L, Hammer R, Taurog J, Cytokine profiles and site-directed B27 mutation in the HLA-B27 transgenic rat.(Abstract). Hum Immunol 1996; 47(1-2): 74.

NZ J Med Lab Science 20148