Portal de Periódicos da CAPES

MOLECULAR HLA TYPING-THE BRAVE NEW WORLD

1997; Wolters Kluwer; Volume: 64; Issue: 11 Linguagem: Inglês

10.1097/00007890-199712150-00001

1534-6080

Mike Bunce, Neil T. Young, Ken I. Welsh,

Cytomegalovirus and herpesvirus research

Soon after the first solid organ clinical transplants were performed, it became apparent that some degree of HLA typing was prudent, originally for matching and later as an aid to both cross-matching and antibody screening. Such typing was performed with a panel of antisera specific for different antigenic products of the system in an assay that became known as the complement-dependent cytotoxicity test (CDC*)(1), or serology. CDC is robust and can provide results within 3 hr, critical for reducing cold ischemia times in cadaveric transplantation. The close cooperation between HLA laboratories throughout the world in terms of serum exchange kept the cost of such tests to a minimum while ever-increasing the resolution. There are drawbacks to CDC; viable lymphocytes are required, the antibodies needed are generally nonrenewable, and the technique has limited powers of resolution, particularly at HLA class II. This problem of serological discrimination of class II antigens, arguably the most important antigens for solid organ transplantation, led to the first comprehensive molecular class II typing system using restriction fragment length polymorphism (RFLP)(2). Briefly, RFLP entails the restriction endonuclease digestion of genomic DNA followed by electrophoretic resolution of the endonucleolytic fragments which are denatured in situ and hybridized to a nylon membrane. The membrane is then probed with homologous labeled cDNA or genomic probes which yield hybridization signals characteristic of various HLA alleles. The relevance of RFLP was shown by Opelz et al. in the Collaborative Transplant Study (3) and by Mytilineos et al.(4, 5). These retrospective studies showed that RFLP-defined HLA-DR-matched renal grafts had improved survival compared with serology-defined HLA-DR-matched grafts, and that serology for class II was inadequate in many typing centers, with an overall discrepancy rate of up to 25% between serologically and RFLP-defined antigens. It soon became apparent that RFLP had too many drawbacks to be the method of choice for the 1990s. The main problem was the lengthy time from sampling to results (up to 10 days), but in addition, the resolution often depended on a strong association between neighboring sequences rather than on polymorphism within the DRβ1 domain. This restricted its use on mixed ethnic populations because of differences in linkage disequilibrium values. The development of the polymerase chain reaction (PCR)(6) allowed the evolution of improved molecular HLA-typing techniques. The PCR can generate specific amplified stretches of DNA sequences in vitro through repeated cycles of DNA denaturation, annealing of specific primer to a single strand, and nucleotide extension from primer pairs using a DNA polymerase. The PCR allowed extensive sequencing of HLA alleles from 1990 onward, as shown in Table 1. Transplantation laboratories were, for the first time, able to use PCR to amplify the polymorphic regions of HLA genes which could subsequently be analyzed for the polymorphism within the amplicon, thus establishing the tissue type. The new PCR tissue-typing techniques now allow histocompatibility scientists the choice of whether to use low-, medium-, or high-resolution methods: Low-resolution methods generally only identify broad specificities or groups of specificities. The term “medium resolution” can be used to describe a typing system that discriminates between all serological specificities but may also give some allele-specific results, whereas high-resolution typing is a term generally used to describe a typing system that discriminates between greater than 90% of the alleles in the loci analyzed. Thus, in solid organ transplantation, medium-resolution typing is generally the method of choice, whereas in unrelated bone marrow transplantation, low- or medium-resolution typing is always followed up with high-resolution typing. Obviously, high-resolution methods are not used routinely because they tend to be more time consuming and more costly compared with medium-resolution methods. Techniques for analyzing polymorphisms in amplified DNA can be divided into two basic groups: probe hybridization and direct amplicon analysis. Currently there are many different PCR-based tissue typing assays and it is up to individual laboratories as to which test or combination of tests they use and for which loci they choose to apply the tests. Most new molecular methods of tissue typing were initially developed for class II typing, particularly DQA1 and DQB1 (because of the relative simplicity of these loci compared with DR and class I). However, most molecular methods are now applicable to both class I and class II, the only difference being that in class I analysis, exons 2, 3, and sometimes exon 4 are required for discrimination of most alleles, whereas most class II alleles can be discriminated between by analyzing exon 2 only. A diagram illustrating PCR-sequence-specific oligonucleotide probing(SSOP), reverse PCR-SSOP, PCR-sequence-specific primers (SSP), and PCR-single-stranded conformational polymorphism (SSCP) is shown inFigure 1. PROBE HYBRIDIZATION TECHNIQUES Probe hybridization techniques rely on amplification of a target DNA sequence which is generally immobilized onto a support membrane (known as dot blotting). The initial amplification is normally generic but may be a mosaic of amplifications which when used together amplify all possible alleles of a given locus. The polymorphisms in the immobilized amplified DNA are subsequently detected by using specific single-stranded DNA probes in combination with highly stringent washes to remove non-specifically bound probe (7). In addition to the use of radioactive isotopes, hybridized probes can be detected by a variety of nonradioactive methods, such as horseradish peroxidase (8) and digoxigenin labeling(9). This technique became known as PCR-SSOP (PCR followed by sequence-specific oligonucleotide probing). PCR-SSOP was first applied to histocompatibility testing in HLA-DQA1 by Saiki et al.(7); subsequently, PCR-SSOP was applied to HLA-DRB1(10-14) and HLA-DQB1(15) and a combination of DR, DQB1, DQA1, DPA1, and DPB1(16). The assays were standardized and gained wide acceptance through the 1991 International Histocompatibility Workshop(13). In 1989, nonradioactive methods of PCR-SSOP analysis were introduced (15) which facilitated the spread of the technique to many laboratories. HLA class I PCR-SSOP techniques were slower to develop because complete sequence data for class I alleles was lacking. As more class I alleles were sequenced, it became clear that serology would not be able to discriminate between certain specificities, so researchers turned their attention to developing class I molecular techniques. The first PCR-SSOP class I dot blot techniques were described for differentiating HLA-B35 and HLA-B53 (17) and subtyping HLA-A2 and A28 (18). Complete single-locus typing systems were then developed for HLA-A (19, 20), HLA-B(19-24), and HLA-C(25, 26) such that a combination of these techniques could be used to completely type an individual for class I. PCR-SSOP is a generally accurate method of genotyping(27) and is especially suitable for large numbers of individual samples due to the fact that generic amplifications from many individuals can be hybridized to a single membrane. There is, however, a problem if you wish to obtain complete results on one individual within 24 hr, a problem that has led to the development of the reverse PCR-SSOP assay(28). In the reverse dot blot, the sequence-specific oligonucleotide probes are bound to a solid support membrane via an incorporated poly-(T) tail, which leaves the detection end of the probe free to interact with target DNA. When labeled DNA target is applied to the reverse dot blot membrane, it will only hybridize to those oligonucleotides that are complementary in sequence. Once hybridized, biotinylated products are detected by the addition of a reporter molecule: antibiotin antibody linked to streptavidin-horseradish peroxidase complex, which induces a color change in the substrate tetramethylbenzidine. Reverse dot or line blot methods have been published for class II typing (28, 29) and HLA-A(30). HLA-B and HLA-C reverse SSOP methods are currently being developed by various commercial companies. When combined, these reverse SSOP methods should allow complete results on one individual within 4-5 hr, which may be suitable for genotyping cadaver donors. However, the large number of probes needed for a parallel class I and II reverse SSOP system requires considerable quality assurance input, which means that complete reverse SSOP systems are most likely to be available only from commercial companies and may be too expensive for routine typing of all samples in a laboratory. DIRECT AMPLICON ANALYSIS TECHNIQUES Considerable impetus for rapid HLA typing comes from solid organ transplantation, where HLA matching is generally regarded as being beneficial and long cold ischemia time as damaging. Current PCR-SSOP approaches require lengthy post-PCR steps, which means that even the quickest reverse SSOP technique can not be completed within 5 hr, double that required for serology. There are molecular methods which, while not so efficient for large sample numbers, are more suitable for rapid limited sample number through-put. These include PCR-RFLP, PCR-SSP, nested PCR-SSP, heteroduplex analysis, and other conformational assays. PCR-RFLP As discussed earlier, restriction endonucleases have been used to detect polymorphism with significant success on genomic DNA, but the development of the PCR enabled the enzymes to be used in a much more precise way on much smaller blood samples without the necessity for detection probes. The principle of PCR-RFLP is that a precise region of target DNA is amplified and then cut. Since the sequences of the most common HLA alleles were already known, enzyme restriction sites could be defined within the PCR amplicon at sites of polymorphisms. Thus, a PCR amplicon of unknown HLA type would be digested with a variety of endonucleases, giving rise to PCR fragments of varying size depending on the alleles present. PCR-RFLP has been described for DQA1 (31), DQB1 (32-34), DPB1 (35), DRB1 (36-38), HLA-B44 subtyping (39), and HLA-C(40). PCR-RFLP can rival more established class II genotyping methods for sensitivity and accuracy (41), but complex methodology and interpretation have prevented its widespread use. PCR UTILIZING SSP In 1989, two separate groups (42, 43) described a system of PCR in which the specificity of the PCR reaction entailed matching the 3′ end of one primer with the target DNA sequence, thereby allowing the identification of any point mutation within one or two PCR reactions. Newton (43) described the detection of a single point mutation using one generic sense primer and two antisense primers: one antisense primer was specific for the “normal” form and was refractory to PCR on “mutant” DNA, and the other antisense primer for the “mutant” was refractory to PCR on “normal” DNA. This was termed the amplification refractory mutation system (ARMS). ARMS works because Taq polymerase lacks 3′ to 5′-exonucleolytic proof-reading activity (44, 45). Such an activity would correct the mismatched terminal base of an ARMS primer in a mismatched primer-template complex and subsequently permit efficient priming with the “repaired” primer. For efficient ARMS amplification without false priming, the conditions need to be highly stringent as it is possible for 3′-mismatch extension(42, 43, 46). Stringency is multifactorial, relying on the concentration of all the PCR constituents, such as target DNA,Taq, dNTPs, Tris, and free magnesium. PCR stringency kinetics also relies on individual primer factors, such as primer sequence and length and type of primer-template mismatches. An important feature of ARMS is that each individual reaction contains primers to amplify a so-called“housekeeping” gene which detects possible PCR inhibition and thus acts as a positive control. Without this positive control, it would be impossible to discriminate between a failed PCR reaction and a negative PCR reaction and hence all homozygous results would be questionable. ARMS was predicted to be applicable to HLA analysis in 1989(42) and was first applied in a limited fashion to HLA typing DPB in 1990 by Fugger et al. (47), while Lanchbury et al. (48) utilized the 3′-mismatch concept to obtain DR4 group-specific amplification prior to DR4 SSOP subtyping. The first comprehensive ARMS HLA typing system was described in 1992 by Olerup and Zetterquist (49) for low-resolution HLA-DRB1 typing, including group-specific detection of DRB3 and DRB4 by ARMS using 19 PCR reactions. Olerup and Zetterquist renamed the assay PCR-SSP (PCR using sequence-specific primers). Modern PCR-SSP features multiple PCR reactions where each reaction is specific for an allele, or more commonly a group of alleles that correspond to a serologically defined antigen. To type an individual completely at any given locus, multiple PCR-SSP reactions are set up and subjected to PCR under identical conditions. The presence or absence of PCR amplification is detected in a gel electrophoresis step with visualization by ethidium bromide incorporation. PCR-SSP was soon applied to detect polymorphisms in other class II loci, such as HLA-DQB1 (50, 51), HLA-DQA1(51), and DPB1 (52). Identification of class I alleles by PCR-SSP was first reported by Browning et al.(53), who described low-resolution typing of HLA-A. Thereafter, medium-resolution PCR-SSP systems were rapidly described for HLA-C(54, 55) and HLA-B (56), with low-resolution HLA-B typing systems also being described(57). Recently, a PCR-SSP method known as phototyping was developed that allows the simultaneous detection of HLA-A, -B, -C, -DRB1,-DRB3, -DRB4, -DRB5, and -DQB1 alleles (58). This method has a resolution and accuracy far greater than average serology, and only takes 3 hr to complete, making it suitable for genotyping cadaver donors. The development of automatic dispensing equipment and better electrophoresis equipment has facilitated the use of phototyping in many laboratories. In some laboratories, including our own (since September 1994), it has completely replaced serology. An advantage of PCR-SSP is that it detects polymorphisms linked on an individual chromosome (cis) whereas PCR-SSOP detects polymorphisms on both DNA chromosomes (cis or trans). Thus, PCR-SSP has greater power for discriminating between heterozygosity involving two closely related alleles, as illustrated in Figure 2. In addition, the nucleic acid substitutions present on novel unsequenced allelic variants can be identified by PCR mapping. In this technique, a potential new allele is mapped by using a sense primer that recognizes the allelic variant (normally from the initial molecular typing) in combination with multiple antisense primers in multiple reactions. The positive and negative reactions with these primer mixes can then be used to map out many of the allelic variants polymorphisms. While this approach is certainly no substitute for sequencing, it has been used to identify many new class I(54-56, 58, 59) and class II alleles (60, 61), alleles which have subsequently been proved correct by sequencing. PCR-SSP is typically used as a medium-resolution technique, but recently we described allele-specific one-step SSP for HLA-C (62), while others have applied allele-specific PCR-SSP to DRB1(63) and to HLA-A (64). Allele-specific PCR-SSP is usually achieved by using a combination of group-specific amplifications coupled with some allele-specific reactions. However, while this approach to allele-specific typing is theoretically applicable to all HLA antigens, it would only be possible by using a large array of primer mixes. Therefore, to obtain allele-specific typing, techniques such as nested PCR-SSP and conformational assays have been developed to be used either as individual complete methods or methods used to supplement medium-resolution typing obtained by PCR-SSP or -SSOP. NESTED PCR-SSP Nested PCR-SSP is a two-step approach to HLA typing whereby the region of interest is amplified in the first step and this amplicon is used instead of genomic DNA for the second, sequence-specific amplifications, using primers which are internal to the first pair of amplification primers. Nested PCR-SSP was first described by Bein et al. (65) as a method for complete medium-resolution typing of HLA-DRB1. In Bein's method, exon 2 of all HLA-DRB1 alleles was amplified and subjected to a second-round amplification using 18 PCR-SSP reactions which, like one-step PCR-SSP, used an internal control amplification to prove the success of each individual reaction. The results were similar to one-step PCR-SSP, but it was the latter, simpler, quicker technique that became popular. Nested PCR-SSP does have distinct advantages over conventional PCR-SSP in that the amount of DNA required is very small. In addition, allele-specific nested PCR-SSP is applicable to subtyping highly polymorphic alleles, such as HLA-A*02 (66). For antigen-specific nested PCR-SSP, only the alleles of interest are amplified, so in the case of HLA-A*02 only the A*02 alleles are amplified and these amplicons are then subjected to a second round with internal HLA-A*02 primers. This antigen-specific approach has the advantage of allowing all HLA-A*02 alleles to be detected in all combinations without the interference from closely related alleles (such as HLA-A*68), which would not be amplified in the initial reaction. CONFORMATIONAL ASSAYS: HETERODUPLEX AND SSCP ANALYSIS At the end of any PCR cycle, the individual strands may re-anneal with each other to form homoduplexes, or they may re-anneal with an unrelated DNA strand to form a heteroduplex, or they may remain as single-stranded structures(67). These different forms of PCR product have unique conformational structures which may be differentiated by their electrophoretic mobilities in a temperature or denaturing gradient gel. This PCR phenomenon has been applied to HLA to produce tissue typing methods that theoretically can detect not only existing polymorphisms but new polymorphisms as well. PCR heteroduplex analysis has never gained popularity for identifying HLA polymorphisms due to the complexity of the gel analysis and the technically challenging conditions. However, it has been used to match individuals for HLA-DR and HLA-DP by “DNA cross-matching”(68), whereby donor and recipient DNA are mixed before the final stage of the PCR to allow donor-recipient hetero- and homoduplexes to form, which may indicate whether the pair are indeed HLA identical. The discrimination of heteroduplex analysis has been enhanced by the use of a universal heteroduplex generator (UHG) (69). The UHG is a synthetic DNA strand that is similar to the polymorphic sequence of the target gene but contains substitutions at the polymorphic sites to make the UHG different to all known alleles. When the UHG strand is incorporated into the PCR, the number of informative heteroduplexes, and hence the discriminatory power of the technique, is increased. Heteroduplex analysis has been used for class II typing, but like conventional PCR-SSOP, it cannot achieve separation of single alleles. Discrimination of heterozygotes has been addressed successfully in class I DNA typing by Arguello et al.(70) using a novel method of allele separation called complementary strand analysis (CSA). CSA uses a biotinylated primer in the initial locus-specific amplification that labels one DNA strand. The DNA strands are then chemically dissociated and separated using streptavidin-coated magnetic beads and then allowed to anneal to a previously prepared reference DNA strand. The reference strand is generated from locus-specific amplification of an unrelated sample of known HLA type. The resulting hybridized alleles can then be separated using a two-layer gel system consisting of agarose and polyacrylamide layers. The different hybridized products migrate to different positions in the gel according to the conformational mobility of the different products. The separated allelic strands can then be excised out of the gel layer and dot-blotted onto a membrane and subsequently analyzed by PCR-SSOP methods. The CSA technique has two advantages; first, the alleles are separated and thus heterozygous individuals can be SSOP-typed more easily than with conventional SSOP methods. Second, the gel electrophoresis of the separated alleles can itself be used as a typing system. SINGLE-STRANDED CONFORMATIONAL POLYMORPHISM SSCP analysis depends on the fact that single-strnaded DNA molecules of differing sequences exhibit conformational changes as a result of intrastrand complementary base pairing (71). The different single-stranded products exhibit different mobilities during nondenaturing polyacrylamide gel electrophoresis that can be used to ascertain the genotype of an individual. So far, SSCP has been applied successfully to HLA-A typing(72), HLA-DRB1, DQB1 and DQA1 typing(73-75), DPA1 and DPB1 typing(76), and HLA-DR4 subtyping (77). However, like heteroduplex analysis, the complexity of both the technique and the interpretation prevented the widespread application of this technique. SEQUENCE-BASED TYPING PCR-SSP and PCR-SSOP methods are limited within the context of existing sequence information; novel alleles may be missed or mistyped by both PCR methods. New alleles are often identified by PCR-SSOP and -SSP by unique patterns, but these always require confirmation by sequencing before a new allele is accepted by the nomenclature committee. PCR-SSOP and -SSP can never identify unique polymorphisms, whereas sequencing can. The principles of sequence-based typing (SBT) are that the polymorphic regions of any given allele are amplified by flanking PCR primers. The resulting PCR product is sequenced by any one of a variety of subtly different methods and analyzed by computer to ascertain the type. Computer analysis is required because the sequenced product from a heterozygous individual will contain two superimposed sequences that need to be aligned with all previously known sequences in order to be identified and separated. SBT was initially described for HLA DRB1, DQB1, and DQA1 by Santamaria et al. (78) and for HLA-DPB1 by Rozemuller et al.(79). SBT is admirably suited to class II typing because the vast majority of the functional polymorphism is located in exon 2. This means that a single sequencing reaction is adequate to ascertain the class II type. SBT of class I exons 2, 3, and 4 was first described by Santamaria et al. (80), but it is not widely used because of perceived complexity and occasional inaccuracies (81). Elegant alternative SBT approaches for class I have been developed that concentrate on exons 2 and 3, where the majority of class I polymorphisms are located(82, 83). To alleviate the complexity of SBT of the highly polymorphic HLA-B locus, Petersdorf and Hansen(83) used a combination of group-specific amplification followed by SBT. The main drawbacks of SBT are the equipment costs and the time required to fully sequence one individual. A laboratory with a single automated sequencer would struggle to fully class I- and II-type more than five individuals a day. Offset against this is the tremendous advantage of having high-resolution typing, but the question remains of how relevant high-resolution typing is in transplantation and disease, a question which can not now be answered without some form of extensive high-resolution typing. One cautionary note is that like all molecular methods, sequencing is not infallible and many sequenced alleles have had to be retracted due to errors, most commonly GC inversions. In addition, even recent SBT methods do not always discriminate between all heterozygous combinations of alleles (82). CONCLUSIONS There are many reasons for tissue typing: matching for solid organ and bone marrow transplantation, as an important aid to alloantibody definition, anthropological studies, disease association studies, drug reactions, forensic studies, and facilitation of investigations into T-cell mediated immunity. The tissue typing method most suited to each application is a balance of resolution, sample numbers, time, money, sample material, the method(s) used on necessary control populations, and the expertise of the individuals performing the typing. It is now possible to use molecular methods for all tissue typing applications without recourse to serology. For the majority of applications, a medium-resolution method is the optimum start. It allows the researcher to quickly focus on candidate loci, antigens, or alleles, which can then be further defined by higher-resolution techniques. It allows the transplant surgeon to define antibody specificities or define bone marrow panels and provide a level of discrimination easily suitable for solid organ exchange programs. In most cases, it does not matter which medium-resolution technique is used, so long as it is performed accurately. Allele-specific typing, whether by SBT, SSOP, or SSP, presents the researcher with a powerful analysis tool. In disease studies, a weak association with serological antigens may become clear when individual amino acids are considered. In transplantation, the analysis of immunodominant epitopes present on multiple alleles can only really be investigated by high-resolution typing of all donors and recipients. The application of high-resolution typing requires a different breed of computer analysis. Instead of correlating disease or rejection with a list of antigens, the computer programs of the future will be required to correlate disease or rejection with linear and conformational epitopes(84). It is also expected that computer programs will be required to correlate disease/rejection with peptide presentation in either class I or class II molecules. Ultimately, this may lead to the discovery that certain combinations of class I and class II antigens present within an individual will predispose that individual to disease or rejection given exposure to certain pathogens or transplanted antigens. Generally, molecular typing methods used without serological backup can not identify null alleles or expression variants that may cause problems in certain transplant situations. Most null alleles are extremely rare, although the DR53 null allele (85) is common in most populations. It is important that null alleles are detected, especially in unrelated bone marrow transplantation. The phototyping PCR-SSP method(58) incorporates a PCR-SSP reaction to identify the DR53 null allele (86), and it is theoretically possible to develop PCR-SSP reactions or PCR-SSO probes to detect the rarer null alleles, such as the HLA-A2 null allele A*0215N (87), as and when these null alleles are sequenced. In time, greater sophistication will be required so that not only will null alleles be identified by molecular techniques, but low-and high-expression variants will be identified as well. The best current general-purpose method for transplantation, typing all HLA loci to a medium resolution, is arguably the PCR-SSP method. It has the advantage of flexible resolution coupled with ease of interpretation and speed of result, but has the disadvantage that multiple reactions are needed that require time-consuming dispensing of reagents and lots of gel electrophoresis to detect PCR products. However, it is now possible to dispense PCR mixtures and primer mixes entirely with multichannel automatic dispensing pipettes or dedicated dispensing machines, and gel electrophoresis equipment has been manufactured to allow direct loading from 96-well formats. This means that dispensing and detection of multiple reactions is now as easy as or easier than dispensing multiple serological reactions. Indeed, it is now possible to completely automate PCR-SSP using Biomek workstations (Biomek 2000, Beckman Instruments, Fullerton, CA) (Patrick Merel, ETSA, Bordeaux, France, personal communication). Eventually, the electrophoresis step will be eradicated completely as enzyme-linked immunosorbent assay-based detection systems become commonplace (88, 89).Figure 1: Basic principles of four commonly used typing methods: PCR-SSOP, reverse PCR-SSOP, PCR-SSP, and PCR-SSCP. The amplification primers in each case are shaded on the representative HLA-A*0101 and A*0201 nucleotide sequences. The polymorphisms are underlined.Figure 2: Difference between PCR-SSOP and PCR-SSP techniques in distinguishing between heterozygotes. PCR-SSOP methods cannot link two polymorphisms on an allele, whereas PCR-SSP can. This can cause problems in discriminating between some heterozygote combinations.