The dissimilarity score concept focuses on the amino acid variations of
the highly polymorphic antigen-presenting α1 and α2 domains in HLA class I,
and of the ß1 domain in HLA class II. Residues polymorphic
between donor and recipient are evaluated, and assigned either to regions of major
(peptide-binding groove and/or region contacting the T-cell receptor) or minor
allogeneic potential (remaining amino acid residues) according to HLA-A2
(Garboczi et al.; Saper et al.) and HLA-DR1 (Brown et al.; Hennecke et al.; Stern et al.).
Furthermore, the dissimilarity between the
single pairs of exchanged amino acids is measured by the amino acid distance
matrix as proposed by Risler et al.
The basic idea of Risler's score was, that
two distinct amino acids are more similar, the more often they substitute
each other in functionally related proteins. Accordingly, pairs of
amino acids with low substition rates (thus representing functional dissimilarity),
yield a higher score, with the maximum value of 100. HistoCheck encorporates Rislter's scores in its
dissimilariy score calculation by summing the Rislter scores on each amino
acid mismatch in the key domains (in HistoCheck, Rislter scores are divided by 100, putting them in the range
0 to 1). When the mismatch is
located at a position with major allogeneic potential, the value 1 is added in
order to gain a weighted score for this position. Residues that contribute to
both peptide binding and contact with the T-cell recptor (TCR) are rated as if
they possess only one function. The described algorithm results in higher scores
even for low mismatch at functionally relevant positions compared to any mismatch
at positions that are probably of functionaly minor importance. A mathematical description of the
dissimilarity score is shown below:
where i ... n1 are mismatching amino acids without
a critical funtion, with Ri being the corresponding Risler score.
Similarly, j ... n2 are mismatching amino
acids which are important for either peptide binding or TCR contact, with Rj being
the corresponding Risler score.
This program may only be used for research purposes. Its algorithm will
be modified according to new perceptions. HistoCheck provides information
about the possible magnitude of allogenicity. It is based on theoretical
data, which the clinical outcome might not correspond to. The software is
still at a developmental stage, and there may be errors that will
interfere with the calculation of results. Any person who uses HistoCheck
accepts full responsibility for using and applying the information
computed by HistoCheck in any clinical setting. The authors deny
responsibility for any harm that may occur related to to the usage of
HistoCheck evaluates amino acid differences between pairs of HLA
alleles, which are relevant for stem cell transplantation (including alleles of different molecular groups).
Currently, HLA class I (HLA-A,B,C) and class II (DRB1, DQB1, DPB1) are available for this comparison. HistoCheck indicates the positions of exchanges, and checks whether these positions can
be assigned to the peptide binding site or to the region likely to have contact
with the T-cell receptor.
Newly described HLA sequences are regularly integrated into the database according to the nomenclature
for factors of the HLA system updates.
Due to the strong structural homology between HLA class I molecules, the assignment of amino acid residues to functional
regions also of HLA-B and C is based on HLA-A2 (Madden et al.). The assignment of positions is given in Table 1.
The same constellation applies to HLA class II, where the the assignment of the amino acid residues of the DQB1 encoded
DQ ß chains refers to the the ß chain of the HLA-DR1 molecule [Brown et al; Hennecke et al.; Stern et al.]. The assignment
of positions is given in Table 2.
Functional dissimilarity of exchanged residues is shown
according to the score proposed by Risler (See Table 3).
These data are used to calculate an
overall dissimilarity score (See Algorithm).
A low score indicates a low probable allogenicity between HLA variants.
In addition to these calculations
HistoCheck visualizes the HLA amino acid differences between donor
and recipient in a graph which is shown by the integrated RasMol software (Sayle et al.).
is based upon either the PDB file for HLA-A*0201 and as reported by Madden, et al.
(See PDB File Ref. #1) or the PDB file for DRB1*0101 as reported by Stern
et al. (See PDB File Ref. #2).
After the query, the RSM files containing
the highlighted differences can be downloaded for further studies.
The transplantation of hematopoietic stem cells is a potentially curative therapy for a
variety of hematologic and non-hematologic diseases. HLA matching of donor and recipient
is essential in order to reduce the risk of severe acute graft-versus-host disease (GVHD).
Despite a continuously growing number of potential stem cell donors and cord blood units
it may be impossible to find an HLA-compatible donor for carriers of rare antigens or
unusual haplotypes. Furthermore, the advance of HLA class I typing at the allelic level
will probably reveal far more subtype mismatches presently not detected.
When there is no genotypically identical sibling and there are several alternative potential
donors that all have a mismatch at an HLA class I or II locus, the allogenicity of mismatches may
be estimated using the Sequence Similarity Matching concept described by our working group.
In this concept the amino acid differences between HLA alleles are evaluated and rated with
regard to position within the molecule (peptide binding, contact with the T-cell receptor)
and with regard to functional similarity of amino acids within proteins.
This procedure led to a dissimilarity score (allogenicity index)
whereby high values represent high dissimilarity. When there are several mismatched donors,
dissimilarity scores may be calculated for any of them, and the donor with the least may be preferred.
The HistoCheck HLA sequence interpreter is the transformation of the dissimilarity score
concept into a an internet-based computer program, which is intended to provide
an easy-to-administer tool for comparing the allogenicity of different HLA class
I and class II mismatches. Although the HLA match is only one player in the donor selection algorithm,
the dissimilarity score may help to weight the immunogenetic variable in the selection procedure.
Studies focusing on the systematic analysis of a reduced number of mismatches may be
helpful in order to further improve the theoretical model proposed here. However,
as long as such studies are not available, the present knowledge on the structure
and function of MHC molecules, as in part reflected by the dissimilarity score,
may be utilized in order to bring
HLA matching a step away from being a matter of chance.