Predicting the immunogenicity of structural HLA class I epitopes Reyna Goodman

1
Predicting the immunogenicity of structural HLA
class I epitopesReyna Goodman
2
Introduction

• Between 5-10 of patients waiting for a renal
  transplant are classified as highly sensitised
  (Panel Reactive Activity 85 IgG)
• Highly sensitised renal dialysis patients wait
  longer for a suitable crossmatch negative donor
  compared to non-sensitised patients
• Identification of acceptable HLA mismatches
  increases the likelihood of transplantation

3
HLA specific antibody screening using single
antigen beads

• Beads are coated with single HLA specificities
• Each bead population has a different ratio of two
  dyes which allows identification of up to 100
  individual HLA specificities
• Analysed using a Luminex platform

4
HLAMatchmaker

• A computer algorithm that determines HLA
  compatibility at a structural level by comparing
  differences between polymorphic amino acid
  triplets in the antibody accessible region of the
  HLA molecule
• HLAMatchmaker performs inter-locus and
  intra-locus comparisons between the patients HLA
  type and each mismatched HLA specificity and
  calculates the number of amino acid triplet
  mismatches

5
Aim

• To determine whether HLAMatchmaker could be used
  to predict the immunogenicity of structural
  epitopes
• comparing acceptable HLA mismatches identified
  using single antigen HLA specific antibody
  detection beads with those predicted using the
  HLAMatchmaker computer algorithm
• To identify acceptable HLA mismatches for highly
  sensitised patients

6
Patient cohort
Of 406 patients on the Addenbrookes Hospital
renal transplant waiting list, 24 were identified
as highly sensitised and selected for study
7
Data analysis

• Single antigen beads detect antibody to 65
  individual HLA-A and -B specificities of which 64
  are represented in the HLAMatchmaker algorithm
• Patient HLA class I types and each mismatched HLA
  specificity represented on the single antigen
  beads were entered into the HLAMatchmaker program
  to determine the number of triplet mismatches
• Logistic regression analysis was used to
  determine the relationship between the detection
  of antibody using single antigen beads and the
  number of amino acid triplet mismatches for each
  HLA specificity determined by HLAMatchmaker

8
HLA specific antibody screening using single
antigen beads

• The serum sample with the highest PRA was
  selected for study
• Of 1,451 mismatched HLA specificities, 972 (67)
  were antibody positive
• A mean of 19 (range 1 - 41) acceptable mismatches
  were identified for each patient
• For 19 patients there were 10 or more acceptable
  mismatches identified including some common
  antigens

9
Presence of antibody for each mismatched HLA
specificity categorised by triplet mismatch
10
Logistic regression analysis of antibody binding
to single HLA specificities stratified by number
of triplet mismatches
antibody positive
40
20
0
Plt0.001
Number of amino acid triplet mismatches
11
Individual logistic regression analyses of
antibody binding to single HLA specificities
stratified by number of triplet mismatches
Proportion antibody positive
Number of triplet mismatches
12
Conclusions

• HLAMatchmaker is an effective tool for predicting
  the immunogenicity of a particular HLA mismatch
• The number and nature of triplet mismatches for a
  given HLA type correlates with the risk of
  humoral sensitisation
• The presence of a single triplet amino acid
  mismatch is often sufficient to invoke a strong
  antibody response
• In combination with single antigen beads,
  acceptable mismatches were identified for highly
  sensitised patients, increasing access to the
  donor pool
• Goodman et al Transplantation 2006811331-1336

13
Triplets versus eplets

• Examine the extent to which the structural
  information provided by triplet and eplet epitope
  analysis of HLA compatibility enables the
  prediction of
• Acceptable mismatches
• Magnitude of the antibody response

14
Data analysis

• Sera obtained from 34 highly sensitised patients
  were screened using single antigen beads, to
  determine the presence and magnitude of HLA
  alloantibody
• Patient HLA class I types and each mismatched HLA
  specificity represented on the single antigen
  beads were entered into the HLAMatchmaker program
  to determine the number of triplet and eplet
  mismatches
• A total of 85 sera were screened (median 2 sera
  per patient, range 1 to 6) and 2,088 mismatched
  combinations examined
• The qualitative and quantitative relationship
  between alloantibody levels to each HLA
  specificity and the number of triplet and eplet
  mismatches was determined

15
Number of triplet eplet mismatches present
within mismatched HLA specificities
Plt0.001
16
Relationship between triplet eplet mismatches
and alloantibody production
17
Relationship between triplet eplet mismatches
and alloantibody levels
Pgt0.001
Pgt0.001
18
Conclusions

• The number of triplet and eplet mismatches
  between an alloantigen and the recipient HLA type
  correlates closely with both the development and
  strength of an alloantibody response
• Eplets offer additional epitope discrimination
  but do not improve HLA immunogenicity prediction
• Self triplets and eplets may form immunogenic
  epitopes when expressed in a different
  conformation on mismatched HLA alloantigens

19
Acknowledgements
HI department Cambridge University Hospitals NHS
Foundation Trust, Addenbrookes Hospital Dr
Craig Taylor Miss Cheryl ORourke Mr Tim Key
Department of Surgery, University of Cambridge,
Addenbrookes Hospital Prof J Andrew Bradley Mr
Vasilis Kosmoliaptsis
Centre for Applied medical statistics, University
of Cambridge Mr Andrew Lynch Dr Linda Sharples
20
Patient follow-up

• 5/24 patients transplanted