ADV/15/0002/EUj January
2015
Paris Translational Research Center
for Organ Transplantation
Complement Activating anti-HLA DSA and Solid Organ
Transplant Survival
Olivier Aubert (MD, PhD)
INSERM U970, Equipe 12
Hôpital Necker, Service de Transplantation Rénale
www.paristransplantgroup.org
THE PREVALENT ORGAN TRANSPLANT UNIVERSE
1,000,000(700,000 kidneys)
80,000 new kidney transplants per year
About 15% have DSA: 150,000
About half have ABMR: 75,000
NATUREL HISTORY OF DSA MEDIATED KIDNEY ALLOGRAFT
INJURY
Loupy, Hill & Jordan, Nat reviews 2012
NATUREL HISTORY OF DSA MEDIATED KIDNEY ALLOGRAFT
INJURY
Loupy, Hill & Jordan, Nat reviews 2012
NATUREL HISTORY OF DSA MEDIATED KIDNEY ALLOGRAFT
INJURY
Loupy, Hill & Jordan, Nat reviews 2012
NATUREL HISTORY OF DSA MEDIATED KIDNEY ALLOGRAFT
INJURY
Loupy, Hill & Jordan, Nat reviews 2012
NATUREL HISTORY OF DSA MEDIATED KIDNEY ALLOGRAFT
INJURY
Loupy, Hill & Jordan, Nat reviews 2012
NATUREL HISTORY OF DSA MEDIATED KIDNEY ALLOGRAFT
INJURY
Loupy, Hill & Jordan, Nat reviews 2012
ABMR IS THE MAIN CAUSE OF LATE ALLOGRAFT LOSS
Loheac C & Aubert O et al. Unpublished data
WHAT NEED TO BE ACHIEVED
Treatment
Risk stratification
Diagnosis – Activity – Stage
Washington, Sept 28th 2015
✓ ABMR = current outstanding matter of
concern
✓ Reponse to therapy in ABMR is unknown
✓ PHENOTYPES ARE MANDATORY
✓ Precision composite end point is needed
Lefaucheur & Loupy, JASN, 2010
Lefaucheur, AJT, 2008
Presence/absence of anti-HLA antibody is not enough
Thresholds of HLA-DSA
ARE DSA EQUAL?
Miss C.3rd graft
DSA DQ2
MFI 2460
ARE DSA EQUAL?
Mr D1st graft
DSA DR 7
MFI 3800
WHAT NEED TO BE ACHIEVEDPresence/absence of anti-HLA Ab:Not enough
Thresholds of HLA-DSA:Not enough
➢ Affinity/Avidity of Anti-HLA antibodies➢ Complement activation: C1q➢ Subclasses: IgG1-4➢ Pre-existing vs. de novo DSA
We need new tools for risk stratification !
Graft failures
Disease progression
Risk prediction
Validation
• Tyan et al; Human Immunol 2011
• Sicard et al; JASN 2014
• Taupin et al; JASN 2015
• Viglietti et al: JASN 2015
• Akalin et al; KI 2015
• Canër et al; Transplantation 2015
• Aubert O et al, JASN 2017
Integrative and multiplex assessment of DSA
C1q
binding
Preformed
De novo
Affinity
Strengh
MFI
levels
Class
Eplets
IgG
Subclass
Anti-HLA DSA characteristics associated with
risk of allograft loss
• Complement binding capacity
• IgG subclass composition
Meta-Analysis: Clinical Significance of Complement Activating Anti-HLA DSA in
Kidney TransplantationSc
ree
nin
gEl
igib
ility
Incl
ud
ed
Records identified through databases (Mayo clinic research) (n=5,513)
Full-text articles assessed for eligibilityperformed by 2 blinded reviewers (n=62)
(n patients=10,510)
Articles selected on abstracts: duplicates were removed, no related abstracts
performed by 2 blinded reviewers (n=286)
Exclusion criteria- Methodology articles - No data on primary or
secondary outcomes
Studies included in quantitative synthesis(n=36, 7,547 patients)
Kappa 0.9941
Inclusion criteria-NOS ≥ 6-HR available
Iden
tifi
cati
on
Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009)
Loheac C, Aubert O et al PLOS MEDICINE IN PRESS
Meta-Analysis: Clinical Significance of Complement Activating Anti-HLA DSA in
Kidney Transplantation
3.05
C’ activating Ab: multiorgan relevance
C’ activating Ab: Timing of Ab detection
C’ activating Ab: type of test
Factors influencing complement fixing
(CF) HLA-Ab in vitro (C1q reactivity)
1. Presence of complement fixing (CF) IgG subtypes
(IgG1/IgG3)
2. Level of IgG subtypes (weak /strong MFI)
3. Mixture of CF and non CF (NCF) (C1q reactivity)
4. Impact of antibody removal therapy:
Loss of C1q reactivity – diminished IgG subtype reactivity-
NOT Switch of IgG subtype
DSA IgG subclass distribution
0
0
54%
1411%
1310%
97%
0
0
3125%
86% 0 17
14%
22%
32%
22%
IgG1N=94
75%
IgG4N=33
26%
IgG2N=55
44%
IgG3N=35
28%
IgG1 N=94 75%
IgG2 N=55 44%
IgG3 N=35 28%
IgG4 N=33 26%
DSA IgG SUBCLASS COMPOSITION ACCORDING TO C1q STATUS
IgG subclass compositionAll patients C1q- DSA C1q+ DSA
N=157 N=113 N=44
None 24 (15) 24 (21) 0
1 28 (18) 21 (19) 7 (16)
2 6 (4) 6 (5) 0
3 6 (4) 4 (4) 2 (5)
4 5 (3) 5 (4) 0
1+2 24 (15) 21 (19) 3 (7)
1+3 28 (18) 13 (12) 15 (34)
1+4 5 (3) 3 (3) 2 (5)
2+4 4 (3) 4 (4) 0
1+2+3 8 (5) 2 (2) 6 (14)
1+2+4 17 (11) 10 (9) 7 (16)
1+2+3+4 2 (1) 0 2 (5)
All C1q-binding DSAs were positive for IgG1 and/or IgG3 subtypes
C1q negative DSA does not indicate absence of IgG1 or IgG3
1) IgG1-4, C1q and pan-IgG MFI segregate presence/absence of ABMR
2) IgG3 and IgG4 segragate ABMR phenotype
Identification of distinct patterns of injury according
to DSA characteristics
Lefaucheur C, JASN, 2015
DSA: N=157
C1q DSA: N=44
OR 95%CI P
MFI level 1 [1.00-1.00] <0.001
IgG1
No 1
Yes 5.59 [0.90-34.60] 0.064
IgG3
No 1
Yes 3.66 [1.40-9.54] 0.008
IMMUNOLOGIC DETERMINANTS OF C1q POSITIVITY: MULTIVARIATE
MODEL
Univariate analysis considered: MFI level, HLA class, IgG1, IgG2, IgG3 and IgG4 subclasses
Systematic monitoring and characterization of DSA
could add to the predictive value for allograft loss of
the conventional approach based on their detection
and strength
STUDY DESIGN
TX 01/2016
Clinical indication 1 & 2 years
screening
Graft loss
2008
- 2010
Day-0 DSA• HLA class, specificity, MFI
• C1q-binding
• IgG1-4 subclasses
Graft Biopsy
GFR
Prot U
Prospective DSA monitoring strategy
Post-Tx DSA• HLA class, specificity, MFI
• C1q-binding
• IgG1-4 subclasses
DYNAMIC MODELING TO ASSESS IMPROVEMENT IN RISK PREDICTION ACCORDING
TO DSA MONITORING AND CHARACTERIZATION
Donor
characteristics
Model 1
Reference
Model
Recipient
characteristics
Transplant
characteristics
Immunological
Anti-HLA DSA
characteristics
Model 2
= Reference Model
+
DSA detection
Post transplant
Kid
ney a
llogra
ft s
urv
ivalModel 3
= Model 2
+
DSA characteristic 1
Post transplant
Model 4
= Model 2
+
DSA characteristic 2
Post transplant
Day 0 Year 2Year 1
Post-TX prospective anti-HLA DSA monitoring strategy
IgG1
MFI
IgG4
IgG2
C1q-binding
IgG3
Relative Variable Importance
0.0 0.2 0.4 0.6 0.8 1.0
HLA class
01000
2000
3000
4000
5000
Number of trees
Out-
of-
bag
err
or
rate
Preformed
0.15
0.20
0.25
IgG1
MFI
IgG2
IgG4
C1q-binding
IgG3
Relative Variable Importance
0.0 0.2 0.4 0.6 0.8 1.0
HLA class 01000
2000
3000
4000
5000
0.25
Number of trees
0.16
0.18
0.20
0.22
Out-
of-
ba
g e
rro
r ra
te
DSA CHARACTERISTICS AND ABILITY TO CLASSIFY ALLOGRAFT LOSS
Day-0 (N=110) Post-Tx (N=186)
Hierarchical ranking by multivariate survival random forest modeling
Viglietti D, JASN, 2016
0.5
0.6
0.7
0.8
1.0
0.0
0.9
Pre-transplant Post-transplant
Day-0
Reference
Model
Day-0
Reference
Model
+
C1q
Day-0
Reference
Model
+
IgG3
Post-Tx
DSA
Model
Post-Tx
DSA
Model
+
C1q
Post-Tx
DSA
Model
+
IgG3
*
*
* *
*
C-s
tati
sti
c
* 95%CI for mean difference in C-statistics
between two models does not include zero
PREDICTIVE VALUE OF DSA MONITORING AND CHARACTERIZATION
Overall population
N=851
0.00
0.25
0.50
0.75
1.00
Pts without graft loss (N=89) Pts with graft loss (N=21)
Day-0
MFI
Day-0
MFI
+
IgG3
Day-0
MFI
Day-0
MFI
+
IgG3
Pts without graft loss (N=89) Pts with graft loss (N=21)
Day-0
MFI
Day-0
MFI
+
C1q
Day-0
MFI
Day-0
MFI
+
C1q
Pre
dic
ted
pro
ba
bil
ity o
f g
raft
lo
ss
0.00
0.25
0.50
0.75
1.00
Pre
dic
ted
pro
bab
ilit
y o
f g
raft
lo
ss
80%
67%
RISK RECLASSIFICATION BY IgG3 AND C1q BEYOND MFI
NRI=1.326, P<0.001 NRI=0.948, P<0.001
In patients with Day-0 DSA
84%
81%
RISK RECLASSIFICATION BY IgG3 AND C1q BEYOND MFI
Pts without graft loss (N=149) Pts with graft loss (N=37)
0.00
0.25
0.50
0.75
1.00
Pre
dic
ted
pro
ba
bil
ity o
f g
raft
lo
ss
Post-Tx
MFI
Post-Tx
MFI
+
IgG3
Post-Tx
MFI
Post-Tx
MFI
+
IgG3
0.00
0.25
0.50
0.75
1.00
Pre
dic
ted
pro
ba
bilit
y o
f g
raft
lo
ss
Post-Tx
MFI
Post-Tx
MFI
+
C1q
Post-Tx
MFI
Post-Tx
MFI
+
C1q
Pts without graft loss (N=149) Pts with graft loss (N=37)
62%
NRI=1.326, P<0.001 NRI=0.948, P<0.001
In patients with Post-Tx DSA
85%91%
76%
CONCLUSION
• Prospective, systematic monitoring of DSA improved risk stratification for allograft loss
beyond traditional determinants
• IgG3 positivity and C1q-binding capacity were the most informative DSA characteristics
for classifying patients according to their risk of allograft loss
• IgG3 positivity or C1q-binding capacity improved risk stratification at the population level
and also in patients with DSA beyond MFI level
• Further studies are needed to determine the most cost efficient DSA screening policies
Anti-HLA DSA characteristics induce distinct injuries in
kidney allografts
• Complement binding capacity
• IgG subclass composition
INTRAGRAFT GENE EXPRESSION ACCORDING TO
DSA C1q STATUS
10-310-210-11000.0
0.5
1.0
1.5
2.0
2.5
Association with C1q binding DSA (adjusted p value)
Fo
ld c
han
ge
(C1q
+ v
s C
1q-)
NK genes
NK genes with CD16 engagement
Interferon gamma genes
Macrophages genes
Endothelial genes
CXCL11
CXCL11
CXCL11FCGR3A
LILRB2
CD163
FCGR3A/3B
MS4A6A
MS4A7
C1QB
PTPRC
FYB
GBP5CCL4
CCL4
C1QC
FCGR1A
CD84
C1QA
CTLA4C3AR1
CD86
Effector T cell
CD72
MS4A6AMS4A7
9954 IQR-filtered genes
0 5 10 15 200
10
20
30
40
50
60
70
80
90
100
Rank of Feature
Re
lati
ve
Im
po
rta
nc
e f
or
C1
q+
DS
A (
%)
CXCL11
CCL4
MS4A6A
CTLA4CTSS
CYBBC1QB
MS4A7
GBP5CD86
FCGR3A
GBP1
LST1
ITKC1QAKLRC1
CRTAMPTPRCFYBHISTOLOGY
RELATIVE IMPORTANCE OF C1q SELECTIVE GENES AND HISTOLOGY
IN DETERMINING DSA C1q STATUS
Hierarchical ranking (random forest)
Histology: g+ptc+v+i+t+C4d Banff scores
Top genes Biological association
CXCL11 ENDOTHELIAL IFNG RESPONSE
CCL4
NK CELL CD16-
ENGAGEMENT/MACROPHAGE
IFNG RESPONSE
MS4A6A MONOCYTE/MACROPHAGE
MS4A7 MONOCYTE/MACROPHAGE
FCGR3A NK CELL
BIOLOGICAL RELEVANCE OF C1q DSA TRANSCRIPTS
Human rejection effector cells culture
CCL4
CXCL11
FCGR3A
MS4A6AMS4A7
Anti-HLA DSA characteristics & gene expression to
identify responders to Eculizumab therapy
THERAPEUTIC STUDY: EFFECT OF COMPLEMENT INHIBITION
TX Complement C5 INH (N=52)
2011
- 2014
Allograft BiopsyHistology
Gene expression
Multi-center, international study in HLA incompatible kidney recipients
11 centers in the US and Europe: NCT01567085 & NCT01399593
DSAC1q status
FCM+
Standard of care (N=64)
Day 0 Day 14 Day 90
Incidence of
biopsy-proven
AMR
Standard of care: PE and IVIG according to local centers’ protocol
C5 INH: Eculizumab 1200 mg at Tx, 900 mg/week x4 and 1200 mg at week 5, 7, 9
gptc v i t
cgC4d
0.0
0.5
1.0
1.5
2.0
2.5
NS
NS
* *
NS
NS
* *
NSNS
NS
NS
SOC/C1q+ (N=32)
Eculizumab/C1q+ (N=37)
SOC/C1q- (N=32)
Eculizumab/C1q- (N=15)
* P<0.01
Ba
nff
sc
ore
EFFECT OF COMPLEMENT INHIBITION ON HISTOLOGY
Eculizumab specifically decreased acute injury in C1q+DSA patients (Day 14
allograft biopsy)
C1q+ DSAC1q- DSA
CXC
L11
CCL4
MS4A
6A
MS4A
7
FCG
R3A
0
5
10 *
NS
* *
*
*
NS
NS
NSNS
SOC/C1q+ (N=32)
Eculizumab/C1q+ (N=37)
SOC/C1q- (N=32)
Eculizumab/C1q- (N=15)
* P<0.001O
pti
ca
l d
en
sit
y (
log
2)
EFFECT OF COMPLEMENT INHIBITION ON THE C1q DSA GENE SET
Eculizumab specifically decreased the C1q gene set expression
C1q+ DSAC1q- DSA
IMPACT OF A THERAPEUTIC STRATEGY BASED ON DSA C1q STATUS
vs. DSA DETECTION
Response rate to complement inhibition improved when characterizing DSA C1q
status at transplantation
CONCLUSION
C’-binding HLA DSA as a biomarker in the era of new technologies
Prognosis
Risk stratification
Enrichment of trials
Diagnosis
Rejection heterogeneity
Precision diagnosis
Gene expression
- NEJM 2013
- Meta-analysis (Plos Med
in press)
Therapy
efficacy
Response to complement INH
Personalized medicine
JASN 2018
CONCLUSION
Lee S & Aubert O et al. JASN in press
CONCLUSION
Kaplan-Meier curves of kidney allograft survival by donor type Kaplan-Meier curves of kidney allograft survival by donor type and the presence of DSA
Aubert O et al. BMJ 2015
88%
80%
90%
85%
73%
44%
ECD/DSA-
ECD/DSA+
paristransplantgroup.org
@ParisTxGroup
Acknowledgments
