Lymphoid and myeloid biomarkers for clinical outcome of ipilimumab and Prostate GVAX treatment
Tanja D. de Gruijl PhD
Dept Medical OncologyVU University medical centerAmsterdam, The Netherlands
2
Tanja de Gruijl-disclosures
Research support from Cell Genesys Inc.
SITC 6 Nov 2011
• Most common malignancy in elderly men• Second leading cause of cancer deaths in western countries• 1 in 6 men affected(www.cancer.org)
Prostate cancer
Clinicallylocalized
Relapsedand
Metastatic
HormoneRefractory
AsymptomaticRising PSA
HormoneRefractory
symptomatic
Local treatment Endocrine Investigational Chemo-
therapy
Progression of disease
Clinicallylocalized
Relapsedand
Metastatic
HormoneRefractory
AsymptomaticRising PSA
HormoneRefractory
symptomatic
Local treatment Endocrine Investigational Chemo-
therapy
Progression of disease
immunotherapy
Drake Nat Rev Immnol 2010
Prostate GVAXVaccine consisting of twoAAV-GM-CSF transduced, irradiated prostate cancercell lines (LNCaP, PC-3)
CELL GENESYSCELL GENESYS
Ipilimumab (Yervoy)*anti-human CTLA-4 Antibody*high affinity and specificity*fully human IgG1k antibody *blocks the binding of CTLA-4 to B7*does not mediate ADCC
Treatment and sampling scheme
* * *
* blood collection forimmunomonitoring
GVAX every 2 weeks for a total of 13 i.d. doses
anti-CTLA4 mAb (Ipilimumab) every 4 weeks for a total of 6 infusions
* ** * *w0v1
w4v3
w8v5
w16v9
w20v11
w24v13
w12v7
fu1
Dose level patient #Cohort 1 1-3 Cohort 2 4-6Cohort 3 7-9Cohort 4 10-12
Cohort 5 13-28
anti-CTLA4 dose0.3 mg/kg1.0 mg/kg3.0 mg/kg5.0 mg/kg
3.0 mg/kg
0
50
100
150
200
250
0
30
60
90
120
150
0
10
20
30
40
50
Clinical results
Partial Response (PR); >50% PSA decline
Stable Disease (SD); No PR or PD
Progressive Disease (PD); >25% PSA increase
Ser
um P
SA (n
g/m
l)
11 / 28
12 / 28
5 / 28
Number of patients
n.a.
85 (82-190)
305 (51-919)
Duration of response (median and range in days)
PSA Progressive Disease (PD)
PSA Stable Disease (SD)
PSA Partial Response (PR)
Category:
>25% on-study PSA increase
No PR or SD
>50% on-study PSA decline
Response
Clinical results
• PSA declines were durable: 6 to 31 months
• Stable disease by bone scan was observed in 11 patients (>5 mns)
• Regressing bone and lymph node metastasis were observed in 2 patients
15-9-2005 29-3-2006
Toxicity: Auto-immune Breakthrough Events (irAE) in 9 patients•7 patients (5/5 PR!) showed hypophysitis with:
- secondary adrenal insufficiencies
- secondary hypothyroidism
• 1 PR patient developed a dose limiting grade 3 alveolitis (5 mg/kg Ipilimumab)
• 2 patients experienced low grade colitis; 1 patient grade 3 hepatitis
• irAE were successfully treated with standard hormone replacement
therapy (endocrinopathies) or steroids.
Clinical results:
Treatment response correlated withsurvival
0 10 20 30 40 50 60 700
20
40
60
80
100
months
Perc
ent s
urvi
val
PR + SDPD
p=0.0034
N=17; med.surv. 41 mths;
N=11; med.surv. 21 mths;
0 10 20 30 40 50 60 700
20
40
60
80
100
months
Perc
ent s
urvi
val
PR + SDPD
p=0.0034
N=17; med.surv. 41 mths;
N=11; med.surv. 21 mths;
0 10 20 30 40 50 60 700
25
50
75
100 actual survivalpredicted survival
monthsPe
rcen
t sur
viva
l
p=0.0099
Med.surv. 31.8 mths; 20 dead, 8 alive
Med.pred.surv. 19 mths
Actual survival was longer thanHalabi-predicted survival
Prostate Cancer as a learning modelCan we identify immune parameters that correlate with clinical activity and may
be useful for clinical response prediction?
Immunomonitoring: principal question
…or treatment resistance prediction? >>avoid autoimmune side effects
NB: Phase I study with non-randomized Phase II study: hypothesis generating.
Further validation of identified immune biomarkers in randomized trials withGVAX and/or ipilimumab required!
Prostate Cancer as a learning modelCan we identify lymphoid and myeloid immune parameters that correlate with
clinical activity and may be useful for clinical response prediction?
Immunomonitoring: principal question
…or treatment resistance prediction? >>avoid autoimmune side effects
1. Serology- tumor-specific antibodies
2. Peripheral blood Teff-/Treg cells- frequency- activation status- effector/memory phenotype
3. T cell Functionality- TAA-specific reactivity- suppression assays- cytokine profiles
4. Peripheral Blood DC (PBDC) and Myeloid Derived Suppressor Cells (MDSC) - frequency- activation status
T cell activation: ICOS, FoxP3, CTLA-4, PD-1
ICOS
w0v1 w8v5 w16v9 w24v13 fu0
25
50
75
100
****
**
*
0 10 20 30 40 50 60 700
25
50
75
100
months
N=11; med.surv. 37 mths
N=7; med.surv. 34 mths
n.s.Pe
rcen
t sur
viva
lCut-off: 2-fold
ICOShi sustained
ICOSlo
0
10
20
30
40
***
w0v1 w8v5 w16v9 w24v13 fu
% C
D4+
/CD
25in
t/Fox
P3+
0 10 20 30 40 50 60 700
25
50
75
100
months
p = 0.030
N=13; med.surv. 41 mths
N=11; med.surv. 21 mths
CD4/CD25int/FoxP3
Cut-off: 50%
Perc
ent s
urvi
val
% C
D4+
/ICO
S+
CD4/ICOS
FoxP3increase
no increase
0
10
20
30
40
**
***
PD-1
% C
D4+
/PD
-1+
w0v1 w8v5 w16v9 w24v13 fu
% C
D4+
/CTL
A-4+
0
20
40
60
80
****
**
w0v1 w8v5 w16v9 w24v13 fu
0 10 20 30 40 50 60 700
25
50
75
100
months
ns
N=7; med.surv. 52 mths
N=17; med.surv. 24 mths
0 10 20 30 40 50 60 700
25
50
75
100
months
ns
N=8; med.surv. 28 mths
N=16; med.surv. 31.5 mths
CD4/PD-1
Cut-off: 40%
Cut-off: 40%
CTLA-4
Perc
ent s
urvi
val
Perc
ent s
urvi
val
CD4/CTLA-4
Consistent upregulation of activation markers upon treatment: little association with survival0
25
50
75
100 **
% C
D4+ H
LA-D
R+
w0v1
w8v5
w16v9
w24v13
fuw12v7
w20v11
w4v3
HLA-DRHLA-DR high
HLA-DR low
n.s.0 10 20 30 40 50 60 70
0
25
50
75
100
months
N=22 med surv. 31.5 mths
N=6; med.surv. 32.5 mths
n.s.Cut-off: 2.0 fold
Per
cent
sur
viva
l
HLA-DR high
HLA-DR low
n.s.0 10 20 30 40 50 60 70
0
25
50
75
100
months
N=22 med surv. 31.5 mths
N=6; med.surv. 32.5 mths
n.s.Cut-off: 2.0 fold
Per
cent
sur
viva
l
n.s.0 10 20 30 40 50 60 70
0
25
50
75
100
months
N=22 med surv. 31.5 mths
N=6; med.surv. 32.5 mths
n.s.Cut-off: 2.0 fold
Per
cent
sur
viva
l
T cell activation: effector/memory phenotype
Increased Th differentiation on treatment: relation with survival
w0 w4 w8 w12 w16 w20 w24 fu0
20
40
60
80
100
** ** ** ** ** ***
% n
on-n
aive
of C
D4+
cells
0 10 20 30 40 50 60 700
25
50
75
100
p = 0.036
months
Perc
ent s
urvi
val
N=12; med.surv. 41 mths
N=16; med.surv. 20 mths
Cut-off: 30%
w0 w4 w8 w12 w16 w20 w24 fu0
20
40
60
80
100
* * *
0 10 20 30 40 50 60 700
25
50
75
100
ns
months
Perc
ent s
urvi
val
≥ 30% increase
< 30% increase
N=7; med. surv. 41 mths
N=21; med. surv. 29 mths
Cut-off: 30%* p<0.05
CD8+ T cells
% n
on-n
aive
of C
D8+
cells
≥ 30% increase
< 30% increase
CD4+ T cells
Healthy donor range
Healthy donor range
High Treg rates: associated with SD/PD and reduced survival
0 10 20 30 40 50 60 700
25
50
75
100 50% increase in Tregs at w24no increase in Tregs at w24
p = 0.023
months
Perc
ent s
urvi
val
N=18; med. surv. 37.0 mths
N=6; med. surv. 20.5 mths
Regulatory T cells (nTregs)
0
3
6
9
12
15
% C
D4+
/CD
25hi
gh
0
3
6
9
12
15 ***
PR SD/PD
w0v1 w8v5 w16v9 w24v13 fu w0v1 w8v5 w16v9 w24v13 fu
0 10 20 30 40 50 60 700
25
50
75
100
months
Perc
ent s
urvi
val N=18; med.surv. 36.0 mths
N=6; med.surv. 20 mths
High Treg levels before Tx
Low Treg levels before Tx
p = 0.089Cut-off: 6.3%
0 10 20 30 40 50 60 700
25
50
75
100
months
Perc
ent s
urvi
val N=18; med.surv. 36.0 mths
N=6; med.surv. 20 mths
High Treg levels before Tx
Low Treg levels before Tx
p = 0.089Cut-off: 6.3%
Tumor-related elevated pre-treatment frequencies of CD4+CTLA4+ Tcellshave predictive value for survival on treatment
Pre-treatment CTLA4+ Th cells
IgG1 CTLA-4C
D4
CD
4
15.9 %
0
5
10
15
20p=0.0216
% C
D4+ C
TLA
-4+
(of “
conv
entio
nal”
CD
4s)
prostate cancerpatients
healthy donors
IgG1 CTLA-4C
D4
CD
4
15.9 %
IgG1 CTLA-4C
D4
CD
4
15.9 %
0
5
10
15
20p=0.0216
% C
D4+ C
TLA
-4+
(of “
conv
entio
nal”
CD
4s)
prostate cancerpatients
healthy donors
0 10 20 30 40 50 60 700
25
50
75
100
p = 0.011
months
High level of CD4+CTLA-4+
Low level of CD4+CTLA-4+
N=13; med.surv. 52 mths
N=12;med.surv. 20.5 mths
Cut-off: 2.4%
Per
cent
sur
viva
l
T cell activation profile
Potential biomarkers:predictive vs prognostic; no relation according to Halabi-predicted survival
On-treatment predictive Immune parameter Median Survival
between groups P-value Median Halabi
Predicted Survival P-value
Non-naïve CD4+ cells 41.0 vs. 20.0 0.036 16.5 vs. 21.4 0.086 CD8+ICOS+ 21.0 vs. 57.0 0.043 16.7 vs.19.3 0.622 CD4+CD25intFoxP3+ 41.0 vs. 21.0 0.030 19.6 vs. 15.0 0.401 CD4+CD25hiFoxP3+ Tregs 37.0 vs. 21.0 0.045 15.0 vs 19.6 0.201
Pre-treatment predictive Immune parameter Median survival
between groups P-value Median Halabi
Predicted Survival P-value
Non-naïve CD8+ cells n.r. vs. 20.5 0.028 21.6 vs. 17.3 0.222 Non-naïve CD4+ cells 19.0 vs. 41.0 0.02 21.4 vs. 16.7 0.021 CD4+PD-1+ 41.0 vs. 18.0 0.014 20.5 vs. 15.9 0.194 CD4+CTLA-4+ (conv. T cells) 52.0 vs. 20.5 0.011 19.0 vs. 15.9 0.097 CD4+CD25hiFoxP3+ Tregs 20.0 vs. 36.0 0.087 20.9 vs. 19.0 0.230 n.r.= not reached
T cell activation profile
Unsupervised cluster analysis:CD4+CDTLA4+ as dominant predictor of survival on treatment
Min. Max.Stat. cut-off
004
006
011
013
018
014
003
022
002
008
017
007
009
023
015
026
010
021
025
020
027
016
019
005
028
004
006
011
013
018
014
003
022
002
008
017
007
009
023
015
026
010
021
025
020
027
016
019
005
028
Treg increaseNon-naïve CD4+ T cell increaseCD4+CD25intFoxP3+ increaseTreg preNon-naïve CD4+ T cell preNon-naïve CD8+ T cell preCD4+CTLA-4+ preCD4+PD-1+ pre
group 1 group 2group 3
Patient codes
0 10 20 30 40 50 60 700
25
50
75
100 group 3group 2
p=0.036
N=14; med.surv. 46.5 months
N=9; med.surv. 21 months
0 10 20 30 40 50 60 700
25
50
75
100 group 3group 2
p=0.036
N=14; med.surv. 46.5 months
N=9; med.surv. 21 months
months
perc
ent s
urvi
val
HSC CMP IMC
moMDSC
grMDSC
MO DC
Granulocytes
MΦ/M2
STAT3IL-6/VEGF
STAT3IL-6/VEGF
IL-10
Cancer DC
IL-6
IL-6/VEGF
IL-4IL-13
HSC CMP IMC
moMDSC
grMDSC
MO DC
Granulocytes
MΦ/M2
STAT3IL-6/VEGF
STAT3IL-6/VEGF
IL-10
Cancer DC
IL-6
IL-6/VEGF
IL-4IL-13
Myeloid subsets: also targets for GVAX and ipilimumab?
Oosterhoff Immunother 2011
Liu CII 2009
Suzuki Cell Transplant 2010
GM-CSF
PBDC: subset activation
pDCBDCA2/CD123
cDC1BDCA1/CD1c
cDC2BDCA3/CD141
cDC3CD14dim/MDC8
w0 w4 w8 w12 w16 w20 w24 fu0
5
10
15
20
25
***
*
**
w0 w4 w8 w12 w16 w20 w24 fu0
10
20
**
*
w0 w4 w8 w12 w16 w20 w24 fu0
10
20
30
40
50
*
w0 w4 w8 w12 w16 w20 w24 fu0
5
10
15
20*
*
CD
40 M
edFI
PBDC: cDC1 and cDC3 activation
Increased activation of cDC1 and cDC3 (also known as 6-sulfo LacNAc+ or SLAN-DC ) is related to survival
Per
cent
sur
viva
l≥ 70% increase< 70% increase
N=22; med. surv. 38.5 mths
N=6; med. surv. 15.5 mthsCut-off 70% p = 0.0004
Per
cent
sur
viva
l
≥ 70% increase< 70% increase
N=19; med. surv. 40 mths
N=9; med. surv. 19 mths
Cut-off 70% p = 0.0031
Per
cent
sur
viva
l
cDC1 and cDC3 activationActivation of one or none
N=15; med. surv. 52 mths
N=13; med. surv. 16 mths
p < 0 .0001
cDC1 activation cDC3 activation
MDSC: monocytoid
Monocytic MDSC (Lin-CD14+HLA-DR-)Li
neag
e(C
D3/
16/1
9/56
)
R3
CD14HLA-DR
R2
0 10 20 30 40 50 60 700
25
50
75
100
months
high pretreatment
low pretreatment
N=13; med. surv. 52 mths
N=11; med. surv. 20 mthsCut-off 0.3% p = 0.0062
Per
cent
sur
viva
l
% L
in- C
D14
+ HLA
-DR
-
w0v1 w8v5 w16v9 w24v13 fu10.0
0.5
1.0
1.5
2.0*
% L
in- C
D14
+ HLA
-DR
-
w0v1 w8v5 w16v9 w24v13 fu10.0
0.5
1.0
1.5
2.0*
High pre-treatment levels of mMDSC are associated with poor survival
Filipazzi et al. JCO 2007
MDSC: granulocytic
CD14
Granulocytic MDSC (CD11b+CD14-CD15+)C
D11
b
R2
CD15
R3
% C
D11
b+ CD
14- C
D15
+
0
1
2
3
4
5
w0v1 w8v5 w16v9 w24v13 fu1
0 10 20 30 40 50 60 700
25
50
75
100
months
= 120% increase< 120% increase
N=12; med. surv. 52 mths
N=9; med. surv. 21 mths
Cut-off 120% p = 0.015
Per
cent
sur
viva
l
On-treatment increases in grMDSC are associated with poor survival
Zea et al. Cancer Res 2005
A predictive T cell and myeloid marker profile
Unsupervised cluster analysis:High DC activation and Th CTLA4 and PD-1 expression and low suppressive MDSC and Treg levels together predict survival on GVAX+ipilimumab treatment
Group 1 Group 2 Group 3 Group 4Group 1 Group 2 Group 3 Group 4
Patient codes
CD4+CD25intFoxP3+ increaseTreg preNon- naive CD4+ T cell preNon-naive CD8+ T cell preCD4+CTLA4+ preCD4+PD-1+ precDC1 CD40 activationcDC3 CD40 activation
Non-naive CD4+ T cell increasegrMDSC increaseTreg increasemMDSC pre
Monocyte CD40 activation
Conclusions
Potential immune biomarkers for patient selection prior to treatment:
mMDSC, Tregs, effector/memory and CD4+PD-1+/CD4+CTLA4+ T cell rates
Next: validation• Treatment specific? (GVAX, ipilimumab monotherapies; other therapies?)
• Disease stage specific? (Early versus advanced prostate cancer?)
• Disease specific? (Melanoma vs prostate cancer)
Anita Stam & Saskia SantegoetsFighting the Blues…
…with wine…
…and awards
Immunotherapy lab
Saskia SantegoetsAnita StamSinéad LougheedPetra Scholten Martine ReijmMary von BlombergRik ScheperTanja de Gruijl
Medical Oncology Clinic
Helen GallFons van den EertweghWinald Gerritsen
Karin Jooss
Natalie Sacks
Kristen Hege
Israel Lowy
CELL GENESYSCELL GENESYS
Jean-Marie Cuillerot