Document 1052

Advances in Cancer Immunotherapy:
Key Translational Research Opportunities
Michael B. Atkins, M.D.
Deputy Director
Georgetown-Lombardi Comprehensive
Cancer Center
Disclosures
♦ Consultant
•  BMS
•  Genentech
•  Prometheus
♦ SAB
•  Co-Stim
•  cCAM
Outline
♦ Advances in tumor immunology
♦ Immune Checkpoint inhibitors
•  Anti-CTLA4
• PD1 Pathway blockade
Outline
♦ Advances in tumor immunology
♦ Immune Checkpoint inhibitors
•  Anti-CTLA4
• PD1 Pathway blockade
♦ Key translational questions
•  PD1 vs PDL1?
•  Combinations?
•  Biomarkers?
•  Setting?
•  Tumor types?
Immunotherapy activity may be limited by the ability of T
cells to infiltrate the tumor
Metastatic Melanoma
Immunotypes
A (40%)
C (24%)
B(36%)
Proportion surviving (%)
Dead
100
100
Cutoff
CD8
CD8
Highfor
>50
> 50
= high
CD8
Low
< 50 (17)
< 50 = low (13)
8080
6060
4040
p = 0.0504
20
20
00
0
0
Courtesy of Dr Craig Slingluff
Number at risk
Group: H
17
20
40
60
7
6
3
20
40
Time
Time (months)
60
High expression of vascular markers, macrophages, fibroblasts +
Low inflammation and chemokines, few lymphocytes =
Poor effector cell trafficking
Gajewski, Curr Opin Immun 2011
High levels of innate immune signals, chemokines for T cell recruitment
But, negative immune regulators dominate
Gajewski, Curr Opin Immun 2011
x
High levels of innate immune signals, chemokines for T cell recruitment
But, negative immune regulators dominate
Gajewski, Curr Opin Immun 2011
$'#"& >K&>H
Tumor Microenvironment
Activation
(cytokines, lysis, proliferation,
migration to tumor)
MHC
Dendritic
c
cell
B7
B7
B
TCR
TCR
++
++
CD28
CTLA-4
C
+++ T cell
---
M
MHC
++
++
T cell
PD-1
P
PD-L1
P
PD
) "$
--anti-PD-1
ant
PD 1
anti-CTLA-4
PD-1
P
PD
PD-L2
--anti-PD-1
CTLA-4 Blockade (ipilimumab)
PD-1/PDL-1 Blockade
PD-1/PD-L1 Pathway Agents in Development
Company
Agent
Structure
Status
BMS*
BMS-946559
Phase I trial
BMS*
Nivolumab
PDL1 Ab, fully
human, IgG4
PD1 Ab Fully
human, IgG 4 Ab
Curetech/Teva
CT-011
Genentech/Roche*
MPDL3280A
PD 1 Ab, IgG4,
humanized
PD-L1 Ab
Phase III RCC,
Lung Ca; Ph II/III
Melanoma
Phase II
melanoma, RCC
Phase I
Medimmune
MEDI 4736
PD-L1 Ab
Phase I
Merck*
Lambrolizumab
MK-3475
PD 1 Ab, IgG 4
Ab, humanized
Phase I, Phase II/
III Melanoma
Study Design: Phase I Multi-dose Regimen
8-wk treatment cycle
Day 1
*
15
*
29
*
43
*
Rapid PD or clin.
deterioration
57
SCANS
*Dose administered IV Q2wk
Unacceptable
toxicity
CR/PR/SD or PD
but clinically
stable
Eligibility: Advanced MEL, RCC, NSCLC, CRC, or CRPC
with PD after 1-5 systemic therapies
1Topalian
S, et al. NEJM 2012;366:2443-2454.
Off Study
Follow-up
every 8 wks
x 6 (48 wks)
Treat to
confirmed CR,
worsening PD,
unacceptable
toxicity, or 12
cycles
(96 wks)
Clinical activity of nivolumab
Tumor type
(dose, mg/
kg)
No.
pts
NSCLC
(1-10)
129
22 (17)
13 (10)
PFS
(months,
median)
2.3
MEL
(0.1-10)
107
33 (31)
7 (7)
3.7
RCC
(1 or 10)
34
10 (29)
9 (27)
7.3
OR (CR/PR) SD ≥24 wk
No. pts (%) No. pts (%)
 All patients initiated treatment 2008-2012, ≥14 months before
data analysis in March 2013
 OR = CR/PR by RECIST 1.0
 No ORs observed in 19 CRC or 17 CRPC patients
Durability of objective responses induced by nivolumab
in patients with advanced NSCLC, MEL and RCC
NSCLC
Maximum
treatment
duration
Sixty-five of 306 patients
had ORs (CR/PR):
  30 of 65 (46%) responses
were evident at first tumor
evaluation (8 weeks)
  42 of 65 (65%) patients had
responses lasting >1 year
  35 of 65 (54%) responses
were ongoing at time of data
analysis
  Responses persisted off-drug
MEL
RCC
Weeks since treatment initiation
Persistence of nivolumab-induced tumor
regressions after treatment discontinuation
 27 responding patients discontinued treatment for reasons other than PD
(CR, toxicity, maximum dosing 96 wks) and were followed
 19 of 27 (70%) maintained responses off-drug for 16-59 weeks; 14 of 19
responses ongoing at time of analysis
Proportion progression-free
Duration of response in patients receiving nivolumab
1.0
NSCLC
Melanoma
RCC
0.9
0.8
Med., mo. (95%CI)
(n=22) 17.0 ( 9.7 - NE)
(n=33) 24.0 (17.0 - NE)
(n=10) 12.9 (11.4 - NE)
0.7
0.6
Censored
0.5
0.4
0.3
0.2
0.1
0.0
0
3
6
12
15
18
21
24
27
30
33
1
0
0
0
0
0
Months since initiation of response
No. at Risk
NSCLC
MEL
RCC
9
22
33
10
NE, not estimable
21
33
10
16
30
10
14
27
9
12
19
5
8
15
3
5
12
2
2
9
2
2
5
2
1
0
1
Change in target lesions from baseline (%)
Unconventional “immune-related” responses in
13 patients with NSCLC, MEL, and RCC
OG
HGG
)!!$
!" !-!$
MG
$%&")$$!
"!+%"!
KG
IG
G
?IG
?KG
?MG
?OG
?HGG
G
HG
IG
JG
KG
LG
MG
NG
OG
PG
HGG
HHG
HIG
HJG
HKG HLG
HMG
Weeks since treatment initiation
 13 of 270 pts (5%) with NSCLC/MEL /RCC had unconventional responses
 irResponse durability and persistence off-drug were similar to conventional
RECIST responses
100
Med.OS
(months)
Died/treated
88/129
80
Overall survival (%)
60
40
NSCLC 9.6
20
(95% CI);
pts at risk
0
100
% survival
1 YR 2YR
42
14
(7.8, 12.4)
(33, 51);
43
(4, 24);
5
MEL
16.8
62
43
(95% CI);
pts at risk
(12.5, 31.6)
(53, 72);
55
(32, 53);
26
70
50
(55, 86);
23
(31 ,70);
8
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51
Died/treated
60/107
80
60
40
20
0
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51
100
Died/treated
15/34
80
60
40
20
RCC
>22
(95% CI);
pts at risk
(13.6, NE)
0
0
3
6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51
Months since treatment initiation
Drug-related adverse events of special interest
(AEOSI) occurring in ≥ 3% of 304 patients
AEOSI (“irAE”)
All grades (n, %)
Grade 3-4 (n, %)
Any
138 (45)
Rash
41 (14)
0
Diarrhea
36 (12)
3 (1)
Pruritis
31 (10)
1 (0.3)
ALT 13 (4)
2 (1)
Pneumonitis*
10 (3)
3 (1)
Infusion reaction
8 (3)
TSH 11 (4)
Vitiligo
9 (3)
AST 11 (4)
18 (6)
0
1 (0.3)
0
2 (1)
*There were 3 (1%) deaths in patients with pneumonitis (2 NSCLC, 1 CRC).
AEOSIs in ≤1% of pts included colitis, hepatitis, hypophysitis, nephritis and
thyroiditis. Analysis as of July 3, 2012.
Topalian et al, ESMO 2012
PD1/PDL1:
Key Translational Questions (1)
♦ PD1 vs PDL1 blockade?
A2<
/!'6
#"#&"'
%!!
&) "$%
6<
6<
<
6= !'66<
!'66<
/!'6
#"#&"'
%!!
&) "$%
768%!
A2<
6<
768%!
<
6<
6=
PD1 vs PD-L1 Abs: Observations
♦ Anti-PD1 appears more active
•  Better or more accessible target?
•  Better antibodies?
•  Is blocking PDL2 important?
♦ Anti-PDL1 appears less toxic
•  Less effective antibody?
•  Keeping PDL2 unblocked prevents pneumonitis?
PD-1/PDL-1:
Key Translational Questions (2)
♦ How do we enhance the efficacy of PD-1/
PDL-1 blockade?
Combination with
•  T cell agonists?
•  Other checkpoint inhibitors?
•  Tumor specific therapy?
Ipi/Nivo- Tumor Responses
""$&=1< 5!*") G> 5# ) #'!&%!"!)$$!&""$&%
First occurrence of new lesion
Wolchok et al. ASCO #9012
JB;F) "$)'"!
# ) AF
I>F
!*") =BF
I=F
" !'"!7""$&=8
@>F
?<F
(">"3'*',''$*+('&+,HI4#+
PD-1/PDL-1:
Key Translational Questions (2)
♦ How do we enhance the efficacy of PD-1/
PDL-1 blockade?
Combination with
•  T cell agonists?
•  Other checkpoint inhibitors?
•  Tumor specific therapy?
♦ How does one identify what is best for a
specific tumor?
•  Analysis of baseline biomarkers?
•  Analysis of escape mechanisms?
PD-1/PDL-1:
Key Translational Questions (3)
♦ Use of biomarkers
•  Can one identify the patients most likely to
benefit?
•  What is the optimal biomarker?
Melanoma
Preliminary molecular marker
studies: Correlation of PD-L1
expression in pretreatment
tumor biopsies with clinical
outcomes
$"#"$'"!"#'!&%
18/18
B
Kidney
18/31
13/31
#H;2;;=
Lung
0/18
49 patients include 20 with melanoma,13
NSCLC, 7 colon, 6 kidney, and 3 prostate
cancer.
9'*%$*&$ $'%*2$2+
%%)%+&<%" $$7<8
PD-L1 negative stain can’t be relied on
– Assays are technically difficult, imperfect; results may
be different depending on the antibody/assay
– 5% expression, tumor heterogeneity, and inducible
gene = sampling error (false negs)
– Primary not equiv to metastasis
– PD-L2 could be dominant factor in some tumors
– May be less relevant in combination trials
$&,6+('&++"&>H& ,+
Biomarker Questions:PD-L1 expression
♦ Which antibody?
♦ Measure on tumor, immune cells or both?
♦ Cut-off parameters? more stringent = higher
ORR, but miss more responders
♦ Which samples-primary vs metastasis?
PDL-1 Expression and Response
<G
"%'*
<6
'*
"3'$2%
@'($"&99IGHIA
KI
[email protected]
[email protected]
"3'$2%
@*DPGHHA
KK
[email protected]
[email protected]
JIOG
@%"DPGHGA
@
A
JG
[email protected]
[email protected]
"3'$2%=
("$"%2%
@$$!&DJGGJA
IN
[email protected]
[email protected]
"3'$2%
@*'++'DJGHMA
JK
[email protected]
[email protected]
Presented by: Walter J. Urba, MD, PhD
%%)%+&<%" $$7=8
•  Why don’t all PDL1+ tumors respond?
– Expression could be constitutive due to oncogene
(? no immune infiltrate)
– Expression of other checkpoint inhibitors
Is a measure of immune infiltration a
better predictor of responsiveness?
Which immune parameter?
-  CD8/Treg ratio?
-  IFNγ?
-  chemokine expression?
Correlation of PD-L1 expression with the
presence of TILs in 150 melanocytic lesions
F"%%
*,!+,!H& *+('&*+8
+,!"+,*2
'',!*
,2%'*
,6(+8
VG<GGGH
 TILs are necessary but not sufficient for PD-L1 expression
in melanoma.
 Still only 38% express PDL1
(Taube et al., Science Transl Med 2012)
12-Chemokine Gene Signature: Non-Lymph Node Melanoma
Metastases
Mule et al Nature Scientific
Reports 2012
Ectopic Lymph Node Structures in
Melanoma
Tumor
PD-1/PDL-1:
Key Translational Questions (4)
♦ Is PD-L1 expression•  Influenced by prior systemic therapy?
•  Predictive of benefit for that other therapy?
•  Is some other biomarker better?
♦ Is PD-1 blockade better applied in the treatment naïve
or resistant setting?
♦ What are the mechanisms of resistance to PD-1/
PDL-1 pathway blockade?
•  Are they key to rationale combination therapy?
•  Particularly relevant if other immunoregulatory pathways go
up with PD1 blockade.
Phase II PD1Ab RCC Study
Eligibility: metastatic clear cell RCC, no prior systemic rx
Baseline
primary
tumor
biopsy
R
a
n
d
o
m
i
z
e
VEGF
TKI Y
Biopsy
PD-1 Ab X
PD-1 Ab X
Biopsy
VEGF TKI Y
Assess:
•  T cell infiltrate
•  PD-L1 expression
•  Chemokine signature
*Serum samples at baseline, dose 4 and at time of progression
(Tregs, MDSCs, sPD-L1)
Primary EP: PFS
Secondary endpoints:
Safety
Correlatives
-CD8 T cell infiltration of kidney
-CD8/Treg ratio, CD4/Treg ratio
-Tumor PD-L1 expression
PD-1/PDL-1:
Key Translational Questions (5)
♦ What tumors are best to treat?
No. patients
"$$'"!"&) "$6<,#$%%"!+&&) "$&-#
!?C#'!&%&$&+&!'66<
G
No. OR/total: 8/20
3/13
2/6
0/7
0/3
-&,+4*'&+"*>[email protected]"&6,2%'*"'(+6!WLS',2%'*$$+
5(*++"& $$+2*>H92+"& %LH&%&2$+,"&"& ,!&")2<
>[email protected]>HA5(*++"'&"&2$-($&*+
'2*;'2<&!&<,2*3"4+
%%2&'$' 6<IGGO:O;KMN>KNN<U',,*%"&<
Interrogation of a 12-Chemokine Gene Expression
Signature across Solid Tumors of Differing Histology:
Principal Component Analysis
)&&)$!'
#"$&%=;<=
# Above
Total 90th
% Above
# CEL percentil 90th
Tissue Type
files e
percentile
Bladder
212
27
12.74
Brain
438
4
0.91
Breast
3705
401
10.82
Cervix
75
19
25.33
Endometrium
333
14
4.20
Esophagus
90
6
6.67
Kidney
850
61
7.18
Large_Bowel
2111
169
8.01
Larynx
56
7
12.50
Liver
116
5
4.31
Lung
2708
488
18.02
Oral_Cavity
98
25
25.51
Ovary
670
21
3.13
Pancreas
468
15
3.21
Peritoneum
30
1
3.33
Prostate
981
5
0.51
Rectum-Anus
188
10
5.32
Renal_Pelvis
62
1
1.61
Skin
569
115
20.21
Small_Intestine
52
2
3.85
Soft_Tissue
97
14
14.43
Stomach
133
13
9.77
Thyroid
71
5
7.04
Tongue
32
8
25.00
Uterus
377
13
3.45
"+%&&")%" $$%/
♦ !'-%%&-#%&"&$&
•  "%+&<7!5"$=8,#$%%"!!
<!5"$ "!%!&)$
•  "%!"!"!$*$"<
•  "%+$<,#$%%"!%%"&+&!"
!&+&"!*!'"!&$#-
♦ &!$%"#%-#$"$&"&$& !&"$
,#"$&"$-0)&!"&%'"!"$
%&$''"!#)$#"%%
PD-1/PDL-1:
Key Translational Questions (6)
♦ Why are some tumors with immune
infiltrates not responsive to PD1/
PDL1 blockade?
• Expression of other immunoinhibitory molecules?
• Wrong immune cells
,)%&,#$%%
)'#
)!"!&"$-#&"$%1
%$$)&$&%"$&) "$
)!"&$#-
58
PD-1/PDL-1:
Key Translational Questions (7)
♦ Can one turn a non-inflamed tumor into an
inflamed tumor with other treatment (RT,
chemo, vaccine, molecularly targeted
therapy)?
♦ If so, how do you coordinate/sequence
therapies?
"&!""-'&"+++'",4",!"&*+%$&'6,>++'",
&- &5(*++"'&C$$"&$,*,
gp100
100
Fold increase
>[email protected]
A
10
1
&>*,%&,
CD8+ T cells (40X)
819
1089 2367
ILG
# CD8+ T cells / hpf
*>*,%&,
8440 3988 9382 2373 7198 9297 9395
IGG
HLG
HGG
LG
G
JPOOPJOIIJNJPIPNPJPL OHP HGOP LGP KPKG
Boni et al. Cancer Res 2010 & Wargo et al. AACR 2011
Vemurafenib Serial Bx Protocol:
Schema
Eligibility
Stage III or IV
melanoma with:
• 
• 
• 
BRAFV600
Mutation
Measurable
disease AND
Disease accessible
for repeat Bx and/or
resection
Vemurafenib 960 mg bid
Day
PBMC
Tumor Bx
Tumor Meas
Treat until
PD
1
8
15
29
x
x
x
x
x
x
x
x
x
43
x
Patient 1 Biopsies: CD8 Staining
Intratumoral
lymph node
structure
Necrosis
PD-1/PDL-1:
Key Translational Questions (8)
♦ Adjuvant therapy: Do you need the T cells in
the tumor microenvironment for adjuvant
anti-PD-1/PDL1 to work?
•  Adjuvant vs neoadjuvant approach
PD1 pathway/checkpoint inhibitor therapy:
Future Directions
♦ Activity relative to standard therapies
♦ Prospective study of optimal sequencing with
standard therapies (move to first line)
♦ Predictive biomarker refinement
♦ Combinations
♦ Mechanisms of resistance
♦ Identification of other therapeutically relevant
checkpoints
♦ Adjuvant therapy
♦ Expansion of utility to other patients/diseases