Transcript view slides - PFF Summit

MOLECULAR AND CELLULAR
BIOMARKERS OF PULMONARY FIBROSIS
NAFTALI KAMINSKI, MD
TRANSLATIONAL SCIENCE: PROGRESS TOWARDS
PERSONALIZED MEDICINE FOR IPF – BIG DATA MEETS
PATIENT CARE
NOVEMBER 13, 2015
Naftali Kaminski
Chief of Pulmonary, Critical Care & Sleep Medicine
Yale School of Medicine
[email protected]
Molecular and Cellular Biomarkers in
Pulmonary Fibrosis
•
•
•
•
Biomarkers
Peripheral Blood
BAL
Lung
What is a biomarker ?
• A variable that is objectively measured and
indicates normal/pathogenic processes or
pharmacologic responses
• Types: Predisposition, Diagnostic, Prognostic,
Treatment response
• Simple, technically accurate, broadly
reproducible (multiple
cohorts), standardized,
multiple cohorts,
have an acceptable risk (Blood > BAL > Lung
tissue), Reflective of disease pathogenesis
Strimbu K, Curr Opin HIV AIDS, 2010
Normal Alveolar
Structure
Usual Interstitial
Pneumonia Lesion
Recurring microinjuries
Epithelial cell injury
Epithelial and fibroblast
activation
Matrix Remodeling
Cell migration and
accumulation
WNT
LET-7
SPA2
TERT
TERTC
UPR
MUC5B
MMP7
SPC
ER
stress
Epithelial cell
dysfunction
Short
telomeres
Apoptosis
Alveolar
macrophage
Mitochondrial
dysfunction
Integrins
TGFβ
MMP7
MUC
miRNA 29
ASMA
COL3
COL1
COMP
Fibroblasts
ECM
Collagen
Fibrogenesis
TLR3
Myofibroblasts
Integrins
:Th2
EMT
LOXL2
COMP
ECM degradation products
MMP1
Lymphocytes
OPN
YKL40
CCL18
CXCL13
Immune
dysregulation
Leukocytes
Fibrocytes
Sema7a
Treg
CD4+
AntiHSP70
IgG
Modified from Ley et al, AJPLung. 2014
Circulating
biomarkers
SPC
ER
stress
SPA2
SPC
WNT
TERT
MUC5B
SPA2
LET-7
TERTC
TERT
TOLLIP UPR ER
MUC5B
Short
TERTC
ELMOD2 stress telomeres
MMP7 Apoptosis
Short
TLR3
telomeres
UPRMitochondrial
dysfunction
Apoptosis
Epithelial cell
dysfunction
Alveolar
macrophage
SPA,
ECM
miRNA 29
ASMA
COL3
COL1
COMP
MMP7
TGFβ
SPA
SPD
MUC
TLR3
KL6
Fibroblasts
cCK18
Myofibroblasts
Fibrogenesis
Collagen
SPD, MMP7, KL6,
MUC1, cCK18,
YKL40, Anti-HSP70,
CXCL13, SPP1,
COMP, VCAM, Periostin,
MMP1, LOXL2, matrix
neoepitopes,
microRNAs
:Th2
EMT
LOXL2
COMP
ECM degradation
products
MMP1
Lymphocytes
SPP1
OPN
YKL40
CCL18
CXCL13
Immune
dysregulation
Leukocytes
Fibrocytes
Sema7a
Treg
CD4+
AntiHSP70
IgG
MMP7
• A small MMP that degrades casein,
proteoglycans and fibronectin
• A WNT/β-catenin target
• Promotes epithelial cell migration & apoptosis
• Multiple potential roles in regulating local
inflammation and growth factor activation
• Role in epithelial Cancers
• KO mice relatively protected
• Highly expressed in epithelial cells in IPF
• The most validated marker so far – >5 cohorts
MMP7 is increased in the blood and lungs of IPF Patients
Rosas et al. PLos Med 2008
Zuo et al. PNAS 2002
Increased MMP7 levels at presentation are associated
with increased mortality
Derivation (n=140)
Replication (n=101)
PCMI - personal clinical and molecular
mortality prediction index, derived
from derivation cohort
PCMI = 114*I(Male) + 2*(100% - FVC
% Predicted) + 3* (100% – DlCo %
Predicted) + 111*I (MMP7 >= 4.3
ng/mL)
>330 median survival 1.5 year
<330 median survival >5 years
P=0.0021
P=0.0111
OS
TFS
>threshold
<threshold
>threshold
<threshold
Richards et al. AJRCCM 2012
P-value = 0.003
The predictive performance PCMI :
area under the curve ranges from
0.74–0.84, C statistic from 0.73–0.84.
This slide has been removed at the request of the
presenter because it contains unpublished data.
11
MMP7 may be an indicator of early disease
MMP7
Rosas et al. PLos Med 2008
ET1
Kropsky
al. AJRCCM
2015
Doyle et
AJRCCM
2015
SPD
TIMP2
MMP7
• Consistently increased in IPF and other ILDs
(multiple cohorts)
• Increased in patients at risk or with preclinical
disease (Multiple cohorts)
Challenges:
• Predicts mortality in IPF (multiple cohorts including
• Standardization,
levels depend on collection matrix
Inspire)
•• Mechanistic
rolea unclear
Most probably
marker of epithelial changes in
Promise:
IPF
• Very robust signal
• Cheap and easy
• Most reproducible
Probably best: In combination with KL-6, SPD
What about targets of MMPs?
• Prospectively collected serum samples at baseline, 1
month, 3 months, and 6 months
• Analysed by panel of novel matrix metalloprotease
(MMP)-degraded ECM proteins, by ELISA-based,
neoepitope assay. CDP – Collagen Degradation Product
• 11 CDP were tested in a discovery cohort of 55
patients
• 8 were assessed in a validation cohort of 134 patients
with 50 age-matched and sex-matched controls
• Changes in CDP concentrations were related to
subsequent progression (defined as death or decline in
forced vital capacity >10% at 12 months after study
enrolment)
Jenkins et al LRM 2013
• In the discovery cohort, 7 CDPs differed significantly
between controls and IPF.
• Baseline concentrations of 6 CDPs were significantly
higher in patients with progressive IPF (n=32) than in
those with stable disease (n=23).
• In the validation cohort, 4 CDPs at baseline were
higher in patients with IPF controls.
• 6 CDPs were  in patients with progressive IPF than
in patients with stable disease.
• Baseline concentrations of 2 CDPs were associated
with subsequent increased mortality.
The rate of change magnitude of change in CDPs
concentrations over time correlated with overall
survival
•
•
•
•
Summary
Concentrations of CDPs are increased in IPF
Increased CDPs concentrations are
associated with disease progression
The rate of this increase predicted
survival.
Is this a measure of disease activity?
SPC
ER
stress
SPA2
SPC
WNT
TERT
MUC5B
SPA2
LET-7
TERTC
TERT
TOLLIP UPR ER
MUC5B
Short
TERTC
ELMOD2 stress telomeres
MMP7 Apoptosis
Short
TLR3
telomeres
UPRMitochondrial
dysfunction
Apoptosis
Epithelial cell
dysfunction
Blood cells:
CD4+ CD28+ T cells
ECM
Tregs
Sema7a+
T cells
TGFβ
Fibrocytes
miRNA 29
ASMA
COL3
COL1
COMP
Fibroblasts
Alveolar
macrophage
[CD4+ CD28+ ]
[CD4+ CD25+ FOXP3+ ]
[CD4+ CD25+Collagen
FOXP3+ Sema7a+ ]
Fibrogenesis
[CD45+ CD34+ Coll1+ FN1+ ] :Th2
TLR3
EMT
Lymphocytes
OPN
YKL40
CCL18
Myofibroblasts
CXCL13
Immune
dysregulation
Leukocytes
Fibrocytes
Sema7a
Treg
CD4+
AntiHSP70
IgG
Microarrays in PBMCs identify 52 genes associated with
survival in IPF
Cluster 1
Cluster 2
Chicago cohort
Pittsburgh N=75
Cluster 1
Cluster 2
Survival Probability
Chicago N=45
Cluster 1
Cluster 2
P=0.019
HR:3.30
Years
Survival Probability
Pittsburgh cohort
HR:1.96
Years
52 Genes (FDR<5%)
52 Genes
Herazo-Maya JD, et al. Sci Transl Med. 2013
Demographic and clinical characteristics of the IPF subjects in two clusters
Characteristics
‡
Age – yr
Mean
Range
Gender – no. (%)
Males
Females
Race – no. (%)
Caucasian
Others
Smoking status – no. (%)
Ever smoker
Never smoker
§
Pulmonary function tests
FVC%
DLCO%
FEV1%
Diagnostic Strategy – no.
Clinical +
¶
Clinical + HRCT
(%)
Immunosuppressive
HRCT+ UIP Proven biopsy
No
**
yes
therapy
– no. (%)
†
Cluster 1 (N=45)
Cluster 2 (N=30)
P Value
67.9 (± 7.3)
56 – 82
70.5 (± 9.2)
50 – 84
0.18
28 (62.2%)
17 (37.8%)
24 (80%)
6 (20%)
0.12
65% (± 16)
49% (± 15)
76% (± 16)
66% (± 18)
48% (± 22)
77% (± 18)
34 (75.6%)
11 (24.4%)
18 (60%)
12 (40%)
0.88
0.76
0.65
0.2
41 (91.1%)
4 (8.9%)
23 (76.7%)
7 (23.3%)
0.51
(95.6%)
30 (100%)
No significant 432clinical
differences
(4.4%)
0
0.46
between patients 30in(63%)the two
clusters
17 (62.1%)
15 (37%)
13 (37.9%)
except for transplant
free
survival
0.1
This slide has been removed at the request of the
presenter because it contains unpublished data.
21
This slide has been removed at the request of the
presenter because it contains unpublished data.
22
This slide has been removed at the request of the
presenter because it contains unpublished data.
23
Conclusions PBMC gene expression data
• 52 gene signature predicts outcome in 5 cohorts
regardless of technique differences
• Pooled Analysis using scoring algorithm of molecular
subphenotypes (SAMS) confirm the results
• The signature improves the predictions by GAP
• Over time up-regulated genes have an up-trend over
time
• Down-regulated genes have a down-trend over time
• 41 genes in the signature were significantly (P<0.05)
associated with changes in forced vital capacity by a
linear mixed model
• No shift in expression profiles was seen between high
and low risk patients
Can we get closer to the
lung?
This slide has been removed at the request of the
presenter because it contains unpublished data.
26
Validation: BAL gene expression
signatures predict outcome in IPF
(nCounter)
A
All Cohorts
B
All Cohorts
The genes that drive the BAL
signature are not what you expect
What about the lung itself?
The extent of fibrosis (Fibroblastic
Foci) in the lung is a predictor of
IPF mortality
King T, et al, AJRCCM, 2001
Gene expression profiles in lungs
identify genes associated with IPF
progression
Mortality based on time
since onset of symptoms
< 6 months ≥ 24 months
Slow
Rapid
Boon K, et al. PLoS One. 2009
Selman M, et al. PLoS One. 2007
This slide has been removed at the request of the
presenter because it contains unpublished data.
31
This slide has been removed at the request of the
presenter because it contains unpublished data.
32
This slide has been removed at the request of the
presenter because it contains unpublished data.
33
This slide has been removed at the request of the
presenter because it contains unpublished data.
34
We have the data – but we have a challenge
Cluster 1
Cluster 2
PF community
Everett M. Rogers' Diffusion of Innovations
Collaborators
Kevin Gibson
Kathy Lindell
Yingze Zhange
Antje Prasse
Gisli Jenkins
Toby Maher
Richard Marshal
Imre Noth
Wim Wuyts
Paola Rottoli
Ivan Rosas
Fernando Martinez
And many others..