Personalised Medicine: experiences in MS

Giancarlo ComiDept. of Neurology & Institute of Experimental Neurology

Università Vita Salute S.Raffaele, Milano

Premio SAPIO Per la Ricerca Italiana

Firenze, 15 Aprile 2013

Personalised medicine is an emerging model that will revolutionise our current healthcare system.

In the last decade, several genomic aberrations were discovered that are now used as predictive markers for treatment with targeted therapeutics.

The technological advances in the last few years, such as the development of high resolution DNA microarrays or second generation sequencers, have led to a dramatic increase in the number of ongoing genomic profiling studies.

Epidermal growth factor receptor (EGFR) tyrosine kinase

inhibitors and their molecular modes of binding to the target

Gonzales de Castro 2013

Bases for individualised treatment in MS

• Complexity and heterogeneity of MS– Polygenic inheritance– Multifaced gene-gene and gene-environment interaction

• Large intraindividual variability of MS courses– Early long term prognostic factors– Short term prognostic factors

• Treatments with different mechanisms of action and different efficacy/safety profile

• Interindividual variability of the response to treatments– Clinical and MRI predictors– Pharmacogenomics

• Multiple treatment algorithms– Induction– Escalation– Combination

“Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis” Nature 2011

This GWAS study involved 9,772 cases of European descent collected by 23 research groups working in 15 different countries: almost all of the previously suggested associations have been replicated and at least a further 29 novel susceptibility loci identified.

Disease prognosis is mostly defined in the early phase of the disease

Better Prognosis Poorer Prognosis• Female

• Younger (<40 years) age at onset

• Caucasian

• Optic neuritis or isolated sensory symptoms at initial presentation

• Monoregional onset

• Infrequent attacks within the first 5 years

• Long interval between first and second relapse

• No or low disability at 5 years

• Male

• Older (>40 years) age at onset

• Afro-American or non-white

• Motor, cerebellar, or sphincter symptoms at initial presentation

• Poor recovery from the first attachs

• Polyregional onset

• Frequent attacks within the first 5 years

• Short interval between first two attacks

• Short time to reach EDSS level of 4

• Disability at 5 years

Kantarci et al. Prognostic Factors in Multiple Sclerosis. In Handbook of Multiple Sclerosis (3rd edition). Cook SD editor. New York: Marcel Dekker. 2001. 449-463.

MRI prognostic factors

• MRI activity predicts relapses (Kappos 1999; Sormani 2003, 2007)

• MRI activity predicts brain atrophy (Paolillo 2004; Filippi 2004)

• MRI T2 lesion load predicts relapses and long term disability (O’Riordan 1998; Sormani 2003; Fisnisku 2008)

• Cortical lesions predict long term disability (Calabrese 2009; 2010)

• “MT MRI provides useful prognostic markers for the prediction of the long-term evolution of multiple sclerosis (Agosta 2006)

• Spinal cord atrophy predicts EDSS (Rocca 2011; Filippi 2011)

• fMRI correlates with cognitive dysfunction (Rocca 2010)

Pelayo, Montalban et al. Multiple Sclerosis 2010

CIS: n. abnormal EPs - time to EDSS 3.0

SEPsVEPsBAEPs

no MEPs

n: 247 (out of 335)

TeriTeripopo

Disease Modifying Drugs in the market or under evaluation Disease Modifying Drugs in the market or under evaluation by EMA:by EMA:

ALZALZiviv

FTY720FTY720po qdpo qd

LAQLAQpopo

BG12BG12

Wolinsky, modified

CLADCLADiviv

Targets of new MS treatments (Adapted from Linker 2008)

Teriflunomide

Comparative effects of MS therapies in disease activity, disease burden and safety

Relapses Disability Active lesions Brain atrophy

IFN/GA + +/- ++ -

Natalizumab +++ +++ +++ +

Alemtuzumab +++ +++ +++ +++

Fingolimod +++ ++ ++ +++

Teriflunomide ++ + ++ ?

Fumarate +++ ++ +++ +

Laquinimod + +++ + +++

Safety & tolerability

+++

+/-

+/-

+

++

++

+++

How to determine the risk-benefit profile of a drug?

• On the drug side, risks & benefits to be taken in account are:

• Efficacy, short and long-term

• Safety (trusting the drug)

• Tolerability

• Monitoring requirements

• Convenience

• Patient’s personal experience of previous therapies

• Doctor’s personal experience of previous therapies

• Cost of the drug

• (New MOA: FTY, cladribine is trying)

New drug must be evaluated in comparison with existing first-line therapies though the times are not the same, nor the trials’ protocols

IFN-GA

TeriflunomideLaquinimod

Efficacy

Burden of TherapyFactors affecting burden of therapy include convenience, monitoring, tolerability, and safety.

Alemtuzumab

Selecting Optimal Therapy in Relapsing MS: Potential Options?

BG12

Natalizumab

Fingolimod

Mitoxantrone

Individualised MS treatment

• When start treatment

• Which treatment start first

• Early detection of non responders

• Quick change of treatment in non responders

• Change of treatment strategy

Degeneration

Disability

time

EDSS 3

Inflammation

?

Response to treatment

Clinical onset

“Delaying treatment in MS:What is lost is not regained”

BR

B

R

B

R

Escalating therapy

1st line therapy

More aggressive approach

3rd line therapy

4th line therapy

5th line therapy

Mitoxantrone / Cyclophosphamide

CampathRituximab

BMT

Beta-Interferons / Glatiramer acetateLaquinimod/BG12/Teriflunomide

Natalizumab2nd line therapy

Fingolimod

Induction therapy

1st line therapy

2nd line therapy

3rd line therapy

4th line therapy

Natalizumab / IFN / GA/ Laquinimod/ Teriflunomide/BG12

Combination therapy

BMT

Mitoxantrone/Cyclophosphamide Fingolimod/ Campath/Rituximab

Preventive strategies

• Cancer risk screening procedures focused on oncogenic infections

• Baseline risk infection risk management

• Optimization of metabolic control

• Cardiovascular risk modification

• Vaccine preventable diseases

• ……………..

NATALIZUMABProgressive multifocal leukoencephalopathy

PML Risk Stratification in patients treated with Natalizumab

Evaluation of Anti-JCV AbsEvaluation of Anti-JCV Abs

Negative

≤0.10/1,000 (95%CI: 0-0.56)

Positive

0-24 months 1.50/1,000(95%CI: 0.83-2.50)

IS+

25-48 months

10.60/1,000(95%CI: 7.70-14.20)

Positive

0-24 months 0.53/1,000(95%CI: 0.33-0.81)

IS-

25-48 months 3.90/1,000(95%CI: 3.00-4.90)

Adapted from Sölberg-Sørensen P et al. Mult Scler 2012.

Biomarkers

• MRI

• EPs

• OCT

• Body fluid

Geurts et al, Radiology 2005

Double inversion recovery (DIR) - cortex

MS: spinal MR imaging

Brain atrophy – early and profound

De Stefano, Neurology 2010

R=0.6p<0.0001 (Spearmann rank correlation)

Global basal EP* score vs EDSS*VEP, MEP, SEP, BAEP

glo

bal

bas

al E

P s

core

EDSS

N=84

Leocani et al, JNNP 2006

Trip et al. 2005

OCT in chronic ONRNFL thickness correlates with VEP amplitude (axonal loss)

unaffected affected

body fluid biomarkers in MS

Markers for a disrupted blood-brain barrier (BBB): s-ICAM, s-

VCAM (IFN- tx), MMP-9/TIMP-1

Markers for demyelination: myelin breakdown products (?)

Markers for axonal damage: NSE, NFL, CK-BB, not clear

Markers for gliosis: S-100, GFAP, not clear

Markers for remyelination: NCAM, CNTF, not clear

Definition of treatment response

Time

MR

I A

ctiv

ity

New Enhancing Lesions

New T2 lesions have a finite time when they enhance, but timing the scan frequency to catch all of these is near impossible in practice

Monitoring Disease Activity Using MRI: A Recipe for Misinterpretation

Criteria for classification of complete response

No brain active lesions

+ No spinal cord active lesions

+ No relapses

+ No disease progression

+ No brain atrophy increase

+ No spinal cord atrophy increase

+ No increase of NAWM abnormalities

+ No EPs changessensitivity specificity

++

--

--

++

Pozzilli et al., Neurol Sci 2005

MRI MARKERS IN MS RRMS / Identification of individual responders

242 RRMS, 4.3 years IFN treatment

Patients with vs. without one-year MRI active scan (i.e., at least 1 EL)

Patients with vs. without one-year new T2 lesions (i.e., at least 1 new T2 L)

Predictivity of the response to DMTs

• Clinical and demographic

• MRI

– The drug effects on MRI activity predicts the effects on disease activity (Rio 2008; Rio 2009;Durelli 2008; Sormani 2009) and on brain atrophy (Filippi 2004)

• Laboratory

• Pharmacogenomic

• At treatment onset

• Early during treatment

Predictive value of biomarkers in MS TREATMENT

Treatment effects according to Activity/Dissemination of the Disease at First

Event in BENEFIT Parameter indicating

more dissemination or activity at onset

Treatmenteffect*

Parameter indicating less dissemination or activity

at onset

Treatmenteffect*

Multifocal 37% Monofocal 55%

> 9 T2 lesions 43% < 9 T2 lesions 60%

> 1 Gd+ lesion 38% No Gd+ lesion 57%

Treatment effects were significant in all subgroups by log rank tests

*Risk reduction according to unadjusted hazard ratios

EPs global score at treatment onset and treatment response

4,3 4

10,7

0

2

4

6

8

10

12

full-responders partial-responders non-responders

valo

re m

edio

sco

re g

lob

ale

bas

ale

EP

36

30

30

Mann Withney test, * p<0.001

*

*

Biomarkers predictive of the reponse to IFN

• NABs• Overexpression of genes of the NFKb network• increase in the frequency and suppressive activity

of CD4+CD25+Foxp3+ regulatory T cells • baseline overexpression of type I IFN-responsive

genes in nonresponders to IFN-β• High baseline levels of Il-17 in non responders• …………….

Poor predictive value in individual patients

Genome Wide Association Study (GWAS) ForPredicting Response To GA: Results

We selected SNPs with a P-Value < 10-4 31 SNPs

P-value = 10-4

Chr: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

The Genetic Model

• As every SNP predicted response only partially, we used the 31 significant SNPs from the GWAS to develop a genetic predictive model:

the best and most parsimonious model includes only 6 SNPs

 

Super Responders

Super Non-Responders Total

Genetic prediction of response**

Super Responders 45 6 51

Super Non-Responders

6 55 61

Total 51* 61  112

* One of the 52 subjects was removed because of missing genotype for this model** Patients were genetically classified responders if probability of super-response is ≥ 0.5

Sensitivity 87.9%Specificity 92.7%Positive predictive value 90.6%Negative predictive value 90.5%Correctly classified 90.5%

• At treatment onset

• Early during treatment

Predictive value of biomarkers in MS TREATMENT

MRI MRIDMT

12 mMS

Predicts:

Relapses

Disability

MRI activity: at least 2 new T2 / gad lesions

AND

or

Río et al. Mult Scler 2010

36 m

Relationship between active lesions on MRI and treatment response

Rio 2009

Take home message

• Prognostic information are available to orient treatment choice (better value at disease onset!)

• At present we have moderate possibility to predict which patients will respond to the various disease modifying agents in advance of therapy onset

• Pharmacogenomic information is still immature to allow its translation to clinical practice

• In treatment monitoring, MRI is of help in predicting therapeutic response

M. Comola

Neurorehabilitation

L. Leocani

Neurophysiology V. Martinelli

Neurology

M. Filippi

Neuroimaging

M. Falautano

Neuropsychology

G. Martino

Nuroimmnology

I. Sangion M. Galdabini