Art's Notes From Cold Spring Harbor Labs Meeting 2/3/08

art — Wed, 13 Feb 2008 11:03:51 -0500

Thanks to generous financial and organizational support from Stan Trotman, the Banbury Center, and Jim Watson, Accelerated Cure Project was able to convene a meeting of 20+ researchers to discuss how to address a disease that has both genetic susceptibility and environmental triggers.

We invited MS neurologists and researchers, researchers from related fields, and from other diseases to come together from 2/3/08 thru 2/6/08 at the Banbury Center on Cold Spring Harbor on Long Island and talk about what they work on and how it could be applied to cracking a disease like MS.

My sketchy notes are included below - I didn't do the same level of detail we do for AAN and ECTRIMS because the pace was too fast, but I figured it wouldn't hurt to share them.

It was a great meeting and we are working on a high-level proposal to circulate among the attendees for a major study that incorporates what we learned. We'll follow that up with a detailed proposal once we get feedback from the participants.

continued...

Cold Spring Harbor Laboratory Meeting 02/03/2008

Genes and the environment: new strategies for research on MS

Funded by Stanley Trotman
Organized by Ben Greenberg, Accelerated Cure Project, Banbury Center

Sun 02/03/08:

Reception, Dinner, Super Bowl XLII

Mon 02/04/08:

Jan Witkowski, Director of Banbury Center

Welcome, announcements

Ben Greenberg, Johns Hopkins
Knowns and Unknowns

Overview of agenda, purpose of meeting: to develop a protocol to
unearth causative factors in a disease where genetics and
environment play a role together.

There are known knowns. These are things we know that we know. There
are known unknowns. That is to say, there are things that we know we
don't know. But there are also unknown unknowns. There are things we
don't know we don't know.
-- Donald Rumsfeld

MS Known knowns:

Only involves CNS
Classic definition as white matter disease (in question)
Charcot description from 1860's is still state of the art
Epidemiology
Classic MRI findings
Gray matter pathology exists
Inflammation and axonal loss over time

MS Known Unknowns:

What causes/triggers MS?
Geographic distribution of MS - what does it mean?
Is it a syndrome rather than single disease?

MS Unknown Uknowns:

How to design a study when you don't know all the variables?
Prospective vs retrospective?
Patient vs Clinician generated data?
What are the relevant subsets of patients?
Importance of outliers in data analysis?
Inter- and Intra-observer reliability?

Amit Bar-Or, Montreal Neurological Institute
Immunology of MS

Dogma: MS as an autoimmune 'Th1' disease leading to inflammatory
demyelination of CNS white matter.

Autoimmune?
Th1/Th17
Inflammatory and degenerative
Demyelinating and axonal/neuronal injury
Both white and gray matter
Pathological heterogeneity

MS does not fit technical definition of autoimmune disease as does
myasthenia gravis.

Neuro-immunology of MS could go:
A. dysregulated immune response -> inflammation -> degeneration
B. degeneration->dysregulation->inflammation

Model 'A' above supported by animal models such as EAE. Activated Th
cell adheres to BBB and eventually enters the CNS where it is
reactivated and attacks myelin. Tysabri's efficacy supports this
model.

Model 'B' supported by discovery of axonal transection outside areas
of inflammation. Presence of macrophages could indicate cause of
damage, or they could just be cleaning up damage done previously by
something else. Cortical lesions seem to require a different
explanation than peri-vascular lesions. Immunoablation does not stop
atrophy of the brain, which suggests possible other mechanism for
degeneration.

New data indicating that B cells are involved in ways that are not
completely understood - separate from production of antibodies. Also,
keep in mind that inflammation can be protective as well as
destructive, so all inflammation may not be causing problems, but
addressing them.

A "multi-hit" model considers a combination of immune abnormalities
required to produce disease, e.g. T cell, B cell, myeloid cell, etc.
imbalance.

Hollie Schmidt, Accelerated Cure Project
Genetics of MS

Hollie gave a presentation on genetics to an audience that included
one of the discoverers of the structure of DNA - Jim Watson. She
bravely fielded his question.

The content of her talk was a distillation of the genetics portion of
our Cure Map.

W. Richard McCombie, Cold Spring Harbor Laboratory
Next Generation Sequencing Technology and Applications

Sequencing technology is advancing almost too fast to keep up.
1980: 1kb per year
2008: 100,000,000 kb per year

Re-sequencing targeted areas of the genome is most efficient use of
CSHL Illumina machines. Whole genome sequencing not that useful due
to huge amount of noise. For small regions, long-range PCR.

Discussed various technologies used at CSHL for sequencing. Noted that
amount of data produced is impediment to analysis.

Sriram Subramaniam, Vanderbilt Stallworh Rehabilitation Hospital
Role of DNA damage pathway in MS

Eukaryotic organisms have developed ability to detect and repair
DNA-DSB (double strand breaks). ATM is a protein that recruits repair
proteins. People with mutations in ATM show increased sensitivity to
radiation, cancer, and immune disfunction.

Serendipitous observation in patient who underwent radiation therapy
showed limited ATM and this patient had a son with MS who also did not
have ATM. So looked at MS vs controls and found difference in ATM and
P53 (regulated by ATM).

People with low ATM showed mRNA for ATM. Appears that the protein is
being degraded rapidly in ATM- people. Don't know why yet. Certain
pathogens use ATM during their life cycle, thus this mutation might be
defensive, but results in a situation that can cause disease like MS.

Steven Jacobson, NINDS/NIH
Virus Associations and MS

List of viruses associated with MS is 19+ long. Steve currently
working on HTLV-1. Possible interpretations of list of viruses: none
of these are associated with MS, some (but not all) are, or the
associated agent has yet to be identified.

Exogenous model: disease due to infection, those without disease do
not have agent.

Ubiquity model: many are infected with the agent but the disease only
appears in a subset of patients.

HHV-6 found associated with MS in oligos when subtracting known
genetic sequences from samples. HHV-6 is a herpesvirus, 90% of
population exposed, causes roseola, can have neurological
complications. Two variants (A & B) are probably different
viruses. Variant A increases risk of neurovirulence, B is the one most
people have. Variant A prevalence is not known.

Problem: How to demonstrate association between ubiquitous agent and
clinical disorder? Association is not causation, however.

Clinical trials: antiviral compounds?
Shared Receptor hypothesis (CD46 used by a number of viruses)
Can HHV-6 be detected in the CNS in the absence of inflammation?
Virus can be in CNS tissue, but not CSF or serum

W. Ian Lipkin, Columbia University
Identifying viral pathogens in neurologic disease

Reviewed a series of methods for identifying viruses. Then gave
examples of diseases that present very differently depending at what
point in development the virus is introduced.

Discussion after talk led to claim the only way to identify a causal
ubiquitous pathogen is a treatment trial.

Autism study following people prospectively from before birth:
www.abc.columbia.edu

Colleen E. Hayes, University of Wisconsin - Madison
Vitamin D and MS

Vit D3 inhibits development of EAE (in females, not males), but not
established EAE. 1,25-(OH)2D3 inhibits EAE and reversal of established
disease. Evidence for Vit D role in MS due to prevalence gradient with
latitude. Ultraviolet light gradient with latitude could link with
Vit-D as it causes production of Vit D3 in the skin. It appears that
estrogen allows Vit D3 to be protective to EAE in females.

Small trial in RRMS subjects using 1,25 (and low calcium
intake due to hypercalcemia w/Vit D) and saw stable disability and
reduced relapses.

Without IL-10, Vit D3 is not protective. Some viruses can inhibit
IL-10.

Bianca Weinstock-Guttman, SUNY University of Buffalo
Gender and MS

"Sex" is biological term, "gender" is a psycho-social term, so
properly we should be considering "Sex and MS".

78% of roughly 80 medical conditions believed to be autoimmune are
more prevalent in women. Might be a consequence of reproductive
capabilities.

Female:Male ratio is increasing with time. No plausible confounders
identified. Effect independent of diagnostic changes.

Women have different immune response to infectious agents or antigens
than men. Hormonal differences: estrogens, progesterone, testosterone,
PRL, GH, ILGF-1, menstruation, pregnancy, menopause, andropause, birth
control. Genetic differences: XX, XY, X inactivation.

No clear difference in response to therapy between men and women,
although Copaxone in PPMS showed increased response in men. Female
twins more likely to be concordant than males.

David C. Mohr, Northwestern University
Stress and MS

Stress: stressful life event-> perceived stress-> symptomatic distress->
biomedical distubance.

Stress generally involves a perceived threat to something valued and a
sense that it can't be fully controlled. Generally about 50% of people
with MS think stress affects their MS.

A number of studies have shown that there is a trend to more
exacerbations when people feel stress. Some people may be predisposed
to perceiving stress, or to being able to cope with it.

Small study looking at IFN-gamma levels in people being treated with
therapy for depression and IFN-gamma dropped when depression
dropped. Those with more social support had more moderate depression.

Hypotheses: Hypothalamic-Pituitary-Adrenal (HPA) Axis - stress causes
increase in corticoids which might induce resistance to cortisol which
reduces inflammation.

Silke Schmidt, Duke Center for Human Genetics
Gene-environment interaction in age-related macular degeneration: A
success story

Age-related macular degeneration (AMD) has a number of associated
genes and has smoking as a risk factor. Silke described methodology
used to analyze the data to find the gene/environment interaction that
shows this relationship.

Reception and Dinner

Tue 02/05/08:

Brenda Banwell, Hospital for Sick Kids
Pediatric MS: insights into the disease

F:M ratio in pediatric MS under 10 is close to 1:1, but after 10 it
is predominantly female. Age also influences first presentation. Very
rare for pediatric MS to have EDSS > 3, however language processing and
visual memory problems can be severe.

Looking at kids with MS can help isolate environmental influences that
may trigger disease due to fewer influences prior to diagnosis.

80% of kids with MS have been exposed to EBV vs 50-60% of kids with
other diseases or healthy controls. Kids with MS also have a higher
antibody titer to EBV. No difference in seroprevalence of other
common childhood viruses. MS was not associated with number of
siblings, sibling order, or month of birth.

T-cell reactivity to a variety of self-antigens increased in CIS and
MS compared to diabetic and healthy controls.

Currently following 300 children at 23 sites prospectively from first
attack of demyelination.

Suzanne D. Vernon, CFIDS Association of America
Empiric delineation of patient heterogeneity to improve diagnosis and
intervention.

Chronic Fatigue Syndrome (CFS) identified 20 years ago as a chronic
viral infection (usually of EBV). Affects est. 4M in US, 4:1 F:M
ratio. No objective diagnosis, no treatments.

Process taken to get to objective diagnosis and treatments:

Series of meetings including novel pathogens, CNS/Immune/Endocrine
mechanisms, integration of disparate data and modeling
Launched CFS Computational Challenge w/5 teams of 5-6 people (each
with an MD, biologist, and computationalist)
Each team had same data and 6 months, results evaluated by expert
panel
Data from 227 people in 2 day study incl. epidemiological
data, nervous system data, sleep lab, endocrine/immune/genetic
data (million+ data variables)
Produced 16 papers that showed ability to isolate non-fatigued
from insuffiently fatigued (ISF) from CFS, subtypes, markers,
treatment hypotheses

Timothy Vollmer, Barrow Neurological Institute
MS Databases and Biorepositories: Challenges and applications

Overview of questions about MS such as is it really T cell mediated,
what role do B cells play, are the inflammatory and degenerative
phases separate diseases, etc.

Idea: collect materials from participants in studies of drugs that we
believe to be highly specific.

Repositories: COSTART, EDMUS, NY State MS Database, SLIFKA, NARCOMS,
MGH/Harvard, ACP, UCLA/Wadsworth VAH MS Brain Bank, RMMS Brain Bank

NARCOMS: 33K participants, 16.6K active, patient derived data

Need common definitions of data points collected
What is the core data that all databases need to have in common?
Patient vs. MD vs. interviewer collected data
Need to validate patient driven vs clinician driven
What are the key biological specimens? (BRAIN!)
What are the key controls?
How to get participants to contribute over the long term?

Alberto Ascherio, Harvard School of Public Health
Use of prospective cohorts in investigating causes of disease

How to "prove" causality between an almost ubiquitous virus and
disease? ("prove": provide sufficient evidence to justify some form of
action).

Something is "causally associated" with a disease when its
modification results in a change in disease frequency.

Two axes of classification of diseases: 1. by symptoms, signs,
pathology 2. causal: grouped according to some common prior
experience, judged to be of a causal nature.

EBV is different from most other viruses because:

more prevalent later in life than childhood
latitude gradient affects age of acquisition
more childhood infection in lower income children
these points overlap with MS epidemiology

Is EBV *infection* associated with a higher risk of MS? Need highly
specific and sensitive assays and large cohort. Strong association
found in recent study.

Alternative explanations:

artifact (false +/-): unlikely because of high Sn/Sp of
serology
reverse causation: unlikely because all women were EBV+ before MS
diagnosis, long latency between mono and MS, recent EBV infection not
seen in MS
confounding by genetic or environmental factor: could be, but
would have to be very strong to explain. EBV and pediatric MS can't be
explained by confounding by genetic facor

EBV titers prior to onset of MS significantly higher than those
without MS.

Recent paper detected more than 50% EBV infected B cells and plasma
cells in the brains of people with MS (which is tremendously higher
than usual infection rate outside of MS).

Epidemiological aspects of MS not explained by EBV: Faroe epidemics,
decreased risk in migrants from high to low risk areas (unless some
EBV strains more likely to cause MS than others).

Donald E. Ganem, UCSF
Discovering novel pathogens

Many acute infectious illnesses still lack etiologic agents (URI,
meningitis). Many diseases once thought to be noninfectious in fact
have a viral origin (liver cancer, cervical cancer, KS).

Traditional approaches: culture, serology, electron microscopy,
PCR. Limitations: insensitivity, requirement for foreknowledge, bias,
low throughput

UCSF solution: Virochip, contains 20K DNA sequences from all known
viruses. 1915 fully sequenced viruses currently.

You will find: pathogens, saprophytes, normal flora (recognized and
unrecognized), contaminants

How to tell them apart? Traditionally, Koch's postulates. Not useful
in practice in many instances. Solution: post-array epidemiology (use
array to find targets, epidemiology to measure risk and interventions
to test hypothesis.

To detect new pathogens, chip uses cross-hybridization of common
conserved 70mers that form the basis of the chip. Use conventional PCR
or array hybridization to physically isolate viral sequence and
amplify and sequence.

MS and viruses: from virologic perspective, MS is a distinctive and
rather stereotypic illness. Unlike many AI syndromes (SLE,
Sjogren's,ITP,etc.) no epidemiologic overlap with other known
inflammatory/AI states. All that favors a unitary precipitant. Biggest
problem is: is precipitant still around when demyelination begins?
Betting that copy number decline is rapid once immune attack starts
(HBV analogy).

Use early cases or fresh exacerbations. Brain biopsy from progressive
disease on off-chance that this is due to persistent viral antigen
(avoid: end-stage RRMS). Use methods capable of detecting low copy
numbers (Solexa?). Will need to be married to deep sequence evaluation
of the normal flora of CSF. If religious, pray to God that resulting
candidate viral genome not ubiquitous.

Estimated cost: $160K/yr x3 years plus provision of well-pedigreed
materials, full-time bioinformaticist, full-time postdoc/technician,
supply costs for deep sequencing, recognition that the effort will
likely fail.

E. Warwick Daw, WUSTL
Oligogenic modeling with environmental covariates for association and
linkage studies.

"To consult the statistician after an experiment is finished is often
merely to ask him to conduct a post mortem examination. He can perhaps
say what the experiment died of."
--Robert Fischer

Oligogenic Model is a parametric method to search over a large
statistical model space, allows for the effects of multiple genes and
covariate effects (e.g. environment). All effects are estimated from
the data, number of genes is estimated as well, allows for the effects
of unknown genes, effects add together but do not "interact", somewhat
tolerant of heterogeneity.

Mike Wigler, Cold Spring Harbor Laboratory
Analysis of large gene sets

Mike is looking at copy number variation in Autism. CNV good for
sporadic cases with strong genetic component and reduced fecundity, or
by somatic mutation in a subset of cells in the body (like cancer).

Douglas W. Bigwood, Diogenix, Inc.
Analysis of large gene expression data sets: strategies for
identifying novel pathogenic pathways

Gene chips give high dimensionality data (e.g. 50K-500K+ data
points). Need to collect samples properly, process them, QC results
from chip.

Common statistical methods produce far too many significant findings
by chance due to large size of data set. Can augment analysis with
existing annotation that is available for some genes. Can use
predictive modeling to learn biological relationships.

Terri Beaty, Johns Hopkins
Gene-environment Interaction with a special focus on MS:
Environment trumps genes or genes trump environment?

Genes: HLA, IL2, IL17, APOE...
Env: Smoking, Vit D, viruses

Biological GxE does exist - falls under ecogenetics, even
pharmogenetics. Ignoring GxE will obscure effects of genes.

Case Control: Case-only, case-family designs (case-related,
case-parent trios, case family tests), cohort designs (really
expensive).

In broad terms, GxE means phenotypic manifestation of genotype depends
on environment. In statistical terms, you must first specify a model
for main effects of G&E plus their combined effects (GxE).

Common requirements for case-control studies: representative of
affected individuals, sample of controls from same population. Look at
main effects of G & E separately, interaction exists if statistical
interpretation deviates from predicted combined effect.

Incomplete designs: case-only, partial case-control designs (partial
info on controls). Family based designs: related controls, case-parent
trio.

Case-parent trios allow to test for maternal vs paternal transmission,
or imprinting. Can't do that in all case-control designs.

Group Discussion
Developing a Protocol

Pre-MS -> CIS -> MS -> disease phenotype and outcome

Tue 02/05/08:

Group Discussion
Developing a Protocol

"The only true wisdom is in knowing you know nothing"
-- Socrates

Challenges:

define phenotype
disease duration
2 stages: trigger vs promulgation
practicality of rare disease studies:
- sample size
- funding

Ben Greenberg and Accelerated Cure Project will pull together a
high-level proposal for a study, circulate it to participants, and
create a detailed proposal for comment from everyone.

MS Organizations

MSNews Updates On Hold

Accelerated Cure Project Fall Newsletter Available

MSNews Summer Vacation

Accelerated Cure Project unveils new program to support volunteers

Starfish TV

Accelerated Cure Project is One Thousand Steps Closer to a Cure for MS

Accelerated Cure Project repository overview video

Accelerated Cure Project 2007 Annual Report Available

Accelerated Cure Project Spring Newsletter Available

Art's Notes From Cold Spring Harbor Labs Meeting 2/3/08