2. Multi-omics
signatures of the
early life exposome
and
Climate factors
Léa Maitre et al.
Courtesy of Zoe – March 2023
Ultra-processed food
Pet allergens
and microbes
UV light
Temperature
Combustion
smoke
Natural
spaces
3. Early Life
Vulnerable periods of rapid organ development
Many chronic diseases have part of their origin in
childhood
Lifetime influence
Effective prevention if programming can be reshaped
5. Initial promises
Integrating omics in birth cohort studies will allow:
to understand biological mechanisms and prove causality:
-> Bradford Hill’s guidelines for causality assessment in epidemiology, i.e.
biological plausibility
to predict past exposures (e.g. from pregnancy)
to predict future health outcomes
6. Exposome cohort follow up
The Exposome in Early Life – 1.Implementation of the exposome in birth cohorts
Pregnancy Birth/Infancy Childhood
6-11 yrs
Health
phenotypes
Adolescence
12-18 yrs
Vrijheid et al. 2022
7. Outline
1. Building an exposome-omics association atlas
2. Follow-up developments to improve causality, health
prediction and open science
3. Climate and child health
9. Building the early life exposome and the multi-omics
phenotypes in HELIX children
Maitre, Bustamente, et al. Nature Communication 2022
Léa Maitre – ISGlobal
10. Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
12. Childhood
Pregnancy
Maternal tobacco smoking shows robust associations and long-lasting effects on
the child methylome …
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
Vives-Usano et al. BMC Med 2020
13. Childhood
Pregnancy
Maternal tobacco smoking shows robust associations and long-lasting effects on
the child methylome …
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
Vives-Usano et al. BMC Med 2020
A clear dose-response
14. Childhood
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
… and novel signatures for prenatal cadmium and indoor air pollution are
detected
Prediction models
based on indoor
measurements by
sensors and
questionnaire data
NO2
TEX
Benzene
PM10
PM2.5abs
Branched amino acids (leucine, valine)
Carnitines
Higher BMI observed in children (Vrijheid 2023)
15. The serum and urinary metabolome reveal principal dietary routes of
exposure to chemical pollutants
▪ Pesticides: DMTP DETP DMP DMDTP
with fruit intake (proline betaine and
hippurate)
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
Arsenic, Hg, PFUNDA with TMAO and fish intake
markers such as PUFA phosphocholine lipids
16. The serum and urinary metabolome reveal principal dietary routes of
exposure to chemical pollutants
▪ Pesticides: DMTP DETP DMP DMDTP
with fruit intake (proline betaine and
hippurate)
▪ Arsenic, Hg, PFUNDA with TMAO and
fish intake markers such as PUFA
phosphocholine lipids
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
Pesticides: DMTP DETP DMP DMDTP with fruit intake
(proline betaine and hippurate)
17. Open-access webserver – “online supplements”
• Open-access webserver to facilitate future research and meta-analyses.
https://helixomics.isglobal.org/
18. Outline
1. Building an exposome-omics association atlas
2. Follow-up developments to improve causality, health
prediction and open science
3. Climate and child health
20. Childhood exposure to non-persistent endocrine
disrupting chemicals and multi-omic profiles
-> Repeat sampling -> Multi-omics
-> Gaussian Graphical Models
Network integration
Fabbri L. et al. Environment International 2023
21. Fabbri L. et al. Environment International 2023
Multi-omic
signatures
Multi-omics
Multi-omics
Multi-omic signatures
Childhood exposure to non-persistent endocrine
disrupting chemicals and multi-omic profiles
23. Leveraging machine learning to compute
exposome risk scores in children
XGBoost: Extreme gradient
boosting decision trees
Manuscript under review Guimbaud et al.
~15% ~45%
~5%
Mental health: local explanations (SHAP)
HELIX Exposome
+ extra
variables and
omics
To predict health outcomes
26. Ressources
Exposome data challenge
Dataset: https://github.com/isglobal-exposomeHub/ExposomeDataChallenge2021/blob/main/README.md
Data description: https://docs.google.com/document/d/1ul3v-sIniLuTjFB1F1CrFQIX8mrEXVnvSzOF7BCOnpQ/edit
Scientific publication Maitre et al 2022 Env. Int.
Slides and videos of the presentations https://www.isglobal.org/-/exposome-data-analysis-challenge
https://www.youtube.com/channel/UC0F3hR04UzUeKkcfAyikltA/featured
Code used shared on GitHub: https://github.com/isglobal-
exposomeHub/ExposomeDataChallenge2021/tree/main/R_Code_Presentations
HELIX project
Data inventory: https://www.projecthelix.eu/index.php/es/data-inventory
Tamayo-Uria I, et al. The early-life exposome: Description and patterns in six European countries. Environ Int. 2019.
Maitre L, et al. Human Early Life Exposome (HELIX) study: a European population-based exposome cohort. BMJ Open. 2018.
Vrijheid M, et al. The human early-life exposome (HELIX): project rationale and design. Environ Health Perspect. 2014
Léa Maitre - ISGlobal - Exposome data challenge
Publicize methods
for developers
Code source for
analysists
27. Outline
1. Building an exposome-omics association atlas
2. Follow-up developments to improve causality, health
prediction and open science
3. Climate and child health
30. Biological insights of climate impact on health?
There are no studies that systematically assessed this
Pregnancy heat exposure and epigenetic modification
found in cord blood, placenta tissue
Direct effect of heat
Direct effect of cold temperature
31. Weather factors in HELIX children
Temporal variability in
• Temperature
• Humidity,
• Cloud coverage
• Atmospheric pressure
• UV radiation
averaged 1 month before sampling
UV radiation at home at 0.5 x 0.5 degree resolution were obtained from the Global Ozone Monitoring Experiment onboard the
ERS-2 (European Remote Sensing) satellite (Temis)
Daily measurements of temperature and humidity were obtained from a local weather stations
32. Childhood meteorological variables are associated with multiple
molecular layers
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
33. Childhood
Proteins
Adiponectin increases with humidity (winter)
and decreases with UV (summer) (Wei et al. 2017;
Imbeault et al. 2009; Jankovic et al. 2013).
Adiponectin -/- Thermogenesis
120 min of cold exposure Adiponectin
Weather factors and adiponectin: a key regulator of thermogenesis
Léa Maitre – ISGlobal Maitre, Bustamente, et al. Nature Communication 2022
34. Childhood
Weather and infectious disease markers
Léa Maitre – ISGlobal
• CpGs associated with
weather conditions
overlapped with CpGs
reported for infections
• Genes related to
temperature were enriched
for cellular response to
type I interferon (antiviral
response)
• Infectious diseases follow
seasonal patterns and are
more prevalent under
particular meteorological
35. Weather and sleep deprivation and circadian rhythm markers
• Serum metabolites
associated with
meteorological variables
included taurine,
asymmetric
dimethylarginine (ADMA),
acylcarnitine C5, and
serotonin
• Related to sleep
deprivation, circadian
rhythm and in the aetiology
of depression
Childhood
36. Weather and BPA exposure?
• Associations between some
molecular markers (carnitine
C5, adiponectin, serotonin)
and weather conditions were
attenuated by adjusting for
exposure to bisphenol A
(BPA).
• BPA was previously found to
reduce adiponectin release.
Hugo et al EHP 2008
37. • Coordination: Martine Vrijheid, Oliver Robinson, Léa Maitre
• Cohort (PIs): John Wright, Remy Slama, Jordi Sunyer, Regina Gražulevičienė,
Leda Chatzi
• Exposome: Cathrine Thomsen, Mark Nieuwenhuijsen,. Maribel Casas
• Methylation: Mariona Bustamante, Ángel Carracedo
• Transcriptomics: Marta Vives, Xavier Estivill
• Proteomics: Eva Borras, Eduard Sabidó
• Metabolomics: Chung Ho Lau, Alexandros P Siskos, Hector Keun
• Data analysis: Carles Hernández, Carlos Ruiz, Jose Urquiza, Juan R González
ACKNOWLEDGEMENTS
Léa Maitre - ISGlobal
Funding:
FP7 and H2020 European program
Juan de la cierva – Spanish ministry
fellowship
38. Omics and Endocrine Disruptor Chemicals (EDCs):
a scoping review of existing studies
Maitre, Jedynak, et al. Env Res 2023
• Focus on “non-persistent” EDCs (excluding organochlorine pesticides
such as DDT, PCBs,...)
• 98 studies identified (2004-2021) out of 1411 papers screened
39. • Current studies lack power to detect subtle effect of ED, as expected
in general population. This effect is likely even more diluted based on
the small sample size, high exposure misclassification due to high ED
intra-individual variability.
• We recommend to use stronger study design with pooled repeat
samples for exposure monitoring, targeted omics based on biological
hypothesis in relevant tissues (e.g. placenta), or even pooled meta
analyses across cohorts to increase sample size.
Omics and Endocrine Disruptor Chemicals (EDCs):
a scoping review of existing studies
Manuscript in preparation Maitre, Jedynak, et al.
40. Open-access webserver – “online supplements”
• Already exist for other omics: EWAS atlas
• Challenge for metabolomics: different nomenclature for metabolites,
different platforms and coverage.
• Essential
• Provide clinical context
• Biological interpretation
• Systematic evidence maps and regulation
Example: https://omicscience.org/apps/mwasdisease/)
41. Limitations of the concept: viewed by
sociologists
• Holistic?
Limitation in coverage
A focus on molecular effects of physical environment, limiting
the (broad) integrative approach
“(…) there is a very real risk that exposome research could molecularize complex social phenomena
reducing the social experiences that condition population-level variations in exposures to individual-level
molecular-level differences.” (Senier et al. 2017)
• More precise?
One specific mechanistic vision of causation. A unidirectional and linear vision (e.g. from the social to the
biological):
back to the “chain of causation” model?
what about interaction and reverse causation ? i.e. biological factors influence social behaviours -> e.g.
prenatal smoking
42. Pierce et al 2016 “Causal inference—so much
more than statistics”
Causal inference with omics, the case of prenatal smoking
Front. Genet., 2020
Sec. Epigenomics and Epigenetics
https://doi.org/10.3389/fgene.2020.00903
43. First mention
of the
exposome
Wild CEBP
Commentary
Rappaport
& Smith
Science
Perspective
Phenome
center
(Imperial
College
London)
US program
NIEHS Center
US program
NIEHS RFA
CHEAR->
children
US program
NIEHS
Human Health
Exposure Analysis
Resource (HHEAR)
European
programs
H2020
Exposome research: a short history
Léa Maitre – ISGlobal
National hubs: e.g.
France exposome
IREINE - EU
infrastructure
44. First mention
of the
exposome
Wild CEBP
Commentary
Rappaport
& Smith
Science
Perspective
Phenome
center
(Imperial
College
London)
US program
NIEHS Center
US program
NIEHS RFA
CHEAR->
children
US program
NIEHS
Human Health
Exposure Analysis
Resource (HHEAR)
European
programs
H2020
Exposome research: a short history
Léa Maitre – ISGlobal
National hubs: e.g.
France exposome
IREINE - EU
infrastructure
European Cluster to
Improve
Identification of
Endocrine Disruptors
NEW TESTING AND
SCREENING
METHODS TO
IDENTIFY ENDOCRINE
DISRUPTING
CHEMICALS (EDCS)
45. Exposures during critical fetal and
infant periods lead to
developmental adaptations, which
predispose individuals to
development of chronic diseases in
later life.
46. Follow up in vitro and in vivo studies
90% of serotonin produced in intestinal tract. Levels are regulated by gut
microbiota determines levels in colon and blood. Serotonin is made by colonic
enterochromaffin cells, which supply 5-HT to the mucosa, lumen, and circulating
platelets activates platelet function and myenteric neurons increases GI
motility and platelet function across body. Peripheral serotonin dysregulation
Gut-Microbiota-Brain Axis and Its
Effects on BBB permeability
47. Lessons learnt
Any omics may vary across time within an individual, not always
relevant to characterize long term exposure effect
Cross-omics integration can be strongly confounded by shared
determinants (BMI, sample handling)
Need for validation: triangulation of evidence
Biological context for omics markers in early-life studies is
missing (most data come from cancer or late life NCD studies)
Importance of temporality and socio-economical context