Zosen LabNCMBM · University of Oslo

Developmental neurotoxicology · Human organoid models

Are sensitive periods in brain development intervals of time, or transient cell states?

Every judgement about whether a medicine is safe at 22 weeks of pregnancy but not at 28 assumes that vulnerability can be located on a calendar. In human tissue, that assumption has never been measured. We are building the tools to measure it.

The research Publications

Each line is one cell over 150 days of human differentiation. Teal marks are written while it divides, orange marks after it stops. Dotted lines are the days a drug is given.

Principal investigator
Denis Zosen, PhD, MBA
Host
NCMBM, University of Oslo
Models
Human iPSC organoids · chicken embryo
Status
Programme in formation, 2026

Three questions

A drug reaches the developing brain within minutes. Whether it does harm appears to depend less on the week of pregnancy than on what the individual cell was doing at that moment. These three questions test that idea in order, and each one has to pass before the next begins.

  1. 01

    We are engineering human neural cells to keep a molecular logbook: a genomic tape that gains an ordered mark each time a defined programme switches on. Reading the tape at the end of an experiment returns, cell by cell, which programmes ran and in what order.

    Tested against a known schedule of drug pulses before anything else proceeds

  2. 02

    The same exposure, at the same gestational age, in the same health system, can produce opposite signals in different populations. We test head to head whether susceptibility is better explained by the date of exposure or by the state the cell was in, in the same cells, with lineage and cell type held constant.

    Two exposures: a synthetic glucocorticoid, timed to the hour; an organophosphate, chronic and undatable

  3. 03

    If a brief cell state really carries the risk, then intervening inside that state should blunt the damage and intervening outside it should do nothing. That asymmetry is the test, and it is also the first hint of where a protective intervention could act.

    Normal differentiation, cell proportions and survival must be preserved

How the question was arrived at

One line of enquiry, carried across four countries, two model systems and two research groups. Each step raised the problem the next one had to solve.

  1. 2016 – 2018

    The direction of a drug effect depends on the state of the cell

    Work on p53 and the MAPK cascade placed p53 at a defined position in the programme controlling neuronal differentiation. Blocking ERK1/2 then raised dopamine release: the same cascade that drives a cell to differentiate also restrains what it secretes.

    Sechenov Institute, St Petersburg · first author on both
  2. 2017 – 2018

    Learning to build human tissue

    Wnt signalling in central nervous system development at Biocenter Oulu, on funding I secured myself, adding CRISPR editing and iPSC-derived organoid culture. A short mission at Oslo University Hospital added cerebral organoid and spinal cord production.

    University of Oulu · Oslo University Hospital · own grants, ~€31,500
  3. 2021

    Exposure has to be resolved by developmental stage, not by dose

    I designed a chicken embryo model to ask where drugs actually go in the developing brain. They arrive within minutes at clinically relevant concentrations, and how far they penetrate depends on the stage of development. The group went on to use the model for further studies.

    University of Oslo · first author · my own methodological contribution
  4. 2022 – 2023

    One endpoint can point the wrong way

    Two antiepileptics with different morphological signatures converged on a single developmental transcription factor, PAX6. Then two antidepressants raised the reassuring molecular marker while reducing neurite outgrowth — for one drug and not the other. A single-endpoint study would have drawn the opposite conclusion.

    University of Oslo · first author on both
  5. 2025 – 2026

    Carrying one question across unrelated systems

    A drug-specific structural effect reproduced across a human cell line and an embryonic model, with a microRNA emerging as the bridge between them. In parallel, the first electrophysiological assessment of neuronal maturation in human iPSC-derived spinal cord organoids, where I was responsible for the differentiation the study rests on.

    University of Oslo · second author, co-conceptualisation and design

Track record

Figures as recorded in Scopus on 21 September 2026. The full list, with a note on my role in each paper, is on the publications page.

12Peer-reviewed papers
5As first author
73Citations · h-index 5
~€100KResearch funding won as applicant

Recent

  • Spinal cord organoid paper published in ACS Chemical Neuroscience — the first electrophysiological assessment of neuronal maturation in this system.
  • Guest Researcher at the Norwegian Centre for Molecular Biosciences and Medicine, the Norwegian node of the Nordic EMBL Partnership.
  • Returned to the Global Biotech Revolution forum, University of Cambridge, as one of 100 selected global leaders.
  • Venlafaxine and miR-92 paper published in Neurochemistry International.

Working with people outside the lab

The programme needs exposure biomarker chemistry, obstetric registry data and single-cell infrastructure. Those partnerships give access and specialist expertise; the hypothesis and the experimental architecture stay here.

Open to new collaborations — get in touch

Replacing animal experiments

Two of my first-author papers exist because pregnant patients are excluded from trials and the evidence has to come from models that do not use mammals. The current programme works entirely in human tissue.

In the press

Science communication in Norwegian, on translational neurobiology, experimental models and the reduction of animal experiments.