Publications
Twelve papers, one question
Each entry says what problem the work addressed, what it found and why that matters. Author position is given for every entry. Five are first-author papers, across two research systems and two supervisors, and the most recent preprint is sole-authored. The complete record is on ORCID.
Key papers
Six papers that carry the argument: where drugs reach the developing brain, what they change in human neurons and in the embryo, the organoid model the next step builds on, and the newest work, a preprint on what DNA recorders can resolve in human brain organoids.
-
2026Zenodo preprint, October 2026Sole author
Molecular recorders promise to write the history of each cell into its DNA, but nobody had asked how much a recorder could capture in a growing human brain organoid. A model of cortical progenitors shows that recording capacity runs out early, half of it before day 27, and that a one-day exposure is 5.3 times more likely to be recorded on day 15 than on day 75, a ratio running from 1.5 to 6.6 across parameterisations and 2.4 under the spread matching measured clonal structure. For four one-day windows a signal-responsive channel would need a dynamic range near 1,700, against 11.5- to 22.6-fold in channels characterised so far. Editing rate, not array length, is the limit. A recorder can now be tested on paper before months are spent building it.
Zosen D. The writable window: division-coupled DNA recorders carry a built-in early bias in human cortical organoids. Preprint. Zenodo. 2026. doi:10.5281/zenodo.23375596
AMA style
-
2026ACS Chemical Neuroscience 17(16), 3048–3057Second author
Spinal cord organoids are grown from human stem cells to study the spinal cord and to prepare cell therapies, but nobody had shown whether their neurons work like real ones. Recording from single neurons between days 15 and 32, we found steady maturation: more negative resting potentials, larger currents, more refined action potentials and, early on, the firing pattern typical of motor neurons. The neurons also carried motor neuron and interneuron markers. The organoids are a credible model of human spinal cord development.
Ievglevskyi O, Zosen D, Requena Osete J, et al. Human iPSC-derived spinal cord organoids: the first electrophysiological assessment of neuronal maturation. ACS Chem Neurosci. 2026;17(16):3048-3057. doi:10.1021/acschemneuro.5c00823
AMA style
-
2025Neurochemistry International 190, 106056Second author
Antidepressants are used in pregnancy, yet their effect on the developing brain is poorly understood. In chicken embryos, both drugs reached clinical concentrations in the brain. Only venlafaxine raised 13 microRNAs in the cerebellum, including miR-92, whose human counterpart miR-363 also rose in human neurons, and only venlafaxine cut neurite outgrowth. Both drugs thinned the cerebellum's germinal layer. Escitalopram and venlafaxine do not act alike, and miR-92 and miR-363 are a molecular lead that holds across species.
Kaplan-Arabaci O, Zosen D, Domanska D, et al. The antidepressant venlafaxine induces upregulation of miR-92 and histological alterations in a chicken embryo model. Neurochem Int. 2025;190:106056. doi:10.1016/j.neuint.2025.106056
AMA style
-
2023Neurochemistry International 169, 105571First author
Antidepressants are used in pregnancy, but their direct effects on developing human neurons were unknown. In purely neuronal human cultures, escitalopram and venlafaxine raised BDNF promoter activity in a dose-dependent way, a result that looks reassuring, while venlafaxine reduced neurite outgrowth. A single marker would have missed that. Safety testing needs several readouts in the same cells.
Zosen D, Kondratskaya E, Kaplan-Arabaci O, Haugen F, Paulsen RE. Antidepressants escitalopram and venlafaxine up-regulate BDNF promoter IV but down-regulate neurite outgrowth in differentiating SH-SY5Y neurons. Neurochem Int. 2023;169:105571. doi:10.1016/j.neuint.2023.105571
AMA style
-
2022Neurotoxicology and Teratology 90, 107057First author
Valproate in pregnancy raises the risk of autism. Lamotrigine is considered safer, but recent data suggest some risk, and the reasons were unclear. In chicken embryos, both drugs lowered PAX6, a transcription factor that steers cerebellar development, and MMP-9, and reduced the density of PAX6-positive cells. Valproate also reduced cerebellar mass and cell proliferation. PAX6 is a candidate shared mediator, which gives researchers a mechanism to examine for the autism-linked effects of both drugs.
Zosen D, Austdal LPE, Bjørnstad S, Lumor JS, Paulsen RE. Antiepileptic drugs lamotrigine and valproate differentially affect neuronal maturation in the developing chick embryo, yet with PAX6 as a potential common mediator. Neurotoxicol Teratol. 2022;90:107057. doi:10.1016/j.ntt.2021.107057
AMA style
-
2021Journal of Pharmacological and Toxicological Methods 112, 107105First author
Finding out where a drug goes in the developing brain normally needs pregnant rodents, which is slow, costly and hard to stage precisely. Injected into the egg, valproate and lamotrigine both reached the embryo's brain within 5 to 15 minutes at clinically relevant concentrations, and lamotrigine levels were higher when dosed at day 13 than at day 16. The model is fast, inexpensive and needs no pregnant animals, and it tells you what dose reaches the brain before you interpret a toxicity result.
Zosen D, Hadera MG, Lumor JS, Andersen JM, Paulsen RE. Chicken embryo as animal model to study drug distribution to the developing brain. J Pharmacol Toxicol Methods. 2021;112:107105. doi:10.1016/j.vascn.2021.107105
AMA style
All other publications
Eight more: earlier papers, collaborations and preprints, with filters
-
2022Heliyon 8(8), e10291Second author
Chordoma with tuberous sclerosis complex is so rare that this is only the 14th case on record since 1975, so treatment guidance barely exists. Sequencing the tumour found loss of both TSC1 copies, chromosomal instability, involvement of epigenetic regulators, and amplification of CBX7 and APOBEC3 genes, which had not been seen in chordoma before. A gain of KRAS suggests that EGFR-blocking drugs alone may not work. The findings nominate targets for personalised therapy and argue against one option.
Anoshkin K, Zosen D, Karandasheva K, et al. Pediatric chordoma associated with tuberous sclerosis complex: a rare case report with a thorough analysis of potential therapeutic molecular targets. Heliyon. 2022;8(8):e10291. doi:10.1016/j.heliyon.2022.e10291
AMA style
-
2022Toxicology and Applied Pharmacology 449, 116130Contributing author
Studies link long-term paracetamol use in pregnancy to ADHD in the children, but how the drug might act on developing neurons was unknown. In chicken and human neurons grown in the dish, paracetamol at concentrations up to 1,600 µM reduced neurite branching and the cytoskeletal protein β2-spectrin, and made β3-tubulin clump. Human cells also migrated less. Damage to the neuron's internal scaffold is a mechanism worth testing against the epidemiological signal.
Labba NA, Wæhler HA, Houdaifi N, et al. Paracetamol perturbs neuronal arborization and disrupts the cytoskeletal proteins SPTBN1 and TUBB3 in both human and chicken in vitro models. Toxicol Appl Pharmacol. 2022;449:116130. doi:10.1016/j.taap.2022.116130
AMA style
-
2021Toxicology Letters 338, 85–96Contributing author
Persistent pollutants such as PFOS are suspected to harm the developing brain. We exposed neuron-like PC12 cells to a mixture at the levels measured in Scandinavian blood. It did not block neurite growth. It increased neurite length induced by nerve growth factor, PFOS alone accounted for about half of that effect, and measures of nuclear and mitochondrial health did not change. Developmental neurotoxicity can appear as changed growth and not as cell death, so screens that only count dying cells can miss it.
Yadav A, Amber M, Zosen D, et al. A human relevant mixture of persistent organic pollutants (POPs) and perfluorooctane sulfonic acid (PFOS) enhance nerve growth factor-induced neurite outgrowth in PC12 cells. Toxicol Lett. 2021;338:85-96. doi:10.1016/j.toxlet.2020.12.007
AMA style
-
2021OSF preprintPreprint, contributing author
A scoping review that sets out to identify and summarise the clinical and basic research on drugs and biopharmaceuticals proposed for engineered negligible senescence, the deliberate prevention of age-related decline. It gives researchers one place to see what has been tested.
Zhelnov P, Parfenenkova A, Zosen D, et al. Pharmaceuticals and biopharmaceuticals for engineered negligible senescence in humans: a scoping review of clinical and basic research. OSF Preprints. 2021. doi:10.17605/OSF.IO/7MFJ8
AMA style
-
2019Neuroscience and Behavioral Physiology 49(6), 765–772Second author
Epilepsy is linked to impaired neurogenesis, and it was unclear whether faulty neurogenesis causes seizures or follows them. Neural stem cells from seizure-prone Krushinskii–Molodkina rats divided less and differentiated faster than cells from control rats, mostly into glutamatergic and catecholaminergic neurons. That supports aberrant neurogenesis as a genetically determined cause of epilepsy, not only a consequence of it.
Saparova VB, Zosen DV, Nasluzova EV, et al. Analysis of neural stem cells from Krushinskii–Molodkina rats, which have a genetic predisposition to audiogenic seizures. Neurosci Behav Physiol. 2019;49(6):765-772. doi:10.1007/s11055-019-00799-9
AMA style
-
2018Neuroscience Letters 684, 6–12First author
How kinase signalling controls dopamine release was contradictory in the literature. Blocking ERK1/2 in nerve growth factor-differentiated PC12 cells made dopamine build up in the medium within four hours, as PKG activity rose and synapsin I phosphorylation and SNAP25 fell. ERK1/2 acts as a brake on catecholamine release, and the balance between PKG and ERK1/2 helps set how much dopamine a neuron secretes.
Zosen DV, Dorofeeva NA, Chernigovskaya EV, Bachteeva VT, Glazova MV. ERK1/2 inhibition increases dopamine release from differentiated PC12 cells. Neurosci Lett. 2018;684:6-12. doi:10.1016/j.neulet.2018.06.056
AMA style
-
2017Russian Journal of Genetics: Applied Research 7(3), 217–225Contributing author
How a seizure starts in a brain with a genetic predisposition to seizures was largely unexplored. In seizure-prone Krushinsky–Molodkina rats, audiogenic seizures raised ERK1/2 activity and synapsin I activation in the striatum and substantia nigra, which drove glutamate release. At a later stage, dopamine release and receptor changes appeared that could restrain the seizure. The map shows ERK1/2 signalling as a control point in both the drive and the brake.
Dorofeeva NA, Nikitina LS, Zosen DV, Glazova MV, Chernigovskaya EV. Functional state of the nigrostriatal system of Krushinsky–Molodkina rats during audiogenic seizure expression. Russ J Genet Appl Res. 2017;7(3):217-225. doi:10.1134/S2079059717030029
AMA style
-
2016Neuroscience and Behavioral Physiology 46(5), 559–565First author
p53 is known as a cell-death protein, and its role in how neurons differentiate was unclear. Activating p53 with nutlin-3 sped up the differentiation of PC12 cells, with longer neurites and more tyrosine hydroxylase, through the cRaf, ERK and CREB cascade. A protein known for stopping cells also helps neurons mature, so drugs that act on p53 may affect neuronal development.
Zosen DV, Glazova MV. The role of the interaction of p53 and the MAPK cascade in controlling neuronal differentiation in the PC12 cell line. Neurosci Behav Physiol. 2016;46(5):559-565. doi:10.1007/s11055-016-0277-y
AMA style
In the press
Science communication in Norwegian, on translational neurobiology, experimental models and the reduction of animal experiments.
- Titan · UiOStuderer hjernen med modellar for å finna årsaker til alvorlege og utbreidde sjukdommar2023
- Titan · UiOVanleg epilepsimedisin påverka hjerneutviklinga til kyllingfoster2022
- Titan · UiONyttig å sjå forskinga si frå industrien si side2021
- forskning.noForskarar vil erstatta dyreforsøk med hønseegg2021
- Journal coverBiochemistry and Cell Biology, 2020: cover image of a neurosphereConfocal image from our work