29Although the intricate and prolonged development of the human brain critically 30 distinguishes it from other mammals 1 , our current understanding of 31 neurodevelopmental diseases is largely based on work using animal models. Recent 32 studies revealed that neural progenitors in the human brain are profoundly different 33 from those found in rodent animal models [2][3][4][5] . Moreover, post-mortem studies 34 revealed extensive migration of interneurons into the late-gestational and post-natal 35 human prefrontal cortex that does not occur in rodents 6 . Here, we use cerebral 36 organoids to show that overproduction of mid-gestational human interneurons 37 causes Tuberous Sclerosis Complex (TSC), a severe neuro-developmental disorder 38 associated with mutations in TSC1 and TSC2. We identify a previously 39 uncharacterized population of caudal late interneuron progenitors, the CLIP-cells. In 40 organoids derived from patients carrying heterozygous TSC2 mutations, 41 dysregulation of mTOR signaling leads to CLIP-cell over-proliferation and formation 42 of cortical tubers and subependymal tumors. Surprisingly, second-hit events 43 resulting from copy-neutral loss-of-heterozygosity (cnLOH) are not causative for but 44 occur during the progression of tumor lesions. Instead, EGFR signaling is required 45 for tumor proliferation, opening up a promising approach to treat TSC lesions. Our 46 study demonstrates that the analysis of developmental disorders in organoid models 47 can lead to fundamental insights into human brain development and neuropsychiatric 48
disorders. 49Tuberous sclerosis complex (TSC) is a rare autosomal dominant disorder 50 characterized by pathological malformations in multiple organs 7 . Among those, brain 51 defects leading to severe neuropsychiatric symptoms like autism spectrum disorder 52 (ASD), intractable seizures and intellectual disability (ID) are most debilitating and 53 seen in the majority of patients 8 . Most patients have cortical tubers 9 , focal dysplastic 54 regions in the cortex that are diagnosed by MRI and consist of dysmorphic neurons 55 and giant cells. In addition, 80% of the patients display subependymal nodules 56 (SEN) that form along the lateral ventricle and can develop into subependymal giant 57 cell astrocytomas (SEGAs) in 10-15% of the patients 7,10 . It was thought that TSC 58 pathogenesis is initiated by constitutive mTOR activity resulting from inactivation of 59 the second allele 11 along the lines of the classic Knudson two-hit hypothesis of 60 tumorigenesis 12 . This is supported by existing mouse models and a spheroid model 61 for TSC, as characteristic brain alterations are observed exclusively in Tsc1 or Tsc2 62 homozygous mutant mice and spheroids [13][14][15][16][17][18] . Genetic analysis in patients, however, 63 revealed that loss of the second allele is frequent in SEN/SEGA, but rare in cortical 64 tubers 19-22 , conflicting with the two-hit hypothesis. In addition, the cellular origins of 65 cortical tubers and SEN/SEGAs remain unclear. Interesting...