STUDY QUESTION Does the endometrial functionalis have the potential to undergo self-renewal after menstruation and how is this process controlled by ovarian steroids? SUMMARY ANSWER Endometrial xenografts subjected to withdrawal of estradiol and progesterone shrink but also show signs of proliferation and tissue repair; new estradiol supply prevents atrophy but is not sufficient to increase graft volume. WHAT IS KNOWN ALREADY Menstruation, i.e. cyclic proteolysis of the extracellular matrix of endometrial functionalis, is induced by a fall in estrogen and progesterone concentration and is followed by tissue regeneration. However, there is debate about whether regenerating cells must originate from the basalis or from stem cells and whether new estrogen supply is required for the early repair concomitant with menstruation. STUDY DESIGN, SIZE, DURATION Fragments from human endometrial functionalis (from 24 hysterectomy specimens) were xenografted in ovariectomized SCID mice and submitted to a 4-day estradiol and progesterone withdrawal (to mimic menstruation) followed by re-exposure to estradiol (to mimic the proliferative phase). We measured signs of proliferation and changes in graft volume. PARTICIPANTS/MATERIALS, SETTING, METHODS Endometrium was collected from spontaneously cycling women. Cell proliferation was examined by immunolabeling Ki-67, cyclin D1 and phosphorylated-histone H3. Xenograft volume was measured by magnetic resonance imaging. Xenograft histomorphometry was performed to determine how the different tissue compartments contributed to volume change. MAIN RESULTS AND THE ROLE OF CHANCE Hormone withdrawal induced a rapid decrease in graft volume mainly attributable to stroma condensation and breakdown, concomitant with an increase of proliferation markers. Reinsertion of estradiol pellets after induced menstruation blocked volume decrease and stimulated epithelial and stromal growth, but, surprisingly, did not induce graft enlargement. Reinsertion of both estradiol and progesterone pellets blocked apoptosis. LIMITATIONS, REASONS FOR CAUTION Mechanisms of endometrial remodeling are different in women and mice and the contribution of circulating inflammatory cells in both species remains to be clarified. Moreover, during human menstruation, endometrial fragments resulting from tissue proteolysis can be expelled by the menstrual flow, unlike in this model. WIDER IMPLICATIONS OF THE FINDINGS Menstruation is a multifocal event within the functionalis. This is the first evidence that endometrial fragments that are not shed after menstrual tissue breakdown can support endometrial regeneration. Endometriosis is commonly thought to result from the retrograde migration of menstrual fragments of the degraded functionalis into the peritoneal cavity. Our study supports their potential to regenerate as ectopic endometrium. STUDY FUNDING/COMPETING INTERESTS This work was supported by the Fonds de la Recherche Scientifique Médicale, Concerted Research Actions, Communauté Française de Belgique, Région wallonne, Région bruxelloise and Loterie nationale. P.H. and B.F.J. are research associates of the Belgian Fonds de la Recherche Scientifique (F.R.S.-F.N.R.S.). E.M. is Associate Editor at Human Reproduction. There is no conflict of interest to declare.