Establishment of Novel Cell Lines Derived from Renal Erythropietin-Producing Cells

Abstract

Erythropoietin (Epo) is the principal humoral factor regulating red blood cell homeostasis by directly influencing the survival, proliferation and differentiation of erythrocyte progenitors in the bone marrow. Epo is almost exclusively regulated on the transcriptional level, mainly by the transcription factor hypoxia-inducible factor (HIF)-2α. HIF-2α stands at the centre of the cellular oxygen sensing machinery: under conditions of sufficient oxygen, it is constantly hydroxylated by the prolyl hydroxylase enzymes PHD1, PHD2, and PHD3 as well as by the asparagine hydroxylase factor inhibiting HIF (FIH), ultimately resulting in its proteasomal degradation. PHD2 and PHD3 both are induced in hypoxia, which serves as a negative feedback regulation of HIF activity and therefore also Epo expression. The main source of Epo in the human adult are peritubular specialized renal Epo-producing (REP) cells located at border of the kidney cortex and medulla. REP cells are of mesenchymal origin with fibroblast properties, and have a neuron-like phenotype with long, thin, mitochondria-containing protrusions, which is reminiscent of the recently described telocyte cell type. Active REP cells are rare in number, present with a telocyte-like phenotype, and do not possess a known marker expression pattern clearly distinguishing them from other kidney cells. Upon a reduction in tissue oxygenation resulting in local hypoxia, HIF-2α in REP cells is stabilized, resulting in the de novo transcription, synthesis, and secretion of Epo. In diseases such as chronic kidney disease (CKD), production of Epo is impaired, resulting in a pathological reduction of erythrocyte production. This reduction in blood oxygen carrying capacity, also called anaemia, results in an undersupply of oxygen and thus to complications in organs throughout the body. It is suggested that in this condition, REP cells either differentiate into myofibroblasts, thereby losing their Epo-producing capabilities, or that their ability to sense changes in blood oxygenation is hampered e.g. due to changes in tissue composition and structure. The mechanisms by which REP cells lose their capability to produce Epo in response to hypoxia are not well understood. This is owed to the fact that up to date, the generation of a REP cell-derived cell line still capable of hypoxic Epo expression has not been achieved. While several groups have attempted to, all of the published cell lines mostly seized Epo production briefly after the start of in vitro culture. As a tool to investigate the fate of REP cells, we have generated a novel mouse model, expressing a tamoxifen-inducible Cre recombinase under the control of Epo-regulatory elements. We then crossed this Epo-CreERT2tg/tg mice with both, a reporter mouse expressing tdTomato downstream of a floxed transcriptional STOP cassette, as well as further crossing it with a mouse expressing a floxed diphtheria toxin receptor to facilitate enrichment. In this manner, we are able to efficiently and reproducibly label actively Epo-producing cells, extract them from the kidney and to take them in culture. We have generated a number of REP cell derived (REPD) cell lines, either with constitutive or conditional SV40 large T immortalization. While these cells present with sporadic hypoxic Epo expression, its degree and reliability are not sufficient for downstream applications, such as investigating the genetic regulation of Epo by HIF-2α, or elucidating the intricacies of its negative feedback regulation, which might be one of the reasons for the transient nature of Epo expression. It has been suggested that cellular quiescence is required for REP cells to produce Epo in response to hypoxia. Indeed, AB-REPD2-22 cells, which were immortalized using a temperature-labile mutant form of the SV40 large T antigen, stop proliferation when cultivated at 37°C and show marked changes in marker expression. However, also long-term culture at 37°C does not restore hypoxic Epo induction. Still, we managed to obtain several interesting insights from studying REPD cells. First, interfering with the negative feedback loop of HIF signalling is not sufficient to restore hypoxic Epo induction in REPD cell lines in vitro. Second, when cultivated at non-permissive temperatures, REPD cells have an increased expression of mesenchymal stem cell (MSC)-associated markers, such as Sca1 or Cd81. This is in line with recent findings suggesting Sca1 as a marker for a REP cell population. REPD cells also present a certain potential for neurogenic differentiation. This is in support of their mesenchymal origin, as well as their neuron-like phenotype and their plasticity observed in vivo as shown e.g. by their myofibroblast differentiation potential during CKD. Third, we realized that some of the generated REPD cells underwent only incomplete Cre-mediated recombination of the floxed STOP cassette in vivo, which resulted only in partial tdTomato fluorescence. This was reproducible in vitro and demonstrates that experiments involving the recombination of this widely used reporter construct need to be evaluated and controlled carefully. Fourth, REP cells as well as REPD cells have characteristic long, thin, mitochondria-containing cellular protrusions which appear to increase in number upon hypoxic exposure. There also appears to be a shedding and perhaps even an uptake of mitochondria-containing membrane compartments. However, at least in vitro, these cellular protrusions or the mitochondria contained within do not seem to be relevant for oxygen sensing. To summarize, while not suitable to answer open questions regarding the hypoxic regulation of REP cells, REPD cells still present a useful and interesting tool to investigate various aspects of this enigmatic cell type. This includes, among others, the mechanisms behind the transient nature of Epo expression, their surprising plasticity, and the role of cellular projections in oxygen sensing.