Many regenerative medicine approaches are based on the reprogramming of somatic cells to iPS cells, followed by directed differentiation to the desired cell type. This is a powerful possibility, in the decade since the first publication on iPS cells the reprogramming has become pretty much routine, and the directed differentiation protocols are rapidly becoming more efficient, more specific and more diverse. Differentiation towards renal fates is no exception to this. However, this method of generating cells of a specific type is, of course, indirect, so will take longer and every step needed that does not have 100% efficiency (and what in stem cell biology is 100% efficient?) will lead to reduced overall efficiency. Moreover, going via a fully pluripotent state followed by directed differentiation will always run the risk of incomplete differentiation and teratoma or other types of cancer development. In this respect, it is also important to remember that incomplete in vivo reprogramming with the Yamanaka factors leads to tumours resembling Wilms’ tumours.
The alternative is to transdifferentiate cells directly from a somatic cell type to the desired type without a pluripotent intermediate.Several examples of this are now known. Previously, the lab of Melissa Little presented data on the transdifferentiation of HK2 cells, an adult human proximal tubule cell line, into CITED1-positive nephron progenitors. However, so far there have been no follow-up publications, and especially the demonstration of this on primary cells will be essential for therapeutic use. And obviously, for this additional routes to use nephron progenitors in a therapeutic setting are required.
Now, in the December issue of Nature Cell Biology, Kaminski et al identify a method to transdifferentiate mouse and human fibroblasts into kidney tubule cells, which they refer to as induced tubular renal epithelial cells, or iRECs. In a reprogramming factor discovery approach reminiscent of the original iPS publication they first use a bioinformatics approach to identify 55 candidate reprogramming factors, which using expression analysis in Xenopus and involvement in human congenital renal disease and mouse phenotypes was further reduced to 13 candidates. These 13 were tested in their capacity to active GFP expression from a Ksp/Cdh16-Cre driven reporter MEFs. Use of all 13 factors gave a low (0.1%) efficiency of reporter activation. Subsequent removal of factors identified four factors, Emx2, Hnf1b, Hnf4a and Pax8 which combined were sufficient to increase efficiency to 0.6% after transduction and 11.2% after 31 days of culture. Further optimisation led to efficiencies of 23.8% after 5 weeks or 5.4% after one week, although the latter required co-expression of SV-largeT which might not be desirable for therapeutic purposes. The iRECs showed different epithelial characteristics and their expression profile resembled that of different segments of primary tubules. Interestingly, further removal of some of the factors resulted in more cells resembling more specific markers. Functionally, iRECs closely resemble tubule cells, they are polarised with correct expression of different polarity markers, they take up fluorescent albumin and they are sensitive to nephrotoxins. the cells can integrate into tubules of disaggregated embryonic kidneys, and they could form tubules in decellularized kidneys. Finally, as icing on the cake, the authors showed it is possible to make iRECs from postnatal (tail) mouse fibroblasts and human fibroblasts.
Their data presented in this paper convincingly show that it is possible to directly reprogram fibroblasts into renal tubular cells. This gives important new opportunities for regenerative medicine and further illustrates the plasticity of fully differentiated cells when transfected with the right reprogramming factors.
Peter Hohenstein