Direct transcriptional reprogramming of adult cells to embryonic nephron progenitors

Abstract

Direct reprogramming involves the enforced re-expression of key transcription factors to redefine a cellular state. The nephron progenitor population of the embryonic kidney gives rise to all cells within the nephron other than the collecting duct through a mesenchyme-to-epithelial transition, but this population is exhausted around the time of birth. Here, we sought to identify the conditions under which adult proximal tubule cells could be directly transcriptionally reprogrammed to nephron progenitors. Using a combinatorial screen for lineage-instructive transcription factors, we identified a pool of six genes (SIX1, SIX2, OSR1, EYA1, HOXA11, and SNAI2) that activated a network of genes consistent with a cap mesenchyme/nephron progenitor phenotype in the adult proximal tubule (HK2) cell line. Consistent with these reprogrammed cells being nephron progenitors, we observed differential contribution of the reprogrammed population into the Six2(+) nephron progenitor fields of an embryonic kidney explant. Dereplication of the pool suggested that SNAI2 can suppress E-CADHERIN, presumably assisting in the epithelial-to-mesenchymal transition (EMT) required to form nephron progenitors. However, neither TGFβ-induced EMT nor SNAI2 overexpression alone was sufficient to create this phenotype, suggesting that additional factors are required. In conclusion, these results suggest that reinitiation of kidney development from a population of adult cells by generating embryonic progenitors may be feasible, opening the way for additional cellular and bioengineering approaches to renal repair and regeneration.

Figure 1.

Combinatorial screening for transcription factors is able to reprogram to a nephron progenitor. (A) Combinatorial screening with 15 different factors was performed in triplicate in 96-well plates. Criteria of morphologic EMT and detection of Cited1⁺ cells in two of three wells allowed progression to the qRT-PCR phase, which assessed a broader panel of NP markers. qRT-PCR results provided the basis for input populations for the recombination assay, which assessed the capacity of reprogrammed cell populations to behave as bona fide kidney progenitors. UB; ureteric bud. (B) Ten pools were identified based on induction of CITED1 protein as assessed by immunofluorescence. E1, EYA1; H11, HOXA11; O1, OSR1; P2, PAX2; S1, SIX1; S2, SIX2; Sn2, SNAI2. (C) qRT-PCR to determine the degree of reprogramming in HK2 cells infected with viral pools. qRT-PCR screening identified only one factor combination (pool 8; black) that showed consistent activation of a variety of NP markers coupled with downregulation of E-CADHERIN. Gene expression is normalized to glyceraldehyde-3-phosphate dehydrogenase, and it is expressed relative to HK2 cells infected with GFP-containing lentivirus and cultured without VPA.

Figure 2.

Identification of a pool of genes able to induce EMT and endogenous expression of NP markers. (A) Compared with HK2 cells infected with an empty lentiviral vector without the addition of VPA (HK2-VPA), pool 8-reprogrammed HK2 cells (HK2+Pool 8) became elongated and spindle-shaped by day 7 of reprogramming, consistent with an EMT. (B) qRT-PCR revealed that EMT markers MMP2 and MMP9 were upregulated ∼30- and 3-fold, respectively, in HK2+Pool 8 cells relative to HK2-VPA control cells, whereas E-CADHERIN levels were unchanged. (C) qRT-PCR showing coordinated upregulation of all NP markers in HK2+Pool 8 compared with HK2-VPA control cells, with simultaneous downregulation of E-CADHERIN expression. qRT-PCR results are expressed as the average relative fold change compared with control from three biologic replicates ± SEM. (D) RT-PCR analysis of the unmanipulated HK2 cell line (HK2 parental), HK2+VPA, and HK2+Pool 8 to examine (1) the endogenous expression level of markers of kidney development (SIX2, CITED1, SALL1, and FOXD1), (2) the marker of three germ layers (mesoderm [goosecoid; GSC], endoderm [FOXA1], and ectoderm [PAX6]), and (3) a pluripotency marker (NANOG). (E) Immunofluorescence analysis of the induction of PAX2 and CITED1 protein in the GFP⁺ lentivirally transfected cells within cultures of HK2+Pool 8 cells versus HK2 parental cells. Merge includes 4',6-diamidino-2-phenylindole to visualize nuclei. Scale bars, 30 µm.

Figure 3.

Development of a stringent assay for NP potential. (A) E12.5 mouse kidneys were dissociated to single cells, mixed with GFP-labeled test cells, and processed to form a recombined metanephric kidney ex vivo. (B–D) Single cells from a GFP-labeled E12.5 kidney can be detected in the NP compartment marked by Wt1 and Six2 and also in the calbindin⁺ ureteric bud (UB) compartment. (E–G) Sall1-GFP NPs contribute to developing nephron compartments but not to the UB-derived calbindin⁺ compartment. Arrows highlight integration sites of the test cell population. (H–K) Non-NP cell lines M15, MSCs, HEK293Ts, and HK2s cannot contribute to the Six2⁺ NP compartment. Cell lines were identified by anti-GFP antibody. Immunofluorescence for Six2 was used to assess incorporation into NP compartments. Scale bars, 120 µm in B–G; 360 µm in H–K.

Figure 4.

Contribution of reprogrammed cells to the NP compartment in recombination assays. (A) GFP⁺ HK2+Pool 8 cells (block arrows) contributed to both the Wt1⁺ and Six2⁺ NP compartments but not the Calbindin⁺ compartment ex vivo. Additional Six2⁺/GFP⁻ cells surround the reprogrammed cells, indicating true integration into the NP compartment. Scale bars, 30 µm. (B) Although unable to efficiently integrate into the Six2+ compartment, HK2+VPA cells did integrate into Wt1⁺ structures likely to represent early nephrons. Scale bars, 30 µm. (C) Quantitation of integration events revealed that, although HK2 cells treated with VPA alone or together with pool 8 integrated into Wt1⁺ structures, HK2+Pool 8 cells preferentially integrated into the endogenous ex vivo Six2⁺ compartment. Both HK2 parental and HK2+TGFβ1-treated cells showed poor integration into either compartment. (n=3±SEM). (D) Schematic representation of the expression domain in the murine developing kidney of the six reprogramming factors in pool 8. MM, metanephric mesenchyme; RV, renal vesicle.

Figure 5.

The role of EMT in reprogramming to a NP phenotype. (A) Removal of SNAI2 from the reprogramming pool leads to dramatic upregulation of E-CADHERIN, with no effect on CITED1 levels. (B) Reprogramming using SNAI2 alone or HK2 cells induced to undergo EMT through treatment with TGFβ1 failed to generate cell populations able to integrate into embryonic kidney recombinations. (C) Schematic diagram illustrating the observed outcomes of the reprogramming screen. (C, i) Addition of either recombinant TGFβ1 or SNAI2 transduction alone resulted in a fibrogenic EMT, with the resulting cells unable to contribute to the nephron progenitor compartment of the developing explants. (C, ii) Undirected chromatin relaxation mediated by VPA treatment biases the HK2s to the NP/cap mesenchyme fate, but the cells cannot overcome existing epigenetic barriers; therefore, they are only partially reprogrammed. (C, iii) Lineage-instructive transcriptions factors together with SNAI2 reduce the existing epigenetic barriers, which then allows (C, iv) transition from the HK2 proximal tubular phenotype to the NP/cap mesenchyme phenotype.