Chromatin-based Mechanisms of Renal Epithelial Differentiation

Abstract

Successful regenerative renal medicine depends on understanding the molecular mechanisms by which diverse phenotypes of epithelial cells differentiate from metanephric mesenchyme to populate nephrons. Whereas many genes are maintained in a poised state within the population of pluripotent progenitors, specialized epithelial functions reflect the selective expression of a subset of genes and the repression of all others. Here we highlight some common mechanisms of cell differentiation and epigenetic regulation to discuss their implications for renal epithelial development, repair, and disease.

How tissue morphogenesis and cellular differentiation occur in multicellular organisms remains a central question in cell and developmental biology. The process of renal epithelial differentiation arguably begins when a few transcription factors specify the posterior intermediate mesoderm into the metanephric mesenchyme, creating a cell population competent to respond to signals emanating from the ureteric bud (UB).^1,2 The UB induces groups of mesenchymal cells to gain epithelial properties and organize into spherical structures termed renal vesicles (RV). Differential expression of additional transcription factors within the RV triggers further differentiation into broad distal and proximal segment identities.³ As proliferation and possibly cell rearrangements morph the RV into a mature nephron, finer cell fate differentiation occurs along the proximal-distal axis to produce the complete repertoire of cells comprising the mature nephron. Here we illustrate some of the common mechanisms regulating renal epithelial differentiation and the implications of their reversible nature to understanding renal epithelial regeneration and disease.

The Epigenetic Aspects of Renal Epithelial Differentiation

Differentiation reflects the expression of a selective subset of genes and the repression of others and might be achieved by irreversible alterations to the DNA, perhaps even by removing unnecessary genes. We now know that DNA in differentiated adult cells can be reprogrammed to a multipotent embryonic state, enabling cloning or the recreation of the entire organism from one somatic cell.^4,5 Thus, differentiation is unlikely to involve deletions or DNA rearrangements. In fact the stepwise differentiation of renal epithelia involves progressive and reversible restriction of poised or bivalent genomic regions into chromatin regions accessible or inaccessible to the transcription machinery.

The characterization of histone-modifying enzyme complexes reveals the reversible mechanisms involved in cell differentiation. Analysis of histone methylation patterns in the pluripotent embryonic stem (ES) cells and their differentiated counterparts⁶ suggests that chromosomal regions consisting of both repressed (defined by lysine 27 trimethylation on histone 3, or H3K27me3) and active (defined by lysine 4 trimethylation on histone 3, or H3K4me3) chromatin marks contain genes coding for developmentally important transcription factors. Genes located in bivalent regions are in a poised state, an uncommitted transcriptional state with their transcripts present at very low levels. In response to external signals, these poised regions can be modified either into actively transcribed regions or silenced by packing into inaccessible heterochromatin, creating unique combinations of expressed transcription factors as cells differentiate.⁶ During differentiation of kidney progenitors, genomic regions switching from poised in ES cells to active chromatin in renal progenitors contain genes encoding for factors specifying the intermediate mesoderm and renal epithelial progenitors, such as Osr1, Eya1, and Six2 (Figure 1A). Nephron segment-specifying genes, however, remain in poised chromatin domains within renal progenitors. As the progenitors differentiate into epithelia, lineage-dependent epigenetic modifications mediated by unique combinations of transcription factors created in response to morphogen gradients activate genes that will enjoy stable transcription, whereas others become residents of silent chromatin (Figure 1A). Recent analysis of H3K4me3 and H3K27me3 marks in chromatin isolated from Wilms tumor cells agrees with this model.

Figure 1.

(A) A model of the stepwise differentiation of the renal epithelia illustrates the progressive restriction of poised chromatin into regions open (more accessible) or closed (less accessible) to the transcription machinery. This figure depicts only two: histone 3-lysine 4 trimethylation and histone 3-lysine 27 trimethylation, of the many types of epigenetic modifications that regulate the accessibility of the transcription machinery to different regions of the genome. During differentiation of renal progenitors, genomic regions switching from poised to open chromatin contain genes coding for factors specifying the intermediate mesoderm and renal epithelial progenitors, such as Six-2. In contrast, the regions containing genes not involved in kidney development, for example FoxF2, are in a poised state in ES cells and become fully closed in renal progenitor cells. Accordingly, these genes are not expressed in the fetal or adult kidneys. Genes expressed in specific nephron segments in the developing and/or the mature kidney, such as Kcnj3, remain in poised chromatin domains within renal progenitors and become residents of open chromatin in mature nephron segments. The reversible nature of differentiation requires that genes coding for chromatin regulators, such as Jmjd2b, are expressed during kidney development and are silenced in the adult tissue. (B) A hypothetical model of a morphogen gradient produced by the ureteric bud. The distal cells of the RV close to the ureteric bud are exposed to a higher concentration of morphogen when compared with the proximal RV cells. The differential exposure to morphogen translates to differential expression of genes in the distal versus proximal RV. One mechanism by which this occurs is the presence of low affinity morphogen response elements in the vicinity of genes expressed in the distal RV. For example, if Wnt9b acts as morphogen, the hypothetical presence of low affinity TCF/LEF elements in the Dll1 and Brn1 genes might explain why these genes are expressed only in the distal RV and not in the proximal RV. Alternatively, only the cells in the distal RV may be exposed to a signal from the ureteric bud that converts genomic regions containing distal segment–specific genes from a poised state into an open state.

Like ES cells, tumor cells keep differentiation-promoting genes embedded in poised chromatin.⁷ However, tumors retained a transcriptional landscape that resembles renal progenitors, not of the less differentiated ES cells, consistent with findings elsewhere indicating that lineage restrictions are retained in transformed cells. Specifically, genes expressed only in the adult kidney are caged in poised chromatin within both ES and Wilms tumor cells. In contrast, genes involved in brain development, for example, are poised in ES cells but fully silenced in Wilms tumor cells and in renal progenitor cells (Figure 1A). Accordingly, these genes are not expressed in fetal or adult kidneys.⁷ Another important finding of this study is that many genes coding for chromatin regulators, such as the jumonji domain containing 2b (Jmjd2b) are expressed only in embryonic tissues, including the fetal kidneys, but are silenced in adult tissue (Figure 1A). Several other types of histone modifications, including methylation and acetylation at other residues, cytosine methylation, and noncoding RNA expression all epigenetically regulate the accessibility of the transcription machinery to different parts of the genome to define the differentiated state. Precisely how these epigenetic modifiers contribute to renal epithelial differentiation remains largely unexplored.

Organizing Centers and Morphogen Gradients Initiate Differentiation

What are the external mechanisms that trigger epigenetic changes? A common mechanism that can pattern a field of equivalent cells into communities of cells with different identities utilizes an organizing center. The organizer is a tissue or a group of cells within the tissue that generates diffusible signals. The strength of the organizer signal received by a cell within the field will depend on its position relative to the organizer, with cells closest to the organizer receiving the highest concentration of the signal and cells farthest from it receiving the lowest.⁸ A gradient forms, asymmetrically impacting cells within the field because different signal levels will initiate different gene expression patterns within the receiving cells, which will translate into various differentiated fates.⁹ During kidney development, morphogen gradients are utilized in at least two separate instances: first in the specification of a region within the intermediate mesoderm called the metanephric mesenchyme and then in the induction of polarized gene expression within the RV.

The renal epithelial cells of the metanephric kidney are derived from the metanephric mesenchyme, located within the posterior intermediate mesoderm.¹⁰ The positioning and specification of the metanephric mesenchyme appear dependent on two morphogen gradients arranged along the mediolateral axis (bone morphogenetic protein [BMP] signals originating from the lateral side and activin-like signals from the neural tube) and a perpendicular retinoic acid gradient aligned along the anterior-posterior axis.^11–13 A particular concentration of BMP and retinoic acid will induce expression of transcription factors (Pax2, Wt1, Osr1, Lhx1, Eya1, Six1, and Hox11 paralogs) to specify the metanephric mesenchyme.^14–19 Interestingly, Pax2 interfaces with the histone 3-lysine 4 methyltransferase machinery²⁰ and can thus modify the epigenetic landscape to promote differentiation. For instance, Pax2 along with Eya1 and Hox11 cooperatively upregulates the expression of a secreted ligand, glial cell derived neurotrophic factor (GDNF), and a transcription factor, Six2.²¹ Recent findings indicate that Six2 within the metanephric mesenchyme maintains a pool of self-renewing renal epithelial progenitors in this state.^22–24 Hence, morphogen gradients induce expression of position-specific transcription factors, which in turn cooperate to activate genes required for the further differentiation of the renal epithelia.

A second, more speculative example of how morphogens trigger differentiation during kidney development occurs as Wnt9b secretion from the ureteric bud initiates the lateral metanephric mesenchyme to aggregate on the medullary side of the ureteric bud tips.²⁵ These aggregates express and secrete Wnt4 and Fgf8, which are necessary for the further acquisition of epithelial properties, proliferation, and organization into spherical RV structures.^26,27 After epithelialization, the polarized gene expression patterns that appear within the RV (Figure 1B) are always oriented relative to the UB, suggesting that the UB acts as an organizer, secreting signals that induce differential gene expression within the RV.^22,28 The region closest to the UB expresses one set of genes (distal compartment), whereas the region furthest from the UB (proximal compartment) expresses a different set of genes.

The expression of Wnt/β-catenin-responsive genes, Lef1²⁹ and Delta-like 1 (Dll1),^30,31 also occurs in the distal compartment,^32,33 perhaps in response to a Wnt9b gradient emanating from the UB and patterning the RV (Figure 1B). In fact, differential Lef1 expression continues in the S-shaped structures with high levels in distal regions.³³ At least one downstream mediator of the Wnt morphogen, pygopus,³⁴ mediates H3 K4 trimeylation³⁵ and may perhaps explain how the Wnt pathway converts poised regions to active regions.

Local Signaling at Compartment Boundaries Promotes Differentiation of New Cell Fates

The use of positional information provided by morphogen gradients creates distinct compartments of gene expression, often containing lineage-restricted cells. This then allows for new asymmetric interactions to occur at the boundary between the two different cell types, which in turn creates new cell fates that could act as a secondary organizer.^36,37 During the conversion of RV into an S-shaped structure, a middle compartment is established in a Notch-dependent manner. Whereas Notch2 is expressed throughout the RV,³⁸ Dll1, a ligand capable of activating Notch receptors, is restricted to the distal compartment.³² Notch2 activation is likely to initially occur at the boundary between distal Dll1⁺ and proximal Dll1⁻ segments of the RV because ligand/receptor interaction within the same cell can be inhibitory.^39,40 Although the boundary compartment itself can act as an organizer by producing another morphogen that triggers further differentiation of the surrounding compartments, we do not know whether this occurs at the boundary within the late RV.

Without Notch2, the distal segment develops properly,³² however, failing to establish a boundary results in the absence of mature glomeruli, proximal tubular segments, and parts of the loop of Henle (LOH). The boundary cells may provide new positional information; alternatively, they may secrete survival factors supporting the surrounding segments during nephrogenesis. How activation of Notch signaling alters the epigenetic landscape in the context of differentiation is unclear. However, it is known that the histone demythylase KDM5A physically interacts with Notch pathway component RBP-J, and increased H3K4 trimethylation occurs at target genes after Notch activation.⁴¹

Nephron Segment Specific Transcription Factors Promote Differentiation

Segment-specific transcription factors are required for the formation of specialized epithelial cell types. For example, Brn1, a POU domain-containing transcription factor, is expressed only in the cells of the distal RV and the S-shaped body.⁴² Brn1 expression continues during the maturation of distal tubules and LOH and persists in the thick ascending limb of the loop, the distal convoluted tubules, and the macula densa of the mature nephron. Genetic inactivation of Brn1 in mice results in truncated nephrons lacking the LOH and distal tubules, indicating that Brn1 specifies the distal domain within the renal vesicle and likely is required for subsequent differentiation of the distal cell types. How segment-specific transcription factor combinations promote epithelial differentiation along the nephron is unclear, but it is likely that chromatin modifiers are directed through interactions with factors like Brn1 and Lef1 in the distal RV to specific regions of the genome to further parse the genome into active and silent domains.

Altered Epigenetic Marks Cause Cells to Forget Their Fate and Initiate Renal Diseases

The epigenetic basis of differentiation predicts that each mature epithelial cell type in the adult kidney will acquire a unique pattern of chromatin modifications to be endowed with a transcriptional landscape that supports a unique set of cell functions. Under normal physiologic conditions, the transcriptional landscape is in a stable equilibrium,⁴³ but what happens after injury? Epithelial repair involves the dedifferentiation, proliferation, and redifferentiation of surviving cells.⁴⁴ This process triggers re-expression of Pax2,⁴⁵ which may reinitiate chromatin modifications after dedifferentiation. The repair process also requires HNF1β/TCF2, which ensures proper inheritance of the differentiated epigenetic state.⁴⁶ HNF1β is a DNA-binding factor expressed initially in the distal RV and later in all renal epithelia except podocytes, even after terminal differentiation.⁴⁷ HNF1β ensures that chromatin modifications are correctly reinstated after the mitotic silencing of gene expression.⁴⁶

The pitfall of the reversible nature of chromatin modifications is that failure to maintain chromatin modifications will result in kidney diseases, as in the absence of HNF1β.⁴⁶ Hence, it should not come as a surprise that persistent renal diseases after improper epithelial repair are likely associated with abnormal chromatin modifications.^48–52 Conversely, chromatin modification allows epithelial repair to occur without a stem cell niche, instead relying on re-expression of chromatin-modifying factors such as Pax2 to permit surviving epithelial cells at the site of tubular injury to dedifferentiate. Of course, histone modification may also reactivate progenitor populations.⁵³ A precise understanding of how the epigenetic landscape of a disease state deviates from the terminally differentiated state, along with an understanding of the molecular mechanisms that regulate the epigenetic landscape during nephrogenesis, may provide new therapeutic options to reverse the progression of chronic renal diseases.

DISCLOSURES

We wish to thank Drs. Ma. Xenia Ilagan, Scott Boyle, and Hila Barak for their helpful comments during the preparation of this manuscript.