Parietal Epithelial Cells Regenerate Podocytes

In response to acute damage, such as poisons, some fish display an effective neo-nephrogenic potential in adulthood whereby new nephrons form from a residual progenitor pool.¹ In contrast, the response of kidney parenchyma in mammals seems restricted to proliferation of surviving tubule cells without new nephron production. Nephron formation is restricted to organogenesis in mammals because the progenitor population is exhausted before birth in humans and shortly afterward in mice.² This scenario severely restricts the renal restoration of functional capacity after damage in mammals. So how does the adult kidney replenish cells lost through damage, aging, or, as occurs for podocytes, ongoing regular loss of viable cells into the urine?

Although epithelial progenitor cells may exist in adult kidney, as shown for several other organs, the evidence remains equivocal. Simple proliferation may be sufficient to maintain most tubular segments; however, highly specialized epithelial cells of the glomerulus, the podocytes, are hardly ever observed to divide yet remain crucial for renewal and maintenance of renal function. As a result of their close developmental relatedness, there are almost no markers that discriminate parietal from visceral epithelium during renal development, yet there is clear specialization of the visceral epithelium into podocytes as the glomerulus matures. Adult glomeruli often possess cells within the lining of Bowman's capsule that bear podocyte markers,³ leading to the proposal that bridging between the tuft and capsule contributes to pathology such as glomerular crescent formation⁴; however, why parietal cells resembling podocytes are centered on the vascular pole⁵ has been challenging to explain.

As described in this issue of JASN, Appel et al.⁶ used electron microscopy and immunofluorescence to observe parietal cells, intermediate peripolar cells, and podocyte phenotypes as cells “progressed” onto the tuft. They then used an elegant transgenic approach to mark permanently parietal cells (and their progeny) and monitor the fate of the parietal epithelium. This approach exploited a serendipitous observation that a specific region of the promoter of the podocalyxin-like gene (a protein normally expressed in podocytes and parietal cells among others) specifically drove marker gene expression to the parietal (but not visceral) epithelium. In this way, they were able to show that the parietal cells do indeed migrate onto the vascular tuft and differentiate into podocytes.

Also in this issue of JASN, Ronconi et al.⁷ extend their previous research characterizing a progenitor population in human nephrons.⁸ They identified cells around the urinary pole of the Bowman's capsule that coexpressed two previously proposed “stemness” markers of progenitor cells, CD133 (using the antibody capable of enriching for hematopoietic stem cells) and CD24. Ironically, murine CD24 had been proposed as a marker of renal stem cells on the basis of its early expression in the metanephric mesenchyme, but this protein is not equivalent to human CD24.² Confocal fluorescence imaging of human kidney tissues stained for stemness markers (CD24 and CD133) and markers of differentiated podocytes revealed the lining of Bowman's capsule to be formed by heterogeneous cells that follow a fairly strict spatial organization. Progressing from the urinary pole to the surface of the glomerular tuft, there was loss of stemness markers and gain of markers indicating commitment to a podocyte phenotype.

To study this commitment further, these authors isolated CD24⁺CD133⁺ cells from dissociated kidney and sorted them according to their expression of podocalyxin-like (PODXL) protein, before expanding them clonally in culture. Clones expressing all three markers seem committed to producing only podocytes and grow relatively poorly, whereas CD24⁺CD133⁺PODXL⁻ cells show a dual potential as they generate podocytes or tubular epithelium (bearing appropriate differentiation markers) when cultured in selective media. A remarkable property specific to CD24⁺CD133⁺PODXL⁻ cells is their integration into damaged glomeruli and tubules after injecting them into SCID mice that ameliorates the proteinuria and histologic damage from previous exposure to adriamycin. Approximately 11% of all podocytes and 7.5% of proximal tubule cells in regenerating mouse kidney were of exogenous origin, so the benefits gained may have resulted from true integration and transdifferentiation from a common bi-potential precursor.

How do these new studies complement what we understand already, and how might they alter future expectations? Some morphologic clues supporting a relationship between podocytes and parietal cells exists from previous scanning electron microscopy studies. In atubular glomeruli and glomerular cysts in which the vascular tuft became atrophied, parietal cells were replaced by podocytes, and some bosselated cells of intermediate morphology were seen⁹; however, the dynamic view now made possible by lineage tracing is key evidence for the existence of a flux of cells to compensate for wear and tear of the podocytes.⁶ It will be exciting indeed to see whether the precursors shown to have bi-potential can contribute more broadly to regeneration.⁷

Hughson et al.¹⁰ first described a phenomenon in human “end stage” kidneys after long-term dialysis whereby cells of Bowman's capsule seemed to proliferate and take on a more “embryonal” phenotype. Termed embryonal hyperplasia of the Bowman's capsule epithelium (EHBCE), this was proposed to involve de-differentiation to a more primitive state. Ogata et al.¹¹ further characterized EHBCE in association with obsolescent glomeruli in long-term dialysis patients and concluded, on the basis of ultrastructure, that these cells resembled anlage of the glomerular epithelium. This view was reiterated from a molecular perspective by Fukuzawa et al.,¹² who found reactivation of WT1 and PAX2 coexpression in EHBCE in two dialysis patients with mutated WT1. In light of the new observations from Appel et al.⁶ and Ronconi et al.,⁷ we might ask whether EHBCE is an abnormal manifestation of a normal reparative response—with accumulation of the progenitors of both parietal cells and podocytes.

Have we now found the only way podocytes are replenished? Other highly specialized cells, the Purkinje cells of the brain, seem to rejuvenate by fusion with macrophages from the blood stream, without a requirement for the structural disruption of mitosis.¹³ In addition, others¹⁴ proposed that in response to chronic injury, both the parietal and visceral epithelium of the glomerulus reverts to a more primitive macrophage-like phenotype that displays an inflammatory response. Some evidence for the incorporation of bone marrow–derived cells into podocytes has been reported in a mouse model of Alport syndrome, a process that would require the incoming cells to traverse the glomerular basement membrane.^15,16 Although this may occur rarely, even with a genetic selective pressure, the concept of a progenitor pool contiguous with the podocytic side of the glomerular basement membrane seems more feasible for regular turnover. The question of whether this population also contributes substantially to turnover of proximal tubule cells and, if so, how far along the tubules is as yet unknown.

Bi-directional flux of cells from a renal progenitor pool would resemble that proposed many years ago for the stomach.¹⁷ Furthermore, the dependence of specific epithelial cell fates on β-catenin/Wnt, Notch, and Hedgehog signaling pathways in the epidermis and hair follicle¹⁸ leads one to speculate that modulating such signaling in the kidney would also alter the balance of cell types produced. Hence, even if no new nephrons can form, it may prove possible to encourage the formation of new podocytes, hopefully in the right place.

The two articles in this issue^6,7 suggest there is normally a turnover of podocytes, emanating from the tubular neck region and flowing through the parietal epithelium and onward to the tip of the glomerular tuft; if so, then it is clear that in states of chronic renal damage, this reparative mechanism fails. The ability of harvested Bowman's capsule progenitors to assist repair after adriamycin in mice is therefore an exciting finding,⁷ because it opens up the possibility of seeding immunologically compatible precursors to promote functional repair.

DISCLOSURES

RP is Deputy Editor of the Journal of Pathology. ML is a National Health and Medical Research Principal Research Fellow. ML has no financial interests to disclose.