When he identified the Bowman capsule, Sir William Bowman believed that it consisted of epithelial cells close to the urinary pole, and a membrane at the vascular pole.1 Joseph von Gerlach demonstrated in 1845 that the capsule was entirely lined by parietal epithelial cells, which were considered to be a simple, homogeneous cell population for more than a century.1 In the 1970s, however, electron microscopy revealed that parietal epithelial cells are heterogeneous, and can also have features of either tubular cells or flat podocytes (named 'parietal podocytes').2 In 2006, Bariety and colleagues described the existence of two types of parietal podocyte. One type, found in healthy adult human kidneys, is characterized by a transitional phenotype with coexpression of podocyte and parietal epithelial cell markers. The other type, observed mostly in atubular glomeruli or glomerular cysts, is characterized by expression of podocyte markers only.3

The function of the transitional population of parietal podocytes was clarified by studies that demonstrated that adult human kidney parietal epithelial cells can act as progenitors for podocytes, and that a subset of parietal epithelial cells, which coexpress parietal and podocyte markers, represent podocyte-committed progenitors (Figure 1a).4 The same transitional population had also been identified in healthy mice by Appel and co-workers who demonstrated, using lineage tracing strategies, that parietal epithelial cells can become podocytes during kidney growth (Figure 1a).5 The function of the other type of parietal podocytes, found in atubular glomeruli and glomerular cysts, has now been described by applying lineage tracing strategies in mouse models of unilateral ureteral obstruction, and in a coagulation model established to induce generation of atubular glomeruli.6

Figure 1: The phenotypes and functions of podocyte progenitors and ectopic podocytes.
figure 1

The directions of migration of cells are indicated by the arrows. a | Parietal epithelial cells that display podocyte as well as parietal epithelial cell markers are localized within the Bowman capsule and act as podocyte progenitors. Parietal epithelial cells can generate novel podocytes by (1) progressively differentiating toward the vascular stalk or (2) migrating over the glomerular tuft and acquiring podocyte markers. Alternatively, parietal epithelial cells that already express podocyte markers can directly migrate over the tuft, which might allow a quick replacement of lost podocytes (3). b | Ectopic podocytes are observed in atubular glomeruli and glomerular cysts. Following injury to the tubuloglomerular junction and the loss of the connection between the tubule and the glomerulus, parietal epithelial cells die. Their loss is followed by relining of the empty Bowman capsule with differentiated podocytes.

Full size image

Schulte and co-workers genetically tagged podocytes or parietal epithelial cells to show that parietal podocytes in atubular glomeruli and glomerular cysts are derived by the migration of podocytes from the adjacent tuft onto the Bowman capsule (Figure 1b).6 Interestingly, both murine models in their study are characterized by primary tubular injury, aberrant proliferation of cells of the urinary pole, tubular obstruction and tubule–glomerular disconnection with consequent parietal epithelial cell death (Figure 1b).6 Podocyte migration on the capsule is, therefore, likely to be a secondary event that occurs after parietal epithelial cell death to fill the empty area and avoid filtrate leakage. Indeed, once progenitors are lost, the empty area of the tissue is usually colonized by differentiated cells. These observations are consistent with those reported in other progenitor systems.For example, in bone marrow aplasia the empty niche left by depleted haematopoietic progenitors is filled by migration of differentiated adipocytes.7 Thus, the study by Schulte and colleagues demonstrates that the parietal podocytes observed in atubular glomeruli and glomerular cysts are not derived from differentiation of parietal epithelial cells.6 These particular parietal podocytes represent podocytes that dislocate following chronic tubular injury and thus, would be better defined as 'ectopic podocytes' (Figure 1b). Ectopic podocytes are clearly distinct from the population of parietal podocytes that represent podocyte progenitors and coexpress both parietal epithelial cells and podocyte markers.3,4

...the ability of podocyte progenitors to generate novel podocytes might have important clinical implications

Although interesting, the phenomenon of podocyte migration onto the capsule in atubular glomeruli is unlikely to be clinically relevant because it occurs only once the nephron is already irremediably lost.3 By contrast, the ability of podocyte progenitors to generate novel podocytes might have important clinical implications. Understanding the mechanisms that drive parietal epithelial cell differentiation, and in which clinical situations this phenomenon can eventually occur, is crucial if we are to attempt to modulate cellular differentiation for therapeutic purposes. However, models of tubule–glomerular disconnection cannot provide an answer regarding the capacity of parietal epithelial cells to differentiate into podocytes. Rather, models of podocyte injury are necessary, because activation of podocyte regeneration requires an induction of primary podocyte loss. Investigation of this phenomenon will also need sensitive analytical tools as this process is likely to be quantitatively limited, although clinically relevant.

Loss of even a small percentage of podocytes is, in fact, enough to induce proteinuria.8 However, when the percentage of lost podocytes remains below 20%, proteinuria can undergo remission and glomerular integrity can still be restored,8 which suggests that in such conditions regenerative mechanisms can occur and be relevant for regression of the disorder. Indeed, in this situation, generation of only 5–10% of new podocytes from parietal epithelial cells allows replacement of more than half of the lost podocytes, which might be enough to induce remission of proteinuria and restoration of glomerular integrity. By contrast, when podocyte loss is over 20%, severe, persistent proteinuria and glomerular scarring occurs,8 suggesting that regenerative mechanisms are either no longer sufficient, or might not be functional. In agreement with the latter possibility, severe albuminuria was recently demonstrated to inhibit differentiation of parietal epithelial cells into podocytes, and promote the generation of focal segmental glomerulosclerosis lesions by sequestering retinoic acid.9 This finding explains why regeneration cannot occur after severe and persistent podocyte injury. However, studies have shown that lowering levels of albuminuria with angiotensin-converting-enzyme inhibitors induces regression of glomerulosclerosis,10 and that treatment with retinoic acid restores parietal epithelial cell differentiation into podocytes and reduces proteinuria.9 In the latter study, about 10% of total glomerular podocytes were newly generated from parietal epithelial cells in response to treatment,9 and the outcome was a restoration of glomerular structure, proteinuria remission and improvement of functional injury.9 Parietal epithelial cell differentiation into podocytes might, therefore, have an important clinical effect and its enhancement through pharmacological agents could be an attractive therapeutic option. More studies are needed to clarify the contribution of parietal epithelial cells to generation of novel podocytes in various glomerular disorders.