Podocyte Aging: Why and How Getting Old Matters

Abstract

The effects of healthy aging on the kidney, and how these effects intersect with superimposed diseases, are highly relevant in the context of the population’s increasing longevity. Age-associated changes to podocytes, which are terminally differentiated glomerular epithelial cells, adversely affect kidney health. This review discusses the molecular and cellular mechanisms underlying podocyte aging, how these mechanisms might be augmented by disease in the aged kidney, and approaches to mitigate progressive damage to podocytes. Furthermore, we address how biologic pathways such as those associated with cellular growth confound aging in humans and rodents.

Scope of the Problem

The mechanisms underlying the aging process are diverse, and include general and cell type–specific aspects.¹ Such mechanisms also apply to kidney aging, and several lines of evidence highlight the importance of studying aging and the kidney. The US population is growing older, and the US Census Bureau predicts that the number of Americans aged ≥65 years will more than double over the next 40 years. Similarly, Eurostat, the statistical office of the European Union, forecasts that 28% of Europeans will be aged >65 years by 2060. As life expectancy increases, the effect of advanced age on kidney health and function is becoming an increasingly important medical and socioeconomic factor. The GFR declines after age 40 years by 0.8%–1.0% per year,^2,3 and kidneys from healthy donors aged 70–75 years have 48% fewer intact nephrons compared with young donors aged 19–29 years.⁴ This is consistent with an estimated annual loss of 6000–6500 nephrons after age 30 years.^4,5 The progressive decrease in nephron number affects the kidney at multiple levels: (1) kidney function declines with age^6–8; (2) glomerular diseases are disproportionately more common in people aged >65 years^9,10; (3) the elderly have worse outcomes compared with those of younger individuals with the same disease^11–14; (4) older age accelerates the transition from AKI to chronic kidney injury¹⁵; (5) new starts for chronic dialysis are now predominantly in the elderly population^16–18; and (6) the kidney’s drug-metabolizing capacity declines in older people.

Clinical Relevance of Aged Podocytes

Characteristic physiologic, histologic, and molecular changes to the aged kidney have been well described and reviewed, and will not be summarized here.^7,8,19–21 This review focuses on aging of the glomerular podocyte. The glomerular structure is a microcosm within the kidney, where its different cell types interact and communicate extensively. Aging results in age-dependent glomerulosclerosis and an accompanying decline in GFR, caused by changes in number, structure, and function to all four resident glomerular cell types.^{5,7,22,23,24–26} Compelling evidence points toward podocytes as the critical cell type in aging. Floege et al.²⁷ were the first to describe age-related glomerulosclerosis as a podocyte disorder, subsequently confirmed by others.^25,28 Clinically, the development of age-associated glomerulosclerosis is accompanied by a predictable decrease in podocyte number and density with advancing age.^28–30 This contributes, for example, to a reduced life expectancy of transplanted kidneys,³¹ and is why glomerular diseases have a higher likelihood for a poorer outcome in older patients than in younger ones.^12,32

But there are still many connections between age and podocyte function that have not been sufficiently studied. These include the direct pleiotropic cellular effects of therapeutics such as corticosteroids, or renin-angiotensin-aldosterone system inhibitors that alter biologic process/pathways in young podocytes; genetic predispositions such as the apolipoprotein L1 genotype; and infection with certain viruses, such as HIV and possibly severe acute respiratory syndrome coronavirus 2. Older age also negatively affects the normal function of podocytes in their role as important players in the health of other glomerular cell types, including making survival factors for glomerular endothelial cells,³³ in regulating parietal epithelial cell differentiation,³⁴ and in recruiting mesangial cells.

Finally, a critical clinical question that requires mechanistic understanding is how podocyte aging and disease intersect. Experimentally, disease-induced podocyte loss compounds the pre-existing podocyte depletion in the aged kidney, associated with increased glomerulosclerosis.³² As such, from a clinical treatment standpoint, it is important to know whether changes in gene expression during podocyte aging overlap or are distinct from gene expression changes in common podocyte diseases. One analysis comparing data from published studies showed the regulated pathways were largely distinct from those perturbed in diabetic kidney disease and focal segmental glomerulosclerosis.³⁵ Yet, more studies are clearly needed to identify distinct and common pathways between aging and disease.

In this review, we do not discuss the normal structure or function of podocytes, but rather examine what is specifically known about podocyte aging. As expected, podocyte aging is a synthesis of pathways involved in the general aging process, with additional cell type–specific aspects.¹ We limit our discussion to published data and do not apply information extrapolated from studies on general aging to podocyte aging.

Approaches to Study Podocyte Aging

Humans

Studying the biology of kidney aging in humans is the gold standard to assure clinical and therapeutic usability of the data. Yet, human studies face numerous barriers. First, longitudinal analyses are not possible because only a single time point of tissue collection is typically available from a given person. Second, proper sex- and race-matched controls are mostly limited. Third, comorbid conditions (such as obesity, diabetes, or hypertension) or environmental factors (such as exposure to cigarette smoke or air pollution) increase with age and confound changes directly attributed to aging alone,³⁶ as do epigenetic changes.³⁷ Fourth, although approaches have been optimized for small sample sizes, collecting sufficient kidney tissue for mRNA and protein studies is still challenging.

Over the years, several creative attempts have been made to overcome some of these barriers for human studies. Wiggins and Rule used implantation biopsies from donor kidneys for transplantation or radical nephrectomy for renal tumors.^8,28,38,39 Bertram et al. used contrast-enhanced computed tomography and a renal biopsy to determine glomerular density and glomerular and tubular area in living people.²¹ Finally, the Kidney Precision Medicine Project is optimizing conditions to acquire material for cutting-edge technologies such as single-nucleus RNA-sequencing (RNA-seq).^40,41 These workflows will become available to study aging in healthy people.

Animal Models

Because use of human kidneys for longitudinal and interventional studies of aging and comorbid conditions remains a challenge, animal models continue to be essential to advance understanding of podocyte aging. Yet, it is important to state that the interpretation of experimental data with respect to human aging always requires careful crossvalidation in human kidneys. Over the past few years, it has become increasingly clear there are significant species-dependent differences in kidney development and disease, and it is likely differences also exist for aging. Early important insights into podocytes and kidney aging were obtained from studying rats.^30,42,43 Marmosets,⁴⁴ canines,⁴⁵ and minipigs⁴⁶ have also been studied. However, largely because of the availability of transgenic animals, mice, and to a lesser extent, rats are now commonly used experimentally to study kidney and podocyte aging.

Importantly, several caveats need to be considered when using mice or rats for podocyte studies of aging (and for studies of other kidney cell types). First is the consideration of age. For example, mice aged for 12 months, 18 months, or 24 months correspond to human aged 42.5 years, 56 years, and 69 years, respectively.^47,48 Thus, because a 12-month-old mouse is definitely not considered aged, and an 18-month-old mouse is really only middle-aged, we recommend only mice that are ≥24 months be considered a true model of “human” kidney aging. Rats that are aged 28 months are equivalent to a 70-year-old human.⁴⁹ Second, species- and strain-dependent glomerular physiology affect age-associated glomerular changes. Similar to findings in aged humans reported by Rule et al.,⁵⁰ aged mice typically do not develop proteinuria.^51–53 They exhibit a 35% decrease their GFR between age 4 months and 24 months, at a rate of 6.3 ml/min per 1.73m² per decade, but show only a very small increase in serum creatinine.^7,54 In contrast, aged rats develop more overt changes, including proteinuria and lower GFR.⁵⁵ Importantly, different mouse strains also are not equivalent with respect to aging. C57BL/6 mice are most susceptible to the development of glomerulosclerosis, spontaneous amyloidosis, and proteinuria as they age.⁵⁶ Conversely, mice on a mixed genetic background have podocyte depletion by age 12 months,^57,58 which is earlier in onset than in C57BL/6 mice. The reasons for these differences are unknown. Third, as mentioned above, an animal’s sex affects aging, with male sex typically associated with an earlier kidney aging phenotype, for reasons that are not well understood. One possibility lies in kidney morphology/homeostasis. Females have more cortical glomeruli per unit volume than males.⁵⁹ Similarly, nitric oxide production, which affects renal hemodynamics, is better preserved in aged females than males.^60,61 When males are housed together, they tend to fight more than females, often resulting in subclinical low-grade inflammation from wounds, which can affect aging. Fourth, housing conditions specific to a given vivarium can influence all strains, due to husbandry, diet, infectious and parasitic diseases, and differences in the microbiome. Fifth, the high costs for healthy aging of mice and rats for up to 24–27 months can be prohibitive and prevent exploratory experiments regularly performed in other scientific areas. Thus, considering models exhibiting accelerated podocyte aging are a more cost-effective possibility. These include genetically manipulated mice, such as those with reduced telomere lengths,⁶² with overexpression of p16 to induce senescence,⁶³ or Inducible Changes in the Epigenome mice that have undergone induced DNA breaks to drive epigenetic changes that accelerate aging.^64(preprint),^65–67 A more physiologic alternative is inducing premature birth, which results in a low nephron mass that is more susceptible to aging effects.⁶⁸ Finally, one might also consider replicating in models comorbid conditions that affect a large proportion of the human population, such as hypertension and/or obesity.

Cell Culture Models

Despite obvious limitations, cultured primary and immortalized human and mouse podocytes have advanced our understanding of podocyte biology and have allowed mechanistic dissections that require large number of pure cell populations.^69,70 Unfortunately, with respect to aging, a standardized model to study this process in vitro is lacking. The most straightforward approach so far has been comparing primary podocytes harvested from aged and young mice. For example, this has provided information on the induction of replication-associated telomere shortening,^71,72 genotoxic stress,^73–76 and how low–molecular weight compounds⁷⁷ or overexpression of senescent pathway regulators⁷⁸ can induce senescence-like states. Finally, stem cell–derived kidney cells/organoids have offered a new model for study of kidney development and disease, although there are no published findings yet toward understanding podocyte aging.

Sex as a Biologic Variable

Over all, aging-related changes in the kidney typically occur earlier in males than females.^60,79–82 Indeed, studies are increasingly identifying sex differences between male and female kidney cell types, such as proximal tubule cells or cells of the renin lineage.^83–86 Unfortunately, in the current gold standard single cell/nuclear RNA-seq studies, glomerular cell types—and podocytes in particular—are numerically underrepresented unless they are specifically enriched.^87–89 To circumvent this and obtain more detailed information on podocyte aging, we recently performed a study using an inducible podocyte-specific reporter mouse (NPHS2-rtTA|tetO-cre|tdTomato) to permanently label the podocyte population in juvenile animals and follow them through their lifetime.³⁵ A comparison of labeled podocytes from young mice (aged 3 months, corresponding to approximately 20 years in humans)^47,48 with podocytes from aged mice (22–24 months, corresponding to approximately 70+ years in humans)^47,48 identified nearly 2500 genes with significant differential expression between young and aged podocytes.³⁵ However, when analyzed with respect to sex differences, age differences explained most of the expression variance and overwhelmed the small expression variance of 7.2% associated with sex. Thus, this study suggests that in mouse podocytes, age-related changes may be largely independent of sex. However, further studies are needed to address this in more detail using human glomeruli, additional mouse strains, and a wider range of ages.

Podocyte Lifespan versus Healthspan

As we review the molecular and cellular changes to aged podocytes and their physiologic and pathologic consequences, consideration needs to be given to distinguishing a podocyte’s lifespan and healthspan (Figure 1). We view the lifespan of the individual podocyte as the length of time it lives, which translates to the total number of podocytes comprising the glomerular filtration barrier. Lifespan is binary (a podocyte is living or not). Healthspan is the part of the podocyte’s life that is spent in good health and able to perform normal cellular and molecular functions to maintain its physiology, structure, and function. Healthspan is not binary, and is best considered a continuous variable, changing dynamically throughout life.

Figure 1.

Podocyte lifespan and healthspan. The length of time that an individual podocyte lives is its lifespan (green arrow). Lifespan is binary, where cells are either alive or lost due to death and/or detachment. Conversely, healthspan (orange arrow) is the part of the podocyte’s lifespan spent in good health. It is not binary as it changes dynamically throughout life (gradient orange-green arrow).

Podocyte Depletion

It is generally accepted that terminally differentiated podocytes are unable to self-renew, yet some studies have shown postnatal nephron endowment with podocyte gain.⁹⁰ Animal model studies have shown a direct correlation between podocyte depletion and glomerulosclerosis.^91–93 Indeed, kidneys that exhibit glomerulosclerosis increasing with age^50,94–96 also feature a decrease in absolute podocyte numbers (reduced lifespan) and density.^{23,30,32,43,57,58,97,98,99} It is important to note that this absolute loss of podocytes is accompanied by an increase in podocyte size (i.e., hypertrophy). Podocyte hypertrophy can partially compensate for the decreased podocyte numbers, yet the relative podocyte depletion is ultimately insufficient to match an overall reduction of podocyte density.^29,100,101 In fact, podocyte density decreases in both aged rats⁴³ and mice.^{23,32,58,97–99} Moreover, this effect is region specific, because the kidney’s cortical versus juxtamedullary compartments differ in age-related changes in podocyte numbers/density. Compared with young mice, aged mice have podocyte densities that are 45.6% and 39% lower in the cortex and juxtamedulla, respectively.³² This contrasts with the other glomerular cell types, where data lack consistency. Some studies have found the number of mesangial and glomerular endothelial cells increase on aging, so the density—the ratio between the glomerular volume and cell number—stays constant.^30,102,103 Conversely, studies in aged Wistar/Lou rats have observed no change in mesangial hypercellularity,¹⁰⁴ and research in Sprague-Dawley rats found a glomerular capillary loss with reduced proliferation.¹⁰⁵

These findings from animal studies are supported by data in humans. Wiggins et al. reported 589±166 podocytes per glomerulus, with a density of 1166±310 per 10⁶ µm³,¹⁰⁶ which is similar to the 558 podocytes per glomerulus reported by Puelles et al.²⁹ Moreover, aging is associated with both an absolute and relative depletion of podocytes.²⁹ The podocyte reserve drops from more than 300 podocytes per 10⁶ µm³ in young kidneys to less than 100 per 10⁶ µm³ by age 70–80 years, resulting in an annual podocyte loss of approximately 0.9%.²⁸ This age-dependent podocyte depletion has been shown to have high clinical relevance. A decrease in absolute podocyte number and density often result in glomerulosclerosis,^102,107 and thus remains the best predictor for glomerulosclerosis risk in healthy aged kidneys.^28,29,31,100 Moreover, podocyte number and integrity are also critical in transplanted kidneys,¹⁰⁸ because kidney transplants with low podocyte endowment develop transplant glomerulopathy at a higher rate compared with those with normal podocyte density.¹⁰⁹ Several mechanisms have been implicated with respect to the causes of podocyte depletion in aged kidneys.

Apoptosis

Although apoptosis occurs in aging podocytes (Figure 2),^110,111 its overall contribution to podocyte depletion is technically difficult to quantify over a lifespan.¹¹² Established assays (e.g., terminal deoxynucleotidyl transferase–mediated digoxigenin-deoxyuridine nick-end labeling) can capture only a short time span in the apoptotic program. Similarly, the underlying causes and inducers of age-dependent podocyte apoptosis are poorly understood. Transcriptomic data from aged podocytes showed both a decrease in survival factors and an increase in proapoptotic pathways (Table 1).³⁵ Yet, their mechanistic contributions have not been experimentally tested. Other candidates can be extrapolated from published studies in nonaged podocytes.¹¹² These include the age-dependent decrease of nephrin and podocin that likely reduces podocyte survival¹¹³; the age-dependent decrease of Wilms’ Tumor 1 that leads to increased WNT/β-catenin signaling, which in turn can induce podocyte death¹¹⁴; the age-dependent increase of TGF-β that directly induces apoptosis¹¹⁵; and the balance in autophagy, because age-dependent decrease in autophagy induces apoptosis and restoring autophagy therefore attenuates it.¹¹⁶

Figure 2.

Changes to podocytes in the aged human and rodent kidney. (A) Schema of the microcosm of cell types comprising a young glomerulus. (B) Glomerular size (hypertrophy) is increased in the aged kidney (ruler shown below). Podocytes detach, hypertrophy, efface, apoptose, and undergo pyroptosis. Parietal epithelial cells (PECs) become activated and decrease in number due to loss, mesangial cells are expanded, and glomerular endothelial cells become injured. Created with BioRender.com.

Table 1.

Differences in podocyte gene expression between young and old mice³⁵

Biologic Process	Genes Increased	Gene Decreased
Extracellular matrix proteins	Mmp13, Fn1	Sema5a, Col4a1, Col4a5, Col4a4, Col18a1, Lamb2, Lamc1, Npnt, Dag1, Myh11
Apoptosis	p53 pathway, TNF pathway, Faslg, Bim, Prf1erforin, Apaf1, Casp, Casp4.	Pou3f3, Fgfr2, Wt1, Erbb3, c-Met, Smad6
Survival factors	Tgfb1, Tgfb3	Pou3f3, Fgfr2, Wt1, Erbb3, c-Met, Egfr, Smo
Pyroptosis	Casp1, Nlrc4, Aim2, Gsdmd.	None
Cell cycle regulation	Tgfb1, Myc, Ccnd3, p21, Notch2, Schlafen-1Slfn1	Fgfr2, Edn3, c-Met, Tbx3, Pkd2, Sox4, Meis2
Hypertrophy	Akap13, Rgs2, P2rx4	Lmcd1, Igfbp5
Inflammasome and inflammation	Nlrp3LRP3, Casp1, Casp4, Nlrc4, Naip2, Naip5, and Naip6, Tnfa, multiple interferon genes, Il6, Il2, multiple complement component genes, Nfkb1, Nfkb2	Nlrp6
Senescence and SASPs	Irf5, Cebpb, and Jdp2, Ccl2, Ccl3, Ccl5, Cc8, Mmp12 Mmp13	Igfbp5, Igfbp7
Podocyte canonical and slit diaphragm genes		Nhsp1, Nhsp2, Lmx1B, Cdkn1c, Wt1, Podxl, Synpo,Vegfa, Cd2ap, Actn4, Ptpro, Trpc6
Tight junction genes	Vasp, Amica1	Tjp1, Pard3, Pard6, Nedd4l, Pkci, multiple occludins, multiple claudins (including Cldn4, Cldn8), Cgn, Magi1, Magi2
Transcription factors	Cebpa, Cebpb, Prdm1, Nr4a3, Tbx21, Bach1, Tcf7, Myc, Irf4, Runx3, Stat4, Nfkb2, Fosl2, Bcl6, Nfkb	Hnf1b, Grhl2, Ssim1, Lrx1, Creb3l1, Zbtbl6, Osr2, Foxo4, Pax8 and Tref1, Hsp27, Foxc2, Arf2, Itga3, Yap, Coro2b, Trpc6, Ptk2
Actin and regulators, motility	Was, Wipf1, Flna, Evl, Myo9b, Myo5a	Kank1-Kank4, Dsp, Parva, Syne2, Myh10, Tpm1, Emp2, Myh14, Gja5, Vil1
Crosstalk with other cells	Il1a/bIl1a, Il1b, C1qa, C1qb, Tgfb1, Mmp13, Cxcl16, Ccl4, Ccl5, Tnf, Vim, Ifng, Il18, Vcam1, Sema4d	Receptor ligand pairs: Vegf, Fgf, Wnt, Bmp, Egf, Erbb4, Slit2, Robo2
Mitochondrial OXPHOS	Ncf1, Ncf2, Tlr2, Aif1, Nox2, P2rx7, Tnf, Prkcd Syk, Nrros	Complex I (Ndufa5, Ndufs2, Ndufs4 and Ndufb8), Complex II (Sdha, Sdhd, and Sdhc), Complex III (e.g., Uqcrb, Uqcrh, Uqcr11), Complex IV Cytochrome C oxidase components (e.g., Cox5b, Cox6c, Cox7b)
Glycolysis	Hk2, Hk3, Eno3	Pfkm, Fbp1, Gpd1, Aldob
Lipids	Cd36, Apoe, Apobr, Soat1, Ptgs2, Fabp5, Ldlr, Cerk, Dgat1	Elovl7, Elovl6, Scd2, Acadsb, Asah2, Ppargc1a, Mgll, Hadh, Acadm, Acad11, Acat1, Acsm, Acox1, Eci3, Ehhadh
Cell-cell junction		Cdh1, Chd2, Cdh3, Chd4, Rab25 along with Grhl2
Integrin-collagen interactions		Collagen chains (Col4a3, Col4a4, Col4a5) and Integrins (Itga3, Itgb6 Itbg8)
DNA damage	Samhd1, Bach1, Eepd1, Hmga1, Uhrf1, Dna2, Dclre1c, Taok3, Topbp1, Uvrag, Rad9a	Spire2, Xpc, Tnks1bp1

SASP, senescence-associated secreted phenotype; OXPHOS, oxidative phosphorylation.

Pyroptosis

Our recent study showed that genes required for pyroptosis, a highly inflammatory form of lytic programmed cell death, are markedly increased in aged podocytes (Table 1, Figure 2). Pyroptosis is characterized by the activation of caspase-1, with cells undergoing lysis, swelling, pore formation, blebbing, and DNA fragmentation, and is typically initiated by the inflammasome.¹¹⁷ It generates an inflammatory response due to the activation and release of IL-1β and IL-18, followed by IFN-γ production.¹¹⁷

Podocyte Detachment

The detection of podocytes in the urine is considered a sensitive marker of ongoing glomerular damage (Figure 2).¹¹⁸ Viable podocytes can be detected in the urine.¹¹⁹ But attempts to determine their actual number is likely to result in an underestimation, because detached podocytes can undergo secondary apoptosis in the urine by the process called anoikis.¹²⁰ Moreover, detachment of one podocyte can result in subsequent detachment of neighboring podocytes.^121–123 Hodgin et al. showed that compared with younger people, humans aged >60 years had a 3.3-fold increase in podocytes in the urine, and many of podocytes were also stressed.²⁸ Mechanistically, RNA-seq data identified several candidate mechanisms for podocyte detachment (Table 1). These include the significant downregulation of several cell-cell junction genes and the weakening of integrin-collagen interactions.

DNA Damage and Cell Cycle Control

Terminally differentiated podocytes cannot proliferate and self-renew, in part because of increased cell cycle inhibitors.¹²⁴ During aging, several inhibitory cell cycle regulatory genes are further increased,³⁵ further limiting any possibility for cell cycle re-entry and self-renewal. Moreover, those podocytes that do manage to enter the cell cycle are unable to assemble an efficient mitotic spindle, thereby triggering a mitotic catastrophe and causing DNA damage.¹²⁵ As in injured nonaged podocytes,¹²⁶ it is likely such DNA damage is an additional mechanism that limits proliferation and increases loss in aged podocytes due to changes in genes associated with cell cycle regulation (Table 1).

Podocyte Hypertrophy

Normal glomerular function requires podocytes to cover the underlying glomerular basement membrane (GBM). Glomerular volume increases up to 700% between childhood and adulthood, at a rate of 2.7% per year, although the mechanisms have not been well defined.²⁹ This is accompanied by an increase in total podocyte cell mass, at a rate of 1.8% per year. However, as described above, podocyte numbers decrease with age and glomerular volume increases, resulting in reduced podocyte density.²⁸ This creates a mismatch between the number of podocytes and the increasing glomerular volume, and, because podocytes cannot proliferate to increase their number, they cannot compensate for this mismatch.¹²⁴ Thus, to continue to cover the GBM during aging, podocytes rely on compensatory hypertrophy, defined as an increase in size due to an increase in protein to DNA ratio (Figure 2).³⁰ Although the average human podocyte volume is 335±136 µm³,¹⁰⁶ the podocyte nuclear volume increases at a rate of 2% per year, and average podocyte cell volume increases by 3.2% per year.²⁸

Wiggins et al. showed in rats that to cover the increased surface area, aging podocytes undergo five stages of hypertrophy.³⁰ Stage 1 is the normal adult podocyte state; stage 2 is characterized by nonstressed hypertrophy with normal function; stage 3 features adaptive hypertrophy, in which podocyte function remains normal despite signs of podocyte stress; stage 4 involves decompensated hypertrophy, in which podocytes are stressed, have decreased machinery for normal function, and start losing function and exhibiting signs of cellular stress; and stage 5, in which failure of podocyte hypertrophy to meet the size demand leads to both relative and absolute podocyte loss. Over time, this compensatory hypertrophy becomes maladaptive, leading to glomerulosclerosis.^25,28,30 In particular, GBM lacking podocytes has been critically linked to the scarring process.¹²⁷ It is also noteworthy that unlike podocytes, two other glomerular cell types, endothelial and mesangial cells, proliferate in response to increased glomerular size.²⁸

At the molecular level, the mechanisms of podocyte hypertrophy in the context of aging are less well understood. It is unclear to what extent they parallel the mechanisms of hypertrophy in nonaged podocytes. However, using cues from hypertrophy studies in nonaged podocytes, candidate pathways for aged podocytes include increased p21, p27, TGF-β, mammalian target of rapamycin (mTOR), and IL-6.³⁵ We acknowledge that aging is inherently confounded with other biologic pathways. These include growth of the glomerulus as it adapts to changes in filtration needs. To further dissect these age-dependent adaptations, future aging studies would benefit from including several time points over the lifespan of the kidney.

Changes in Canonical Podocyte Gene Expression

Canonical genes are defined as proteins that are constitutively active in a particular cell type and are required for its unique shape and function. Podocytes contain numerous canonical genes, many of which are highly specific to this terminally differentiated epithelial cell type. The normal functioning of the slit diaphragm as a size and charge barrier requires constitutive expression of a set of specific proteins.¹²⁸ Slit diaphragms are highly specialized adherens junctions that contain unique membrane proteins, such as nephrin and podocin; typical adherens junction proteins, such as cadherins and catenins; and scaffold proteins such as ZO-1, CD2AP, and MAGI-2.¹²⁹ Decreased levels of these proteins lead to impaired signaling to the actin cytoskeleton, resulting in weakening of the foot processes, proteinuria, and at times, glomerulosclerosis.¹³⁰ The expression of several critical proteins that constitute the slit diaphragm is significantly reduced in aged podocytes (Table 1). For example, Wiggins et al.³⁰ showed that in 24-month-old rats (approximately age 60 years in humans), expression for Tcf21, Nphs1, Col4a5, Glepp1, and Wt1 was strongly reduced. Similarly, podocytes of aged mice have decreased mRNA and protein expression of many critical slit diaphragm, tight junction, and other canonical podocyte genes (Table 1).^23,30,35,58 Considering the role of these proteins in nonaged podocytes, it is likely that expression changes will biologically affect the healthspan of podocytes as they progressively age.

Foot Process Effacement and Podocyte Stress

In effacement, a characteristic feature of podocyte injury, podocytes undergo a flattening in shape. It is a dynamic, ATP-consuming signaling-based process that involves the rearrangement of the actin cytoskeleton and integrin–focal adhesion dynamics. Effacement is accompanied by loss of the interdigitating pattern of the foot processes between neighboring podocytes, loss of adherence to the GBM, and loss of the subpodocyte space between the podocyte cell body and the GBM.^128,131,132 Foot process effacement has been associated with reduced GFR due to changes in the ultrafiltration coefficient Kf.¹³³ The same process is observed in aged human kidneys, in which the remaining podocytes in sclerosed glomeruli tend to be effaced, likely contributing to a decrease in kidney function (Figure 2).

The causes of age-associated podocyte effacement are not fully elucidated, but clearly involve remodeling of the cytoskeleton. In fact, many proteins involved in podocyte effacement in different clinical settings are also altered during aging.^{23,35,134,135} RNA-seq data³⁵ identified actin and several other regulators involved in the regulation of actin polymerization and cell motility as altered in aged podocytes⁶ (Table 1). One major player appears to be desmin, a type III intermediate filament protein. Desmin connects the nuclear and cell membranes with actin filaments and microtubules and contributes to cell morphology and mechanics. Although desmin is present in healthy podocytes, it is increased on podocyte stress, and in aging.^27,30 More importantly, increased desmin in aged podocytes in rats and mice precedes glomerulosclerosis, suggesting its dysregulation is a critical prerequisite for podocyte dysfunction.^27,30 Together, these studies suggest the cytoskeleton is significantly affected during the aging process, and these effects likely contribute to marked disturbances in normal podocyte physiology in the aged kidney.

Reduced Synthetic Function and Reduced Crosstalk with Other Glomerular Cell Types

Resident glomerular cells communicate to maintain glomerular integrity and function. For example, a critical podocyte synthetic function is the production of angiopoietic factors required for the survival and health of neighboring glomerular endothelial cells.¹³⁶ This crosstalk is best exemplified by a meta-analysis of recent RNA-seq data.³⁵ We identified 55 ligand-receptor interactions that were specifically enriched in healthy podocytes. Although the individual contributions of these to podocyte health require evaluation, it is noteworthy that 26 of these receptor-ligand pairs were significantly reduced in aged podocytes (Table 1). These data suggest that aging podocytes are characterized by the lack or reduced activity of autocrine and perhaps paracrine signaling networks, such as the VEGF signaling pathway (Figure 3).³⁵ This may explain why glomerular endothelial cell numbers are reduced in aged rats.¹⁰⁵ Surprisingly, the opposite, an upregulation of ligand-receptor pairs in aged podocytes, was not observed, suggesting podocyte aging is rather characterized by the disappearance of homeostatic signaling events. Further studies are needed to assess the biologic effects of reduced podocyte ligand-receptor pairs on podocytes themselves and on neighboring glomerular cells in aging.

Figure 3.

Loss of autocrine and paracrine functions in aged podocytes. (A) In young human and rodent glomeruli, podocytes produce ligands (represented by circles, squares, and rectangles) that bind to receptors on podocytes for autocrine effects to maintain podocyte health. Likewise, podocyte ligands bind receptors on neighboring parietal epithelial cells and fenestrated glomerular endothelial cells for paracrine effects that maintain their health. (B) In the aged mouse kidney, data support a decrease in podocyte autocrine and paracrine functions, which adversely affects the health of podocytes, parietal epithelial cells, and glomerular endothelial cells. This is further compounded by an increased inflammatory and senescent-associated secretory phenotype (SASP), which in aged podocytes exert an injurious autocrine effect on podocytes and paracrine effect on neighboring parietal epithelial cells and glomerular endothelial cells. Created with BioRender.com.

Mechanisms Contributing to Podocyte Aging

In previous sections, we discussed multiple cellular and molecular manifestations of podocyte aging. However, when addressing the question of why podocytes age, at least four considerations that should be heeded:

• The mechanisms underlying the podocyte aging process likely involve common aging-related factors shared with many other cell types (e.g., senescence, loss of autoph agy, p53, mTOR, NAD, AMPK, and NF-κB), and podocyte-specific factors (e.g., changes to specific transcription factors such as Grh2). Some of those are known (see below); others remain to be discovered.

• Podocyte aging is a multifactorial process/pathway and not simply linear. Yet, the interactions between the pathways, their hierarchy, and chronology are mostly unexplored.

• Although not firmly established, podocyte aging is almost certainly accelerated by age-related comorbid stressors on podocytes, such as obesity, diabetes, and hypertension, and environmental factors. These might act using pathways shared between aging and injury (e.g., mTOR, p16, AMPK signaling) or they could be distinct, culminating in the ultimate phenotype.

• Crosstalk likely exists between the different glomerular cell types. For example, aged podocytes might elicit changes in the endothelial cells and vice versa (Figure 3).

Below we discuss published candidate mechanisms and their contribution to podocyte aging (Figure 4).

Figure 4.

Summary of mechanisms and subsequent consequences in aged podocytes. Multiple mechanisms (listed in the red box) contribute to podocyte aging in humans and rodents. These molecular and cellular changes decrease the podocyte’s lifespan and healthspan each leading to physiologic and pathologic consequences. Created with BioRender.com.

Senescence, Senescence-associated Secreted Phenotypes, and Inflammasome

Cellular senescence is a permanent cell cycle arrest, with telomere shortening and macromolecule accumulation, and is multifactorial in origin.¹³⁷ Senescence develops acutely as a safeguard to prevent damaged cells from proliferating and replicating damaged DNA. Because senescent cells are resistant to apoptosis, they tend not to be cleared by macrophages, but instead accumulate within an organ. Chronic accumulation of senescent cells is damaging to an organ, largely because these cells release proteins comprising the senescence-associated secreted phenotype and other factors, which in turn directly or indirectly damage the surrounding cells.¹³⁸ Although short-term exposure to these proteins is beneficial to cellular regeneration in certain tissues, chronic exposure can induce tissue damage,¹³⁹ and also induce senescence of neighboring cells.¹⁴⁰ The clinical relevance of the accumulation of senescent cells in aged mice was best illustrated in the INK-ATTAC mouse model that specifically ablated senescent (i.e., p16^INK4-positive) cells.¹⁴¹ These aged mice exhibited improved kidney health and a reduction of age-associated glomerulosclerosis.

Although the classic senescence markers p16^INK4 and SA-ßgal have been detected occasionally in podocytes from healthy nonaged human glomeruli,¹⁴² they are clearly increased in healthy aged kidneys and in nonaged kidneys with superimposed glomerular or diabetic kidney disease.^39,142,143 Similarly, podocyte senescence has also been demonstrated in aged mice.^23,144–146 Our RNA-seq study of aged podocytes identified many senescence-associated genes, including 11 SASP genes (Table 1).³⁵ Surprisingly, classic inflammation-associated genes, and components of the inflammasome pathway, were significantly increased (Table 1).³⁵ Both Nfkb1 and Nfkb2 are increased in aged podocytes and have been shown to be involved in glomerular aging.¹⁴⁷ Thus, it is likely they are the transcriptional drivers for this inflammatory phenotype. However, because there is little recruitment of immune cells into aging glomeruli,¹⁴⁸ aging podocytes are likely in a state of sterile inflammation (i.e., inflammation not caused by infection).^149,150

Mitochondrial Dysfunction, Oxidative Stress, and Reduced Metabolism

Mitochondria generate most of the chemical energy needed for cellular homeostasis, and their dysfunction is often coupled to an increase in reactive oxygen species (ROS), a hallmark of aging in many cell types. We reported that in mice aged 24 months, podocytes had extensive mitochondrial damage, including disruption of the inner mitochondrial membranes, loss of cristae, changes to the mitochondrial matrix density, outer mitochondrial rupture, and increased expression of the ROS marker Nox4.²³ These changes were progressive, because mitochondria of 26-month-old mice were even more severely affected. A transcriptomic study³⁵ of aged mouse podocytes showed global downregulation of genes involved in mitochondrial oxidative phosphorylation (OXPHOS), including mitochondrial complexes I–IV, and the upregulation of several genes known to be activated by ROS, supporting a concomitant increase in ROS production (Table 1).

In addition, key metabolic pathways, including OXPHOS, the citric acid cycle, fatty acid oxidation (FAO), and amino acid catabolism, were reduced in aged podocytes (Table 1).³⁵ Such changes in metabolism are well described for a majority of aged cells. Yet, in podocytes, the focus has largely been on OXPHOS.¹⁵¹ Indeed, aged podocytes show dramatically reduced expression of enzymes involved in energy production via OXPHOS (Table 1). In particular, the peroxisome proliferator-activated receptor γ coactivator 1-α, a central regulator of mitochondrial biogenesis and OXPHOS,^152,153 was significantly downregulated in aged podocytes, and may thus be responsible for the overall decline in the mitochondrial energy production in aged podocytes.

It has recently been reported that anaerobic glycolysis—not OXPHOS—is the predominant energy source of podocytes.¹⁵⁴ Surprisingly, we did not observe significant age-dependent changes in genes of the glycolysis pathway.²³ We hypothesize that a reduction in OXPHOS might render aged podocytes even more reliant on glycolysis for energy production. This may be even further aggravated by the observed downregulation of Fatty Acid Oxidation (FAO) genes (Table 1). Podocytes primarily oxidize fatty acids instead of glutamine or pyruvate,¹⁵⁴ and inhibition of FAO has been shown to increase the susceptibility of podocytes to palmitic acid–induced cell death in diabetic nephropathy.¹⁵⁵ A similar process may be at play in podocyte aging. Taken together, decreased energy sources and metabolism accompanied by increased ROS likely cause progressive injury to aging podocytes.

Transcriptional Changes

Transcriptional regulation is critical for normal podocyte development and homeostasis.¹⁵⁶ In line with the observed DNA damage in aged podocytes, the transcription factor p53 is upregulated, although this has not been functionally evaluated. In contrast, the majority of other transcription factors studied decrease in aged podocytes. A virtual inference of protein activity by enriched regulon (VIPER) analysis,¹⁵⁷ which studies transcription factor gene expression and their regulatory networks, showed that many candidate transcription factors, their regulators, and their downstream targets were significantly downregulated in the aging podocyte (Table 1).³⁵ One of the best examples, the transcription factor CAAT enhancer-binding protein alpha (Cebpα), a member of the basic-region leucine zipper family of proteins, is highly expressed in young podocytes¹⁵⁸ but decreases in 20-month-old mice.¹⁴⁴ Functionally, podocyte-specific deletion of Cebpα in aged mice induces podocyte senescence (increased p16 and SA-ßgal staining), reduced podocyte numbers, and reduced expression of several canonical podocyte markers.¹⁴⁴ Moreover, in vitro studies show that adriamycin-induced aging of immortalized mouse podocytes reduces Cebpα levels and that Cebpα overexpression reversed these changes.¹⁴⁴ Although the precise downstream mechanisms are still unclear, the authors hypothesize that Cebpα acts via the AMPK-mTOR pathway. Finally, the RE1-silencing transcription factor protects podocytes from injury in aging mice.¹⁵⁹ Taken together, a robust decrease in transcription factors and their downstream targets is likely a major cause underlying the morphologic and physiologic changes in aged podocytes.

Sirtuins

The sirtuins (Sirt) comprise a family of nicotinamide adenine dinucleotide–dependent deacetylases that are involved in several cellular processes such as autophagy, and act largely through transcription repression by deacetylation.¹⁶⁰ Sirt1 and Sirt6 have been implicated in normal podocyte function. A decrease in Sirt1 expression augments podocyte aging,¹⁴⁵ and knockdown of Sirt1 in mouse podocytes resulted in an overall reduction in the number of podocytes, increased glomerulosclerosis and albuminuria, and reduced levels of key podocyte proteins.¹⁴⁵ The changes were accompanied by elevated expression of the senescent markers p16^INK4a and p19^ARF. The underlying mechanism for Sirt1 is unknown, but the authors speculated that the effect on podocyte aging is dysregulation of proliferator-activated receptor γ coactivator 1-α, FOXO3, FOXO4, and NF-kB. Yet, Sirt1 knockdown mice also show elevated urinary levels of the cellular oxidative stress marker 8-OHdG, suggesting the overall effects of reduced Sirt1 levels are caused by increased oxidative stress and mitochondrial injury, as has been observed in other cell types.¹⁶¹ The deletion of Sirt6 from mouse podocytes were performed in 7-month-old mice (not considered aged), and studies need to be analyzed in older cohorts to solidify the authors’ claim that Sirt 6 is critical for the podocyte aging process per se.¹⁶²

Finally, the missing aspect of these studies is that no study actually shows an age-dependent downregulation of Sirt mRNAs. In fact, our transcriptomic analysis showed that the only change in sirtuins between old and young podocytes was an age-dependent upregulation of Sirt7 (1.9-fold). Although counterintuitive, this may be a protective/adaptive response, as Sirt7-deficient mice show an accelerated aging phenotype in other tissues.¹⁶³ Finally, sirtuins have been implicated in the beneficial effects of calorie restriction via the regulation of autophagy, which are described below.

Reduced Autophagy

Autophagy is a healthy, biologic “housekeeping” process that helps cells maintain homeostatic balance between the synthesis, degradation, and recycling of cellular proteins and organelles, with the ultimate purpose of supplying nutrients for survival.¹⁶⁴ Podocytes, as other postmitotic cells, have a very high basal autophagocytic activity, which is essential for energy homeostasis.¹⁶⁵ When podocytes are unable to perform autophagy, damage occurs, resembling an aging phenotype. Moreover, because decreased autophagy enhances the extent of injury in podocyte diseases such as FSGS,¹⁶⁵ it is likely that decreased autophagy is augmenting podocyte injury in aged patients with podocyte disorders.

This correlation is also supported by the podocyte-specific deletion of autophagy-related 5 (Atg5), one of the key regulatory genes in the autophagy process. These mice exhibit reduced autophagy and characteristics of enhanced aging, namely, a decline in podocyte numbers and accumulation of damaged organelles and protein aggregates.¹⁶⁵ Surprisingly, this phenotype appears to be independent of one of the major regulators of autophagy, the mTOR complex. This pathway has been shown to inhibit autophagy⁹⁸ and, as expected, mTOR activity is higher in podocytes of aged mice.⁹⁸ In aged podocytes, expression of Prkaa2, which encodes a protein that is part of the AMPK complex, and Pkhd1¹⁶⁶—two known inhibitors of mTOR activity—is downregulated, whereas Card11, an activator of mTOR signaling,¹⁶⁷ is upregulated 6.9-fold.³⁵ Yet, administering the mTOR inhibitor rapamycin to 26.5-month-old mice (approximately age 70+ years in humans)^47,48 reduced mTOR activity, but did not alter podocyte density.⁹⁸ Thus, further studies are needed to understand the mTOR-independent regulation of autophagy in aged podocytes and whether this can be harnessed toward antiaging approaches.

Mitigating Podocyte Aging

There is a large effort to identify treatments and drugs that mitigate the effects of aging in humans. In our opinion, alleviating or reversing podocyte aging in the healthy aged kidney should not be the focus for therapy development. Instead, we believe the goal should be maintaining and improving podocyte lifespan and healthspan in states of stress. As previously reported,³² this would allow one to interfere with the acceleration of the aging process mediated by diseases. Such diseases include a wide array of disorders that either directly injure podocytes (e.g., FSGS and diabetic kidney disease), glomerular diseases in which podocytes are secondarily injured (e.g., IgA nephropathy, hypertension, obesity, and vasculitis), and conditions (and treatment of conditions) that result in a reduction of kidney mass or function (e.g., as a result of surgery, transplantation of older kidneys, and CKD). Thus, mitigating aging in podocytes will likely (as has been seen in other cell types¹⁶⁸) require different drug combinations or personalized, disease-specific therapies. To date, several approaches or compounds have been reported to reduce podocyte aging.

Calorie Restriction

This approach is the oldest and most robust way to extend rodent life span systemically. In the kidney, Wiggins et al.³⁰ showed that calorie restriction prevented age-associated podocyte hypertrophy in rats. Moreover, aged calorie-restricted rats exhibited lower mRNA levels of Wt1, Podxl, Nphs1, and Glepp1 compared with age-matched rats given an ad libidum diet, and they also displayed upregulation of the podocyte stress marker desmin.³⁰ Similarly, we reported that calorie restriction of male F344 rats during aging limited their podocyte loss compared with their counterparts fed an ad libidum diet.⁴³ The precise mechanisms underlying the effects of calorie restriction on podocyte biology is unknown, as is whether those mechanisms differ from those described for other aging cell types. One possible link between calorie restriction and its effects in podocytes is the sirtuins, which (as mentioned above) have been implicated as effectors of calorie restriction.¹⁶⁹

SS-31

The mitochondrial antioxidant elamipretide (SS-31) is a small cell-permeable synthetic peptide that is in clinical trials for heart failure, primary mitochondrial myopathy, and other mitochondrial diseases.¹⁷⁰ SS-31 improves mitochondrial function by enhancing electron transfer through cytochrome c, improving OXPHOS coupling, increasing ATP synthesis, and reducing ROS production.^171,172 Interestingly, it is also linked to changes in Sirt1 levels.¹⁷³ To explore SS-31’s role in podocyte aging, 24-month-old mice were treated for 8 weeks with either SS-31 or vehicle²³; SS-31 preserved podocyte mitochondrial integrity, lowered the expression of the ROS-generating enzyme NOX4, and increased expression of genes encoding NduFA and Cox IV, enzymes in the mitochondria electron transport chain. At the cellular level, SS-31 reduced podocyte hypertrophy and foot process effacement, but did not affect podocyte density. It also reduced the injury marker desmin and preserved the expression of synaptopodin. Together, these results provide a strong rationale for the possibility of interfering with age-dependent decline in podocyte function. The most impressive aspect of this study was finding that despite administering the treatment to aged mice, it was still possible to preserve the remaining mitochondrial integrity and significantly improve podocyte health.

Rapamycin

The mTOR pathway regulates many essential cellular processes, including translation, transcription, and autophagy. Inhibiting mTOR induces autophagy, thereby sustaining cell survival under conditions of nutrient deprivation. It has been shown that increased mTOR signaling in podocytes causes disease and augments injury, whereas `reducing it specifically in podocytes limits injury.¹⁶⁴ However, as mentioned above, rapamycin treatment had no effect on podocyte density in aged mice.⁹⁸ However, it is noteworthy that, as in injury scenarios, inhibiting mTOR signaling can revert podocyte hypertrophy.¹⁷⁴

Senolytics

The cancer drugs quercetin and dasatinib inhibit a broad spectrum of protein kinases and tyrosine kinases and have been shown to reduce markers of aging.^175–177 This effect was also extended to the podocytes, in which administration of dasatinib and quercetin in a diabetes model of senescence improved kidney function, increased expression of WT1, and decreased p16 levels.¹⁷⁸

Endothelin Receptor

Darusentan is a selective endothelin ET_A receptor antagonist that is clinically evaluated as a treatment for congestive heart failure and hypertension. Interestingly, treating 23-month-old male Wistar rats with darusentan for 28 days resulted in substantial recovery of podocyte structure, accompanied by a decrease in expression of p21 and MMP9, but no changes in blood pressure or renin levels.¹⁷⁹

Novel Targets

Our recent RNA-seq analysis comparing young with aged podocytes,³⁵ using the Connectivity Map database,¹⁸⁰ identified potential interventions to reverse the aged podocyte phenotype. This analysis yielded candidate genes that have been implicated in podocyte diseases. For example, Stat5 has been shown to be upregulated in FSGS.¹⁸¹ The analysis also identified therapeutics that have previously been used, such as the poly (ADP-ribose) polymerase inhibitor PJ-34, which blocks ROS species generation and protects podocytes in diabetic nephropathy.¹⁸²

Because podocyte depletion is a critical pathophysiologic process in the aged kidney, one might ask if replacement can be augmented from the putative adult podocyte progenitors, glomerular parietal epithelial cells,^{34,58,183,184} and renin-producing cells.⁹⁹ Both progenitors are detected in individual aged glomeruli in which podocytes are depleted, albeit in low numbers.^{34,58,99,183,184} It is noteworthy that these aged glomeruli have a normal morphology with no scarring. However, both types of podocyte progenitors decline in number with advancing age, and they acquire additional adverse changes, such as mesenchymal transformation, that limit their progenitor function.^97,146 Thus, podocyte progenitor augmentation is not fully capable of countering the age-associated decrease in podocyte numbers. On the basis of this observation, boosting adult podocyte progenitors should be considered a target for future therapies to improve podocyte numbers.

Future Directions

Although certain pathways, such as senescence and genomic instability, are considered general hallmarks of aging, recent single-cell studies show changes occurring across multiple organs, and within each are cell type–specific changes.¹⁸⁵ Yet, a unique podocyte aging signature is not currently available. Fortunately, the emergence of single-cell technologies and an associated ever-evolving analysis toolbox are providing the opportunity to study cell types such as the podocyte that are less abundant and to resolve the aging process at an unprecedented level of detail.

To understand the molecular and cellular mechanisms of podocyte aging, future studies should integrate longitudinal transcriptomic, proteomic, metabolomic, and epigenetic changes, and translate experimental studies to clinical settings. This would result in the identification of specific aging signatures for podocytes and their temporal profile. Such signatures will likely include pathways implicated in the general aging process (e.g., insulin/IGF, p53, mTOR, AMPK, and NF-κB^147,186,187), and podocyte-specific ones. A holistic analysis of these pathways will allow researchers to determine if and how aging in podocytes differs from that of other cell types and whether developing therapeutics specifically targeting aged podocytes is possible. Are the pathways controlling podocyte lifespan similar to or different from those of podocyte healthspan?

A wide-open opportunity exists to fill an important clinical gap, namely our understanding of the mechanistic intersection between podocyte disease and aging. Definitive data remain scarce for determining if disease-related stressors might hasten natural podocyte aging,³² and/or if aging lowers the threshold to disease-induced podocyte injury. In practical terms, this means the efficacy of currently used clinical therapeutics for podocyte disorders should be thoroughly evaluated in elderly adults. This is particularly important if, as we suspect, pathways are common to podocyte injury and aging. Such knowledge will have substantial therapeutic implications in clinical practice for the elderly patient population with superimposed glomerular disease. Finally, new therapeutics limiting or even reversing the rate of podocyte aging might have beneficial effects in podocyte diseases that lack effective treatment options.

Disclosures

A. Shaw reports being empl oyed by Genentech; reports having an ownership interest in Roche; and reports other interests/relationships with NephCure. J. Pippin reports having patents and inventions with University of Washington TechTransfer Invention Licensing Unit. All remaining authors have nothing to disclose.

Funding

This work is supportedby National Institute of Diabetes and Digestive and Kidney Diseases grants 5 R01 DK 056799-10, 5 R01 DK 056799-12, and 1 R01 DK097598-01A1.