TGF-β1–p53 cooperativity regulates a profibrotic genomic program in the kidney: molecular mechanisms and clinical implications

Abstract

Chronic kidney disease affects >15% of the U.S. population and >850 million individuals worldwide. Fibrosis is the common outcome of many chronic renal disorders and, although the etiology varies (i.e., diabetes, hypertension, ischemia, acute injury, and urologic obstructive disorders), persistently elevated renal TGF-β1 levels result in the relentless progression of fibrotic disease. TGF-β1 orchestrates the multifaceted program of renal fibrogenesis involving proximal tubular dysfunction, failed epithelial recovery and redifferentiation, and subsequent tubulointerstitial fibrosis, eventually leading to chronic renal disease. Recent findings implicate p53 as a cofactor in the TGF-β1–induced signaling pathway and a transcriptional coregulator of several TGF-β1 profibrotic response genes by complexing with receptor-activated SMADs, which are homologous to the small worms (SMA) and Drosophilia mothers against decapentaplegic (MAD) gene families. The cooperative p53–TGF-β1 genomic cluster includes genes involved in cell growth control and extracellular matrix remodeling [e.g., plasminogen activator inhibitor-1 (PAI-1; serine protease inhibitor, clade E, member 1), connective tissue growth factor, and collagen I]. Although the molecular basis for this codependency is unclear, many TGF-β1–responsive genes possess p53 binding motifs. p53 up-regulation and increased p53 phosphorylation; moreover, they are evident in nephrotoxin- and ischemia/reperfusion-induced injury, diabetic nephropathy, ureteral obstructive disease, and kidney allograft rejection. Pharmacologic and genetic approaches that target p53 attenuate expression of the involved genes and mitigate the fibrotic response, confirming a key role for p53 in renal disorders. This review focuses on mechanisms whereby p53 functions as a transcriptional regulator within the TGF-β1 cluster with an emphasis on the potent fibrosis-promoting PAI-1 gene.—Higgins, C. E., Tang, J., Mian, B. M., Higgins, S. P., Gifford, C. C., Conti, D. J., Meldrum, K. K., Samarakoon, R., Higgins, P. J. TGF-β1–p53 cooperativity regulates a profibrotic genomic program in the kidney: molecular mechanisms and clinical implications.

Chronic kidney disease (CKD), a frequent outcome of repeated episodes of acute kidney injury (AKI), is a progressive incurable condition affecting >850 million people worldwide. Approximately 30–40 million adults in the US alone have some form of CKD (1, 2), with the incidence increasing most notably in younger individuals (3). The annual financial impact on the health care system for patients with CKD and end-stage renal disease is estimated at >$100 billion (4). Although the pathologic basis varies, lethal or recurring sublethal epithelial trauma and prolonged TGF-β1 pathway activation initiate and sustain the fibrotic process (5–7). Diabetes is the predominant pathophysiologic driver of CKD; other contributors include hypertension, sepsis, ischemia/reperfusion injury, obstructive nephropathy, metabolic disorders, and exposure to nephrotoxins (8–10).

Injury-induced proximal tubular dysfunction, failed epithelial redifferentiation, and subsequent tubulointerstitial fibrosis are key to the development of CKD (5, 11–16). Functional impairment following AKI may be transient (adaptive repair), reflecting compensatory regeneration of surviving tubular elements. Severe or persistent injury, however, triggers inflammatory cell infiltration, reduced microvascular density, epithelial dedifferentiation, and G₂/M growth arrest (11, 16–21) (Fig. 1A). Capillary rarefaction due to pericyte loss results in hypoxia, tubule degeneration, and nephron dropout (22, 23). G₂/M-stalled cells, moreover, secrete several major profibrotic effectors, including connective tissue growth factor (CTGF) and TGF-β1, that stimulate epithelial-mesenchymal crosstalk and fibroblast differentiation to matrix-producing myofibroblasts (Fig. 1B). Progressive disease eventually culminates in organ failure, for which dialysis or allograft transplantation are the only clinical options.

Figure 1.

Repair outcomes in the injured kidney. Mild or sublethal AKI initiates a process that involves resolution of inflammation and restoration of tubular architecture with a regain of renal function (A, left; B, top). In contrast, recurring AKI or severe tubular injury results in a prolonged inflammatory response, stalling of proximal tubular epithelial cells in G₂/M with increased expression of TGF-β1 and CTGF, recruitment of vascular pericytes into the interstitial compartment and differentiation into matrix-secreting myofibroblasts, tubular atrophy, capillary loss, failure to regenerate a functional epithelium, and progressive interstitial fibrosis (A, right; B, bottom).

THE RENAL FIBROTIC MICROENVIRONMENT

Increased deposition and reduced clearance of extracellular matrix (ECM) elements in the interstitial, glomerular, and vascular compartments are causative factors in renal scarring. The primary source of ECM synthesis during renal fibrogenesis are activated fibroblasts or myofibroblasts (24, 25) that drive the pathophysiology of maladaptive repair. Lineage tracing suggests that these disease-causative cells arise predominantly from glioma-associated oncogene homolog 1⁺–forkhead box (FOX) D1⁺ vascular pericytes and fibroblasts (26–36), a conclusion supported by sine oculis homeobox homolog 2 fate mapping of the injured renal epithelium (35, 37). TGF-β1 governs the myofibroblastic differentiation of pericytes and resident fibroblasts while coordinating a program of pathologic ECM synthesis and advancing fibrosis (38–44). Indeed, mice engineered to express epithelial-targeted, inducible, constitutively active TGF-β receptor I exhibit acute tubular injury and apoptosis, phenotypic dedifferentiation, synthesis of profibrotic factors, and interstitial inflammatory cell recruitment (45). In transplant-associated renal ischemia/reperfusion injury, TGF-β1 initiates early repair and regeneration but in the setting of chronic inflammation drives continual matrix deposition and remodeling, leading to fibrotic disease, compromised renal function, and eventual organ rejection (46). Excessive ECM accumulation also provides a source of damage-associated molecular pattern ligands for the inflammatory response TLR (47–50). Progressive fibrosis, moreover, leads to enhanced tissue stiffness. Loss of elasticity stimulates the conversion of latent to bioactive TGF-β1 as a result of integrin engagement and cellular force generation (51–53). This increasingly noncompliant TGF-β1–rich microenvironment promotes myofibroblastic differentiation while activating the Hippo pathway mechanosensitive effectors yes-associated protein (YAP) and transcriptional coactivator with PDZ binding motif (TAZ), which in turn reinforce expression of genes encoding the profibrotic factors plasminogen activator inhibitor-1 (PAI-1), endothelin-1, and CTGF (54–58). Increasing stromal stiffness and cytoskeletal tension create a feed-forward mechanosensitive circuit (59) contributing to myofibroblast persistence and continued organ damage. A threshold may be reached when progressive fibrosis becomes self-sustaining (in the absence of ongoing or intermittent injury), involving both cell autonomous and ECM-driven mechanisms and resulting in a permanent change in the mechanical properties of the supporting stroma that exacerbate disease progression (60).

INVOLVEMENT OF p53 IN RENAL DISEASE

Binding of TGF-β1 to its type II receptor drives dimerization with and subsequent phosphorylation of a type I (activin-like kinase) receptor that in turn phosphorylates receptor small worms and mothers against decapentaplegic homologs (SMADs) (predominately SMAD2 and 3 in fibrotic disease) at the distal C-terminal SxS motif. Phosphorylated (p-)SMADs form a complex with the shuttle SMAD4 for nuclear entry, activating the transcription of genes with SMAD binding motifs. Because SMAD2 does not bind DNA and DNA-SMAD3 interactions are of relatively low affinity, TGF-β1 target gene transcription requires the involvement of additionally recruited cofactors (61). One such coactivator is p53, a sequence-specific DNA-binding protein and a key regulator of the cell growth, apoptosis, and senescence programs as well as the cellular stress response (62–65). p53 is activated in the injured renal epithelium, initiating cell cycle arrest at the G₁ and G₂/M checkpoints depending on the participating effectors (e.g., ataxia telangiectasia mutated, ataxia telangiectasia and Rad3-related serine/threonine-protein kinase, casein kinase 1 and 2, p21, and TGF-β1) and the extent of tissue hypoxia (18, 63, 66). TGF-β1 signaling in the damaged kidney increases p53 levels and phosphorylation, particularly at p53^S9/15 (Fig. 2A), triggering p53-SMAD2 and 3 interactions and resulting in transcription of the growth inhibitor p21 and subsequent p21-dependent G₁ arrest (62, 67, 68). p53 up-regulation in hypoxic tubular cells also suppresses cyclin-dependent kinase 1, cyclin B1, and cyclin D1 expression, potentially involving the p53–retinoblastoma protein (RB)-like, E2F and multi-vulval class B complex (69) and thereby increasing residence time in G₂/M (50). p53 induction and increased p53^S15 phosphorylation are prominent in nephrotoxin-induced AKI, ischemia/reperfusion injury, unilateral ureteral obstruction (UUO)–mediated fibrosis, and renal allograft rejection (21, 63, 70–74) (Fig. 2B–E). p53 is also implicated in the development of insulin resistance and diabetes (75). Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.genome.jp/kegg/kegg1.html) analysis of the transcriptome of diabetic and nondiabetic mice indicated, in fact, that significant differentially expressed genes closely associate with the p53 signaling network as well as the MAPK and TGF-β pathways (76). Pharmacologic or genetic p53 inhibition appears to attenuate and p53 activation exacerbates the diabetic phenotype, suggesting that targeting pathways and genes regulated by p53 may have therapeutic promise.

Figure 2.

The p53 protein phosphotome. A) Schematic illustrating p53 phosphorylation sites and the specific kinases involved. Kinases responsible for p53^S9 and p53^S15 phosphorylation in response to TGF-β1 are highlighted in orange. B) Nuclear p-p53¹⁵ immunoreactivity was not evident in tissue sections from a normal (4-wk sham-treated) rat kidney. C) In contrast, tubular dilation and frequent nuclear p-p53^S15 localization (red signal; white arrows) was prominent in the flattened epithelium in kidneys from rats 4 wk post-UUO. D, E) Compared to a normal-appearing patient allograft (D) in which only infrequent p-p53^S15 tubular epithelial cells were detectable, abundant nuclear and cytoplasmic p-p53^S15 immunoreactivity was readily apparent (brown stain; black arrows) in a renal transplant exhibiting dysmorphic tubules with a flattened and occasionally denuded epithelium and a markedly expanded interstitial region (E). ATM, ataxia telangiectasia mutated; ATR ataxia telangiectasia and Rad3-related serine/threonine-protein kinase; CDK, cyclin-dependent kinase 1; CK, casein kinase; DBD, DNA binding domain; PRD, proline-rich domain; REG, regulatory domain; TAD, transactivating domain; TET, tetramerization domain.

TGF-β1 stimulates p53^S9/15 phosphorylation, albeit with different kinetics and functional implications (73, 74). p-p53^S15 is involved in p53 stabilization and transactivation of p53-responsive promoters (including PAI-1), whereas p-p53^S9 participates in p53-SMAD signaling (77, 78). Pharmacologic inhibition (with pifithrin-α) or genetic deletion in the proximal tubular epithelium confirmed involvement of p53 in prolonged G₂/M residence and the fibrotic response to cisplatin, UUO, or ischemic injury (18, 21, 79, 80). Similarly, cisplatin- or bilateral ischemia–induced AKI in streptozotocin-treated mice or genetically susceptible (Akita) diabetic animals is significantly diminished by pifithrin-α, p53 siRNA, or proximal tubule–targeted p53 ablation (81). Intravenously delivered p53 siRNA, moreover, mitigates the structural and functional damage to transplanted kidneys upon ischemia/reperfusion injury in 2 syngeneic rat models (82), suggesting the potential clinical utility of targeting p53 in patients with failing renal allografts.

Although the mechanism whereby p53 contributes to renal disease remains to be fully clarified, 1 level of control involves the dual-site mouse double minute 2 homolog (MDM2) E3 ubiquitin ligase, a major negative regulator of p53 transcriptional activity and stability (83). MDM2 deletion in the tubular epithelium of Pax8rtTA-cre/MDM2^f/f mice indicated that specific renal anomalies associated with MDM2 loss (tubular damage; cell death; expression of the injury markers kidney injury molecule-1, neutrophil gelatinase-associated lipocalin, and tissue inhibitor of metalloproteinases-2; and impaired renal function) are p53 mediated (84). This was confirmed using an inducible MDM2/p53–double knockout strain (Pax8rtTA-cre/MDM2^f/f/p53^f/f); upon doxycycline administration, the typical single MDM2^−/−-dependent abnormalities were no longer apparent in the p53-MDM2–codeficient mice.

Although TGF-β1 activates p53, the mechanism is complex, involving increases in both MDM2 and the p53 stabilizing factor Numb (85–89). TGF-β1 stimulates Numb expression in proximal tubular cells, and Numb is elevated in the fibrotic kidney. Adenoviral-mediated Numb delivery up-regulates p21 and TGF-β1–CTGF expression, increases p53 and p-p53 levels, and promotes G₂/M stalling (89). G₂/M arrest, transcription of profibrotic genes, and the extent of fibrogenesis are reduced in proximal tubule Numb-deficient mice following UUO or ischemia/reperfusion injury (88, 89). Pifithrin-α attenuates Numb-induced G₂/M arrest (89), suggesting the involvement of p53. p53 and Numb are each targeted for ubiquitination and subsequent degradation by MDM2 (90, 91). One model proposes that Numb dissociates MDM2-p53 complexes or inhibits complex formation (89), effectively maintaining p53 engagement of target genes. Numb also promotes ubiquitination of phosphatase and tensin homolog, a regulator of the C-terminal SMAD phosphatase protein phosphatase, Mg2⁺/Mn2⁺ dependent 1A (92–95). Collectively, these findings suggest that Numb-mediated attenuation of MDM2-dependent p53 degradation coupled with phosphatase and tensin homolog loss and the resulting maintenance of p-SMAD2 and 3 signaling may predispose the kidney to development of tubulointerstitial fibrosis (94, 95).

p53 IS REQUIRED FOR EXPRESSION OF A SUBSET OF TGF-β1 PROFIBROTIC TARGET GENES

The cooperative p53–TGF-β1 genomic cluster includes genes involved in cell growth control and ECM remodeling (96–99). Although the molecular basis for this codependency requires clarification, many TGF-β1–responsive genes possess p53 binding motifs (100, 101). p53 participates in the transcription of several renal disease–causative genes, including PAI-1, CTGF, and collagen I, underscoring the complexities of noncanonical pathways in TGF-β1–induced fibrosis (21, 73, 74, 96, 102). Similar p53 and TGF-β1 crosstalk jointly regulates PAI-1 expression in γ-radiation–initiated fibrosis (103). Mutation of the p53 binding site in the PAI-1 promoter (104–106) inhibits γ-radiation–induced PAI-1 transcription and attenuates the dual γ-radiation + TGF-β1 synergy (103). Consistent with criteria for identification of p53 target genes (73), γ-radiation did not induce PAI-1 expression in p53-deficient cells, whereas vector-driven expression of p53 in a p53-null genetic background largely restored the PAI-1 response to γ-radiation + TGF-β1 costimulation. Interestingly, in a different cell type, γ-radiation also stimulates PAI-1 expression as well as the binding of p53 and the avian reticuloendotheliosis oncogene homolog A (RelA) (p65) to the PAI-1 gene. p53 knockdown, as expected, mitigates (whereas RelA silencing enhances) radiation-dependent PAI-1 expression (107). Because p300–cAMP response element binding protein-binding protein (CBP) coactivator complexes mediate expression of genes competitively regulated by both p53 and RelA, 1 model suggests that RelA interactions with p300-CBP may inhibit target gene expression, whereas binding of p53 to p300-CBP results in loss of (suppressive) NF-κB activity (107, 108). Gain-of-function p53 variants also promote acquisition of the fibroproliferative phenotype by complexing with and suppressing the activity of the PAI-1-CD44-cyclin D1 transcriptional repressor Kruppel-like factor 17 (109).

TGF-β1 impacts p53 activity largely by serine phosphorylation in the N-terminal transactivation domain and serine/lysine acetylation and methylation in the tetramerization and regulatory domains in the C terminus (110), facilitating the binding of p53-SMAD3 complexes to the promoter of the potent profibrotic PAI-1 gene (73). PAI-1, 1 of the most highly up-regulated genes in the TGF-β1–stimulated proximal tubular epithelial cell dataset (73, 76), is a canonical and noncanonical TGF-β1 pathway-response gene (Fig. 3) (111–115) and a pifithrin-α–sensitive p53 target (Fig. 4A) (111–113). Depending on the specific renal epithelial cell, TGF-β1–induced PAI-1 expression ranges from 15- to >90-fold and approximates >100-fold in various forms of renal fibrosis, in which it contributes to the creation of a profibrotic milieu by negatively regulating the plasmin-generating urokinase-type plasminogen activator/tissue-type plasminogen activator pericellular proteolytic cascade (Figs. 4B–D and 5), eventually altering stromal mechanics. PAI-1 is specifically implicated in the renal and vascular complications of diabetes. Indeed, hierarchical clustering of differentially expressed genes between diabetic and nondiabetic mice positioned PAI-1 among the top 3 of 15 protein-protein interaction nodes based on degree, between-ness, and subgraph centrality (76, 116).

Figure 3.

Convergence of the TGF-β and collateral signaling pathways regulates transcription of the profibrotic PAI-1 and CTGF genes. TGF-β1 activates a canonical network, leading to SMAD2 and 3-dependent PAI-1 expression. Collateral noncanonical pathway mobilization (i.e., Hippo) also contributes to profibrotic gene activation (CTGF, PAI-1) via YAP and TAZ interaction with receptor-activated SMAD2 and 3. KEGG database (111–113). CK, casein kinase; Gli1, glioma-associated oncogene homolog 1.

Figure 4.

Spectrum of p53-dependent genes and the PAI-1 interactome in human proximal tubular epithelial cells. Transcriptome profiling indicated that PAI-1 [serine protease inhibitor, clade E, member 1 (SERPINE1)] is a prominent member of the TGF-β1–induced, pifithrin-α–sensitive p53 target gene subset (A), a major hub in the regulation of the immediate pericellular proteolytic cascade (B), and, more distally, functions in the control of matrix adhesion (C). Proximal tubular cells were preincubated with pifithrin-α or left untreated prior to stimulation with TGF-β1 for 24 h; p53-dependent genes were identified by comparative expression profiling (73) (A). PAI-1 suppresses the conversion of plasminogen (Plg) to plasmin by binding to and inhibiting the catalytic activity of the urokinase-type plasminogen activator (PLAU) and tissue-type plasminogen activator (PLAT), effectively limiting the extent and locale of stromal proteolysis (B). This interaction promotes matrix accumulation and onset of fibrotic disease in several organ systems, including the kidney. PAI-1 also regulates cellular attachment and migration largely by altering the conformation of the PLAU–PLAU receptor (PLAUR) complex with its associated integrins. This results in a decrease in their affinity for ECM binding sites, thereby promoting cellular detachment (B, C) (114, 115). Search of the Gene Ontology (GO) Database (http://geneontology.org/) provided several key biologic processes related to fibrotic disease (D) that are impacted by PAI-1 and its interacting network of proteins. Nedd, neural precursor cell expressed developmentally down-regulated protein; Tgfbr1, TGF-β receptor I.

Figure 5.

PAI-1 negatively regulates the coagulation and cellular proteolytic cascades. PAI-1, induced as a consequence of tissue damage (likely in response to injury-related TGF-β1 release) or during activation of the extrinsic or intrinsic pathways of the coagulation cascade, inhibits both urokinase and tissue-type plasminogen activators. Fundamental cellular processes, including cell-to-ECM adhesion, migration, and proliferation, are affected while the generation of plasmin and the downstream fibrinolytic and pericellular proteolytic cascades are suppressed. KEGG database (111–115). tPA, tissue-type plasminogen activator; uPA, urokinase-type plasminogen activator.

A total of 2 canonical decameric bp p53 binding half-site motifs in the PAI-1 promoter map to nt −224 to −204 relative to the transcription start site (104–106); both conform to the >90 p53 transcription factor DNA binding site algorithm score threshold required for the identification of potential p53-responsive genes (117). PAI-1 is, in fact, a member of the “high-confidence” cohort of 151 p53 target genes (118) and number 131 among 343 of the most prominent p53-activated genes identified in multiple datasets (64). In view of the extensive repertoire of direct and indirect p53-responsive genes (e.g., 109), stringent criteria are required for identification of actual p53 genomic targets (reviewed in 73, 119, 120). p53, however, can also transactivate genes using noncanonical sequences with transcriptional functionality independent of the 0–13 bp spacer and half-site requirements (64). The 2 10-bp p53 elements in the PAI-1 promoter, which provide a platform for p53 docking in a typical dimer-of-dimers configuration, are not separated by a nucleotide spacer (105, 106). These data as well as the inclusion of PAI-1 as a member of the p53 target database, the topographic proximity of the p53 binding motifs to the transcription start site, the chromatin immunoprecipitation–confirmed p53 occupancy of the PAI-1 promoter, and the pharmacologic and genetic findings reviewed earlier (71, 73) conclusively implicate p53 as a major cofactor in PAI-1 gene transcription.

Modeling p53-SMAD interactions with accessory coactivators

Several p53-response genes are also targeted by FOXO as well as the E box recognition factor microphthalmia-associated transcription factor and members of the upstream stimulatory factor (USF) family, suggesting that p53 coordinates with FOXO, microphthalmia-associated transcription factor, and USF to impact transcriptional outcomes (121). USF2 occupancy of the PAI-1 promoter proximal E box 2 (PE2) region E box (CACGTG) site, which adjoins the 3 5′ SMAD binding elements (SBEs) that mediate TGF-β1–induced gene expression (Fig. 6A), is essential for the growth state–dependent transcriptional activation of the PAI-1 gene (71, 122–124). USF1 and -2 involvement in PAI-1 gene control has significant implications regarding both renal fibrosis and p53 function (124–130). Indeed, USF1 directly interacts with p53 and inhibits MDM2-mediated p53 nuclear export and degradation, resulting in p53 protein stabilization (131). USF1, therefore, may function cooperatively with p53 to maintain p53 function as a coactivator of profibrotic gene transcription, particularly at promoters containing E box motifs and SMAD binding elements (e.g., PAI-1).

Figure 6.

Construction of a transcriptional competent complex on the PAI-1 promoter. The PE2 region E box is a docking site for USF1 and -2. A) Occupancy of the immediate 5′′ upstream SBEs with SMAD2, 3, and 4 and complex formation with p53 is required for TGF-β1–induced PAI-1 expression (73, 122–124). The Mad homology (MH) 1 domain of SMAD2, 3, and 4 contains a β-hairpin structure for binding to DNA, although SMAD2 has a relatively low affinity for the SBE. B) Several bHLH-LZ proteins (including USF) redirect DNA minor grove orientation, promoting interactions between tetramerized p53 bound to its half-site motifs, with SMAD2 and 3 tethered to the PE2-region SBEs. This conformation facilitates interactions between the MH1 N-terminal domain of SMAD2 and the p–N-terminal transactivation domain of p53 (79, 96) and, perhaps, between the C terminus of p53 and the MH2 region of SMAD3. It appears that both SMAD2 and 3 but not SMAD4 bind p53 (68). p53^S15 phosphorylation recruits the histone acetylase p300 (77). p53 acetylation is evident within 1 h of TGF-β1 stimulation of human proximal tubular epithelial cells. This is consistent with the increased complex formation between p53 and p300 (73) and correlates with the time frame of PAI-1 transcriptional activation. The Hippo pathway transcriptional effectors YAP and TAZ interact with SMAD2 and 3 in response to TGF-β1, whereas the AP-1 transcription factor appears to form a complex with YAP-TAZ-TEAD.

Recent findings, however, suggest an alternative possibility. The SMAD ubiquitination regulatory factors (Smurfs) Smurf1 and 2 are neural precursor cell expressed developmentally down-regulated protein 4 family homologous to the E6AP carboxyl terminus–type E3 ubiquitin ligases that target p-SMAD1, 5, and 8 (Smurf1) as well as p-SMAD2 and 3 and TGFβRI (Smurf2) for degradation (132–134). Smurf1 and 2 also promote p53 degradation; this does not require, however, the E3 activity of Smurfs but involves instead a Smurf-dependent stabilization of MDM2 via stimulation of heterodimerization MDM2 and murine double minute X while preventing MDM2 homodimerization (132). A serine-arginine-phenylalanine motif in Smurf1 and 2 is required for binding to MDM2, fostering MDM2–murine double minute X complex formation and enhancing p53 degradation. Importantly, USF2 is a transcriptional repressor of Smurf1 and 2 (135). As a negative regulator of Smurf2, which ubiquitinates SMAD2 and 3, USF2 likely promotes expression of p53 target profibrotic genes by 2 mechanisms involving the E Box-SMAD–dependent transcriptional activation of the PAI-1, CTGF, and collagen genes and repression of Smurf2, thereby inhibiting both p-SMAD2 and 3 and p53 degradation.

The phosphorylation-dependent formation of transcriptionally active complexes between p-p53 and SMAD2 and 3 are involved in the expression of a subset of TGF-β1 target genes (e.g., Mix2, p21, PAI-1, and matrix metalloproteinase 2) (68, 71, 73, 74, 76, 136) (Fig. 6B). This is consistent with the topographic requirement that p53 selectively activates TGF-β1 response genes with both SBEs and p53 binding motifs in their promoter regions. Such interactions between p53 and SMAD2 and 3 at their respective DNA binding sites or perhaps collectively with USF2 at the PE2 motif (71) recruits CBP-p300 to the PAI-1 promoter to increase histone [³H] acetylation and gene transcription (73, 137). Phosphorylation of the p53 amino-terminal serines 15 and 20 and threonine 18 increases association of p53 with members of the p300-CBP histone acetyltransferase coactivator family while stimulating p53 transactivation (77). CBP-p300 also acetylates SMAD2 and 3 in response to TGF-β1 (138), facilitating the creation of a multicomponent p53-SMAD-USF2 transcriptional complex necessary for optimal TGF-β1–dependent induction of the PAI-1 gene (73, 139–141).

CONCLUSIONS

GenBank (https://www.ncbi.nlm.nih.gov/genbank/) annotation and computational methodologies identified ≥1 p53 binding motif within 2000 bp upstream of the transcription start site in ∼1100 human genes; use of a transcript mapping approach added considerably to the number of candidate p53–regulated promoter sequences (100). Comparative analysis of microarray-classified TGF-β1–induced genes with the p53 target database further indicated that the majority of those responsive to TGF-β also possesses p53 binding sites. An increasing number of key signaling intermediates (e.g., the small GTPase Ras homolog family member B) are subject to joint regulation by TGF-β1 and p53 (142), and recent studies provide insights into the mechanistic basis of TGF-β1-p53 crosstalk (71,73.96,137, 143). For this cohort, which includes the fibrosis-causative PAI-1 and CTGF genes, pharmacologic inhibition of p53 function may have significant clinical implications in the management of fibrotic disease regardless of organ site.

Glossary

AKI

acute kidney injury

CBP

cAMP response element binding protein-binding protein

CKD

chronic kidney disease

CTGF

connective tissue growth factor

ECM

extracellular matrix

FOX

forkhead box

KEGG

Kyoto Encyclopedia of Genes and Genomes

MDM2

mouse double minute 2 homolog

PAI-1

plasminogen activator inhibitor-1

PE2

proximal E box 2

RB

retinoblastoma

RelA

the avian reticuloendotheliosis oncogene homolog A

SMAD

small worms and mothers against decapentaplegic homolog

Smurf

SMAD ubiquitination regulatory factor

TAZ

transcriptional coactivator with PDZ binding motif

USF

upstream stimulatory factor

UUO

unilateral ureteral obstruction

YAP

yes-associated protein

AUTHOR CONTRIBUTIONS

R. Samarakoon and P. J. Higgins designed the research; C. E. Higgins, J. Tang, S. P. Higgins, and C. C. Gifford performed the research; C. E. Higgins, S. P. Higgins, and K. K. Meldrum contributed new reagents; and C. E. Higgins, B. M. Mian, D. J. Conti, R. Samarakoon, and P. J. Higgins analyzed the data and wrote the manuscript.