The adducin saga: pleiotropic genomic targets for precision medicine in human hypertension—vascular, renal, and cognitive diseases

Abstract

Hypertension is a leading risk factor for stroke, heart disease, chronic kidney disease, vascular cognitive impairment, and Alzheimer’s disease. Previous genetic studies have nominated hundreds of genes linked to hypertension, and renal and cognitive diseases. Some have been advanced as candidate genes by showing that they can alter blood pressure or renal and cerebral vascular function in knockout animals; however, final validation of the causal variants and underlying mechanisms has remained elusive. This review chronicles 40 years of work, from the initial identification of adducin (ADD) as an ACTIN-binding protein suggested to increase blood pressure in Milan hypertensive rats, to the discovery of a mutation in ADD1 as a candidate gene for hypertension in rats that were subsequently linked to hypertension in man. More recently, a recessive K572Q mutation in ADD3 was identified in Fawn-Hooded Hypertensive (FHH) and Milan Normotensive (MNS) rats that develop renal disease, which is absent in resistant strains. ADD3 dimerizes with ADD1 to form functional ADD protein. The mutation in ADD3 disrupts a critical ACTIN-binding site necessary for its interactions with actin and spectrin to regulate the cytoskeleton. Studies using Add3 KO and transgenic strains, as well as a genetic complementation study in FHH and MNS rats, confirmed that the K572Q mutation in ADD3 plays a causal role in altering the myogenic response and autoregulation of renal and cerebral blood flow, resulting in increased susceptibility to hypertension-induced renal disease and cerebral vascular and cognitive dysfunction.

INTRODUCTION

The adducin saga began early in the 1970s with the development of the Milan Hypertension Strain (MHS) of rats by a research group led by Giuseppe Bianchi in Milan, Italy (1, 2). The MHS and its normotensive counterpart, the Milan Normotensive Strain (MNS), were created by brother-and-sister mating of Wistar rats selected in each generation from those with the highest and lowest blood pressures. The development of these strains paralleled worldwide efforts to develop genetic rat models of hypertension. These efforts led to the creation of the spontaneous hypertensive rat (SHR) (3), the New Zealand Genetically Hypertensive strain (4), Lyon Hypertensive rats (5), and the Dahl salt-sensitive (Dahl SS) rat (6). Later, other groups generated Sabra Hypertensive Prone (7) and Fawn-Hooded Hypertensive (FHH) strains (8). The underlying premise was that causal hypertensive genes were fixed in the strains selected for high blood pressure and that the same evolutionary variants might contribute to hypertension in man. Investigators have since been characterizing strain differences in neural, hormonal, and vascular pathways to identify potential mechanisms contributing to hypertension. Before the development of the genetically hypertensive strains, hypertension research primarily used renovascular experimental models (9, 10), including the two-kidney Goldblatt hypertension model (11), reduced renal mass rats (12, 13), renal wrap hypertension model (14, 15), and mineralocorticoid hypertensive animals (16–18).

The study of genetically hypertensive rat models led to the discovery of the roles of elevated sympathetic tone, the renin-angiotensin-aldosterone system, prostaglandins, bradykinin, nitric oxide, altered renal hemodynamics, vascular hypertrophy, and sodium retention in the development of hypertension (10, 19). However, in 1987, a landmark paper by John Rapp (20) emphasized that strain differences in phenotypes between normotensive and hypertensive animals do not mean they contribute to hypertension. He clearly established that the validation of causal genes requires the identification of sequence variants that alters the expression or activity of a protein and that the replacement of the defective allele normalizes arterial pressure. This conclusion revolutionized the field and led to the advent of genetic studies of hypertension throughout the 1990s, followed by genome-wide association studies (GWAS) after the sequencing of the human genome. Studies in transgenic and knockout (KO) mouse and rat models have confirmed that some of the nominated candidate genes have the potential to alter blood pressure. However, identification and validation of causal variants have proven to be very difficult with few successes (21–23). This review chronicles one success story regarding the discovery of a mutation in ADD related to hypertension in rats that was translated to man. However, the journey was long, and 40 yr have passed since the initial report that a sequence variant in Add1 was linked to hypertension in MHS rats (24, 25).

IDENTIFICATION OF ADD1 AS A CANDIDATE GENE FOR HYPERTENSION

Soon after the creation of the MHS and MNS rats in 1974 (1), Bianchi and his group characterized these rats for differences in the renin-angiotensin-aldosterone system and kidney function. This was done since renal vascular hypertension was the prevailing hypertension model at the time (10), and Guyton and associates (26) had just advanced the ground-breaking hypothesis for the dominant role of the kidney in hypertension. Bianchi’s laboratory reported that the mean arterial pressure was similar in MHS and MNS rats at 3–4 wk of age, but by 12 wk of age, blood pressure increased to 160–180 mmHg in MHS versus 120–125 mmHg in MNS (1). Renal transplant studies confirmed a role for the kidney in the development of hypertension (2). The development of hypertension was associated with greater sodium retention in young MHS rats. Plasma renin activity was lower in MHS than in MNS rats during the development of hypertension, and MHS rats exhibited reduced adrenal, heart, kidney, and body weights versus MNS rats (1). Glomerular filtration rate (GFR) and renal interstitial pressure were elevated during the development of hypertension in young MHS rats in association with elevated sodium reabsorption in the proximal tubule, thickening of ascending limb, and decreased influence of tubuloglomerular feedback on the afferent arteriole (Af-art) (27–31). In adult MHS, GFR was reduced due to overactivation of tubuloglomerular feedback (27, 32). It is interesting that MHS rats (33–35), like most of the commonly studied commercially available SHR strains derived from the SHR-B and C lines, are highly resistant to the development of hypertensive renal disease (36). However, it should be noted that the SHR-A3 line, often referred to as the SHR stroke-prone strain, is genetically distinct and is susceptible to hypertensive renal disease (23, 36).

A difference in the erythrocyte size and Na⁺/K⁺/Cl⁻ cotransporter (NKCC) expression was noted in MHS versus MNS rats. The difference in NKCC expression was resolved by bone marrow transplants (37) and after the removal of membrane cytoskeletal proteins in erythrocyte inside-out vesicles (38). This finding led to the idea that the differential expression of NKCC in the erythrocytes and possibly the kidney was caused by an abnormality in a membrane protein. This hypothesis was explored by immunizing one strain with membrane extracts from erythrocytes of the opposite strain. The MHS developed an antibody against a 105 kDa membrane protein from MNS (39), which was later identified as adducin (40).

Subsequent studies revealed that adducin is a heterodimeric cytoskeleton protein that is ubiquitously expressed. It consists of an α-subunit (ADD1) that binds to either β (ADD2) or γ (ADD3) subunits in a tissue-dependent fashion. ADD1 forms heterodimers with ADD3 in most tissues, including the kidney and vasculature (41). ADD1 forms heterodimers with ADD2 in erythrocytes and neurons (41, 42). Adducin is critical for actin-spectrin interactions and functions to cap the barbed ends of actin filaments to inhibit actin polymerization (Fig. 1A) (44, 45). Adducin also plays a key role in many biological processes such as cytoskeletal organization, membrane trafficking, signal transduction, cell-to-cell contact formation, and cell migration (41, 46–48). The activity of adducin and its association with the cell membrane is dependent on calcium and calmodulin and regulated by protein kinase A and C (49–52), tyrosine kinase (41, 53), and Rho-kinase (54).

Figure 1.

Adducin in the regulation of actin cytoskeleton. A: the adducin tetramer serves as an anchoring protein to bind the actin and spectrin to the membrane and regulates the polymerization of F-actin. Under normal conditions, wild-type adducin plays a vital role in regulating endocytosis of membrane large-conductance calcium-activated potassium channel (BK) channels in vascular smooth muscle cells and Na⁺/K⁺ATPnase in renal tubules. B: adducin-actin complex forms contractile units with myosin and is regulated by myosin light chain regulatory protein. C: adducin with inactivating mutation dissociates the protein complex from the membrane, activates the actin-nucleating actin-related protein 2/3 (Arp2/3) complex, enhances F-actin polymerization, and forms F-actin nets. Mutant adducin has been associated with hypertension, renal, and cognitive dysfunction. [Republished from Gao et al. (43).]

Tripodi et al. (40) cloned a gene for an adducin-like protein from a mouse cDNA library using the antibodies generated in the MHS and MNS cross-immunization experiments. Sequencing revealed that this clone was a major splice variant of Add2 and identified an adenine-to-guanine sequence variant that resulted in a glutamine-to-arginine substitution at amino acid position 529 of ADD2 (Q529R) (Fig. 2). Additional studies by this group identified another variant in ADD1 (F316Y) in MHS versus MNS rats (Fig. 2) (25, 55). The MHS was homozygous for the ADD1 (316YY) and ADD2 (529RR) variants. The MNS rat was homozygous for the ADD1 (316FF) allele but heterozygous for ADD2 (529QR) (25). However, more recent sequencing (http://rgd.mcw.edu/) indicates that MNS rats (MNS/Gib) are fully inbred and homozygous for the 529RR allele in ADD2, the same as seen in MHS (MHS/Gib). Subsequent studies also identified a variant in ADD3 (Q572R) in MNS versus MHS rats (56).

Figure 2.

Summary of sequence variants in adducin associated with hypertension, renal disease, vascular, and neurological dysfunction in genetic studies in humans and rat models.

In an F2 cross of MHS and MNS rats, the ADD1 Y allele cosegregated with a 6–12 mmHg increase in blood pressure (Fig. 2) (25). This finding was replicated in an independent genetic mapping study using an MNS × MHS F2 population (57). A subsequent sequencing study compared the genotype of 17 normotensive and hypertensive inbred rat strains of rats (55, 56). The results indicated that the ADD1 316FF allele was shared by most normotensive strains, whereas hypertensive SHR, SHR-Stroke-Prone, and MHS rats contain the YY allele. This led to the conclusion that the ADD1 316FF was the mutant hypertensive allele (25, 56). Add1 was confirmed as a candidate gene for hypertension by the creation of reciprocal congenic strains (58). Transfer of Add1 from MHS rats raised pressure in MNS by 10 mmHg, whereas introgression of Add1 from MNS into MHS rats lowered pressure by the same extent (58). The variants in ADD2 and ADD3 were not associated with hypertension but epistatically interacted with ADD1 to alter blood pressure (25, 46, 57).

The association of the ADD1 variant with hypertension in MHS led to the cloning and sequencing of ADD1 in humans. These studies identified a G460W variant in ADD1 (Fig. 2) that was associated with increased susceptibility to essential hypertension (32, 46), salt-sensitive hypertension (59, 60), altered sodium handling (61), and enhanced sodium reabsorption in the proximal tubule of the kidney (62) compared with patients homozygous for the ADD1 460GG allele. Initially, negative results were obtained for hypertension in a follow-up study in Caucasians (63), but meta-analysis studies in the Han Chinese population yielded significant associations with the ADD1 variant and hypertension (64). After additional genetic and environmental modifiers were included in the analysis, numerous studies confirmed an association between the ADD1 G460W variant and hypertension and cardiovascular complications in more than 160 studies in various populations (32, 46, 48, 65–77). An S586C variant (rs4963) in ADD1 (Fig. 2) has been reported to be linked to hypertension in an Asian population (74). However, its mechanism of action has not been explored or identified.

Several studies have investigated the cellular mechanisms by which genetic variants in ADD1 and ADD2 contribute to hypertension. Tripodi et al. (24) extracted adducin protein from red blood cells from MHS and MNS rats and studied its effects on actin polymerization in vitro. They also cloned Add1 genes from MHS, MNS, and transfected a renal epithelial cell line. They observed enhanced actin polymerization and increased levels of filamentous actin in cells transfected with Add1 of MHS, suggesting the actin capping protein function was diminished (24). Similar results were found in cells transfected with ADD1 containing the G460W variant (46, 48), suggesting that both Y316F in MHS and G460W in man in ADD1 are loss-of-function mutations that diminish actin capping and increase actin polymerization. Changes in the actin cytoskeleton in association with increased Na⁺/K⁺ ATPase activity were ascribed to the ADD1 Y316 allele from MHS in this study. They speculated that the kidney was involved in the differences in blood pressure in MHS and MNS rats and in hypertensive patients with ADD1 G460W mutation due to cytoskeletal disruption that elevated Na⁺/K⁺ ATPase activity and enhanced renal sodium retention (46, 48).

The ADD1 variants in man and MHS rats are localized at the neck and tail domains of the adducin protein. The tail of ADD contains a highly conserved 22-residue myristoylated alanine-rich C kinase substrate (MARCKS)-related domain, which is required for the actin capping function along with the neck domain (41). The human ADD1 G460W variant has been reported to increase the expression of Na⁺/K⁺ ATPase in the basolateral membrane of renal tubular cells by impairing endocytosis (46, 48, 78–81). The increase in Na⁺/K⁺ ATPase activity in the kidney promotes sodium retention and volume expansion (80), which was suggested to elevate the production and levels of endogenous ouabain (46, 48, 80), which in turn, enhances Na⁺/Ca²⁺ exchange in vascular smooth muscle cells (VSMCs) and vascular reactivity and tone (82, 83). Interestingly, the plasma levels of endogenous ouabain measured in MHS rats are elevated (32). Moreover, Linde et al. (84) reported that the myogenic response and vascular contractile response of small arteries were greater in MHS than in MNS rats. This was associated with upregulation of the expression of Na⁺/Ca²⁺ exchanger and the Transient Receptor Potential Cation Channel Subfamily C Member 6 expression and enhanced calcium responses in VSMCs isolated from MHS than from MNS rats (84, 85). These results support the view that elevated renal Na⁺/K⁺ ATPase activity and sodium retention leading to an elevation in circulating levels of endogenous ouabain could be the mechanism by which mutations in ADD1 increase vascular reactivity and the development of hypertension (86). This interpretation is supported by the findings that rostafuroxin, which is a digitoxigenin derivative specifically disrupts binding to the Src-SH2 domain of mutant forms of ADD1 and ouabain-activated Na⁺/K⁺-ATPase to decrease Src-dependent phosphorylation and increased activity of Na⁺/K⁺-ATPase in the membrane. It lowers blood pressure and affords renal and cardiac protection in rat models (MHS and congenic MNS strains) expressing the mutant F316Y form of ADD1 but not in controls (87, 88). Similarly, rosafuroxin lowers pressure in hypertensive patients with mutations in ADD1 and/or ADD3 or variants in genes encoding for endogenous ouabain (89, 90). However, the concept of endogenous ouabain in the pathogenesis of hypertension requires further investigation since the exact chemical structure of endogenous ouabain has not been fully defined, and the levels measured in animal and human studies were only determined by immunoreactivity to an ouabain-specific antibody (32).

The ADD1 G460W polymorphism has also been implicated in stroke (91, 92). Van Rijin et al. (91) reported a significant increase in the incidence of ischemic and hemorrhagic stroke in a population of 6,471 patients with the ADD1 G460W variant in the Rotterdam study. They noted that the carriers of the ADD1 variant commonly presented with silent brain infarcts and white matter lesions in subcortical areas. These findings were later replicated in a much larger study of 15,236 Dutch women (93). However, the incidence of stroke was higher in women with hypertension, so the question remains as to whether the ADD1 G460W increases the susceptibility to stroke, secondary to hypertension or by other independent effects.

VARIANTS IN ADD2 ARE NOT ASSOCIATED WITH HYPERTENSION BUT ARE LINKED TO COGNITIVE DYSFUNCTION

ADD2 is predominantly expressed in erythrocytes and the brain (41). Studies in Add2 KO mouse indicated that the loss of ADD2 results in the production of osmotically fragile, microcytic red blood cells (RBCs) associated with a 70% drop in ADD1 (94) in erythrocytes that were partially compensated for by a fivefold increase in ADD3 (93). KO of Add2 was found to increase MAP by 18 mmHg in C57BL/6 mice (95) but it did not alter blood pressure in a later study in 129/Sv mouse genetic background (96). KO of Add3 had no effect on RBC function, platelet structure, or systemic blood pressure in mice (97). Nor, did it exacerbate the spherocytosis of erythrocytes in Add2 KO mice (97). However, KO of Add1 caused RBC spherocytosis and lethal hydrocephalus in 50% of the animals, which was surprisingly associated with complete loss of both ADD2 and ADD3 in erythrocytes (98). Other phenotypes such as blood pressure, renal, and cognitive function have not been studied in Add1 KO mice.

ADD2 plays a more significant role in the synaptic formation in neurons, the loss of which affects long-term cognitive and motor function (99). The assembly of new synapses was impaired in Add2 KO mice (53), and they exhibited diminished long-term hippocampal memory following environmental enrichment (100). In humans, a silent exonic mutation has been detected in ADD2 (C1797T, rs4984) that is linked to changes in learning and memory (Fig. 2). The ADD2 1797CC genotype was found to have significant effects on almost every measurement of cognitive function in 342 Caucasian schizophrenia patients (101). Individuals with ADD2 1797CC homozygotes performed better in verbal fluency, memory, cognitive flexibility, and attention tests than those with the 1797TT genotype, most likely due to a neurodevelopmental defect. Furthermore, patients with both ADD1 (G460W, rs4961) and ADD3 (G386A, rs3731566) variants (Fig. 2) exhibited impairment of symbol coding and verbal memory (101). Bosia et al. (101) hypothesized that the cognitive impairment in this group might be secondary to a vascular insult. More studies are needed to examine these fascinating interactions.

IDENTIFICATION OF ADD3 AS A CANDIDATE GENE FOR RENAL DISEASE IN RAT MODELS

The FHH rat is a genetic model of renal disease, which develops mild hypertension, focal glomerulosclerosis, and proteinuria as they age (8, 102–106). The onset of chronic renal disease (CKD) in FHH rats is associated with altered renal hemodynamics, including reduced preglomerular vascular resistance, impaired myogenic response of preglomerular arterioles, elevated glomerular capillary pressure (P_GC), and hyperfiltration relative to disease-resistant strains (107–111).

Using crosses of FHH rats and August-Copenhagen Irish (ACI) or Brown-Norway (BN) rats, linkage analysis studies identified five quantitative trait loci (QTLs), termed (renal failure) Rf-1–5, that cosegregate with proteinuria and focal glomerulosclerosis following L-N^G-Nitro arginine methyl ester (L-NAME) induced hypertension, uninephrectomy, or aging (112–114). Rf-1 and Rf-2 are located on rat chromosome 1, Rf-3 on chromosome 3, Rf-4 on chromosome 14, and Rf-5 on chromosome 17. The Rf-1 region remains of particular interest since it is homologous to a QTL on human chromosome 10 associated with diabetic nephropathy, especially in African Americans (115–117). Subsequent mapping studies using congenic strains identified a mutation of Rab38 in the Rf-2 region on chromosome 1, which downregulates albumin reuptake in the proximal tubule and contributes to the development of proteinuria in FHH rats (118). Shroom3 in the Rf-4 region was also linked to proteinuria in FHH rats and CKD in humans by enhancing podocyte effacement and breakdown of the glomerular filtration barrier (119). Lazar et al. (120) found that introgression of a 1.5-Mbp region in the FHH Rf-1 locus containing Sorcs1 into the ACI.1^FHH congenic strain increased proteinuria, whereas KO of Sorcs1 in the proximal tubular cells impaired protein trafficking and increased proteinuria in an FHH.1^BN strain. However, the role of Sorcs1 as the causal gene for proteinuria has not been confirmed since no variants in the coding sequence or the renal expression of Sorcs1 were found among FHH, ACI, or BN rats (120).

Substitution mapping studies using overlapping FHH.1^BN congenic strains revealed that impaired myogenic response of preglomerular arterioles and autoregulation of renal blood flow (RBF), elevated P_GC, and hypertension-induced proteinuria in FHH rats were associated with the Rf-1 region. Lopez et al. (121) reported that introgression of a 12.9 Mbp QTL in the Rf-1 region (D1Mgh13-D1Rat89) along with a 99.4 Mbp QTL in the Rf-2 locus (D1Rat183-D1Rat76) in a double-congenic FHH.1^BN strain restored autoregulation of RBF and reduced proteinuria. In contrast, introgression of the Rf-2 region (D1Rat183-D1Rat76) alone did not rescue the RBF autoregulation. Williams et al. (122) later narrowed the region in the Rf-1 down to 4.7 Mbp (D1Rat376-D1Rat225). In 2013, Burke et al. (123) created a series of overlapping FHH.1^BN congenic strains that narrowed the locus that rescued autoregulation of RBF down to 2.4 Mbp in the Rf-1 region. This region contains 15 genes, and Add3, Dual Specificity Phosphatase 5 (Dusp5), and Suppressor of Clear Homolog 2 (Shoc2) were identified to have coding sequence variants in FHH versus BN rats. Shoc2 is involved in the extracellular signal-regulated kinase (ERK) signaling pathway (124); however, it was not considered as a candidate gene because of the lack of evidence that it is related to the control of vascular function. Dusp5 dephosphorylates and inactivates ERK and was investigated as a candidate gene involved in vascular function (125). KO of Dusp5 in the FHH.1^BN congenic strain enhanced, rather than inhibited, the myogenic response of renal and cerebral arteries and arterioles (125–127). Thus, Add3 emerged by default as the most likely candidate gene.

Recently, Fan et al. (128) identified a K572Q mutation in ADD3 in FHH rats. Interestingly, a polymorphism (Q572K) at amino acid (AA) 572 in ADD3 was originally reported by Tripodi et al. (56) in MHS, that are susceptible to hypertension, compared with MNS rats, that are susceptible to hypertensive and diabetic nephropathy. AA 572 of ADD3 is located on exon 13 at a critical ADD3-ACTIN-binding site. Molecular dynamic simulations indicated that this mutation in FHH rats destabilizes the secondary structure of ADD3 in the region from AA 547 to 596 (127). These changes were predicted to alter the binding affinity of ADD3 to actin and actin-cytoskeleton dynamics (43, 128, 129), which in turn regulate signal transduction and membrane protein trafficking (41, 130). Our findings supported this prediction that membrane expression of ADD3 was reduced in primary VSMCs and podocytes isolated from FHH rats (128). ADD3 was redistributed from the cell membrane to a perinuclear location in isolated and cultured VSMCs from FHH rats in association with the loss of F-actin stress filaments and the formation of excessively branched F-actin networks (Fig. 1) (43).

Confirmation that a sequence variant is causal to the alteration in renal hemodynamics and CKD requires demonstration that the expression of the WT protein restores function. To validate whether the K572Q mutation of ADD3 was responsible for the impaired myogenic response and autoregulation of RBF in FHH rats, congenic and transgenic strains were created in which WT Add3 from BN rats was introduced into FHH genetic background. Add3 was globally knocked out in the FHH.1^BN congenic strain and normal Sprague Dawley (SD) rat. The myogenic response of the Af-art and RBF autoregulation was impaired in FHH and MNS rats carrying the mutant 572QQ ADD3 allele and in two strains of Add3 KO rats (128). These responses were intact in SD and MHS rats and were rescued in FHH. 1^BN congenic and FHH.Add3^K572 transgenic strains that express WT ADD3 (122, 127, 131, 132). Furthermore, the role of the K572Q mutation in altering the myogenic response was confirmed in a genetic complementation study by crossing FHH and MNS rats. We found that the Af-art constricted normally in an F1 cross of FHH and FHH. 1^BN rats with one copy of the 572 K allele when perfusion pressure was increased, but it did not constrict in an F1 cross of FHH and MNS rats which share the mutant 572QQ allele (128). Similarly, autoregulation of RBF was impaired in FHH × MNS F1 rats but was intact in a cross of FHH and FHH. 1^BN rats (128). These results provide strong genetic evidence that the mutant ADD3 572Q allele underlies renal vascular dysfunction in both FHH and MNS rats.

The mechanism by which loss of ADD function impairs the myogenic response was also explored. Knockdown of ADD3 using dicer-substrate-short-interfering RNA (DsiRNA) markedly impaired not only the myogenic response of renal Af-art but also middle cerebral arteries (MCA) in the brain of normal SD rats (130). Similarly, strains that exhibit reduced ADD3 function, including FHH and two Add3 KO rats, displayed not only poor autoregulation of RBF (123, 128, 131, 132) but also cerebral blood flow (CBF) (133–136). The loss of myogenic response was associated with the increased activity of the large-conductance calcium-activated potassium channel (BK). The BK channel activity was fourfold greater in the VSMCs isolated from renal and cerebral arteries after the knockdown of Add3 (130). Administration of iberiotoxin, a BK channel inhibitor, restored the myogenic response and vascular reactivity following the knockdown of ADD3 (130). Similarly, impaired myogenic responses and elevated BK channel activities were found in VSMCs isolated from renal and cerebral arteries of FHH rats with the K572Q mutation in ADD3 relative to FHH. 1^BN rats that express the WT ADD3 (127, 136, 137).

EVIDENCE THAT A MUTATION IN ADD3 CONTRIBUTES TO THE DEVELOPMENT OF RENAL DISEASE

The recently identified K572Q mutation of ADD3 in FHH rats disrupts the actin cytoskeleton, dislocates the ADD3 protein from the membrane, and impairs endocytosis of other membrane proteins, thereby increasing membrane distribution of BKα in VSMCs (43, 128, 129, 135, 137). The Q572 allele is also found in MNS but not MHS rats, and both FHH and MNS rats are genetic models of renal disease (33, 35, 56, 128).

Disruption of the actin cytoskeleton in VSMCs by ADD3 dysfunction could explain why both FHH and MNS rats that exhibit impaired myogenic response of the Af-art and autoregulation of RBF are susceptible to the development of hypertension-induced CKD (33, 128). The impaired myogenic response allows for greater transmission of systemic pressure to the glomeruli, raising P_GC and inducing hyperfiltration. Hyperfiltration and elevated distention of glomerular basement membranes promote podocyte injury, effacement, and loss, all of which contribute to increasing the filtered load of protein and proteinuria. With time, increases in the reuptake of filtered protein promote tubulointerstitial injury, especially following the onset of hypertension or diabetes. The hypothesis that elevated P_GC and glomerular hyperfiltration initiate podocyte injury and proteinuria has been confirmed in a variety of hypertensive and diabetic nephropathy models (131, 132, 138–140). Results from our recent studies are consistent with these findings. We found that FHH and Add3 KO rats do not develop hypertension before the onset of proteinuria (128, 131, 132). Expression of WT ADD3 in an FHH.1^BN congenic or an FHH.Add3 transgenic strain restored the myogenic response of Af-art and autoregulation of RBF (128). Both DOCA-salt and N^ω-Nitro-l-arginine methyl ester hydrochloride (L-NAME) elevated blood pressure to the same extent in FHH compared with FHH.Add3 transgenic and FHH.1^BN compared with Add3 KO rats on the FHH.1^BN genetic background (128, 131, 132). We found that P_GC and glomerular permeability to albumin were elevated after the induction of DOCA-salt hypertension in FHH rats (128). Transmission of pressure to the glomerulus was markedly elevated in hypertensive Add3 KO rats, leading to glomerulosclerosis and renal fibrosis and a progressive fall in GFR following DOCA-salt hypertension (131, 132). Interestingly, we observed that GFR and proteinuria were even higher in L-NAME hypertensive Add3 KO rats than in the control FHH.1^BN strain expressing the WT ADD3 (131, 132). We speculated that was due to the impaired myogenic response in Add3 KO rats that maintain an elevated P_GC and GFR relative to the control rats with intact RBF autoregulation. We found renal injuries were markedly increased in hypertensive FHH and Add3 KO rats with higher proteinuria, glomerular injury, and reduced nephrin expression. Renal interstitial fibrosis and protein cast formation were markedly reduced in hypertensive FHH.Add3 transgenic rats. Moreover, FHH.Add3 rats exhibited attenuated CKD and proteinuria with aging as compared with FHH rats (128).

In addition to the contribution of impaired renal hemodynamics to renal injury in FHH and ADD3 KO rats, disruption of the actin cytoskeleton by ADD3 dysfunction may also play a direct detrimental role in podocyte function (141–143). In this regard, we recently reported that mutant ADD3 in FHH rats enhanced the formation of branched F-actin nets and reduced unbranched F-actin in podocytes isolated from FHH rats compared with cells isolated from FHH Add3 transgenic rats (43). The predominant branched F-actin imbalances the intracellular pushing and pulling forces that maintain the podocyte foot process and predisposes the podocytes to detachment and effacement following hypertension or diabetes (43). Thus, the podocyte injury in FHH rats is a downstream pathologic effect resulting from a combination of impaired renal hemodynamics and direct podocyte damage caused by cytoskeletal dysfunction. Other studies (96) also indicate that the mutation in ADD2 in MHS rats and KO of Add2 in mice alters podocyte function and promotes proteinuria and renal disease. In contrast, mutant ADD1 from MHS rats has a renal protective effect in MNS congenic strains.

Overall, these results support the view that the mutant K572Q in ADD3 contributes to the susceptibility to kidney disease, at least in FHH and MNS rats. Our results are consistent with previous findings, indicating that MHS contains the K572 but MNS contains the Q572 allele of ADD3 (55). However, the definition of “mutant” and “WT” alleles is different between our studies and previous reports. Bianchi and Tripodi studied hypertension; thus, their “mutant” allele was defined in MHS (hypertensive) rats in comparison with MNS (normotensive) rats. More studies have been reported since then, and now strain sequence variants of Add3 in 90 strains of rats have been screened (https://rgd.mcw.edu/rgdweb/report/gene/main.html?id=2043). Thirteen normotensive strains (such as BN, SD, and ACI) and five hypertensive strains (such as SHR and MHS) contain the K572 allele of ADD3, and the Q572 allele exhibits in FHH, MNS, Wistar Kyoto (WKY), Buffalo, and Lewis rats (56, 128). Although it is clear that FHH, MNS, and Buffalo rats are genetic models of renal disease (33, 35, 56, 128, 144–146), the influence of the Q572 allele of ADD3 in the susceptibility to hypertensive and diabetic nephropathy in the latter group of rats has not been studied. However, there has been some evidence indicating that the myogenic response of the Af-art and autoregulation of RBF is reduced relative to their genetic control strains (146–151). Based on the available evidence that most inbred strains, including the Brown Norway (BN/NHsdMcwi) strain that is the reference sequence for Rattus norvegicus in the GenBank at the National Center for Biotechnology Information (NCBI), express the K572 allele of ADD3, it appears that Q572 is the variant associated with impaired RBF autoregulation and podocyte structure that promotes the development of CKD in FHH rats.

ADDUCIN VARIANTS AND CEREBROVASCULAR DYSFUNCTION, STROKE, AND DEMENTIA

Whether ADD variants play a causal role in the development of hypertension, cardiovascular and renal disease in man remain uncertain. A summary of the available genetic association information is presented in Fig. 2. A G460W mutation in ADD1 was associated with the incidence of hypertension (25, 46, 48, 77, 152, 153), stroke (91–93), and cardiovascular diseases (92, 154, 155) in various human populations, along with the corresponding F316Y mutation in MHS rats (25, 46, 57). There is evidence that variants in ADD2 and ADD3 exert an epistatic interaction to enhance the influence of the ADD1 genotype on blood pressure in man and MHS rats (25, 46, 57). Patients in the European Project on Genes in Hypertension (GPOGH) that are homozygous for the ADD3 (386GG) and also express ADD1 460 GW or 460WW genotypes had a significantly elevated pulse pressure (156).

There are a few reports suggesting that variants in ADD3 may be linked to cardiovascular and renal disease in humans (Fig. 2). A rare variant (G367D) in ADD3 was found in a family line that was associated with cardiomyopathy and steroid-resistant nephrotic syndrome (157). Finally, a region on human chromosome 10, homologous to Rf-1 containing the mutant Add3 region in rats, has been associated with diabetic and nondiabetic CKD in African Americans (115, 117) and a large population from Utah (116). However, the QTLs in these studies were broad (40 cM), and the most recent GWAS meta-analysis for CKD reported that ADD3 is 5–6 Mbp from the nearest highly polymorphic marker associated with a reduction in eGFR (158). Unfortunately, these analyses did not specifically address linkage to proteinuria, and the density of polymorphic markers is not dense enough to fine map the region. Analysis of the Genome Aggregation Database of CKD identified 492 nonsynonymous variants in human ADD3, 227 of which are predicted to damage ADD3 protein function. Many of these variants are in the same ACTIN-binding region as the K572Q variant in FHH and MNS rats, and the corresponding R571H variant in ADD3 is the most common mutant found in man (https://macarthurlab.org/2019/10/16/gnomad-v3-0/). However, these are all rare variants with less than 1% frequency; thus, they have all been excluded in previous GWAS studies of renal disease and hypertension (128). Most of these variants are in the tail region of the ADD3 protein, indicating they can alter ADD3 hetero- and tetramer assembly with ADD1, interfere with ADD interactions with actin and spectrin (41, 44), and thus may be involved in the pathogenesis of CKD. In addition, multiple splice variants and three isoforms of ADD3 have been identified, one of which deletes amino acid 576–607 across the critical ACTIN-binding region (159). The influence of the multiple isoforms on the susceptibility to renal vascular dysfunction and disease has not been evaluated. Thus, we believe that additional genetic association studies that are sequence-based and pool results from diabetic and hypertensive subjects with damaging genotypes are warranted to discern if rare variants in ADD3 are associated with CKD in susceptible populations. Alternatively, one could fine map the ADD3 region in the previous GWAS studies for CKD with a higher density of polymorphic markers. Nevertheless, present results indicate that variants in ADD3 that alter the actin cytoskeleton and impair renal hemodynamics can increase the susceptibility to diabetic and nondiabetic CKD in patients with one of these rare mutations.

More recently, we have reported that the K572Q mutation in ADD3 in FHH rats is associated with cerebral vascular dysfunction in aging and after the development of hypertension (134–136, 160). We found that this mutation reduced ADD3 expression in the membrane fraction of cerebral VSMCs isolated from FHH rats compared with the FHH.1^BN congenic strain. FHH rats exhibited altered cerebral hemodynamics, including impaired myogenic response in the MCA, increased transmission of pressure to the terminal pial artery, larger capillary diameters, and reduced capillary density (rarefaction) compared with the FHH.1^BN congenic strain (129, 133–136, 160, 161). Autoregulation of CBF was impaired in the FHH rats and rescued in FHH.Add3 transgenic strain. The impaired cerebral hemodynamics in the FHH rats contributed to greater blood–brain barrier (BBB) leakage, neurodegeneration and vascular remodeling in the neocortex and hippocampus, and spatial learning and memory dysfunction compared with the FHH.1^BN congenic and FHH.Add3 transgenic strains following the development of hypertension (134–136, 160). We have reported that impaired myogenic response and autoregulation of CBF are also associated with increased BBB leakage, neurodegeneration, and loss of cognitive function following the development of hypertension in Dahl SS rats and type-II diabetic models (162–167). The impaired CBF autoregulation in Dahl SS rats is associated with a genetic deficiency in the expression of cytochrome P450 4 A (CYP4A) and the formation of 20-hydroxyeicosatetraenoic acid (20-HETE) (168). Inactivating mutations in CYP4A have been associated with increased incidence of stroke (169–171) and loss of cognitive function in a recent preliminary report (170, 172–174). Given that loss of function in ADD3 and CYP4A causes similar changes in cerebral hemodynamics (163, 168, 172, 174), it will be important to test whether mutations in ADD3 are associated with the incidence of ischemic or hemorrhagic stroke, vascular dementia, or Alzheimer's Disease (AD) and Alzheimer's Disease-Related Dementias (ADRD) in man.

ADD2 is highly expressed in the dendrites, and KO of Add2 induced deficits in behavioral, motor coordination, and learning (53, 99, 100, 175) in mice. The behavioral effects in ADD2 KO mice are also associated with changes in the expression and phosphorylation levels of ADD1 and ADD3 (53). A cleavage fragment of ADD3 has been associated with synaptic dysfunction, cognitive dysfunction, and AD-like pathology in transgenic mice (Fig. 2) (176). Overexpression of a miR-135a-5p has been reported to cause memory and synaptic impairments via the Rock2/ADD1 signaling pathway in an AD mouse model (177). Bosia et al. (101) reported that the C1797T (rs4984) variant in ADD2 had significant detrimental effects on almost every cognitive function domain in a GWAS study of schizophrenia. Significant interactions between ADD1 G460W and ADD3 G386A mutations were observed in symbol coding and verbal memory in these patients (101). Charney et al. (178) also identified that a region of chromosome 10 containing ADD3 is associated with an enhanced risk for the bipolar type of schizoaffective disorder (Fig. 2). This effect of adducin variants in schizophrenia may result both from a direct mechanism affecting synaptic building and plasticity or indirectly as a consequence of vascular insults.

In other studies, Gonçalves et al. (157) identified three families with ADD3 G386A mutation displayed intellectual disability, microcephaly, cataracts, and skeletal defects (Fig. 2). Finally, a mutation G367D (rs564185858) in ADD3 has been found in individuals with inherited cerebral palsy in a consanguineous Jordanian family (Fig. 2) (179). Fibroblasts isolated from these patients demonstrated that the mutant ADD3 had reduced ability to form heterodimers with ADD1, impaired the actin-capping function, suggesting that alterations in ADD-related cytoskeleton dynamics may be responsible for the neurological abnormality. Finally, genetic polymorphisms near and within the Add3 gene have also been linked to biliary atresia (180), but the relationship to the cardiovascular and neurological abnormalities is unclear.

SUMMARY AND CONCLUSIONS

The Adducin saga parallels the evolution of hypertension research over the last 40 yr, moving from the development of genetic rat models of hypertension through GWAS in man to studies using transgenic and KO in rodent models. Sequence variants in ADD1, 2, and 3 that alter cytoskeleton dynamics have been identified that are linked to hypertension, vascular dysfunction, renal, and neurodegenerative diseases. The underlying mechanisms of a sequence variant in ADD3 in FHH rats that contributes to the hypertension-induced renal disease include impaired renal hemodynamics and podocyte structure and function. There is some evidence that similar abnormality in CBF autoregulation may also be involved in developing the vascular contribution of AD/ADRD (Fig. 3).

Figure 3.

Mechanisms by which loss of ADD3 function promotes chronic kidney disease and cognitive impairment by altering vascular function. Loss of function mutations in ADD3 in FHH, MNS, and man that impair actin capping function disrupt the cytoskeleton in podocytes and vascular smooth muscle cells (VSMC). This leads to impaired myogenic response (MR) and autoregulation of renal and cerebral blood flow (RBF/CBF) that increases the transmission of pressure to the glomerulus and cerebral microcirculation. The rise in glomerular capillary pressure especially with aging and hypertension causes hyperfiltration and together with changes in the podocyte cytoskeleton causes podocyte effacement and elevated filtration of protein. Increased delivery of protein to the proximal tubule causes epithelial-to-mesangial transition, renal fibrosis, and chronic kidney disease. In the brain, elevated capillary pressure causes endothelial damage, blood–brain barrier (BBB) leakage, and impaired neurovascular coupling (NVC) leading to capillary stalling, rarefaction, and focal hypoperfusion that promote neurodegeneration, white matter damage, and cognitive impairments. FHH, Fawn-Hooded Hypertensive; MNS, Milan Normotensive.

GRANTS

This study was supported by National Institutes of Health Grants AG050049, AG057842, AG066245, P20GM104357, DK104184, and HL138685; American Heart Association Grants 16GRNT31200036 and 20PRE35210043; and the Medical Student Research Program (MSRP) from the University of Mississippi Medical Center.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.F., and F.F. prepared figures; E.G-F., L.F., and R.J.R. drafted manuscript; E.G-F., L.F., K.N.T., F.F., and R.J.R. edited and revised manuscript; E.G-F., L.F., S.W., Y.L., W.G., K.N.T., F.F., and R.J.R. approved final version of manuscript.