Phosphate in cardiovascular disease: From new insights into molecular mechanisms to clinical implications

Abstract

Hyperphosphatemia is a common feature in patients with impaired kidney function and is associated with increased risk of cardiovascular disease. This phenomenon extends to the general population, whereby elevations of serum phosphate within the normal range increase risk; however, the mechanism by which this occurs is multifaceted and many aspects are poorly understood. Less than 1% of total body phosphate is found in the circulation and extracellular space, and its regulation involves multiple organ-crosstalk and hormones to coordinate absorption from the small intestine and excretion by the kidneys. For phosphate to be regulated, it must be sensed. While mostly enigmatic, various phosphate sensors have been elucidated in recent years. Phosphate in the circulation can be buffered, either through regulated exchange between extracellular and cellular spaces, or through chelation by circulating proteins (i.e., fetuin-A) to form calciprotein particles, which in themselves serve a function for bulk mineral transport and signalling. Either through direct signalling, or through mediators like hormones, calciprotein particles or calcifying extracellular vesicles, phosphate can induce various cardiovascular disease pathologies: most notably, ectopic cardiovascular calcification, but also left ventricular hypertrophy, as well as bone and kidney diseases, which then propagate phosphate dysregulation further. Therapies targeting phosphate have mostly focused on intestinal binding, of which appreciation and understanding of paracellular transport has greatly advanced the field. However, pharmacotherapies that target cardiovascular consequences of phosphate directly, such as vascular calcification, are still an area of great unmet medical need.

Graphical Abstract

Introduction: Development and complexity of phosphate physiology

Rock existed long before any lifeforms on Earth. Elemental phosphorus is unstable and highly combustible, and phosphorus-containing gases are infinitesimally minuscule; therefore, one is dealing primarily with phosphate, a phosphorus atom bound to four oxygen (PO₄). Inorganic phosphate can be released from minerals and non-enzymatically transformed into simple organophosphate molecules in the prebiotic planet, a reaction vessel poetically mused as “Darwin’s warm little pond”.¹ In a paper by Frank Westheimer titled “Why nature chose phosphates”,² fundamental properties of phosphate were highlighted to justify its central role in life. Organic phosphate, phosphate bound to carbon moieties, is ubiquitously present as nucleic acid derivatives, phosphoamino acids and phosphoproteins, phosphoglycans, and phospholipids, which contribute to heredity, catalysis, metabolism, and membrane integrity. Inorganic phosphate is biologically active, and forms calcified tissues capable of turnover, and thus maintained carefully at an intracellular, organ, and system level. Here we focus on regulation and pathophysiological consequences of inorganic phosphate dysregulation in cardiovascular disease (CVD).

Given the diverse biologic functions of phosphate, its homeostasis is critical and staunchly preserved. However, there are scenarios in human pathobiology that breach the homeostatic safety zone. The most prevalent setting of phosphate dysregulation conferring CVD consequences is in patients with chronic kidney disease (CKD), specifically a sub-population exhibiting mineral bone disorder (CKD-MBD) with overt hyperphosphatemia and related sequelae, in part, due to the inability of kidneys to remove excess phosphate³. However, recent evidence is expanding this understanding of phosphate toxicity to the setting of normophosphatemia and in those with heathy kidney function. In this review, we explore recent advances and burning questions in phosphate homeostasis and sensing as it relates to CVD.

Circulating and microenvironmental phosphate comes in different forms

The form of extracellular phosphate present in the circulation and microenvironment is intrinsic to how it impacts pathophysiology. When one refers to circulating or extracellular phosphate, one is typically referring to the pool of inorganic phosphate ions comprising less than 1% of total body phosphate⁴. Much of this phosphate is absorbed from food sources in the small intestine, however it is also released from bone, the primary storage site of phosphate (and calcium) or generated through hydrolysis of organic phosphate moieties⁵. Serum phosphate levels are consistently associated with increased risk of all-cause mortality and CVD in not only CKD and dialysis patients^6–10, but also in the general population within the normal range of serum phosphate^11–13.

Circulating phosphate, while under the levels of spontaneous precipitation with calcium, is above the solubility of bulk bone mineral (hydroxyapatite), which will readily sustain crystal growth if nucleated¹⁴. Thus, mechanisms exist to buffer the amount of free circulating phosphate. Mediation of free phosphate from the circulation can include both altering protein bound and unbound phosphate in the circulation, forming calciprotein particles (CPPs), as well as regulated movement of phosphate to different compartments of the body. These buffering systems were initially considered merely as a mechanism to inhibit ectopic calcification; however, recent evidence suggests these mechanisms have signaling implications¹⁵, and the extraction of nascent phosphate may also protect against or cause adverse cardiovascular effects.

CPPs are colloidal mineral-protein complexes posited to buffer fluxes and chaperone calcium and phosphate, in a form not amenable to ectopic calcification, to sites where it can be utilized or disposed of. The most primitive form of CPP consists of single fetuin-A molecules laden with amorphous calcium-phosphate, which are called calciprotein monomers (CPMs) and are present in the circulation constitutively¹⁶. In response to various stimuli, CPMs spontaneously undergo self-aggregation and amorphous-to-crystalline phase transition of calcium-phosphate over time to mature into CPPs with larger particle size and higher density, often termed CPP-I and CPP-II¹⁷. CPMs, as well as these more mature CPPs, elevate in the blood after dietary phosphate intake and constitutively in the setting of CKD^15,18,19. Standardization of both nomenclature and measurement is an active area of research. Transition of CPMs to amorphous CPPs (CPP-I) and crystalline CPPs (CPP-II) are dependent on a number of factors, notably increasing organic phosphate, as well as reductions in pyrophosphate (PPi) and CPP-bound proteins, fetuin-A and albumin, all of which are characteristic of CKD^16,20. CPMs are different from the mature CPPs not only in colloidal properties but also in biological activity. CPMs are cleared from the blood mainly through glomerular filtration²¹. In contrast to CPMs, mature CPPs are not filtrated through glomeruli but endocytosed by reticuloendothelial system cells²². In vitro studies suggest that CPPs exert pathogenic activity inflammatory responses partly through binding to toll-like receptor 4²². In clinical studies, blood CPP levels correlate with serum phosphate levels²³ and parameters for vascular calcification, aortic stiffness, chronic inflammation, and cardiovascular events in patients with CKD^24,25. The elucidated molecular mechanisms and clinical associations with disease motivate research into potential direct causative roles of CPPs in initiating or propagating cardiovascular complications in this population.

While in the exchangeable pool of the biological system, phosphate must move between bone and soft tissue extracellular and intracellular pools. Following intravenous administration of phosphate, healthy rats only excreted 50% after four hours, despite normalization of serum phosphate²⁶. Similarly, in healthy humans full clearance of a large phosphate load required approximately 120 hours despite only mild changes in serum phosphate that normalized by 48 hours²⁷. Compartmental modelling predicts kinetics of this prolonged but not permanent extra circulatory phosphate capacitance and identifies gaps in knowledge through discrepancies between literature-informed predictions and mammalian kinetic experiments^28,29. Using calcium studies as a basis, bone is assumed to have both a rapidly and slowly exchangeable pool, and an intracellular pool. Current modelling suggests that 10 minutes following IV administration, up to 90% of phosphate has accrued in the bone’s rapidly exchangeable pool³⁰. Recent models now include CPP formation and clearance and more recently discovered phosphate-regulating hormones, such as fibroblast growth factor 23 (FGF23)²⁹.

In CKD, this prolonged but not permanent storage that homeostatically buffer phosphate in the circulation would likely have unique implications due to the necessity of increased duration of storage due to delayed excretion³¹. Further, the acute elevations after oral phosphate consumption, without hyperphosphatemia per se³², in both early CKD and healthy individuals is a means by which phosphate may potentiate adverse effects that requires investigation. For example, recent studies identify a role for calcifying vasculature as an acute influx depot³¹. Considering perturbations in efflux, in patients undergoing hemodialysis, the acute changes in circulating phosphate could not account for the acute bulk removal of phosphate in the dialysate and identified intracellular phosphate from muscle, a bioenergetically active tissue, as the transient exchangeable pool in these circumstances³³. Broad strokes of our understanding of phosphate movement can be outlined by these models, but precise locations (i.e., which cells), governing mechanisms, and consequences of prolonged but not permanent extra-circulatory phosphate storage are not well understood.

Organ crosstalk and developments in phosphate sensing

To allow for effective regulation of circulating phosphate availability, phosphate must be quantifiable in the various compartments of the body, including the extracellular and intracellular space^34–36. While the existence of phosphate sensors has been predicted for a long time³⁷, it is only recently that the first molecules involved have been identified. We can consider separately how the sensing and signalling of phosphate (Figure 1A), in its inorganic and CPP form, coordinate its regulation and movement in, out, and within the body (Figure 1B).

Figure 1: Multi-organ crosstalk in phosphate homeostasis.

A) Sensing and signalling of phosphate and regulatory hormones. B) Regulation of the movement and transformation of phosphate between different compartments. Abbreviations: phosphate (P_i), calcium-sending receptor (CaSR), fibroblast growth factor receptor 1 (FGFR1), parathyroid hormone (PTH), glycerol-3-phosphate (G-3-P), sodium phosphate co-transporter encoded by slc20 (PiT), lysophosphatidic acid receptor 1 (LAR1), reticuloendothelial system (RES), fibroblast growth factor 23 (FGF23), calciprotein particle (CPP), calciprotein monomer (CPM), PP_i (pyrophosphate)

Parathyroid hormone

Bioactive parathyroid hormone (PTH) is a phosphaturic and bone-regulating hormone released from parathyroid gland (PTG) chief cells. The acute response of PTH is critical for timely renal excretion of consumed phosphate²⁶. Through PTH receptor 1 binding, it inhibits renal phosphate reabsorption through reducing membrane sodium-phosphate transporters, Npt2a and Npt2c, encoded by slc34a1 and slc34a3, respectively, in the proximal tubule and stimulating the synthesis of 1,25(OH)₂D₃, active vitamin D hormone. Vitamin D is a critical mineral-regulating hormone; however, its primary regulatory target is calcium³⁸ and therefore it will not be discussed in detail.

It has long been understood that changes in circulating inorganic phosphate modulates PTH secretion from PTG^39,40, however until recently the mechanism was unknown. A recent study has shown that elevations in inorganic phosphate non-competitively antagonize the calcium-sensing receptor (CaSR) to increase PTH release⁴¹. The mechanism regulating PTH reduction in response to hypophosphatemia is still unknown. Longer term regulation of PTH and phosphate involves organ crosstalk and other phosphate hormones, as the PTG expresses important phosphate-related signaling molecules, such as fibroblast growth factor receptor 1 (FGFR1), co-receptor klotho, and the vitamin D receptor⁴².

PTH has critical but opposite effects on bone turnover dependent on exposure timing. Intermittent exposure to PTH is anti-osteoporotic and stimulates bone formation through canonical wnt-beta catenin-signaling activation⁴³. In contrast, continuous exposure to elevated PTH increases osteoclast activity, mainly through receptor activator of nuclear factor κB ligand (RANKL) and osteoprotegerin signaling, promoting demineralization and a continuous release of phosphate into the extracellular fluids. Secondary hyperparathyroidism is characteristic of renal osteodystrophy, the bone component of CKD-MBD and an important therapeutic target for both modulating bone-related outcomes but also cardiovascular events⁴⁴.

Fibroblast growth factor 23

FGF23 is a phosphaturic hormone mainly produced and secreted by bone and exerts negative feedback on PTH and active vitamin D production, and thus appears hierarchically as a primary regulator of phosphate homeostasis, over calcium. FGF23 binds to the binary complex of FGFR1c and klotho expressed in renal tubules to suppress phosphate reabsorption, thereby promoting urinary excretion⁴⁵.

Current data indicate that the PiT2 protein, a sodium phosphate co-transporter encoded by slc20a2, is instrumental in the appropriate secretion of FGF23 in response to changes in extracellular phosphate concentration^36,46,47. The absence of PiT2 in mice blunted the phosphate-dependent FGF23 secretion in whole animals, as well as in isolated ex vivo bone shaft cultures, indicating that this sensing role is independent of PTH or vitamin D endocrine regulations⁴⁷. However, the cellular mechanisms underlying the sensing are largely unknown. PiT2 can heterodimerize with its ortholog PiT1, encoded by slc20a1, at the plasma membrane and only these PiT1-PiT2 heterodimers allow for signal transmission⁴⁸. These heterodimers may have other binding partners, and whether the composition of this protein complex may vary between cell types and be responsible for phosphate sensing across other tissues remains to be studied.

Another putative phosphate sensor regulating FGF23 secretion is FGFR1c⁴⁹. However, it remains to be understood how it is activated by extracellular phosphate since FGFRs do not bind phosphate^46,50. FGFR1c may therefore represent an intracellular rather than extracellular phosphate sensor with a role in FGF23 secretion, or it may functionally interact with PiT2 or another phosphate transporter to control FGF23 secretion.

Lastly, CPMs extravasate through bone marrow sinusoids to induce FGF23 release¹⁵. Increasing levels of both phosphate and calcium induced FGF23 osteoblast expression, correlating with increasing CPP generation and inhibition of CPP formation attenuated FGF23 induction in osteoblasts in vitro regardless of free calcium and phosphate. Interestingly, amorphous CPMs were the major inducer of FGF23, and amorphous-to-crystalline CPP transformation attenuated FGF23, further adding nuance to our understanding of CPP signaling. The mechanism by which CPM-FGF23 signaling occurs in still unclear.

Klotho

Klotho is a ~140 kD single-pass transmembrane glycoprotein with a small intracellular domain and large (~130kD) extracellular domain⁵¹. Formation of the FGF23-klotho-FGFR ternary complex is required for activation of the canonical FGF signaling pathway, thus conferring the tissue specificity of FGF23 with the kidney and PTG^52,53. Klotho is cleaved on the cell surface by membrane-anchored secretases to release the entire extracellular domain. While it has been speculated that secreted klotho can confer responsiveness to FGF23 on cells that do not endogenously express klotho but do express FGFR1, it is unlikely as circulating levels of secreted klotho and FGF23 are too low to form the functional ternary complex^54–56. However, secreted klotho has been reported to exert multiple functions as a humoral factor independent of FGF23, including independent downregulation of Npt2a and modulation of wnt signaling pathways with known roles in ectopic calcification⁵⁷. Recent studies have raised the possibility that these seemingly promiscuous functions may be explained by lectin-like activities of secreted klotho binding to α2,3-sialyllactose of sugar chains in glycoproteins and gangliosides GM1/GM3 in lipid rafts^58–60. Indeed, therapeutic effects of secreted klotho on mouse models of cardiac hypertrophy, as discussed later, were recapitulated by reduction of GM1/GM3 in lipid rafts in the heart tissues⁶¹. Identification of lectin-like activity of secreted Klotho has opened a new field in klotho research.

Glycerol-3-phosphate

Kidney glycolysis has recently been identified as a phosphate sensing pathway, and glycerol-3-phosphate (G-3-P) as a matched phosphate hormone^62,63. Specifically, phosphate availability is rate-limiting for the production and release of glycerol-3-phosphate (G-3-P) into the circulation generated by kidney glycolysis⁶³. Circulating G-3-P then regulates FGF23 production. G-3-P was first identified as an FGF23 regulator in the context of acute kidney injury through renal vein molecular screening⁶², whereby it was found that G-3-P acetyl transferase mediates lysophosphatidic acid synthesis and lysophosphatidic acid receptor 1 (LAR1) to induce FGF23 release in bone. In response to a phosphate load, ¹⁸F-FDG imaging showed robust kidney-specific glycolysis and the in vivo release of G-3-P was Npt2a-dependent, perhaps leading to the proximal tubule specificity of this response⁶³. This landmark finding, linking cellular energy metabolism to extracellular phosphate homeostasis, opens a field of research for phosphate sensing. The potential role of this molecule or other similar reactions provides interest in a potential mechanistic link explaining why sodium-glucose co-transporter 2 inhibitors increase phosphate and FGF23.^64,65

Vascular consequences of homeostatic phosphate dysregulation

Through both direct and indirect effects, elevations in systemic or local phosphate promote and sustain multiple mechanisms of CVD progression (Figure 2).

Figure 2: Direct and secondary effects of elevations in phosphate on the cardiovascular and renal system.

Abbreviations: fibroblast growth factor 23 (FGF23), calciprotein particle (CPP), parathyroid hormone (PTH), phosphate (P_i), reactive oxygen species (ROS), endothelial nitric oxide synthase (eNOS), valvular interstitial cells (VICs), smooth muscle cells (SMCs), PP_i (pyrophosphate), left ventricular hypertrophy (LVH)

Endothelial dysfunction

Endothelial dysfunction encompasses an array of maladaptive alterations in cellular phenotype, which have been associated with the onset and progression of CVD^66,67. Human studies investigating the impact of phosphate-related stimuli on the endothelium have focused on impaired endothelium-dependent vasodilation as the outcome^68–70. Hyperphosphatemia and acute post-prandial phosphate elevation impair endothelium-dependent vasodilation^68–70, and intestinal phosphate binders can improve these measures in patients with CKD and mouse models with impaired kidney function, possibly by reducing serum phosphate, FGF23 and/or PTH levels^71,72.

High inorganic phosphate induces oxidative stress: increasing reactive oxygen species (ROS) and decreasing endothelial nitric oxide synthase (eNOS) levels, thus reducing bioavailability of nitric oxide (NO) in endothelial cells^69,73, leading to decreased endothelial-dependent vasodilation that is essential in maintaining vascular tone. Notably, PiT1 and PiT2, mediate high phosphate-induced endothelial dysfunction, through mechanisms related to mitochondrial oxidative stress^69,74. Given the role of phosphate in oxidative phosphorylation, modulation of mitochondrial driven pathways is reasonable. In cultured endothelial cells, high phosphate promotes endothelial dysfunction via several mechanisms, including inhibiting phosphoprotein phosphatases, activating protein kinase C (PKC) and AMP-activated protein kinase signaling pathways, impeding activation of the AKT/mTOR signaling pathway, and promoting apoptosis^69,74,75. In human coronary artery endothelial cells, high phosphate downregulates annexin 2⁷⁵, a multifunctional regulator for signal transduction, fibrinolysis, cell adhesion, and angiogenesis. Furthermore, high phosphate can induce endothelial-mesenchymal transition (EndMT) even in the absence of EndMT-inducer TGF-β⁷⁶. Mature CPPs have also been implicated in endothelial dysfunction ex vivo and in vitro, through reductions in NO availability^77,78. Endothelial cells are capable of internalizing CPPs, which increase cytosolic calcium, leading to ROS formation and generation of pro-inflammatory cytokines⁷⁹. These may have paracrine effects to local vascular smooth muscle cells (SMCs). Considering the constitutive and acute elevations of CPPs in CKD and following acute phosphate exposure, respectively, which may be recapitulated inadvertently in vitro, the role of free phosphate and CPPs-mediated outcomes may be conflated. Despite this, both clinical and experimental evidence support a role of phosphate elevations in endothelial dysfunction.

Cardiovascular calcification

Ectopic calcification across the vascular tree and in heart valves is the most prominent cardiovascular consequence of phosphate toxicity. Calcification of these tissues are characteristic of patients with CKD-MBD, but also occurs with healthy aging and is accelerated by diabetes^80,81. In a subgroup of patients from the Multi-Ethnic Study of Atherosclerosis who had moderate CKD and no clinical CVD, higher serum phosphate levels, even within the normal range, were associated with a greater prevalence of vascular and valvular calcification⁸². Further, in the general population higher levels of serum phosphate within the normal range are associated with increased coronary artery calcification⁸³. While statistics and clinical measurement techniques vary, there is a marked increased risk and, perhaps more importantly, a higher burden of cardiovascular calcification in patients with impaired kidney function⁸⁰.

Cardiovascular calcification is strongly associated with an increased risk of cardiovascular morbidity and mortality in both CKD patients and in the general population^84–87. Of note, systemic phosphate dysregulation or hyperphosphatemia is not necessary for the development of valvular and vascular calcification and in many cases occurs as a result of local inorganic phosphate abundance regulated by other factors, such as inflammation⁸⁸.

There is marked heterogeneity in how this pathology manifests across the valves and in the tunica intima or media of the vascular tree. However, despite this diversity, cardiovascular calcification is generally an active cell-mediated process that occurs as a result of vascular SMCs in the vascular wall or valvular interstitial cells (VICs) in the valve, transdifferentiating into an osteoblast-like cell. These cells, or others in the microenvironment, release calcifying extracellular vesicles (EVs) into the extracellular matrix, that forms a nidus for calcification, thus providing the substrate for homologous nucleation in the setting of microenvironment phosphate abundance^89–91.

It Is well-established that high phosphate directly induces osteogenic differentiation and calcification of vascular SMCs^92–94 and VICs^95,96, which is inhibited by calcification inhibitors fetuin-A and PPi^97–99: molecules also involved in CPP maturation. In CKD, expression of these inhibitors is largely reduced^18,100–102. In vitro studies have demonstrated that PiT1 mediates high phosphate-induced SMC calcification, via upregulation of Runx2⁹⁴, a key osteogenic transcription factor that is essential for SMC osteogenic differentiation and calcification^103–106. High phosphate induces SMC oxidative stress^106,107, a potent inducer of Runx2 upregulation and SMC calcification^103,104,108. Furthermore, activation of multiple pro-osteogenic signaling pathways, including the wnt/β-catenin, PKC, and bone morphogenic protein signaling^109–111, regulates high phosphate-induced SMC osteogenic differentiation and calcification. Further, it has been shown that in response to a phosphate load, calcifying vasculature acutely accrue inorganic phosphate, buffering phosphate in the circulation, identifying a participatory role of vasculature in systemic phosphate homeostasis^31,112.

Secondary to a direct effect of phosphate, FGF23 is thought to influence the vasculature, however the directionality is controversial. It has been posited that klotho presence in the artery and subsequent klotho-mediated FGF23 signaling was vasculo-protective¹¹³; however, another study has reported that FGF23 promoted vascular calcification¹¹⁴. Local expression of vascular klotho is also contentious^115,116; however, reduced soluble klotho correlates with vascular calcification in CKD patients and has been shown to reduce phosphate uptake of SMCs^117,118.

Calciphylaxis is a rare soft tissue calcification disorder occurring predominantly in dialysis patients whereby microvessels in the skin or subcutaneous adipose calcify leading to extremely painful lesions with poor survival^119–121. Elevations in serum phosphate, PTH, and FGF23 are risk factors, but mechanisms remain poorly understood^120,122. While it is a process associated with necrosis, it is still believed to be mostly a cell-mediated related to osteogenic differentiation of vascular SMCs^123,124.

Extracellular vesicles and calciprotein particles in ectopic calcification

In addition to increasing the expression of bone-related markers and inhibiting SMC marker genes, high phosphate-induced vascular SMC calcification is linked to increased secretion of EVs^98,125. EVs are membrane-bound nanoparticles released from cells that contain biologically active cargo (e.g., proteins, miRNAs, phosphate/calcium) that can mediate cell-cell communication and serve as critical initiators of calcification. In bone mineralization, the role of matrix vesicles, a type of EV released by osteocytes, is well-known, and EVs released by cardiovascular cells serve similar role in ectopic calcification¹²⁶. The released EVs are enriched in surface proteins that facilitate binding to the extracellular matrix and each other, as well as calcium and phosphate. Electron microscopy imaging supports the working understanding that intra-EV calcium and phosphate elevate until hydroxyapatite forms and bursts the cell membrane, which then serves as a nidus for homologous nucleation of hydroxyapatite (HAP)⁸⁹. At the initial stages of this pathology, microcalcifications form (1–5 um), which progress to macrocalcifications that cause hemodynamic stress¹²⁷. Large macrocalcification do not meaningfully regress, thus the microcalcification stage of disease progression presents a window of therapeutic opportunity whereby the lesions may regress and be amenable to cellular clearance. Further, intimal microcalcifications in atherosclerotic plaques are destabilizing leading to cardiovascular events^128,129.

We can make the mechanistic distinction between settings of cardiovascular calcification where there is abundant serum phosphate and extracellular phosphate (i.e., CKD), and settings where the source of phosphate that necessarily must incorporate into HAP crystals needs to be locally enzymatically generated (Figure 3). This typically occurs through phosphatase activity, such as tissue non-specific alkaline phosphatase (TNAP)⁹⁶, which also cleaves and thus locally reduces potent mineralization inhibitor PPi. Depending on the source of phosphate, evidence suggests that the EVs released from SMC are distinct. In the setting of high calcium and inorganic phosphate in vitro, SMC produce calcifying vesicles that are released through multivesicular bodies. However in setting of high organic phosphate, whereby local calcification will be dependent on appropriate phosphatase activity, EVs produced are larger with a likely different egress method (i.e., microparticles)^91,130,131. In the setting of high calcium and phosphate, EV loading of calcification inhibitors matrix Gla protein and fetuin-A, as well as TNAP is attenuated, and they are enriched with extracellular matrix binding proteins^91,98,130. However, in settings of organic phosphate abundance in the form of β-glycerol phosphate, TNAP is enriched through sortilin-mediated trafficking^90,132. It should be noted that local inflammatory cells in calcification, such as macrophages, are also capable of secreting calcifying EVs loaded with phosphate modulating activity¹³³, providing a putative link between inflammation and accelerated calcification⁸⁸. Given EVs appear to be a consistent factor of ectopic calcification initiation, therapeutic avenues exploiting EV-loaded cargos or synthetic EVs have gained interest.

Figure 3: Phosphate-centric mechanisms of ectopic cardiovascular calcification mediated through inorganic and organic phosphate sources.

Abbreviations: Phosphate (P_i), calciprotein particle (CPP), reactive oxygen species (ROS), extracellular vesicle (EV), PP_i (pyrophosphate), valvular interstitial cells (VICs), smooth muscle cells (SMCs), tissue non-specific alkaline phosphatase (TNAP), extracellular matrix (ECM).

CPPs have been shown to consistently induce SMC calcification in vitro^18,134–136, however there is controversy as to whether this effect is through a cytotoxic apoptosis-mediated pathway or through induction of osteoblastic cell differentiation following internalization¹³⁷. It has been posited that when CPPs are taken up by cells, lysosomes release the calcium and phosphate from the CPPs into the cytosol, which are then packaged into calcifying EVs and released as a cellular pass-through system^130,137. Elevations in intracellular calcium may also induce apoptosis. Apoptotic bodies are a subtype of EV that can also serve as calcification nidi. If occurring, in vitro studies indicate induction of osteoblastic differentiation appears limited to crystalline CPPs¹³⁶. In EVs and CPPs isolated from serum of patients with stage 5 CKD, both were able to independently potentiate calcification of vascular SMCs in vitro¹⁸. CPPs, whether created locally or extravasated from the circulation, may also directly incorporate into developing calcification, as they have been isolated from calcified atherosclerotic plaques⁷⁹. There is clear evidence of a correlation of circulating CPPs with cardiovascular calcification, however our understanding of how CPPs mechanistically coordinate with other factors is limited. Standardization of the synthesis and characterization of CPPs, as well as other nanoparticles, such as EVs, is an area of rapid progression in the field^16,138,139.

Extravascular consequences of dysregulated phosphate homeostasis

Bone

Bone demineralization is fundamental to CKD-MBD, characterized by deleterious effects on bone histomorphometry, growth and bone strength, leading to greater fracture risk^3,140–142. The mechanisms underlying the CKD-MBD related bone fragility are complex and multifactorial¹⁴³, and due to this complexity, very few data examine the direct role of phosphate. Early studies proposed high phosphate may induce osteoblast apoptosis, reduce bone formation, and inhibit bone resorption^144,145. Current treatment to manage the adverse effects of MBD in CKD, particularly bone fragility, includes anti-resorptive drugs used for the treatment of classical osteoporosis, which have recently been approved due to the ageing CKD population^146,147.

In both aging populations with healthy kidney function and CKD-MBD patients, we observe the interesting phenomenon of demineralization and impaired bone metabolism with concurrent mineralization of the vascular tree: the bone-vascular paradox. In a population-based longitudinal study of postmenopausal women, the progression of bone loss and vertebral fracture were associated with increases in aortic calcification¹⁴⁸. In hemodialysis patients, arterial calcification correlated with the prevalence of vertebral fractures, which were positively associated with mortality¹⁴⁹. Additionally, high phosphorus diets reduce bone density whilst simultaneously increasing vascular calcification in animal models of CKD¹⁵⁰, suggesting that phosphate activates different intrinsic signals in the tissue-specific microenvironments.

The mechanism of this paradoxical effect is not well understood; however, evidence supports elevated serum phosphate directly, as well as PTH and FGF23 as factors¹⁵¹. Secondary hyperparathyroidism induces phosphate and calcium release from bone and promotes osteoclast-mediated bone resorption that accelerates bone remodeling and high turnover, the most common form of metabolic bone disorder in CKD¹⁵². Vitamin D analogues used to reduce hyperparathyroidism and secondary bone resorption are associated with hypercalcemia and vascular calcification^153,154. Klotho or FGF23 knock out mice, which are both considered models of accelerated aging phenotypes, also exhibit simultaneous vascular calcification and decreased bone mineral density^51,155. Hyperphosphatemia and increased serum vitamin D levels in the FGF23 and klotho deficient mice may mediate the opposite regulatory role of FGF23-klotho axis in bone and vascular mineralization^51,155. A recent study has shown that EVs derived from the aging bone matrix released during bone resorption are transported to the circulation where they exacerbate vascular calcification¹⁵⁶. EVs from aged mice contained markedly more calcium and phosphate compared to young mice, with miRNA cargo as critical mediators of the effect, potentially mediating bulk mineral movement from bone to vasculature.

This paradox may also be attributed to the loss of PPi. Mice deficient in the ecto-nucleotide pyrophosphatase/phosphodiesterase, which enzymatically generates PPi, exhibit decreased bone density and increased vascular calcification¹⁵⁷. Bisphosphonates, analogues of PPi, increase bone density and inhibit vascular calcification in CKD^158–160. Bisphosphonates have also shown some efficacy in calciphylaxis progression¹⁶¹. Osteoblasts have over 100-fold higher alkaline phosphatase activity than calcifying osteoblastic SMCs¹⁶². In an in vitro study, TNAP inhibition markedly hindered bone formation in cultured osteoblasts in a concentration-dependent manner, but had no effect on SMC calcification¹⁶². Accordingly, the differences in the basal PPi-phosphate metabolism and their responses to local PPi-phosphate changes may account for the inverse mineralization observed in bone and vascular cells. However, in TNAP over expressing mice and in an adenine-induced CKD mouse model, orally administered TNAP inhibitor (SBI-425) attenuated vascular calcification without measurable secondary skeletal effects^50,163. Better understanding of distinct regulations and effectively managing local and systemic PPi and phosphate levels may alleviate adverse mineral outcomes in both bone and vasculature.

Heart failure and left ventricular hypertrophy

Cardiomyocyte mitochondrial changes disrupt cardiac function leading to heart failure^164–166. High phosphate has been shown to directly induce mitochondrial dysfunction and myocardial energy metabolism remodeling (i.e., shift towards glycolysis from oxidative phosphorylation) in cardiac cells in in vitro models contributing to heart failure¹⁶⁷. Phosphate critically downregulates peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC1α) which induces mitochondrial energy metabolism dysfunction in vitro and CKD-associated heart failure in mice¹⁶⁷. While the exact sensing mechanism is unclear, the phenomena was PiT1/2 dependent and mediated by epigenetic regulation of transcriptional factor interferon regulatory factor 1 (IRF1).

Secondary to phosphate itself, FGF23, klotho, and PTH have been implicated as direct effectors of cardiomyocyte function and subsequent LVH. In the setting of CKD and phosphate dysregulation, FGF23 levels can reach 1000x higher than would be observed in health, and have been consistently associated with LVH and cardiovascular mortality in CKD patients^168–172. FGF23 induces LVH through klotho-independent FGFR4-mediated activation of the PLCγ–calcineurin–NFAT signaling pathway^{170,173–175}. Intramyocardial or intravenous injection of FGF23 in mice induced LVH independent changes in cardiac contractility and blood pressure¹⁷⁰. Several studies in CKD-models show modulation of this pathway can be protective (FGFR4 KO and antibodies)¹⁷⁵, but direct phosphate regulation of the phenomena has not been shown. In a model of high FGF23 but hypophosphatemia and normal kidney function, elevated FGF23 did not result in CVD¹⁷⁶. Klotho-deficient animals display substantial LVH, however separating this phenomenon from the robust vascular calcification and FGF23 elevations is complicated^51,177. Klotho is not expressed in the heart or required for FGFR4 signaling; however, soluble klotho has been shown to be protective against cardiac fibrosis in various circumstances^178–180, potentially through modulation of TRPC6¹⁸¹. Lastly, PTHR1 activation in cardiomyocytes trigger calcium influx, activation of phospholipase C-PKC pathways, resulting in cell proliferation and myocardial hypertrophy^182–184.

Acute phosphate toxicity and chronic kidney disease

Acute phosphate nephropathy can occur in patients treated with phosphate-containing enemas¹⁸⁵, which is generally accepted to be a consequence of elevated tubule phosphate precipitating and thus causing luminal obstructions, epithelial injury, and fibrosis. Along this line of thinking, it has been shown that similar pathologic feed forward responses occur in CKD, whereby as functional renal mass declines (i.e., decreased number of nephrons), each nephron is required to handle and concentrate more phosphate mediated through reductions of Npt2a/c by elevating FGF23 and PTH, thus the nephrotoxicity of phosphate increases. Reductions in dietary phosphorus in rats can slow the decline of kidney function^186–188. This increased concentration of phosphate in the tubular fluid also facilitates maturation of intratubular CPMs into mature CPPs¹⁸⁹. The mature CPPs induce proximal tubular cell damage through binding to TLR4^189,190, followed by interstitial inflammation/fibrosis and nephron loss. Reduction of nephron number causes a reciprocal escalation of FGF23 and PTH to maintain the phosphate balance and triggers a deterioration spiral towards progressive nephron loss^189,191. Accordingly, human and epidemiological studies have shown higher serum phosphate or FGF23 predict accelerated decline in kidney function in CKD patients¹⁷² and in the general population, higher serum phosphate levels were predictive of the development of end stage renal disease¹⁹².

Potential role of sex in phosphate-mediated cardiovascular disease

A direct link between the sex differences observed in phosphate metabolism and CVD presentation is plausible. Serum phosphate declines and is similar in both sexes with aging, until approximately 45, where the serum phosphate levels in females increase, whilst males continue to decrease¹⁹³. Females and males have differences in CVD presentation: females demonstrate greater arterial stiffness and valvular fibrosis, rather than calcification per se^194,195, and have a higher prevalence of LVH^196–199, whereas obstructive coronary artery disease and calcification, characteristic of intimal calcification specifically, is more prevalent in males²⁰⁰. Further, risk factors for calcification progression are different^201,202. Females also have a lower incidence of CVD compared to males however it rises following menopause^203,204. It is attractive to speculate that sex differences in phosphate metabolism after menopause may, in part, underlie this rise in CVD incidence.

Clinical and experimental evidence suggest that sex hormones, in part, mediate these differences in serum phosphate. In an analysis of 2688 Caucasian participants >45 years of age in the Rotterdam Study, estradiol was inversely associated with serum calcium and phosphate in both pre- and post-menopausal females²⁰⁵. Further, adjustment for serum estradiol diminished sex differences in serum calcium while only adjustment for serum testosterone diminished sex differences in serum phosphate. These data suggest an impact of these hormones on phosphate metabolism and the presence of sex-specific factors.

Both testosterone and estrogen can influence mineral metabolism, and their respective receptors are expressed in the kidney²⁰⁶. Estradiol downregulates renal Npt2a to reduce phosphate reabsorption²⁰⁷. This observation is in line with lower estrogen levels and post-menopausal elevations in serum phosphate. The relationship of testosterone with phosphate is less clear. GnRH inhibitors-mediated testosterone suppression increases serum phosphate^208,209 and higher testosterone levels associate with more efficient elimination of phosphate following an oral challenge independent of changes in PTH or FGF23²¹⁰. The Osteoporotic Fractures in Men study revealed a significant inverse association between testosterone levels and serum phosphate levels, independent of FGF23 levels²¹¹. Accordingly, in adult mice, orchiectomy elevated the expression of renal phosphate transporters responsible for phosphate reabsorption, which would support a phosphaturic effect of testosterone²¹².

Sex hormone deficiency alters mineral metabolism by inducing bone lost in both males and females, leading to osteoporosis and associated increased fracture risk^213,214. These effects may indirectly participate in the bone-vascular paradox. Sex differences in the acute response to an oral phosphate challenge over four hours identified that females excreted endogenous calcium and more phosphate in response to the challenge compared to males²¹⁰. This phenomenon was present in both pre- and post-menopausal females, however correlated with higher resorptive bone biomarkers relative to resorptive markers suggesting a role of bone in mediating these sex differences in phosphate metabolism²¹⁰.

Aging is often linked to phosphate dysregulation, whereby hyperphosphatemia or mineral metabolism dysregulation is considered by some an accelerated aging phenotype⁵⁷. Both klotho and FGF23 KO mice are utilized as pre-mature aging models, with a phenotype that is associated with osteoporosis, systemic vascular calcification, hyperphosphatemia, and premature death, all of which can be attenuated by a low phosphate diet⁵⁷. Healthy aging is associated with vascular calcification, mild reductions in kidney function, reductions in klotho, and elevations in FGF23 despite lower serum phosphate^54,193,215. In aging mice, increases in bioavailable phosphate induced robust elevations of FGF23 that was greater in females compared to males, and in older females compared to younger²¹⁶. Thus, the impact of aging and sex on phosphate homeostasis are intermingled. These data are all very suggestive of direct effects of sex on phosphate metabolism, however a direct link to cardiovascular disease in humans has not been established.

Dietary phosphate exposure and homeostatic balance

Intestinal phosphate transport

Of the methods to modulate CVD consequences of phosphate dysregulation, attempts to attenuate phosphate absorption from the small intestine is the most direct, and our understanding of how phosphate is absorbed is rapidly expanding to inform novel pharmacologic interventions. Phosphate uptake along the digestive tract occurs through two independent pathways: an active component that is saturable at low luminal phosphate concentrations (in the micromolar range), and a passive unsaturable component^217,218.

Based on mice and rat studies, active transcellular phosphate transport was thought to be primarily driven by Npt2b^218–221. However, current evidence describes substantial interspecies heterogeneity^{219,222–224} and supports that its role in humans is limited. Individuals with inactivating mutations do not have a defect in intestinal absorption^225–227 and a Npt2b inhibitor, ASP3325, failed to influence phosphate absorption in late CKD patients^228,229. This raises the possibility for roles of other transporters, thus PiT1 and PiT2 in the gut and their regulation by dietary phosphate undoubtedly require further clarification^{217,230–232}. A pan inhibitor of Npt2b and PiT1 and PiT2 transporters had improved efficacy²³³ but has not been further developed.

While the activity of the transcellular pathway predominates at low luminal phosphate concentrations, the unsaturable paracellular pathway contributes to the vast majority of intestinal phosphate absorption above a luminal phosphate concentration of 1 mM, at which the transcellular transporter capacity is saturated. Measurement of luminal phosphate concentrations are in the millimolar range^223,234,235, making the unsaturable paracellular phosphate uptake quantitatively the most important overall mechanism of phosphate absorption under typical conditions of dietary phosphate availability. Recently, a direct measurement of intestinal phosphate absorption in humans demonstrated that the efficiency of intestinal phosphate absorption is retained in patients with moderate CKD and low levels of 1,25(OH)₂D₃²³⁶, illustrating the importance of paracellular phosphate absorption and the need to identify the molecules responsible for this transport in the gut. Accordingly, tenapanor, a drug that blocks the intestinal sodium-hydrogen exchanger to block paracellular transport of phosphate in CKD patients, was FDA approved in October 2023. It functions by lowering intracellular pH and modulating the tight junctions between the epithelial cells increasing the transepithelial electrical resistance thus impeding paracellular phosphate transport^237,238. The same compound has been FDA approved since 2019 for use in irritable bowel syndrome with constipation. Whilst this first-in-class drug generates potential to further reduce phosphorus levels in adults who have an inadequate response to phosphate binders, long term safety assessments are warranted as pharmacological modulation of tight junctions may have unanticipated secondary effects, such as impacting paracellular movement of other solutes.

Dietary phosphate and the role of acute phosphate elevation in CVD

Though phosphorus is an essential nutrient, dietary phosphorus deficiency is very rare due to the widespread presence of phosphorus in the food supply. Mean dietary phosphorus intake is estimated ~1400 mg/d, compared to the recommended daily allowance of 700 mg/d for adults in the United States^239,240. This value is likely underestimated due to limitations in the nutrient databases for phosphorus content²⁴¹, owing in part to the wide-ranging use of phosphate-containing food additives and that phosphorus is not required to be quantified or included on the food label²⁴². Note that the nutrient requirement is for the element phosphorus, even though phosphorus exists as phosphates in the diet. This distinction is important when referring to phosphorus in milligrams or grams, as is standard in the field of nutrition. Phosphorus bioaccessibility, the amount of phosphorus accessible for intestinal absorption, varies greatly by source. It is lowest from plant sources, estimated at ~30–50%, followed by animal sources ~60–70%, and the highest from inorganic phosphate-containing food additives ~90–100%^243–245. Plant sourced phosphate, mostly in the form of phytates, must be hydrolyzed prior to absorption²⁴⁶, whilst inorganic phosphate-containing food additives (e.g., sodium phosphates) depend only on solubility to release the phosphate ions for absorption.

Dietary phosphorus restriction in CKD patients is recommended by both KDIGO³ and KDOQI²⁴⁷ guidelines in the treatment of hyperphosphatemia. However, clinical trials that test the effectiveness of dietary phosphate restriction are scarce and are especially difficult to conduct over long enough duration to observe effects on clinical outcomes. Some observational data indicate associations with higher dietary phosphate intake and worsened mortality and cardiovascular outcomes in CKD patients. A secondary analysis of the multinational DIET-HD cohort study of hemodialysis patients²⁴⁸ found that each 896 mg/day higher of dietary phosphorus intake was associated with an increased all-cause mortality and cardiovascular mortality of 15 and 18%, respectively, and greater consumption of phosphorus from poorly bioavailable plant sources was associated with lower mortality²⁴⁹. Two RCTs in patients on hemodialysis with hyperphosphatemia found that limiting inorganic phosphate-containing food additives significantly reduced serum phosphate over 3 months compared to the control of standard care^250,251. While these data support current recommendations for patients with advanced CKD to limit highly bioaccessible forms and overall phosphate intake, consideration of overall nutritional status and avoiding malnutrition is paramount, as some studies have also shown greater mortality risk associated with lower dietary phosphorus intakes^252–254. Thus, greater strength of evidence for the effectiveness and risk-benefit of dietary interventions in CKD patients must be pursued. In alignment with these gaps in knowledge, the potential negative consequences of excess intake of highly bioaccessible phosphate in the general population has yet to be sufficiently studied.

While in advanced CKD, patients present with hyperphosphatemia, the role of acute elevations of serum phosphate after meals may provide a linking factor of the associations CVD risk in normophosphatemic conditions. Mild reductions in kidney function with normophosphatemia lead to higher and more persistent acute elevations of phosphate in response to an oral challenge³². There are also clinically recognized differences in acute phosphate homeostasis in sex²¹⁰ and race²⁵⁵ that should be investigated further. Relating to vascular calcification, transient increases in dietary phosphate in a rat model of CKD induced more robust vascular calcification than those animals fed the same amount of phosphorus over the same time periods at a consistent level²⁵⁶. Acute fluctuations in circulating phosphate may have a more important role than previously recognized.

Treatment strategies and new frontiers

As our knowledge of phosphate homeostasis evolves, so does the knowledge of the complexity of homeostasis with multiple organ systems and hormones involved. Unfortunately, our treatment approaches have not similarly advanced although there is hope. Novel therapies that extend the half-life of pyrophosphate or increase tissue activity are under development, offering hope for genetic disorders and severe calcification disorders such as calciphylaxis^257–261 and potentially for cardiovascular calcification. While data strongly suggest vascular calcification confers substantial cardiovascular risk, due to a lack of therapeutics that directly modulate cardiovascular calcification in humans, the magnitude of risk reduction that would be achieved by preventing or reversing vascular calcification has yet to be definitively established. Further, while early stages of intimal calcification/microcalcifications are destabilizing, certain advanced ectopic calcification lesions have been predicted to have stabilizing effects on plaques²⁶². Thus, therapies that reverse or regress cardiovascular calcification would need to be carefully considered in disease and lesion localization context.

In the absence of dietary phosphate reduction efficacy or adherence, the next strategy is to prevent phosphate absorption. The traditional approach of phosphate binders in reducing serum phosphate levels remains efficacious, but difficult for patients to follow, and evidence for influencing hard outcomes, such as mortality is lacking despite their widespread use³. Our increased understanding of the intestinal absorption of phosphate has led to novel therapies that directly inhibit this absorption and potential efficacy with respect to prevention of hard outcomes in hyperphosphatemic patients is an exciting prospect.

Recent discoveries in the CPP field are reframing how we think about circulating phosphate and the importance of buffering circulating phosphate systemically. Previous research with respect to molecular phosphate signalling must now be re-contextualized accounting for inadvertent CPPs-related effects. Further, identification of its potential role as a phosphate signalling moiety with unique sensors promotes research into CPPs as a therapeutic target. Current research is dedicated to investigating whether circulating CPP measurements will provide valuable clinical information into CKD patient prognosis, and investigating these measurements into the general non-CKD population in different demographics is also exciting^23–25,137. Standardization of these measurements are needed to push this endeavor forward.

Therapeutics that directly target calcification initiation and interfere with systemic FGF23 effects are exciting new frontiers of active interest, accelerated by technological advances in -omics and single cell data approaches. Additionally, advancing nanoparticles technologies will also push the field forward with respect to EVs and CPPs, which clearly have important roles in both soft tissue and bone mineralization. Advances in our understanding of phosphate in CVD are rapidly progressing and with hope, so too will treatment strategies.

Highlights:

• Either through direct signalling, or through mediators such as hormones, calciprotein particles or calcifying extracellular vesicles, phosphate contributes to various cardiovascular disease pathologies.

• Recent advances in phosphate sensing are accelerating our understanding of homeostasis.

• Buffering of circulating free phosphate, in the form of calciprotein particles or through regulated movement into extra-circulatory spaces, is an adaptive mechanism that may become maladaptive in settings of impaired kidney function.

• The negative consequences of phosphate dysregulation on the cardiovascular system are predominantly considered within the context of advanced chronic kidney disease, however developments suggest aspects may also extend to mildly impaired kidney function and the general population.

Non-Standard Abbreviations

CKD-MBD

chronic kidney disease-mineral bone disorder

CPP

calciprotein particle

CPM

calciprotein monomer

EndMT

endothelial-to-mesenchymal transition

EV

extracellular vesicle

eNOS

endothelial nitric oxide synthase

FGF23

fibroblast growth factor 23

FGFR

fibroblast growth factor receptor

G-3-P

glycerol-3-phosphate

Npt2

sodium phosphate co-transporter encoded by slc34a

PiT

sodium phosphate co-transporter encoded by slc20

PKC

protein kinase C

PPi

pyrophosphate

PTH

parathyroid hormone

PTG

parathyroid gland

ROS

reactive oxygen species

VIC

valvular interstitial cell

SMC

smooth muscle cell