Intestinal Phosphorus Absorption: Recent Findings in Translational and Clinical Research

Abstract

Purpose of review:

The purpose of this review is to discuss recent findings in intestinal phosphorus absorption pathways, particularly the contributions of paracellular versus transcellular absorption, and the differential findings from studies using in vitro versus in vivo techniques of assessing phosphorus absorption in experimental animal studies.

Recent findings:

Experimental animal studies show that in vivo effects of low phosphorus diets, 1,25D, and chronic kidney disease on intestinal phosphorus absorption efficiency contradict effects previously established ex vivo/in vitro. Recent in vivo studies also suggest that the paracellular pathway accounts for the majority of phosphorus absorption in animals across very low to high luminal phosphate concentrations. The data from experimental animal studies correspond to recent human studies showing effectiveness of targeted inhibition of paracellular phosphate absorption. Additionally, recent human studies have demonstrated that NaPi-2b inhibition alone does not appear to be effective in lowering serum phosphate levels in patients with CKD. Pursuit of other transcellular phosphate transporter inhibitors may still hold promise.

Summary:

In vivo animal and human studies have added to our understanding of intestinal phosphorus absorption pathways, regulation, and mechanisms. This is beneficial for developing effective new strategies for phosphate management in patients with chronic kidney disease.

Introduction

Phosphorus is an essential mineral that humans must consume through their diets. However, phosphorus is found so widespread in our food supply that deficiency is rare. Conversely, dietary phosphorus intake in excess of requirements is common. Phosphorus homeostasis is controlled by a multi-tissue axis including intestine, bone, kidney, and parathyroid glands. The major site of phosphorus regulation occurs at the kidneys, under the influence of phosphaturic hormones: fibroblast growth factor 23 (FGF23) and parathyroid hormone (PTH). Mechanisms of extracellular phosphate sensing have been largely enigmatic(1, 2); however, a recent study suggests a role of fibroblast growth factor receptor 1 (FGFR1) and the gene product of Galnt3 in increasing FGF23 in response to high phosphorus(3).

Intestinal phosphorus absorption occurs by two pathways: transcellular (“active”) and paracellular (“passive”). The transcellular pathway is saturable, predominates at low intestinal luminal phosphate concentrations, and relies on sodium-dependent phosphate transporters located in the BBM. The paracellular pathway is sodium-independent, non-saturable displaying a linear relationship to intestinal luminal phosphate concentration, and is dominant at high phosphate loads. The transcellular pathway has been considered the regulated component of phosphorus absorption, whereas the paracellular pathway has been considered unregulated. The two factors best understood to affect transcellular phosphorus absorption are 1,25D and low phosphorus diets, both of which increase BBM levels of the main known intestinal phosphate transporter, sodium-dependent phosphate co-transporter protein 2b (NaPi-2b). While low dietary phosphorus indeed increases 1,25D levels, there is a 1,25D-independent effect of low dietary phosphorus to increase NaPi-2b and phosphate transport, as has been demonstrated in vitamin D receptor null mice(4, 5). Though, the mechanism by which low dietary phosphorus increases NaPi-2b independently of VDR is still unclear.

As kidney function declines with chronic kidney disease (CKD), the kidney’s ability to regulate phosphorus homeostasis is impaired. Rising PTH and FGF23 increase fractional phosphorus excretion which effectively keeps serum phosphorus within normal range until the late stage of disease, when this compensatory mechanism is eventually overwhelmed, and serum phosphorus rises. These hormonal changes also include reduced 1,25-dihydroxyvitamin D (1,25D) as a result of rising FGF23 suppressing 1,25D. The primary effect of 1,25D on phosphorus homeostasis is understood to be to increase intestinal phosphorus absorption by upregulating NaPi-2b. Thus, declining 1,25D that occurs with progressive CKD would be assumed to be accompanied by decreased intestinal phosphorus absorption, which would presumptively provide at least some compensation for the failing kidney. However, it is unclear whether and to what extent this occurs (6).

Importantly, much of the current understanding of intestinal phosphorus absorption and its regulation comes from studies using ex vivo/in vitro methods to assess phosphate transport. However, some recent studies that have used in vivo methods to assess intestinal phosphorus absorption have produced results that contradict knowledge established in vitro.

In Vivo Intestinal Phosphorus Absorption: Effects of 1,25D, Dietary Phosphorus, and Chronic Kidney Disease

Table 1 summarizes various techniques and approaches used to assess intestinal phosphorus absorption ex vivo/in vitro and in vivo and clinically. Ex vivo and in vitro methods, while having the advantage of tightly controlled conditions, may be less physiologically relevant than in vivo and clinical approaches. Indeed, the definition of absorption requires the substance to be transported into blood. A limited number of intestinal phosphorus absorption studies in animal models have been conducted using in vivo methods. However, several themes are emerging from these investigations that are at odds with knowledge gained from studies that have used ex vivo/in vitro methods. Using the jejunal ligated loop method in healthy Sprague-Dawley males, we found that a low phosphorus diet of 0.1%P increased plasma 1,25D compared to diets of 0.6%P and 1.2%P; but there were no differences in intestinal phosphorus absorption efficiency among the three diet groups(7). Notably, the phosphate concentration injected into the lumen of the rats was very low ([0.1mM]), where active absorption would be maximized. Similarly, in an earlier study in 5/6^th nephrectomized rats, Marks et al.(8) also showed no effect of a phosphorus deficient diet of 0.02%P compared to a moderate phosphorus diet of 0.52%P on intestinal phosphorus absorption efficiency using the same jejunal ligated loop method.

Table 1.

Techniques used to assess intestinal phosphorus uptake/absorption

Technique	Type	Brief Description	Advantages	Disadvantages	Example of Use
Oral gavage	in vivo Measures absorption	A transport solution with 33P* is injected via a small tube directly through the esophagus and into the stomach or small intestine via a syringe. Absorption of 33P into blood is measured.	Measures absorption with all tissues intact. Does not need to be a terminal procedure.	Cannot completely isolate paracellular absorption component. Measurement may be affected by tissue sequestration of circulating phosphorus.	(35)
Ligated loop	in vivo/in situ Measures absorption	Ligatures are placed to close off an intestinal segment (~5cm). A transport buffer with 33P is injected into the lumen and tied off. Absorption of 33P is measured into blood and from the loop.	Keeps circulation and nervous system intact. Can test individual intestinal segments.	Cannot completely isolate paracellular absorption component. Anesthesia may affect absorption. Measurement may be affected by tissue sequestration of circulating phosphorus. Terminal procedure.	(10)
Everted gut sac	ex vivo/in vitro Measures transport	Excised intestinal segment (~5cm) is everted around a glass rod to form a sac that is filled with a buffer. Both ends are sutured closed and sac is incubated in buffer with 33P. Transport of 33P into the sac is determined.	Can test individual intestinal segments and isolate transcellular and paracellular components. Measures transport through both apical and basolateral membranes.	Tissue is detached from nervous system and circulation. Variability in measurement may occur if sac volumes differ.	(8)
Everted sleeve	ex vivo/in vitro Measures uptake	Excised intestinal segment (~2–4 cm) is everted around a glass rod and incubated in buffer with 33P followed by an uptake termination step. Tissue is digested, and 33P uptake into the tissue is measured.	Can test individual intestinal segments and isolate transcellular and paracellular components.	Tissue is detached from nervous system and circulation. Measures uptake into the tissue, not transport through both membranes.	(15)
Ussing chamber	ex vivo/in vitro Measures transport	A segment of epithelial membrane is placed between two chambers and ion movement across the membrane is measured by the change in the potential difference upon crossing the epithelium.	Can tightly control temperature, pH, and buffer concentration. Can measure both active and passive components.	Tissue is detached from nervous system and circulation.	(12)
BBMV Rapid filtration	in vitro Measures uptake	The mucosa is scraped from an intestinal segment and the brush border membrane vesicles isolated. The vesicles are incubated in uptake buffer with 33P. The reaction mixture is transferred to a filter and 33P is counted.	Can test individual intestinal segments and isolate transcellular and paracellular components.	Measures uptake into vesicles and not absorption per se.	(36)
24h urine	in vivo/clinical Absorption biomarker	Urine is collected from subjects over a timed 24h period and phosphorus is measured.	Noninvasive, easy to collect and measure.	Prone to timing errors/incomplete collections. May not reflect absorption, could also be affected by renal excretion rat and bone turnover.	(37)
Metabolic balance	in vivo/clinical Measures absorption	Animals are placed in individual metabolic cages. Phosphorus is measured from collected urine and feces, as well as food consumed over the time period. Net absorption can be calculated from intake and fecal excretion. In humans, balance measurements are made from controlled study diets and complete urine and fecal collections.	Does not require radioactive substances and can be performed in addition to other absorption methods in animals.	Net absorption measurement is not as precise and reflects both transcellular and paracellular pathways.	(10)
Fractional Absorption with Radiotracer	in vivo/clinical Measures absorption	33P is administered orally and by IV. Fecal, urine, and blood collection allow for 33P measurement and kinetic modeling.	Allows for direct absorption measurement in vivo.	Burdensome to participants, requires radioactivity dosing.	(11)

^*32P may also be used with all techniques. ³³P has a lower beta emission energy so it is safer for use in humans.

Transport = movement across both the apical and basolateral membranes.

Uptake = movement across the apical membrane into the cell.

Absorption = movement across both the apical and basolateral membranes and into circulation.

The effect of CKD progression on intestinal phosphorus absorption efficiency using in vivo models is also contrary to what would be predicted based on presumed effects of decreased 1,25D. This has now been shown independently in three different rat models of CKD by different research groups. Marks et al.(8) first reported that intestinal phosphorus absorption efficiency measured by jejunal ligated loop was not different between 5/6^th nephrectomized and sham rats, despite markedly lower 1,25D in the CKD rats. Turner et al.(9) used an oral gavage method in an adenine-induced rat model of CKD and also found no difference in intestinal phosphorus absorption in rats with mild or advanced CKD compared with control rats. Finally, in the Cy/+ rat model of progressive CKD-mineral and bone disorder, we recently showed that jejunal intestinal phosphorus absorption was slightly yet statistically higher in the CKD rats compared to controls, using the ligated loop method with a low phosphate concentration injected into the lumen ([0.1mM P])(10). Again, this was despite lower 1,25D in the CKD rats. In the same study, we also found that sodium-independent absorption was similarly higher in CKD rats compared to controls, as was % net phosphorus absorption from a 4-day metabolic cage study – which, on the relatively high phosphorus diet (0.7%P) would be expected to primarily reflect paracellular absorption. In a clinical study using a radioisotopic tracer method, we have also recently reported that fractional intestinal phosphorus absorption was not noticeably different between moderate stage CKD patients compared to control participants on a controlled high phosphorus diet, though the CKD patients had lower 1,25D levels as expected(11). Together, results from these in vivo technique studies suggest a lack of effect of low phosphorus diets to increase intestinal absorption efficiency, and a lack of an expected reduction in intestinal phosphorus absorption from reduced 1,25D levels in the context of CKD, even when assessed in conditions providing phosphate concentrations that in vitro evidence suggests would favor transcellular transport.

Transcellular versus Paracellular Pathways of Intestinal Phosphorus Absorption

The relative contributions of the transcellular and paracellular pathways and their mechanisms and regulation are of great importance in developing effective approaches for phosphate management. Transcellular absorption is considered to be dominant at low luminal phosphate concentrations, and then decreases in importance as luminal phosphate concentration rises, such as with a high phosphorus diet. With high phosphorus diets, it is generally thought that paracellular absorption is dominant and that transcellular absorption plays a very minimal role. Therefore, because 1,25D regulates the transcellular component, it is reasonable to infer that 1,25D levels would have a greater influence on phosphorus absorption efficiency when dietary phosphorus intake is low and have a lower or even negligible influence on phosphorus absorption efficiency when dietary phosphorus intake is high. Would calcitriol or its analogues increase intestinal phosphorus absorption but only at low dietary phosphorus intake levels? This is what was shown in a recent study by Hernando et al.(12), where WT or Slc34a2 (NaPi-2b) null mice were administered calcitriol and intestinal phosphate flux was measured by Ussing chamber under low and high apical phosphate concentrations to assess conditions where transcellular transport and paracellular transport would be favored, respectively. With a low apical phosphate concentration, phosphate flux in jejunal tissue from WT mice was increased in response to calcitriol, but this effect was absent in the Slc34a2 null mice. However, with a high apical phosphate concentration favoring paracellular transport, there was no effect of calcitriol on phosphate flux in either WT or Slc34a2 null mice. These results were similar using the ileal ligated loop method. As observed in the accompanying commentary(13), this study provides evidence that the effect of calcitriol to increase intestinal phosphorus absorption is entirely through its effect on transcellular absorption, not paracellular, and that this is completely mediated by NaPi-2b.

Understanding how these two pathways function in relation to each other is of practical clinical importance. In studying intestinal phosphorus absorption, the choice of absorption assessment technique, whether in vivo or ex vivo/in vitro, selection of animal models or humans, intestinal segment studied, background diet phosphorus intake, and luminal phosphate concentration all can affect the estimates of transcellular versus paracellular absorption(14) (Table 2). Using the in vitro everted sleeve technique in rat jejunum, Marks et al.(15) demonstrated that the transcellular (sodium-dependent) pathway predominates at 73% of total phosphate uptake at the very low luminal phosphate concentration of 0.1 mM (about 10-fold lower than what they measured as typical luminal concentrations). The proportion of total phosphate uptake at a high luminal phosphate concentration of 10 mM was lower, but still accounted for the majority – 53% – of phosphate uptake. In the same study, the investigators used identical luminal phosphate concentrations with the in vivo jejunal ligated loop method to compare with their results obtained in vitro. The transcellular pathway accounted for only approximately 1/3 of total phosphorus absorption in vivo and was relatively stable across the very wide range of luminal phosphate concentrations (0.1 mM to 10 mM). Thus, the vast majority (approximately 2/3) of total phosphorus absorption in the rat jejunum was attributed to the paracellular (sodium-independent) pathway and did not change from very low to very high luminal phosphate concentrations. Recently, we corroborated this finding in a rat model of CKD (Cy/+ rat and normal littermates), also using the in vivo jejunal ligated loop technique with the very low injected phosphate concentration of 0.1 mM. Similar to Marks et al., we found that transcellular (sodium-dependent) absorption accounted for approximately 1/3 (23–33%) of total phosphorus absorption(10). Some fraction of measurements of in vivo sodium-independent absorption may be an artifact of residual luminal sodium, but the implications of a predominant paracellular pathway are relevant to the quest for targeted phosphate lowering medications.

Table 2.

Recent studies highlighting differences among techniques in estimating the proportion of sodium-dependent (transcellular) intestinal phosphorus uptake/absorption.

			Na+ dependent (transcellular), % of total
Species	Technique	[P] (mM)	Duodenum	Jejunum	Ileum	Ref
Rat (SD)	Everted gut sac	0.1	48	73	~0	(15)
Rat (SD)	Everted gut sac	10	48	53	~0	(15)
Rat (SD)	Ligated loop	0.1	~0	32	52	(15)
Rat (SD)	Ligated loop	10	~0	29	74	(15)
Rat (Han:SD)	Ligated loop	0.1	-	23–33	-	(10)
Rat (Cy)	Ligated loop	0.1	-	24–25	-	(10)
Rat (SD)	Oral gavage	0.5	~0			(35)

SD = Sprague Dawley; Cy = Cy model of CKD

Inhibition of Transcellular versus Paracellular Absorption

The NaPi-2b gene was first identified as highly expressed in human lung, small intestine, and kidney(16), and has since been shown as the transporter responsible for the vast majority of transcellular (sodium-dependent) intestinal phosphorus absorption and about half of total phosphorus absorption in rodents when measured with an ex vivo method(17). Thus, NaPi-2b has been a major focus of drug targeting for reducing intestinal phosphorus absorption to treat hyperphosphatemia in patients with CKD. This year, results from a human phase 1b trial of a NaPi-2b inhibitor, DS-2330b – alone and in combination with sevelamer, were reported(18) with an accompanying commentary(19). Despite preclinical rat studies showing effectiveness of the drug to reduce phosphorus absorption, the 14-day human trial did not show efficacy in reducing serum phosphate. This comes following similarly discouraging results from a clinical trial of another NaPi-2b inhibitor, ASP3325 (20) – which had likewise shown effectiveness for reducing serum phosphate in an adenine-induced rat model of CKD as well as healthy rats on a high phosphorus diet(21). An effect of this NaPi-2b inhibitor in rats consuming a high phosphorus diet – presumably when transcellular absorption would be diminished – is notable. Still, the clinical trial for ASP3325 did not show efficacy in reducing serum phosphate in humans(20). These clinical trials highlight three main unknowns: 1) the relative contribution of NaPi-2b compared with other phosphate transporters to transcellular phosphorus absorption in humans, 2) how this might change with kidney disease, and 3) the relative importance of transcellular versus paracellular absorption at luminal phosphate concentrations one might observe with any reasonable human diet(22, 23).

Recently, a novel pan-inhibitor of sodium-dependent phosphate transporters, EOS789, has been shown to reduce fecal phosphorus excretion and urinary phosphorus excretion in normal rats in a dose-dependent manner, indicating effectiveness in reducing intestinal phosphorus absorption on a relatively high phosphorus diet of 0.81%P(24, 25). Significant reductions in serum phosphate, PTH, and FGF23 were also observed with higher EOS789 dosing. EOS789 inhibits not only NaPi-2b, but also sodium-dependent phosphate transporter (PiT)-1 and PiT-2. Further, unlike the studies of NaPi-2b inhibition alone, EOS789 had an indication of efficacy for reducing intestinal phosphorus absorption in hemodialysis patients when assessed directly by radioisotopic tracer method(26). However, this was only observed in a secondary analysis between the higher dose versus placebo, so further confirmation is needed. These findings indicate that efforts to inhibit transcellular phosphorus absorption may yet be fruitful, but it will likely not be through NaPi-2b inhibition alone.

As discussed above, paracellular phosphorus absorption appears to predominate in vivo across a broad range of luminal phosphate concentrations – accounting for up to ~ 2/3 of total phosphorus absorption. Thus, inhibiting paracellular absorption is an approach with great potential. Indeed, the sodium/hydrogen exchanger (NHE3), tenapanor, which decreases paracellular phosphate permeability(27), significantly reduced serum phosphate in hyperphosphatemic hemodialysis patients in a phase 3 randomized controlled trial(28). Further, a recent rat study showed that tenapanor combined with sevelamer had synergistic effects, with greater reduction in 24-h urine phosphorus in rats on the combination treatment compared to either drug alone(29).

Other mechanisms to target the paracellular pathway may exist as well. A study recently demonstrated that lithocholic acid, a bile acid, increased serum phosphate and vascular calcifications in mice with CKD but did not affect levels of intestinal sodium-dependent phosphate transporters and only increased sodium-independent (paracellular) phosphate permeability in everted gut sacs(30). These effects were found to be vitamin D-dependent and mediated through the tight junction protein, claudin 3. These findings highlight the complexity of known and yet to be discovered regulators of dietary phosphorus absorption.

Species Differences in Intestinal Phosphate Transporters

Recently, Ichida et al.(31) measured mRNA expression of NaPi-2b, PiT-1, and PiT-2 transporters across intestinal segments among rats, dogs, Cynomolgus monkeys, and humans. In human intestine, they found that PiT-1 was the most highly expressed of the three transporters in all segments. Species differences may explain at least in part why NaPi-2b inhibitors ASP3225 and DS-2330b had shown promise in preclinical rat studies, but then were unsuccessful in reducing serum phosphate in human clinical studies(18, 20). Further, evidence of a still unidentified sodium-dependent transporter has been reported in rats(32), as well as an alternative lower affinity phosphate transporter like PiT-1 playing a greater role in intestinal phosphorus absorption with disease progression in CKD rats(33). In addition, transporters may play differing roles depending on dietary phosphorus conditions. In a murine study of PiT-2 ablation, expected changes such as increased 1,25D were only observed under the condition of dietary phosphorus restriction(34).

Conclusion

In conclusion, in vivo and clinical studies of phosphorus absorption have augmented our understanding of absorption pathways and their mechanisms, which is beneficial for developing effective strategies to reduce phosphorus burden in CKD. It appears that the paracellular pathway is predominant in vivo at low and high dietary phosphorus intakes, but that inhibiting transcellular phosphate absorption also continues to have potential to provide meaningful phosphorus reduction. Modification of dietary phosphorus intake, phosphate binders, paracellular inhibitors, and transcellular inhibitors would make for a powerful toolkit. In addition, clinical endpoints beyond serum phosphate reduction must continue to be pursued as new therapies are developed.

Key Points.

• Within species, in vivo and clinical phosphorus absorption assessment methods have yielded contradictory results compared with in vitro/ex vivo methods regarding effects of low phosphorus diets, CKD, and 1,25D.

• In vivo methods suggest that paracellular absorption accounts for the majority – up to 2/3 – of total phosphorus absorption across very low to high luminal phosphate concentrations.

• In humans, inhibition of transcellular phosphate absorption, but beyond NaPi-2b inhibition alone, remains a promising approach.

• Inhibition of paracellular absorption offers a new alternative or synergistic target in combination with phosphate binding, dietary restriction, or transcellular inhibition as mechanisms are better understood.