Effects of ferric citrate and intravenous iron sucrose on markers of mineral, bone, and iron homeostasis in a rat model of CKD-MBD

ABSTRACT

Background

Anemia and chronic kidney disease–mineral and bone disorder (CKD-MBD) are common and begin early in CKD. Limited studies have concurrently compared the effects of ferric citrate (FC) versus intravenous (IV) iron on CKD-MBD and iron homeostasis in moderate CKD.

Methods

We tested the effects of 10 weeks of 2% FC versus IV iron sucrose in rats with moderate CKD (Cy/+ male rat) and untreated normal (NL) littermates. Outcomes included a comprehensive assessment of CKD-MBD, iron homeostasis and oxidative stress.

Results

CKD rats had azotemia, elevated phosphorus, parathyroid hormone and fibroblast growth factor-23 (FGF23). Compared with untreated CKD rats, treatment with FC led to lower plasma phosphorus, intact FGF23 and a trend (P = 0.07) toward lower C-terminal FGF23. FC and IV iron equally reduced aorta and heart calcifications to levels similar to NL animals. Compared with NL animals, CKD animals had higher bone turnover, lower trabecular volume and no difference in mineralization; these were unaffected by either iron treatment. Rats treated with IV iron had cortical and bone mechanical properties similar to NL animals. FC increased the transferrin saturation rate compared with untreated CKD and NL rats. Neither iron treatment increased oxidative stress above that of untreated CKD.

Conclusions

Oral FC improved phosphorus homeostasis, some iron-related parameters and the production and cleavage of FGF23. The intermittent effect of low-dose IV iron sucrose on cardiovascular calcification and bone should be further explored in moderate–advanced CKD.

Graphical Abstract

Graphical Abstract.

KEY LEARNING POINTS.

What is already known about this subject?

• Anemia and chronic kidney disease–mineral and bone disorder (CKD-MBD) are associated with higher morbidity and mortality and lower quality of life in patients with CKD.

• In moderate CKD, there is limited concurrent assessment of the effects of intravenous (IV) iron versus ferric citrate.

What this study adds?

• Our work shows that in rats with moderate–advanced CKD, a diet enriched with ferric citrate compared with intermittent IV iron sucrose, with doses lower than previously reported, had similar improvement in cardiovascular calcification.

• Only ferric citrate improved biomarkers of iron status, phosphorus and the coupling of fibroblast growth factor-23 (FGF23) formation and cleavage, with no overall effects of either intervention on bone outcomes.

What impact this may have on practice or policy?

• In moderate-to-advanced CKD, ferric citrate may improve phosphorus homeostasis and lower cardiovascular calcifications, but the effect of low-dose IV iron sucrose on cardiovascular calcification may be due to intermittent effects on phosphorus homeostasis and should be explored further.

• The reduction in both forms of FGF23 with ferric citrate may be due to its dual role on intestinal phosphate binding and more steady-state iron repletion due to constant administration and should be studied in patients with CKD and its relationship with hard clinical outcomes.

INTRODUCTION

Chronic kidney disease (CKD) is a prevalent disease that affects 9.1% of the population worldwide [1]. Common manifestations of CKD include alterations in iron homeostasis and mineral and bone metabolism. Anemia is twice as prevalent in individuals with CKD compared with healthy adults, and its prevalence increases along with the progression of kidney dysfunction [2]. A systematic disorder of mineral and bone metabolism, known as chronic kidney disease–mineral and bone disorder (CKD-MBD), is also highly prevalent [3]. Both anemia and CKD-MBD are associated with poor quality of life [4, 5], increased cardiovascular morbidity [6, 7] and death [8]. Therefore therapies that prevent and treat anemia and CKD-MBD are of interest.

Ferric citrate (FC) is a relatively new therapy that is US Food and Drug Administration–approved as an oral iron replacement product to treat iron-deficiency anemia in nondialysis CKD and as a phosphate binder in dialysis-dependent CKD, therefore it has a plausible beneficial effect for both CKD-MBD and anemia [9]. Once FC is ingested, ferric iron dissociates from citrate in the duodenum and a portion of ferric iron binds to phosphorus, creating an insoluble compound that is excreted in feces [9]. The serum phosphorus–reducing effect has been observed in animal models of CKD [10, 11], patients with nondialysis-dependent CKD [12–15] and individuals on maintenance dialysis [16–18]. Another portion of the ferric iron can be absorbed by being reduced to ferrous iron by the action of duodenal cytochrome-B reductase (DcytB) and then transported into the duodenal enterocyte through divalent metal transporter-1 (DMT1) [9, 19]. As a result, FC has also been shown to improve iron-related parameters in experimental models of CKD [11], patients with non-dialysis-dependent CKD [13, 14] and those undergoing dialysis [20].

While FC is approved for the treatment of anemia in nondialysis CKD, intravenous (IV) iron is another treatment option recommended by the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines for the management of anemia [8]. There are reports of a differential effect of IV versus oral iron on iron homeostasis, where IV iron may be associated with enhanced oxidative stress [21] and increased hepcidin [22], but this remains controversial and varies with the type of IV iron [23]. Similarly, the effects of IV iron on fibroblast growth factor-23 (FGF23) may depend on the IV preparation, some having opposite effects on the intact or active form of FGF23 (iFGF23) or C-terminal FGF23 (cFGF23) [24]. Concurrent assessment of markers of CKD-MBD and anemia utilizing oral versus IV therapies is limited. Therefore our objective was to compare the effects of 10 weeks of FC versus IV iron in the form of iron sucrose on a comprehensive assessment of CKD-MBD, iron homeostasis and oxidative stress in rats with moderate and progressive CKD. We hypothesized that FC would have a lower effect on oxidative stress and end-organ manifestations of CKD-MBD compared with IV iron sucrose.

MATERIALS AND METHODS

Experimental design

Cy/+_IU rats (CKD hereafter) are characterized by an autosomal dominant progressive cystic kidney disease that is not orthologous to human Autosomal Dominant Polycystic Kidney Disease (ADPKD) [25–27]. Male CKD rats progressively develop kidney dysfunction, reaching kidney failure by 30–35 weeks of age with all three components of CKD-MBD manifesting from 28 weeks on [25]. Female rats do not develop azotemia, even after ovariectomy [28, 29] and thus only males were used for these studies.

Rats were bred in-house and weaned at 3 weeks of age; at 10 weeks they were single-housed with free access to food (autoclaved Teklad 2018SX; Envigo, Indianapolis, IN, USA) and tap water. Rats were phenotyped by plasma blood urea nitrogen (BUN) at 10 weeks of age [25, 30]. At 17 weeks of age, CKD and unaffected normal (NL) littermates were placed on a nonautoclaved diet of 18% casein-based protein, 0.7% phosphorus (85% as phosphate additives) and 0.7% calcium (Teklad TD.04539; Envigo), which leads to a more consistent and reproducible CKD-MBD phenotype [25]. CKD rats were randomly assigned into three groups (n = 11 − 14/group): CKD control, CKD + 2% FC (CKD + FC) and CKD + IV iron sucrose, while NL animals remained untreated. All procedures were reviewed and approved by the Indiana University School of Medicine Institutional Animal Care and Use Committee and adhere to the Guide for the Ethical Treatment of Animals to minimize pain and suffering.

CKD rats assigned to the CKD + FC group received a 2% FC diet with the same casein diet background (Teklad TD.180 350; Envigo). CKD rats were placed on the FC diet at 18 weeks and consumed the diet for 10 weeks until euthanasia at 28 weeks of age. Rats assigned to the CKD + IV iron group were injected via tail vein with 1 mg/kg iron sucrose dose biweekly [31] for 10 weeks until sacrifice (total of five injections, last dose 3 days before euthanasia). Initial dosing of weekly injections led to early mortality in three rats consecutively assigned to weekly IV iron and thus the dosing interval was changed for all remaining rats in the group. These rats with early mortality were not included in any outcome measures.

Study endpoints

Animals were anesthetized at 28 weeks of age with isoflurane, blood was collected via cardiac puncture and euthanasia was completed by exsanguination and bilateral pneumothorax. Heart, aortic arch, kidney, tibia, a portion of the liver and femurs were collected, weighed as appropriate and stored for analysis. Duodenum was dissected 1 cm proximal to the pyloric junction until the suspensory muscle of the duodenum contents were flushed using a gavage needle with 0.9% saline solution, cut transversally and the mucosa was scraped and flash-frozen in liquid nitrogen. Samples were stored at − 80°C until analyses.

Blood was collected in lithium heparin-coated blood collection tubes (BD 367884; Becton Dickinson, Franklin Lakes, NJ, USA). Plasma was analyzed using colorimetric assays for BUN (DIUR-100; BioAssay Systems, Haywood, CA, USA), calcium and phosphorus (Pointe Scientific, Canton, MI, USA). Plasma intact parathyroid hormone (PTH), intact FGF23 (iFGF23) and C-terminal FGF23 (cFGF23, which captures both the iFGF23 and cFGF23, measuring total FGF23) were determined by enzyme-linked immunosorbent assay (ELISA; Quidel, San Diego, CA, USA). Serum creatinine was analyzed using a colorimetric assay (BioAssay Systems). Serum levels of the oxidative stress marker 8-hydroxy-2′-deoxyguanosine (8-OHdG) were measured using an ELISA kit (Enzo Life Sciences, Farmingdale, NY, USA).

Additionally, blood was obtained at 18, 23 and 28 weeks of age via tail bleed in capillary tubes for hematocrit quantification. For hematocrit, capillary tubes were centrifuged at 11 000 rpm for 10 min in a microhematocrit centrifuge (M24; LW Scientific; Lawrenceville, GA, USA). Hemoglobin was assessed via an Hb 201 system (HemoCue, Ängelholm, Sweden). Total iron, unsaturated iron-binding capacity (UIBC), total iron-binding capacity (TIBC) and transferrin saturation rate (TSAT) was performed with an AU 5822 analyzer (Beckman Coulter, Brea, CA, USA).

Half of the aortic arch and 100–200 mg of cardiac tissue were washed in sterile 0.9% saline solution and then incubated in 0.6 N HCl for 48 h (4 µL/mg of aortic arch tissue and 1.25 µ/mg of heart). The supernatant was analyzed for calcium using the o-Cresolphthalein complex 1 method (Calcium kit; Pointe Scientific) and normalized by tissue dry weight [32].

For gene expression, total RNA was isolated using the miRNeasy Mini Kit (Qiagen, Venlo, The Netherlands). All real-time polymerase chain reactions (PCRs) were performed using TaqMan gene expression assays with ViiA 7 Real-Time PCR Systems (Applied Biosystems, Waktham, MA, USA): dmt1 (Rn01533109_m1), fpn1 (Rn00591187_m1), hepcidin (Rn00584987_m1), GAPDH (Rn01775763_g1), cyp27b1 (Rn01647147_g1) and β-actin (Rn00667869_m1). The ΔΔCT method was used to analyze the relative change in gene expression normalized to β-actin (intestine and kidney) and GAPDH (liver).

Bone assessments

For microarchitecture, proximal tibias were scanned using microcomputed tomography (SkyScan 1172; Bruker, Billerica, MA, USA) at 12-µm resolution using published methods [30]. Trabecular parameters were obtained from a 1-mm region of interest located ∼0.5 mm distal to the proximal growth plate. Trabecular bone volume (BV/TV, %), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp) and trabecular number (Tb.N) were measured following standard recommendations [33]. Cortical porosity was measured from an average of five slices ∼4 mm below the trabecular region of interest. Cortical porosity was determined by assessing the void area between the periosteal and endosteal surfaces, presented as a percentage of overall cortical volume.

All animals received an injection of calcein (30 mg/kg) 14 and 4 days prior to euthanasia. For dynamic bone histomorphometry, undemineralized proximal tibia was fixed in neutral buffered formalin then subjected to serial dehydration and embedded in methyl methacrylate (Sigma Aldrich, St. Louis, MO, USA). Serial frontal sections were cut 4-µm thick and left unstained for analysis of fluorochrome calcein labels. Histomorphometric analyses were performed using BIOQUANT Image Analysis; BIOQUANT, Nashville, TN, USA. We utilized a standard region of interest of trabecular bone excluding primary spongiosa and endocortical surfaces. Total bone surface (BS), single-labeled surface (sLS), double-labeled surface (dLS) and interlabel distances were measured at magnification ×20. Mineralized surface to bone surface [MS/BS; (dLS + sLS/2)/BS*100], mineral apposition rate (MAR; average interlabel distance/10 days) and bone formation rate [BFR/BS; (MS/BS*MAR)*3.65] were calculated. An additional section was von Kossa stained with McNeal tetrachrome counterstain for assessment of osteoclast-covered trabecular surfaces (OcS/BS). All nomenclature for histomorphometry followed standard usage [34].

Additional 4-µm sections of the methyl methacrylate–embedded tibia were deplasticized with acetone and then rehydrated. The cancellous bone sections were subsequently stained with Perl's stain for iron and nuclear fast red counterstain according to established protocols. Images were collected on a DM 3000 light microscope (Leica Microsystems, Deerfield, IL, USA). Slides were semiquantitatively evaluated by a single researcher blinded to the treatment group assignments.

For mechanics, at the time of tissue collection, femora were frozen in phosphate-buffered saline–soaked gauze. Frozen femoral midshafts were scanned at 18 µm (SkyScan 1176) to obtain geometric data for the normalization of mechanical properties. Prior to mechanical testing, samples were thawed, hydrated in saline and tested in three-point bending (TestResources, Shakopee, MN, USA) with the posterior and anterior surfaces in compression and tension, respectively. Geometry values (polar moment of inertia and anteroposterior diameter) were obtained from CTAN software (Bruker) and were used to generate mechanical data in a custom MATLAB script (MathWorks, Natick, MA, USA) [30]. All data are presented using standard nomenclature.

Statistical analyses

Each rat represented an experimental unit. Outliers were assessed and removed before analysis by the robust regression and outlier removal method based on a false discovery rate (Q) of 1% [35]. All outcomes were assessed for normality and variance using Brown–Forsythe's test and log transformed before analyses as appropriate. Results are reported as mean ± standard deviation (SD) unless otherwise noted. One-way analysis of variance (ANOVA) was performed for all outcomes utilizing least square means with Tukey/Sidak's post hoc comparisons. When 18- and 28-week biochemical parameters were assessed, a repeated-measures ANOVA was performed to assess the mean effect of the group (NL, CKD, CKD + FC and CKD + IV iron), time (18 or 28 weeks) and its interaction. Statistical significance was set at α < 0.05. All statistical analyses were performed in GraphPad Prism 8.2.0 (GraphPad Software, San Diego, CA, USA).

RESULTS

Iron treatment does not affect kidney function

At 18 weeks of age before the beginning of the iron interventions, BUN was higher in the CKD rats (NL 19.5 ± 1.60 versus CKD 41.64 ± 6.29 mg/dL; P < 0.0001). At 28 weeks (Table 1), BUN, creatinine and kidney weight were increased in all CKD groups, regardless of treatment, and there was no effect of either iron treatment.

Table 1.

Plasma biochemistries at 28 weeks

Biochemistries	NL (n = 14)	CKD (n = 12)	CKD + IV iron (n = 13)	CKD + FC (n = 12)	ANOVA P-value
BUN (mg/dL)	19.96 ± 3.11	48.37 ± 6.77a	48.97 ± 7.74a	46.13 ± 5.03a	<0.0001
Creatinine (mg/dL)	0.33 ± 0.06	0.86 ± 0.31a	0.85 ± 0.19a	0.98 ± 0.34a	<0.0001
Calcium (mg/dL)	7.18 ± 1.12	7.80 ± 1.22	7.49 ± 1.34	7.01 ± 1.48	0.462
Phosphorus (mg/dL)	5.09 ± 0.57	9.14 ± 4.34a	7.42 ± 1.53	5.77 ± 1.62b	0.001
PTH (pg/mL)	145.6 ± 39.48	1870 ± 1637a	996.0 ± 800.5a	1868 ± 1921a	<0.0001
iFGF23 (pg/mL)	458.6 ± 267.1	2215 ± 1457a	1127 ± 879	519.7 ± 455.9b	0.0004
cFGF23 (pg/mL)	470.8 ± 145.1	2814 ± 1248a	3422 ± 3392a	1361 ± 496.9a	<0.0001
iFGF23/cFGF23	1.07 ± 0.73	0.96 ± 0.56	0.57 ± 0.39	0.42 ± 0.44a,b	0.002

^aTukey's post hoc different than NL rats, P < 0.05. ^bTukey's post hoc different than CKD, P < 0.05.

CKD-MBD

Iron treatment reduces plasma phosphorus and intact FGF23. Biochemical parameters at 28 weeks are shown in Table 1 and Figure 1. Plasma phosphorus was higher in the untreated CKD rats compared with NL animals (Figure 1A). Both iron treatments had plasma phosphorus concentrations comparable with NL rats, but only the CKD + FC rats had lower levels compared with untreated CKD rats (Figure 1A). When assessing the phosphorus over time at 18, 23 and 28 weeks of age, we found a group × time interaction (P interaction = 0.02) driven primarily by the effect at 28 weeks (Figure 1B). However, phosphorus over time showed divergent trajectories depending on the treatment, where the untreated CKD rats increased at 28 weeks, CKD + FC rats showed a downward trajectory and CKD + IV iron had steady phosphorus concentrations.

FIGURE 1:

FC and IV iron improved serum biochemistries and hormones related to CKD-MBD. Except for panel B, at 28 weeks of age plasma was assessed for (A) phosphorus, (B) phosphorus at 18, 23 and 28 weeks of age, (C) iFGF23, (D) C-terminal FGF23 and (E) PTH and calcification in the (F) aorta and (G) heart. Compared with untreated CKD rats, FC-treated rats had lower phosphorus, iFGF23, aorta and heart calcification and tended to have lower cFGF23, while IV iron–treated rats had fewer aorta and heart calcifications. Compared with normal littermates, IV iron and FC-treated rats had similar levels of plasma phosphorus, iFGF23 and aorta and heart calcification and higher cFGF23 and PTH. There was a group × time interaction on the phosphorus primarily driven by the change at 28 weeks and there were divergent trajectories of the circulating phosphorus in the groups, including a downward trajectory in the CKD + FC group and maintenance overtime in the CKD + IV iron. For the analyses, iFGF23 had one outlier removed in the CKD group; aorta calcification CKD + FC had two outliers removed and heart calcification in the CKD group had one outlier removed.

Similarly, both iron treatments had comparable levels of iFGF23 compared with NL, but only CKD + FC rats had levels lower than untreated CKD rats (Figure 1C). Conversely, the cFGF23 concentrations were higher in the CKD and CKD + IV iron group versus NL rats, with a trend toward lower cFGF23 concentrations in CKD + FC versus untreated CKD rats (P = 0.07; Figure 1D). Calcium concentrations did not differ between the groups. Finally, intact PTH concentrations were higher in all CKD rats compared with NL rats and there was no effect of either iron treatment (Figure 1E).

FGF23 and PTH have opposite effects on vitamin D activation, as FGF23 inhibits cyp27b1 and PTH increases its expression. Since we were unable to measure 1,25-dihydroxyvitamin D due to the needed volume, we measured the renal gene expression of cyp27b1 as an indirect measure. Despite the higher levels of FGF23, untreated CKD rats had similar cyp27b1 expression compared with NL rats. However, compared with untreated CKD rats, both iron treatments resulted in higher renal expression of cyp27b1 and the expression was not different between iron treatments.

Iron treatment does not improve renal osteodystrophy. Bone parameters are shown in Table 2. For dynamic histomorphometry, all CKD rats had a higher mineral apposition and bone formation rate compared with NL, consistent with secondary hyperparathyroidism with no effect of iron treatment. Trabecular bone volume was lower in untreated and iron-treated CKD rats, with concurrent reductions in the trabecular number and greater trabecular separation. Trabecular thickness was higher in the CKD + IV iron rats compared with NL and CKD rats, but it was not different from the CKD + FC rats. Osteoclast-covered trabecular surfaces were higher in all CKD rats and iron treatment had no effect. While we have observed cortical porosity in our animals at 30 and 35 weeks of age [36], cortical porosity was not different at 28 weeks of age between the NL rats and all CKD groups.

Table 2.

Bone parameters in normal and CKD rats untreated or treated with FC or IV iron sucrose

Parameters	NL (n = 13)	CKD (n = 11)	CKD + IV iron (n = 10)	CKD + FC (n = 11)	ANOVA P-value
Dynamic histomorphometry
BFR/BS (µm3/µm2/year)	78.75 ± 32.71	214.30 ± 43.46a	196.7 ± 73.88a	209.50 ± 38.50a	<0.0001
MS/BS (%)	22.3 ± 4.05	26.66 ± 5.26	27.47 ± 4.92	27.26 ± 5.83	0.046
MAR (µm/d)	0.95 ± 0.26	2.22 ± 0.31a	1.98 ± 0.68a	2.15 ± 0.41a	<0.0001
Bone resorption
OcS/BS (%)	1.86 ± 0.62	7.34 ± 1.84a	6.15 ± 0.12a	6.66 ± 1.48a	<0.0001
Structure proximal tibia
BV/TV (%)	6.58 ± 2.23	3.95 ± 1.63a	3.83 ± 1.47a	4.82 ± 2.36	0.0035
Tb.Th (mm)	0.08 ± 0.005	0.08 ± 0.007	0.09 ± 0.01a,b	0.08 ± 0.007	0.0008
Tb.sp (mm)	0.55 ± 0.09	0.71 ± 0.05a	0.73 ± 0.06a	0.69 ± 0.08a	<0.0001
Tb.n (n/mm)	0.86 ± 0.27	0.50 ± 0.19a	0.42 ± 0.13a	0.58 ± 0.26a	<0.0001
Cortical area (mm2)	6.97 ± 0.32	6.25 ± 0.81a	6.88 ± 0.36b	6.31 ± 0.53a	0.002
Cortical thickness (mm)	0.52 ± 0.02	0.42 ± 0.11a	0.50 ± 0.05	0.48 ± 0.06	0.011
Cortical porosity (%)	0.71 ± 0.39	0.95 ± 0.89	0.77 ± 0.45	0.58 ± 0.16	0.614
Midshaft femur
Cortical area (mm2)	8.24 ± 0.30	7.23 ± 0.59a	7.78 ± 0.50b	6.91 ± 0.49a,c	<0.0001
Cortical thickness (mm)	0.72 ± 0.02	0.59 ± 0.12a	0.68 ± 0.05	0.57 ± 0.12a,c	0.0002
Structural properties
Yield force (N)	109.8 ± 6.73	91.77 ± 13.28a	109.2 ± 11.75b	91.04 ± 12.34a,c	0.0001
Ultimate force (N)	158.6 ± 7.86	127.1 ± 24.11a	147.9 ± 15.55b	126.1 ± 20.00a,c	<0.0001
Displacement to yield (µm)	344.9 ± 29.65	335.8 ± 16.18	362.3 ± 25.48	337.1 ± 37.37	0.1437
Postyield displacement (µm)	550.7 ± 86.42	378.8 ± 142.1a	372.5 ± 92.34a	341.5 ± 129.6a	0.0002
Total displacement (µm)	895.6 ± 75.01	714.6 ± 144.9a	734.7 ± 86.59a	678.6 ± 145.5a	0.0009
Stiffness (N/mm)	389.2 ± 22.29	323.7 ± 45.46a	359.8 ± 29.70a,b	320.6 ± 26.79a,c	<0.0001
Work to yield (mJ)	19.27 ± 1.98	16.16 ± 2.71a	20.40 ± 2.72	16.09 ± 3.18a,c	0.0005
Postyield work (mJ)	77.99 ± 12.02	45.43 ± 21.15a	49.88 ± 16.19a	39.88 ± 20.81a	<0.0001
Total work (mJ)	97.26 ± 11.15	61.59 ± 23.10a	70.29 ± 16.96a	55.97 ± 22.67a	<0.0001
Material properties
Yield stress (MPa)	38.57 ± 4.49	37.13 ± 3.83	40.02 ± 4.83	39.64 ± 4.80	0.4550
Ultimate stress (MPa)	55.54 ± 3.99	51.16 ± 6.94	54.20 ± 6.08	54.69 ± 6.38	0.3189
Strain to yield (me)	23 048 ± 2707	21 497 ± 1121	23 196 ± 1731	21 186 ± 2279	0.0558
Total strain (me)	59 913 ± 7456	45 699 ± 9110a	47 105 ± 6147a	42 650 ± 9051a	<0.0001
Modulus (GPa)	2.045 ± 0.15	2.046 ± 0.21	2.059 ± 0.18	2.22 ± 0.18	0.0687
Resilience (MPa)	0.45 ± 0.08	0.42 ± 0.05	0.48 ± 0.08	0.44 ± 0.08	0.2979
Toughness (MPa)	2.28 ± 0.36	1.55 ± 0.46a	1.64 ± 0.35a	1.50 ± 0.47a	<0.0001

^aTukey's post hoc different than NL rats, P < 0.05. ^bTukey's post hoc different than CKD, P < 0.05. ^cTukey's post hoc different than IV iron, P < 0.05.

Midshaft femur cortical area and thickness (same site as mechanics) were similar between NL rats and CKD + IV iron but were lower in the CKD and CKD + FC rats. The mechanical properties of the midshaft femur are shown in Table 2. For structural properties, yield force, ultimate force and work yield were lower in the CKD and CKD + FC rats compared with NL, but CKD + IV iron rats were not different from NL. Post-yield displacement, total displacement, stiffness, post-yield work and total work were not different due to iron treatment in CKD groups. For the bone material properties, total strain and toughness were lower in all CKD rats versus NL with no impact of iron treatment.

Iron treatment reduces vascular calcifications. At 28 weeks there was greater aorta and heart calcification in the untreated CKD rats compared with their NL littermates. However, CKD + FC and CKD + IV iron rats had fewer aortic and heart calcifications than the untreated CKD rats and were not different from the NL rats (Figure 1F, G).

Iron homeostasis and oxidative stress

Ferric citrate improves iron parameters without a further increase in systemic oxidative stress. At 18 weeks, before the interventions started, hematocrit was already lower in all CKD rats compared with NL rats (NL 51.97 ± 1.57%; versus CKD 45.05 ± 1.90%; P < 0.0001). At 28 weeks (Table 3), untreated and treated CKD rats had lower hematocrit than NL rats. At 28 weeks, hemoglobin and hematocrit remained lower than NL in CKD and iron-treated rats.

Table 3.

Iron and oxidative stress biochemistries at 28 weeks

Biochemsitries	NL (n = 14)	CKD (n = 12)	CKD + IV iron (n = 13)	CKD + FC (n = 12)	ANOVA P-value
Hemoglobin (mg/dL)	15.67 ± 0.58	13.00 ± 0.84a	13.06 ± 0.53a	13.08 ± 0.55a	0.0002
Hematocrit (%)	52.08 ± 2.31	43.06 ± 1.27a	43.45 ± 3.15a	43.17 ± 2.32a	<0.0001
Total iron (µg/dL)	325.3 ± 45.32	287.5 ± 88.94	217.1 ± 57.85a	266.7 ± 50.61	0.017
UIBC (µg/dL)	427.5 ± 35.85	325.3 ± 28.87a	305.2 ± 53.22a	202.6 ± 41.89a,b,c	<0.0001
TIBC (µg/dL)	752.8 ± 46.47	612.8 ± 76.75a	499.7 ± 118.1a,b	470.4 ± 83.58a,b	<0.0001
TSAT (%)	43.17 ± 4.49	46.08 ± 8.78	42.33 ± 4.12	56.00 ± 5.29a,b,c	<0.0001
8OHdG (ng/mL)	18.59 ± 1.95	26.49 ± 3.68a	26.61 ± 2.79a	25.38 ± 2.22a	<0.0001

^aTukey's post hoc different than NL rats, P < 0.05. ^bTukey's post hoc different than CKD, P < 0.05. ^cTukey's post hoc different than IV iron, P < 0.05.

After 10 weeks of treatment and at 28 weeks of age, serum total iron was significantly lower than the NL only in the CKD + IV iron rats (Table 3, Figure 2A). UIBC and TIBC were lower in all the CKD animals compared with NL rats (Figure 2B). Compared with untreated CKD rats, TIBC was lower in both iron-treated groups and UIBC was lower only in the CKD + FC rats (Figure 2B). Conversely, TSAT was higher only in the CKD + FC rats compared with NL and untreated CKD rats (Figure 2C).

FIGURE 2:

FC had a greater impact on iron homeostasis parameters compared with IV iron. At 28 weeks of age, (A) total serum, (B) TIBC, (C) TSAT, (D) duodenal dmt1 mRNA expression, (E) hepatic total iron and (F) femur bone marrow iron staining were evaluated. Compared with untreated CKD rats, FC-treated rats had similar levels of total serum iron, lower TIBC and higher TSAT, while IV iron sucrose–treated rats had lower TIBC and tended to have lower total iron but similar TSAT. Compared with untreated CKD rats, the mRNA expression of the main duodenal transporter of nonheme iron, dmt1, was lower in the FC and IV iron sucrose–treated rats, while hepatic iron was numerically higher in both iron-treated rats. Finally, femur bone marrow iron staining was consistently higher in the FC-treated rats. For the analyses, duodenal dmt1 had one outlier removed in the NL group and one in the CKD + IV iron group.

To assess the effects of the iron supplementation on intestinal iron homeostasis, we measured the messenger RNA (mRNA) expression of nonheme iron transporters in the duodenal mucosa and hepcidin in the liver (Table 4 and Figure 2D). Dmt1 was not different between the CKD groups and NL but was lower in the CKD + FC and CKD + IV iron rats compared with the untreated CKD rats (Figure 2D). Conversely, fpn1 was not different between NL and the untreated and iron-treated CKD rats. Liver hepcidin mRNA expression did not differ between NL and all CKD rats, although it was numerically the highest in the CKD + FC.

Table 4.

Intestinal, hepatic and renal gene expression

Genes	NL	CKD	CKD + IV iron	CKD + FC	ANOVA P-value
Duodenum dmt1	0.84 ± 0.52	1.01 ± 0.38	0.43 ± 0.33b	0.59 ± 0.26b	0.005
Duodenum fpn1	2.63 ± 2.54	1.41 ± 0.47	1.76 ± 1.13	1.89 ± 0.89	0.632
Duodenum Dcytb	0.87 ± 0.85	3.27 ± 4.05	1.61 ± 2.72	0.53 ± 0.31	0.07
Liver hepcidin	0.69 ± 0.21	0.72 ± 0.33	0.82 ± 0.34	0.90 ± 0.26	0.252
Kidney cyp27b1	0.99 ± 0.30	0.89 ± 0.61	1.57 ± 0.56b	1.82 ± 0.37a,b	0.0001

Dmt1, divalent metal transporter-1; fpn1, ferroportin-1; Dcytb, duodenal cytochrome B-reductase; cyp27b1, 1α-hydroxylase. ^aTukey's post hoc different than NL rats, P < 0.05. ^bTukey's post hoc different than CKD, P < 0.05.

Since duodenal expression of dmt1 was lower in the iron-treated rats compared with untreated CKD rats and total iron was modestly altered in the iron-treated rats, we explored tissue iron deposition in the liver quantitatively and bone marrow cells in the tibia semiquantitatively (Figure 2E–F). Compared with NL rats, hepatic total iron was higher in the CKD + FC rats (P = 0.03) and trended higher in the CKD + IV rats (P = 0.06; Figure 2E). Similarly, using Perls Prussian blue staining in the tibia, we found that 70% of the CKD + FC rats had a score of 2+ or 3+ staining in the bone marrow, compared with 45% in the CKD + IV iron rats, 36% in the untreated CKD rats and 29% in the NL rats (P = 0.07; Figure 2F). No significant iron staining was observed on mineralized bone surfaces, indicating no evidence of excess tissue deposition.

Finally, to assess the effects of the iron supplementation route on oxidative stress, we measured the circulating levels of 8-hydroxy-2-deoxyguanosine (8OHdG), a marker of DNA oxidation (Table 3). Untreated and iron-treated CKD rats had a higher plasmatic concentration of 8OHdG than the NL rats, but there was no effect of iron treatment.

DISCUSSION

CKD-MBD and anemia are highly prevalent in CKD and recent data demonstrate a connection between iron, intestinal phosphorus binding and FGF23 [37]. In our study we compared the effects of 10 weeks of FC and IV iron sucrose on the three components of CKD-MBD, iron homeostasis and oxidative stress in a rat model of CKD-MBD, the Cy/+ rat, at an age where the rats have moderate–advanced CKD (∼50–25% kidney function of NL). The Cy/+ rat developed slowly progressive CKD and complications of kidney dysfunction, such as CKD-MBD [25], left ventricular hypertrophy and cardiac arrhythmias [38], diastolic dysfunction (unpublished data), oxidative stress [39] and muscle wasting at 35 weeks of age [40, 41]. With moderate–advanced CKD at 28 weeks, we see that the CKD rats developed abnormalities within the three components of CKD-MBD, left ventricular hypertrophy, higher oxidative stress and anemia, with some abnormalities in iron homeostasis. Compared with untreated CKD rats, CKD rats treated with FC had lower circulating phosphorus and iFGF23, leading to less cardiovascular calcification, but with no impact on bone. Meanwhile, IV iron did not lead to the same marked effects as FC, as well as fewer cardiovascular calcifications. Of note, however, was that the doses were lower than those previously published: 4–5% FC and 1 mg/kg every 4 days with IV iron sucrose [10, 11, 31]. Further, iron treatments did not impact kidney function measured by plasma BUN, creatinine and kidney weight and both iron supplementation routes did not further increase systemic oxidative stress. Since the changes in iron status biomarkers were limited, the unique FC responses in this animal model appear to be driven by a strong lowering of phosphorus, likely due to lower intestinal phosphorus absorption, inducing lower FGF23 and increased cleavage shown by a lower iFGF23:cFGF23 ratio. An alternative or additive hypothesis for the lower ratio is that the lower phosphorus with FC may decrease Galnt3, the enzyme responsible for the O-glycosylation of FGF23 that prevents the cleavage of FGF23 [42], as in vitro studies show that high phosphate upregulates Galnt3 gene and protein expression [43].

In individuals with nondialysis CKD, FC lowers iFGF23 and cFGF23, but these changes may depend on higher phosphorus and lower TSAT levels at baseline [12, 13, 44]. A study in the Col4a3^KO mouse model of CKD showed that 5% FC lowered phosphorus and both iFGF23 and cFGF23 levels in CKD [11]. In our study, a diet with 2% FC also lowered iFGF23 and tended to lower cFGF23 (P = 0.07). In contrast, we found that IV iron sucrose numerically lowered phosphorus and iFGF23, albeit nonsignificantly, whereas cFGF23 remained similar to that in untreated CKD rats. Since both iron treatments did not affect oxidative stress (at least by 8OHDG), these differences remain unexplained. One hypothesis may be that the IV iron dose may not have been enough to restore iron levels; however, hepatic iron concentrations were similar between IV iron sucrose and FC. Another hypothesis may be the differences in the pharmacodynamics of iron levels between the two preparations.

In addition to the improvements in CKD-MBD biochemical parameters, iron-treated CKD rats had fewer aortic and heart calcifications compared with untreated CKD rats and similar levels to NL rats. Hyperphosphatemia is a strong inducer of vascular calcification [45, 46], therefore therapies that reduce phosphorus may limit vascular calcification formation and propagation. In vitro, Ciceri et al. [47] showed that in vascular smooth muscle cells incubated with a high-phosphate media, iron citrate limited calcification by ∼50% by reducing cell apoptosis and increasing autophagy. Iida et al. [48] showed that in rats fed a 0.75% adenine diet to induce CKD, adding 1% or 3% FC for 28 days reduced the calcium deposits in the aorta. With the results of our study, it is intriguing to suggest that iron supplementation may prevent vascular calcifications and limit continued increases once it has already occurred. The lower aortic and heart calcification in the IV iron–treated rats suggest that an intermittent effect of lowering phosphorus may also be reducing soft tissue calcification. There are reports of hypophosphatemia secondary to IV iron, and this appears to be due to an increase of iFGF23 and FGF23-induced phosphaturia [49]. In a network meta-analysis comparing the risk of hypophosphatemia with IV iron formulations, ferric carboxymaltose led to a higher incidence of hypophosphatemia compared with other treatments, including iron sucrose [50]. However, hypophosphatemia with IV iron sucrose is variable, with a range of 0–40% [51]. We hypothesize that since the treatment with IV iron sucrose was every other week, there may have been an intermittent increase in iFGF23, an increase in phosphaturia and a concomitant lowering of circulating phosphorus. Unfortunately, we did not assess urinary phosphorus excretion. However, when we evaluated the treatment effects over time there were divergent trajectories of circulating phosphorus (Figure 1B). Intriguingly, with IV iron there was a steady phosphorus concentration over time, which may have prevented the development of cardiovascular calcification.

The Cy/+ rat, even at moderate CKD stages as presented here, has high bone turnover consistent with secondary hyperparathyroidism bone disease [25, 36, 52]. However, we did not observe the typical severity of cortical porosity as we typically see in more advanced CKD [36]. Additionally, CKD impaired bone mechanical properties despite the absence of porosity, supporting the idea of material-level differences in matrix properties. Of interest, we observed that some of the mechanical parameters, particularly yield force, ultimate force and stiffness, were higher in the CKD + IV iron rats compared with untreated CKD. Cortical area and thickness were also similar between NL and CKD + IV iron rats and there was a numerical decrease in PTH, albeit not significant, which may help explain the mechanical properties as these are related to bone mineralization. Additionally, iron is involved in collagen synthesis, the most abundant protein in bone, where it serves as the reducing agent in the hydroxylation of pro-collagen [53]. It is therefore plausible that the IV iron somehow improved collagen structure. However, because the CKD + FC rats had greater improvement in iron-related parameters, but no improvement in bone mechanical properties compared with untreated CKD, it seems unlikely that iron alone led to these improvements in bone parameters. Alternatively, weight differences may be a contributing factor in these differences [54]. At 28 weeks of age, NL and CKD + IV iron rats tended to have an ∼7% higher body weight compared with the untreated CKD rats and the CKD + FC rats but consumed similar amounts of food (data not shown). In our animal model, weight loss toward the end seemed to correlate with faster progression of the disease, but also a lower starting weight at baseline (18 weeks of age). At 18 weeks, the CKD + FC rats had a significantly lower weight compared with the NL and CKD + IV iron rats and a trend towards lower weight compared with the untreated CKD rats. But when we looked into the percent weight change from 18 to 28 weeks of age, both iron treatments gained similar weight (Supplementary data, Figure S1). Therefore we believe the impact of weight differences at 28 weeks on bone may be limited.

Intestinal absorption of nonheme iron is an important component of systemic iron homeostasis [55]. Oral iron supplements and replacement products, such as FC, may lead to downregulation of the duodenal apical dmt1 transporter [56]. In our study, while mRNA expression of dmt1 was lower in the iron-treated CKD groups, TSAT and hepatic iron were highest in the FC-treated CKD rats. Furthermore, while hepatic hepcidin mRNA numerically increased in the FC-treated rats, this was not statistically significant. Transcellular iron transport is the primary route of absorption and occurs mainly in the duodenum [55]. Recently, however, Vaziri et al. [57] showed that in 5/6 nephrectomized rats, a diet containing 4% FC led to increased deposition of iron in the epithelial and subepithelial walls in the colon coupled with lower expression of the tight junction proteins JAM-1 and occludin, suggesting the possibility of increased paracellular iron transport. Since the CKD + FC rats had improved TSAT but low dmt1, we cannot rule out that colonic absorption of iron occurred, and this needs to be further explored.

Our study had some limitations. First, our results are limited to male rats with moderate–advanced CKD using a single model of progressive CKD. Second, it is possible that a higher frequency or doses of IV iron may have further improved iron homeostasis, but given the high mortality observed in the first three rats treated weekly, it is also possible that other adverse events would have occurred. However, to our knowledge, this is the first study that assesses the comprehensive effects of lower doses of FC and IV iron sucrose on CKD-MBD and anemia in moderate–advanced CKD.

Overall, our results indicate that in moderate–advanced CKD, FC and IV iron sucrose, in doses lower than previously published, may confer similar benefits for some measures of CKD-MBD and iron homeostasis without further increasing oxidative stress. An additional benefit of FC is the improved coupling of formation and degradation of FGF23 due to its dual role in intestinal phosphate binding and more steady-state iron repletion due to constant administration.

FUNDING

The study was funded by Akebia Therapeutics (grant to S.M.M.), but Akebia employees were not involved in the study procedures, data collection, statistical analyses or manuscript preparation. A.B. was supported by the National Institutes of Health (NIH; T32 DK120524-02 and AR065971). C.E.M. was supported by NIH F32DK122731. N.X.C. was supported by Veteran's Administration (VA) Merit I01BX001471. E.L.C. was supported by an award from Indiana University School of Medicine. S.M.M. was funded by NIH RO1DK110871, P30AR072581, UL1TR002529 and VA Merit I01 BX001471. M.A. was funded by VA BX003025, NIH RO1DK110871 and DK119266 and Amgen.

CONFLICT OF INTEREST STATEMENT

The results presented in this article have not been published previously in whole or part, except as a poster at Kidney Week 2020 (PO0317). A.B. has received honoraria from Amgen. M.A. has received grant support from Amgen. S.M.M. has received grant support from Chugai and Keryx (for current study) and honoraria from Amgen, Sanifit and Ardeylx.

DATA AVAILABILITY STATEMENT

The data underlying this article will be shared upon reasonable request to the corresponding author.

AUTHORS’ CONTRIBUTIONS

A.B, N.X.C and S.M.M designed the study. A.B., C.E.M., E.A.S., C.V., E.L.C., S.S., K.O. and K.G.A. carried out experiments. A.B and C.E.M. analyzed the data. A.B. made the figures. A.B. and S.M.M. drafted the article. C.E.M., N.X.C., E.A.S., C.V., E.L.C., S.S., K.O., K.G.A. and M.R.A. revised the article. All authors approved the final version of the article.