Major Determinants of BMP-2 Serum Levels in Hemodialysis Patients

Bone morphogenetic proteins (BMPs) are members of the transforming growth factor-β superfamily, and more than 20 different isoforms have been identified in mammals and Drosophila.¹ BMP-2, one of that isoforms, is a multifunctional regulator of cell growth and differentiation and appears to play an important role in nephropathogenesis, by inhibiting epithelial growth factor-induced biological effects, such as DNA synthesis in kidney glomerular mesangial cells and cell proliferation of collecting tubule cells, and by stimulating apoptosis. Moreover, it was reported that abnormal preurine calcium oxalate triggers the expression of BMP-2 in long Henle’s loop cells, inducing differentiation of these cells toward the osteogenic lineage.²

The 25-amino acid peptide hepcidin has emerged as a central regulator of iron homeostasis, being upregulated by pro-inflammatory cytokines, such as interleukin- (IL-) 1α, IL1-β, and IL-6, and by high levels of iron; it is downregulated by hypoxia and anemia.³ Recently, it was reported that BMP-2, BMP-4, and BMP-9 are also involved in hepcidin expression.⁴

As hemodialysis (HD) patients present high serum levels of hepcidin and inflammatory markers, as well as a disturbed iron metabolism,⁵ we wondered how BMP-2 and all these factors interact in HD patients.

In chronic kidney disease patients under HD and recombinant human erythropoietin (rhEPO) therapies, we evaluated the serum levels of BMP-2, hematological data, inflammatory markers, and iron metabolism; the values of rhEPO doses used to treat patients and the type of vascular access used for HD procedure were also considered in data analysis.

Forty-nine patients (28 males and 21 females; 67.4 ± 15.0 years old) were enrolled in the study. A group of 24 healthy volunteers presenting normal hematological and biochemical values, without history of kidney or inflammatory diseases, matched as far as possible for age and gender with patients, was also included in the study.

Patients were under therapeutic HD three times per week, for 3–5 h, for a median period of time of 28.3 months. All patients used the high-flux polysulfone FX-class dialyzers of Fresenius. Fourteen patients used central venous catheter (CVC) and 35 arteriovenous fistulas (AVFs), as vascular access for HD. The causes of renal failure in these patients were as follows: diabetic nephropathy (n = 17), hypertensive nephrosclerosis (n = 5), obstructive nephropathy (n = 2), amyloidosis (n = 1), chronic nephritis syndrome (n = 3), tubular nephropathy (n = 1), chronic glomerulonephritis (n = 2), renal myeloma (n = 1), tuberculosis (n = 1), and uncertain etiology (n = 16). Patients with autoimmune disease, malignancy, hematological disorders, and acute or chronic infection were excluded. All participants gave their informed consent to participate in this study.

The HD patients included 34 responders [0.33 (0.12–0.55) μg darbepoetin/kg/week] and 15 nonresponders [1.98 (1.76–2.34) μg darbepoetin/kg/week] to rhEPO therapy, in accordance with the European Best Practice Guidelines.⁶

Blood sample collection was performed immediately before starting HD procedure. Red blood cell (RBC) count, hematocrit, hemoglobin (Hb) concentration, hematological indices, and red cell distribution width (RDW) were measured by using an automatic blood cell counter (Sysmex K1000; Sysmex, Hamburg, Germany). Reticulocyte count was measured by microscopic counting on blood smears after vital staining with new methylene blue (reticulocyte stain; Sigma, St. Louis, MO, USA). The reticulocyte production index (RPI) was calculated to measure the effective reticulocyte production. Serum iron concentration was determined using a colorimetric method (Iron, Randox Laboratories Ltd., North Ireland, UK), whereas serum ferritin (Laboratories Ltd., North Ireland, UK) and serum transferrin (Transferrin, Laboratories Ltd.) were measured by immunoturbidimetry. Transferrin saturation (TS) was calculated by the formula: TS (%) = 70.9 × serum iron concentration (μg/dL)/serum transferrin concentration (mg/dL). Enzyme-linked immunosorbent assays were used for measurement of plasma soluble transferrin receptors (s-TfRs) (Human sTfR immunoassay, R&D Systems, MN, USA), IL-6 (human IL-6, Bender Med Systems), and BMP-2 (BMP-2 human, R&D Systems). Serum C-reactive protein (CRP) was determined by nephelometry (N high sensitivity CRP, Dade Behring, Marburg, Germany). Serum hepcidin concentration was performed by using an enzymatic immunoassay (Hepcidin-25, EIA Kit Extraction-Free, Peninsula Laboratories, LLC, San Carlos, CA, USA).

For statistical analysis, the Statistical Package for Social Sciences, version 17.0, was used. For comparisons between groups, we used the Student’s t-test whenever the parameters presented a Gaussian distribution and the Mann–Whitney U-test in the case of a non-Gaussian distribution. Spearman’s and Pearson’s rank correlation coefficients were used to evaluate relationships between sets of data. Multiple regression analysis using the stepwise method was used to determine independent factors affecting BMP-2 serum concentration. Significance was accepted at p < 0.05.

Table 1 shows hematological data, iron status, inflammatory markers, and BMP-2 serum levels, for controls and HD patients, according to the vascular access and sensitivity to rhEPO therapy. When compared with controls, HD patients showed significantly lower Hb concentration, hematocrit, RBC count, and mean cell hemoglobin concentration, and significantly increased mean cell volume. HD patients also presented significantly decreased reticulocyte and lymphocyte counts, significant changes in iron metabolism (higher ferritin, s-TfR and hepcidin, and lower iron and transferrin levels), and a significant increase in inflammatory markers (IL-6 and CRP) and in BMP-2 serum levels. AVF and responder patients showed similar median values that were lower than those presented by CVC and responder patients, respectively; only HD patients using CVC as vascular access presented significantly increased BMP-2 serum levels, when compared with those using AVF. In HD patients, negative correlations were found between BMP-2 serum levels and erythrocyte (r = −0.390; p = 0.006) and reticulocyte counts (r = −0.305; p = 0.033). No correlations were found between BMP-2 serum levels, and rhEPO doses, inflammatory and iron metabolism markers, including hepcidin serum levels.

Table 1. Hematological data, iron status, inflammatory markers, and BMP-2 serum levels, for controls and HD patients.

			HD patients by vascular access		HD patients by response to rhEPO
	Controls (n = 24)	All HD patients (n = 49)	AVF (n = 35)	CVC (n = 14)	Responders (n = 34)	Nonresponders (n = 15)
rhEPO dose (IU/kg/week)	–	0.51 (0.16–1.62)	0.36 (0.14–1.63)	0.96 (0.39–1.60)^b	0.33 (0.12–0.55)	1.98 (1.76–2.34)^c
Hb (g/dL)	14.94 ± 0.88	11.95 ± 1.79^a	11.95 ± 1.77	11.92 ± 1.92	12.42 ± 1.70	10.86 ± 1.53^c
Hematocrit (%)	44.32 ± 2.85	37.38 ± 5.35^a	36.79 ± 5.03	38.85 ± 6.00	38.22 ± 5.37	35.46 ± 4.93
RBC (× 10^12/L)	4.90 ± 0.46	3.93 ± 0.56^a	3.86 ± 0.54	4.14 ± 0.56	4.00 ± 0.54	3.78 ± 0.58
MCV (fL)	88.25 ±11.95	95.03 ± 5.86^a	95.41 ± 5.52	94.07 ± 6.76	95.58. ± 5.54	93.76 ± 6.54
MCH (pg)	30.62 ± 1.61	30.39 ± 2.44	31.08 ± 2.10	28.62 ± 2.43^b	31.00 ± 2.32	29.19 ± 6.54^c
MCHC (g/dL)	33.75 ± 0.47	32.05 ± 1.48^a	32.47 ± 1.42	31.01 ± 1.06^b	32.54 ± 1.27	30.95 ± 1.34^c
RDW (%)	15.00 (14.80–15.50)	15.91 (13.67–17.90)	14.30 (13.60–16.45)	17.90 (14.35–20.25)^b	14.00 (13.60–16.38)	17.15 (14.98–18.85)^c
Reticulocytes (× 10^9/L)	61.09 ± 32.01	43.40 ± 27.77^a	42.92 ± 27.18	46.38 ± 29.77	43.41 ± 25.78	43.37 ± 32.81
RPI	1.21 ± 0.62	0.92 ± 0.60^a	0.84 ± 0.59	0.98 ± 0.64	0.92 ± 0.55	0.92 ± 0.71
Leucocytes (× 10^9/L)	6.55 ± 1.15	6.46 ± 2.13	6.51 ± 2.22	6.34 ± 1.95	6.68 ± 2.32	5.92 ± 1.48
Neutrophils (× 10^9/L)	3.98 (3.16–4.90)	3.81 (2.97–4.91)	3.83 (3.08–4.93)	3.48 (2.64–4.88)	3.78 (2.93–5.04)	4.09 (2.96–4.53)
Lymphocytes (× 10^9/L)	1.96 ± 0.37	1.45 ± 0.65^a	1.43 ± 0.61	1.50 ± 0.75	1.59 ± 0.64	1.12 ± 0.54^c
Neutrophil/lymphocyte ratio	1.72 ± 0.11	3.21 ± 0.44^a	3.34 ± 1.80	2.88 ± 0.48^a	2.17 ± 0.14	4.05 ± 0.34^c
Iron metabolism
Iron (μg/dL)	54.50 (43.25–70.75)	32.00 (26.50–51.00)^a	37.00 (28.00–54.00)	27.00 (19.50–32.25)	37.00 (28.75–55.00)	24.00 (15.00–30.00)^c
Ferritin  (ng/mL)	103.06 ± 61.86	415.50 ± 166.00^a	461.28 ± 157.07	301.00 ± 132.21^b	434.80 ± 147.66	371.76 ± 200.31
Transferrin  (mg/dL)	312.17 ± 44.47	172.14 ± 36.74^a	171.66 ± 30.56	173.35 ± 59.40	174.00 ± 36.86	167.93 ± 37.40
Transferrin  saturation (%)	14.23 ± 2.71	13.32 ± 4.75	13.32 ± 4.75	8.77 ± 1.26	20.97 ± 13.95	13.97 ± 4.75
sTfR (nmol/L)	12.82 (11.06–15.25)	20.97 (15.16–38.27)^a	20.10 (14.14–27.59)	37.44(22.05–56.78)^b	20.73 (14.16–27.83)	41.26 (17.22–47.87)^c
Hepcidin  (ng/mL)	218.40 (124.46–318.63)	979.74 (514.40–1698.18)^a	1175.80 (565.68–1942.00)	841.26 (273.59–1354.90)^b	1052.60 (663.36–1772.85)	791.13 (326.09–1484.30)^c
Inflammatory markers and BMP-2 serum levels
CRP (mg/dL)	0.68 (0.42–1.47)	6.14 (2.68–17.23)^a	4.96 (2.31–16.94)	10.45 (5.28–20.74)	5.46 (2.27–17.08)	9.73 (2.97–27.75)
IL-6 (pg/mL)	0.42 (0.27–0.59)	3.19 (1.23–4.69)^a	2.20 (1.18–4.44)	2.65 (1.23–5.87)	2.15 (1.06–4.60)	3.40 (1.24–4.93)
BMP-2 (pg/mL)	26.49 (18.58–34.26)	46.41 (31.17–62.62)^a	43.41 (24.92–55.33)	55.31 (35.78–85.65)^b	43.40 (30.39–53.11)	55.32 (34.25–66.95)

Notes: Results are presented as mean ± SD or as median values (inter-quartile range). Hb, hemoglobin concentration; RBC, red blood cell count; MCV, mean cell volume; MCH, mean cell hemoglobin; MCHC, mean cell hemoglobin concentration; RDW, red cell distribution width; RPI, reticulocyte production index; sTfR, soluble transferin receptor; CRP, C-reactive protein; IL-6, interleukin-6; BMP-2, bone morphogenetic protein-2; AVF, arteriovenous fistula.

^ap < 0.005 versus controls.

^bp < 0.05 versus AVF.

^cp < 0.05 versus responders.

Multiple regression analysis identified hemoglobin concentration (β = −0.489; p < 0.001), type of vascular access (β = 0.457; p < 0.001), transferrin concentration (β = −0.253; p = 0.027), and age (β = −0.244; p = 0.03) as independent variables associated with BMP-2 serum concentration in HD patients.

Our data, in accordance with literature, show that HD patients present high levels of ferritin and low levels of seric iron and transferrin, alongside with high levels of inflammatory markers, namely IL-6 that is known to upregulate hepcidin synthesis.^5,7 Indeed, hepcidin is significantly higher and contribute, at least in part, to the rise in ferritin (ferritin is an acute phase protein and HD patients receive iron supplementation), via reduction of iron mobilization from the macrophages; hepcidin is involved in the degradation of ferroportin, which is present in the membrane of macrophages and enterocytes, reducing iron mobilization from the macrophages and iron absorption through enterocytes.⁸ It is known that BMP-2 is upregulated by iron overload and downregulated by hypoxia/anemia. In HD patients, both conditions coexist. There is also a lower erythroid cell turnover (low RPI and reticulocyte count) that has been associated with downregulation of hepcidin synthesis,⁹ but on the other hand the synthesis of hepcidin is also upregulated by BMP-2.¹⁰ Therefore, in HD patients, there are different counteracting factors concerning the modulation of hepcidin and BMP-2 synthesis. In these patients, there is also another factor that is important to consider, the dose of rhEPO used to correct anemia. Actually, increasing rhEPO has been associated with decreasing hepcidin synthesis;⁹ in accordance, we found a significant negative correlation between rhEPO doses and the levels of hepcidin in our patients.

The use of CVC and resistance to rhEPO has been associated with higher levels of inflammation.¹¹ In our study, both groups presented higher levels of IL-6 and PCR (although without reaching statistical significance), but were associated with lower levels of hepcidin, when compared with patients with AVF and responders to rhEPO, respectively. Both the groups needed higher doses of rhEPO, suggesting that rhEPO could have contributed to downregulate hepcidin. In CVC and nonresponders to rhEPO groups, the lower erythroid turnover lead to lower hemoglobin values and hypoxia, underlying the higher levels observed for BMP-2 and the lower hepcidin values. The multiple regression analysis showed the severity of the anemia, low transferrin, the use of CVC and age, as the most important determinant factors for BMP-2 levels in HD patients.

In conclusion, in HD patients, there are exogenous factors, namely the type of vascular access for the HD procedure, the iron supplements, and the therapeutic dose of rhEPO, that interfere with iron homeostasis, hepcidin synthesis, and BMP-2 levels.

ACKNOWLEDGMENTS

This work was supported by national funds “Fundação Portuguesa para a Ciência e Tecnologia” (FCT: PIC/IC/83221/2007) and co-financed by FEDER (FCOMP-01-0124-FEDER-008468).

Elísio Costa^1,2, Joana Coimbra³, Cristina Catarino^1,2, Sandra Ribeiro^1,2, Flávio Reis⁴, HenriqueNascimento^1,2, João Fernandes^2,4, Vasco Miranda⁵, Maria do Sameiro Faria⁵, Luís Belo^1,2 and Alice Santos-Silva^1,2

¹Departamento de Ciências Biológicas, Faculdade de Farmácia, Universidade do Porto, Porto, Portugal; ²Instituto de Biologia Molecular e Celular (IBMC), Universidade do Porto, Porto, Portugal; ³Instituto de Ciências da Saúde, Universidade Católica Portuguesa, Porto, Portugal; ⁴Laboratorio de Farmacologia e Terapêutica Experimental, IBILI, Universidade de Coimbra, Coimbra, Portugal; ⁵Fresenius Medical Center, Dinefro-Diálise e Nefrologia, Maia, Portugal