Assessing and influencing the fractional contribution of erythrocyte-bound 59Fe to individual 59Fe tissue content in murine 59Fe distribution studies

Available online at www.sciencedirect.com

Toxicology 244 (2008) 198–208

Assessing and influencing the fractional contribution of erythrocyte-bound 59Fe to individual 59Fe tissue content in murine 59Fe distribution studies B. Szegner a , N. Herbach b , T. Ettle c , B. Elsenhans d , K. Sch¨umann a,∗ a

Wissenschaftszentrum Weihenstephan, Technische Universit¨at M¨unchen, Germany Institut f¨ur Tierpathologie, Ludwig-Maximilians-Universit¨at M¨unchen, Germany Abteilung Tierische Lebensmittel, Tierern¨ahrung und Ern¨ahrungsphysiologie, Universit¨at f¨ur Bodenkultur Wien, Austria d Walther-Straub-Institut for Pharmacology and Toxicology, LMU, M¨ unchen, Germany b

c

Received 17 October 2007; received in revised form 19 November 2007; accepted 19 November 2007 Available online 28 November 2007

Abstract Background: Murine proteins of iron homoeostasis are frequently manipulated to investigate the mechanisms of iron-distribution and their toxicological consequences. Beyond subtracting erythrocyte-bound 59 Fe of the residual blood content determined for each tissue (subtraction method), procedures are needed to determine 59 Fe distribution in murine models of, e.g. inflammation or diabetes that cause local hyperaemia and changes in microcirculation. Aim: Two new methods were developed to correct total 59 Fe tissue content individually for erythrocyte-bound 59 Fe-labelled haem iron. Methods: Iron-deficiency and iron-overload was induced in male C57BL6 mice by feeding of respective diets. Distribution of 59 Fe between different tissues was determined 24 h, 14, and 28 days after intravenous injection of 59 Fe trace amounts. Haem-bound 59 Fe was separated from non-haem 59 Fe in homogenates from all tissues by dispersion in a mix of lipophilic cyclohexanone and hydrophilic H3 PO4 (separation method). Moreover, the reduction of 59 Fe-labelled tissue blood content was determined in all organs after in vivo saline perfusion via the left ventricle (perfusion method). Results and discussion: 59 Fe-labelled non-haem iron determined by the separation method was not significantly different from values determined by the subtraction method, except for the iron-deficient spleen 14 and 28 days after 59 Fe injection when the separation method yielded ∼20% higher values. Approximately 20% of 59 Fe-labelled haem iron spilled over into the hydrophilic phase. The impact of this error decreases in parallel to 59 Fe radioactivity in the residual tissue blood content: thus, it is higher in iron-deficient mice which accumulate more 59 Fe in their erythrocytes than iron-adequate and iron-rich mice. For the same reason this type of error is more marked after long distribution periods and in organs with high residual blood content. Saline perfusion via the left ventricle reduced total blood content in mice to less than 10%. Liver (95%) and duodenum (94%) showed the highest removal of blood while it is lowest in spleen (66%) and lungs (69%). Conclusions: The separation and the perfusion method can be used to correct the impact of erythrocyte-bound haem iron individually. A margin of error below 10% was determined for all organs except for spleen, lungs, and fat. Both methods can be applied sequentially to obtain satisfactory results in spleen, lungs, and fat. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Iron-distribution; Mouse; Residual blood; In vivo perfusion; Haem iron separation

∗

Corresponding author at: Lehrstuhl f¨ur Ern¨ahrungsphysiologie, TUM, Am Forum 5, D 85350 Freising, Germany. E-mail address: [email protected](K. Sch¨umann).

0300-483X/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.tox.2007.11.015

B. Szegner et al. / Toxicology 244 (2008) 198–208

1. Introduction Distribution kinetics between iron storage compartments and functional iron compartments are closely regulated by corresponding expression of the proteins of iron transport and regulation (Hentze et al., 2004). Homoeostatic mechanisms optimize iron-distribution in situations of scarcity and reduce formation of reactive oxygen species by excessive iron which can oxidize lipids, DNA, and proteins. Deregulated iron-distribution is discussed to participate in the pathogenesis of atherosclerosis, inflammatory bowel disease, colon cancer, diarrhoea, and significantly increases the prevalence of serious courses of malaria (see ref. Sch¨umann, 2001; Sch¨umann et al., 2007a; Sch¨umann and Solomons, 2007). Therefore, the mechanisms of iron homoeostasis are of considerable toxicological interest. An increasing number of mice with genetically manipulated proteins of iron metabolism are developed to elucidate these mechanisms in vivo. Among these are mice with differential expression of hephaestin (Vulpe et al., 1999), hepcidin (Nicolas et al., 2002), iron regulatory protein-2 (IRP-2) (Grabill et al., 2003; Galy et al., 2005), divalent metal transporter (DMT-1) (Fleming et al., 1998), transferrin receptor-1 (TfR-1) (Levy et al., 1999), transferrin (Trenor et al., 2000), haemojuvelin (Huang et al., 2005), ferritin (K¨uhn et al., 2007), or tissue-specific HFE knock-out (Spasic et al., 2007), to name a few. The complex concert of interactions between proteins of iron homoeostasis can only be assessed in vivo, and the mouse is by far the most widely used mammalian species for this purpose. A common method to assess trace metal shifts between different compartments of the organism is to inject radioactive isotopes and to pursue their distribution over time (e.g. Hunder et al., 2000). However, a large fraction of body iron is channelled rapidly into erythropoiesis and, thus, into circulating erythrocytes (Bothwell et al., 1979; Sch¨umann et al., 2007b). Such erythrocyte-bound 59 Fe will contribute to total 59 Fe radioactivity in each tissue, although its quantity is independent from the mechanisms that regulate irondistribution and will disturb their assessment. A recently published method subtracts 59 Fe-bound iron in the residual blood content from total 59 Fe radioactivity in each organ (Sch¨umann et al., 2007b). However, the “subtraction method” cannot account for exceptional shifts of blood between organs, which may, e.g. be due to local hyperaemia in murine models of inflammation, such as TNFARE -mice or IL-10-deficient mice (Kontoyiannis et al., 2002; Kuhn et al., 1993), or for changes in the

199

organs microcirculation that may affect residual tissue blood content, e.g. in diabetic mice (Srinivasan and Ramarao, 2007). The impact of such changes on body iron-distribution is of high interest, e.g. in order to elucidate the pathophysiology of anaemia of inflammation (Weiss and Goodnough, 2005) or to investigate the involvement of iron metabolism in obesity (Kennedy et al., 1986). Such cases require individual assessment of haem-bound and non-haem-bound 59 Fe in tissues and organs. Therefore, we adapted and tested two new methods for this purpose: the first approach separates haem iron from non-haem iron in 59 Fe-labelled tissue homogenates by means of their different lipophilicity. The spillover of 59 Fe into the reciprocal fraction was assessed by spiking tissue homogenates with 59 Fe-labelled erythrocytes and aqueous non-haem iron, respectively. The results of the “separation method” were compared to corresponding values of the “subtraction method” (Sch¨umann et al., 2007b). The second approach reduces residual tissue blood content by in vivo whole-body perfusion via the left ventricle (“perfusion method”). The “separation method” and the “perfusion method” and their margins of error are compared. 2. Materials and methods 2.1. Animals The experiments were approved by the Tierschutzkommission der Regierung von Oberbayern (AZ 209.1/211-2531109/03). Conventionally bred male C57BL6 mice [Charles River, Sulzfeld, Germany; 30–40 days old (=18–20 g body weight)] were housed in macrolone cages (5–8 animals/cage, 12:12 h light–dark cycle: 06:00–18:00 h; 22 ± 1 ◦ C, 60 ± 5% humidity). Iron-deficiency was induced by feeding an irondeficient diet (C1038, Altromin, Lage, Germany; total Fe content: 6 mg/kg) and distilled water ad libitum for 5 weeks (age 7–12 weeks, Sch¨umann et al., 1989). Control mice and iron-overloaded animals were fed an iron-adequate diet of the same composition (C 1000, Altromin, total Fe content: 180 mg/kg) or an iron-rich diet for 2 weeks (25 g Fe/kg diet, added as carbonyl iron [C3518, Sigma, Munich, Germany]; age 10–12 weeks). 2.2. Analysis of body iron status Haematocrit and haemoglobin concentrations were measured by the microhaematocrit method (No. 749311, Brand, Wertheim, Germany; Hematocrit-centrifuge 2104, Hettich, Tuttlingen, Germany) and the cyanmethaemoglobin method (reagent: Bioanalytic 4001, Umrich, Freiburg, Germany; photometer: UV-DK-20, Beckmann, M¨unchen, Germany). To determine unlabelled non-haem iron, liver samples

200

B. Szegner et al. / Toxicology 244 (2008) 198–208

(∼300–400 mg) were cut to pieces (∼1 mm3 ) and dispersed in an acid-mixture (6 M HCl, 20% TCA (v/v); 65 ◦ C, 20 h). The solution was centrifuged (3000 rpm, 3 min) and the non-haem iron concentration in the supernatant was determined photometrically (Feren-B, Bioanalytic, Umrich/Freiburg, Germany) (Torrance and Bothwell, 1968; Frazer et al., 2002). 2.3.

59

Fe injection

Nitrilotriacetic acid (NTA), complexed with Fe(NO3 )3 at a molar ratio of Fe:NTA of 1:2, was solved in isotonic HEPESbuffered saline (pH 7.2) and labelled with ∼2 ␮Ci 59 Fe/animal (NEZ 037, NEN, Dreieich, Germany) (Hunder et al., 2000). For routine distribution studies a dose of 0.2 ␮mol Fe/kg body weight was injected intravenously in ether anaesthesia. Mice used as donors for 59 Fe-labelled blood were killed 14–28 days after injection of the same solution labelled with 10–20 ␮Ci 59 Fe/animal and blood was sampled. 59 Fe radioactivity was determined per ␮L of donor blood. Only traces of 59 Fe radioactivity were found in donor serum, i.e. all 59 Fe radioactivity in the blood had accumulated in the erythrocytes (Sch¨umann et al., 2007b). 2.4. Use of 59 Fe-labelled erythrocytes Receiver mice were anaesthetised (Dormitor® , 0.33 mg/kg medetomidinhydrochlorid; plus Ketavet® , 66 mg/kg ketaminhydrochlorid, i.m., Pharmacia GmbH, Karlsruhe, Germany) and i.v. injected with 200 ␮L of 59 Fe-labelled blood, while 200 ␮L of receiver blood was subsequently drawn to provide isovolumetric conditions for erythrocyte distribution. Injected erythrocyte-bound 59 Fe radioactivity in the mice was determined in a whole-body counter for small animals (Typ AW3, M¨unchener Apparatebau, Unterf¨ohring, Germany). After 15 min, 59 Fe-labelled erythrocyte distribution reached a steady state (Everett et al., 1956) and receiver animals were killed. Moreover, we analysed (1) the reduction of residual 59 Fe-labelled blood content in the whole animal and in all of its organs after in vivo PBS perfusion, and (2) the spillover of erythrocyte-bound 59 Fe into the hydrophilic phase after snapfreezing of liver tissue and storage at −80 ◦ C as compared to storage at room temperature for up to 14 days. 2.5. 59 Fe tissue content after subtraction of haem-bound 59 Fe in residual tissue blood content (“subtraction method”) The “subtraction method” extrapolates the 59 Fe concentration in the blood to the average residual blood volume in each organ, which was determined earlier by use of 59 Fe-labelled mouse erythrocytes in the same strain. The extrapolated haembound 59 Fe radioactivity was subtracted from total 59 Fe in the tissue to derive a value for the non-haem iron content of each organ (Sch¨umann et al., 2007b).

2.6. Separating haem iron from non-haem iron in tissue homogenates (“separation method”) Tissue samples were homogenised in digitonine solution (130 mg digitonine/L, No. 4946.1, Roth, Karlsruhe, Germany; 20% (w/w), 2 min, 20,000 rpm, IKA-Ultraturrax, Janke & Kunkel, Staufen, Germany). The homogenates were immersed in an ultrasonic bath for 10 min to crack erythrocyte membranes. 59 Fe radioactivity was determined in 313 ␮L homogenate samples in 2 mL disposable plastic tubes (Sorenson, Salt Lake City, USA). The homogenates were incubated with 312 ␮L H3 PO4 (85%, 9 mol/L, 37 ◦ C, 20 min; Sigma, No. 466123) and mixed with 125 ␮L of a saturated KH2 PO4 solution for 1 min (Sigma, No. 0662), adding up to a total volume of 750 ␮L. After addition of 750 ␮L cyclohexanone (Sigma, No. 398241), the solution was intensely agitated for 5 min (Vortex, Janke & Kunkel IKA Labortechnik VF, Staufen, Deutschland) to permit the distribution of 59 Fe-labelled haem iron and non-haem iron between the lipophilic and hydrophobic phases of the turbide mixture (Fig. 1a). After centrifugation (14,000 rpm, 15 min, Centrifuge Sigma 1–15 K, Steinheim, Germany), the mixture separated into a lower, inorganic phase (hydrophilic; specific weight: 1.2 g/mL) containing 59 Fe-labelled non-haem iron, and an upper, organic phase (lipophilic; cyclohexanone, specific weight: 0.95 g/mL) containing haem-bound 59 Fe (Fig. 1b). A thin layer of tissue debris floats on the dense H3 PO4 /KH2 PO4 phase. When the layer of hepatic tissue debris was subjected to the separation process for a second time, 79 ± 3% (n = 5) of the residual 59 Fe content moved into the hydrophilic inorganic phase. Therefore, 80% of 59 Fe radioactivity in the debris was regarded to represent inorganic iron. One, 14, and 28 days after 59 Fe injection, the fraction of 59 Fe-labelled haem iron was assessed in different samples of the same tissue homogenates by the “separation method” and by the “subtraction method” in parallel. 2.7. Control of 59 Fe spillover to the reciprocal phase during separation Fifty microlitres of a 59 Fe saline solution (20 ␮mol Fe/L), spiked with ∼25,000 CPM 59 Fe/50 ␮L, were added to 313 ␮L of a 20% (w/v) liver homogenate and subsequently subjected to the separation process to determine the spillover of non-haem iron into the lipid phase (n = 3). To determine the spillover of 59 Fe-labelled haem iron into the hydrophilic fraction, we added 50 ␮L of 59 Fe-labelled donor blood (∼25,000 CPM 59 Fe/50 ␮L) to 313 ␮L of a 20% liver homogenate and subjected it to the separation process a second time (n = 4). 2.8. Reduction of residual tissue blood by in vivo PBS perfusion via the left ventricle Fifteen minutes after i.v. injection of 59 Fe-labelled blood into anaesthetised receiver animals, ∼100 ␮L blood was

Table 1 59 Fe content [nmol 59 Fe/g wwt] in different organs after applying the “separation method” (upper lines in bold letters) and the “subtraction method” (lower lines) Iron-deficient (days)

Iron-adequate (days)

14

28

1

14

28

1

14

28

Brain

0.019 ± 0.006 0.022 ± 0.009

0.018 ± 0.003 0.022 ± 0.004

0.028 ± 0.008 0.026 ± 0.005

0.018 ± 0.010 0.021 ± 0.011

0.018 ± 0.004 0.022 ± 0.004

0.015 ± 0.005 0.015 ± 0.011

0.007 ± 0.002 0.014 ± 0.006

0.007 ± 0.002 0.009 ± 0.008

0.012 ± 0.003 0.013 ± 0.003

Muscle

0.046 ± 0.032 0.032 ± 0.015

0.032 ± 0.012 0.019 ± 0.007

0.027 ± 0.013 0.020 ± 0.006

0.022 ± 0.021 0.022 ± 0.018

0.038 ± 0.015 0.026 ± 0.008

0.021 ± 0.004 0.014 ± 0.012

0.031 ± 0.006 0.037 ± 0.012

0.020 ± 0.014 0.016 ± 0.014

0.020 ± 0.010 0.025 ± 0.015

Stomach

0.195 ± 0.137 0.284 ± 0.181

0.059 ± 0.003 0.107 ± 0.032

0.037 ± 0.015 0.032 ± 0.181

0.122 ± 0.088 0.151 ± 0.125

0.046 ± 0.008 0.051 ± 0.008

0.044 ± 0.009 0.029 ± 0.020

0.241 ± 0.129 0.242 ± 0.095

0.062 ± 0.047 0.088 ± 0.072

0.064 ± 0.029 0.086 ± 0.053

Kidney

0.483 ± 0.263 0.403 ± 0.164

0.266 ± 0.039 0.287 ± 0.048

0.142 ± 0.054 0.132 ± 0.039

0.218 ± 0.137 0.260 ± 0.182

0.192 ± 0.042 0.139 ± 0.092

0.315 ± 0.229 0.355 ± 0.212

0.394 ± 0.178 0.490 ± 0.240

Lung

0.638 ± 0.389 0.742 ± 0.397

0.213 ± 0.169 0.104 ± 0.138

0.173 ± 0.076 0.040 ± 0.087

0.162 ± 0.078 0.091 ± 0.090

0.352 ± 0.303 0.054 ± 0.131

0.475 ± 0.319 0.596 ± 0.458

0.310 ± 0.187 0.498 ± 0.401

+

+ +

0.303 ± 0.134 0.351 ± 0.200 0.540 ± 0.332 0.784 ± 0.484

+ +

+ +

1.075 ± 0.379 1.149 ± 0.506 0.555 ± 0.434 0.831 ± 0.731

 

−

− −

Results showed no significant differences between both methods (p < 0.05, Student’s t-test). Values 1, 14, and 28 days after59 Fe injection (M ± S.D., n = 3–4) are presented in iron-deficient, iron-adequate, and iron-rich mice. The symbols −, o, and + signify significant differences between corresponding points in time, for the comparison between the iron-deficient (−), iron-adequate (), and iron-rich (+) counterparts (p < 0.05, ANOVA).

201

Fig. 1. Separation of 59 Fe-labelled haem iron from 59 Fe-labelled nonhaem iron in tissue homogenates. Tissue homogenate is mixed with the lipophilic and hydrophilic phase, a turbid white fluid (a). After centrifugation, the lipophilic and hydrophilic phase are separated (b). 59 Fe-labelled haem dissolves in the upper (lipophilic) phase and produces a brownish colour. 59 Fe-labelled non-haem iron is found in the water-clear lower (hydrophilic) phase. The optically dense layer between both phases consists of homogenised tissue debris floating on the more dense hydrophilic phase, and ∼80% of the 59 Fe-content of which is non-haem iron in this layer.

sampled to determine its 59 Fe concentration. The ribs were cut right and left from the sternum at 1–2 cm distance from each other. The distal end of the sternum was lifted and a 20 G × 2 5/5 in. cannula was inserted into the left ventricle. After cutting the vena cava inferior, 2 mL of lidocaine hydrochloride (2%, Astra-Zeneca, Wedel, Germany) was injected. After 2 min of perfusion (PBS, 37 ◦ C, 120 cm H2 O-column), tissues and organs became markedly pale and the efflux from the v. cava contained almost pure PBS. Perfusion was stopped and 59 Fe in the carcass was determined in a whole-body counter before dissection. 59 Fe radioactivity was measured in each organ after dissection in a well-type ␥-counter (1282 Compugamma CS, LKB, Wallac, Finland). These results were compared to 59 Fe radioactivity in corresponding organs of a parallel group of mice which were dissected after sampling ∼100 ␮L of blood without previous PBS perfusion. Longer perfusion periods did

B. Szegner et al. / Toxicology 244 (2008) 198–208

1

Iron-rich (days)

202

B. Szegner et al. / Toxicology 244 (2008) 198–208

Table 2 Total 59 Fe content [nmol 59 Fe/g wwt] (bold letters, first line), content (italic letters, third line)

59 Fe

content of the residual blood (second line), and

1d

Iron-deficient Duodenum

Liver

Spleen

Iron-adequate Duodenum

Liver

Spleen

Iron-rich Duodenum

Liver

Spleen

14 d

[nmol

59 Fe/g]

0.383 0.031 0.352 0.679 0.108 0.571 1.126 0.247 0.879

± ± ± ± ± ± ± ± ±

0.204 0.007 0.199 0.303 0.024 0.280 0.290 0.054 0.285

0.301 0.027 0.275 0.557 0.094 0.463 0.864 0.214 0.650

± ± ± ± ± ± ± ± ±

0.164 0.010 0.154 0.309 0.036 0.274 0.193 0.083 0.221

0.084 0.010 0.273 1.578 0.036 1542 0.541 0.081 0.459

± ± ± ± ± ± ± ± ±

0.094 0.002 0.093 0.859 0.005 0.856 0.147 0.012 0.156

[%]

8

16

22

[nmol

± ± ± ± ± ± ± ± ±

0.016 0.006 0.011 0.050 0.022 0.062 0.066 0.051 0.040

0.062 0.036 0.026 0.289 0.126 0.166 0.389 0.288 0.101

± ± ± ± ± ± ± ± ±

0.034 0.008 0.022 0.130 0.028 0.081 0.187 0.063 0.137

0.084 0.014 0.070 0.672 0.051 0.621 0.367 0.116 0.251

± ± ± ± ± ± ± ± ±

0.059 0.002 0.057 0.132 0.005 0.127 0.188 0.012 0.200

17

25

4

2

15

non-haem iron

28 d 59 Fe/g]

0.095 0.056 0.039 0.315 0.196 0.119 0.394 0.447 −0.053

9

59 Fe-labelled

[%]

59

62

100

58

44

74

17

8

32

[nmol 59 Fe/g] 0.042 0.040 0.002 0.183 0.140 0.043 0.315 0.320 −0.005

± ± ± ± ± ± ± ± ±

0.012 0.011 0.005 0.053 0.039 0.031 0.091 0.088 0.044

0.061 0.029 0.031 0.173 0.103 0.070 0.472 0.235 0.236

± ± ± ± ± ± ± ± ±

0.066 0.016 0.057 0.109 0.057 0.058 0.288 0.130 0.183

0.050 0.009 0.041 0.759 0.030 0.729 0.253 0.069 0.184

± ± ± ± ± ± ± ± ±

0.019 0.003 0.022 0.261 0.012 0.267 0.107 0.027 0.080

[%]

94

77

100

48

60

50

17

4

27

Values in duodenum, liver, and spleen are given 1, 14, and 28 days after 59 Fe injection (M ± S.D., n = 3-4). The fraction of 59 Fe in the residual blood is also given in percent.

not decrease the 59 Fe-labelled blood content of the animals any further (n = 5). All chemicals were purchased from Sigma (M¨unchen, Germany) and Merck (Darmstadt, Germany).

2.9. Statistics Values of the different groups are presented as means ± standard deviation. The results of two methods (Table 1) were compared by use of the unpaired Student’s t-test (p < 0.05). Corresponding iron-deficient, iron-adequate, and iron-rich groups, as well as changes in 59 Fe distribution over time within each group were compared by ANOVA, followed by a Scheff´e-test (WinSTAT® , Microsoft, USA). The level of significance was p < 0.05.

3. Results 3.1. Iron status Iron-deficiency and iron loading were successfully induced (unlabelled liver non-haem iron content in iron-deficient, iron-adequate, and iron-rich mice: 27.3 ± 6.7 ␮g Fe/g wet weight (wwt), 49.0 ± 16.2 ␮g Fe/g wwt, 820.2 ± 194.6 ␮g Fe/g wwt). In accordance to earlier results (Sch¨umann et al., 1989), haemoglobin (Hb) and haematocrit (Hk) showed no significant differences between the three different feeding groups (Hb and Hk; iron-deficient: 11.5 ± 1.7 g/dL, 39.1 ± 5.1%, iron-adequate: 12.3 ± 1.0 g/dL, 42.0 ± 2.1%, iron-rich: 11.7 ± 0.5 g/dL, 39.8 ± 2.6%).

B. Szegner et al. / Toxicology 244 (2008) 198–208

3.2. Assessment of 59 Fe spillover during separation Spiking 20% liver homogenate with ionic 59 Fe or with 59 Fe-labelled erythrocytes revealed a spillover of 4.5 ± 0.25% for ionic 59 Fe into the lipophilic medium (n = 3). 19.8% of 59 Fe-labelled haem iron from solubilised erythrocytes spilled over into the hydrophilic phase (n = 8). After injection of 59 Fe-labelled erythrocytes in vivo, 59 Fe spillover into the hydrophilic fraction of liver homogenates was 25% when the organ was snap-frozen immediately after dissection and stored at −80 ◦ C for up to 14 days. The spillover after 14 days of storage at room temperature was 69% (n = 3). 3.3. Haem and non-haem-bound 59 Fe in different organs The results obtained after separation of haem and nonhaem iron between the hydrophilic and lipophilic phases in duodenum, liver, and heart (“separation method”) were not significantly different from those obtained by the “subtraction method” for most distribution intervals (Fig. 2). However, after 28 days of distribution, iron-deficient spleen and duodenum showed significantly less 59 Fe-labelled non-haem iron, determined by the “subtraction method” as compared to the “separation method” (Fig. 2a and c). In the iron-deficient spleen this was also found after 14 days of 59 Fe distribution (Fig. 2c). Both methods yielded no significant differences in 59 Fe-labelled non-haem tissue iron in brain, muscle, stomach, kidney, and lungs (Table 1). Total 59 Fe, and erythrocyte-related 59 Fe values extrapolated on the basis of previously determined residual tissue blood content (“subtraction method”) are given for duodenum, liver, and spleen in Table 2. The difference between total 59 Fe and erythrocyte-related 59 Fe was called 59 Fe-labelled non-haem iron. Erythrocytebound 59 Fe was expressed as the percentage of total 59 Fe (Tables 2 and 3). Table 3, moreover, assesses the error of the “separation method” caused by the spillover of erythrocyte-bound 59 Fe into the hydrophilic phase. 3.4. Reduction of residual tissue blood by in vivo whole-body perfusion (“perfusion method”) In vivo perfusion with PBS mobilised 92 ± 2% (n = 5) of the entire 59 Fe-labelled blood from the murine organism. The residual blood content in single tissues was reduced significantly, though not to the same extent in

203

all organs (Table 4). The largest fractions of residual blood were retained in fur, spleen, and lungs. In contrast, perfusion mobilised the largest amount of blood from liver, brain, and testis. Table 5 expresses the fraction of residual erythrocyte-bound 59 Fe after perfusion in percent of total 59 Fe tissue content. This fraction cannot be mobilised by the “perfusion method” and represents its margin of error. 4. Discussion The chosen iron dose of 0.2 ␮mol 59 Fe/kg of body weight corresponds to 0.5% of the daily absorbed dietary iron in mice and presumably neither influences body iron-distribution nor the animals’ iron status. 59 Fe is cleared almost entirely from the plasma within 12 h after injection (Hosain and Finch, 1966; Sch¨umann et al., 2007b). The shortest 59 Fe distribution period chosen in this study was 24 h. So, 59 Fe radioactivity in blood can be regarded as ∼100% haem-bound in all mice of this study. Twenty-eight days were chosen as the longest distribution interval because it corresponds to the lifetime of murine erythrocytes (Jasinski et al., 2004). Thus, after 28 days a steady state for 59 Fe distribution can be expected (Sch¨umann et al., 2007b). 4.1. Advantages and disadvantages of the “separation method” Teale (1959) used acid methylethylketone as a lipophilic phase to cleave the haem–protein link. Ehtechami (1987) improved the process by using orthophosphoric acid instead of HCl and cyclohexanone instead of methylethylketone to separate haem from globin in 59 Fe-labelled rat erythrocytes. The latter procedure was applied in the “separation method” to separate 59 Fe-labelled haem from 59 Fe-labelled nonhaem iron in blood containing tissue homogenates. Alternatively, the 59 Fe-labelled non-haem iron fraction can be determined by subtracting erythrocyte-bound 59 Fe radioactivity from total 59 Fe radioactivity in tissue homogenates; erythrocyte-bound 59 Fe being assessed on the basis of previously determined residual blood content in each organ (“subtraction method”; Sch¨umann et al., 2007b). Both methods yielded no significant differences in most organs (Fig. 2). Changes in 59 Fe-distribution in iron-deficient, iron-adequate, and iron-rich mice over time as revealed by use of the “subtraction method” were in accordance to physiological expectation in a recent, more detailed study of murine iron-distribution kinetics (Sch¨umann et al., 2007b). Thus, the agreement of results shows that both methods lead to an appropriate assess-

204

B. Szegner et al. / Toxicology 244 (2008) 198–208

Fig. 2. 59 Fe content in different tissues after separation of haem and non-haem iron in tissue homogenates (“separation method” = white) and after subtraction of 59 Fe in the residual blood content (“subtraction method” = black) 1, 14, and 28 days after 59 Fe injection. Significant differences between the two methods are marked with “*” (p < 0.05, t-test). Data for duodenum, liver, spleen, and heart are given in panels a–d. Each panel compares iron-deficient, iron-adequate, and iron-rich mice (M ± S.D., n = 3–4). The symbols −, , and + over the columns signify significant differences between corresponding points in time, for the comparison between iron-deficient (−), iron-adequate (), and iron-rich (+) counterparts (p < 0.05, ANOVA).

Table 3 59 Fe fractions in residual blood and estimated margins of error for the “separation method” in different organs Iron-deficient (days) [%]

Duodenum Liver Spleen Heart Kidney Fat Brain Muscle Lung Stomach

Iron-adequate (days) [%]

Iron-rich (days) [%]

1

14

28

1

14

28

1

14

28

8 (2) 16 (3) 22 (4) 25 (5) 26 (5) 24 (5) 45 (9) 37 (7) 26 (5) 9 (2)

59 (12) 62 (0) 100 (20) 60 (12) 47 (9) 100 (7) 60 (12) 64 (3) 82 (2) 32 (6)

94 (19) 77 (15) 100 (20) 55 (1) 58 (12) 100 (7) 47 (9) 55 (3) 89 (2) 52 (10)

9 (2) 17 (3) 25 (5) 19 (4) 26 (5) 92 (20) 42 (8) 42 (8) 22 (4) 14 (3)

58 (12) 44 (9) 74 (3) 47 (9) 41 (10) 100 (7) 52 (10) 51 (10) 76 (15) 33 (7)

48 (10) 60 (0) 50 (3) 44 (9) 49 (10) 100 (7) 52 (10) 56 (3) 82 (2) 47 (9)

4 (1) 2 (0) 15 (3) 7 (1) 4 (1) 35 (7) 29 (6) 14 (3) 9 (2) 4 (1)

17 (3) 8 (2) 32 (6) 6 (1) 16 (3) 67 (13) 47 (9) 36 (7) 17 (3) 13 (3)

17 (3) 4 (1) 27 (5) 6 (1) 7 (1) 56 (11) 27 (5) 17 (3) 13 (3) 8 (2)

The fraction of 59 Fe in the residual blood content is given as a percentage of total 59 Fe tissue content (in analogy to Table 2, percent values given in middle lines). Applying the “separation method”, 20% of the residual blood content spilled over into the hydrophilic phase. This leads to corresponding margins of error of the “separation method”, which depends on the residual blood content in each organ (i.e. 20% of the estimated residual blood content). These values are given in brackets. The error increases with the 59 Fe activity in the residual tissue blood content. The error decreases in the order: iron-deficiency > iron-adequate > iron-rich, and 28days > 14days > 1day, because erythrocyte-bound 59 Fe decreases in this order (M ± S.D., n = 3–4).

B. Szegner et al. / Toxicology 244 (2008) 198–208

205

Table 4 Residual blood content [␮L/g wwt] (M ± S.D., n = 5) Blood content without perfusion [␮L blood/g organ] Fur Spleen Lungs Bones Muscle Kidney Heart Stomach Duodenum Liver Brain Testis

13.8 238.9 169.1 35.9 12.1 92.0 68.2 17.7 20.0 70.1 11.6 11.9

± ± ± ± ± ± ± ± ± ± ± ±

Blood content after perfusion [␮L blood/g organ]

2.3 90.6 59.9 11.1 4.4 43.8 36.5 7.1 10.6 26.1 3.7 4.3

4.9 81.7 52.1 6.0 1.8 10.4 7.4 1.4 1.2 3.6 0.4 0.0

± ± ± ± ± ± ± ± ± ± ± ±

Mobilised fraction of erythrocytes [%]

59 Fe-labelled

1.2 36.2 33.4 6.2 1.4 10.2 1.4 0.6 0.6 1.4 0.5 0.0

65 66 69 83 85 89 89 92 94 95 97 100

The mobilised fraction of 59 Fe-labelled total erythrocytes content is given in percent.

ment of changes in iron-distribution and are applicable for this purpose in genetically modified mice. However, significant differences between both methods were observed in duodenum and spleen after long distribution periods (Fig. 2a and c). This finding urges to analyse possible sources of error for each method and to compare their advantages and disadvantages in more detail. Subtraction of the estimated erythrocytebound 59 Fe in the residual blood of the spleen yielded a 59 Fe-labelled non-haem iron fraction that was not significantly different from zero. This finding was associated with increased 59 Fe-shifts into the erythropoietic system, as shown before (Sch¨umann et al., 2007b), and reflects early observations of empty spleenic iron stores in iron-deficiency (Noyes et al., 1960). In contrast to

these findings, the separation method found an endogenous non-haem 59 Fe content of 119 and 78 pmol 59 Fe/g wwt 14 and 28 days after 59 Fe injection, respectively (Fig. 2c). Although we used optimized separation media (Ehtechami, 1987), there was a spillover of ∼20% of 59 Fe-labelled haem iron into the hydrophilic phase. Related to the erythrocyte-bound haem iron content estimated in Table 2, a spillover of ∼20% haem iron accounts for 79 and 63 pmol 59 Fe/g wwt 14 and 28 days after 59 Fe injection. These results explain 60% and 80% of the differences between both methods at these points in time (Fig. 2). Moreover, part of 59 Fe-labelled erythrocytes in the spleen leaves the intravascular space to be degraded when senescent (Dijkstra and Veerman, 1990), and ∼10% of total body iron is incorporated in

Table 5 Non-mobilised residual tissue blood content after in vivo saline perfusion (“perfusion method”) Non-mobilised residual blood after perfusion [%]

Fraction of non-mobilised erythrocyte-bound 59 Fe after perfusion given in percent of total 59 Fe content Iron-deficient (days) [%]

Duodenum Liver Spleen Heart Kidney Fat Brain Muscle Lung Stomach

6 5 34 11 11 13 3 15 31 8

Iron-adequate (days) [%]

Iron-rich (days) [%]

1

14

28

1

14

28

1

14

28

0 1 7 3 3 3 2 5 8 1

3 3 34 7 5 19 2 9 25 2

5 4 34 6 7 23 2 8 27 4

1 1 8 2 3 12 1 6 7 1

3 2 25 5 5 14 2 7 23 3

3 3 17 5 6 19 2 8 25 4

0 0 5 1 0 4 1 2 3 0

1 0 11 1 2 8 2 5 5 1

1 0 9 1 1 7 1 3 4 1

The fraction of non-mobilised 59 Fe after perfusion is given in percent of total 59 Fe content. Assuming that erythrocyte-bound 59 Fe tissue content represents the 59 Fe-labelled non-haem iron fraction of each organ, this percentage reflects the margin of error of the “perfusion method” after removal of all erythrocyte-related activity.

206

B. Szegner et al. / Toxicology 244 (2008) 198–208

haem-containing tissue enzymes (Lynch, 1984). These fractions of 59 Fe-labelled haem iron are not accessible to the “subtraction method”; they are, however, accessible to the separation method. This is a major advantage of this approach over the “subtraction method” and also over the “perfusion method”, both of which exclusively account for erythrocyte-bound 59 Fe and not for 59 Fe in haem-containing tissue enzymes. Thus, based on the data presented here, it cannot be decided to which extent the spillover of haem-bound 59 Fe into the hydrophilic phase and/or the 59 Fe in haem-containing tissue enzymes contributes precisely to the differences between both methods in the spleen and, to a lesser extent, in the duodenum after 28 days of distribution (Fig. 2). The contribution of 59 Fe in the residual blood content to total 59 Fe radioactivity in a tissue can be assessed by the “subtraction method” (Tables 2 and 3). A maximum spillover of 59 Fe-labelled erythrocyte-bound haem iron will cause an error of up to 20% when close to 100% of 59 Fe in a tissue is bound to erythrocyte-related haem iron, as seen, e.g. in the iron-deficient spleen 14 and 28 days after 59 Fe injection (Tables 2 and 3). However, the error will account for only ∼6% in the iron-overloaded spleen at the same point in time (i.e. 20% of the 30% erythrocyte-bound haem iron) (Table 2 and Fig. 1). The error due to spillover was estimated for each organ at each point in time in parallel to this example (Table 3). Haem oxigenase can enzymatically release 59 Felabelled non-haem iron from the haem iron fraction during storage of dissected tissues. However, separation after injection of 59 Fe-labelled erythrocytes yielded no change in 59 Fe spillover to the non-haem iron fraction when organs were frozen at −80 ◦ C for up to 14 days. As this interval seems sufficient to process the samples, longer intervals were not checked. Some tissues are not suitable for the “separation method”. A pilot trial showed that homogenised testis emulsified the lipophilic and hydrophilic phases and prevented no clear separation. Bones and fur are difficult to use as they can hardly be homogenised. The lung has a high residual blood content of 16% of tissue wet weight (Sch¨umann et al., 2007b) due to the low specific weight of lung tissue. Thus, small differences in residual blood have a high impact on results which may be responsible for the high pulmonary standard deviations observed by use of both methods. 4.2. Advantages and disadvantages of the “perfusion method” Whole-body perfusion with lidocaine and PBS via the left cardiac ventricle is a common procedure for in

vivo fixation with glutaraldehyde in electron microscopy (Hoeflich et al., 2002). Two minutes of perfusion mobilised 92% of previously injected 59 Fe-labelled blood from the body. Mobilisation of 59 Fe-labelled erythrocytes differs significantly between organs (Table 4). It is very low in the spleen, most likely because erythrocytes leave the vascular bed in this organ and pass through the spleenic reticuloendothelial tissue (Dijkstra and Veerman, 1990). The lungs are supplied with blood via the right cardiac ventricle, so that blood will not be directly replaced by saline after injection via the left ventricle. This explains incomplete pulmonary wash-out of 59 Fe-loaded erythrocytes by the “perfusion method”. The fur is able to reduce its microcirculation for thermoregulation. Thus, comparably low mobilisation of 59 Fe-labelled blood from these organs is in line with expectation. In contrast, large fractions of 59 Fe-labelled blood were mobilised from brain and testis. The tissue in these two organs is known to have a blood barrier, so that the 59 Fe-labelled erythrocytes remain largely intravascular and can be easily washed out by PBS perfusion. Due to the wide sinus of the liver, 95% of residual blood content can be mobilised from hepatic tissue. In iron-deficient mice 14 days after 59 Fe injection, total hepatic 59 Fe radioactivity is 315 pmol 59 Fe/g wwt, while erythrocyte-bound 59 Fe in the liver corresponds to 196 pmol 59 Fe/g wwt (Table 2). The “perfusion method” reduces residual blood content to 5% of the original load, so that erythrocyte-bound 59 Fe radioactivity is as low as 9.8 pmol 59 Fe/g wwt. This corresponds to 3.1% of the 315 pmol of total 59 Fe-labelled iron content of the liver 14 days after injection (Table 5). In this case, the “perfusion method” provides a smaller margin of error than the “separation method” with its spillover rate of ∼20%. Corresponding figures for the margins of error estimated for the “perfusion method” in the other organs are given in Table 5. The “perfusion method”, though tedious, may be advantageous for the determination of endogenous 59 Fe tissue content, if the residual blood content can be reduced to less then 10% (Table 3 vs. Table 5). 5. Conclusion The “perfusion method” and the “subtraction method” will exclusively assess erythrocyte-bound 59 Felabelled haem iron, while the “separation method” will assess all 59 Fe-labelled tissue haem iron including that in haem-dependent tissue enzymes. The “separation method” and the “subtraction method” agree nicely in most experimental situations and both seem highly appropriate to determine 59 Fe distribution in genetically

B. Szegner et al. / Toxicology 244 (2008) 198–208

modified mouse models. The “separation method” and the “perfusion method” permit individual correction for 59 Fe in the residual tissue blood content which is highly advantageous when blood distribution is likely to be changed, such as in local inflammatory tissue hyperaemia. The “separation method” shows a spillover rate of ∼20% which can lead to corresponding errors, in particular in iron-deficiency and after long distribution intervals in organs with high residual blood content, e.g. in the spleen (Table 3). In this case, both methods can be combined, i.e. the mice are perfused to reduce residual tissue blood content before the separation method is applied. This reduces the margin of error in the spleen from 34% (“perfusion method”) and 20% (“separation method”) to ∼7% (i.e. 20% of 34%) (Tables 3 and 5). This permits to analyse distribution-related problems of iron toxicity also in mouse models of common chronic disease states such as inflammatory bowel diseases, rheumatism, diabetes, or adipositas. References Bothwell, T.H., Charlton, R.W., Cook, J.D., Finch, C.A., 1979. Iron metabolism in man. In: Internal Iron Kinetics. Blackwell Scientific Publications, Oxford, pp. 327–349 (Chapter 15). Dijkstra, C.D., Veerman, A.J.P., 1990. Spleen. Normal anatomy, histology, ultrastructure, rat. In: Jones, T.C., Ward, J.M., Mohr, U., Hunt, R.D. (Eds.), Hematopoietic System. Springer-Verlag, Berlin, pp. 185–193. Ehtechami, C., 1987. Zur Rolle von Ferritin bei der Eisenresorption. Tierexperimentelle Untersuchungen an Ratten mit verschiedenen Eisenderivaten. Dissertation der Fakult¨at f¨ur Chemie und Pharmazie der LMU, M¨unchen. Everett, N.B., Simmons, B., Lasher, E.P., 1956. Distribution of blood (Fe59 ) and plasma (I131 ) volumes of rats determined by liquid nitrogen freezing. Circ. Res. 4, 419–424. Fleming, M.D., Romano, M.A., Su, M.A., Garrick, L.M., Garrick, M.D., Andrews, N.C., 1998. Nramp2 is mutated in the anemic Belgrade (b) rat: evidence of a role for Nramp2 in endosomal iron transport. PNAS 95, 1148–1153. Frazer, D.M., Wilkins, S.J., Becker, E.M., Vulpe, C.D., McKie, A.T., Trinder, D., Anderson, G.J., 2002. Hepcidin expression inversely correlates with the expression of duodenal iron transporters and iron absorption in rats. Gastroenterology 123, 835–844. Galy, B., Ferring, D., Minana, B., Bell, O., Janser, H.G., Muckenthaler, M., Sch¨umann, K., Hentze, M.W., 2005. Altered body iron distribution and microcytosis in mice deficient in iron regulatory protein 2 (IRP2). Blood 106, 2580–2589. Grabill, C., Silva, A.C., Smith, S.S., Koretsky, A.P., Rouault, T.A., 2003. MRI detection of ferritin iron overload and associated neuronal pathology in iron regulatory protein-2 knockout mice. Brain Res. 971, 95–106. Hentze, M.W., Muckenthaler, M.U., Andrews, N.C., 2004. Balancing acts: molecular control of mammalian iron metabolism. Cell 117, 285–297. Hoeflich, A., Weber, M.M., Fisch, T., Nedbal, S., Fottner, C., Elmlinger, M.W., Wanke, R., Wolf, E., 2002. Insulin-like growth factor binding protein 2 (IGFBP-2) separates hypertrophic and hyperplastic

207

effects of growth hormone (GH)/IGF-I excess on adrenocortical cells in vivo. FASEB J. 16, 1721–1731. Hosain, F., Finch, C.A., 1966. A study of internal distribution of iron in man. Acta Med. Scand. Suppl. 445, 256–263. Huang, F.W., Pinkus, J.L., Pinkus, G.S., Fleming, M.D., Andrews, N.C., 2005. A mouse model of juvenile hemochromatosis. J. Clin. Invest. 115, 2079–2082. Hunder, G., Javdani, J., Elsenhans, B., Sch¨umann, K., 2000. Use of ␥spectrometry for simultaneous determination of 210 Pb, 73 As, 109 Cd 203 Hg and 59 Fe distribution and excretion in rats at low doses. Toxicology 150, 69–82. Jasinski, M., Pantazopoulos, P., Rother, R.P., van Rooijen, N., Song, W.C., Molina, H., Bessler, M., 2004. A novel mechanism of complement-independent clearance of red cells deficient in glycosyl-phosphatidylinositol-linked proteins. Blood 103, 2827–2834. Kennedy, M.L., Failla, M.L., Smith, J.C., 1986. Influence of genetic obesity on tissue concentrations of zinc, copper, manganese and iron in mice. J. Nutr. 116, 1432–1441. Kontoyiannis, D., Boulougouris, G., Manoloukos, M., Armaka, M., Apostolaki, M., Pizarro, T., Kotlyarov, A., Forster, I., Flavell, R., Gaestel, M., Tsichlis, P., Cominelli, F., Kollias, G., 2002. Genetic dissection of the cellular pathways and signaling mechanisms in modeled tumour necrosis factor-induced Crohn’s-like inflammatory bowel disease. J. Exp. Med. 196, 1563– 1574. K¨uhn, L.C., Richman, L., Darshan, D., 2007. The iron storage function of ferritin H protects cells from iron toxicity by oxidative damage. In: Proceedings of the Second Congress of the International BioIron Society, Kyoto, pp. 0–23. Kuhn, R., Lohler, J., Rennick, D., Rajewsky, K., Muller, W., 1993. Interleukin-10-deficient mice develop chronic enterocolitis. Cell 75, 263–274. Levy, J.E., Jin, O., Fujiwara, Y., Kuo, F., Andrews, N.C., 1999. Transferrin receptor is necessary for development of erythrocytes and the nervous system. Nat. Genet. 21, 396–399. Lynch, S.R., 1984. Iron. In: Solomons, N.W., Rosenberg, U.H. (Eds.), Absorption and Malabsorption of Mineral Nutrients. Alan R. Liss Inc., New York, pp. 89–124. Nicolas, G., Chauvet, C., Viatte, L., Danan, J.L., Bigard, X., Devaux, I., Beaumont, C., Kahn, A., Vaulont, S., 2002. The gene encoding the iron regulatory peptide hepcidin is regulated by anemia, hypoxia, and inflammation. J. Clin. Invest. 110, 1037–1044. Noyes, W.D., Bothwell, T.H., Finch, C.A., 1960. The role of the reticulo-endothelial cell in iron metabolism. Br. J. Haematol. 6, 43–55. Sch¨umann, K., Elsenhans, B., Hunder, G., Strugala, G., Forth, W., 1989. Increase of the intestinal iron absorption in growing rats and mice after 8 days of iron-deficient feeding. Z. Versuchstierkd. 32, 261–267. Sch¨umann, K., 2001. Safety aspects of iron in food. Ann. Nutr. Metabol. 45, 91–101. Sch¨umann, K., Ettle, T., Szegner, B., Elsenhans, B., Solomons, N.W., 2007a. On risk and benefits of iron supplementation—recommendations for iron intake revisited. J. Trace Elem. Med. Biol. 21, 147–168. Sch¨umann, K., Szegner, B., Kohler, B., Pfaffl, M.W., Ettle, T., 2007b. A method to assess 59 Fe in residual tissue blood content in mice and its use to correct 59 Fe-distribution kinetics accordingly. Toxicology 241, 19–32. Sch¨umann, K., Solomons, N.W., 2007. Safety of interventions to reduce nutritional anemias. In: Kraemer, K., Zimmermann, M.B.

208

B. Szegner et al. / Toxicology 244 (2008) 198–208

(Eds.), Nutritional Anemia. Sight and Life Press, Basel, pp. 285– 314 (Chapter 18). Spasic, M.V., Kiss, J., Herrman, T., Kessler, R., Stolte, J., Galy, B., Rathkolb, B., Wolf, E., Stremmel, W., Hentze, M.W., Muckenthaler, M.U., 2007. Physiologic systemic iron metabolism in mice deficient for duodenal Hfe. Blood 109, 4511– 4517. Srinivasan, K., Ramarao, P., 2007. Animal models in type 2 diabetes research: an overview. Indian J. Med. Res. 125, 451– 472. Teale, F.W., 1959. Cleavage of the haem-protein link by acid methylethylketone. Biochim. Biophys. Acta 35, 543.

Torrance, J.D., Bothwell, T.H., 1968. A simple technique for measuring storage iron concentrations in formalinised liver samples. S. Afr. J. Med. Sci. 33, 9–11. Trenor 3rd., C.C., Campagna, D.R., Sellers, V.M., Andrews, N.C., Fleming, M.D., 2000. The molecular defect in hypotransferrinemic mice. Blood 96, 1113–1118. Vulpe, C.D., Kuo, Y.M., Murphy, T.L., Cowley, L., Askwith, C., Libina, N., Gitschier, J., Anderson, G.J., 1999. Hephaestin, a ceruloplasmin homologue implicated in intestinal iron transport, is defective in the sla mouse. Nat. Genet. 21, 195–199. Weiss, G., Goodnough, L.T., 2005. Anemia of chronic disease. N. Engl. J. Med. 352, 1011–1023.